Metabolomics and the pig model reveal aberrant cardiac energy metabolism in metabolic syndrome

ABSTRACT Although metabolic syndrome (MS) is a significant risk of cardiovascular disease (CVD), the cardiac response (MR) to MS remains unclear due to traditional MS models’ narrow scope

around a limited number of cell-cycle regulation biomarkers and drawbacks of limited human tissue samples. To date, we developed the most comprehensive platform studying MR to MS in a pig

model tightly related to human MS criteria. By incorporating comparative metabolomic, transcriptomic, functional analyses, and unsupervised machine learning (UML), we can discover unknown

metabolic pathways connections and links on numerous biomarkers across the MS-associated issues in the heart. For the first time, we show severely diminished availability of glycolytic and

citric acid cycle (CAC) pathways metabolites, altered expression, GlcNAcylation, and activity of involved enzymes. A notable exception, however, is the excessive succinate accumulation

despite reduced succinate dehydrogenase complex iron-sulfur subunit b (SDHB) expression and decreased content of precursor metabolites. Finally, the expression of metabolites and enzymes

from the GABA-glutamate, GABA-putrescine, and the glyoxylate pathways significantly increase, suggesting an alternative cardiac means to replenish succinate and malate in MS. Our platform

discovers potential therapeutic targets for MS-associated CVD within pathways that were previously unknown to corelate with the disease. SIMILAR CONTENT BEING VIEWED BY OTHERS THE CARDIAC

MOLECULAR SETTING OF METABOLIC SYNDROME IN PIGS REVEALS DISEASE SUSCEPTIBILITY AND SUGGESTS MECHANISMS THAT EXACERBATE COVID-19 OUTCOMES IN PATIENTS Article Open access 05 October 2021

INTEGRATED LANDSCAPE OF CARDIAC METABOLISM IN END-STAGE HUMAN NONISCHEMIC DILATED CARDIOMYOPATHY Article 29 August 2022 METABOLIC CHARACTERIZATION OF HYPERTROPHIC CARDIOMYOPATHY IN HUMAN

HEART Article 09 May 2022 INTRODUCTION One of the most concerning implications of the rapidly expanding MS is the increased risk of CVD1. Many processes underlying the cardiac response to

MS, however, are not yet clarified. There is little data on cross-talk between metabolites alterations and cardiac functionality. MS is a state of simultaneously appearing at least three of

the medical conditions elevated triglycerides, low-density lipoprotein (_LDL_), blood pressure, hyperglycemia, and obesity1,2. Each of these factors independently and synergistically

increase the risk of developing CVD1. Because of the pleiotropic nature of CVD and MS, it is unlikely that single metabolites or metabolic pathways underlie the development of CVD in MS. To

this end, understanding the global metabolic state of the MS heart, using high throughput metabolomics, transcriptomics, and proteomics, may provide novel insights. Although progress has

been made in characterizing some signaling mechanisms and proteins involved in the cardiac response to MS, metabolic processes implicated in the altered cardiac energy demands may contribute

to the CVD response in MS as well3,4. In a recent metabolomics study, some amino acids and derivatives were shown to be altered in blood samples from adult humans with MS1,2. Given the

unique metabolic properties of the heart, analyses of metabolite levels in the blood and other body fluids, however, likely do not accurately reflect changes in this tissue. Furthermore,

limited human tissue samples availability and metabolic pathway discrepancies between human and small animal species complicate the clarification of the cardiac response to MS, necessitating

different approaches in the characterization of MS5. Therefore, the present study aimed to use a pig-based model that recapitulates the human MS criteria when fed a high-fat high-calorie

diet6,7. Control lean diet (LD) and MS pig heart were harvested and used to generate profiles of 283 polar metabolites by applying liquid chromatography-tandem mass spectrometry (LC/MS-MS)

and thin-layer chromatography (TLC) as well as to generate gene expression profile by whole transcriptome shotgun sequencing (RNA-Seq). We then apply unsupervised machine learning (UML) to

differentiate metabolite profiles from MS and LD pigs together with transcriptomics, enzymatic activities, and physiological data. We discovered that the heart adopts a strategy to defend

against MS by severely altering glycolysis, the availability of several Krebs’s cycle (Citric Acid Cycle, CAC) intermediates, and changes in CAC related enzyme expression, GlcNAcylation, and

activity. Our studies suggest that increased activity of the gamma-aminobutyric acid (GABA) and glyoxylate pathways replenish CAC intermediates in MS. These observations point towards a

cardiac obligation in MS to modulate the expression of enzymes involved in that GABA-glutamate, GABA-putrescine, and glyoxylate cycles that could explain the accumulation of succinate and

malate in MS. Overall, with our platform, we discovered unknown pathway connections and correlations on numerous biomarkers across all five MS-associated factors leading to a deeper and

broader insight into CVD pathogenesis. We plan to continue to leverage our pig-based platform’s capabilities to discover and investigate potential therapeutic targets to MS-associated CVD

within pathways that were previously unknown to corelate with the disease. RESULTS A HIGH-FAT/HIGH-CALORIE DIET LEADS TO MS-RELATED ALTERATIONS IN OUR PIG MODEL The MS pigs showed elevated

blood glucose (162 ± 14 mg/dl vs 96 ± 10 mg/dl, _p_ < 0.02), triglycerides (1.66 ± 0.5 mmol/L vs 0.66 ± 0.22 mmol/L, _p_ < 0.03), plasma LDL (2.64 ± 0.25 vs 0.46 ± 0.12, _p_ < 

0.01), and total cholesterol (5.8 ± 0.7 mmol/L vs 1.08 ± 0.2 mmol/L, _p_ < 0.007I) in MS vs LD. In MS pigs, we further observed increased weight gain (Fig. 1a) and blood pressure (Fig. 

