Metabolomics reveals the impact of type 2 diabetes on local muscle and vascular responses to ischemic stress.

Abstract

Objective:

Type 2 Diabetes Mellitus (T2DM) reduces exercise capacity, but the mechanisms are incompletely understood. We probed the impact of ischemic stress on skeletal muscle metabolite signatures and T2DM-related vascular dysfunction.

Methods:

We recruited 38 subjects (18 healthy, 20 T2DM), placed an antecubital intravenous catheter, and performed ipsilateral brachial artery reactivity testing. Blood samples for plasma metabolite profiling were obtained at baseline and immediately upon cuff release after 5 minutes of ischemia. Brachial artery diameter was measured at baseline and 1 minute after cuff release.

Results:

As expected, flow-mediated vasodilation was attenuated in subjects with T2DM (p<0.01). We confirmed known T2DM-associated baseline differences in plasma metabolites, including homocysteine, dimethylguanidino valeric acid and β-alanine (all p<0.05). Ischemia-induced metabolite changes that differed between groups included 5-hydroxyindoleacetic acid (Healthy: −27%; DM +14%), orotic acid (Healthy: +5%; DM −7%), trimethylamine-N-oxide (Healthy: −51%; DM +0.2%), and glyoxylic acid (Healthy: +19%; DM −6%) (all p<0.05). Levels of serine, betaine, beta-aminoisobutyric acid and anthranilic acid were associated with vessel diameter at baseline, but only in T2DM (all p<0.05). Metabolite responses to ischemia were significantly associated with vasodilation extent, but primarily observed in T2DM, and included enrichment in phospholipid metabolism (p<0.05).

Conclusions:

Our study highlights impairments in muscle and vascular signaling at rest and during ischemic stress in T2DM. While metabolites change in both healthy and T2DM subjects in response to ischemia, the relationship between muscle metabolism and vascular function is modified in T2DM, suggesting that dysregulated muscle metabolism in T2DM may have direct effects on vascular function.

INTRODUCTION

Individuals with Type 2 Diabetes Mellitus (T2DM) have significantly reduced exercise capacity compared to age-matched healthy individuals[1]. It is thought that dysregulated metabolism in T2DM may contribute to the attenuation in functional capacity. Systemic blood metabolomic signatures are known to be altered in T2DM compared with healthy subjects, even before the onset of overt disease[2-11]. However, little is known about how local tissue metabolomic signatures differ in T2DM.

It is known that T2DM impairs the blood flow response to exercise, to infusion of endothelium-dependent vasodilators, and to flow-mediated vasodilation[1, 12-14], but the impact of local blood metabolomic signatures on vascular function has not been studied. The T2DM-associated vascular pathophysiological state may be affected by metabolite production and signaling.

We hypothesized that biochemical profiling of blood might illuminate specific pathways underlying altered skeletal muscle metabolism and maladapted vascular response to ischemia in the setting of T2DM. We investigated the local plasma metabolome of the upper extremity and upper extremity vascular function, at rest and after experimental ischemia in T2DM (n=18) and healthy control subjects (n=20). We induced experimental ischemia by inflating a sphygmomanometric cuff on the upper arm, and obtained samples from the cubital vein immediately below the cuff, reflecting local arm skeletal muscle metabolism at rest and after ischemia. We identified local limb metabolite signatures of rest and the response to ischemia, and observed relationships between metabolites and the vascular response in subjects with T2DM compared to healthy subjects.

SUBJECTS AND METHODS

Subjects

Thirty-eight subjects, including 18 subjects with T2DM and 20 healthy, nonsmoking controls were recruited through newspaper advertisement. All subjects underwent screening, medical history, physical examination, and laboratory analysis including complete blood count, serum electrolytes, glucose, blood urea nitrogen, creatinine, transaminases, alkaline phosphatase, total cholesterol, and LDL. Healthy subjects with hypertension, diabetes mellitus, LDL or total cholesterol greater than the 50th percentile for age and gender, cardiovascular disease, or other significant disease were excluded from the control group. All participants provided written informed consent. The protocol was approved by the Human Research Committees of the Brigham and Women’s Hospital.

