Differential effects of acute versus chronic dietary fructose consumption on metabolic responses in FVB/N mice

Abstract

Increased human consumption of high-fructose corn syrup has been linked to the marked increase in obesity and metabolic syndrome. Previous studies on the rapid effects of a high-fructose diet in mice have largely been confined to the C57BL/6 strains. In the current study, the FVB/N strain of mice that are resistant to diet-induced weight gain were used and fed a control or high-fructose diet for 48 h or for 12 wk. Many of the previously reported changes that occurred upon high-fructose feeding for 48 h in C57BL/6 mice were recapitulated in the FVB/N mice. However, the acute increases in fructolytic and lipogenic gene expression were completely lost during the 12-wk dietary intervention protocol. Furthermore, there was no significant weight gain in FVB/N mice fed a high-fructose diet for 12 wk, despite an overall increase in caloric consumption and an increase in average epididymal adipocyte cell size. These findings may be in part explained by a commensurate increase in energy expenditure and in carbohydrate utilization in high-fructose-fed animals. Overall, these findings demonstrate that FVB/N mice are a suitable model for the study of the effects of dietary intervention on metabolic and molecular parameters. Furthermore, the rapid changes in hepatic gene expression that have been widely reported were not sustained over a longer time course. Compensatory changes in energy expenditure and utilization may be in part responsible for the differences obtained between acute and chronic high-fructose feeding protocols.

INTRODUCTION

The introduction and popularization of the western diet into society has long been correlated with increases in obesity and many metabolic diseases (1). Increasingly, home-cooked meals have been replaced by calorically dense foods, which are high in both fat and sugar content but may lack other desired nutrients (2). Although found naturally in fruits and honey, fructose, especially high-fructose corn syrup, has been used increasingly in the past 40 years as an artificial sweetener, paralleling the rise in obesity in the past 40 years (3). Although both fructose and glucose are six-carbon sugar molecules, differences in their functional groups and metabolizing enzymes differentiate the absorption and utilization of these sugars. For reasons that are only partially understood, significant consumption of fructose has been linked to increased rates of metabolic syndrome, type 2 diabetes mellitus, and many corresponding comorbidities (4–6).

Fructose consumed in low concentrations is easily digested in the small intestines into glucose and other intermediates, which are easily catabolized through the normal glycolytic pathway once it reaches the liver (7, 8). However, in higher concentrations much of the fructose is metabolized by the liver. In the liver, fructose is rapidly phosphorylated in a completely unregulated process by ketohexokinase (KHK) (9, 10). The intermediates of fructose catabolism can be shunted toward gluconeogenesis to replenish glycogen stores, but remaining carbon intermediates are utilized for de novo lipogenesis and for the production of fatty acids (11, 12). In sedentary individuals this process can lead to excessive lipid accumulation, both in the liver and in peripheral tissues, resulting in obesity and insulin resistance (13).

Animal models have been used for a more thorough analysis of how metabolism is altered after inclusion of high-fructose corn syrup (usually 55%–60% fructose and 40%–45% glucose) in the diet. Because of the soluble nature of sugars, including fructose, dietary interventions involving high-fructose feeding can be implemented both through solid food or sugar-supplemented drinking water (7, 14). Although fructose supplementation through drinking water appears to be more commonly utilized in recent studies, similar results were obtained between 4 wk of fructose consumption in food or drinking water on hepatic steatosis in C57BL/6 male mice (15). In this study, dietary fructose was incorporated in a pelleted chow and all components were matched except for the fructose versus corn starch in the control diet.

Over the past decades, the C57BL/6 mouse strains have been widely utilized in a large variety of metabolic studies. For example, the C57BL/6 background is highly susceptible to diet-induced obesity that is highly advantageous for many studies. However, the onset of obesity and associated metabolic perturbations can also cloud analysis of specific metabolic changes arising from the dietary intervention (16, 17). For this reason and others, the study of multiple mouse strains is essential when studying systemic and tissue-specific disturbances caused by specific diets.

Recent studies have begun to shed light on how different genetic backgrounds of host mice can alter metabolic responses and gene expression through large-scale, unbiased approaches (18, 19). While almost all agree that C57BL/6 mice are prone to developing diet-induced obesity on both high-fat and high-fructose diets, the effects of specific diets in FVB/N mice appear more nuanced and understudied (20, 21). FVB/N mice are known to consume greater amounts of food when exposed to ad libitum feeding conditions and appear to gain weight on a high-fat diet (HfrD); however, most studies have not shown significant weight gain for FVB/N mice when fed a high-fructose diet. In addition to differential metabolic effects including weight gain in various genetic backgrounds, the sex of the animal is known to impact many metabolic parameters in mice and rats, including fructose-induced hepatic and renal gene expression and/or liver steatosis (14, 22, 23). Studies including female animal models are more limited, and only have become a focus in recent years.

In this present study, female FVB/N mice were used to interrogate both the short- and long-term metabolic consequences of high-fructose feeding. FVB/N mice are known to be resistant to the development of fructose-induced obesity (21), which allows for a direct and thorough investigation into the metabolic consequences of high-fructose feeding without the confounding effects of obesity. The current work has explored the effects of acute and chronic high-fructose consumption on FVB/N mice. Most recent published reports in C57BL/6 mouse lines have focused on short-term consequences of high-fructose feeding in male and female mice (8, 14, 15, 23). However, the negative health outcomes associated with high-fructose feeding in humans are seen after months or years. In the short-term arm of this study, we aimed to validate the use of FVB/N female mice and confirm previous results of short-term high-fructose feeding without significant weight gain in Sprague–Dawley male rats (24). Furthermore, for the long-term arm of this study, we hypothesize that there would be significant differences to fructose-related metabolic pathways after 12 wk on diet. These potential differences may be critical to understanding the progression of metabolic disease associated with high-fructose feeding.