1b). Additionally, our histology stains demonstrated changes at the tissue level that have been associated with CVD pathogenesis7,8 including diminished capillary density, as shown by the

difference in the immunologically detectable CD31 in the cardiac tissue (Fig. 1c, **_p_ < 0.01); the picro sirius red stain for collagen (Fig. 1d, **_p_ < 0.01); and accumulation of

intracellular lipid bodies as shown with Oil Red O staining (Fig. 1e, **_p_ < 0.01). Together, the phenotype observed in pigs on the high-fat, high-calorie diet meets all five metabolic

syndrome diagnostic criteria in humans: obesity, elevated fasting blood sugar, elevated triglycerides and LDL, and increased blood pressure. The high-fat diet, introduced to In meet the

clinical criteria for MS, has demonstrable effects on heart tissue composition in pigs: specifically, increased fat, fibrous tissue, and diminished vascularity, suggesting that high-fat diet

leads to histopathological changes at the cellular and molecular levels. METABOLOMIC ANALYSES, TOGETHER WITH UNSUPERVISED LEARNING APPROACHES, UNCOVER DIET-RELATED ALTERATIONS IN THE

ABUNDANCE OF SEVERAL CARDIAC POLAR METABOLITES To create metabolite signatures from MS and LD heart, we first applied targeted LC/MS-MS for polar metabolites, quantifying the relative

abundance of 283 metabolites in each cardiac sample9. Tissues were isolated from eight MS and seven LD animals and analyzed by mass spec within a single experiment to avoid potential batch

effects. As the variables (283 metabolites) are significantly higher than the replicates (15 pigs), we applied unsupervised machine learning (UML) to compare differences in profiles of

metabolites in MS vs. LD10,11,12. UML recognized five distinct profiles (signatures S1, S2, S3, S4, and S5), each with different proportions of the 283 assayed metabolites. Differential

weighting of each profile accounts for variations between samples (Fig. 2a). Subsequent clustering of the weights of the metabolite profiles results in two clear groups of samples, one group

containing eight MS pigs and another group with the seven LD pigs (Fig. 2b). The centroids of the identified clusters demonstrate the relative contribution of the individual metabolic

signatures in the corresponding clusters (Fig. 2c). UML identified S5 as the primary metabolic signature in the LD pigs, while S3 is prevailing in MS. The overrepresented metabolites in S5

are shown in Table S1 (supplementary). The difference between S3 and S5 is further demonstrated by the probability of individual metabolites to appear in the signatures (Fig. 2d). CAC (_p_ =

 4e-5), together with glucose and pyruvate (p = 3.1e-4) metabolism predominate in the LD related S5. CAC and pyruvate are, however, significantly diminished in the MS corresponding signature

S3 (Fig. 2e). Processes that are involved in the metabolism of butyrate (_p_ = 1.1e-4) and the Warburg effect (_p_ = 3.3e-5) are diminished in S3 as well. Conversely, processes that are

related to glutamate-gamma-aminobutyric acid (GABA) biosynthesis (_p_ = 4e-5) and the urea cycle (_p_ = 0.004) are overrepresented in S3. LD AND MS PROFILES SHOW A DISTINCT METABOLITE

DISTRIBUTION The S3 and S5 profiles likely differ in the content of numerous metabolites related to both CAC and glycolysis (Fig. 3). Therefore, we next compared metabolite abundance in the

LD-related S5 and the MS-related S3 using the quantitative mass spec data. The quantitative mass spec results (Fig. 3a) showed a decrease in the CAC entry metabolites acetyl CoA (_p_ < 

0.001) and pyruvate (_p_ < 0.01) in MS. The decreases in the CAC intermediates α-ketoglutarate (_p_ < 0.01), succinyl-CoA (_p_ < 0.0001), fumarate (_p_ < 0.01), malate (_p_ < 

0.01), and oxaloacetate (_p_ < 0.01) are all statistically significant. Glyoxylate (GLY), an intermediate in an alternative CAC pathway was diminished in MS (_p_ < 0.01) as well.

Contrariwise, succinate (_p_ < 0.01), citrate (_p_ < 0.01), and isocitrate (_p_ < 0.01) were strikingly elevated in MS. Additionally, glutamine (_p_ < 0.05), glutamate (_p_ < 

0.05), and gamma-aminobutyric acid (GABA, _p_ < 0.05) as well as putrescine (_p_ < 0.01) from the urea cycle were all increased in MS. Moreover, while the AMP/ATP ratio is elevated in

MS, the NADH content (_p_ < 0.01), as well as the NAD + /NADH value (_p_ < 0.01) are diminished. Two-dimensional thin-layer chromatography (2dTLC) was applied to validate observations

(Fig. 3b). The semiquantitative 2dTLC of MS (n = 4) and LD (n = 4) samples supported the mass spec data by consistently showing more than a threefold increase in succinate (_p_ < 0.01)

and significantly less malate (_p_ < 0.02) vs LD. Changes in metabolites between the two groups are graphically summarized in Fig. 3c. Altogether, the comparison of S3 and S5 signatures

reveals an imbalance in CAC intermediates in MS. Specifically, we observed significant decreases in the availability of pyruvate and acetyl-CoA at the entry of the CAC, as well as decreases

in several downstream CAC intermediates, including alpha-ketoglutarate, malate, and oxaloacetate. The critical CAC cofactor NADH and the NAD + /NADH ratio are altered as well. The

disproportionally large accumulation of succinate, the fifth intermediate in CAC, in comparison to succinyl-CoA, fumarate, and malate, coincides with an increase of GABA; increases in the

GABA cycle intermediates glutamine and glutamate; as well as elevated putrescine from the urea cycle in MS. Moreover, the considerable accumulation of citrate and isocitrate coincides with a

decrease of acetyl-CoA. CAC, GABA CYCLE, AND GLYOXYLATE SHUNT ENZYMES RESPOND TO MS AT THE TRANSCRIPTIONAL AND PROTEIN LEVEL To determine the mechanisms of cardiac CAC dysregulation and MS

adaptation, we determined and compared at the mRNA levels the expression of enzymes that drive the traditional CAC, GABA cycle, and the glyoxylate shunt (Fig. 4a). RNA-seq data from four of

the MS and four of the LD pigs demonstrated diminished levels of several CAC enzymes in MS: isocitrate dehydrogenase (_IDH2_, _p_ < 0.05); glutamate dehydrogenase, mitochondrial (_GluD1,

p_ < 0.05); subunit B of the mitochondrial succinate dehydrogenase (_SDHB_, n = 4, _p_ = 0.07), and the mitochondrial malate dehydrogenase (_MDH2_, MS n = 4, _p_ = 0.08). We also observed

a trend of decreased expression of the mitochondrial succinyl-CoA synthetase (_SCS-A_, MS n = 4, _p_ = 0.1) and the cytosolic aconitase (_ACO1_, MS n = 4, _p_ = 0.5) in MS vs. LD.