Study Design

The effect of ischemia on skeletal muscle metabolism and endothelium-dependent vasodilation of the brachial artery was studied using a cross-sectional design. All subjects were studied in the morning in the post-absorptive state, fasting since the previous midnight. Cyclooxygenase inhibitors, alcohol, and caffeine were prohibited for 24 hours before the study. Upon arrival to the laboratory, subjects lay supine and underwent cubital vein intravenous catheterization. After a minimum of 15 minutes of rest, baseline venous blood samples were obtained. Next, baseline brachial artery diameter was measured. Then, a sphygmomanometric cuff was placed on the upper arm and inflated to suprasystolic blood pressure for 5 minutes (“ischemic challenge”). Venous blood samples were obtained at 5 seconds and 55 seconds after cuff release. Brachial artery diameter was measured from 60 to 70 seconds after cuff release.

Vascular Reactivity Studies

Brachial artery vascular function was assessed according to standard methods and as we have previously performed[15-18]. Participants were studied in a quiet, temperature-controlled, dimly lit room after resting supine for a minimum of 15 minutes. High-resolution B-mode ultrasonography of the brachial artery was performed with a 7.5-MHz linear array probe (GE VIVID 7; GE Healthcare, Chicago, IL). The brachial artery was imaged longitudinally just proximal to the antecubital fossa. The transducer position was adjusted to obtain optimal images of the near and far walls of the intima. The video output and electrocardiographic signal of the ultrasound machine were connected to a computer equipped with a Data Translation Frame-Grabber video card (Dataviz, Westport, CT). The R wave on the electrocardiogram served as a trigger to acquire frames at end diastole. After baseline image acquisition, a sphygmomanometric cuff placed on the upper arm was inflated to suprasystolic pressure (~200 mm Hg) for 5 minutes. Flow-stimulated endothelium-dependent vasodilation of the brachial artery was determined by acquiring images from 60 to 70 seconds after cuff deflation. This timepoint was selected based on previous work; flow-mediated vasodilation at this time point is largely endothelium dependent and nitric oxide mediated and can be inhibited by administration of the nitric oxide synthase antagonist NG-monomethyl-L –arginine[19,20]. Ten minutes after cuff release, the brachial artery was imaged again to reestablish basal conditions.

Metabolomics

Metabolite measurements were performed on stored plasma samples. Metabolite profiling of 158 metabolites was performed as previously described[2,6,9], using two well-established, complementary methods to capture metabolites that are measured by mass spectrometry in the positive and negative ion modes[2,21]. Together, these methods measure a reasonably broad swath of key metabolic pathways, including amino acids and amines, acylcarnitines, organic acids and nucleotides with excellent precision (CVs generally <10%) in small sample volumes (40 microliters). Thus, our overall goal was to efficiently capture as many “sentinels” as possible from these key pathways including many low abundance compounds, though when setting up the platform we did enrich for metabolites associated with cardiovascular disease and vascular biology (e.g., nitric oxide metabolites). Previously aliquoted plasma samples, stored at −80°C since collection, were thawed and centrifuged for 5 min at 1,500 rpm. All samples were exposed to one freeze-thaw cycle before measurement. Briefly, central metabolites including sugars, sugar phosphates, organic acids, purine, and pyrimidines were extracted from 30 μL of plasma using acetonitrile and methanol and separated using a 100 x 2.1 mm XBridge Amide column (Waters). A high sensitivity Agilent 6490 QQQ MS (Agilent) was used to profile metabolites in the negative ion mode via multiple reaction monitoring (MRM) scanning. We used hydrophilic interaction liquid chromatography coupled to tandem mass spectrometry (MS) to analyze polar metabolites in the positive ion mode in a targeted manner. This platform assesses amino acids, amines, acylcarnitines, nucleotides, and other compound classes. All liquid chromatography–MS analyses were performed using a 4000 QTRAP triple quadrupole mass spectrometer (AB Sciex, Foster City, CA) coupled to either an 1100 Series pump (Agilent Technologies, Santa Clara, CA) or an HTS PAL autosampler (Leap Technologies, Carrboro, NC) equipped with a column heater. Raw data were processed using MassHunter Quantitative Analysis Software (Agilent). Metabolite measurements were normalized to pooled plasma samples and internal standards. After removing known drug metabolites and duplicates, 138 metabolites were analyzed in the current study.

