Physiological Genomics

A combined 1H-NMR spectroscopy- and mass spectrometry-based metabolomic study of the PPAR-α null mutant mouse defines profound systemic changes in metabolism linked to the metabolic syndrome

Helen J. Atherton, Nigel J. Bailey, Wen Zhang, John Taylor, Hilary Major, John Shockcor, Kieran Clarke, Julian L. Griffin


The mobilization of triacylglycerides from storage in adipocytes to the liver is a vital response to the fasting state in mammalian metabolism. This is accompanied by a rapid translational activation of genes encoding mitochondrial, microsomal, and peroxisomal β-oxidation in the liver, in part under the regulation of peroxisome proliferator-activated receptor-α (PPAR-α). A failure to express PPAR-α results in profound metabolic perturbations in muscle tissue as well as the liver. These changes represent a number of deficits that accompany diabetes, dyslipidemia, and the metabolic syndrome. In this study, the metabolic role of PPAR-α has been investigated in heart, skeletal muscle, liver, and adipose tissue of PPAR-α null mice at 1 mo of age using metabolomics. To maximize the coverage of the metabolome in these tissues, 1H-NMR spectroscopy, magic angle spinning 1H-NMR spectroscopy, gas chromatography-mass spectrometry, and liquid chromatography-mass spectrometry were used to examine metabolites in aqueous tissue extracts and intact tissue. The data were analyzed by the multivariate approaches of principal components analysis and partial least squares. Across all tissues, there was a profound decrease in glucose and a number of amino acids, including glutamine and alanine, and an increase in lactate, demonstrating that a failure to express PPAR-α results in perturbations in glycolysis, the citric acid cycle, and gluconeogenesis. Furthermore, despite PPAR-α being weakly expressed in adipose tissue, a profound metabolic perturbation was detected in this tissue.

  • metabolic profiling
  • metabonomics
  • gluconeogenesis
  • lipid metabolism
  • metabolic regulation
  • peroxisome proliferator-activated receptor-α

during periods of fasting, when glucose availability is low, triacylglycerides (TG) in adipose tissue are hydrolyzed to free fatty acids and mobilized to reach the liver (7). In the liver, fasting causes rapid translation of genes encoding mitochondrial, microsomal, and peroxisomal β-oxidation (19). This generates ketone bodies, which provide energy to other tissues including cardiac and skeletal muscle and the brain (41).

The peroxisome proliferator-activated receptors (PPARs) are a group of nuclear receptor transcription factors (20) involved in lipid metabolism regulation. Three isoforms have been identified: PPAR-α, PPAR-δ (or PPAR-β), and PPAR-γ. PPAR-α is highly expressed in tissues with a high catabolic rate of fatty acids such as the liver, skeletal muscle, the heart, and kidneys, whereas PPAR-γ is mainly expressed in adipose tissue, where it is important in adipogenesis (39). PPAR-δ is ubiquitously expressed throughout the body (2, 43). Numerous fatty acids and their derivatives, including eicosanoids and prostaglandins, are known ligands of PPARs (4). PPAR-α and -γ are also the primary targets of numerous classes of synthetic compounds used in the successful treatment of diabetes and dyslipidemia (1). These include fibrates, which are weak PPAR-α agonists used to treat dyslipidemia (27, 40), and thiazolidinediones (TZDs), insulin-sensitizing agents that exert their therapeutic actions via PPAR-γ binding (45). In some individuals and animal models, mutations in the genes that encode the PPAR-γ isoforms result in resistance to insulin regulation of muscle glucose uptake and lipolysis in adipose tissue (36). The abnormality in insulin regulation is often seen in conjunction with atherogenic risk factors including obesity, dyslipidemia, hyperinsulinemia, type II diabetes, and hypertension (8, 33, 42). The coexistence of several of these symptoms constitutes the “metabolic syndrome.” Although no naturally occurring mutations in the gene encoding PPAR-α have been found in humans suffering from metabolic syndrome, the PPAR-α null mouse is an important research tool for understanding the regulatory actions of the PPAR-α receptor and its role in controlling systemic metabolism. For example, the loss of this receptor has been shown to significantly alter PPAR-γ expression in a number of tissues (31), demonstrating that these pathways are inextricably linked. Furthermore, PPAR-α is the main receptor target of fibrates (23, 33), demonstrating the central role this receptor plays in treating type II diabetes and obesity. Despite this knowledge, the relationship between PPARs and insulin regulation remains to be fully elucidated (10).