MATERIALS AND METHODS

Animals

Female virgin FVB/N mice were used in all experiments. Mice were initially purchased from Jackson laboratories and then a breeding colony was maintained on site. This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. All animal work was approved by the University of Chicago Institutional Animal Care and Use Committee. At 4 wk of age, FVB/N mice were separated and randomly assigned to either a control-diet group or a high-fructose-diet group. These diets were purchased from Harlan Lab (Madison, WI). The high-fructose diet was custom-made to match the control group in all aspects except for the presence of 100 g/kg fructose that was replaced by corn starch (a polymer of glucose found in corn that is slowly digested to glucose) in the control diet (Table 1). FVB/N mice were provided the high-fructose diet or the matched control diet for either 48 h or 12 wk. The 48-h study began in mice at 15 wk of age, whereas the 12-wk study began immediately after weaning (at 4 wk of age). Following both studies, mice were euthanized by isoflurane narcosis and cervical dislocation at 16 wk of age.

Table 1.

Dietary information for control diet and matched high-fructose diet

	Formula, g/kg
	Control Diet (AIN-93G Purified Diet) TD.94045	High-Fructose Diet (60% kcal) TD.140022
Sucrose	100.0	100.0
Fructose	0.0	529.386
Corn starch	397.486	0.0
Maltodextrin	132.0	0.0
Casein	200.0	200.0
l-Cystine	3.0	3.0
Soybean oil	70.0	70.0
Cellulose	50.0	50.0
Mineral mix (94046)	35.0	35.0
Vitamin mix (94047)	10.0	10.0
Choline bitartrate	2.5	2.5
TBHQ antioxidant	0.014	0.014
Green food color	0.0	0.1

The only significant difference to the high-fructose diet compared with the control diet is the presence of fructose, which is replaced by corn starch and maltodextrin in the control diet.

Metabolic Rate and Activity

Metabolic cages (LabMaster by TSE Systems, Chesterfield, MO) were used on a smaller cohort of female FVB/N mice at 16 wk of age to measure and directly compare metabolic differences between control (n = 4) and fructose-fed mice after 12 wk (n = 4) on their respective diets. Mice were individually housed in the metabolic chambers while they had ad libitum access to food and water and were acclimated to the cages for 3–4 days before measurements were collected over the next two days. Circadian activity (energy expenditure and activity) and feeding behavior of the mice were measured by calculating food and water consumption, movement, and fuel utilization by measuring oxygen consumption and carbon dioxide production every 30 min. From these results, the average respiratory exchange ratio (RER) was calculated as a ratio of CO₂ produced versus O₂ consumed. Energy expenditure was further calculated as caloric value (CV) multiplied by O₂ consumed, where CV is equal to 3.815 + 1.232 × RER. After 5 days of acclimatization, measurements were taken every 30 min for 2 days.

Dual-Energy X-Ray Absorptiometry

Total body fat percentage was noninvasively measured by dual-energy X-ray absorptiometry (DEXA). Mice were anesthetized before DEXA imaging via injection with a mixture of ketamine (80 mg/kg body wt) and xylazine (5 mg/kg body wt) solution. Body composition was assessed by DEXA (Lunar PIXImus densitometer system, GE Healthcare) using PIXImus 2 software. These experiments were completed immediately before the metabolic cage testing, on the same cohorts of mice, at 16 wk of age.

Glucose Tolerance Tests

Glucose tolerance tests (GTTs) were conducted on the FVB/N mice after 8 and 12 wk on both the control (n = 4) and high-fructose diet (n = 4). After an overnight fast, the GTTs were conducted via an intraperitoneal injection of 20% dextrose at a volume of 10 μL/g body wt. Blood glucose was measured preinjection, postinjection, and at 15-, 30-, 60-, and 120-min postinjection. Blood, and subsequent serum via ultracentrifugation, was also collected from the tail vain for insulin measurements via the ALPCO mouse ultrasensitive insulin ELISA (80-INSMSU-E01).

Blood Serum Collection and Metabolite Measurement

Upon animal euthanasia, blood was collected via cardiac puncture. After blood was allowed to clot for 30 min, serum was separated by centrifugation for 10 min at 2,000 g. Samples were stored at −80°C until experimentation. Serum concentration of fibroblast growth factor-21 (FGF-21) was measured with Abcam’s fluorescent mouse FGF-21 ELISA kit (ab229382) from the stored samples.

Histology

As part of a larger study on mammary cancer, after euthanasia, mammary gland white adipose tissue (WAT) from each mouse was placed in a histology cassette, formalin fixed, and paraffin embedded. During pathological sectioning, tissue was removed until a full face of tissue was found, and serial sections were taken to be placed on histological slides for subsequent analysis.