Conversely, in the GABA-glutamate-glutamine cycle, mitochondrial succinic semialdehyde dehydrogenase (_ALDH5A1_) is significantly increased in MS vs. control (n = 4, _p_ < 0.05).

Furthermore, the monoamine oxidase B (_MAOB_), which is known to use putrescin for GABA production increases in MS as well13. Additionally, the mitochondrial citrate lyase beta-like

(_CLYBL_) that is involved in the glyoxylate shunt, and could convert glyoxylate to malate, is also increased in MS (n = 4, _p_ < 0.01)14. Although we noted significantly less fumarate in

MS vs. LD, expression of fumarase (FH) did not change in response to diet. Additionally, expression levels of neither glyceraldehyde-3-phosphate dehydrogenase (_GAPDH_) nor the

mitochondrial _ACO2_ variant showed diet-related mRNA variations in content and served as internal controls for the observations noted above. Western blots with total cardiac tissue lysates

(50 µg protein) from four MS and four LD pig and specific antibodies against FH, SDHB, IDH2, MAOB, and GAPDH supported the RNA-seq observations related to the expression of the corresponding

genes (Fig. 4b). Fumarase and GAPDH protein levels broadly reflected the mRNA levels without significant diet-related alterations. The SDHB (*_p_ < 0.05) and IDH2 (****_p_ < 0.0001)

protein contents are diminished in MS, reflecting their mRNA levels. Conversely, analogous to the mRNA level, there is significantly more MAOB (***_p_ < 0.001) protein in MS vs. LD. The

membranes with the transferred proteins were stained with ponceau S to control for equal transfer and protein loading (Fig. S1; supplementary). Gene expression results are summarized in Fig.

 4c. Combined, these observations support the hypothesis that MS alters the flux of multiple CAC metabolites by changing the expression of genes encoding traditional CAC enzymes along with

enzymes that are involved in the _GABA_ and glyoxylate shunts. FUMARASE O-GLCNACYLATION COINCIDES WITH ALTERED ENZYMATIC ACTIVITY IN MS Given the apparent absence of changes in FH protein

and mRNA abundance in MS vs LD, we investigated the possibility that posttranslational modification may lead to disruption of FH function and activity, and thus corresponding disturbances in

fumarate and malate fluxes in MS. This notion is supported by the precursor-to-product proportion distinction in MS vs. LD (Fig. 5a). Based on the mass spec determined fumarate and malate

quantities we determined that their ratio is lower in MS vs. LD (MS, n = 8, 1:2.82 ± 0.14, _p_ < 0.01; LD, n = 7, 1:3.56 ± 0.13, _p_ < 0.01) by fumarate as 100%. Next, we used total

tissue lysates to measure and compare the FH activity in hydration of fumarate to malate reaction (Fig. 5b). In reactions assembled with 10 μg protein/reaction, at 37 °C FH was found to be

significantly less active in MS (n = 4, ****_p_ < 0.0001) vs LD (n = 4). As FH activity is dependent on enzyme oligomerization15, posttranslational modifications such as O-GlcNAcylation

could alter the conformation of the enzyme and/or prevent oligomerization of enzyme monomers, thereby diminishing its activity. Previously, we and others have observed that in hyperglycemic

conditions, the abundance of O-GlcNAcylated proteins is correspondingly greater than in normoglycemic conditions16,17,18,19,20. Because a change in FH activity could explain the significant

accumulation of succinate which we registered in the MS heart, we tested the hypothesis that in MS tissue whole-cell lysates, the O-GlcNAcylation state of FH protein will be altered. Wheat

germ agglutinin (WGA) affinity binding of O-GlcNAcylated proteins in MS and LD lysates (100 mg) and subsequent western blot (0.05 mg/lane) with FH -specific antibodies revealed significant

raise of O-GlcNAcylated FH variant in MS (n = 4, ***_p_ < 0.001) vs. LD (n = 4) (Fig. 5c, lanes B). The finding is consistent through separate studies with the lysates of four MS and four

LD pigs. The O-GlcNAcylated FH in the MS and LD fractions was quantified based on the optical density of the immunostaining reactions in the western blots. These outcomes, considered

together, allude to MS-specific FH O-GlcNAcylation that favors a derangement of fumarate and malate levels in MS. DISCUSSION Despite the considerable investigations into physiological,

epigenetic, and signaling pathway changes in MS, specific disturbances characterizing the cardiac metabolic state remain unclear. Because of MS’s multifaced nature, single molecular markers

and processes may not adequately comprehend its impact on the heart. Essential mechanisms that drive the development of MS remain to be established. To date, we developed the most

comprehensive platform studying the cardiac response to MS. Our application of UML to 283 metabolites from MS and LD cardiac pig tissues differs from the previously used pattern-recognition

approaches. It allows us the identification of yet undescribed metabolic connections. It licenses us the distinctive confirmation of yet undescribed metabolic affiliations. By incorporating

mass spec-based comparative metabolomics, transcriptomics, and functional analyses together with UML, in this study, we discovered unknown pathway links and correlations on numerous

biomarkers across all five human MS-associated factors: obesity, elevated fasting blood sugar, triglyceride, LDL level, and blood pressure. Specifically, our UML-based clustering of polar

metabolites revealed unforeseen, contrasting changes in CAC-related metabolites, concurrent declines in glycolysis, and evidence of CAC replenishment by GABA-glutamate, GABA-putrescine, and

