Genomic and Metabolic Responses to Methionine ...

J Nutrigenet Nutrigenomics 2012;5:132–157 DOI: 10.1159/000339347 Received: January 24, 2012 Accepted: May 4, 2012 Published online: October 9, 2012

© 2012 S. Karger AG, Basel 1661–6499/12/0053–0132$38.00/0 www.karger.com/jnn

Original Paper

Genomic and Metabolic Responses to Methionine-Restricted and MethionineRestricted, Cysteine-Supplemented Diets in Fischer 344 Rat Inguinal Adipose Tissue, Liver and Quadriceps Muscle Carmen E. Perrone a Dwight A.L. Mattocks a Jason D. Plummer a Sridar V. Chittur b Rob Mohney c Katie Vignola c David S. Orentreich a Norman Orentreich a a Orentreich

Foundation for the Advancement of Science, Inc., Cold Spring-on-Hudson, N.Y., for Functional Genomics, State University of New York at Albany, Rensselaer, N.Y., and c Metabolon Inc., Durham, N.C., USA b Center

Key Words Methionine restriction ⴢ Cysteine supplementation ⴢ Liver ⴢ Inguinal adipose tissue ⴢ Muscle ⴢ Gene expression ⴢ Metabolism ⴢ qPCR ⴢ Microarrays ⴢ Mass spectrometry

Abstract Background/Aims: Methionine restriction (MR) is a dietary intervention that increases lifespan, reduces adiposity and improves insulin sensitivity. These effects are reversed by supplementation of the MR diet with cysteine (MRC). Genomic and metabolomic studies were conducted to identify potential mechanisms by which MR induces favorable metabolic effects, and that are reversed by cysteine supplementation. Methods: Gene expression was examined by microarray analysis and TaqMan quantitative PCR. Levels of selected proteins were measured by Western blot and metabolic intermediates were analyzed by mass spectrometry. Results: MR increased lipid metabolism in inguinal adipose tissue and quadriceps muscle while it decreased lipid synthesis in liver. In inguinal adipose tissue, MR not only caused the transcriptional upregulation of genes associated with fatty acid synthesis but also of Lpin1, Pc, Pck1 and Pdk1, genes that are associated with glyceroneogenesis. MR also upregulated lipolysis-associated genes in inguinal fat and led to increased oxidation in this tissue, as suggested by higher levels of methionine sulfoxide and 13-HODE + 9-HODE compared to control-fed (CF) rats. Moreover, MR Carmen E. Perrone, PhD

Cell Biology Laboratory 855 Route 301 Cold Spring-on-Hudson, NY 10516 (USA) Tel. +1 845 265 4200, ext. 235, E-Mail perrone @ orentreich.org

132

J Nutrigenet Nutrigenomics 2012;5:132–157 DOI: 10.1159/000339347

© 2012 S. Karger AG, Basel www.karger.com/jnn

Perrone et al.: Genomic and Metabolic Responses to Methionine-Restricted and Methionine-Restricted, Cysteine-Supplemented Diets in Fischer 344 Rat Inguinal Adipose Tissue, Liver and Quadriceps Muscle

caused a trend toward the downregulation of inflammation-associated genes in inguinal adipose tissue. MRC reversed most gene and metabolite changes induced by MR in inguinal adipose tissue, but drove the expression of Elovl6, Lpin1, Pc, and Pdk1 below CF levels. In liver, MR decreased levels of a number of long-chain fatty acids, glycerol and glycerol-3-phosphate corresponding with the gene expression data. Although MR increased the expression of genes associated with carbohydrate metabolism, levels of glycolytic intermediates were below CF levels. MR, however, stimulated gluconeogenesis and ketogenesis in liver tissue. As previously reported, sulfur amino acids derived from methionine were decreased in liver by MR, but homocysteine levels were elevated. Increased liver homocysteine levels by MR were associated with decreased cystathionine ␤-synthase (CBS) protein levels and lowered vitamin B6 and 5-methyltetrahydrofolate (5MeTHF) content. Finally, MR upregulated fibroblast growth factor 21 (FGF21) gene and protein levels in both liver and adipose tissues. MRC reversed some of MR’s effects in liver and upregulated the transcription of genes associated with inflammation and carcinogenesis such as Cxcl16, Cdh17, Mmp12, Mybl1, and Cav1 among others. In quadriceps muscle, MR upregulated lipid metabolism-associated genes and increased 3-hydroxybutyrate levels suggesting increased fatty acid oxidation as well as stimulation of gluconeogenesis and glycogenolysis in this tissue. Conclusion: Increased lipid metabolism in inguinal adipose tissue and quadriceps muscle, decreased triglyceride synthesis in liver and the downregulation of inflammation-associated genes are among the factors that could favor the lean phenotype and increased insulin sensitivity observed in MR rats. Copyright © 2012 S. Karger AG, Basel

