Article pubs.acs.org/JAFC
Alteration of Gut Bacteria and Metabolomes after Glucaro-1,4lactone Treatment Contributes to the Prevention of Hypercholesterolemia Baogang Xie,*,† Aihong Liu,‡ Xuejun Zhan,§ Xinming Ye,⊥ and Jing Wei† †
School of Pharmacy, Nanchang University, Nanchang, 330006, People’s Republic of China Center of Analysis and Testing, Nanchang University, Nanchang, 330047, People’s Republic of China § Jiangxi Provincial Institute of Medical Science, Nanchang, 330006, People’s Republic of China ⊥ Hospital of Nanchang University, Nanchang, 330006, People’s Republic of China ‡
ABSTRACT: D-Glucaro-1,4-lactone (1,4-GL) has been shown to have a hypocholesterolemic effect in rats and human subjects. However, little information is known concerning the alteration of metabolome associated with the effect. Here, we show that 1,4GL delays the development of hypercholesterolemia with the coadministration of a high-fat, high-cholesterol diet (HFHC) in rats. Metabonomic results based on proton nuclear magnetic resonance indicate that urinary trimethylamine N-oxide, trimethylamine, lactate, acetate, formate, and creatinine are significantly altered after 1,4-GL and HFHC treatments. Colonic flora test results reveal that the quantity of Bif idobacterium and Lactobacillus in the intestines respectively increase by about 1.7- and 4.2-fold in rats treated with 1,4-GL compared with those in the control group. Rats that were coadministered with HFHC and 1,4-GL exhibit normal levels of lactate and acetate in serum and display urinary excretions of lactate and acetate that are 2 to 3 times higher compared with those treated with HFHC alone. The results imply that the increased probiotic quantities and urinary excretion of breakdown products of fat/cholesterol after 1,4-GL treatment contribute to the prevention of hypercholesterolemia. Our study offers insights into the model of action for 1,4-GL in preventing hypercholesterolemia. KEYWORDS: D-glucaro-1,4-lactone, hypercholesterolemia prevention, metabonomics, gut bacteria creatinine clearance
■
INTRODUCTION A long-term high-fat diet poses a substantial cardiovascular risk due to its association with several common diseases such as fatty liver and atherosclerosis,1,2 which lead to one of the major causes of morbidity and mortality in developed countries. It is well known that an elevated circulating cholesterol level is an independent risk factor for the development of atherosclerosis.3 Statins are the most widely prescribed drugs that are used to treat hypercholesterolemia. As inhibitors of 3-hydroxy-3methylglutaryl-coenzyme A (HMG-CoA) reductase, statins block cholesterol biosynthesis and lower serum low-density lipoprotein cholesterol (LDL-c). However, some patients have a low tolerance to statins. Moreover, some patients under statin treatment alone do not achieve the goal of reduction of plasma LDL-c.4 Therefore, natural products have gained attraction in the treatment and prevention of hypercholesterolemia as well as in the control of related metabolic diseases. D-Glucarate (GT), the precursor of D-glucaro-1,4-lactone (1,4-GL), is a natural substance found in fruits and vegetables.5 1,4-GL is a specific competitive inhibitor of β-glucuronidase, an enzyme produced by colonic microflora and involved in phase II liver detoxification. The enzyme increases the glucuronidation and excretion of potentially toxic compounds, such as exogenous estrogen.6 Studies have established that oral supplementation of 1,4-GL and GT help control the progressions of breast, prostate, and colon cancers at particular stages, and they exhibit antiproliferative effects.7 Recent studies reveal that 1,4-GL could ameliorate renal dysfunction in diabetic rats,8 and treatment of diabetic rats with it could be a © 2014 American Chemical Society
promising approach in lessening diabetes-mediated hepatic tissue damage.9 Indeed, in an aqueous solution, GT converts into 1,4-GL10,11 (Figure 1), and formation of 1,4-GL is regarded as the prerequisite for their positive benefits on health.12 In addition, GT significantly reduces total serum cholesterol by as much as 12% and LDL-cholesterol by 35% in rats;5 similar results have been obtained in humans.13 No
Figure 1. Chemical structure of D-glucaro-1,4-lactone and D-glucaric acid. Received: Revised: Accepted: Published: 7444
