Caloric restriction of db/db mice reverts hepatic steatosis and body weight with divergent hepatic metabolism

Non-alcoholic fatty liver disease (NAFLD) is one of the most frequent causes of liver disease and its prevalence is a serious and growing clinical problem. Caloric restriction (CR) is commonly recommended for improvement of obesity-related diseases such as NAFLD. However, the effects of CR on hepatic metabolism remain unknown. We investigated the effects of CR on metabolic dysfunction in the liver of obese diabetic db/db mice. We found that CR of db/db mice reverted insulin resistance, hepatic steatosis, body weight and adiposity to those of db/m mice. 1H-NMR- and UPLC-QTOF-MS-based metabolite profiling data showed significant metabolic alterations related to lipogenesis, ketogenesis, and inflammation in db/db mice. Moreover, western blot analysis showed that lipogenesis pathway enzymes in the liver of db/db mice were reduced by CR. In addition, CR reversed ketogenesis pathway enzymes and the enhanced autophagy, mitochondrial biogenesis, collagen deposition and endoplasmic reticulum stress in db/db mice. In particular, hepatic inflammation-related proteins including lipocalin-2 in db/db mice were attenuated by CR. Hepatic metabolomic studies yielded multiple pathological mechanisms of NAFLD. Also, these findings showed that CR has a therapeutic effect by attenuating the deleterious effects of obesity and diabetes-induced multiple complications.


Blood glucose and the insulin tolerance test (ITT)
Blood glucose was measured every two weeks using an Accu-Chek glucometer (Roche Diagnostics GmbH, Mannheim, Germany). For the ITT, mice were given intraperitoneal injections of insulin (0.75 U/kg, Humulin-R; Eli Lilly, Indianapolis, IN, USA), and blood samples were collected before and 15, 30, 45, and 60 minutes after the injection. Blood glucose was measured using an Accu-Chek glucometer (Roche Diagnostics).

Sircol collagen assay
The Sircol collagen assay is a quantitative dye-binding method for the analysis of acid and pepsin-soluble collagens. The collagen concentration from frozen liver tissues was determined by using a Sircol assay kit (Bioclor Ltd., Northern Ireland, UK).

Immunofluorescence
For immunofluorescence staining, deparaffinized sections from the liver were incubated with goat anti-LCN2 or mouse anti-NF-Kb at 4°C for one day. After washing three times with 0.1 M PBS, sections were incubated with Alexa Fluor 594-conjugated donkey anti-goat or mouse secondary antibody (Invitrogen, Carlsbad, CA, USA). Nuclei were stained with DAPI (1:10,000, Invitrogen). Fluorescence was visualized under a BX51-DSU microscope (Olympus).

Metabolite analysis of liver tissue
For analyzing both aqueous and organic metabolites, we followed a slightly modified two-step solvent extraction [1]. Frozen liver tissues (80 mg) were loaded into appropriate bead beating tubes. Then, 1.5 ml 4 of pre-chilled methanol:water (1:1, v:v) was added. For extraction, samples were homogenized twice at 5,000 rpm for 10 min and centrifuged at 16,000 g for 10 min. After centrifuging, the supernatant of the aqueous phase was transferred to an Eppendorf tube and completely dried using a speed vacuum. Then, 1.6 ml of dichloromethane:methanol (1:1, v:v) were added to remaining solid precipitates. Samples were homogenized as before, followed by centrifugation. Then, the supernatant of the organic phase was transferred to a new tube and dried under a stream of oxygen-free N2.

H-NMR measurements of polar metabolites
For the NMR experiment, each extract was dissolved in a 700-μl buffer solution (0.1 M phosphate buffer and pH 7.0), and the pH was adjusted to 7.0 0.1. A 1-mM solution of 4,4-dimethyl-4-silapentane-1sulfonic acid (DSS) dissolved in 99.8% D2O was added and a 600-μl aliquot was placed in a 5-mm NMR tube (Wilmad Lab Glass, Buena, NJ). The 1 H-NMR spectra were acquired on a VNMRS-600 MHz NMR spectrometer (Agilent Technologies, Inc., Santa Clara, CA, USA) using a triple resonance 5-mm HCN salt-tolerant cold probe. The water-suppressed Carr-Purcell-Meiboom-Gill spin-echo pulse sequence [RD-90°-( -180°-) n-ACQ] with a total T2 relaxation time of 60 ms was used to attenuate broad signals from metabolites. For each sample, the 1 H-NMR spectrum was collected with 128 transients into 65K data points using a spectral width of 12019.2 Hz with a relaxation delay of 2.0 s and an acquisition time of 2.7263 s. Free induction decays were weighted by an exponential function with a 0.3-Hz line-broadening factor prior to Fourier transformation. Signal assignment for representative samples was facilitated by acquisition of two-dimensional (2D) total correlation spectroscopy (TOCSY), correlation spectroscopy (COSY), H J-resolved (JRES), spiking experiments, and comparisons to literature.