Statistical Methods

Descriptive and anthropomorphic measures are reported as mean ± SD. Wilcoxon signed rank tests were conducted to examine the baseline differences in plasma metabolites between T2DM and healthy subjects. Paired t tests were conducted to examine the change in metabolites following ischemia for the overall sample and by T2DM status. To examine the inter-relationships between metabolite changes in response to ischemia, we calculated the change in metabolite from baseline to 5-seconds post-ischemia within T2DM and healthy participants, and calculated Pearson correlation analysis on the change across all metabolites measured. We visualized these results in a heatmap and identified patterns in the correlation structure suggesting differences between T2DM and healthy individuals. Regression analyses were conducted to examine the relationship between metabolite changes and change in vessel diameter. Regression models were adjusted for age, sex, baseline metabolites, and baseline vessel diameter. We additionally included T2DM and T2DM by metabolite change interaction in the regression models to explore different associations for healthy and T2DM subjects. Secondary models included additional adjustment for smoking history and BMI to account for heterogeneity between groups. Adjustment for multiple testing in metabolomics studies is challenging, and there is currently no widely-accepted consensus approach[22,23]. We chose to use unadjusted p<0.05 as our threshold for suggestive biological relevance that may merit further consideration[24], and we present these data to the reader for their interpretation. We consider p<0.017 as the threshold for Bonferroni correction for multiple comparisons (Baseline, Post-ischemia, and Vasodilation). We also present results for the Benjamini-Hochberg (BH) false discovery rate adjusted p values, to further inform the reader. MetaboAnalyst[25,26] was used for functional enrichment and pathway analysis of significant metabolites. For each enrichment analysis, the suggestive (unadjusted p<0.05) metabolites were entered and compared to all platform metabolites (N=138) as background.

RESULTS

Subject characteristics are presented in Table 1. As expected, healthy subjects were leaner than subjects with T2DM, and had lower heart rate, mean arterial pressure, fasting glucose, HDL cholesterol, and triglycerides at baseline (P<0.05). Because healthy subjects were younger than T2DM, all statistical models were adjusted for age, with secondary models adjusting for smoking history and BMI.

Table 1:

Demographics of the Study Subjects

	Healthy Subjects (SD)	Type 2 Diabetes Mellitus Subjects (SD)	p
N	20	18
Age (y)	46±11	58±8	<0.001
Female	45%	28%	0.27
BMI	26±3	30±5	0.003
MAP (mm Hg)	91±10	99±15	0.03
HR (BPM)	66±8	74±16	0.03
Former Smokers	21%	50%	0.065
Creatinine (mg/dl)	0.8±0.1	0.8±0.2	0.816
Fasting glucose (mg/dl)	80±10	136±45	<0.001
Total cholesterol (mg/dl)	158±31	155±29	0.418
LDL (mg/dl)	93±26	82±22	0.161
HDL (mg/dl)	48±15	39±7	0.072
Triglycerides (mg/dl)	86±59	182±166	0.016

Abbreviations: BMI: body mass index; MAP: mean arterial pressure; HR: heart rate; BPM: beats per minute; LDL: low density lipoprotein; HDL: high density lipoprotein

P values from Wilcoxon test for continuous variables and Pearson test for categorical variables (sex and smoking).

Resting limb metabolomic signatures

At baseline, 16 of 138 measured metabolites (12%) were significantly different between T2DM and healthy control subjects at p<0.05, with 11 metabolites significant at Bonferroni-adjusted p<0.017 (Supplement Table S1)These included metabolite changes previously associated with diabetes or cardiometabolic disease phenotypes including lower homocysteine (p=0.003)[11], higher dimethylguanidino valeric acid (DMGV, p=0.028)[7], lower beta alanine (p=0.001)[10], and higher gluconic acid (p=0.049)[5] in T2DM compared to healthy control subjects (all p<0.05). There was a trend for the expected differences in other previously reported diabetes-related metabolites including higher branched-chain amino acids valine, leucine and isoleucine[2], and 2-aminoadipic acid[8] in diabetes compared with healthy subjects, although these did not reach statistical significance in our sample.

Skeletal muscle ischemia and metabolism

We examined the change in plasma metabolite profiles following experimental ischemia in healthy and T2DM subjects. Following cuff release, 33 of 138 (24%) metabolites were significantly altered immediately (11 using BH adj. P), while 24 of 138 (17%) metabolites were significantly altered at one-minute following cuff release (8 using BH adj. P), compared with baseline (Supplement Table S2). Consistent with expected effects of ischemia, there were changes in markers of glycolysis/gluconeogenesis, including pyruvic acid, lactic acid, and sugar (composite measure of glucose/fructose/galactose) (all p<0.001). Many of the metabolites which were altered immediately after ischemia had returned to baseline levels after one minute, and there was limited overlap between metabolites altered at 1 minute vs. those changed immediately after cuff release (7 metabolites), highlighting the rapid and dynamic nature of the metabolic changes.

