Abstract
Hyperaminoacidemia is an early hallmark of insulin resistance, with aromatic and branched chain amino acids particularly associated with insulin resistance and type 2 diabetes. We previously showed that healthy adults with obesity exposed to acute hyperglycemia have lower brain glucose levels measured by magnetic resonance spectroscopy (MRS) than lean controls, suggesting that a blunted brain response to hyperglycemia may be an early marker of insulin resistance. Here, in a secondary analysis of our prior study, we used targeted mass spectrometry-based metabolomics to measure plasma amino acids in participants with and without obesity to determine if changes in peripheral metabolites associated with early insulin resistance such as amino acids were associated with changes in brain glucose levels during hyperglycemia. There were few differences in baseline amino acids between groups, but acute hyperglycemia unveiled higher plasma concentrations of amino acids including cysteine, cystine, glutamic acid, glutamine, methionine, and aromatic amino acids in obesity. Plasma glucagon levels were also higher in obesity during acute hyperglycemia. Higher plasma concentrations of aromatic amino acids and glucagon were significantly correlated with lower brain glucose levels, illustrating parallel development of central and peripheral metabolic changes in obesity.
Keywords: obesity, amino acids, glucagon, brain glucose, magnetic resonance spectroscopy
Graphical Abstract
New & Noteworthy
We related early insulin resistance-associated peripheral factors with brain glucose measured by 13C magnetic resonance spectroscopy during acute hyperglycemia in young, healthy adults with and without obesity. Plasma amino acids including aromatic amino acids and glucagon were higher in obesity during acute hyperglycemia. There were negative correlations between aromatic amino acids and glucagon with the change in brain glucose. These findings may be related to brain oxidative stress and neurotransmitter synthesis in obesity.
Introduction
The incidence and prevalence of obesity and type 2 diabetes continue to increase at alarming rates, with some estimates projecting that nearly half of Americans will have obesity by 2030 (1), and that more than 1 billion people worldwide will have diabetes by 2050 (2). Furthermore, the sequence of events by which metabolic dysfunction emerges in obesity and translates into risk for type 2 diabetes remains unclear and, if better understood, could advance our understanding of its pathophysiology and treatment.
Changes in the central nervous system (CNS) may be among the earliest pathophysiological changes in obesity, which can include altered brain structure, function, and metabolism (3). Recently published data from our group showed that, compared to lean controls, adults with obesity exhibited nearly 20% lower brain glucose uptake during hyperglycemia, as measured by magnetic resonance spectroscopy (MRS) (4). Notably, these changes occurred in young, otherwise healthy individuals and were observed despite the absence of any overt clinical signs of metabolic dysfunction, including traditional measures of insulin resistance calculated from oral glucose tolerance testing. As obesity is the most significant risk factor for insulin resistance (5), here we have investigated other factors that may be associated with central changes in these participants by looking for additional subtle signs of peripheral insulin resistance that may be revealed during a metabolic stressor such as acute hyperglycemia.
Alterations in plasma amino acid signatures and glucagon have been associated with insulin resistance. Changes in aromatic amino acids and branched chain amino acids (BCAA) were associated with obesity and insulin resistance as early as the 1960s (6–8), but whether these profiles are early manifestations of or contributors to insulin resistance remains under investigation (9). Amino acids stimulate pancreatic α-cell secretion of glucagon, which promotes amino acid uptake and catabolism in the liver. Amino acids are also transported from the periphery across the blood-brain barrier (BBB) for utilization in glucose-fueled neurotransmitter synthesis. The deleterious effects of excess concentrations of aromatic amino acids or BCAA are most dramatically illustrated by phenylketonuria and maple syrup urine disease, inborn errors of metabolism caused by mutations in phenylalanine hydroxylase and branched chain ketoacid dehydrogenase complex enzymes, leading to elevations of phenylalanine and BCAA, respectively. Accumulation of these amino acids can lead to devastating neurotoxicity (10). In other studies, supplementation of obesity-inducing diets with extra BCAA results in increases in anxiety-like behaviors in pre-clinical models (11). As the interplay between CNS and peripheral metabolism is increasingly appreciated, it becomes important to examine both the CNS and the periphery in parallel at singular time points along the path to metabolic dysfunction. Accordingly, here we have investigated the dynamic changes in plasma amino acids and glucagon during hyperglycemia and their association with brain glucose in a cohort of young adults with metabolically healthy obesity.