Introduction

Decreasing methionine content in the diet from 0.86 to 0.17%, also known as methionine restriction (MR), prolongs lifespan and delays the onset of aging-related diseases [1, 2]. Significant reductions in body weight gain, adiposity and dyslipidemia as well as improvement of insulin sensitivity in rodents are hallmarks of MR [3, 4]. Insulin sensitivity in MR rats is associated with reduced fat accretion, a response involving metabolic adaptations that alter lipid metabolism [4]. Decreased adiposity is also associated with increased energy expenditure [3, 5] corresponding with the upregulation of peroxisome proliferator-activated receptor ␥ (PPAR␥), PPAR␥ coactivator 1␣ (PGC1a) and their target genes, which leads to mitochondrial biogenesis in adipose tissues and increased mitochondrial aerobic capacity of liver and quadriceps muscle from MR rats [6]. MR lowers serum levels of methionine and methionine-derived amino acids including cysteine [7]. In agreement with reports showing a positive association between cysteine levels and body mass index (BMI) in humans [8], the supplementation of the MR diet with cysteine was shown to reverse the MR-mediated effects on adiposity and serum parameters associated with adiposity suggesting that lowered cysteine levels contribute to MR’s effects on body composition [9]. Although cysteine supplementation was demonstrated to reverse MR’s effects on the transcription of the stearoyl-CoA desaturase 1 (Scd1) gene [9] which has been linked with obesity and insulin resistance [10], there is little insight regarding the other consequences of MR and, subsequently, lowered cysteine levels that dictate the lean phenotype and insulin sensitivity in F344 rats. Amino acids are building blocks for protein synthesis as well as regulators of signaling cascades that control key metabolic pathways through as yet unknown mechanisms [11]. High-throughput technologies have been proposed to identify molecular, physiological, and pathological effects produced by perturbations of dietary amino acid intake based on changes in DNA, RNA, proteins and metabolites [11]. In fact, RNA, protein and metabolite studies

133




can provide the capacity to measure subtle perturbations of pathways caused by nutrients [12, 13], and combined with mathematical models, can be used to delineate and quantify multiscale responses to nutrients ranging from whole body to subcellular levels of organization [14]. Omic studies were therefore conducted to gain a better understanding of the physiological responses triggered by MR that contribute to reduced adiposity and increased insulin sensitivity.

Materials and Methods Animal Husbandry, Tissue Collection and Serum Parameters Four-week-old male F344 rats obtained from Taconic Farms (Germantown, N.Y., USA) were maintained 1 rat per cage on a 12-hour light-dark cycle and fed Laboratory Rodent Diet 5001 (PMI Nutrition International LLC, Brentwood, Mo., USA) for 2 weeks. At 6 weeks of age, the rats were randomly assigned to chemically defined diets based on the AIN-76 diet with protein replaced by amino acid mixtures containing 0.86% methionine (control; CF), 0.17% methionine (methionine-restricted; MR) or 0.17% methionine + 0.5% cysteine (methionine-restricted, cysteine-supplemented; MRC; n = 8 per treatment group) [1, 15]. Food and water were provided ad libitum. After 3 months on these dietary regimens, the rats were anesthetized and blood was collected from the subclavian vein. The rats were then euthanized and inguinal adipose tissue, liver, and quadriceps muscle were immediately harvested, weighed, frozen in liquid nitrogen, and stored at –80 ° C. The rats were not fasted prior to tissue collection.

Gene Expression Analysis Total RNA was isolated with RNeasy Kits (Qiagen Inc., Valencia, Calif., USA), RNA concentrations were determined spectrophotometrically, and RNA integrity was assessed by agarose gel electrophoresis. Microarray analysis was conducted at the Center for Functional Genomics (Rensselaer, N.Y., USA) where the integrity of the RNA (n = 6 per tissue type and treatment group) was reassessed with an Agilent 2100 Bioanalyzer (Agilent Technologies, Inc., Santa Clara, Calif., USA). The RNA was used as a template for complementary DNA (cDNA) preparation using the WT-Ovation Pico RNA amplification kit (NuGEN Technologies, Inc., San Carlos, Calif., USA), and hybridized to Rat Exon ST 1.0 Arrays according to the manufacturer’s protocol (Affymetrix, Santa Clara, Calif., USA). The raw microarray data was quantile normalized in GeneSpring GX 11 using PLIER16, the log2 signal values were filtered to remove entities that showed signals in the bottom 20th percentile across all samples, and the results were subjected to ANOVA with a Benjamini-Hochberg False Discovery Rate Correction to identify genes with significant differences (p ^ 0.05). A 1.5-fold filter was applied to this ANOVA list to identify genes that were differentially expressed between any 2 conditions. Microarray CEL files were converted into annotated files using Expression Console (Affymetrix) and imported into Pathway Studio 8.0 (Ariadne Inc., Rockville, Md., USA). Enrichment analysis for biological processes was conducted for genes upregulated or downregulated by 1.5-fold and p ^ 0.05 using Fisher’s exact test. The results were displayed with p values indicating whether the imported list was enriched with members from various biological process categories. TaqMan quantitative PCR (qPCR) was conducted as described previously [6]. Total RNA (n = 7–8 per tissue and treatment group) was used for cDNA preparation with the High-Capacity cDNA Reverse Transcription Kit (Life Technologies Corp., Carlsbad, Calif., USA) in a Perkin-Elmer GeneAmp PCR System 9600 and the cDNA was then used in multiplexed qPCR in a StepOnePlus Real-Time PCR System (Life Technologies Corp.) using commercially available primer-probe sets (online supp. table S1; for all online suppl. material, see www.karger.com/doi/10.1159/000339347). Gene expression was assessed via the comparative CT (⌬⌬CT) method with ␤-actin as the reference gene. ␤-Actin CT values were approximately the same in tissues from CF, MR and MRC rats, and efficiencies for ␤-actin and the genes of interest in the multiplexed reactions were 187%. Protein Analysis Levels of selected proteins in tissues were determined by Western blot analysis or ELISA. For Western blot analysis, tissue homogenates (n = 6–8 per tissue and treatment group) were prepared as described previously [4, 6]. Fifty micrograms of protein were electrophoresed in 4–20% SDS-PAGE gradient gels, the

