Figures
Abstract
Excessive fructose consumption inhibits adaptive increases in intestinal Ca2+ transport in lactating and weanling rats with increased Ca2+ requirements by preventing the increase in serum levels of 1,25(OH)2D3. Here we tested the hypothesis that chronic fructose intake decreases 1,25(OH)2D3 levels independent of increases in Ca2+ requirements. Adult mice fed for five wk a high glucose-low Ca2+ diet displayed expected compensatory increases in intestinal and renal Ca2+ transporter expression and activity, in renal CYP27B1 (coding for 1α-hydroxylase) expression as well as in serum 1,25(OH)2D3 levels, compared with mice fed isocaloric glucose- or fructose-normal Ca2+ diets. Replacing glucose with fructose prevented these increases in Ca2+ transporter, CYP27B1, and 1,25(OH)2D3 levels induced by a low Ca2+ diet. In adult mice fed for three mo a normal Ca2+ diet, renal expression of CYP27B1 and of CYP24A1 (24-hydroxylase) decreased and increased, respectively, when the carbohydrate source was fructose instead of glucose or starch. Intestinal and renal Ca2+ transporter activity and expression did not vary with dietary carbohydrate. To determine the time course of fructose effects, a high fructose or glucose diet with normal Ca2+ levels was fed to adult rats for three mo. Serum levels of 1,25(OH)2D3 decreased and of FGF23 increased significantly over time. Renal expression of CYP27B1 and serum levels of 1,25(OH)2D3 still decreased in fructose- compared to those in glucose-fed rats after three mo. Serum parathyroid hormone, Ca2+ and phosphate levels were normal and independent of dietary sugar as well as time of feeding. Thus, chronically high fructose intakes can decrease serum levels of 1,25(OH)2D3 in adult rodents experiencing no Ca2+ stress and fed sufficient levels of dietary Ca2+. This finding is highly significant because fructose constitutes a substantial portion of the average diet of Americans already deficient in vitamin D.
Citation: Douard V, Patel C, Lee J, Tharabenjasin P, Williams E, Fritton JC, et al. (2014) Chronic High Fructose Intake Reduces Serum 1,25 (OH)2D3 Levels in Calcium-Sufficient Rodents. PLoS ONE 9(4): e93611. https://doi.org/10.1371/journal.pone.0093611
Editor: Daniel Tomé, Paris Institute of Technology for Life, Food and Environmental Sciences, France
Received: November 28, 2013; Accepted: March 5, 2014; Published: April 9, 2014
Copyright: © 2014 Douard et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported by the National Science Foundation (IOS-1121049 to RPF). The work also received support from the Foundation of UMDNJ (RPF), and the Benjamin Delessert Foundation (to V. Douard). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: Yves Sabbagh is a full time employee of Genzyme, a Sanofi company, and owns Sanofi stock. However, this does not alter in any way the authors’ adherence to PLOS ONE policies on sharing data and materials.
Introduction
Fructose is one of the key components of the American diet and represents more than 10% of daily calorie intake [1]. This remarkable and recent increase in fructose consumption coincides with an equally striking increase in prevalence of vitamin D deficiency or insufficiency in developed countries [2]. Since excessive fructose consumption and vitamin D deficiency are each associated with similar metabolic diseases (i.e., hypertension, chronic kidney disease, metabolic syndrome, obesity) [3], [4], [5], a better understanding of the interaction between fructose and vitamin D metabolism will contribute to better health recommendations. Fructose absorption across the intestinal apical membrane is mediated by the facilitative glucose transporter GLUT5 while basolateral exit is mediated by GLUT2 [6], [7]. Fructose is then metabolized in the liver and to a lesser extent in the kidney.
In the event of decreased Ca2+ levels in the serum due to low dietary intake or increased demand of Ca2+, the synthesis of the hormonally active form of vitamin D, 1,25-dihydroxyvitamin D3 (1,25(OH)2D3) is enhanced, leading to adaptive increases in intestinal Ca2+ absorption. If normal serum Ca2+ level cannot be maintained by intestinal absorption, then 1,25(OH)2D3 together with the parathyroid hormone (PTH) will mobilize bone Ca2+ and increase reabsorption of Ca2+ from the renal distal tubule [8]. Active intestinal and renal Ca2+ transport involves Ca2+ entry through the apical Ca2+ channel transient receptor potential vanilloid 6 and 5 (TRPV6 and TRPV5) respectively, its intracellular diffusion via Ca2+-binding proteins (CaBP9k and CaBP28k respectively), and its extrusion across the basolateral membrane mainly through the plasma membrane Ca2+-ATPase (PMCA1). 1,25(OH)2D3 stimulates the transcription, via the vitamin D receptor (VDR) [9], of TRPV6 and TRPV5, as well as CaBP9k and CaBP28k.
Our recent work has demonstrated that excessive fructose consumption affects 1,25(OH)2D3 and Ca2+ metabolism [10], [11]. In both lactating and rapidly growing weanling rats, fructose consumption inhibits adaptive increases in intestinal and renal Ca2+ transport, mainly by preventing the upregulation of TRPV6 and CaBP9k expression in the duodenum, and to a lesser extent, of TRPV5 and CaBP28k in the kidney. The mechanism underlying its inhibitory effect on Ca2+ transport is that chronic fructose intake also inhibits the adaptive increases in synthesis and blood levels of 1,25(OH)2D3 regulating intestinal and renal Ca2+ transport systems. Highlighting the key role of 1,25(OH)2D3, treatments with 1,25(OH)2D3 rescued the inhibitory effect of chronic fructose intake on intestinal Ca2+ absorption in weanling rats [11].