Metabolomics describes the comprehensive analysis of the collection of small molecule metabolites associated with a cell, tissue, organ, or organism in a context-dependent manner, being influenced by genetic modification, pathological stimuli, or environment (14, 44), and has been discriminatory even for mild or “silent phenotypes” (16, 35) in a range of organisms including humans (6, 32). The aim of this study was to use a combined 1H-NMR spectroscopy-, gas chromatography-mass spectrometry (GC-MS)-, and liquid chromatography-mass spectrometry (LC-MS)-based metabolomics approach to study the differences between control and PPAR-α null mice at 1 mo of age. At this age, overt pathological changes in these animals are not apparent. This approach also describes the organ-specific effects of PPAR-α dysfunction in terms of systemic metabolism by examining metabolic changes in the liver, heart, adipose tissue, and muscle tissue.


Tissue collection and extraction.

Tissues from wild-type SVEV/129 mice and PPAR-α null mice (n = 5; 1 mo age) were obtained from stable colonies at the University of Oxford. All procedures were carried out in accordance with National Institute of Health guidelines. The University of Oxford Animal Ethics Review Committee and the Home Office (London, UK) approved all procedures. Mice were fed standard laboratory chow ad libitum before death (Special Diet Services, Essex, UK). Mice were injected subcutaneously with a 0.3-ml mixture of medetomidine (10%; Pfizer, Kent, UK), ketamine (7.6%; Vétoquinol UK, Bicester, UK), and sterile water (82.4%), and tissue collection was performed after the loss of corneal and pedal reflexes. Abdominal white adipose tissue, hearts, livers, skeletal muscle (gastrocnemius and soleus), and diaphragms were rapidly dissected (<60 s post mortem time before freezing), snap frozen in liquid nitrogen, and stored at −80°C until extraction.

Samples were extracted using methanol-chloroform (24). Frozen tissue (∼100 mg, except soleus muscle where only ∼20 mg were available) was pulverized with dry ice. Methanol-chloroform (2:1, 600 μl) was added, and samples were sonicated for 15 min. Chloroform-water (1:1) was then added (200 μl of each). Samples were centrifuged (13,500 rpm, 20 min), and the aqueous layer was dried overnight in an evacuated centrifuge (Eppendorf, Hamburg, Germany).

NMR spectroscopy.

The dried extracts were rehydrated in 600 μl of D2O and buffered in 0.24 M sodium phosphate (pH 7.0) containing 1 mM sodium-3-(tri-methylsilyl)-2,2,3,3-tetradeuteriopropionate (TSP; Cambridge Isotope Laboratories, Andover, MA) as an internal standard. The samples were analyzed using an Inova spectrometer operating at 400.13 MHz for the 1H frequency (Varian), using a 5-mm Broadband Inverse probe. Spectra were collected using a solvent suppression pulse sequence based on a one-dimensional nuclear Overhauser effect spectroscopy pulse sequence to saturate the residual [1H]water proton signal (relaxation delay = 2 s, t1= 3 μs, mixing time = 150 ms, solvent presaturation applied during the relaxation time and the mixing time). One hundred twenty-eight transients were collected into 16 K data points over a spectral width of 12 ppm at 37°C.

High-resolution magic angle spinning 1H-NMR spectroscopy of intact tissue.

Intact tissue (∼10 mg) soaked in distilled water was inserted into zirconium oxide rotors alongside 5 μl of distilled water containing 1 mM TSP. The samples were spun at 4,000 Hz at 300 K in a 1H-13C high-resolution magic angle spinning (HRMAS) probe, interfaced to an Avance spectrometer operating at 400.13 MHz for the 1H frequency (Bruker, Rheinstetten, Germany). Spectra were acquired using both the solvent suppression sequence described above and the Carr-Purcell-Meiboom Gill (CPMG) pulse sequence, using a 40-ms total spin echo delay. This sequence uses a T2 filter to attenuate the contribution that motionally restrained metabolites such as lipids make to the resultant spectrum. In both cases, 128 transients were collected.