Adipocyte Sizing

Average mammary gland adipocyte size was measured in female FVB/N mice using an immunofluorescent perilipin antibody to estimate adipocyte size between control (n = 4) and fructose-fed cohorts (n = 4). Perilipin is a protein that surrounds the lipid droplet, and lipid droplet size is a close estimation of adipocyte size (25). For perilipin immunohistochemistry, five serial cut technical replicates of 5-mm sections of WAT were cut and adhered to positively charged glass slides, dewaxed in xylene, and hydrated using graded ethanol washes. For heat-induced antigen retrieval, slides were placed within a steamer for 30 min in a citrate bath (EMS 10× citrate buffer, pH 6.0, Cat. No. 64142-08). Immunostaining was performed using a 1:100 dilution of anti-perilipin antibody (Abcam, ab3526) with a donkey anti-rabbit secondary antibody conjugated to DyLight 594 (Abcam, ab96893). After images are taken using a FV1000 confocal microscope (Olympus), Cell Profiler imaging software was utilized to calculated average adipocyte size. Each biological replicate consisted of five technical replicates from sequential serial cut sections of adipose tissue.

Adipokine Array

Conditioned media (CM) made from mammary fat (MF) was separated from all the mammary glands of each mouse. Conditioned media is produced by placing enriched adipose tissue in media (SH30240.01, Hyclone), containing 1% BSA and 1% penicillin-streptomycin (P/S), at 10 mL media per gram tissue and incubated at 37°C for 8 h (26). The media is subsequently filtered and stored at −80°C until subsequent analysis. Three mouse samples per diet group were thawed at 4°C and analyzed as described in instructions for the adipokine antibody array kit (ARY013; R&D Systems). Each membrane was incubated on a rocker with 250 μL of MF CM, 125 μL of array buffer 4, and 1.125 mL of array 6 overnight at 4°C. Membranes were then imaged after incubation with the corresponding secondary antibody, as well as with streptavidin-horseradish peroxidase and chemiluminescent detection reagents provided. Adipokine levels in the MF-CM were determined with Bio-Rad imager using Quantity One software to image blots and determine densitometry of each antibody spot. Adipokine antibodies were captured in duplicate and the average for each duplicate pair minus the densitometry value of the surrounding background was used to determine the peak average signal density of each antibody.

Quantitative Real-Time Polymerase Chain Reaction

For the analysis of hepatic gene expression, the liver was immediately frozen on dry ice after the animal was euthanized and stored at −80°C until RNA was extracted. The liver was homogenized utilizing the Next Advance Bullet Blender tissue homogenizer and the corresponding Green Rino RNA lysis kit. After tissue homogenization, RNA was extracted using the Omega E.Z.N.A Total RNA Kit II (R6934-01). After RNA extraction, iScript Reverse Transcription Supermix for RT-qPCR (BIO RAD, 1708840) was used to produce cDNA. Gene expression for genes of interest was measured with quantitative real-time polymerase chain reaction (qRT-PCR). 18S Ribosomal RNA was used as a house-keeping gene throughout all measured genes.

Data Analysis

Heteroskedastic two-tailed t tests were used to analyze the differences between control and fructose-fed FVB/N mice.

RESULTS

Acute High-Fructose Feeding in FVB/N Mice

Many recent investigations into high-fructose feedings have focused on short-term metabolic effects in C57BL/6 mice. To verify similar outcomes in FVB/N mice, in the current study female FVB/N mice, at 15 wk of age, were fed a high-fructose (n = 4) or matched control diet (n = 4) for 48 h to compare results with previous studies. Hepatic expression of fructolytic, adipogenic, and fibroblast growth factor-21 (FGF-21) related genes were all assessed. Results from previous studies have overwhelmingly shown an increase in all three of these pathways upon acute feeding of a high-fructose diet. Here, these results were compared with FVB/N mice.

After 48 h on diet, there was a significant and expected increase for all measured hepatic fructolytic genes in the fructose-fed mice compared with the control cohort (Fig. 1). In the fructose-fed FVB/N mice, compared with the control-fed animals, there was a 2.51-fold increase of KHK, the enzyme which initially phosphorylates fructose in the liver into fructose-1-phosphate (P = 0.02). For aldolase B (AldoB), which cleaves fructose-1-phosphate into glyceraldehyde and dihydroxyacetone-phosphate, there was a 3.58-fold increase in the fructose-fed FVB/N mice after 48 h compared with the control-fed animals (P = 0.0009). In the expression of triokinase/FMN cyclase (TKFC), the enzyme that controls the phosphorylation of glyceraldehyde into glyceraldehyde-3-phosphate, there was a 2.14-fold increase for the fructose-fed cohort when compared with the control mice (P = 0.04). In this pathway of fructose metabolism, glyceraldehyde-3-phosphate is shunted toward pyruvate (under enzymatic control from pyruvate kinase), which can be converted into lactate via lactate dehydrogenase (LDHA). Pyruvate kinase (PKL) was also increased 2.92-fold in the fructose-fed cohort compared with the control animals (P = 0.005). Finally, LDHA was also increased in fructose-fed animals, with a 2.29-fold increase in expression compared with the control-fed mice (P = 0.007).

Figure 1.

Effects of short- and long-term high-fructose feeding on hepatic fructolytic gene expression. Major fructolysis genes were analyzed via quantitative real time polymerase chain reaction (qRT-PCR) after 48 h (control n = 4, high fructose n = 4) and 12 wk (control n = 15, high fructose n = 11) on diet in FVB/N mouse liver. Fold-change in gene expression of fructose-fed mice compared with control-fed mice was calculated. After 48 h, there is a significant increase in all measured fructolysis-related genes in high-fructose-fed animals compared with controls [ketohexokinase (KHK) *P = 0.02, aldolase B (AldoB) ***P = 0.0009, triokinase/FMN cyclase (TKFC) *P = 0.04, lactate dehydrogenase (LDHA) **P = 0.007, pyruvate kinase (PKL) **P = 0.005]. There is no significant difference in any measured genes between control and fructose-fed animals after 12 wk on diet.