The precursor vitamin D3 (originating from the skin or diet) is initially hydroxylated by 25-hydroxylase (encoded by CYP2R1) in the liver, producing calcidiol or 25(OH)D3 that is further hydroxylated by 1α-hydroxylase (encoded by CYP27B1) to 1,25(OH)2D3 in the kidney. The circulating level of 1,25(OH)2D3 is balanced between synthesis by 1α-hydroxylase and degradation by 24-hydroxylase (encoded by CYP24A1), also mainly expressed in the kidney [12]. When serum levels of Ca2+ and 1,25(OH)2D3 are normal, 1,25(OH)2D3 inhibits its own synthesis by reducing CYP27B1 expression, while low serum levels of 1,25(OH)2D3 are often associated with a compensatory increase in CYP27B1 expression to restore homeostasis [13]. Chronic fructose-feeding seems to disrupt this balance in rats undergoing a Ca2+ challenge. In fructose-fed lactating rats and in fructose-fed rat models of chronic kidney disease, both 1,25(OH)2D3 levels and CYP27B1 expression are low [10], [11]. In rapidly growing, postweaning rats, fructose feeding not only reduced 1,25(OH)2D3 and CYP27B1 levels, but also increased CYP24A1 expression, suggesting that high fructose intakes enhance the catabolism and impair the synthesis of 1,25(OH)2D3 [11]. However, whether this occurred due to low Ca2+ or some other mechanism could not be determined.
The aim of this study is to test the hypothesis that chronic fructose consumption causes 1,25(OH)2D3 insufficiency independent of its effects on adaptive adjustments in Ca2+ transport. We first demonstrated that adult mice challenged with a low Ca2+ diet failed to upregulate Ca2+ transporter expression if fed a high 43% fructose diet for five wk, because fructose prevented compensatory increases in 1,25(OH)2D3 levels. To minimize potential diet-induced changes in Ca2+ transport, we then fed a second group of adult mice with normal Ca2+ but with a much higher (60%) fructose concentration for a longer duration of three mo, and eventually found the excessive fructose consumption did reduce levels of 1,25(OH)2D3 independent of Ca2+ transporter expression. Finally, we used rats to enable us to sample blood levels of 1,25(OH)2D3 and of other hormones, as well as of Ca2+ and Pi concentrations, each month and determine the time course of fructose effects on 1,25(OH)2D3.
Materials and Methods
All the procedures in this study were approved by the Institutional Animal Care and Use Committee, Rutgers, The State University of New Jersey.
Experimental Design
Experimental diets were modified from a standard American Institute of Nutrition (AIN)-93G formula by the manufacturer (Research Diets, New Brunswick, NJ). Previous work indicated that 4 wk is sufficient for fructose consumption to have significant effects on intestinal Ca2+ transport and serum 1,25(OH)2D3 levels [11]. Male mice and rats were used and were kept under standard conditions: 12-h light-dark cycle and 24°C. Study 1: mice fed fructose for five wk. Four-wk-old C57BL/6 mice (Jackson Laboratory, Bar Harbor, Maine), of similar initial body weights were randomly divided into four groups (n = 5) and pair-fed a 43% glucose or 43% fructose diet each containing normal (0.5%) or deficient Ca2+ (0.02%) levels (body weights in Fig. S1). The levels of Ca2+ in the diet were based on previous studies [14], [15]. The rationale underlying dietary fructose concentrations is in the discussion, under “Limitations”. Study 2: mice fed fructose for three mo. Four-wk-old mice were randomly divided into three groups (n = 5) and fed ad libitum a 63% glucose, 63% fructose, or 63% starch diet containing normal Ca2+ levels for three mo (body weights in Fig. S2). The higher dietary fructose concentration and the longer feeding duration were designed to attempt to determine fructose effects on 1,25(OH)2D3 under conditions of Ca2+ sufficiency. Study 3: rats fed fructose for three mo. Four wk old rats were randomly divided into two groups (n = 6) fed a 43% glucose or 43% fructose diet containing normal Ca2+ levels ad libitum for three mo (body weights in Fig. S3). Rats were utilized to ensure a large enough volume of serum would be collected each month for monitoring of Ca2+, 1,25(OH)2D3, phosphate (Pi), as well as the putative Pi and 1,25(OH)2D3 regulators, fibroblast growth factor 23 (FGF23) and parathyroid hormone (PTH).
In vitro Intestinal Ca2+ Transport
Intestinal segments were everted immediately after isolation and prepared as sacs to determine Ca2+ transport rates at 37°C as described previously [3]. The everted gut sacs were made by using the first 4-cm of proximal duodenum where active transcellular transport of Ca2+ is localized [14] and then incubated in Ca2+ transport buffer [3]. The inner serosal and outer luminal compartments had equal initial concentrations (0.25 mM) of nonradioactive Ca2+, then a tracer concentration of 45Ca2+ was added to the outer mucosal compartment. After 1 h, the active accumulation of 45Ca2+ in the inner serosal compartment was calculated as a ratio of the final concentration of 45Ca2+ in compartments, inside/outside.
Serum
Following earlier work [3], serum Pi concentrations were determined using QuantiChrom Pi Assay Kit (BioAssay Systems, Hayward, CA). The total serum Ca2+ concentrations were determined by previously described techniques using flame atomic absorption spectrophotometry (Perkin-Elmer Model 633, Norwalk, CT) [16].
1,25(OH)2D3, PTH and FGF23 Assays
Following earlier work [3], serum 1,25(OH)2D3 levels were measured by enzyme immunoassay (ImmunoDiagnosticSystems (IDS), Arizona). Serum samples were de-lipidated by adding a solution containing dextran sulphate and magnesium chloride, and then centrifuged at 10000×g for 10 min. The resulting supernatant containing the de-lipidated serum is then 1,25(OH)2D3 immuno-extracted from potential cross-reactants by incubation for 90 minutes with a highly specific solid phase monoclonal anti-1,25(OH)2D3 before assaying by intact enzyme-linked immuno-sorbent assay (ELISA) for 1,25(OH)2D3 levels according to the manufacturer’s instructions [17]. Intact FGF23 measurement was performed by ELISAs according to manufacturer’s instructions (Kainos, Tokyo, Japan). PTH was determined using a 2-site sandwich ELISA (Immunotopics, San Clemente, CA, USA).