Metabolic profiling by GC-MS.

Aqueous samples were derivatized using the procedure reported by Gullberg et al. (18). One hundred fifty microliters of the D2O sample used for 1H-NMR spectroscopy were evaporated to dryness in an evacuated centrifuge, and 30 μl of methoxyamine hydrochloride (20 mg/ml in pyridine) were added. The samples were vortex mixed for 1 min and then left for 17 h. Samples were then silylated with 30 μl of N-methyl-N-trimethylsilyltrifluoroacetamide (MSTFA) for 1 h. The derivatized sample was diluted (1:10) with hexane before GC-MS analysis. Organic phase metabolites were derivatized by acid-catalyzed esterification (30). Lipids were dissolved in 0.75 ml of chloroform-methanol (1:1 vol/vol). BF3-methanol (0.15 ml; Sigma-Aldrich) was added, and the vials were incubated at 80°C for 90 min. Once cool, 0.3 ml of H2O (mQ) and 0.6 ml of hexane were added, and each vial was vortex mixed for 1 min. The aqueous layer was discarded, and the remaining organic layer was evaporated to dryness before reconstitution in 200 μl of hexane for analysis.

The derivatized sample was injected splitless into a Thermo Electron Trace GC Ultra equipped with a 30 m × 0.25 mm-internal diameter 5% phenylpolysilphenylene-siloxane column with a chemically bonded 25-μm TR-5MS stationary phase (Thermo Electron; injector temperature = 220°C, helium carrier gas flow rate = 1.2 ml/min). The initial column temperature was 70°C; this was held for 2 min and then increased by 5°C/min to 310°C and held for 20 min. The column effluent was introduced into a Trace DSQ quadrupole mass spectrometer (Thermo Electron; transfer line temperature = 250°C, ion source temperature = 220°C, electron beam = 70 eV). The detector was turned on after a solvent delay of 120 s, and data were collected in full scan mode using 3 scans/s across a mass range of 50–650 m/z.

Data analysis.

NMR spectra were processed using an ACD SpecManager 1D NMR processor (version 8; ACD, Toronto, Canada). Spectra were Fourier transformed after multiplication by a line broadening of 1 Hz and referenced to TSP at 0.0 ppm. Spectra were phased, and baseline was corrected manually. Each spectrum was integrated using 0.04-ppm integral regions between 0.5–4.5 and 5.1–10.0 ppm. To account for any difference in concentration between samples, each spectral region was normalized to a total integral value of 10,000.

GC-MS chromatograms were analyzed using Xcalibur, version 1.4 (Thermo Electron), integrating each peak individually. Again each integrated peak was normalized so that the total sum of peaks was set to 10,000. Deconvolution of overlapping peaks was achieved by traces of single ions. A 0.1-min threshold window was used for the deviation of peaks away from the predicted retention time across the data set. The chromatograms generated allowed the quantification of ∼120 distinct metabolites in the heart extracts, 70 of which were assigned structures using the National Institute of Standards and Technology (NIST) database of mass spectra (Fig. 1). This assignment rate was similar for all tissues, except the soleus skeletal muscle, where, because of reduced tissue quantity, only 25 metabolites were quantified in these samples, of which 19 were assigned structures.

Fig. 1.

Using gas chromatography-mass spectrometry (GC-MS) for metabolic profiling. A section of the total ion chromatogram (TIC) from the analysis of methoxime- and trimethyl silyl (TMS)-derivatized aqueous liver extracts from the peroxisome proliferator-activated receptor-α (PPAR-α) null mouse; metabolites are identified from exact retention times and comparison of the corresponding mass spectrum with those in the National Institute of Standards and Technology (NIST) database, as shown here for myo-inositol (*).