There was also a significant increase for all measured hepatic lipogenic genes in the fructose-fed FVB/N mice compared with the control cohort after 48 h (Fig. 2). Peroxisome proliferator-activated receptor γ (PPARγ), which regulates the storage of fatty acids in adipose tissue, was increased 2.18-fold compared with control-fed mice (P = 0.02). The expression of acetyl-CoA carboxylase (ACC), which catalyzes the conversion of acetyl-CoA into malonyl-CoA (thereby initiating de novo lipogenesis), was increased 3.51-fold in the fructose-fed cohort compared with controls (P = 0.03). Furthermore, ATP citrate lyase (ACLY), which catalyzes the first step of endogenous fatty acid synthesis, was increased 4.74-fold in the fructose-fed mice when compared with their control-fed counterparts (P = 0.006). Expression of acyl-CoA synthetase short-chain family member 2 (ACSS2), which controls the formation of acetyl-CoA from acetate, was increased 4.87-fold in fructose-fed animals compared with the control-fed mice (P = 0.002). Finally, fatty acid synthase (FASN), a multi-enzyme protein that is critical for fatty acid synthesis, was also increased 2.99-fold in fructose-fed FVB/N when compared with the control-fed mice (P = 0.03).

Figure 2.

Effects of short- and long-term high-fructose feeding on hepatic lipogenic gene expression. Lipogenesis-related genes were analyzed via quantitative real time polymerase chain reaction (qRT-PCR) in livers after FVB/N mice were fed diets for 48 h (control n = 4, high-fructose n = 4) and 12 wk (control n = 15, high-fructose n = 11). Fold-change in gene expression of fructose-fed mice compared with control-fed mice was calculated. After 48 h there is a significant increase in all measured hepatic lipogenesis-related genes in fructose-fed animals compared with controls [peroxisome proliferator-activated receptor γ (PPARγ) *P = 0.02, acetyl-CoA carboxylase (ACC) *P = 0.03, ATP citrate lyase (ACLY) **P = 0.005, acyl-CoA synthetase short chain family member 2 (ACSS2) **P = 0.002, fatty acid synthase (FASN) *P = 0.03]. There is no significant difference in any measured genes between control and fructose-fed animals after 12 wk on diet.

Expression of FGF-21-related genes was also significantly upregulated in the livers of fructose-fed cohort after 48 h on diet (Fig. 3). The FVB/N mice fed a high-fructose diet expressed a 5.92-fold increase in FGF-21 (P = 0.006), a 1.70-fold increase in its co-receptor β-klotho (P = 0.02), and a 1.40-fold in its established receptor fibroblast growth factor receptor 1 (FGFr1) (P = 006), when compared with the control-fed cohorts. In parallel, circulating levels of FGF-21 were also measured (Fig. 4). The relative difference in serum levels of FGF-21 between fructose and control-fed FVB/N mice corresponded with the hepatic gene expression of FGF-21. After 48 h of high-fructose feeding in FVB/N mice, there is a significant increase in the concentration of circulating FGF-21 when compared with the control-fed cohort, and this difference correlates with the hepatic gene expression in the same study. After 48 h on diet, the control-fed FVB/N mice had an average FGF-21 blood serum concentration of 418.8 ± 22.7 pg/mL, whereas the fructose-fed mice had an average concentration of 596.0 ± 82.3 pg/mL (P = 0.04). These results, while previously reported in C57Bl/6 mice, have not been widely confirmed in FBV/N and further served as baseline measures for longer-term studies on the effects of high-fructose feeding.

Figure 3.

Effects of short- and long-term high-fructose feeding on hepatic fibroblast growth factor-21 (FGF-21) gene expression. Hepatic levels of FGF-21-related genes were analyzed via quantitative real time polymerase chain reaction (qRT-PCR) after 48 h (control n = 4, high fructose n = 4) and 12 wk (control n = 15, high fructose n = 11) on diet in both control and fructose-fed FVB/N mice. Fold-change in gene expression of fructose-fed mice compared with control fed mice was calculated. After 48 h there is a significant increase in all measured FGF-21-related genes in fructose-fed mice compared with controls [FGF-21 **P = 0.006, β-Klotho *P = 0.02, fibroblast growth factor receptor 1 (FGFr1) **P = 0.007]. There is no significant difference in any measured genes between control and fructose-fed animals after 12 wk on diet.

Figure 4.

Effects of short- and long-term high-fructose feeding on serum fibroblast growth factor-21 (FGF-21) levels. Serum FGF-21 was measured via ELISA. After 48 h on diet, fructose-fed (n = 8) mice had a significantly higher serum concentration of FGF-21 when compared with control-fed (n = 8) animals (*P = 0.04). After 12 wk on diet, high-fructose-fed FVB/N mice (n = 26) had significantly lower serum concentration of FGF-21 when compared with control-fed mice (n = 26), (*P = 0.03).