January 2, 2014 June 21, 2014 June 27, 2014 June 27, 2014 dx.doi.org/10.1021/jf501744d | J. Agric. Food Chem. 2014, 62, 7444−7451
Journal of Agricultural and Food Chemistry
Article 1
H NMR Spectroscopy of Urine Samples. Urine samples were prepared for NMR spectroscopy by diluting 300 μL of rat urine with 200 μL of 0.2 M sodium phosphate buffer (pH 7.4) and 50 μL of TSP in D2O to a final concentration of 0.1 mM to calibrate the chemical shift. NMR spectroscopy was carried out using a Bruker AvanceII-600 MHz spectrometer (Germany). All spectra were recorded at 298 K. 1D proton spectra were obtained from 32 scans over a spectral width of 14 ppm. The presaturation method was used to suppress the proton signal of the solvent. All 1H NMR spectra were processed using Topspin 2.1 (Bruker Biospin GmbH, Rheinstetten, Germany). The spectra were processed to a size of 64 k after multiplying with an exponential window function. The spectra were phase and baseline corrected manually. The chemical shifts were referenced to the TSP signal. Full-resolution spectra from 0.2 to 9.5 ppm were saved. Data from 4.5 to 6.5 ppm were excluded to eliminate the effects of imperfect water suppression and cross-saturation of urea. Each spectral intensity data set was normalized to the total spectral intensity over the whole spectrum. The metabolites were identified in the 1H NMR spectra by comparing their chemical shifts and coupling patterns with the corresponding values from the literature and publicly accessible databases (http://www.bmrb.wisc.edu, http://www.hmdb.ca). Total correlation spectroscopy (TOCSY), 13C-heteronuclear single-quantum correlation (13C-HSQC), statistical total correlation spectroscopy (STOCSY), and some standard compounds were also utilized to identify the metabolites. Data Processing. Principal component analysis (PCA) and orthogonal partial least-squares discriminant analysis (OPLS-DA) were performed using SIMCA-P (version 12.0, Umetrics, Umeå, Sweden). The quality of OPLS-DA models was described by R2X, R2Y, and Q2. The parameters were calculated using the default leave 1/7 out cross validation. R2X and R2Y represent the fraction of the sum of squares for the selected component. Q2 represents the predictive ability of the model. To investigate model predictability and data overfitting, a 200-iteration permutation test was performed in the PLSDA model with the same number of components as that in the OPLSDA model.19 Score and loading plots combined with variable importance in the projection (VIP) were used to interpret the model. Gut Microflora Cultivation. After thawing at room temperature, the samples taken from colonic contents were accurately weighed and immediately homogenized using glass beads into a sterile saline solution (1.0 g/10.0 mL). The solution was serially diluted in 10-fold steps. Then 0.1 mL of the aliquots of the diluents (10- and 100-fold) was inoculated in the middle of agar plates containing eosin methylene blue agar, Enterococcus selective agar, Lactobacillus selective agar, and BBL agar media. The sample was spread over the surface of the plate by using an L-rod. Each diluent was inoculated in duplicate. Escherichia coli was aerobically incubated at 37 °C for 48 h. Lactobacillus and Enterococcus were incubated at 37 °C for 48 h in a microaerophilic environment; Bif idobacterium was incubated in an anaerobic glovebox at 37 °C for 72 h. Bacterial colonies were counted after incubation. Quantification of Significant Metabolites in Urine and Serum. The contents of lactate, acetate, formate, creatinine, and bile acid in urine samples were quantified based on their individual 1H NMR spectra. At least six concentrations of the standard solutions were analyzed. The calibration curves were constructed by plotting the ratio of the analyte peak area to TSP area against each analyte. The contents of lactate and creatinine in serum were quantified using commercial kits from the Jiancheng Biotechnology Institute (Nanjian, China). The acetate in serum was measured using commercial kits from R-Biopharm (Germany). Creatinine clearance (CC) was calculated using the formula given as follows: CC = Urine creatinine concentration × Urine volume (24 h)/ Serum creatinine concentration.