Data sets were imported into the SIMCA-P 10.0 (Umetrics, Umeå, Sweden) for processing, using both principal components analysis (PCA) and partial least squares-discriminant analysis (PLS-DA; a regression extension of PCA used for classification). Pareto scaling was used, in which each variable was centered and multiplied by 1/(Sk)1/2, where Sk is the standard deviation of the variable. This scaling increased the importance of low-concentration metabolites without significant amplification of noise. Identification of major metabolic perturbations within the pattern recognition models was achieved by analysis of corresponding loadings plots. Additionally, R2 and Q2 were used as measures for the robustness of a pattern recognition model. R2 is the fraction of variance explained by a component, and cross-validation of R2 gives Q2, which reveals the fraction of the total variation predicted by a component. Both values are indicative of how good the overall model is. Typically, a robust model has R2 >50% and Q2 >40%. Variable importance parameters (VIPs) rank the observations according to their contribution to the model. To confirm which metabolites contributed significantly to each model, the integral regions contributing most to the separation, as identified by the VIP scores, were successively excluded from the analysis, thus producing new models, the predictive power of which could be assessed using the new R2 and Q2 values.


Profiling cardiac metabolism in the PPAR-α null mouse by solution state 1H-NMR spectroscopy, HRMAS-NMR spectroscopy, and GC-MS.

High-resolution 1H-NMR spectroscopy, HRMAS-NMR spectroscopy, and GC-MS were used in conjunction with multivariate pattern recognition to metabolically profile the heart of PPAR-α null mice. Application of PCA to the 1H-NMR data differentiated the control and mutant tissue (R2 = 63%) (Fig. 2A). The corresponding loading plot revealed that this separation was caused by increases in lactate and glycerol concentrations and decreases in creatine, choline, and taurine concentration in PPAR-α null mice.

Fig. 2.

Metabolic profiling of cardiac tissue. Principal components analysis (PCA) score plot showing clustering of cardiac samples following solution state 1H-NMR spectroscopy (A), high-resolution magic angle spinning (HRMAS)-1H-NMR spectroscopy (B), and partial least squares-discriminant analysis (PLS-DA) score plot showing clustering of cardiac samples following GC-MS analysis (C). •, PPAR-α null tissue; □, control tissue. Associated metabolic differences are shown in the adjoining tables.

HRMAS-1H-NMR spectroscopy of intact tissue enables the simultaneous measurement of both lipids and low-molecular-weight metabolites, which is important in a disease such as the metabolic syndrome. In addition to an increase in lactate and decreases in creatine and choline concentrations, also detected by solution state NMR spectroscopy, there was an increase in concentration of -CH2CH2CH2- lipid moieties (R2 = 86%) (Fig. 2B).

PLS-DA, applied to the corresponding GC-MS data, also successfully distinguished between PPAR-α null and control cardiac tissues (R2 = 64%, Q2 = 80%) (Fig. 2C). The loading plot revealed that the difference was again due to increased lactate in the PPAR-α null tissue as well as increases in adenosine, mannitol, and glycerophosphoric acid and decreases in alanine, glutamine, creatine, glucose, proline, isocitrate, oxalic acid, and fructose concentrations. The increased number of metabolites contributing to this model reflected the fivefold increase in the number of metabolites that were quantified using GC-MS compared with 1H-NMR spectroscopy. Proton NMR spectroscopy detects only high-concentration metabolites, and therefore only these will feature as significant metabolic changes identified during pattern recognition. In contrast, GC-MS detects a broader range of metabolites containing OH and NH2 functionalities, and so significant changes are not restricted to high-concentration metabolites. Therefore, although a given change measured by 1H-NMR and GC-MS should be in the same direction, it is plausible that a change may only be significant in one of the data sets.

Profiling skeletal muscle metabolism in the PPAR-α null mouse by 1H-NMR spectroscopy and GC-MS.

To determine whether the metabolic perturbations described above were general to all muscle types, 1H-NMR spectroscopy and GC-MS were used to investigate metabolism in the diaphragm, gastrocnemius, and soleus skeletal muscle. PCA differentiated tissue from PPAR-α null and control gastrocnemius muscle (R2 = 77%, data not shown). While no separation was apparent using PCA for the diaphragm and the soleus tissue, most likely reflecting the low level of PPAR-α expression in these tissues, the supervised approach of PLS-DA did produce robust separation (diaphragm: R2 = 44%, Q2 = 50%; soleus: R2 = 52%, Q2 = 79%). The loading plots for muscle tissues identified increases in lactate and glutamate and decreases in glucose, alanine, and creatine concentrations.