glyoxylate shunting. MS drives alterations in the CAC, as evidenced by the concurrent decline in pyruvate, the entry point of the CAC, and its breakdown product, acetyl-CoA. Others have

previously reported a similar drop in pyruvate and acetyl-CoA accumulation in infarcted hearts of rodents21. Currently, the etiology of diminished cardiac glycolysis and acetyl-CoA

availability in MS is not well understood. The elevated circulating triglycerides and fatty acids that we registered in MS could overwhelm mitochondrial capacity for fatty acid oxidation,

subsequently leading to diminished acetyl-CoA production, though further investigation is warranted. Interestingly, despite the depletion of many CAC intermediates in MS, succinate, the

fifth intermediate in the cycle, is more abundant in MS than in control lean-diet heart. This observation raises two questions: what are the sources of succinate in MS, and what role does

succinate accumulation play in MS? The putrescine-GABA and glutamate-GABA cycle may supply succinate in MS; this is supported by the increased GABA abundance and its precursors’ putrescine,

glutamine, and glutamate that we observed in MS, as GABA is an established succinate precursor. This cycle may provide recourse for the CAC to progress in MS, though further data are needed.

Propionyl-CoA is in like way rich in MS and could give an additional source of succinyl-CoA and its downstream product, succinate. However, given the diminished levels of succinyl-CoA

together with biotin, which is the essential cofactor for propionyl-CoA carboxylase, this is unlikely22,23. Together, our data suggest that the putrescine-GABA and glutamate-GABA cycles are

the most likely candidates for succinate production in MS. Until recently, succinate had been functionally considered only as a CAC intermediate. Recent data, however, suggest it may have

additional, unexpected signaling functions24. As the continuous administration of succinate produces a dose-dependent increase in mean arterial pressure, the observed systolic blood

pressures in our MS model could be linked to succinate accumulation23,24. Interestingly, metoprolol succinate has shown considerable therapeutic promise in reducing mortality related to

heart failure, and it is more effective than other metoprolol formulations. Though succinate is administered as salt to increase the serum half-life of metoprolol, the role of succinate _per

se_ may be an overlooked benefit in this widely-employed therapy25. Given the vital role of succinate in the heart and the dysregulation we report in MS, it has potential as a therapeutic

target. Despite increased succinate, levels of fumarate, the following intermediate in the CAC, are diminished in MS. This decrease may be driven principally by the reduced expression of

succinate dehydrogenase in MS. As previously reported, fumarate functions not only as a malate precursor in the CAC but also could preserve cellular redox potential, improve calcium

homeostasis, diminish free radicals and toxic acyl-CoA derivatives deposition26. The low fumarate also correlates with the diminished expression of the hypoxia-inducible factor HIF1A in MS

(Fig. S2; supplementary) and altered cardiac angiogenesis in MS as previously reported27. Not only is fumarate low in MS, but the possible means of its replenishment through

phenylalanine/tyrosine, which is abundant in MS, likely does not occur. This is evidenced by the low levels of phenylalanine hydroxylase and fumarylacetoacetase enzymes in MS (Fig. S2;

supplementary). Likely, the fumarate level is tightly controlled not only at the level of succinate dehydrogenase expression but also at the level of fumarase activity (to convert fumarate

to malate). Our observation that O-GlcNAcylated inactive fumarase is relatively abundant in MS may explain the diminished fumarate – malate balance. The multimeric fumarase is a highly

potent enzyme that requires no co-substrates or co-enzymes to function28,29. The bulky O-GlcNAcyl moiety, however, may alter fumarase oligomerization and function30. Despite its centrality,

the reaction part of fumarase isn’t inside and out of understanding. As fumarate accumulation could trigger both adaptive and maladaptive responses, appropriately balanced fumarate

supplementation is a future avenue in developing strategies to prevent the heart from the negative consequences in MS. Malate likewise participates not only in CAC but also in other

metabolic pathways and may have a protective effect against cardiac ischemic injury31. Its balance is likely to be controlled not only by fumarase substrate availability and activity but

also by external anaplerotic replenishment. The conversion of pyruvate to oxaloacetate by pyruvate carboxylase (and the subsequent reduction of OAA to malate), however, is less likely to be

efficient in MS, as levels of both the enzyme’s substrate (pyruvate) and obligate co-factor biotin are low. Conversely, increased mitochondrial citramalyl-CoA lyase (_CLYBL_) expression and

the presence of cardiac glyoxylate may replenish malate in MS. It is commonly acknowledged that the glyoxylate cycle is missing in mammals and exists just in plants and microorganisms.

However, its function as an alternate CAC cycle that bypasses the oxidative part of the CAC to replenish malate has been reported to occur in the liver32. Glyoxylate in its metabolic network

presents a new cardiac paradigm for control of malate flux into the CAC in MS. Despite decreased acetyl-CoA and oxaloacetate, levels of citrate and isocitrate, the following intermediates

in the CAC, are significantly more abundant in MS than in control lean diet-heart. The high fatty acids level in MS may likely contribute to the cardiac citrate accumulation, as previously

reported occurring in rat hearts after reperfusion with fatty acids33. Citrate is a key intermediate in CAC and fatty acid synthesis and could significantly contribute to the diminished

glycolysis in MS34,35. Moreover, it could also be a factor in the collateral dysregulation that we registered in the MS heart36. A superior comprehension of components managing heart citrate

accumulation and discharge may recommend medication or diet in MS patients with CVD. Low metabolite abundance may reflect either a higher rate of consumption or a lower rate of formation in

MS, necessitating additional data to suggest which process is occurring. Our RNA-seq data support the latter conclusion, at least in part, as the mRNAs coding for multiple enzymes necessary

for the production of CAC intermediates is diminished in MS. Additionally, as the NAD + /NADH balance regulates the CAC, the lower NAD + /NADH ratio that we observed in MS likely impairs