Materials and Methods
Human subjects
The study was approved by the Yale University Human Investigation Committee, and all participants provided written informed consent. Detailed inclusion and exclusion criteria and screening procedures were described previously (4). Briefly, lean adults with BMI 18–25 kg/m2 (n=11) and adults with obesity with BMI greater than 30 kg/m2 up to the weight limit of the MRI scanner (n=10) were included in the parent study. All participants were without chronic medical conditions and prescription medication use apart from hormonal contraception. All participants underwent a standard 75-gram oral glucose tolerance test (OGTT) (on a separate day) prior to hyperglycemic clamp and 13C MRS scanning. OGTT data were used to calculate: homeostatic model assessment for insulin resistance (HOMA-IR) [(fasting insulin (μU/L) × fasting glucose (nmol/L))/22.5]; Matsuda Index (12); insulinogenic index [Δ plasma insulin (μU/L) / Δ plasma glucose (mmol/L) during the first 30 minutes of the OGTT]; and disposition index (insulinogenic index × Matsuda Index). Participants from each group in the parent study (4) who had available plasma samples (n=8/group) were included in the present study (Figure 1).
Figure 1. CONSORT diagram illustrating participants who completed the parent study (4) and had available samples for secondary analyses presented in the current study.
BMI, body mass index; MRS, magnetic resonance spectroscopy.
Hyperglycemic clamp and 13C MRS scanning
After an overnight fast, participants arrived in the morning at the Yale Magnetic Resonance Research Center. Participants were prepared for concurrent glucose infusion and MRS scanning by inserting an intravenous (IV) catheter into the distal aspect of each arm – one for blood sampling and one for 20% [1-13C] glucose infusion – and then placed in a 4.0 Tesla whole-body magnet. Details of 13C MRS protocols and analyses were described in the parent study (4). Briefly, after tuning, voxel selection, shimming, and acquisition of baseline MRS data, the hyperglycemic clamp was initiated using a variable rate IV infusion of 20% [1-13C] glucose to achieve and maintain a target blood glucose of 180 mg/dL for 120 minutes (13). Steady state plasma percent enrichment was similar between groups (BMI<25: 78.3 ± 18.9%; BMI>30: 82.3 ± 7.4%; p=0.72). Plasma glucose was measured every 5 minutes during the hyperglycemic clamp. Plasma was also collected at 0, 2, 4, 6, 8, 10, 30, 60, 90, and 120 minutes and stored at −80°C to −20° for targeted metabolomics and other hormonal analyses. MRS data were acquired continuously for the duration of the hyperglycemic clamp to quantify the change in brain glucose and calculate the maximum rate of glucose transport (Tmax) relative to the cerebral metabolic rate of glucose (CMRgl) using the following equation:
where Vd is the volume of distribution of the brain water space [0.77 mL/g (14)], Go and Gi are the steady-state plasma and brain glucose concentrations, respectively (mM), and KT is the Michaelis-Menten constant for the glucose transporter [1.1 mM (15)]. The period of hyperglycemia was terminated prior to 120 minutes for two participants (one in each group), therefore these participants were excluded from analyses incorporating data from plasma samples collected at 120 minutes (Figure 1). MRS data from one participant with class 2 obesity had signal-to-noise ratios that did not meet quality control criteria, therefore this participant’s spectroscopy data were excluded (Figure 1).
Targeted metabolomics
Amino acids were measured by liquid chromatography-tandem mass spectrometry by the Duke Molecular Physiology Institute’s Metabolomics Core Laboratory, using a method adapted from the literature (16). Ten μl of plasma was spiked with a mixture of stable isotope-labeled internal standards and deproteinized with methanol. The supernatants were derivatized with AccQTag reagent (Biosynth-Carbosynth) at 55 °C for 10 minutes. Chromatographic separation was performed using a Waters Acquity UPLC system (Milford, MA) equipped with a Waters Acquity UPLC HSS T3 column (1.8 μm, 2.1 × 100 mm). Mobile phase A consisted of 0.1% formic acid in water, and mobile phase B was acetonitrile. The flow rate was set to 0.6 ml/min, and the column temperature was maintained at 40°C. The gradient program was as follows: 0–1.0 min, 0% B; 1.0–6.0 min, linear increase to 95% B; followed by a 1-min wash and a 1-min re-equilibration. Analytes were detected in the positive ion mode using multiple reaction monitoring (MRM) on a Waters Xevo TQ-XS mass spectrometer.
Laboratory analysis
Plasma glucose was measured every 5 minutes during the hyperglycemic clamp by glucose oxidase method (YSI Inc.). Plasma insulin levels were measured by double-antibody radioimmunoassay (Millipore). Plasma glucagon was measured by ELISA (Mercodia) by UNC’s Respiratory TRACTS Core.