Western Blot Analysis
Western blot analyses were performed using 50 μg of intestinal or renal protein extracts following earlier work [3]. For CaBP9k (anti-rabbit, Swant Swiss Antibodies, Switzerland), 4–20% Tris-HCl gels (BioRad, Hercules, CA) were used. For CYP27B1 (anti-rabbit, Santa Cruz Biotechnology Inc., CA) and CYP24A1 (Sigma Chemicals, MO), 12% Tris-HCl gels were used. All membranes were stripped and re-probed with β-actin antibody (anti-mouse, Chemicon International, MA). The membrane was blocked with 5% nonfat milk in Tris-buffered saline-Tween 20 (TBS-T) (0.1% Tween 20, 50 mmol/L Tris-HCL, 137 mmol/L NaCl, pH 7.4) buffer for 1 h. The blots were then incubated with the primary antibody, CaBP9k, CYP24A1 and CYP27B1 diluted 1∶1000, 1∶200 and 1∶1000, respectively, in 5% blocking agent in TBS-T buffer overnight at 4°C. Anti-β-actin antibody was diluted 1∶2000 in the same buffer but incubated with the membrane for one h at room temperature. The membranes were incubated with secondary antibodies, anti-rabbit (1∶20000, GE Healthcare Life Sciences, PA) or anti-mousse (1∶5000, GE Healthcare Life Sciences, PA), for one h at room temperature. The western blot was revealed by enhanced chemiluminescent (ECL) HRP substrate (Thermo Scientific, Rockford, IL USA).
Real Time RT-PCR
Total RNA from homogenized intestinal mucosa or kidney was isolated, reverse transcribed, and real-time RT-PCR performed using Mx3000P (Stratagene, La Jolla, CA) as previously described [18]. The reference gene was α-elongation factor1 (EF1α) whose expression is independent of age and dietary carbohydrate [19]. Previously published primer sequences and annealing temperatures [3] [19] are listed in Table S1. To examine the effects of animal age on expression of intestinal Ca2+ transporters and of renal 1,25(OH)2D3 metabolic enzymes, the glucose-fed mouse samples from Studies 1 and 2 were re-run together through real time RT-PCR with results normalized to the older, 4-mo-old, glucose-fed mice of study 2.
Statistical Analyses
Data are presented as means ± SEM. For studies one and three, a two-way ANOVA was used to determine the effect of fructose and Ca2+ level, and of fructose and feeding time course, respectively. If an initial two-way ANOVA indicated a significant effect, a one-way ANOVA followed by LSD test (STATVIEW, Abacus Concepts) determined if differences existed between groups. Differences were considered significant at P<0.05. In study two, a one-way ANOVA was used to test for fructose effects.
Results
Study 1: Mice Fed Ca2+-deficient Diets for Five Weeks
Differences in intestinal Ca2+ transport and transporter mRNA were associated with a low Ca2+ diet containing fructose as main carbohydrate source. Active transepithelial Ca2+ transport rate was modest in mice fed fructose-based, normal Ca2+ diet and was similar to that in mice fed a glucose-based, normal Ca2+ diet (Fig. 1A). When dietary Ca2+ level was made deficient, active transepithelial Ca2+ transport was enhanced by >2 fold when the carbohydrate source in the diet was glucose. However, no adaptive increases occurred on the fructose-based diet. We investigated the mRNA expression of the three transporters known to mediate active Ca2+ transport in the duodenum: TRPV6, CaBP9k and PMCA1 (Fig. 1B). mRNA levels of TRPV6 and CaBP9k increased significantly by ∼400- and 30-fold, respectively, in mice fed a glucose-based, low Ca2+ diet compared to the glucose-based, normal Ca2+ diet, paralleling the low Ca2+-induced increase in Ca2+ transport. In mice fed a fructose-based, low Ca2+ diet these mRNA levels increased by ∼150- and 10-fold, respectively, compared to those fed normal Ca2+ diet. Thus, for TRPV6 and CaBP9k, dietary fructose dampened adaptive increases of Ca2+ transporter expression by three-fold compared to glucose. Differences in CaBP9k protein levels clearly paralleled those of CaBP9k mRNA (Fig. S4). As we previously demonstrated, intestinal GLUT5 expression increased with dietary fructose, regardless of dietary Ca2+ level (Fig. S5), suggesting that the inhibitory effect of fructose on Ca2+ transporter expression is specific.
(A) Ca2+ transport was measured using everted duodenal sacs from mice fed diets containing either 43% glucose or fructose, in combination with either normal or low Ca2+. Active transepithelial Ca2+ transport from the luminal to the basolateral compartment was expressed as a ratio of the final quantity of 45Ca2+ inside/outside of the everted gut sacs. (B) mRNA expression levels of intestinal Ca2+ (TRPV6, CaBP9k, PMCA1) transporters. All expression data were analyzed by real-time PCR using EF1α as a reference and then normalized to levels in mice fed a normal Ca2+-glucose diet (Nor Ca2+-G). Data are means ± SEM (n = 6 per group). Differences (P<0.05) among means are indicated by differences in superscript letters, as analyzed by 1-way ANOVA LSD. Thus, within a gene of interest, bars with superscript “a” are > bars with “b” which in turn are > bars with “c”. Dietary fructose inhibits compensatory increases in Ca2+ transporter activity and expression induced by dietary Ca2+-deficiency.
Differences in renal mRNA expression of synthesis and degradation enzymes for 1,25(OH)2D3 were associated with fructose and low Ca2+ diets. In the kidney of mice fed low Ca2+, mRNA levels of CYP27B1 increased significantly by 17- and 7-fold, respectively for glucose- and fructose-based diets, compared to a normal Ca2+ diet with glucose or fructose as carbohydrate source (Fig. 2A). Thus, dietary fructose lead to a significantly lower expression of CYP27B1 when associated with low Ca2+ levels, as compared to dietary glucose. CYP24A1 mRNA expression was significantly upregulated by ∼14-fold with low Ca2+, regardless of dietary sugar. Similar patterns of expression were found with protein levels (Fig. 2B).