134



135


proteins were transferred to PVDF membranes, and the membranes were incubated overnight at 4 ° C with primary antibodies (Abs). The membranes were then incubated in corresponding secondary Abs and developed using the West Pico, West Femto, or Immun-Star Western C enhanced chemiluminescence substrates. Images from Western blots were captured using a Bio-Rad ChemiDoc XRS+ Molecular Imager and densitometry measurements of protein bands were conducted with Un-Scan-It Gel version 6.1 (Silk Scientific Corporation, Orem, Utah, USA). Protein levels were normalized using ␤-actin as the loading control. The DUSP4 Ab was from Abcam (Cambridge, Mass., USA), the AQP7 Ab was from Alomone Laboratories (Jerusalem, Israel), CXCL14 and PCK1 Abs were from Aviva Systems Biology (San Diego, Calif., USA), the Insig1 Ab was obtained from BioVision (Mountain View, Calif., USA), EGR1, GADD45 and PCK2 Abs were from Cell Signaling Technology (Billerica, Mass., USA), and the CD36 Ab was from Novus Biologicals (Littleton, Colo., USA). Rabbit polyclonal Ab raised against cystathionine ␤-synthase (CBS) was a gift from Dr. Warren Kruger (Fox Chase Cancer Center, Philadelphia, Pa., USA). The methionine synthase (MTR) Ab was purchased from Acris Antibodies (San Diego, Calif., USA), the betainehomocysteine methyltransferase (BHMT) Ab was from Abcam, and the methylenetetrahydrofolate dehydrogenase 2 (MTHFD2) Ab was obtained from GenWay Biotech, Inc. (San Diego, Calif., USA). Proteins released into the blood were analyzed by ELISA following the manufacturer’s instructions. FGF21 ELISAs were purchased from Millipore (Billerica, Mass., USA), insulin-like growth factor binding protein 1 (IGFBP1) ELISAs were obtained from Uscn Life Science, Inc. (Missouri City, Tex., USA) and resistin ELISAs were from DRG International (Mountainside, N.J., USA).

Metabolomics Metabolomic studies were conducted at Metabolon Inc. (Durham, N.C., USA) using a non-targeted platform that enables relative quantitative analysis of a broad spectrum of molecules with a high degree of confidence [16]. The three mass spectrometry-based platforms utilized were ultra-high performance liquid chromatography/tandem mass spectrometry (UHPLC/MS/MS) optimized for detection and quantification of basic species, UHPLC/MS/MS optimized for acidic species, and gas chromatography/mass spectrometry (GC/MS) [17]. Serum, inguinal adipose tissue, liver and quadriceps muscle samples (n = 5 per tissue and treatment group) were processed as described previously using a series of organic and aqueous extractions to remove the protein fraction while allowing maximum recovery of small molecules [16, 18]. Liquid chromatography/mass spectrometry (LC/MS/MS) was conducted on a Waters ACQUITY UHPLC and a Thermo-Finnigan LTQ spectrometer consisting of an electrospray ionization source and linear ion-trap mass analyzer. Sample extracts were dried under nitrogen and then reconstituted in acidic or basic LC-compatible solvents, each of which contained injection standards at fixed concentrations. Extracts reconstituted in acidic conditions were gradient eluted using water/methanol containing 0.1% formic acid, while the basic extracts, which also used water/methanol, contained 6.5 mM ammonium bicarbonate, pH 8. The MS instrument scanned 99–1,000 m/z and alternated between MS and data-dependent MS/MS scans using dynamic exclusion. For gas chromatography/mass spectrometry (GC/MS) analyses, samples dried under nitrogen and vacuum desiccation were derivatized using equal parts of bistrimethyl-silyl-trifluoroacetamide and a solvent mixture of acetonitrile:dichloromethane:cyclohexane (5: 4:1) with 5% triethylamine at 60 ° C for 1 h. The derivatized samples were separated on a 5% phenyldimethyl silicone column with helium as the carrier gas and a temperature ramp from 60 to 340 ° C and then analyzed in a Thermo-Finnigan Trace DSQ fast-scanning single-quadruple mass spectrometer operated at unit mass resolving power with electron impact ionization and a 50–750 atomic mass unit scan range. Three types of controls were analyzed in concert with the experimental samples: samples generated from pooled experimental samples or a pool of human plasma that has been extensively characterized by Metabolon served as technical replicates throughout the data set, extracted water samples served as process blanks, and a cocktail of standards spiked into every analyzed sample allowed instrument performance monitoring. Experimental samples and controls were randomized across the platform run. Metabolites were identified by automated comparison of the ion features in the experimental sample to a reference library of purified chemical standards or recurrent unknown entities that included retention time, mass (m/z), preferred adducts, and in-source fragments as well as associated MS spectra and were curated by visual inspection for quality control using software developed at Metabolon [19]. Tissue and serum metabolomes consisted of: (a) liver: 262 named and 182 unnamed biochemicals; (b) inguinal adipose tissue: 179 named and 69 unnamed biochemicals; (c) muscle: 219 named and 143 unnamed biochemicals, and (d) serum: 247 named and 85 unnamed biochemicals.




Statistical Analysis Multiple group comparisons were conducted for gene and protein expression analyses by one-way ANOVA using SigmaPlot 12 (Systat Software, Inc., San Jose, Calif., USA). Pair-wise comparisons were performed using Welch’s t tests and/or Wilcoxon’s rank sum tests and ANOVA for metabolomic studies.