CAC dehydrogenases, except for succinate dehydrogenase although it should be noted that expression of _SDHB_ was considerably reduced in our protein study. Dysregulation of CAC enzymes in MS

may further drive CVD pathogenesis; indeed, aberrant CAC enzyme levels and activity have previously been described as harming the heart37,38. Disturbance of CAC was demonstrated to precede

and to be responsible for mechanical failure of the left ventricle37. Isocitrate dehydrogenase 2 is the rate-limiting enzyme of the CAC and is also diminished in MS. In particular, decreased

expression of IDH2 induces contractile dysfunction in the heart, and it correlates with impaired oxidative decarboxylation of α-ketoglutarate, thereby redirecting CAC intermediates39. The

lower expression level of _IHD2_ that we observe in MS could contribute to the increased cardiac hypertrophy in MS animals40. We observe considerable alterations in the MS heart, suggesting

that MS builds the peril of CVD yet, basically changing the cardiac energy balance. Although our steady-state metabolomic and transcriptomic analyses are snapshots supporting our

predictions, any reliable conclusions related to perturbed metabolic fluxes are unavoidably limited. The relatively short time frame of the high-fat diet may limit the long-term effects of

MS in this early metabolic syndrome model. Furthermore, the metabolic response may be gender- and age-specific, and we investigated only intact male animals. However, given the relative

paucity of data on the development of MS in young animals, especially early in the development of MS, this may also be beneficial to our understanding of MS pathogenesis. Overall, we report

several novel findings: (i) unsupervised ML of 283 polar metabolites from MS and LD heart reveals a decline in CAC and glycolysis in MS. (ii) Specifically, there is diminished availability

of metabolites that enter the CAC: pyruvate and acetyl CoA. Downstream CAC intermediates, including succinyl-CoA, fumarate, malate, and oxaloacetate, are also diminished in abundance. (iii)

Accumulation of citrate in comparison to its precursors and its product (α-ketoglutarate) suggests for altered fatty acids synthesis and glycolysis. (iv) Accumulation of succinate in

comparison to its precursor, succinyl CoA, and its products (fumarate and malate) suggests an alternative pathway leading to succinate accumulation. (v) Increased putrescine, glutamine,

glutamate, and GABA further suggest that this alternative method for succinate repletion occurs through the GABA-putrescine and GABA-glutamate shunts. (vi) We additionally observed changes

in glyoxylate accumulation in MS. Observed changes in metabolite levels correlated well with corresponding alterations at the enzyme level. (vii) Several vital enzymes that drive the

traditional CAC (isocitrate dehydrogenase, mitochondrial succinyl-CoA synthetase, subunit B of the mitochondrial succinate dehydrogenase, and malate dehydrogenase) all are transcriptionally

down-regulated in MS. (viii) Conversely, we observed increased expression of enzymes from the GABA and glyoxylate cycle pathways that correlate with the accumulation of succinate in MS. (ix)

Additionally, low levels of α-ketoglutarate in MS could be explained by the low levels of the enzyme _GluD1_ in MS, which usually replenishes this metabolite. (x) Though we do not observe

changes in fumarase at the mRNA or protein level in MS, its post-translational O-GlcNAcylation is increased in MS. We further show that this modification coincides with a diminished

conversion of fumarate to malate, providing a mechanistic explanation for malate and fumarate derangement in MS. Together, these observations allude to a cardiac obligation in early MS to

alter the availability of enzymes from the GABA, glyoxylate, and glutamate dehydrogenase anaplerotic pathways, which could explain the replenished pools of metabolic intermediates in CAC.

Overall, the present data support a model in which high-fat diet-induced transcriptional and post-translational alternations in several standard CAC enzymes results in a paucity (or in the

case of succinate and citrate, overabundance) of their associated metabolic substrates or products in the heart. These alterations could represent adaptive attempts to salvage cardiac

function or may drive pathogenesis in MS. Although the response of patients and our pug model to MS is nearly identical41, further research will clarify which of the reported cardiac effects

are adaptive or pathologic in patients. To this end, our pig-based data may provide a molecular basis for the development of new therapeutic targets, including early interventions for CVD

development in the increasingly prevalent MS. Our pig-based platform can discover and investigate potential therapeutic targets for MS-associated CVD within pathways that were previously

unknown to be associated with the disease. MATERIAL AND METHODS ANIMAL MODEL Male intact Yorkshire pigs (n = 15, four to six weeks old) were provided by the Parsons Research, Amherst, MA.

Eight pigs (MS) were given a hypercholesterolemic (2248 kcal/daily) diet: seven pigs (LD) were given regular chow (1824 kcal/daily, Sinclair Research, Columbia, MO) for 12 weeks. In 12

weeks, pigs were anesthetized and physiologic measurements were taken, followed by euthanasia via exsanguination. Tissue from identical cardiac left ventricular regions is immediately frozen

in fluid nitrogen. All animal tests and procedures were endorsed by the Animal Care and Use Committee at the Rhode Island Hospital based on the animal care and use regulations that are

described in the NIH publication number 5377-3, 1996. SEROLOGICAL ANALYSES Prior animal euthanasia blood samples were drawn from the jugular vein. The serum samples are analyzed at the Rhode

Island Hospital, Providence, RI laboratory. All analyses are conducted in agreement with the biosafety regulations at the Rhode Island Hospital, Providence, RI. IMMUNOLOGICAL PROCEDURES AND

WGA AFFINITY CHROMATOGRAPHY Frozen or formalin-fixed cardiac tissue sections (12 µm in thickness) of left ventricular territory were stained with the following antibodies: anti-OGA,

[EPR7154 (B)], (Abcam); α- Smooth Muscle actin – FITC, (Vector); anti-YY1, ChIP Grade (ab38422) (ABcam); anti-SP1, ChIP Grade (ab13370) (ABcam). WGA affinity chromatography was applied as

previously reported20. Ponceau S membranes staining is used to validate equal protein loading and transfer. BCA Assay kit (Pierce) is used to measure protein concentration. Periodic