(A) mRNA expression levels of renal CYP27B1 (1α-hydroxylase) and CYP24A1 (24-hydroxylase). (B) The protein abundance of CYP27B1 and CYP24A1, using β-actin as a reference. (C) mRNA expression levels of renal TRPV6, CaBP28k, CaBP9k and PMCA1. Nor = normal; G = glucose; F = fructose. All mRNA expression level data were analyzed by real-time PCR using EF1α as a reference and normalized to levels in mice fed glucose and normal Ca2+ diet. Data are means ± SEM (n = 5–6 per group). Differences (P<0.05) among means are indicated by differences in superscript letters, as analyzed by 1-way ANOVA LSD. Thus, within a gene of interest, bars with superscript “a” are > bars with “b” which in turn > bars with “c”. Dietary fructose inhibits compensatory increases in renal expression of Ca2+ transporters and 1,25(OH)2D3 metabolizing enzymes induced by dietary Ca2+-deficiency.
As expected, the circulating level of 1,25(OH)2D3 was significantly increased with low Ca2+ (Table 1) and paralleled the increases in renal mRNA and protein expression of CYP27B1 and CYP24A1 with low Ca2+. However, dietary fructose dampened the low Ca2+-induced increase in 1,25(OH)2D3 level which was not significantly different from those in mice fed normal Ca2+ diets (P>0.075 by one-way ANOVA). A low Ca2+ diet was associated with a significant 40% decrease in circulating level of FGF23, independent of dietary sugar concentrations. However, serum concentrations of Ca2+ and Pi which are primarily regulated by 1,25(OH)2D3 and FGF23, did not vary with diet.
In the kidney, differences in mRNA expression of Ca2+ transporters were associated with low Ca2+ diets. Low dietary Ca2+ induced a ∼two-fold increase in renal TRVP5 and CaBP9k mRNA levels in combination with either glucose or fructose (Fig. 2C). In contrast, CaBP28k and PMCA1 mRNA expression did not vary with diet.
Study 2: Mice Fed Fructose for Three Months
In order to determine if chronic high fructose intake alters CYP27B1 and CYP24A1 expression independently of Ca2+demand and metabolism, mice were challenged for three mo with diets containing higher levels of fructose and sufficient levels of Ca2+. Similar rates of food intake were observed among mice fed glucose, fructose or starch diets (average of 0.09±0.01 g per g of bw per day for all mice). Duodenal Ca2+ transport rates did not vary with carbohydrate source and were similar in glucose-, fructose- and starch-fed mice (Fig. 3A). TRPV6, CaBP9k and PMCA1 mRNA expression did not vary among the three diets (Fig. 3B). CaBP9k protein expression was not detected by Western blot (data not shown), indicating a low expression level in the 4 mo old mice fed sufficient dietary Ca2+.
(A) Active transduodenal Ca2+ transport from the luminal to the basolateral compartment was expressed as a ratio of the final quantity of (45Ca2+ inside/45Ca2+ outside) of the everted sacs of mice fed diets containing 63% fructose, glucose or starch. (B) mRNA expression levels of intestinal Ca2+ (TRPV6, CaBP9k, PMCA1) transporters. All expression data were analyzed by real-time PCR using EF1α as a reference and normalized relative to levels seen in mice fed glucose diet. Data are means ± SEM (n = 5 per group). Chronic consumption of high fructose levels has no significant effect on intestinal Ca2+ transport rate and transporter mRNA expression.
In the kidney, fructose feeding had remarkable effects on the mRNA expression of genes involved in regulating 1,25(OH)2D3 levels. CYP27B1 expression decreased and CYP24A1 expression increased when compared to starch or glucose feeding (Fig. 4A). However, fructose had no effect on TRPV5, CaBP9k, CaBP28k, or PMCA1 expression (Fig. 4B). Taken together, these findings from intestine and kidney indicate that excessive fructose consumption may regulate 1,25(OH)2D3 independent of any effect on Ca2+ transport.
(A) Marked effects of excessive fructose intake on mRNA expression levels of renal 1,25(OH)2D3 metabolic enzymes, CYP27B1 and CYP24A1. (B) mRNA expression levels of renal TRPV6, CaBP28k, CaBP9k and PMCA1. All expression data were analyzed by real-time PCR using EF1α as a reference and normalized relative to levels seen in mice fed glucose diet. Data are means ± SEM (n = 5 per group). Differences (P<0.05) among means are indicated by differences in superscript letters, as analyzed by 1-way ANOVA LSD. Thus, within a gene of interest, bars with superscript “a” are > bars with “b”. Chronic consumption of high fructose levels has dramatic effects on mRNA expression of renal 1,25(OH)2D3 metabolizing enzymes but not on Ca2+ transporter mRNA expression.
To examine possible mechanisms behind the difference in Ca2+ transport rates between Figs. 1A and 3A, we reanalyzed the intestinal transporter and renal 1,25(OH)2D3 regulating enzymes in a direct comparison of mRNA expression levels in mice fed normal Ca2+ and glucose from study 1 (those mice were 2.5 mo old at sacrifice) and mice fed normal Ca2+ and glucose from study 2 (those mice were 4 mo old at sacrifice). Intestinal mRNA expression levels of TRPV6 and CaBP9k were both ∼100-fold lower at 4 than at 2.5 mo of age (Fig. 5A). Low CaBP9k expression at 4 mo-old mice fed normal Ca2+ explains the undetectable CaBP9k protein levels in these mice. Despite the large age-related differences in Ca2+ transport and transporter expression, expression of renal CYP27B1 and CYP24A1 was the same (Fig. 5B).
A comparison of the expression levels of Ca2+ transporters (A) and 1,25(OH)2D3 metabolic enzymes (B) in four mo old mice after three mo of feeding on normal Ca2+ diet containing 63% glucose (from study 2) and in two mo old mice fed a normal Ca2+ diet containing 43% glucose for five wk (from study 1). All expression data were analyzed by real-time PCR using EF1α as a reference and normalized relative to levels seen in four mo old mice. Data are means ± SEM (n = 5–6 per group). Differences (P<0.05) between means are indicated by asterisks. Expression of intestinal Ca2+ transporters decreases with age.