Results

At a p value ^0.05, MR led to the differential expression of 2,170 genes in liver, 1,483 genes in inguinal adipose tissue and 736 genes in quadriceps muscle (online suppl. table S2). When analyzed at cutoff values of 1.5-fold and p ^ 0.05, MR significantly upregulated 172 and downregulated 113 genes in liver, and upregulated 123 and downregulated 101 genes in inguinal adipose tissue (online suppl. table S2). Quadriceps muscle showed the lowest gene expression changes with 39 upregulated and 41 downregulated genes (online suppl. table S2). Because MRC reverses MR’s effects on adiposity and serum parameters [9], microarray analyses were also conducted in tissues from MRC rats. At the same cutoff values of 1.5-fold and p ^ 0.05, MRC significantly upregulated 61 genes and downregulated 25 genes in liver, upregulated 4 genes and downregulated 9 genes in inguinal adipose tissue, and upregulated 15 genes and downregulated 8 genes in quadriceps muscle (online suppl. table S2). Fold-changes and p values of significantly expressed genes in MR versus CF and MRC versus CF groups are listed in online supplementary table S3a–S3l. Gene enrichment analysis using Ariadne’s Gene Ontology revealed a trend toward the upregulation of genes associated with pathways involved in lipid, carbohydrate, nucleotide and amino acid metabolism, and the downregulation of genes associated with pathways involved in fatty acid synthesis, cytoskeleton and cell junction complexes, intracellular transport and inflammation in inguinal adipose tissue from MR rats (table 1). In inguinal adipose tissue, MRC led to the downregulation of genes that encode proteins involved in branchedchain amino acid metabolism, triglyceride biosynthesis, the respiratory chain, fatty acid oxidation and glucose metabolism (online suppl. table S4). MRC, however, did not reverse MR’s effects on genes involved in glycogen synthesis, desmosome assembly, and extracellular matrix and intermediate filament polymerization (online suppl. table S4). In liver tissue, gene enrichment analysis showed significant changes in the expression of genes associated with translation and nucleolar assembly as well as mitochondrial protein transport (table 1). Other genes upregulated in liver by MR were associated with amino acid and carbohydrate metabolism (table 1). Genes downregulated in liver by MR included those encoding proteins associated with nucleotide metabolism, branched-chain amino acid metabolism, inflammation and xenobiotic clearance (table 1). MRC upregulated the expression of genes associated with tetraspanin, ARF, chemokines and IGFBP, while it downregulated genes for cholesterol-isoprenoid and G-protein-coupled receptor (GPCR) pathways in liver (online suppl. table S4). In quadriceps muscle, MR upregulated the expression of genes involved in fatty acid oxidation pathways as well as genes associated with chemokine signaling (table 1), while it downregulated genes from pathways involved in glycogen metabolism and ubiquitin-dependent protein degradation (table 1). Interestingly, MRC significantly upregulated genes involved in the respiratory chain, fatty acid oxidation and reactive oxygen species (ROS) catabolism, among others, and caused a trend toward a decrease in insulin receptor signalingassociated genes in quadriceps muscle (online suppl. table S4). In all tissues examined, there were significant changes in GPCR gene transcription. Among the genes significantly upregulated (1.5-fold and p ^ 0.05) by MR in rat inguinal adipose tissue microarrays are ELOVL family member 6 (Elovl6), acetyl-CoA carboxylase ␣

136




Table 1. Biological processes

identified from enrichment analysis of genes upregulated or downregulated by 1.5-fold and p < 0.05 using Fisher’s exact test in adipose tissue, liver and skeletal muscle

Pathway Inguinal adipose tissue Upregulated Tricarboxylic acid cycle Respiratory chain Glycogen synthesis Glucose metabolism Fatty acid oxidation Glycogen degradation Cholesterol-isoprenoid metabolism Triacylglycerol biosynthesis Prostaglandins, thromboxane synthesis K+ import homeostasis Phospholipid biosynthesis Purine metabolism Pentose-phosphate shunt GPCR-GsCR GPCR-Gq-iCR Branched amino acid metabolism Pyrimidine metabolism Downregulated Tight junction assembly Actin-based cytoskeleton assembly Intermediate filament polymerization Chemokines Actinomysin-based movement Desmosome assembly Co-translational ER protein import Protein folding Amino sugars synthesis ER to Golgi transport Kinetochore assembly Golgi to endosome transport GPI anchor biosynthesis Focal junction assembly Lysosomal H+ import Retrograde Golgi-ER transport Fatty acid biosynthesis ECM-cell adhesion proteins Endocytosis TCR Liver Upregulated Translation Nucleolus organization and biogenesis Mitochondrial protein transport Bile acid metabolism Folate and pterins biosynthesis Glut-Gln-Pro metabolism Glucose metabolism Aromatic amino acid metabolism Urea cycle Steroid metabolism SOCS Glycogen synthesis Tricarboxylic acid cycle Ser-Gly metabolism

MR vs. CF, p value

6.15E-11 4.08E-07 0.0003 0.0011 0.0025 0.0029 0.0069 0.0074 0.0105 0.0150 0.0162 0.0202 0.0232 0.0240 0.0302 0.0337 0.0368 1.50E-05 0.0001 0.0003 0.0007 0.0034 0.0042 0.0093 0.0010 0.0116 0.0133 0.0141 0.0158 0.0183 0.0187 0.0218 0.0347 0.0352 0.0428 0.0458 0.0476

5.15E-12 5.54E-07 0.0006 0.0010 0.0013 0.0015 0.0022 0.0022 0.0046 0.0054 0.0065 0.0082 0.0180 0.0204

137




Table 1 (continued)

