Myocardial metabolic alterations in mice with diet-induced atherosclerosis: linking sulfur amino acid and lipid metabolism

Atherosclerosis is a leading cause of cardiovascular disease (CVD), but the effect of diet on the atherosclerotic heart’s metabolism is unclear. We used an integrated metabolomics and lipidomics approach to evaluate metabolic perturbations in heart and serum from mice fed an atherogenic diet (AD) for 8, 16, and 25 weeks. Nuclear magnetic resonance (NMR)-based metabolomics revealed significant changes in sulfur amino acid (SAA) and lipid metabolism in heart from AD mice compared with heart from normal diet mice. Higher SAA levels in AD mice were quantitatively verified using liquid chromatography-mass spectrometry (LC/MS). Lipidomic profiling revealed that fatty acid and triglyceride (TG) levels in the AD group were altered depending on the degree of unsaturation. Additionally, levels of SCD1, SREBP-1, and PPARγ were reduced in AD mice after 25 weeks, while levels of reactive oxygen species were elevated. The results suggest that a long-term AD leads to SAA metabolism dysregulation and increased oxidative stress in the heart, causing SCD1 activity suppression and accumulation of toxic TGs with a low degree of unsaturation. These findings demonstrate that the SAA metabolic pathway is a promising therapeutic target for CVD treatment and that metabolomics can be used to investigate the metabolic signature of atherosclerosis.


H NMR-based metabolomic analysis
For serum samples, 90 μL of serum was mixed with 510 μL of saline solution (0.9% w/v sodium chloride in deuterium oxide) and transferred to 5-mm NMR tubes. One-dimensional 1 H NMR spectra of serum were acquired at 298 K on a Bruker Avance III HD 800 MHz NMR spectrometer (Bruker BioSpin, Germany) with a Bruker 5 mm CPTCI Z-GRD probe. The water-suppressed CPMG spin-echo pulse sequence (RD-90˚-[τ-180˚-τ] n-ACQ) was used to attenuate broad signals from proteins and lipoproteins with total T2 filter time of 32 ms. For all spectra of each serum sample, 128 transients were acquired with 64 k data points, a spectral width of 16,025.641 Hz, and the relaxation delay of 4 s, and acquisition time of 2.045 s.
For heart tissues, 10-12mg of each sample was transferred into a disposable insert tube. The residual space in the tube was filled with deuterium oxide (99.9 atom % D). And then, the disposable insert tube was inserted into a zirconium HR-MAS rotor (4mm outer diameter). The samples were maintained in a refrigerated tray at -20˚C before the analysis. All HR-MAS 1 H NMR spectra of heart sample were acquired on a Bruker Avance III 700 MHz NMR spectrometer (Bruker BioSpin, Germany) with a Bruker 4 mm TXI HR-MAS probe with z-gradients at 278 K and a spinning rate of 6 kHz using CPMG spin-echo pulse sequence. For all spectra of each heart sample, 128 transients were acquired with total T2 filter time of 37.8 ms, 32 k data points, a spectral width of 14,097.744 Hz, and the relaxation delay of 4 s, and acquisition time of 1.162 s.
Free induction decays were weighted by an exponential function with a 0.3 Hz linebroadening factor prior to Fourier transformation. All acquired 1 H NMR spectra were phase-and baseline-corrected using TopSpin 3.1 software (Bruker BioSpin, Germany) and Chenomx NMR Suite Version 7.1 (Chenomx, Canada). The chemical shift was referenced to the proton signals of lactate at 1.32 ppm and formate at 8.45 ppm. Resonance assignments for serum and heart metabolites were accomplished using the 800 MHz library of Chenomx NMR Suite Version 7.1 (Chenomx, Canada) and literature reports [1][2][3][4] . A set of 2D NMR spectra (Fig. S7-9) was acquired on some representative serum and heart samples and spiked experiments were performed to confirm the metabolite identities.
Representative 800 MHz 1 H NMR spectra of sera and 700 MHz HR-MAS 1 H NMR spectra of heart tissues were shown in Fig. S10 and S11, respectively. The complete resonance assignment is listed in Table S3.
Concentration of 4,4-dimethyl-4-silapentane-1-sulfonic acid (DSS) was estimated using ERETIC2 (Electronic Reference To access In vivo Concentrations) module in TopSpin software 3.1 (Bruker BioSpin, Germany), which is based on the PULCON principle (PULse Length based CONcentrations determination) 5,6 . Targeted metabolic profiling of serum and heart tissue were performed using Chenomx NMR Suite Version 7.1 (Chenomx, Canada) by integrating peak areas of metabolites compared with the areas of the DSS peak. For heart metabolites, tissue mass (g) was used to normalize metabolite levels.