Study 3: Rats Fed Fructose for Three Months
Having demonstrated that prolonged feeding of a fructose diet altered renal expression of 1,25(OH)2D3 metabolizing enzymes independent of changes in expression of Ca2+ transporters and binding proteins, we proceeded to evaluate the time course of the fructose effect on 1,25(OH)2D3 levels. Rats fed a fructose or glucose diet displayed similar rates of food intake as well as similar final body weights (Fig. S3A, B). Similar to Study 2, the expression levels of intestinal and renal transporters with normal Ca2+ were also the same on fructose- and glucose-based diets (Figs. 6, 7). While it might have been possible to get a statistically significant fructose effect with a greater n, another reason we failed to get significance was the known age-related reductions in expression of Ca transporters that would also reduce the magnitude of the fructose effect. However, fructose feeding did lead to a two-fold decrease in mRNA expression of the 1,25(OH)2D3 synthesizing renal enzyme, CYP27B1 (Fig. 7A). Meanwhile, CYP24A1 expression did not change. Protein levels demonstrated similar results (Fig. 7B).
Rats were fed for three mo a normal Ca2+ diet containing either 43% glucose or fructose. mRNA expression levels of intestinal Ca2+ (TRPV6, CaBP9k, PMCA1) transporters remained similar between diets.
Rats were fed for three mo a normal Ca2+ diet containing either 43% glucose or fructose. (A) mRNA expression level of renal 1,25(OH)2D3 metabolic enzymes, CYP27B1 and CYP24A1 (B) The protein abundance of CYP27B1 and CYP24B1 using β-actin as a reference. (C) mRNA expression levels of renal TRPV6, CaBP28k, CaBP9k and PMCA1. All expression data were analyzed by real-time PCR using EF1α as a reference and normalized relative to levels seen in rat fed glucose diet. Data are means ± SEM (n = 6 per group). Differences (P<0.05) between means are indicated by asterisks. Chronic consumption of high fructose diets reduces mRNA and protein expression of CYP27B1.
As expected, the decreased CYP27B1 expression after three mo of consuming a high fructose diet was associated with a significantly lower circulating level of 1,25(OH)2D3 versus the glucose diet group (Table 2). The most pronounced effect was with duration as in both diet groups, 1,25(OH)2D3 serum concentrations fell by over 40% from two to three and four mo of age. Conversely, circulating levels of FGF23 were not affected by the dietary sugar, and increased about 20% between three and four mo of age, suggesting that age-related reductions in 1,25(OH)2D3 levels occur prior to increases in FGF23. Serum PTH, Ca2+ and Pi did not vary with duration or dietary sugar.
Discussion
We show in mice that excessive fructose intake inhibits adaptive increases in intestinal Ca2+ absorption and in 1,25(OH)2D3 levels, in response to dietary Ca2+ restriction. Thus, regardless of the source of Ca2+ challenge (growth [11], lactation [10], renal disease [3], dietary insufficiency (this paper)), excessive fructose consumption will diminish, if not completely abolish, the main physiological responses: increased intestinal absorption and renal reabsorption of Ca2+. Very importantly, we also demonstrate in both rats and mice that chronic fructose intake can decrease 1,25(OH)2D3 levels independent of Ca2+ demand, and can alter CYP27B1 or CYP24A1 expression in the absence of physiological and nutritional Ca2+ challenge.
Excessive Fructose Intake Decreases 1,25(OH)2D3
Results from these current studies and our previous work [10], [11] have consistently shown that fructose feeding prevents adaptive increases in serum levels of 1,25(OH)2D3. This fructose-induced decrease in serum 1,25(OH)2D3 concentrations was not associated with a reduction in serum levels of its precursor 25(OH)D3, suggesting fructose had little effect on the initial steps of vitamin D synthesis. Instead, these findings indicate a specific effect of fructose at the renal step of either 1,25(OH)2D3 synthesis or degradation [10]. The fructose-induced decrease in 1,25(OH)2D3 levels was always associated with a fructose-induced decrease in CYP27B1 expression and less consistently with increased CYP24A1 expression. So far the effect of fructose on 1,25(OH)2D3 had been observed only when Ca2+ had been challenged. We demonstrate in the current study that chronic fructose intake reduces over time normal circulating levels of 1,25(OH)2D3 independently of any increase in Ca2+ requirement.
Chronic Fructose Intake and Intestinal Ca2+ Transport
Total intestinal Ca2+ absorption consists of a passive paracellular and a 1,25(OH)2D3-dependent active transcellular pathway [20]. Since the ultimate source of Ca2+ is intestinal absorption, vertebrates undergoing Ca2+ stress adaptively increase active Ca2+ transcellular transport, achieved by augmenting levels of 1,25(OH)2D3. Here we confirmed in mouse our previous findings in rat models showing that marked increases in active absorption of Ca2+ are associated with equally remarkable increases in TRPV6 and CaBP9K expression, and that high fructose intakes dramatically compromise these adaptive increases in Ca2+ transport rates as well as in TRPV6 and CaBP9k mRNA or protein levels [10], [11].
In the present work, under Ca2+ restricted conditions, the inhibition of adaptive increases in TRPV6 and CaBP9k expression by fructose likely explains the limited upregulation of active intestinal Ca2+transport in rodents with high fructose intakes. CaBP9k and TRPV6 had been shown to play a major role in the adaptation of intestinal Ca2+ transport under low Ca2+ conditions, and after induction by 1,25(OH)2D3 [14]. When dietary supply of Ca2+ is normal, TRPV6 null mice or CaBP9k null mice display a normal Ca2+ transport. However, in double KO TRPV6−/−/CaBP9k−/− mice, the adaptive increase in intestinal active Ca2+ transport in response to low Ca2+ diet or 1,25(OH)2D3 injection was reduced by 50% compared to that of wild-type mice, indicating that both CaBP9k and TRPV6 are needed for optimal Ca2+ transport under a Ca2+ challenge condition [14]. The same adaptive increase was also seriously impaired in the single KO TRPV6−/− mice [14], suggesting that TRPV6 plays a major role in active intestinal Ca2+ absorption.