Pathway

MR vs. CF, p value

Asp-Lys-Thr-Met-Cys metabolism Leukotriene synthesis Mono–di-carboxylate import DUSP Downregulated Pyrimidine metabolism Fatty acid oxidation Purine metabolism Branched amino acid metabolism Cholesterol-isoprenoid metabolism Xenobiotic clearance FZD Low-density lipoprotein receptor Prostaglandins, thromboxane synthesis IGFBP

0.0261 0.0387 0.0402 0.0403

Quadriceps muscle Upregulated Fatty acid oxidation ITGL Chemokines SOCS Leukotriene synthesis Gap junction assembly Maf Actomyosin-based movement Downregulated mRNA processing Ras Endocytosis Protein nucleus import Glycogen degradation Ubiquitin-dependent protein degradation Calcineurin Nuclear pore organization and biogenesis Purine metabolism APC-C TLR

1.35E-05 4.58E-05 0.0005 0.0006 0.0076 0.0120 0.0132 0.0199 0.0367 0.0433

0.0003 0.0025 0.0079 0.0093 0.0268 0.0284 0.0381 0.0394 0.0031 0.0041 0.0139 0.0149 0.0216 0.0292 0.0328 0.0379 0.0430 0.0438 0.0462

(Acaca), mitochondrial glycerol-3-phosphate acyltransferase (Gpam), Scd, lectin, galactoside-binding-soluble, 12 (Lgals12), acetyl-CoA acetyltransferase 2 (Acat2), insulin-induced gene 1 (Insig1), resistin (Retn), peroxisome proliferator-activated receptor ␥ coactivator 1␣ (Ppargc1a), lipin 1 (Lpin1), ATP citrate lyase (Acly), hormone-sensitive lipase (Lipe), glycerol3-phosphate dehydrogenase 1 (Gdp1), diacylglycerol O-acyltransferase 2 (Dgat2) and monoglyceride lipase (Mgll) (online suppl. table S3a). Cysteine supplementation of the MR diet reversed the expression of these genes to control levels, except for Elovl6 and Lpin1, whose expressions were driven below control levels (p ^ 0.05; online suppl. table S3a). MR also increased the expression of genes coding for proteins involved in carbohydrate metabolism such as glycogen synthase 2 (Gys2) and glycogen phosphorylase (Pygl), pyruvate dehydrogenase ␤ (Pdhb), phosphoenolpyruvate carboxykinase 1 (Pck1), isocitrate dehydrogenase 3␣ (Idh3A), lactate dehydrogenase (Ldha), pyruvate dehydrogenase kinase 1 (Pdk1), pyruvate

138




carboxylase (Pc), muscle phosphofructose kinase (Pfkm), 6-phosphofructose-2-kinase/fructose-2,6-biphosphatase 1 and 3 (Pfkfb1 and Pfkfb3), malate dehydrogenase 1 and 2 (mdh1 and mdh2), and mitochondrial aconitase (Aco2) in inguinal adipose tissue (online suppl. table S3a). MRC restored the expression of these genes to control levels, except for Pdk1 and Pc, which had mRNA levels below control levels (online suppl. table S3a). A gene expression pattern similar to those of Pdk1 and Pc was also observed for the regulator of G-protein signaling 7 (Rgs7) (online suppl. table S3a). MR downregulated the expression of genes associated with amino acid, nucleotide and xenobiotic metabolism as well as transporters and signal transduction in inguinal adipose tissue. MR also decreased the expression of genes that encode proteins involved in immune responses and inflammation such as chemokine (C-C motif) receptors 2 and 5 (Ccr2 and Ccr5), chemokine (C-C motif) ligand 7 (Ccl7) and lipopolysaccharide binding protein (Lbp) (online suppl. table S3b). MRC drove the expression of these and other genes to or toward control levels (online suppl. table S3b). Both dietary interventions, MR and MRC, downregulated the expression of fatty acid binding protein 3 (Fabp3), but the effect of MRC on this gene was weaker compared to the effect induced by MR (online suppl. table S3b). MRC also caused transcriptional changes on genes not affected by MR in inguinal adipose tissue. Specifically, MRC significantly upregulated the expression of genes coding for the GPCRs Olr1415, Olr52e8, Olr1590, and downregulated the expression of GDP-forming succinateCoA ligase (Suclg2) and S100 calcium binding protein A9 (S100a9) (online suppl. table S3h). In MR liver tissue, microarrays revealed the significant induction of genes implicated in fatty acid and triglyceride metabolism such as very low density lipoprotein receptor (Vldlr), thrombospondin receptor (Cd36), malic enzyme 1, NADP+-dependent (Me1), Ppargc1a, hydroxyacyl-CoA dehydrogenase/3-ketoacyl-CoA thiolase/enoyl-CoA hydratase (Hadhb) and 1-acylglycerol-3-phosphate O-acyltransferase 9 (Agpat9) (online suppl. table S3c). MRC reversed the expression of these genes to control levels with the exception of Me1, which had mRNA levels that fell below control levels (online suppl. table S3c). MR also upregulated genes associated with carbohydrate, amino acid, nucleotide and xenobiotic metabolism, transport, protein turnover, cell signaling, cell proliferation and apoptosis in liver (online suppl. table S3c). Although MRC brought mRNA levels for some of these genes to control levels, the expression of other genes was upregulated by this dietary intervention (online suppl. table S3c). Finally, MR upregulated the expression of Fgf21 (online suppl. table S3c), a gene that encodes an FGF family protein that protects against diet-induced obesity and insulin resistance [20, 21]. Microarrays also showed that MR and MRC downregulated the expression of genes involved in liver lipid metabolism such as Scd, patatin-like phospholipase domain containing 3 (Pnpla3), acetyl-CoA carboxylase ␤ (Acacb), ATP citrate lyase (Acly), carnitine O-octanoyltransferase (Crot), insulin-induced gene 1 (Insig1), hepatic lipase (Lipc) and ELOVL family member 5 (Elovl5) (online suppl. table S3d); however, the expression changes mediated by MRC were smaller compared to those mediated by MR. MR also downregulated the expression of glucokinase (Gck) and pyruvate kinase, liver and RBC (Pklr, involved in carbohydrate metabolism), ornithine aminotransferase (Oat) and methylmalonyl-CoA epimerase (Mcee, involved in amino acid metabolism), proteasome subunit, ␤ type, 8 (Psmb8) and ISG15 ubiquitin-like modifier (Isg15, involved in protein turnover) as well as genes associated with xenobiotic metabolism. The expression of xenobiotic metabolism genes was reversed to control levels by MRC, with the exception of NADPH oxidase 4 (Nox4), whose expression was significantly upregulated by MRC (online suppl. table S3d). MRC was also found to drive the expression of some genes in the opposite direction of MR-mediated changes. While MR downregulated the transcription of chemokine (C-X-C motif) ligands 9 and 12 (Cxcl9 and Cxcl12), esterase 1 (Es1), insulin-like growth factor bind-