evidence of toxicity to GT has been reported. Patients have shown tolerance to drug complaints of gastrointestinal distress, even at high doses (72 g/kg body weight).14 GT derivatives may thus be useful as additional or adjuvant preventive and therapeutic agents for hypercholesterolemia. The lipid-lowering effect of GT derivatives is considered to be associated with the increased excretion of bile acids.13 However, the metabolic effects of 1,4-GL in hypercholesterolemia are yet to be reported. Metabonomics, an emerging subject of the postgenome era, deals with the quantitative understanding of metabolites in biological fluids and their dynamic responses to changes in endogenous and exogenous factors. Metabonomics can thus provide the information on what is actually happening. Furthermore, this technology adopts a “top-down” strategy to determine the function of organisms from the terminal symptoms of the metabolic network and to elucidate the metabolic changes in a complete system caused by interventions in a holistic context.15 Therefore, metabonomics provides important information on the possible model of action of nutritional agents.16,17 The present study reveals that 1,4-GL significantly prevents or delays the onset and development of hypercholesterolemia when rats were coadministered with a high-fat, high-cholesterol (HFHC) emulsion. The metabolic mechanism of 1,4-GL in hypercholesterolemia prevention was quantitatively investigated via nuclear magnetic resonance (1H NMR)-based metabonomic analysis and colonic flora tests.
■
MATERIALS AND METHODS
Chemicals. 1,4-GL and the sodium salt of (trimethylsilyl)propionic-2,2,3,3-d4 acid (TSP) were purchased from Sigma (USA). All chemicals used were of analytical reagent grade. The HFHC emulsion was prepared as previously described,18 and it was composed of 20% lard, 10% cholesterol, and 2% sodium cholate. Animal Experiment. Forty female Sprague−Dawley (SD) rats weighing 180−200 g were randomly divided into five groups (n = 8 each) after acclimating to the environment for 7 d. The control group received water via oral gavage for 12 d once daily. Rats in the HFHC group were administered with 1.0 mL of HFHC for 12 d once daily to induce hypercholesterolemia. The 1,4-GL treatment group received 1,4-GL (30.0 mg/kg) via oral gavage for 12 d once daily. The HFHC +1,4-GL1 and HFHC+1,4-GL2 groups were treated with HFHC (1.0 mL) in combination with 30.0 or 10.0 mg/kg 1,4-GL, respectively, via oral gavage daily for 12 d. All animals were housed in approved facilities by group at a controlled relative humidity (50−70%) and temperature (22 ± 2 °C). Water and regular rodent chow were available ad libitum. On the 13th day, all rats were placed in individual metabolism cages and fasted for 12 h. Urine and blood samples were collected after 24 h. The rats were immediately sacrificed following the treatment. Rat livers were removed and stored at −20 °C. All urine samples were immediately diluted after collection with H2O to 15.0 mL. About 1.2 mL of the aliquots of the solution were centrifuged at 8000 rpm for 10 min. The obtained supernatant was stored at −20 °C until analysis. Animal experiments were carried out in accordance with the Guidelines for Animal Experiments at Nanchang University (Nanchang, China). The protocol was approved by the Animal Ethics Committee at Nanchang University. Measurement of Lipids in Serum and Extraction of Liver. In this experiment, 0.2 g of liver tissue was homogenized. The lipid was extracted by 4.0 mL of 2-propanol. The lipid levels in the serum and liver extracts, which include total cholesterol, triglyceride, and highand low-density lipoprotein cholesterol, were measured in duplicate by using commercial kits from Jiancheng Biotechnology Institute (Nanjian, China).