Measurement of metabolites in sulfur amino acid metabolism
To determine concentration of sulfur amino acids, previously used methods were modified and adopted 7,8 . In brief, to extract heart metabolites, 20 mg heart were transferred to each 1.5 mL microcentrifuge tube. 500 µL of 50 % aqueous methanol (v/v) containing 0.1 % formic acid and 0.05 % trifluoroacetic acid were added and vigorously mixed for 30 s. Homogenization with 2.8 mm zirconium oxide beads was performed at 5000 rpm twice using a Precellys 24 tissue grinder (Bertin Technologies, France). 250 µL of chloroform was added to each sample tube and mixed for 30 s. And then, the sample solution was incubated at 4 °C for 10 min and centrifuged at 4 °C and 13,000 rpm for 10 min. 400 µL of supernatant was transferred to new 1.5 mL tube and dried in a vacuum centrifuge.
The dried sample was dissolved in 50 µL of 30 % aqueous methanol (v/v) containing 0.1 % formic acid, treated with a reducing agent of 50 µL of 500 mM dithiothreitol (DTT) solution resolved in 0.075 M NaOH to cleave disulphide bonds and then stored for 1 h at room temperature. To remove dust and particle in sample, purification was conducted using 0.2 µm PTFE membrane filter, the sample was transferred to LC glass vial and 500 ng/mL betaine-D11, cysteine-D2, methionine-D3 and S-adenosylmethionine (SAM)-D3 as internal standards were added to the vial. All these compounds were purchased from Cambridge Isotope Laboratories (MA, USA). 2 µL of sample was injected for each LC-MS/MS analysis .   To quantify the level of metabolites, liquid chromatography-mass spectrometry was   performed on an Agilent 1290 Infinity LC and an Agilent 6490 Triple Quadrupole MS system   equipped with Agilent Jet Stream ESI source (Agilent Technologies, Table S4.

LC-MS/MS-based lipidomic analysis
After extraction of sulfur amino acid metabolites in heart tissues, the solution under pellet was transferred to a 1.5-mL Eppendorf tube for lipidomic analysis of heart. Lipid extract of heart tissues was evaporated under a stream of nitrogen, diluted with an isopropanol: acetonitrile: water mixture (2:1:1, v/v/v), and transferred into vials after centrifugation for 10 min at 13,000g and 4°C.
To obtain MS spectral data, LC-ESI-MS/MS analyses of lipid extracts were performed on a triple TOF™ 5600 MS/MS System (AB Sciex, Canada) combined with a UPLC system (Waters, USA). LC separations were carried out on an Acquity UPLC BEH C18 column (100 x 2.1 mm, particle size 1.7 µm, Waters, USA). Colum temperature and flow rate were set to 35 °C and 0.35 ml/min, respectively. The mobile phases used were 10 mM ammonium acetate in an acetonitrile: water mixture (40:60, v/v) (A) and acetonitrile: isopropanol mixture (10:90, v/v) (B). The linear gradients were as follows: 40-65% B for 5 min, 65-70% B for 7 min, 70-99% B for 3 min, 99% B for 2 min, 40% B for 3 min. The injection volume of the sample was 5 μL using partial loop mode for both positive and negative ionization polarity modes. Quality control (QC) samples, pooling identical aliquots of the samples, were measured regularly throughout the run for data reproducibility.
The spectral data were analyzed by MarkerView TM (AB Sciex, Canada), which was used to find peaks, perform the alignment, and generate peak tables of m/z and retention times (min). Spectra were normalized to the median fold-change normalization. To identify reliable peaks and remove instrumental bias, peaks whose intensity in QC sample is lower than those in a blank sample were eliminated, and peaks with coefficients of variation below 20 in QC sample were selected. Lipids were tentatively identified by comparing the experimental data against various databases, including the METLIN (metlin.scripps.edu), Human Metabolome (www.hmdb.ca), MassBank (www.massbank.jp), and Lipid Maps (www.lipidmaps.org) databases. Fragment patterns (MS/MS spectra) were also used to identify lipid metabolites.