139




ing protein, acid labile subunit (Igfals), ISG15 ubiquitin-like modifier (Isg15), solute carrier family 1, member 2 (Slc1a2), solute carrier family 34, member 2 (Slc34a2), major urinary protein pseudogene (Mup), 2ⴕ,5ⴕ-oligoadenylate synthase 1 (Oas1), purinergic receptor P2X, ligand-gated ion channel, 7 (P2rx7) and tumor necrosis factor ligand superfamily, member 13 (Tnfsf13), MRC significantly upregulated the expression of these genes above control levels (online suppl. table S3d). Opposite to the MR effects, MRC also led to the upregulation of amylase ␣ 1A (Amy1A), cysteine sulfinic acid decarboxylase (Csad), glutathione S-transferase mu 3 (Gstm3), aldehyde dehydrogenase 1 family, member b1 (Aldh1b1), aldo-keto reductase family 1, member C4 (Akr1c4), potassium voltage-gated channel, shaker-related subfamily, member 3 (Kcna3), chemokine (C-X-C motif) ligands 13 and 16 (Cxcl13 and Cxcl16), cyclin-dependent kinase inhibitor 2C (Cdkn2c), phosphodiesterase 6C (Pde6C), cadherin 17 (Cdh17) and matrix metallopeptidase 12 (Mmp12) genes (online suppl. table S3i). Interestingly, MRC also upregulated the expression of genes associated with cell growth, differentiation and carcinogenesis such as probasin (Pbsn), tetraspanin 8 (Tspan8), V-myb viral oncogene homologue (avian)-like 1 (Mybl1), caveolin 1 (Cav1), cyclin-dependent kinase inhibitor 1A (Cdkn1a) and embigin (Emb) (online suppl. table S3i). Finally, most genes downregulated by MRC in liver were also decreased by MR, except for glucose-6-phosphate dehydrogenase (G6pd) and Bhmt, in which mRNA levels were not changed by MR (online suppl. table S3j). The microarray analysis in quadriceps muscle showed that MR induced the expression of genes coding for lipid and carbohydrate metabolism proteins, while MRC lowered their expression to control levels (online suppl. table S3e). One exception was NADH dehydrogenase (ubiquinone) flavoprotein 3 (Ndufv3): its transcription levels remained significantly above control levels (online suppl. table S3e). Both dietary interventions caused the transcriptional upregulation of genes coding for the myofibrillar protein troponin I type 1 (Tnni1); the skeletal muscle development proteins myozenin 2 (Myoz2), cysteine and glycine-rich protein 3 (Csrp3) and myomesin family, member 3 (Myom3); the water and glycerol transporter aquaporin 7 (Aqp7); the ubiquitin-conjugating enzyme E21 (Ube21) and the extracellular matrix protein leprecan-like 1 (Leprel1) (online suppl. table S3e). The expression changes for the latter genes were smaller in quadriceps muscle from MRC rats compared to MR rats (online suppl. table S3e). Most genes downregulated by MR in quadriceps muscle were brought to control levels by MRC, except for potassium voltage-gated channel, shaker-related subfamily, ␤1 member (Kcnab1), secreted frizzled-related protein 2 (Sfrp2), phospholipase C, ␦4 (Plcd4), ralA binding protein 1 (Ralbp1) and glycoprotein M6B (Gpm6b) genes, whose expression levels remained below control levels (online suppl. table S3f). MRC also lowered the expression of genes not affected by MR in quadriceps muscle such as the trace amine associated receptor 7C (Taar7), Ras association (RalGDS/AF-6) domain family (N-terminal) member 9 (Rassf9), N-acetyltransferase 8B (Nat8b) and small nuclear ribonucleoprotein D2 polypeptide (Snrpd2) (online suppl. table S3l). Because the majority of significant gene expression changes associated with lipid, carbohydrate and amino acid metabolism was observed in inguinal adipose tissue and liver, validation of gene expression changes induced by MR and MRC was conducted in these two tissues by TaqMan qPCR. Genes were chosen from the lists of differentially expressed genes by MR and MRC. The qPCR studies also included genes whose expression was not modulated by MR and MRC according to the microarray analysis. In inguinal adipose tissue, MR was confirmed to increase mRNA levels of genes involved in lipid metabolism such as Acly, Aqp7, Elovl6, and Insig1, and carbohydrate metabolism such as Aco2, Ldha, Pc, Pck1, Pdk1 and Pfkfb3 (fig. 1, online suppl. table S5). The expression of Lpin1, an amplifier of PPAR␥ action; Retn, an adipokine released by differentiated adipocytes, and genes coding for proteins involved in signaling pathways such as dual spec-