When analyzing the data obtained via GC-MS analysis of the gastrocnemius tissue, we achieved robust separation using PLS-DA (R2 = 52%, Q2 = 53%) caused by increased glycine, lactate, and malic acid and decreased glucose, alanine, threonine, and proline concentrations. Similar clustering was observed following PLS-DA analysis of the diaphragm (R2 = 75%, Q2 = 33%) due to increased myo-inositol, gluconic acid, and lactate and decreased creatine, glutamine, isocitric acid, proline, and glucose concentrations. Soleus skeletal muscle was also analyzed via GC-MS, but, because of the limited amounts of tissue available, only 25 high-concentration metabolites could be quantified, and the resultant PLS-DA model failed to show any significant clusterings (data not shown).

Profiling liver metabolism in the PPAR-α null mouse by 1H-NMR spectroscopy, HRMAS-NMR spectroscopy, and GC-MS.

Visual inspection of the 1H-NMR spectra demonstrated that the largest metabolic differences between the PPAR-α null and control mice were in the liver (Fig. 3), with large decreases in glucose and choline concentrations and increased trimethylamine-N-oxide (TMAO) detected. PCA of the 1H-NMR spectroscopic data robustly separated the mutant and control tissues [R2 = 74% for 2 principal components (PCs)] as a result of increases in concentrations of TMAO and lactate and decreases in glucose, fructose, citrate, glutamine, and choline (Fig. 4A).

Fig. 3.

Metabolic profiling of liver tissue. Comparison of 1H-NMR spectra of 1-mo-old PPAR-α null (black) and control (gray) liver tissue with the metabolites labeled: 1, lactate; 2, alanine; 3, lysine; 4, glutamate + glutamine; 5, glutamate; 6, glutamine; 7, choline; 8, glucose; 9, trimethylamine-N-oxide; 10, taurine; 11, glycerol; 12, myo-inositol + fructose; 13, betaine.

Fig. 4.

Metabolic profiling of liver tissue. PCA score plot showing differentiation of mutant and control liver samples following solution state 1H-NMR spectroscopy (A), PLS-DA score plot showing differentiation of mutant and control liver samples following HRMAS-1H-NMR spectroscopy (B), and GC-MS analysis of aqueous metabolites (C) and organic metabolites (D). •, PPAR-α null tissue; □, control tissue. Associated metabolic differences are shown in the adjoining tables.

Intact liver samples were also analyzed using HRMAS-NMR spectroscopy. While PCA failed to identify any clusterings, PLS-DA demonstrated an increase in TMAO and decreases in glucose and choline, as detected by solution state 1H-NMR spectroscopy. Additionally, increased CH3CH2- and -CH2CH2CH2- lipid moieties and decreased HC=CH lipid moieties contributed to this classification (R2 = 87%, Q2 = 53%; Fig. 4B).

PLS-DA was used to discriminate among data from GC-MS analysis of PPAR-α null and control liver tissue (R2 =84%, Q2 = 63%; Fig. 4C). This classification was again caused by an increase in lactate and decreases in glucose and fructose, as detected by NMR, as well as increases in concentrations of succinate, gluconic acid, adenosine, and ribose and decreased concentrations of glycine and alanine in the PPAR-α null mice.

PPAR-α in mice is predominantly expressed in the liver (7, 31). This is in contrast to human liver, which contains 10-fold lower levels of PPAR-α mRNA when compared with mouse liver (34). The liver is also where the greatest metabolic changes were observed during this study. For these reasons, analysis of the organic fraction of liver metabolites was also performed. PPAR-α tissue could be readily discriminated from control tissue by PLS-DA of the GC-MS analysis of the organic fraction (R2 = 98%, Q2 = 51%, Fig. 4D). Identification of metabolites responsible for the separation was more difficult because of the similarity in the mass spectra of all the fatty acid methyl esters. Only ∼50% of the quantified metabolites were assigned structures. Of those identifiable metabolites, in the PPAR-α tissue, there was a significant increase in pentadecanoic acid, steric acid, cholesterol, and tetradecanoic acid concentration and a decrease in 7,10,13-eicosatrienoic acid and trans-2-hexadecanoic acid.