Acid-Schiff (PAS) staining (Sigma-Aldrich, procedure 395), lipid Oil Red O (Bio Vision, Catalog # K580-24) and picrosirius red staining kit (Polysciences, Inc) were used per manufacturer’s

protocols. All analyses are conducted in agreement with the biosafety regulations at the Rhode Island Hospital, Providence, RI. RNA-SEQ Fresh left ventricular cardiac tissue was extracted

with the RNeasy Mini Kit (Qiagen) per the manufacturer’s protocol. The integrity of RNA of the eight samples (4 MS and 4 LD) is ≥ 9.8 as determined with the RNA 6000 Nano Kit (Agilent

Technologies). High-output mode sequencing was performed by GENEWIZ (South Plainfield, NJ) on Illumina HiSeq 2500. The porcine reference genome (USMARCv1.0), the STAR aligner42 and

HTSeq-count, version 0.5.3p943 were used to map and quantify the reads. The Bioconductor package DESeq was used to perform the differential gene expression analysis. 2DTLC Silica gel G

plates with fluorescent indicator (Sigma-Aldrich) were used to perform 2dTLC as previously reported using standard molecules as a migration reference44. Spots are identified in UV light

based on the migration of standard molecules. The experiments are conducted in agreement with the biosafety regulations at the Rhode Island Hospital, Providence, RI. MASS SPECTROSCOPY

(LC/MS-MS) Water-soluble metabolites were extracted from 100 mg tissue with 1 ml of ice-cold 80% (v/v) methanol and 0.6 ml acetonitrile. Samples were analyzed by LC/MS-MS9. MultiQuant v2.0

software (AB/SCIEX) was used for peak areas integration. LC/MS-MS was run for the 15 individual pig samples (15 independent runs). MetaboAnalyst 4.0 was used to identify known pathways. All

analyses are conducted in agreement with the biosafety regulations at the Rhode Island Hospital, Providence, RI. UNSUPERVISED MACHINE LEARNING (UML) Nonnegative matrix factorization based

UML was applied to analyze the mass spec data as we previously reported11. Hierarchical clustering was performed as in43,45. Linux clusters at the Los Alamos National Laboratory were used to

run the simulations. STATISTICAL ANALYSIS Microsoft Excel and Graphpad Prism7 software were used for data analysis. Differences between 2 datasets were statistically compared with _the

Student t_-test (GraphPad Software, Inc, San Diego, CA). We represent the data as means ± SD; _p_ < 0.05 was considered to be a statistically significant difference. Immunohistochemical

results are presented in pixels as average mean intensity pixels /40XHPF + / − SD or as a fold change vs. LD, analyzing six random fields for each pig. Western blot data are presented as

fold MS change vs. LD (pixels, +/− standard error of the mean, SD). DATA AVAILABILITY The RNA-Seq data accession number PRJNA544355 availability under GEO. The UML predicted metabolites with

higher probability in LD vs. MS (P4 = 0) are shown in the Supporting document (Table S1). The 283 polar metabolites in MS and LD are available upon request. REFERENCES * Sperling, L. S. _et

al_. The CardioMetabolic Health Alliance: Working Toward a New Care Model for the Metabolic Syndrome. _J. Am. Coll. Cardiol._ 66, 1050–1067 (2015). Article  Google Scholar  * Libert, D. M.,

Nowacki, A. S. & Natowicz, M. R. Metabolomic analysis of obesity, metabolic syndrome, and type 2 diabetes: amino acid and acylcarnitine levels change along a spectrum of metabolic

wellness. _PeerJ — J. Life Environ. Sci._ 6, 15 (2018). Google Scholar  * McMillen, I. C. & Robinson, J. S. Developmental origins of the metabolic syndrome: prediction, plasticity, and

programming. _Physiological Rev._ 85, 571–633 (2005). Article  CAS  Google Scholar  * Cheng, S. _et al_. Potential Impact and Study Considerations of Metabolomics in Cardiovascular Health

and Disease: A Scientific Statement From the American Heart Association. _Circulation: Cardiovascular Genet._ 10, e000032 (2017). Google Scholar  * Ahluwalia, N. _et al_. Trends in

overweight prevalence among 11-, 13- and 15-year-olds in 25 countries in Europe, Canada and USA from 2002 to 2010. _Eur. J. Public. Health_ 25, 28–32, https://doi.org/10.1093/eurpub/ckv016

(2015). Article  PubMed  Google Scholar  * Gerrity, R. G., Natarajan, R., Nadler, J. L. & Kimsey, T. Diabetes-induced accelerated atherosclerosis in swine. _Diabetes_ 50, 1654–1665

(2001). Article  CAS  Google Scholar  * Lee, L. _et al_. Nutritional model of steatohepatitis and metabolic syndrome in the Ossabaw miniature swine. _Hepatology_ 50, 56–67,

https://doi.org/10.1002/hep.22904. (2009). Article  CAS  PubMed  PubMed Central  Google Scholar  * Robich, M. P. _et al_. Resveratrol improves myocardial perfusion in a swine model of

hypercholesterolemia and chronic myocardial ischemia. _Circulation Res._ 122, S142–149 (2010). Article  CAS  Google Scholar  * Yuan, M., Breitkopf, S. B., Yang, X. & Asara, J. M. A

positive/negative ion-switching, targeted mass spectrometry-based metabolomics platform for bodily fluids, cells, and fresh and fixed tissue. _Nat. Protoc._ 7, 872–881 (2012). Article  CAS 

Google Scholar  * Cichocki, A., Zdunek, R., Phan, A. H. & Amari, S. I. _Nonnegative Matrix and Tensor Factorizations: Applications to Exploratory Multi‐Way Data Analysis and Blind Source

Separation_. (John Wiley & Sons, Ltd, 2009). * Alexandrov, L. B., Serena, N. Z., Wedge, D. C., Campbell, P. J. & Stratton, M. R. Deciphering Signatures of Mutational Processes

Operative in Human Cancer. _Cell Rep._ 3, 246–259 (2013). Article  CAS  Google Scholar  * Brunet, J. P., Tamayo, P., Golub, T. R. & Mesirov, J. P. Metagenes and molecular pattern

discovery using matrix factorization. _Proc. Natl Acad. Sci. USA_ 101, 64–4169 (2004). Article  Google Scholar  * Yoon, B. E. _et al_. Glial GABA, synthesized by monoamine oxidase B,

mediates tonic inhibition. _J. Physiol._ 592, 4951–4968 (2014). Article  CAS  Google Scholar  * Shen, H. _et al_. The Human Knockout Gene CLYBL Connects Itaconate to Vitamin B12. _Cell_ 171,