NMR data processing
All spectra were phase-adjusted and baseline-corrected using the Chenomx NMR software suite (version 7.1, Chenomx, USA). The spectral region from = 0.12 to 9.35 was segmented into 0.001-ppm-wide 5 regions using the software. The regions corresponding to water (4.58-5.05 ppm), DSS (0.20-0.655 ppm, 1.645-1.808 ppm), lipid (1.246-1.360 ppm) and methanol (3.277-3.366 ppm) were excluded. Then, spectra were normalized to the total spectral area and converted to the ASCII format. ASCII files were imported into MATLAB (R2006a; Mathworks, Inc., Natick, MA, USA), and all spectra were aligned using the correlation optimized warping method. Identified metabolites were quantified using the Chenomx NMR software suite.

UPLC-QTOF-MS measurements of lipid metabolites
Dry residues of lipid extracts were dissolved in 200 μl of water:methanol (1:1, v:v) and filtered through the 0.2 μm hydrophilic PTFE syringe filter (MILLEX ® , MILLIPORE). Then, 100-μl samples were diluted using 300 μl of isopropyl alcohol. An Acquity UPLC system (Waters, Maidstone, UK) including a binary solvent manager, column heater, and photodiode array detector was coupled with a hybrid Q-TOF tandem mass spectrometer (ESI/Triple TOF 5600; AB Sciex, Concord, ON, Canada). LC separation was carried out on BEH-C18 columns (100 mm × 2.1 mm, particle size 1.7 μm; Waters) on the reverse phase mode for 29 min. Column temperature and flow rate were set to 40°C and 0.35 ml/min, respectively. The mobile phase used was 10 mM ammonium acetate in acetonitrile:water (4:6, v:v) (A), and 10 mM ammonium acetate in acetonitrile:isopropyl alcohol (1:9, v:v) (B). The gradient elution program is as follows: 40-65% B at 0-5 min, 65-75% B at 5-20 min, 75-99% B at 20-25 min, 99% B for 2 min, 40% B at 27-27.1 min, and maintained for 1.9 min. Data acquisition was performed with a Triple TOF 5600 system with Turbo V sources and a Turbo ion spray interface. TOF-MS data were acquired both in positive and negative ion modes using an ion spray voltage of 4500 V. The nebulizer gas (Gas 1), heater gas (Gas 2), and curtain gas were set to 50 psi, 60 psi, and 30 psi, respectively. The Turbo spray temperature was 500°C and the declustering potential was 90 V. Information-dependent acquisition (IDA) experiments were conducted to obtain single MS information and MS/MS information. TOF-MS and product ion calibrations were performed every day in both high-sensitivity and high-resolution modes 6 using an automated calibrant delivery system (CDS) before analysis. To check the reproducibility of the data, pooled quality control (QC) samples were measured per nine samples.

UPLC-QTOF-MS data processing
UPLC-QTOF-MS data from the liver extractions were processed by using MarkerView software version 1.2.1.1 (SCIEX, concord, ON, Canada) for peak finding and peak alignment. Data collection parameters in peak finding were set as follows: mass range, 100-1300 m/z; retention time range, 1.5-25.3 min; subtraction offset, 10 scans; subtraction multiplier, 1.8; minimum spectral peak width, 1 ppm; retention time peak width, 5 scans. For peak alignment, retention time and mass tolerance were set to 0.2 min and 10 ppm, respectively. The intensity threshold for peak filtering was set to 100, and peaks detected in fewer than four samples were removed. Peak area matrices were normalized by batch normalization to remove the systematic variation among the samples. The dataset resulted in peak lists containing the accurate mass of the precursor ion, retention time, and peak area of metabolites. To decrease instrumental bias including noise and variation, we eliminated peaks having a large variation in QC (coefficient of variation (CV)>20%). All of the variables were tentatively identified by connecting free accessible metabolite database such as HMDB (http://www.HMDB.ca) and Lipid Maps (http://www.lipidmaps.org) and comparing the literature.

Statistical analysis
A multivariate statistical analysis was performed using SIMCA-P+ software (version 12.0, Umeå, Sweden). Partial least-squares discriminant analysis (PLS-DA) was conducted for model discrimination.
Score plots, loading plots, and variable importance of projection (VIP) values were obtained from the PLS-DA model. VIP>1 was considered to designate the most important metabolites responsible for the differentiation of groups. Student's t-tests were performed using GraphPad Prism version 5.0 (GraphPad 7 Software, Inc., La Jolla, CA) to test the significance of differences in metabolite levels among groups.
The differences were tested to a 95% probability level (p<0.05). For multiple comparisons, Bonferroni's corrections were applied (p<0.017). Differences in protein levels between groups were determined by one-way ANOVA, followed by post hoc analysis by using the Bonferroni's Multiple Comparison test.
Results are presented as mean ± standard error of the mean (SEM), and p<0.05 was considered significant.    Band intensity was normalized to β-actin or elF2α. Data are shown as the mean ± SEM. *p<0.05 for db/db versus db/m mice. †p<0.05 for db/db+CR versus db/db mice.