Profiling adipose tissue metabolism in the PPAR-α null mouse by 1H-NMR spectroscopy, HRMAS, and GC-MS.

Analysis of the aqueous extract of adipose tissue by 1H-NMR spectroscopy was difficult, as spectra were dominated by broad aqueous-soluble lipid resonances, and most of the inherent variation between samples could be attributed to differential tailing of these peaks. Removing these resonances from the PCA model allowed classification of PPAR-α null tissue because of increases in alanine, glutamine, proline, and CH3CH2- lipid moieties and decreases in choline, glucose, methionine, and creatine (R2 = 90%, data not shown).

To maximize the number of metabolites detected in intact tissue, the CPMG pulse sequence was used. PPAR-α null adipose tissue was found to have increased concentrations of CH3CH2- and -CH2CH2CH2- lipid moieties and decreased concentrations of -CH=CH- and COCH2CH2- lipid (R2 = 83% for PCA model). When analyzed by GC-MS and PLS-DA, classification of PPAR-α null adipose tissue was due to increased myo-inositol, acetic acid, cholesterol, stearic acid, malic acid, and palmitic acid concentrations and decreased lactate, creatine, and glucose concentrations (R2 = 68%; Q2 = 7%).

Profiling metabolism in the PPAR-α null mouse by LC-MS.

The aqueous metabolites of adipose tissue, hearts, livers, skeletal muscle tissue (both gastrocnemius and soleus), and diaphragm were analyzed in both positive and negative mode by LC-MS. Results can be found in the Supplemental Materials (the online version of this article contains supplemental data).


One of the major analytical challenges of metabolomics is the ability to measure and quantify metabolites across a concentration range of the order of ∼109, polarity range of ∼1020, and mass range of the order of 1,500 amu. Currently, no one analytical approach can provide universal coverage of even the simplest metabolome. To address this issue, we have examined a range of tissues in the PPAR-α null mouse using different, but complementary, approaches for metabolic profiling. Phenotypic differences were detected in all the tissues examined, despite previous studies only detecting changes in PPAR-α null tissue after fasting/starvation (22, 25, 47). The largest changes were in liver tissue, where PPAR-α is predominantly expressed in the mouse (7, 31), unlike in humans, where liver tissue contains 10-fold lower levels of PPAR-α mRNA when compared with mouse liver, and PPAR-α expression is relatively increased in skeletal muscle (34). These phenotypic changes were also evident at a young age (1 mo) and in adipose tissue, where expression of PPAR-α is low (2) and discriminatory in detecting some of the perturbed metabolic pathways as detailed below.

High-resolution 1H-NMR spectroscopy is the least sensitive of the approaches used, only detecting ∼25 metabolites, and suffers from significant peak overlap in one-dimensional spectra. However, it is capable of detecting all of the high-concentration organic compounds in a solution, is relatively robust in terms of reproducibility, and is a rapid approach, with acquisition of spectra taking ∼7 min. Additionally, in the form of HRMAS-1H-NMR and in vivo magnetic resonance spectroscopy (MRS), the approach is capable of detecting a wide range of metabolites in intact tissue, circumventing the need for extraction procedures that may introduce artifacts in the quantification of some metabolites and allowing the investigation of metabolic compartmentation within tissues (3, 17, 29). Although GC-MS is a more sensitive approach (11), it does require metabolite derivatization beforehand. There is also an upper limit in terms of what metabolites can be made volatile even after derivatization. LC-MS is a more practical approach in detecting hormones and peptides found in cell and tissue extracts. However, this approach crucially depends on good chromatography and may suffer from ion suppression and adduct formation, and metabolite identification is a significant challenge.