■
RESULTS HFHC-Induced Hypercholesterolemia in Rats. Animal models have provided fundamental contributions to the analysis of the onset and progression of complex multifactorial
7445
dx.doi.org/10.1021/jf501744d | J. Agric. Food Chem. 2014, 62, 7444−7451
Journal of Agricultural and Food Chemistry
Article
Table 1. Serum (and/or Liver) Lipid Levels for Various Treatmentsa biochemical indexes LDL-C (mmol/L) HDL-C (mmol/L) TC (mmol/L) (liver, μmol/g) TG (mmol/L) (liver, μmol/g)
control 0.70 1.76 3.06 6.04 0.82 0.21
± ± ± ± ± ±
0.09 0.08 0.26 1.21 0.08 0.02
1,4-GL1 treatment 0.94 1.73 3.08 6.34 0.93 0.18
± ± ± ± ± ±
0.16 0.10 0.18 2.11 0.10 0.04
HFHC group 5.45 1.68 6.51 15.3 1.58 0.78
± ± ± ± ± ±
0.74b 0.14 0.09b 2.41b 0.56 0.04b
HFHC coadministered with 1,4-GL1 1.78 1.45 3.44 7.42 1.04 0.38
± ± ± ± ± ±
0.39c 0.15 0.29c 2.12c 0.15 0.01c
HFHC coadministered with 1,4-GL2 2.58 1.68 4.32 10.6 0.95 0.53
± ± ± ± ± ±
0.33c 0.13 0.32c 1.33c 0.17 0.04c
1,4-GL1 and 1,4-GL2 respectively denote the doses of 1,4-GL, which are 30.0 and 10.0 mg/kg. The values are expressed by mean ± SEM, n = 8. Abbreviations: LDL-C: low-density lipoprotein cholesterol; HDL-C: high-density lipoprotein cholesterol; TC: total cholesterol; TG: triglyceride. b Significant difference from the control group (p < 0.05, one-way ANOVA). cSignificant difference from the HFHC group (p < 0.05, one-way ANOVA). a
Figure 2. Results of PCA and (O)PLS-DA generated by data from 1H NMR spectroscopy of rat urine. (a) PCA score plot; the variables were scaled by unit variance (UV). ●: HFHC group; ■: HFHC coadministered with 1,4-GL (30.0 mg/kg); ○: control group; ▲: 1,4-GL treatment (30.0 mg/ kg). ★: Mean score of t[1] (or t[2]) for each group. (b) Statistical validation of the OPLS-DA model by PLS-DA for the HFHC group and the control group. (c) Statistical validation of the OPLS-DA model by PLS-DA for the 1,4-GL treatment group and the control group.
diseases, such as cardiovascular diseases.20 Animal models have smaller interindividual variation in their genetic and metabolic factors in each strain than in humans. In addition, experiments can be carried out at the same environmental and feeding conditions. In this study, the rats were administered with HFHC18 or coadministered with 1,4-GL for 12 successive days. Table 1 reveals that the biochemical indicators of hypercholesterolemia, particularly serum LDL-c and total cholesterol (TC), significantly increase in the HFHC group compared with those in the control group. Furthermore, notable increases in TC and triglyceride (TG) in the liver are observed in the HFHC group. The dose-dependent decreases in LDL-c, TC, and TG in the serum are detected upon coadministration of HFHC with 1,4-GL (10.0 and 30.0 mg/kg), which suggests strongly that 1,4-GL prevents or delays the onset and development of HFHC-induced hypercholesterolemia. 1 H NMR-Based Metabonomic Analysis for Rat Urine. A 1 H NMR-based metabonomic experiment was carried out to explore the possible mechanisms of 1,4-GL in preventing the
development of HFHC-induced hypercholesterolemia at metabolic levels. PCA is a statistical approach to facilitate an understanding of the relationships of a number of complex objects in a multivariate data set. The resulting data were displayed as “score plots”, which represent the distribution of samples in multivariate space. The score plots of the first two principal components allowed visualization of the data and to establish whether there was any intrinsic difference in the 1H NMR-based metabolic profiling of rat urine. The PCA score plot (Figure 2a) shows that the HFHC group is isolated from other groups; however, the group of HFHC coadministered with 1,4-GL is closer to the normal groups. The result indicates that 1,4-GL restores the urinary 1H NMR phenotype that is disturbed by HFHC-induced hypercholesterolemia. OPLS-DA, a supervised model, can be considered as a modification of the traditional PLS-DA, particularly in extracting information on changes in the metabolic profiling of samples. Thus, in this study, the 1H NMR data of rat urine were employed to establish pairwise OPLS-DA models of the 7446
dx.doi.org/10.1021/jf501744d | J. Agric. Food Chem. 2014, 62, 7444−7451
Journal of Agricultural and Food Chemistry
Article
Figure 3. VIP-coded loadings plot. The color scales (VIP values) show variable importance in the OPLS-DA projection generated by the urinary 1H NMR data of the HFHC and the control group (a: R2X, 0.878, R2Y, 1.0; Q2, 0.89) or the 1,4-GL treatment and the control group (b: R2X, 0.72, R2Y, 0.97; Q2, 0.73). Metabolite keys: 1, lactate; 2, alanine; 3, acetate; 4, pyruvate; 5, succinate; 6, 2-oxoglutarate; 7, citrate; 8, TMA; 9, creatinine; 10, TMAO; 11, hippurate; 12, formate; 13, cholate.