Sustained High-Fructose Feeding in FVB/N Mice

Next, the long-term metabolic effects of high-fructose feeding in FVB/N mice were assessed. After weaning at 4 wk of age, female mice were randomly assigned to either the control (n = 7) or fructose-fed group (n = 8). After 12 wk of feeding, there is no significant difference in all measured fructolytic genes (KHK, AldoB, TKFC, LDHA, and PKL) between the control and fructose-fed cohorts (P = >0.05 for all relationships). There was also no significant difference in any of the adipogenic genes (PPARγ, ACC, ACLY, ACSS2, and FASN) between the two dietary conditions (P = >0.05 for all relationships). Furthermore, after long-term feeding there was no significant change in any of the FGF-21-related genes (FGF-21, β-klotho, and FGFr1) measured between the high-fructose and control-fed cohorts (P = >0.05 for all relationships). These measures of hepatic gene expression after 12 wk on diet stand in sharp contrast to the significant changes seen after only 48 h on diet (Figs. 1, 2, and 3).

Despite no difference in hepatic expression of FGF-21 after 12 wk of high-fructose feeding, there is a significant decrease in the concentration of circulating FGF-21 when compared with the control-fed FVB/N mice (Fig. 4). The control-fed FVB/N mice had an average FGF-21 blood serum concentration of 738.2 ± 127.3 pg/mL, whereas the fructose-fed mice had an average concentration of 434.4 ± 51.2 pg/mL (P = 0.03), suggesting potential downregulation of hepatic FGF-21 production over the longer time course.

Systemic Measures

To further understand differences in hepatic gene expression between short- and long-term high-fructose feeding, systemic measures of body composition and energy expenditure were analyzed in a cohort of female FVB/N mice (n = 4 Control, 4 HFrD) during 12 wk on diet. This approach allowed for simultaneous measures of all animals using metabolic cages. Body weight was monitored weekly from weaning to time of euthanasia, at 16 wk of age (Fig. 5). No significant differences in body weight were measured over this period. Furthermore, after 12 wk on each diet, the female FVB/N mice underwent DEXA analysis for body composition analyses (n = 4 Control, 4 HFrD) (Fig. 6). These mice fed a high-fructose diet had a comparable weight to the mice fed the matched control diet after 12 wk on diet (average body weight of 21.6 ± 0.59 g for control-fed mice compared with 21.6 ± 0.81 g for the fructose-fed mice, P = 0.98). Similarly, there is no significant difference in lean body mass between the fructose-fed and control-fed FVB/N mice after 12 wk on diet (average lean body mass of 18.2 ± 0.48 g for control-fed mice compared with 18.6 ± 0.68 g for their fructose-fed counterparts, P = 0.65). There is also no significant difference in body fat percentage between the two groups (average body fat percentage of 15.6 ± 1.04% for the control-fed FVB/N mice vs. 13.7 ± 0.65% for the fructose-fed mice, P = 0.17).

Figure 5.

Effects of high-fructose feeding on FVB/N body weight. Mouse weight was recorded weekly from weaning to time of euthanasia, at 16 wk of age/12 wk on diet. There are no significant differences between control (n = 8) and high-fructose-fed (n = 8) FVB/N mice.

Figure 6.

Effects of long-term high-fructose feeding on FVB/N body composition. Body composition of control (n = 4) and fructose-fed (n = 4) FVB/N mice after 12 wk on diet. There is no significant difference between cohorts in weight (P = 0.17), lean body mass (P = 0.98), or percentage body fat (P = 0.65).

After both 8 and 12 wk on diet, GTTs were performed on the female FVB/N mice (n = 4 Control, 4 HFrD) (Fig. 7). After an overnight fast, an intraperitoneal injection of dextrose was performed, and blood glucose was monitored. No significant difference in glucose tolerance was measured between control- and fructose-fed animals at either timepoint. There was also no change in serum insulin levels (data not shown).

Figure 7.

Effects of long-term high-fructose feeding on glucose tolerance. A glucose tolerance test was performed at both 8 and 12 wk on diet for control (n = 4) and high-fructose-fed (n = 4) FVB/N mice. There is no significant difference between groups as measured by area under the curve.

After 12 wk of feeding, metabolic cages were used to discern more possible differences between the fructose- and control-fed female mice (n = 4 Control, 4 HFrD). Data from the metabolic cages allowed the analysis of food and water consumption, ambulatory movements, fuel utilization via RER, and energy expenditure. Food and water consumption were measured over a 12-h active (dark) period and found that fructose-fed animals consumed significantly more food and water than their control-fed counterparts (Fig. 8). Control-fed mice consumed an average of 1.1 ± 0.07 g of food, whereas the fructose-fed mice consumed an average of 1.6 ± 0.03 g of food over a 12-h active period for the animals (P = 0.0001). Furthermore, control-fed mice drank an average of 0.87 ± 0.06 mL of H₂O over 12 h, but the fructose-fed mice drank an average of 1.1 ± 0.07 mL of H₂O (P = 0.02).

Figure 8.

Effects of 12 wk of high-fructose feeding on food and water consumption. Total consumption of both food and water over a 12-h period for control (n = 4) and high-fructose-fed (n = 4) FVB/N mice. There is a significant increase in both food (****P = 0.0001) and water (*P = 0.02) consumption for fructose-fed mice compared with mice on control diet.