In the present study, LC-MS provided the largest data set for each tissue with ∼5,000 features being detected in positive mode and 7,000 features in negative mode. However, during investigation of the 10 most important metabolites in distinguishing tissue types, many of the detected features were either artifacts or metabolite adducts. This suggests that, although the LC-MS approach provides the most comprehensive profiling tool of those used in this study, in order for this approach to achieve its full potential, it will be necessary to develop extensive metabolite libraries for mass spectra and retention times. While GC-MS detected a much more modest number of metabolites, of the order of 100 well-defined peaks, the ability to match mass spectra with those contained within the NIST library allowed the rapid identification of the largest number of metabolites. The most significant drawback to this approach was analysis time (70 min compared with ∼7 min for NMR and ∼15 min for LC-MS).

One of the largest and most consistent changes in all tissues detected using NMR and GC-MS was a decrease in glucose concentration (Fig. 5). Hypoglycemia is a feature of the fasting state in the PPAR-α null mouse (22, 25, 47), and this is thought to arise from impaired fatty acid oxidation, resulting in a failure to generate ketone bodies for systemic metabolism during fasting that would otherwise spare glucose metabolism. This is consistent with the well-documented role of PPAR-α in maintaining the constitutive activity of the β-oxidation pathways in tissues including the heart and liver (5, 25). In this study, reduced levels of muscle and hepatic tissue glucose are already evident, demonstrating that a failure to express PPAR-α results in a lower basal concentration of glucose. PPAR-α is known to stimulate pyruvate dehydrogenase kinase-4 (PDK4), which in turn inhibits the activity of pyruvate dehydrogenase and limits glycolytic flux (46). Loss of PPAR-α prevents PDK4 expression, and this may lead to an active form of pyruvate dehydrogenase in the liver, accounting for the increased glucose utilization (28). The decrease in glucose is in contrast to the increased lactate concentrations also detected in all muscle and liver tissue. Tissue lactate may result from post mortem anaerobic glycolysis, but in animals that are killed rapidly, this effect is small. Instead a more plausible explanation is that this lactate arises from increased glycolytic consumption of glucose, because of a failure by the liver to supply peripheral organs with ketone bodies from fatty acid oxidation (Fig. 5). Under normal physiological conditions, lactate produced by muscle tissue is converted back into glucose in the liver as part of the Cori cycle (12). However, Xu et al. (47) have previously shown that, in PPAR-α null mice, the key gluconeogenic enzyme phosphoenolpyruvate carboxykinase (PEPCK) is not increased in expression at the transcriptional level, suggesting that, despite decreased systemic glucose levels and increased peripheral lactate production, the liver does not upregulate gluconeogenesis to compensate (47).

Fig. 5.

The metabolic pathways perturbed in the PPAR-α null mouse across the tissues investigated. Metabolites detected as increased or decreased in concentration relative to control tissue are indicated by adjoining arrows. The increased/decreased width of certain arrows reflects relative increased/decreased concentrations across these pathways, respectively. The data are consistent with increased glycolysis and decreased gluconeogenesis, which in turn affect both the alanine-glucose cycle and the TCA cycle.

A number of amino acids, including glutamine and alanine, were decreased in concentration in skeletal muscle, cardiac tissue, and the liver. PPAR-α suppresses the expression of a number of genes involved in aspects of amino acid degradation, including transamination, deamination, amino acid interconversions, and synthesis of amino acid-derived products (21). During fasting, one of the mechanisms for producing glucose for consumption by the brain, kidney medulla, and red blood cells involves the production and export of alanine from muscle tissue, followed by deamination and gluconeogenesis in the liver (the so-called glucose-alanine cycle) (9), and it is conceivable that the enzymes controlling this may be constitutively more active in the PPAR-α null mouse. In addition, muscle tissue contains a branched-chain 2-oxoacid dehydrogenase for the metabolism of branched-chain amino acids and their subsequent oxidation (12). The catabolism of protein within muscle tissue also produces glutamine for export to the kidneys. Furthermore, the latter-stage citric acid cycle intermediate succinate was increased in concentration, but citrate and isocitrate were decreased in concentration. While succinate-CoA can be produced by deamination of valine, the production of glutamine from 2-oxoglutarate will deplete the first half of the citric acid cycle of intermediates (Fig. 5).