771–782 (2017). Article  CAS  Google Scholar  * Bulku, A., Aleksanyan, N. & Berkmen, M. Oligomerization of wild-type fumarase and a mutant associated with fumarate hydratase deficiency

2016 FASEB journal. _Biochemistry and Molecular Biology_ 30 (2016). * Lange, M. _et al_. Arterial territory-specific phosphorylated retinoblastoma protein species and CDK2 promote

differences in the vascular smooth muscle cell response to mitogens. _Cell cycle_ 13, 315–323 (2014). Article  CAS  Google Scholar  * Marsh, S. A., Collins, H. E. & Chatham, J. C.

Protein O-GlcNAcylation and Cardiovascular (Patho)physiology. _J. Biol. Chem._ 289, 34449–34456 (2014). Article  Google Scholar  * Ma, J. & Hart, G. W. Protein O-GlcNAcylation in

diabetes and diabetic complications. _Expert. Rev. Proteomics._ 10, 365–380 (2013). Article  CAS  Google Scholar  * Groves, J. A., Maduka, A. O., O’Meally, R. N., Cole, R. N. & Zachara,

N. E. Fatty acid synthase inhibits theO-GlcNAcase during oxidative stress. _J. Biol. Chem._ 292, 6493–6511 (2017). Article  CAS  Google Scholar  * Hiromura, M. _et al_. YY1 is regulated by

O-linked N-acetylglucosaminylation (O-glcNAcylation). _J. Biol. Chem._ 278, 14046–14052 (2003). Article  CAS  Google Scholar  * Dodd, M. S. _et al_. Impaired _in vivo_ mitochondrial Krebs

cycle activity after myocardial infarction assessed using hyperpolarized magnetic resonance spectroscopy. _Circ. Cardiovasc. Imaging._ 7, 895–904 (2014). Article  Google Scholar  * Gibala,

M. J., Young, M. E. & Taegtmeyer, H. Anaplerosis of the citric acid cycle: role in energy metabolism of heart and skeletal muscle. _Acta Physiol. Scand._ 168, 657–665 (2000). Article 

CAS  Google Scholar  * Kalousek, F., Darigo, M. D. & Rosenberg, L. E. Isolation and characterization of propionyl-CoA carboxylase from normal human liver. Evidence for a protomeric

tetramer of nonidentical subunits. _J. Biol. Chem._ 255, 60–65 (1980). CAS  PubMed  Google Scholar  * He, W. _et al_. Citric acid cycle intermediates as ligands for orphan G-protein-coupled

receptors. _Nat._ 429, 188–193 (2004). Article  ADS  CAS  Google Scholar  * Fröhlich, H. _et al_. Carvedilol Compared With Metoprolol Succinate in the Treatment and Prognosis of Patients

With Stable Chronic Heart Failure: Carvedilol or Metoprolol Evaluation Study. _Circ. Heart Fail._ 8, 887–896 (2015). Article  Google Scholar  * Laplante, A., Vincent, G., Poirier, M. &

Des Rosiers, C. Effects and metabolism of fumarate in the perfused rat heart. A 13C mass isotopomer study. _Am. J. Physiol._ 272, 74–82 (1997). Google Scholar  * Hewitson, K. S. _et al_.

Structural and mechanistic studies on the inhibition of the hypoxia-inducible transcription factor hydroxylases by tricarboxylic acid cycle intermediates. _J. Biol. Chem._ 282, 3293–3301

(2007). Article  CAS  Google Scholar  * Mescam, M., Vinnakota, K. C. & Beard, D. A. Identification of the catalytic mechanism and estimation of kinetic parameters for fumarase. _J. Biol.

Chem._ 286, 21100–21109 (2011). Article  CAS  Google Scholar  * Ohad, Y., Adi, N. & P., O. Fumarase: a paradigm of dual targeting and dual localized functions. _The FEBS Journal_ 278,

4230–4242 (19 September 2011). * Pereira de Pádua, R. A. & Nonato, M. C. Cloning, expression, purification, crystallization and preliminary X-ray diffraction analysis of recombinant

human fumarase. _Acta crystallographica. Sect. F, Struct. Biol. Commun._ 70, 120–122 (2014). Article  Google Scholar  * Ding, S., Yang, Y. & Mei, J. Protective Effects of L-Malate

against Myocardial Ischemia/Reperfusion Injury in Rats. _Evidence-Based Complementary Alternative Med._ 2016, 9 (2015). Google Scholar  * Dean, J. T. _et al_. Resistance to Diet-Induced

Obesity in Mice with Synthetic Glyoxylate Shunt. _Cell Metab._ 9, 525–536 (2009). Article  CAS  Google Scholar  * Vincent, G., Bouchard, B., Khairallah, M. & Des Rosiers, C. Differential

modulation of citrate synthesis and release by fatty acids in perfused working rat hearts. _Am. J. Physiol. Heart Circ. Physiol_ 286, H257–H266 (2004). Article  CAS  Google Scholar  *

Thomassen, A. R., Nielsen, T. T., Bagger, J. P. & Henningsen, P. Myocardial exchanges of glutamate, alanine and citrate in controls and patients with coronary artery disease. _Clin.