Fructose-fed animals also had a significant increase in ambulatory movements along both the x- and z axes measured in parallel to food and water consumption during the dark period, had a corresponding significant increase in energy expenditure per mouse (Fig. 9). Fructose-fed mice had greater movement along the x-axis (107,509 counts vs. 42,089 counts, P = 4.14e-8). Fructose-fed mice also had greater movement along the z-axis (19,163 counts vs. 8,309 counts, P = 2.98e-9). In congruence with the increase in movement, fructose-fed animals also had greater energy expenditure when compared with the control-fed mice (0.43 kcal/h vs. 0.37 kcal/h, P = 2.55e-9).

Figure 9.

Effects of 12 wk of high-fructose feeding on activity measures. Total movement on the X and Z axes, energy expenditure per mouse, and respiratory exchange ratio was measured via metabolic cages while comparing FVB/N mice fed for 12 wk with either a control (n = 4) or high-fructose diet (n = 4). The high-fructose diet fed cohort had a significant increase in movement along both the x (****P = 4.1e⁻⁸) and z axes (****P = 3.0e⁻⁹) when compared with the control-diet-fed cohort. High-fructose-diet fed animals also have a significant increase in energy expenditure (****P = 2.6e⁻⁹) and respiratory exchange ratio (RER) (**P = 0.008) when compared with control-diet-fed mice.

Finally, the respiratory exchange ratio (RER) was calculated as the ratio of CO₂ production and O₂ consumption every 30 min (Fig. 9). A RER closer to 0.7 indicates that the predominant fuel source is lipid, whereas a RER closer to 1.0 is representative of carbohydrate being the predominant fuel source. Control-fed FVB/N mice had an average RER of 0.86 ± 0.008, whereas the fructose-fed cohort had an average RER of 0.89 ± 0.007 (P = 0.008), indicating preferential carbohydrate oxidation in the high-fructose-fed animals.

Adipocyte Measures

Upon euthanasia of the mice, after 12 wk on each diet and at 16 wk of age, epididymal WAT was fixed, paraffin embedded, and analyzed by histological staining. Adipocyte cell size was estimated using antibodies against perilipin, a lipid droplet-coated protein that is used to estimate primary adipocyte cell size. Despite no measurable difference in body weight between these two groups, fructose-fed mice had a significant increase in adipocyte size when compared with the control-fed mice after 12 wk of feeding (Fig. 10). The control-fed mice had an average adipocyte size of 5,260.5 ± 524.6 μm³, whereas fructose-fed mice had an average area of 8,575.2 ± 682.5 μm³ (P = 0.0005).

Figure 10.

Effects of 12 wk of high-fructose feeding on mammary gland adipocyte cell size. Average adipocyte area was measured for control (n = 4) and high-fructose-diet fed animals (n = 4). Each biological replicate consists of five technical replicates. Using confocal microscopy (×40), anti-perilipin immunohistochemistry staining was used to outline lipid droplets on slides of formalin fixed, paraffin embedded white adipose tissue from FVB/N mice after 12 wk on each diet. Cell Profiler image analysis software was used to calculate average adipocyte size. Mammary gland adipocytes from high-fructose-diet fed mice are significantly larger than adipocytes from control-diet-fed mice (***P = 0.0005). Scale bar = 50 µM.

The increase in average adipocyte size occurred despite no detectable changes in weight or adiposity. To better understand this result and potential alterations in adipocyte function, an adipokine array was conducted on the secreted factors from isolated adipocytes. After adipocyte isolation, mammary adipose tissue was cultured in serum-free media for 8 h, to collect adipocyte-derived factors secreted into the media (termed secretome). Using adipokine arrays with a span of antibodies already bound to the membrane, we investigated the change to the overall secretome of adipose tissue from a high-fructose diet compared with the secretome from the control-diet-fed mice (Fig. 11). Significant changes were seen in many measured adipokines, but the largest changes between high-fructose and control-fed mice were seen in the abundance of FGF-21 in the conditioned media.

Figure 11.

Effects of 12 wk of high-fructose feeding on adipokine secretion. Mammary gland adipose tissue conditioned media from FVB/N mice fed control or high-fructose diet for 12 wk (n = 3 for each diet group) was analyzed using an adipokine array to evaluate changes to the adipose secretome between dietary interventions.

DISCUSSION

In the past half-century, the consumption of fructose, mainly through high-fructose corn syrup, has increased drastically, paralleling the rise in obesity rates seen throughout the world (3). Through mechanisms not completely understood, the metabolism of fructose increases de novo lipogenesis and the production of fatty acids causing lipid accumulation, resulting in obesity and insulin resistance (13). Although this metabolic phenotype occurs in humans after many years of heightened fructose consumption, most studies in animal models have primarily focused on the short-term consequences of high-fructose feeding. In the current study, the effects of acute and more chronic feeding of a high-fructose diet were investigated at the systemic and cellular levels in FVB/N mice.

In these previous short-term, high-fructose feeding studies, fructolytic and adipogenic genes, as well as FGF-21 production have been shown to be significantly upregulated in the widely used C57BL/6 mouse model (24, 27). In one of the first investigations of these pathways in FVB/N mice, this study found a similar upregulation of canonical fructolytic, adipogenic, and FGF-21-related genes. In contrast to the short-term consequences of high-fructose feeding in mice, the long-term effects of a high-fructose diet are not well established. As the negative health effects of fructose consumption seen in humans occurs over months, years, and decades, this study also began to investigate the long-term impact of high-fructose feeding. For the first time, the differences between short-term and long-term consumption of fructose were assessed in FVB/N mice. Importantly, this mouse model allowed for the assessment of metabolic changes resulting from extended high-fructose feeding without the confounding onset of obesity and insulin-resistant that exudes metabolic differences by themselves.