Despite the fact that PPAR-α is weakly expressed in adipose tissue and has a lower relative expression in skeletal muscle compared with liver in mice (8, 37), the metabolomic approach detected a profound perturbation in metabolism in these tissues. This is indicative of the tight metabolic control exerted among organs through a combination of hormones (insulin, glucagon, etc.) and substrate availability. Because of its very low expression in white adipose tissue, it is inconceivable that PPAR-α would regulate many pathways here, but Li et al. (26) have shown that the PPAR-γ expression increases significantly in the PPAR-α null mouse. Because this receptor is highly expressed in the adipose tissue, this compensatory upregulation of PPAR-γ may account for some of the metabolic changes we have detected here (26). One of the major benefits of metabolomics is that the approach is cheap per sample, allowing the analysis of multiple organs. With multifactorial diseases such as diabetes, dyslipidemia, and obesity, it will be imperative to monitor a number of tissues to understand their interactions with one another for any functional genomic approach (e.g., DNA microarrays, proteomics).

The use of HRMAS-1H-NMR spectroscopy and GC-MS allowed the detection of increased saturated and decreased unsaturated lipids in intact liver tissue and lipid tissue extracts, respectively. HRMAS-1H-NMR spectroscopy is particularly appropriate for understanding lipid metabolism in that it appears selective for cytosolic and biologically active lipids. Because of the modest spinning speeds employed, only cytosolic lipid droplets with a relatively high degree of translational or rotational freedom will be detected (15), unlike the lipid extracts that extract both cytosolic and membrane-bound lipids. The increase in saturated lipids in hepatic tissue may represent the start of steatosis, which occurs in the PPAR-α null mouse under certain pathological conditions. For example, Hashimoto and colleagues (18a) have demonstrated a mild steatosis in PPAR-α knockout mice after a 72-h fast and profound steatosis in mice where both PPAR-α and peroxisomal fatty acyl-CoA oxidase have been knocked out, resulting from a combination of impaired mitochondrial and peroxisomal oxidation of fatty acids. The increase in the degree of saturation of fatty acids in the PPAR-α null livers detected in this study is consistent with the findings of Knight et al. (23). Their study showed that PPAR-α activation in wild-type mice caused a doubling in the ratio of C18:1 to C18:0 fatty acids in the hepatic membranes (23). The authors concluded that this was due to an eightfold increase in expression of stearoyl-CoA desaturase (SCD) activity, an enzyme that catalyzes the synthesis of monosaturated fatty acids from saturated fatty acyl-CoAs. In this study, saturated fats were increased in concentration in adipose tissue, despite SCD expression also being controlled by PPAR-γ in this tissue (38). This suggests that, even if PPAR-γ is increased in expression in response to the loss of PPAR-α, the impact of systemic metabolism still produces an overall increase in saturated fats within this tissue.

In conclusion, our combined metabolomic study of heart, liver, muscle, and adipose tissues has identified a range of steady-state changes in metabolism in the PPAR-α null mouse at 1 mo of age. These changes demonstrate a profound perturbation in glycolysis, gluconeogenesis, and amino acid metabolism in addition to fatty acid oxidation.


This study was supported by grants from the British Heart Foundation (K. Clarke and J. L. Griffin), the Wellcome Trust (J. L. Griffin), the Biotechnology and Biological Sciences Research Council, and Selcia Ltd. (H. J. Atherton and J. L. Griffin). J. L. Griffin is a Royal Society University Research Fellow.


We thank Frank Gonzalez (National Institutes of Health) for the kind gift of PPAR-α−/− mice.


  • Article published online before print. See web site for date of publication (

    Address for reprint requests and other correspondence: J. L. Griffin, Dept. of Biochemistry, Univ. of Cambridge, Tennis Court Rd., Cambridge, CB2 0GS, UK (e-mail: Jlg40{at}


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 18a.
  20. 19.
  21. 20.
  22. 21.
  23. 22.
  24. 23.
  25. 24.
  26. 25.
  27. 26.
  28. 27.
  29. 28.
  30. 29.
  31. 30.
  32. 31.
  33. 32.
  34. 33.
  35. 34.
  36. 35.
  37. 36.
  38. 37.
  39. 38.
  40. 39.
  41. 40.
  42. 41.
  43. 42.
  44. 43.
  45. 44.
  46. 45.
  47. 46.
  48. 47.
View Abstract