We [10], [11] have not found rodent PMCA1 levels to vary with Ca2+ stress and other workers have even used this gene for reference or “housekeeping” purposes [21] because of its stable expression even as TRPV6 levels change markedly. This difference in regulation between TRPV’s and PMCA’s is understandable because Ca2+ transport across the apical membrane is the rate-limiting step [22].
Effect of Age
In four mo old mice and rats, fructose did not affect the active Ca2+ transport or Ca2+ transporter expression likely because active Ca2+ transport constitutes a very small component of total intestinal transport in older, Ca2+-sufficient animals. Comparing Fig. 3A and 1A, Ca2+ transport rate was also lower in older mice (study 2) which were fed normal Ca2+ for three mo than in younger mice fed the same sufficient level of dietary Ca2+ in Study 1. Compared to the 2.5 mo old mice, the modest rates of active Ca2+ transport rates exhibited by four mo old mice were also associated with reduced expression of TRPV6 and CaBP9K, as previously demonstrated in rats [23]. In turn, age-related decreases in Ca2+ transport and transporter expression occurring in adult rodents between two and four mo of age may be linked to the age-related decrease in serum 1,25(OH)2D3 levels occurring simultaneously, as shown in this study and as previously demonstrated in rats [24]. The role of age in modulating 1,25(OH)2D3 levels should be further investigated in mice on the same diets as a limitation of our age comparison is that different (43% carbohydrate, Study 1 and 63%, Study 2) diets were utilized.
In this and in previous work, adult rats and mice fed sufficient Ca2+ and Pi levels and not under Ca2+ stress have normal serum Ca2+ and Pi levels even when fed high fructose diets. The active, NaPi2b-mediated component of intestinal Pi transport also decreases dramatically in postweaning rodents as growth rate decreases markedly with age [25]. These findings suggest that paracellular transport of Ca2+ and Pi may be sufficient to maintain serum levels as demand for these minerals in older rodents is low.
Interactions among Blood Levels of 1,25(OH)2D3, FGF23 and PTH
FGF23 is a hormone synthesized mainly in the osteocytes and osteoblasts of bone and regulates blood levels of Pi via removal of Pi transporters from the apical membrane of kidney cells, thus reducing Pi reabsorption and increasing urinary excretion [10], [26]. FGF23 also regulates 1,25(OH)2D3 by inhibiting CYP27B1 expression and increasing CYP24A1 expression [27]. Moreover, FGF23 synthesis is positively regulated by increased serum levels of 1,25(OH)2D3 or of Pi.
In a previous study using growing rats, we found that chronic consumption of fructose was associated with a significant increase in circulating levels of FGF23, which could potentially be the mechanism by which fructose reduces 1,25(OH)2D3 levels [11]. However, in the present study, in mice fed for five wk a fructose-based, low Ca2+ diet and in rats fed for three mo a fructose-based, normal Ca2+ diet (high-fructose conditions that reduce circulating levels of 1,25(OH)2D3), FGF23 production did not increase with fructose feeding. In mice, the failure of fructose to induce FGF23 production may be due to the confounding effects of hypocalcaemia that acutely reduces circulating level of FGF23 [28]. Our data confirm this since in Ca2+ deficiency, circulating levels of FGF23 are low, thus allowing 1,25(OH)2D3 circulating levels to remain elevated and prevent decreases in blood Ca2+. In adult rats, the mechanism of the fructose-induced decrease in 1,25(OH)2D3 levels remains unclear, as fructose-induced increases in FGF23 levels were not significant perhaps due to a statistically insufficient number of samples. Fructose, however, is known to injure the kidneys [3], [10], [29] which are the site of 1,25(OH)2D3 synthesis. PTH levels were not affected by fructose as shown in this as well as in our previous studies [10], [11].
Limitations
There are limitations to this study, both in conduct and extrapolation of the results to the human condition. Despite similar food intakes, fructose feeding was associated with a significantly greater kidney mass index (kidney weight/body weight) in both the longer-term studies (two and three) in both mice and rats (Figs. S2 and S3). Additionally, rodents and humans have different diets. The AIN recommends either 73% (AIN-93M (maintenance) diet) or 63% (AIN-93G (growth) diet) carbohydrate, levels higher than the average for humans (54%) consuming a high carbohydrate diet [30]. Thus, rodents consuming 43 or 63% sugars, though consuming more % carbohydrate than humans, are eating less than the carbohydrate amount and concentration in a standard AIN-93M laboratory rodent diet. The popularity of>60% fructose diets is demonstrated in the fact that >130 published studies in the last five years specifically used pellets containing ∼60% fructose, because such a level typically causes pathological effects in a mouse within 4 wk. Here we used high concentrations of fructose to allow us to detect any potential effect it may have on expected outcomes. While our previous studies found 63% fructose diets to cause a significant decrease in serum 1,25(OH)2D3 levels [10], [11] after 4–6 wk of feeding, the present paper demonstrated that lowering the amount of fructose in the diet could lead to similar decreases in serum 1,25(OH)2D3, but only after a longer period of feeding (3 mo). Humans consume lower concentrations of fructose that are often mixed with other carbohydrates, but studies have shown that mixing with glucose may accelerate fructose effects. Because of the cumulative effects of fructose demonstrated here, lower fructose concentrations will likely cause similar deleterious effects as higher fructose concentrations if consumed for a longer time period [31], [32].