140




ificity phosphatase 4 (Dusp4), phosphodiesterase 3B (Pde3b), Fgf21, Rgs7 and thyroid stimulating hormone receptor (Tshr) were also confirmed in MR rat inguinal adipose tissue by qPCR (fig. 1, online suppl. table S5). MR was also shown to upregulate the expression of the carrier Slc36a2, a gene coding for a protein involved in the transport of small amino acids such as glycine, alanine and proline, and to downregulate the expression Dusp18 in inguinal adipose tissue by qPCR (fig. 1, online suppl. table S5). Although microarrays showed the significant downregulation of Ccr5, an inflammation-associated gene in MR rat inguinal fat, qPCR analysis showed a non-significant trend toward a decrease in the expression of this gene (online suppl. table S5). Other selected genes were not significantly changed in this tissue by MR based on the qPCR analysis (online suppl. table S5). In addition, the qPCR analysis showed that MRC only caused trends toward an increase or decrease in the expression of the selected genes in inguinal adipose tissue, except for Aco2, whose expression was significantly decreased compared to CF rats (online suppl. table S5). In MR rat liver, qPCR confirmed the upregulation of the fatty acid and triglyceride metabolism genes Agpat9, Cd36, cytochrome P450, family 17, subfamily A, polypeptide 1 (Cyp17a1), Fgf21, Hadhb, Me1 and Vldlr (fig. 1, online suppl. table S6). While the expression of Agpat9, Cd36, Cyp17a1, Fgf21 and Vldlr mRNA was brought to control levels by MRC, Hadhb and Me1 mRNA levels were brought below CF levels according to the qPCR analysis (online suppl. table S6). In agreement with the microarray data, qPCR studies showed the downregulation of the lipid metabolism genes Acacb, Acly, Crot, Elovl5, Insig1, Lipc and Pnpla3 by MR and MRC (fig. 1, online suppl. table S6). Although MR liver microarrays showed the decreased expression of Mcee, a gene also associated with the degradation of branched-chain amino acids, no significant expression changes were observed for this gene in MR and MRC rat livers by qPCR (fig. 1, online suppl. table S6). TaqMan qPCR studies in liver confirmed the upregulation of the carbohydrate metabolism-associated genes aconitase 2 (Aco2) and Pck2 as well as the downregulation of Gck, but not of Pklr, by MR suggesting that this dietary intervention could be promoting gluconeogenesis in this tissue (online suppl. tables S3c, S3d and S6). TaqMan qPCR studies in liver from MR rats confirmed the transcriptional upregulation of the stress-response genes growth arrest and DNA-damage inducible 45␣ (Gadd45a) and activating transcription factor 4 (Atf4), as well as the protein turnover genes eukaryotic translation initiation factor 2, subunit 1␣ (Eif2s1), leucyl-tRNA synthetase (Lars) and tyrosine aminotransferase (Tat) (online suppl. table S6). While liver microarrays showed no significant changes (1.5-fold and p ^ 0.05) in the expression of activating transcription factor 6 (Atf6) in MR rats, this stress response gene was upregulated according to the qPCR studies (online suppl. table S6). TaqMan qPCR also confirmed the upregulation of other genes associated with amino acid metabolism by MR in rat liver such as ornithine decarboxylase (Odc1) and spermine synthase (Srm), suggesting increased synthesis of polyamines in this tissue, but, in contrast to the microarray data, the expression of Odc1 was also upregulated by MRC (online suppl. table S6). Increased expression of glutamate-cysteine ligase, modifier subunit (Gclm), which encodes the rate-limiting enzyme in the synthesis of glutathione in liver, was observed in the qPCR studies in contrast to microarrays, which showed that MR and MRC caused no significant changes on the expression of this gene (online suppl. table S6). Similarly, Nox4 expression was upregulated in MRC rat liver microarrays, but the qPCR data revealed the downregulation of this gene in MR rat liver and no change in its expression in MRC rat liver (online suppl. table S6). In addition, while the microarrays revealed that MR and MRC upregulated genes associated with nucleotide metabolism such as CTP synthase (Ctps), uridine monophosphate synthase (Umps), phosphoribosylpyrophosphate amidotransferase (Ppat) as well as genes encoding cell transporters such as solute carrier family 38, member 2 (Slc38a2) and aquaporin 8 (Aqp8), the qPCR studies showed that these genes were only up-

141

J Nutrigenet Nutrigenomics 2012;5:132–157 © 2012 S. Karger AG, Basel www.karger.com/jnn

DOI: 10.1159/000339347


Inguinal adipose tissue

Fold-change relative to control

25

CF MR MRC

20

15

10

5

Rgs7

Slc36a2

Tshr Vldlr

Retn

Srm

Prss15 Pck2

Phgdh

Trib3

Pfkfb3 Pck1

Pdk1

Pde3b

Pc

Pck1

Lpin1

Ldha

Insig1

Fgf21

Elvol6

Dusp4

Aco2

Acly

a

Aqp7

0

Liver


100 CF MR MRC

80 40

20

Odc

Me1

Niban

Lars

Igfbp1

Hadhb

Gldc

Gadd45a

Fgf21

Ctps

Cyp17a1

Cd36

1.0

0.5

Slc6a6

Ppp1r3b

Pnpla3

Mcee

LipC

Insig1

Igfbp3

Fgf1

Crot

Acly

0 Elovl5

b

1.5

Acacb


Agpat9

0

Fig. 1. TaqMan qPCR analysis of significantly changed genes in inguinal adipose tissue and liver of F344

rats fed CF, MR or MRC diets. Values are expressed as the mean 8 SEM of 6–8 samples per treatment group and analyzed by one-way ANOVA.