The ischemia-induced metabolite changes in the arm differed between T2DM and healthy subjects. Specific examples are shown in Figure 1 (all p<0.05), with all changes shown in Supplementary Figure S1. For the branched-chain amino acid metabolite C5-valerylcarnitine (Figure 1A; Healthy: −25%; T2DM −7%), and the atherosclerosis marker trimethylamine-N-oxide (Figure 1B; Healthy: −51%; T2DM +0.2%), there was a significant reduction in healthy subjects, but little change in subjects with T2DM. For glyoxylic acid (Figure 1C), there was a significant increase in healthy subjects, with almost no change in T2DM (Healthy: +19%; T2DM −6%). Notably, some metabolites showed significant changes in different directions in healthy and T2DM, most notably 5-hydroxyindoleacetic acid (Figure 1D; Healthy: −27%; T2DM +14%). Taken together, these data support a role for altered dynamic metabolism within the skeletal muscle vasculature in the setting of diabetes.

Figure 1. Metabolites with strikingly different disease-related profiles in healthy and T2DM.

C5-Valerylcarnitine (A) and Trimethylamine-N-oxide (B) decreased post-ischemia in health subjects with no change in subjects with diabetes. Glyoxylic acid (C) increased post-ischemia in healthy subjects with no change in diabetes. 5-hydroxyindoleacetic acid (5-HIAA) (D) changed in opposite directions in healthy and diabetes subjects.

Correlated metabolite changes that differ between T2DM and healthy subjects reflect potential impairments in biological pathways.

We examined correlation structure between the change in metabolites in healthy and T2DM subjects to explore whether metabolite changes were inter-related, or if they appeared to change independently of each other. As shown in Figure 2, we observed striking differences in how metabolite changes were correlated from baseline to immediately following ischemia. In T2DM, there was evidence for strong inter-relationships between the individual metabolite responses to ischemia (Figure 2A), which was not evident in healthy subjects (Figure 2B). The correlated metabolites in T2DM appear to reflect shared pathways, with clusters of related metabolites correlating strongly within their group (e.g. amino acids, carnitines). Interestingly, many of the metabolites which were not strongly inter-correlated, or inversely correlated, were related to pyruvate metabolism. As shown in Figure 3, subjects with T2DM had significant differences in the change in pyruvic acid, lactic acid, glutamate, alanine, and glyoxylic acid compared with healthy subjects. These metabolites feed into key pathways in energy metabolism and RNA synthesis, and may point towards signaling pathways that are defective in the response to ischemic stress in diabetes.

Figure 2: Heatmaps depicting the inter-relationship between changes in metabolites post-ischemia.

Metabolite changes from baseline to 5-seconds post-ischemia were highly inter-correlated in subjects with T2DM (A), but not in healthy subjects (B)

Figure 3.

Defects in Pyruvic Acid and related pathways may underlie the dysregulated ischemic response in T2DM

Endothelial function and local metabolic signatures after ischemia

Baseline brachial artery diameter did not differ by group and was 3.7 ± 0.7 mm in the healthy subjects and 3.8 ± 0.6 mm in the subjects with T2DM (p>0.2). Flow mediated vasodilation was significantly attenuated in subjects with T2DM compared to healthy control subjects, 6.8% ± 1.6% vs 11.3% ± 2.6%, respectively (p < 0.01) (Figure 4). We hypothesized that local changes in metabolite signaling in response to ischemia would relate to vasodilation in a disease-dependent manner. The change in vessel diameter from rest to 60 seconds after cuff release was significantly correlated with metabolite changes (from baseline to 5 seconds) in a disease-dependent manner. In healthy subjects, only one metabolite (phosphocholine) nominally associated with vasodilation. In contrast, 25 of 138 (18%) metabolite changes associated with vasodilation in T2DM (p<0.05, none reached BH adj. significance). Because vascular function could be affected by obesity or smoking history, which differed between healthy and T2DM, we ran secondary models including BMI and smoking history (in addition to age, sex, and baseline parameters), and observed very similar results. Notably, while change in phosphocholine was significantly correlated with vasodilation in both groups, the effects were in the opposite direction (Figure 5). Because we have examined endothelium-dependent, flow-mediated vasodilation, it is notable that another endothelium-dependent vasodilator, acetylcholine, associates indirectly with the change in vessel diameter. In addition, the phosphocholine dysmetabolism in type 2 diabetes show a consistent association with the change in brachial artery diameter not found in healthy control subjects (Figure 5). Moreover, several pathways intersect with endothelium dependent vasodilation including branched-chain amino acids.