Conclusion and Perspective
Despite these limitations, we have clearly demonstrated that a chronic intake of high levels of dietary fructose can lead to a decrease in circulating levels of 1,25(OH)2D3 independent of dietary Ca2+ levels and of physiological increases in Ca2+ requirement. This work is highly relevant since fructose, a sugar contained in many types of foods that are being consumed at high levels, may contribute to the increasing prevalence of vitamin D insufficiency, especially in sensitive populations with high, processed sugar intakes such as in children and teenagers [33], [34]. Indeed, when consumption is normalized to body weight, these age groups are the highest consumers of fructose, with an average total ingestion reaching ∼2 g kg−1 day−1 occurring at a time when there is a very high demand for vitamin D during growth [35].
Supporting Information
Figure S1.
The body and kidney weight of mice pair-fed diets containing either 43% glucose or fructose, in combination with either normal or low Ca2+. A) Initial and final body weight after 5 wk of feeding on the experimental diets. B) Kidney weight after 5 wk. C) Kidney somatic index. Bars are means ± SEM; n = 5–8. Nor = normal; G = glucose; F = fructose. Differences (P<0.05) among means are indicated by differences in superscript letters, as analyzed by 1-way ANOVA LSD. Thus, bars with superscript “a” are > bars with “b” and similar to those with “ab”.
https://doi.org/10.1371/journal.pone.0093611.s001
(TIF)
Figure S2.
The body weight, feeding rate, and kidney weight of mice fed diets containing normal Ca2+ and 63% glucose, fructose or starch. A) Body weight after 3 mo of feeding on the special diets. B) Feeding rate per day normalized to body weight. C) Kidney weight after 3 mo of feeding on the special diets. D) Kidney somatic index. Bars are means ± SEM; n = 5–8. Differences (P<0.05) among means are indicated by differences in superscript letters, as analyzed by 1-way ANOVA LSD. Thus, bars with superscript “a” are > bars with “b”.
https://doi.org/10.1371/journal.pone.0093611.s002
(TIF)
Figure S3.
The body and kidney weight of rat fed diets containing either 43% glucose or fructose. A) Body weight after 3 mo of feeding of the special diets. B) Feeding rate per day normalized to body weight. C) Kidney weight after 3 mo of feeding of the special diets. D) Kidney somatic index. Bars are means ± SEM; n = 5–8. Differences (P<0.05) among means are indicated by differences in superscript letters, as analyzed by 1-way ANOVA LSD. Thus, bars with superscript “a” are > bars with “b”.
https://doi.org/10.1371/journal.pone.0093611.s003
(TIF)
Figure S4.
CaBP9k protein abundance. The protein abundance of CaBP9k was determined in the small intestine of mice fed diets containing either 43% glucose or fructose, in combination with either normal or low Ca2+. β-actin was used as a reference. Nor = normal; G = glucose; F = fructose.
https://doi.org/10.1371/journal.pone.0093611.s004
(TIF)
Figure S5.
mRNA levels of GLUT5. Expression level of GLUT5 in the intestine of A) mice fed diets containing either 43% glucose or fructose, in combination with either normal or low Ca2+ for 5 wk; B) mice fed diets containing normal Ca2+ in combination with 63% glucose, fructose or starch for 3 mo; and C) rats fed diets containing either 43% glucose or fructose for 3 mo. All expression data were analyzed by real-time PCR using EF1α as a reference and normalized relative to levels seen in rat fed glucose diet. Nor = normal; G = glucose; F = fructose. Data are means ± SEM (n = 6 per group). Differences (P<0.05) among means are indicated by differences in superscript letters, as analyzed by 1-way ANOVA LSD. Thus, for each study, bars with superscript “a” are > bars with “b”.
https://doi.org/10.1371/journal.pone.0093611.s005
(TIF)
Acknowledgments
We are deeply indebted to Frank Kemp for help in analyzing serum Ca2+ and Pi levels.
Author Contributions
Conceived and designed the experiments: VD RPF. Performed the experiments: VD CP JL PT EW YS. Analyzed the data: VD. Contributed reagents/materials/analysis tools: YS. Wrote the paper: VD RPF JCF.
References
- 1. Bray GA (2013) Energy and fructose from beverages sweetened with sugar or high-fructose corn syrup pose a health risk for some people. Adv Nutr 4: 220–225.
- 2. Forrest KY, Stuhldreher WL (2011) Prevalence and correlates of vitamin D deficiency in US adults. Nutr Res 31: 48–54.
- 3. Douard V, Asgerally A, Sabbagh Y, Sugiura S, Shapses SA, et al. (2010) Dietary Fructose Inhibits Intestinal Calcium Absorption and Induces Vitamin D Insufficiency in CKD. J Am Soc Nephrol 21: 261–271.
- 4. vinh quoc Lu‘o’ng K, Nguyen LT (2013) The beneficial role of vitamin D in obesity: possible genetic and cell signaling mechanisms. Nutrition journal 12: 89.
- 5. Dekker MJ, Su Q, Baker C, Rutledge AC, Adeli K (2010) Fructose: a highly lipogenic nutrient implicated in insulin resistance, hepatic steatosis, and the metabolic syndrome. Am J Physiol Endocrinol Metab 299: E685–694.
- 6. Manolescu AR, Witkowska K, Kinnaird A, Cessford T, Cheeseman C (2007) Facilitated hexose transporters: new perspectives on form and function. Physiology (Bethesda) 22: 234–240.
- 7. Thorens B, Mueckler M (2010) Glucose transporters in the 21st Century. Am J Physiol Endocrinol Metab 298: E141–145.
- 8. Hoenderop JGJ, Dardenne O, Van Abel M, Van Der Kemp AWCM, Van Os CH, et al. (2002) Modulation of renal Ca2+ transport protein genes by dietary Ca2+ and 1,25-dihydroxyvitamin D3 in 25-hydroxyvitamin D3-1alpha-hydroxylase knockout mice. FASEB J 16: 1398–1406.
- 9. Haussler MR, Whitfield GK, Haussler CA, Hsieh JC, Thompson PD, et al. (1998) The nuclear vitamin D receptor: biological and molecular regulatory properties revealed. J Bone Miner Res 13: 325–349.