1,4-GL Modifies Gut Flora Intervened by HFHC. Figure 3 reveals the significant downregulation and upregulation of trimethylamine-N-oxide (TMAO) in the HFHC and 1,4-GL groups, respectively, upon comparison with the control group. The excretion of trimethylamine (TMA) is also found to be opposite the alteration by HFHC compared with the 1,4-GL treatment. Metabolites (e.g., TMAO and TMA) are directly related to the statuses of gut microbiota and are urinary biomarkers of the alteration of gut flora.22,23 We therefore presume the differences between the effects of 1,4-GL treatment and HFHC on gut flora. To clarify the hypothesis, the contents in rat colon underwent aseptic manipulation to culture the enterobacteria. The results show that the levels of gut bacteria, such as Bif idobacterium, Lactobacillus, and Enterococcus, markedly decrease in the HFHC group compared with the control group (Figure 4). Similar results to Bif idobacterium species in high-fat-feeding mice have been previously reported.24,25 However, the quantities of Bif idobacterium, Lactobacillus, and Enterococcus significantly increase when the rats with hypercholesterolemia were coadministered with 1,4-GL. 1,4-GL significantly increases the quantities of Bif idobacterium (1.7-fold) and Lactobacillus (4.2-fold) but
HFHC group (and the 1,4-GL treatment group) to the control group, aiming to single out statistically and potentially significant variables (metabolites) responsible for the difference between the HFHC group (and 1,4-GL treatment group) and the control group. Successful discrimination is achieved with a goodness-of-prediction Q2 of more than 0.70. The OPLS-DA models reveal that the HFHC and 1,4-GL treatment groups are statistically distinguishable from the control group (p = 0.031 and p = 0.018, respectively; CV-ANOVA). The validation plot from the PLS-DA model (Figure 2b and c) demonstrates that the original OPLS-DA model is not random and overfitted because the permutated R2 and Q2 values are significantly lower than the corresponding original values; hence, the OPLS-DA model is robust. A bivariate plot, which combines the loading coefficients and VIP, was generated based on the OPLS-DA model21 (Figure 3). The plot characterizes the variation of a peak with changes in intensity and its importance to classification (color code). Upward- and downward-oriented peaks in the bivariate plot reflect the increased and decreased intensities from the intervention of HFHC or 1,4-GL, respectively. Figure 3 lists the important metabolites for classification. 7447
dx.doi.org/10.1021/jf501744d | J. Agric. Food Chem. 2014, 62, 7444−7451
Journal of Agricultural and Food Chemistry
Article
The quantitative results (Table 2) also show that the concentrations of lactate and acetate in serum fall to normal levels when the rats were coadministered with HFHC and 1,4GL. Meanwhile, urinary excretions of lactate and acetate after rats were treated with HFHC and 1,4-GL were found to be about 2 to 3 times higher than those treated with HFHC alone. Kidneys excrete waste products from the body through urine and regulate the concentration of metabolites in the blood. The data indicate that the increased urinary excretion of lactate and acetate by 1,4-GL treatment contributes to the prevention of hypercholesterolemia. Increased Creatinine Clearance after 1,4-GL Treatment. Metabonomic and quantitative results likewise reveal that 1,4-GL treatment significantly increases the urinary excretion of creatinine. Creatinine is constantly produced at a rate proportional to the muscle mass; this substance is then filtered from the blood by the kidneys. In this study, no statistical differences were observed in rat weight between different groups during the experiment. This result suggests that the increased urinary excretion of creatinine during 1,4-GL treatment is not attributed to changes in muscle mass. The excretion of creatinine is primarily determined by glomerular filtration and proximal tubular secretion. Therefore, creatinine levels in blood and urine can be used to calculate the creatinine clearance, which reflects renal function.26 Our results show that the creatinine clearance of 1,4-GL treatment exceeds that of the control group by 77.4%. Creatinine clearance was reduced by 33.6% in the HFHC group and was not statistically significant in the group coadministered with HFHC and 1,4-GL compared with that of the control group (Table 2). The increased creatinine clearance after 1,4-GL treatment increases the urinary elimination of metabolites (e.g., lactate and acetate) and maintains the concentration of metabolites in the blood.