Figure 4.

Brachial artery flow-mediated and endothelium-independent vasodilation was lower in subjects with T2DM than healthy controls.

Figure 5. Evidence for an impaired relationship between phosphocholine metabolism and vasodilation in T2DM.

* Significant association between change in relative plasma metabolite concentration and change in vessel diameter in T2DM ** Significant association between change in metabolite and change in vessel diameter in T2DM and Healthy. ER = Endoplasmic reticulum.

DISCUSSION

T2DM is associated with considerable cardiovascular risk and vascular dysfunction. Individuals with T2DM have differences in resting metabolite profiles, however, whether these differences are associated with acute metabolic responses was unknown. We examined metabolite profiles and conduit artery endothelium-dependent vasodilation during the dynamic response to transient ischemia in healthy and T2DM subjects. Consistent with our hypothesis, we found evidence of dysregulated metabolism in response to muscle ischemia in individuals with T2DM.These data suggest that altered metabolomic signatures in response to ischemia may coordinate an abnormal tissue and attendant vascular response in the setting of T2DM.

Several large cohort studies have found differences in circulating metabolites between healthy and T2DM subjects[2-5,7-11]. Our sample extended many of the expected differences in previously implicated metabolites to the limb, indicating that metabolic dysfunction was established in the T2DM subjects in our sample. Following ischemic challenge, we observed immediate differences in metabolites compared with baseline, with >20% of measured metabolites significantly altered compared with the resting condition. Within one minute, many of these metabolites returned to baseline levels, highlighting the rapid and dynamic nature of these changes. Metabolites that changed in response to ischemia have known function in energy metabolism, including glycolysis and gluconeogenesis, consistent with a shift towards anaerobic glycolysis in response to reduced oxygen availability[27].

We were particularly interested in differences in metabolic responses between healthy and T2DM subjects, to understand whether maladaptive changes in metabolism participate in diabetic complications. Sugars (glucose, fructose and sucrose) declined far more in healthy subjects than in subjects with T2DM, supporting the idea that nutrient uptake by muscle may be a limiting factor in T2DM. Consistent with that observation, other metabolites, including C5-valerylcarnitine and the known atherogenic metabolite trimethylamine-N-oxide[28] both declined significantly in healthy subjects, but were unchanged in T2DM. Glyoxylic acid, an advanced glycation end-product previously implicated as a marker of diabetes in human and animal studies[29,30], increased significantly in healthy subjects, while remaining unchanged in T2DM. Our data suggest that in subjects with T2DM there is a reduced skeletal muscle metabolic flexibility to respond to changes in physiological conditions. Others have reported a preserved flexibility using systemic measurements[31], however the absence of ischemia-induced changes in glucose and other metabolites suggest an inability to respond during ischemia. Supporting this, a previous study in mice found that obesity was associated with reduced capacity to increase glycolysis in response to intermittent hypoxia, and that this was associated with impaired cardiac responsiveness[32].

We observed a striking difference in the serotonin metabolite 5-hydroxyindoleacetic acid (5-HIAA), which was significantly decreased post-ischemia in healthy subjects, but was significantly increased in T2DM. 5-HIAA has previously been reported to be elevated in persons with Metabolic Syndrome[33]. Serotonin metabolism has been studied in animal models in relation to myocardial ischemia-reperfusion, with serotonin catabolism (and 5-HIAA accumulation) suppressed during ischemia, but increased during reperfusion[34,35]. Serotonin metabolism has long been linked to vascular function[36], with serotonin interacting with endothelin to modulate vascular signaling[37]. Our data support a role for serotonin and 5-HIAA in modulating the ischemic response, compensating for the reduction in vasodilation, and extends these findings to skeletal muscle.