Interestingly, this study found that the widely reported increases in fructolytic, adipogenic, and FGF-21-related genes in the liver upon short-term, high-fructose feeding were lost upon continued fructose feeding over a 12-wk period. In the long-term, 12-wk assessment of high-fructose feeding on FVB/N mice, this study found that there are no longer significant increases in these genes. One interpretation of this is that over such an extended period of high-fructose feeding, the small intestine adapts to compensate for such high levels of fructose. It is well established, at least at low levels of fructose consumption, that the small intestine is able to catabolize the consumed fructose into glucose, acetate, and other byproducts that are easily metabolized by the liver and protects the liver from the deleterious effects of hepatic fructolysis (7, 8).

Moreover, it has also been demonstrated that the gut can adapt over time to metabolize more fructose if fructose is consistently consumed (7, 15). In addition to the upregulation of fructolytic enzymes in the gut, the changing microbiome can impact gut metabolism as well (28). Recently, it has been firmly demonstrated that a fructose-enriched diet can alter the microbiome, which potentially alters the amount of fructose metabolized by the liver (29–31). As this current study did not investigate the gut metabolism of fructose, more work is required to determine if it can compensate to catabolize a 60% kilocalorie high-fructose diet.

As the small intestine has never been reported to metabolize greater than 1 g/kg fructose, another method of blunted fructolysis must occur to instigate the downregulation of hepatic fructolytic, adipogenic, and FGF-21-related genes (32). As previously noted, the hepatic metabolism of fructose is an expensive energetic process that requires significant use of ATP, as well as other biological cofactors, to complete the catabolic process. For example, the initial phosphorylation step of fructolysis, catalyzed by KHK, requires ATP and it has been reported that fructose-1-phosphate often builds in hepatocytes after high-fructose feeding, and ATP stores can begin to become depleted (33, 34). If cellular stores of ATP, as well as other biological cofactors continue to become depleted, it is not unrealistic to hypothesize that the normal hepatic fructolysis would not be able to continue. Although not considered the canonical fructolytic pathway, hepatic fructose can be converted directly into fructose-6-phosphate by hexokinase and enter higher into the glycolytic pathway. Furthermore, a separate isoform of KHK is expressed in a variety of tissues that theoretically allows for the extrahepatic metabolism of fructose. If noncanonical version of fructolysis became heightened when the standard pathway could no longer cope, the increase in fructolytic, and possibly adipogenic or FGF-21 related genes would no longer be upregulated as is seen in this 12-wk study.

Previous reports indicated that FVB/N mice are resistant to diet-induced obesity, especially when compared with often-used C57BL/6 mice or other obesity-prone mouse backgrounds (21). In agreement, this study found that the FVB/N mice fed a high-fructose diet for 12 wk had no significant change in body weight when compared with the control-diet-fed cohort. Furthermore, there was also no significant change in glucose tolerance between the fructose- and control-fed cohorts after either 8 or 12 wk of feeding. These findings uniquely allow for the investigation of high-fructose feeding in FVB/N mice without the confounding effects of obesity and systemic insulin resistance. Therefore, the metabolic outcomes of high-fructose feeding were also examined in FVB/N mice after 12 wk on diet, which corresponds to 16 wk of age. This is only one of a few studies to assess the effects of long-term, high-fructose feeding in mice.

One potential reason for the lack of weight gain in these mice on a high-fructose diet could be due to the increase in energy expenditure. The fructose-fed FVB/N mice were shown to have a significant increase in energy expenditure and movement compared with their control-fed counterparts. With the increase in dietary fructose these mice have an increase in movement, energy expenditure, and RER. The high-fructose diet (3.9 kcal/g) had slightly more calories when compared with the control diet (3.8 kcal/g) due to the incorporation of fructose into the food. This corresponds to the average fructose-fed mouse consuming an extra 2.06 kcal every 12 h when compared with the control-fed cohort. The significant increase in energy expenditure in the fructose-fed mice only accounts for an increase of 0.72 kcal every 12 h. As more ATP is utilized in the catabolism of fructose compared with that of glucose, the simple metabolism of a high-fructose diet may be responsible for the increase in caloric intake with no change in average body weight between the fructose- and control-fed cohorts (35). As the metabolism of fructose can diverge in many directions, and some of its byproducts, such as lactate, may be utilized for ATP generation, further studies into the bioenergetics of fructose metabolism are required to conclude that the metabolism of fructose is responsible for the increased caloric intake.

Interestingly, many previous studies have concluded that high-fructose feeding decreases energy expenditure, or that any increase in energy expenditure is due to increased thermogenesis (36, 37). From this study it is evident that FVB/N mice fed a high-fructose diet for 12 wk have a significant increase in RER and movement, which suggests an increase in energy expenditure. However, for this study, energy expenditure was calculated as a function of RER and oxygen consumption, and not directly, so it is possible that the increase in energy expenditure is mostly a function of RER. In general, RER does increase as a function of energy expenditure, as increased exertion will make carbohydrate the preferred fuel source of the body over lipid that predominates at rest. In this study, however, under such high-fructose conditions, carbohydrate may be the predominant fuel source regardless of energy expenditure (38).