- 10. Douard V, Suzuki T, Sabbagh Y, Lee J, Shapses S, et al. (2012) Dietary fructose inhibits lactation-induced adaptations in rat 1,25-(OH)(2)D(3) synthesis and calcium transport. FASEB J 26: 707–721.
- 11. Douard V, Sabbagh Y, Lee J, Patel C, Kemp FW, et al. (2013) Excessive fructose intake causes 1,25-(OH)2D3-dependent inhibition of intestinal and renal calcium transport in growing rats. Am J Physiol Endocrinol Metab 304: 1303–1313.
- 12. Christakos S, Ajibade DV, Dhawan P, Fechner AJ, Mady LJ (2012) Vitamin D: metabolism. Rheum Dis Clin North Am 38: 1–11.
- 13. Sakaki T, Kagawa N, Yamamoto K, Inouye K (2005) Metabolism of vitamin D3 by cytochromes P450. Front Biosci 10: 119–134.
- 14. Benn BS, Ajibade D, Porta A, Dhawan P, Hediger M, et al. (2008) Active intestinal calcium transport in the absence of transient receptor potential vanilloid type 6 and calbindin-D9k. Endocrinology 149: 3196–3205.
- 15. Lieben L, Benn BS, Ajibade D, Stockmans I, Moermans K, et al. (2010) Trpv6 mediates intestinal calcium absorption during calcium restriction and contributes to bone homeostasis. Bone 47: 301–308.
- 16. Bogden JD, Kemp FW, Huang AE, Shapses SA, Ambia-Sobhan H, et al. (2008) Bone mineral density and content during weight cycling in female rats: effects of dietary amylase-resistant starch. Nutr Metab (Lond) 5: 34.
- 17. Yu X, Sabbagh Y, Davis SI, Demay MB, White KE (2005) Genetic dissection of phosphate- and vitamin D-mediated regulation of circulating Fgf23 concentrations. Bone 36: 971–977.
- 18. Douard V, Cui XL, Soteropoulos P, Ferraris RP (2008) Dexamethasone sensitizes the neonatal intestine to fructose induction of intestinal fructose transporter (Slc2A5) function. Endocrinology 149: 409–423.
- 19. Douard V, Choi HI, Elshenawy S, Lagunoff D, Ferraris RP (2008) Developmental reprogramming of rat GLUT5 requires glucocorticoid receptor translocation to the nucleus. J Physiol 586: 3657–3673.
- 20. Hoenderop JGJ, Nilius B, Bindels RJM (2005) Calcium absorption across epithelia. Physiol Rev 85: 373–422.
- 21. Brini M (2009) Plasma membrane Ca(2+)-ATPase: from a housekeeping function to a versatile signaling role. Pflugers Archiv : European journal of physiology 457: 657–664.
- 22. Schedl HP, Wilson HD (1985) Calcium uptake by intestinal brush border membrane vesicles. Comparison with in vivo calcium transport. The Journal of clinical investigation 76: 1871–1878.
- 23. Brown AJ, Krits I, Armbrecht HJ (2005) Effect of age, vitamin D, and calcium on the regulation of rat intestinal epithelial calcium channels. Arch Biochem Biophys 437: 51–58.
- 24. Wood RJ, Fleet JC, Cashman K, Bruns ME, Deluca HF (1998) Intestinal calcium absorption in the aged rat: evidence of intestinal resistance to 1,25(OH)2 vitamin D. Endocrinology. 139: 3843–3848.
- 25. Kirchner S, Muduli A, Casirola D, Prum K, Douard V, et al. (2008) Luminal fructose inhibits rat intestinal sodium-phosphate cotransporter gene expression and phosphate uptake. Am J Clin Nutr 87: 1028–1038.
- 26. Shimada T, Urakawa I, Yamazaki Y, Hasegawa H, Hino R, et al. (2004) FGF-23 transgenic mice demonstrate hypophosphatemic rickets with reduced expression of sodium phosphate cotransporter type IIa. Biochem Biophys Res Comm 314: 409–414.
- 27. Shimada T, Hasegawa H, Yamazaki Y, Muto T, Hino R, et al. (2004) FGF-23 Is a potent regulator of vitamin D metabolism and phosphate homeostasis. J Bone Miner Res 19: 429–435.
- 28. Rodriguez-Ortiz ME, Lopez I, Munoz-Castaneda JR, Martinez-Moreno JM, Ramirez AP, et al. (2012) Calcium deficiency reduces circulating levels of FGF23. J Am Soc Nephrol 23: 1190–1197.
- 29. Nakayama T, Kosugi T, Gersch M, Connor T, Sanchez-Lozada LG, et al. (2010) Dietary fructose causes tubulointerstitial injury in the normal rat kidney. Am J Physiol Renal Physiol 298: F712–720.
- 30. Krauss RM, Blanche PJ, Rawlings RS, Fernstrom HS, Williams PT (2006) Separate effects of reduced carbohydrate intake and weight loss on atherogenic dyslipidemia. Am J Clin Nutr 83: 1025–1031.
- 31. Johnson RJ, Perez-Pozo SE, Sautin YY, Manitius J, Sanchez-Lozada LG, et al. (2009) Hypothesis: could excessive fructose intake and uric acid cause type 2 diabetes? Endocr Rev 30: 96–116.
- 32. Douard V, Sabbagh Y, Suzuki T, Ferraris R (2012) A high fructose intake prevents compensatory increases in intestinal calcium transport in rodent models physiologically and nutritionally deficient in calcium Gastroenterology. 142: S557.
- 33. Marriott BP, Cole N, Lee E (2009) National estimates of dietary fructose intake increased from 1977 to 2004 in the United States. J Nutr 139: 1228S–1235S.
- 34. Douard V, Ferraris RP (2013) The role of fructose transporters in diseases linked to excessive fructose intake. J Physiol 591: 401–414.
- 35. Bischoff-Ferrari HA (2011) Vitamin D: role in pregnancy and early childhood. Ann Nutr Metab 59: 17–21.