Figure 4. Levels of bacterial flora in rat colon after various interventions. CFU: colony forming units. The bar graph represents the mean ± SE for samples derived from five rats in each group. *Significant difference from the control group (p < 0.05, one-way ANOVA). #Significant difference from the HFHC group (p < 0.05, one-way ANOVA).
reduces that of E. coli (31.2-fold) in 1,4-GL-treated rats (Figure 4). The findings indicate that 1,4-GL restores the HFHCinduced imbalance of gut flora and affects their conversion to probiotics (e.g., Bifidobacterium and Lactobacillus). Elevation of Serum Lactate and Acetate in Hypercholesterolemic Rats. NMR-based metabonomic results (Figure 3) indicate that the alterations of particular metabolites (e.g., lactate, creatinine, acetate, and formate) in urine were significantly opposite the direction of regulation caused by HFHC intervention compared with 1,4-GL treatment. To comprehend the possible biological effects of 1,4-GL in preventing hypercholesterolemia, the contents of the key metabolites were determined in the urine and serum samples (Table 2). The HFHC group exhibits higher serum (lactate: 1.77-fold, p < 0.01; acetate: 1.50-fold, p < 0.01) and urinary excretion (lactate: 1.47-fold, p > 0.05; acetate: 4.87-fold, p < 0.01) of lactate and acetate than the control group.
■
DISCUSSION Mammals are “superorganisms” with karyomes, chondriomes, and microbiomes. Research has suggested that the microbiome−host relationship contributes to diabetes, obesity, and increased blood pressure.27,28 The results from metabonomic analysis show that particular urinary biomarkers (e.g., hippurate, TMAO, and TMA) of the changes in gut microbiota are significantly altered after 1,4-GL treatment. Thus, counts of some general gut bacteria in rat colonic contents were
Table 2. Quantitative Results of Particular Metabolites in the Urine (from 0 to 24 h) and Serum of Rats (n = 6−8 in Each Group, Mean ± SE)a metabolite
control
HFHC group
HFHC coadministered with 1,4-GL
1,4-GL treatment
lactate (U, μmol) acetate (U, μmol) formate (U, μmol) creatinine (U, μmol) cholate (U, μmol) lactate (S, μmol/L) acetate (S, μmol/L) creatinine (S, μmol/L) CC (mL/h)
271 ± 48.3 299 ± 40.6 81.2 ± 8.90 13.5 ± 1.30 not detected 189 ± 20.9 15.1 ± 1.40 30.5 ± 2.90 20.8 ± 3.30
± ± ± ± ± ± ± ± ±
± ± ± ± ± ± ± ± ±
146 ± 37.2b 179 ± 50.7b 72.4 ± 25.5 22.2 ± 2.10b not detected 167 ± 18.7 15.3 ± 1.50 26.7 ± 2.70 36.9 ± 3.40c
397 1461 540 12.5 19.4 335 23.8 40.4 13.8
79.5 292c 135c 2.30 1.70 47.1c 1.20c 5.70 3.0b
750 4384 2158 13.9 28.5 219 15.5 30.7 19.8
c
126 608c 431c 3.20 3.60d 31.5 1.50 3.45 2.20
a
Abbreviations: U, urine; S, serum; CC, creatinine clearance; control, rats were administered with water; HFHC, rats were administered with highfat, high-cholesterol emulsion; HFHC coadministered with 1,4-GL, rats were administered with 1,4-GL (30.0 mg/kg) and administered with highfat, high-cholesterol emulsion afterward; 1,4-GL treatment, rats were administered with 1,4-GL (30.0 mg/kg). bSignificant difference from the control group by p < 0.05 (one-way ANOVA). csignificant differences from the control group by