Sci._ 64, 33–40 (1983). Article  CAS  Google Scholar  * Garland, P. B., Randle, P. J. & Newsholme, E. A. Citrate as an intermediary in the inhibition of phosphofructokinase in rat heart

muscle by fatty acids, ketone bodies, pyruvate, diabetes and starvation. _Nat._ 200, 169–170 (1963). Article  ADS  CAS  Google Scholar  * Hassel, B., Ilebekk, A. & Tønnessen, T. Cardiac

accumulation of citrate during brief myocardial ischaemia and reperfusion in the pig _in vivo_ _Acta Physiol Scand_ 164 (1998). * Doenst, T., Nguyen, T. D. & Abel, E. D. Cardiac

metabolism in heart failure: implications beyond ATP production. _Circ. Res._ 113, 709–724 (2013). Article  CAS  Google Scholar  * Russell, R. R. & Taegtmeyer, H. Changes in citric acid

cycle flux and anaplerosis antedate the functional decline in isolated rat hearts utilizing acetoacetate. _J. Clin. Invest._ 87, 384–390 (1991). Article  CAS  Google Scholar  * Karlstaedt,

A. _et al_. Oncometabolite d-2-hydroxyglutarate impairs α-ketoglutarate dehydrogenase and contractile function in rodent heart. _Proc. Natl Acad. Sci. USA_ 113, 10436–10441,

https://doi.org/10.1073/pnas.1601650113 (2016). Article  CAS  PubMed  Google Scholar  * Ku, H. J., Ahn, Y., Lee, J. H., Park, K. M. & Park, J. W. IDH2 deficiency promotes mitochondrial

dysfunction and cardiac hypertrophy in mice. _Free. Radic. Biol. Med._ 80, 84–92, https://doi.org/10.1016/j.freeradbiomed.2014.12.018 (2015). Article  CAS  PubMed  Google Scholar  *

Elmadhun, N. Y. _et al_. The pig as a valuable model for testing the effect of resveratrol to prevent cardiovascular disease. _Ann.N.Y.Acad.Sci._ 1290, 130–135 (2013). Article  ADS  CAS 

Google Scholar  * Dobin, A. _et al_. ultrafast universal RNA-seq aligner. _Bioinforma._ 29, 15–21 (2013). Article  CAS  Google Scholar  * Anders, S. & Huber, W. Differential expression

analysis for sequence count data. _Genome Biol._ 11, R106 (2010). Article  CAS  Google Scholar  * Karimi, M. _et al_. Robust effect of metabolic syndrome on major metabolic pathways in the

myocardium. _PLOS ONE_ 14, 1–16 (2019). Google Scholar  * Nik-Zainal, S. _et al_. Mutational Processes Molding the Genomes of 21 Breast Cancers. _Cell_ 149, 979–993 (2012). Article  CAS 

Google Scholar  Download references ACKNOWLEDGEMENTS The authors are grateful to; Drs. Nassrene Elmadhun, Laura Scrimgeour, Brittany Plotz for conducting the pig surgery and cardiac tissue

dissection; Luke Sergott for helpful discussions and language accuracy suggestions; Nawal Yessuf to assist in the laboratory. This research project was supported by the RO1HL128831 (A.U.);

Los Alamos National Laboratory LDRD 20180060DR grant, DE-AC52-06NA25396 from the US Department of Energy National Nuclear Security to the Los Alamos National Laboratory (B.S.A.); RO1HL128831

(F.W.S.); 5P30CA006516 to A.M.J. and the BIDMC Research Capital Fund for funding the mass spectrometry instrumentation (QTRAP 5500 and 6500). AUTHOR INFORMATION AUTHORS AND AFFILIATIONS *

Division of Cardiothoracic Surgery, Department of Surgery, The Warren Alpert Medical School, Brown University, Providence, RI, 02903, United States Maryam Karimi, Frank W. Sellke & Anny

Usheva * Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA, 02115, United States Victoria Petkova & John M. Asara * Sam Houston State University, College of

Osteopathic Medicine, Huntsville, TX, 77320, United States Michael J. Griffin * Los Alamos National Laboratory, Los Alamos, NM, 87545, United States Alan R. Bishop & Boian S. Alexandrov

Authors * Maryam Karimi View author publications You can also search for this author inPubMed Google Scholar * Victoria Petkova View author publications You can also search for this author

inPubMed Google Scholar * John M. Asara View author publications You can also search for this author inPubMed Google Scholar * Michael J. Griffin View author publications You can also search

for this author inPubMed Google Scholar * Frank W. Sellke View author publications You can also search for this author inPubMed Google Scholar * Alan R. Bishop View author publications You

can also search for this author inPubMed Google Scholar * Boian S. Alexandrov View author publications You can also search for this author inPubMed Google Scholar * Anny Usheva View author

publications You can also search for this author inPubMed Google Scholar CONTRIBUTIONS K.M. conducted bioinformatics and metabolite extractions; S.W.F. physiological data, P.V. assisted with

RNA-sec data evaluation; A.M.J. established the mass spec procedure; G.M. data evaluation and discussion; A.R.B. and B.S.A. conducted UML data analysis; U.A. supervised the project,

designed metabolomics, UML and transcriptomics experiments, and wrote the manuscript with input from G.M., B.S.A. All authors reviewed the manuscript. CORRESPONDING AUTHOR Correspondence to

Anny Usheva. ETHICS DECLARATIONS COMPETING INTERESTS The authors declare no competing interests. ADDITIONAL INFORMATION PUBLISHER’S NOTE Springer Nature remains neutral with regard to

jurisdictional claims in published maps and institutional affiliations. SUPPLEMENTARY INFORMATION SUPPLEMENTARY INFORMATION. RIGHTS AND PERMISSIONS OPEN ACCESS This article is licensed under

a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate

credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article

are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and

your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this

license, visit http://creativecommons.org/licenses/by/4.0/. Reprints and permissions ABOUT THIS ARTICLE CITE THIS ARTICLE Karimi, M., Petkova, V., Asara, J.M. _et al._ Metabolomics and the

pig model reveal aberrant cardiac energy metabolism in metabolic syndrome. _Sci Rep_ 10, 3483 (2020). https://doi.org/10.1038/s41598-020-60387-7 Download citation * Received: 05 July 2019 *

Accepted: 03 February 2020 * Published: 26 February 2020 * DOI: https://doi.org/10.1038/s41598-020-60387-7 SHARE THIS ARTICLE Anyone you share the following link with will be able to read

this content: Get shareable link Sorry, a shareable link is not currently available for this article. Copy to clipboard Provided by the Springer Nature SharedIt content-sharing initiative