We examined the inter-relatedness of metabolite changes, to address the hypothesis that T2DM associates with broad dysfunction in multiple pathways. In subjects with T2DM, there was much greater correlation between metabolite changes post-ischemia, compared with healthy subjects. We suggest that this may be due to dysfunction in key pathway nodes, which affect entire downstream pathways, such as amino acid and carnitine metabolism. Within healthy subjects, the relatively lower degree of correlation between metabolite changes may suggest an appropriate ability to regulate individual metabolites in response to ischemia as needed.

Pyruvate metabolism emerged as a key pathway differentiating T2DM and healthy subjects. Subjects with T2DM had significant differences in the change in pyruvic acid, lactic acid, glutamate, alanine, and glyoxylic acid compared with healthy subjects. These metabolites are crucial in glycolysis, and indicate that the ability to regulate energy metabolism may be defective in disease[38]. Immediately following cuff release, there were greater decreases in pyruvic acid and lactic acid in healthy subjects compared with T2DM. At one minute post cuff release, there was an increase in both pyruvic and lactic acid in T2DM compared with baseline, while in healthy subjects both remained below baseline levels. These suggest differences in the diversion of pyruvic acid towards mitochondrial oxidative metabolism versus lactic acid metabolism or gluconeogenesis[39]. This may result from the metabolic inflexibility known to occur in T2DM[40], and be a consequence of lipotoxicity, or may function as a protective adaptation to reduce accumulation of ROS.

We observed significant associations between acute changes in metabolites and the change in brachial artery diameter in response to 5 minutes of ischemia. However, these associations were predominately in T2DM, with limited evidence for association in healthy subjects. The relative difference between healthy and T2DM subjects in the relationships between vasodilation and changes in metabolites may indicate that healthy individuals are able to mount an appropriate vasodilatory response irrespective of local muscle metabolism. The dysregulated metabolic state in T2DM may preclude an appropriate vasodilatory response to ischemic stress. Phosphocholine, previously associated with the progression towards diabetes in animals[41], as well as with inflammation [42,43], had inverse relationships in T2DM compared with healthy subjects, suggesting an important modulating role in vascular response that may be dysregulated in T2DM.

Our study has several strengths, but also some limitations. To our knowledge, this is the first study of metabolite responses to limb ischemia in healthy and T2DM subjects. Although the sample size was modest, use of an evoked phenotype greatly increases power to detect physiologically-relevant metabolites, even in a relatively small sample. In our sample, healthy subjects were younger and leaner than subjects with T2DM, and had lower cardiometabolic risk profiles at baseline, including a lower proportion of ex-smokers. Observed differences in metabolite profiles could be attributable to these or other differences in risk factors independent of T2DM status. However we expect that the biological differences in the ischemic responses between healthy and individuals with T2DM are informative with acknowledged differences in other risk contributors, and thus our data provide insight into a disease milieu typical of, but perhaps not specifically limited to, a diagnosis of T2DM. Given the exploratory nature of our study, we used a liberal threshold of unadjusted p<0.05 to prioritize metabolites of potential interest, but note that for some analyses no metabolites remained significant after adjustment for multiple testing. Further studies are required to validate these findings in independent samples, and to confirm the precise role of the highlighted signaling pathways in modulating disease processes and reduced exercise capacity. Our study demonstrates the value of rapid and localized sampling in an evoked phenotype setting to uncover the underlying biology of acute metabolic responses.

In conclusion, we have identified metabolomic signatures of vascular function, that suggest a novel pathway of pyruvate- and phosphocholine-mediated vascular dysfunction in T2DM beyond known challenges of oxidative stress, inflammation, hyperglycemia, and free fatty acid excess.

Clinical Perspectives.

• Individuals with Type 2 Diabetes Mellitus (T2DM) have reduced exercise capacity, potentially mediated by maladaptive responses to ischemic stress; however the mechanisms are incompletely understood.

• We probed the impact of ischemic stress on skeletal muscle metabolite signatures and T2DM-related vascular dysfunction, and found impairments in muscle and vascular metabolite signaling at rest and during ischemic stress in T2DM, suggesting that dysregulated muscle metabolism in T2DM, particularly in pyruvate and phosphocholine metabolism, may have direct effects on vascular function.

• Therapies aimed at correcting metabolic perturbations may allow for improved disease treatment and management in T2DM.