142




regulated by MR (online suppl. table S6). Finally, TaqMan qPCR studies confirmed the upregulation of CCAAT/enhancer binding protein (C/EBP), ␤ (Cebpb), Tnfsf13, Cxcl13, Cdh17, Mmp2, Pbsn, and Myb1a, but not of Amy1a, Cxcl16, Tspan8 and G6pd in liver by MRC (online suppl. table S6). Western blot analysis showed the increased expression of AQP7, DUSP4, INSIG1 and PCK1 proteins in inguinal adipose tissue from MR rats (online suppl. fig. S1). Also, MR significantly decreased CXCL14 protein levels in inguinal adipose tissue and increased serum resistin levels in agreement with the MR effects on inguinal adipose tissue gene expression (online suppl. fig. S1 and S2). MRC reversed MR’s effects on these proteins toward control levels, except for CXCL14 and resistin, whose protein levels were decreased below control levels by MRC (online suppl. fig. S1 and S2). In livers from MR rats, Western blots showed higher levels of CD36, early growth response 1 (EGR1) and PCK2 proteins (online suppl. fig. S2). Although MR upregulated the transcription of Gadd45a, no MR effect was observed at the protein level (online suppl. fig. S2). Proteins released from the liver into the blood stream were also examined by ELISA. Corresponding with the gene expression data, MR rats had higher serum levels of FGF21 compared to CF rats (online suppl. fig. S2). In contrast, serum levels of IGFBP1 were significantly lowered in MR rats compared to CF rats (online suppl. fig. S2). In agreement with the gene expression data by qPCR, MRC reversed MR’s effects on FGF21, GADD45a and PCK2 proteins to control levels (online suppl. fig. S2). Only a trend toward a decrease in CD36 protein level was observed in MRC rat liver in contrast to the qPCR data which showed the significant downregulation of Cd36 expression by MRC (online suppl. fig. S2). MR only caused a few changes in the transcription of genes encoding enzymes involved in hepatic methionine metabolism according to the microarray data. No gene expression changes were observed for Cbs, methylenetetrahydrofolate reductase (Mthfr), Bhmt, glutamate-cysteine ligase (Gclm), and methionine synthase (Mtr). Only glutathione synthase (Gss) and methylenetetrahydrofolate dehydrogenases 1 and 2 (Mthfd1 and Mthfd2) expression levels were upregulated in livers according to the microarray data (not shown). In contrast, qPCR showed the upregulation of most genes associated with sulfur amino acid metabolism, except for glutathione S-transferases mu2 and mu3 (Gstm2 and Gstm3) and Mthfr, whose expression remained unchanged (table 2). MRC reversed MR’s effects on the expression of these genes to control levels, except for the significant downregulation of Bhmt (table 2). Western blot analysis showed that CBS protein levels were significantly decreased by MR, while BHMT and MTHFR protein levels were increased (fig. 2). Protein levels for MTHFD2 and MTR remained unchanged by MR (fig. 2). MRC brought the levels of hepatic methionine metabolic enzymes to or toward control levels, with the exception of CBS, whose protein levels remained low (fig. 2). To further define the physiological changes that contribute to MR’s favorable effects, a comprehensive metabolomic analysis was conducted in serum, inguinal adipose tissue, liver and quadriceps muscle from CF, MR and MRC rats. A summary of biochemical changes for each tissue is found in online supplementary table S7. The metabolomic analysis not only revealed changes in amino acid metabolism but also in pathways involved in lipid, carbohydrate and nucleotide metabolism (online suppl. tables S8–S11). As previously shown, MR rats had lower serum levels of methionine and the methionine-derived amino acid cysteine compared to CF rats, although the decrease in cysteine was not statistically significant (table 3). MR rats also had significantly lower levels of serum tryptophan, an amino acid that can be converted to the neurotransmitter serotonin, which was significantly increased compared to serum from CF and MRC rats (table 3, online suppl. table S8). While MR decreased levels of aromatic amino acids, it significantly increased serum levels of serine, threonine, aspartate, glutamine and histidine as well as urea and urea

143

J Nutrigenet Nutrigenomics 2012;5:132–157 © 2012 S. Karger AG, Basel www.karger.com/jnn

DOI: 10.1159/000339347


Table 2. TaqMan qPCR analysis of genes involved in hepatic methionine metabolism

Gene

CF

MR vs. CF

MRC vs. CF

fold-change Bhmt Cbs Gss Gsta3 Gstm2 Gstm3 Mat1a Mthfd1 Mthfd2 Mthfr Mtr

1.0080.06 1.0080.06 1.0080.05 1.0080.03 1.0080.04 1.0080.17 1.0080.02 1.0080.07 1.0080.44 1.0080.09 1.0080.05

p value

fold-change