Despite no change in weight or glucose tolerance between control and fructose-fed FVB/N mice, notable metabolic differences were seen in peripheral tissues between these two cohorts. Interestingly, there is a significant increase in epididymal adipocyte size in the fructose-fed mice, indicating differences in systemic metabolic regulation and local tissue biology. The increase in adipocyte size is not surprising, even without the presence of obesity, as a high-fructose diet is well established to enhance hepatic lipogenic pathways (39). Starting in the liver, the process of fructolysis is completely unregulated, bypassing the rate-limiting steps of glycolysis. This uncontrolled metabolism of fructose leads to the buildup of many metabolic intermediates, and without increased energy expenditure most of the excess carbon is shunted toward lipid production, as only so much energy can be stored as glycogen. Increased hepatic lipogenesis is common with prolonged fructose intake, leading to large amounts of VLDL to be released into circulation and to be stored in peripheral adipose depots (40). Therefore, even in the absence of obesity, after 12 wk of high-fructose feeding it is not surprising to see increases to adipocyte size as lipid continues to be produced and stored at heightened rates.

In addition to the increased adipocyte size seen in fructose-fed mice, further differences in adipose tissue biology are detected in the adipose tissue secretome. Adipose tissue is an active participant in metabolic regulation and can influence cell biology in numerous tissues through endocrine, paracrine, and autocrine signaling (41). In the adipokine array presented here conducted on mammary adipose tissue secretome, it is evident that a high-fructose diet imparts significant changes to adipose tissue and related signaling. The largest change in a measured adipokine is in the release of FGF-21, which is significantly upregulated in the mammary adipose tissue secretome in FVB/N mice on a high-fructose diet. This result was in stark contrast to the loss of any dietary effect of fructose ingestion on circulating levels of FGF-21 at 12 wk and suggested that localized adipose tissue-specific effects of fructose feeding on gene expression may predominate upon longer-term fructose consumption.

FGF-21 is known to be highly influential maintaining metabolic homeostasis in stressful situations, such as prolonged fasting or high-fructose feeding (42–44). FGF-21 is known to induce numerous positive health outcomes in humans including improving insulin sensitivity, reducing hepatic steatosis, and even improving cognition (45–47). FGF-21 drug trials have been conducted in patients with obesity; however, circulating FGF-21 is already elevated in individuals with obesity and is thought to be resistant to the benefits of FGF-21 (48). Although the long-term effects of obesity on the production of FGF-21 have been thoroughly investigated, the impact of continued fructolysis on the regulation of FGF-21 is yet to be fully elucidated.

In addition to an increase in FGF-21 in the adipose secretome of fructose-fed mice, significant differences were also evident in the hepatic expression and circulation of FGF-21 between the control and fructose-fed animals, both after 48 h and 12 wk on diet. Although there is evidence for adipose tissue production of FGF-21, the liver is main source of production. After 48 h of high-fructose feeding in FVB/N, there is a significant sixfold increase in the expression of hepatic FGF-21 when compared with the control-fed mice, and further significant increases to both its co-receptor and receptor β-klotho and FGFr1, respectively. After 12 wk of high-fructose feeding, however, there is no significant difference in FGF-21 or related genes when compared with the control-fed mice.

Significant differences in circulating FGF-21 between the control and fructose-fed mice were also noted at both 48 h and 12 wk on diet. In concert with the trends displayed in the gene expression results, there is a significant increase in the serum levels of FGF-21 after 48 h of high-fructose feeding, but there is a significant decline in serum FGF-21 after 12 wk of high-fructose feeding after comparing these levels with the control-fed FVB/N mice. As hepatic production of FGF-21 is the main source of circulating hepatokine it is not surprising that the hepatic expression correlates with circulating levels. As circulating FGF-21 is known to have protective benefits in terms of metabolic health, the significant reversal in hepatic production and circulating levels after continued fructose consumption may provide evidence for the metabolic decline evident in humans after long-term fructose consumption.

Perspectives and Significance

Overall, work throughout this paper demonstrates the metabolic effects of a high-fructose diet in FVB/N mice, which provides insights into the negative health outcomes of high-fructose feeding evident in humans. Although much previous work has focused on the acute metabolic effects of fructose metabolism, here the differences in short- and long-term fructose metabolism are presented for one of the first times. In congruence with past work, acute fructose ingestion is associated with a significant increase in fructolytic, adipogenic, and FGF-21-related hepatic gene expression. Intriguingly, after continued fructose consumption over 12 wk, these same genes are no longer upregulated, and a decline in circulating FGF-21 is reported. These results shed light on metabolic adaptions to chronic fructose consumption and provide the framework for future studies into the mechanism by which a continued high-fructose diet induces metabolic disease.

GRANTS

This project was financially supported by National Institute of Diabetes and Digestive and Kidney Diseases through the Diabetes Research and Training Center Grant P30-DK-020595 (to M. J. Brady). J. W. Strober was supported by a predoctoral fellowship from the Graduate Training in Disparities Research Grant funded by the Susan G. Komen Foundation.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

J.W.S., S.F., and M.J.B. conceived and designed research; J.W.S., S.F., and H.Y. performed experiments; J.W.S., S.F., H.Y., and M.J.B. analyzed data; J.W.S., S.F., H.Y., and M.J.B. interpreted results of experiments; J.W.S. and S.F. prepared figures; J.W.S. and M.J.B. drafted manuscript; J.W.S., S.F., H.Y., and M.J.B. edited and revised manuscript; J.W.S., S.F., H.Y., and M.J.B. approved final version of manuscript.