Statistics
Analyses were performed using SPSS Version 29 or GraphPad Prism 10. Continuous dependent variables were compared by independent samples t-tests or, if Levene’s test for equality of variances was violated, by Welch’s t-test. Given the small sample size, we accepted non-normally distributed data by Shapiro-Wilk’s test. Plasma glucose, GIR, and plasma insulin time courses during the hyperglycemic clamp were compared by mixed effects analysis. Individual and summed amino acid concentrations were compared by one-way ANCOVA with baseline concentration as a covariate. Correlations were performed by Pearson’s product-moment correlations or Spearman’s rank correlations. A p value <0.05 was considered statistically significant.
Results
Peripheral metabolic characterization of study participants
Demographics, baseline clinical data obtained from participants’ screening visit and labs, and measures of insulin resistance and beta cell function derived from 75-gram oral glucose tolerance test data are displayed in Table 1. By study design, body mass index (BMI) was higher in the participants with obesity (p<0.001). As expected from our prior work in a larger cohort inclusive of these participants (4), the insulinogenic index, a measure of pancreatic beta cell function, and systolic blood pressure were higher in the obesity group (p=0.02 and p=0.04, respectively). Other standard clinical metabolic parameters including A1c, total cholesterol and cholesterol subspecies, and liver enzymes were not different between groups.
Table 1. Demographic, glycemic, and clinical metabolic characteristics of participants.
All laboratory values were obtained under fasting conditions. Glycemic variables, with the exception of A1c, were derived from 75-gram oral glucose tolerance test data. Scale variables are expressed as mean (standard deviation). Categorical variables are expressed as frequency. Comparisons between groups were made by independent samples t-tests or by Welch’s t-test if Levene’s test for equality of variances was violated. ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; BP, blood pressure; h, hour; HDL, high-density lipoprotein; HOMA-IR, homeostatic model of insulin resistance; LDL, low-density lipoprotein; OGTT, oral glucose tolerance test; TSH, thyroid stimulating hormone.
| BMI<25 (n=8) | BMI>30 (n=8) | BMI<25 vs. >30 | |
|---|---|---|---|
| Demographic Data | |||
| Age (years) | 27.3 (5.7) | 27.3 (4.7) | p=1.00 |
| Sex (n) | |||
| Female | 6 | 5 | |
| Male | 2 | 3 | |
| Race (n) | |||
| White | 6 | 5 | |
| Black | 0 | 3 | |
| Asian | 2 | 0 | |
| BMI (kg/m2) | 20.8 (1.8) | 32.3 (2.5) | **p<0.001 |
| Glycemic Data | |||
| A1c (%) | 5.4 (0.2) | 5.4 (0.2) | p=0.74 |
| 2h OGTT (mg/dL) | 103.8 (32.8) | 101.4 (35.0) | p=0.89 |
| HOMA-IR | 2.4 (1.7) | 3.4 (1.4) | p=0.25 |
| Matsuda Index | 5.6 (2.2) | 4.1 (2.6) | p=0.23 |
| Insulinogenic Index | 1.1 (0.8) | 3.5 (2.1) | *p=0.02 |
| Disposition Index | 5.3 (2.4) | 14.3 (15.5) | p=0.15 |
| Other Metabolic Data | |||
| Systolic BP (mmHg) | 107.6 (7.5) | 115.3 (5.9) | *p=0.04 |
| Diastolic BP (mmHg) | 68.0 (6.9) | 72.0 (7.5) | p=0.29 |
| Total Cholesterol (mg/dL) | 162.1 (22.2) | 161.4 (43.4) | p=0.97 |
| LDL Cholesterol (mg/dL) | 91.1 (20.9) | 91.5 (28.9) | p=0.98 |
| HDL Cholesterol (mg/dL) | 56.0 (11.5) | 50.8 (17.0) | p=0.48 |
| Triglycerides (mg/dL) | 74.9 (38.5) | 96.4 (38.7) | p=0.28 |
| AST (U/L) | 12.3 (6.0) | 11.5 (2.5) | p=0.75 |
| ALT (U/L) | 4.1 (8.3) | 4.8 (8.0) | p=0.88 |
| TSH (μIU/mL) | 2.3 (1.2) | 2.0 (0.7) | p=0.55 |
All participants were exposed to acute hyperglycemia via a hyperglycemic clamp of 120 minutes’ duration with target glucose 180 mg/dL, permitting clamp-derived approximations of glucose tolerance, beta cell function, and tissue sensitivity to insulin. Prior to initiation of IV 20% [1-13C] glucose infusion, fasting plasma glucose, insulin, and glucagon were not significantly different between participants with and without obesity (p=0.07, p=0.17, and p=0.16, respectively) (Figure 2A–C). Over the course of the hyperglycemic clamp, plasma glucose levels were similar between groups (Figure 2D). The glucose infusion rate (GIR), an approximation of glucose tolerance, and steady state GIR, or the GIR averaged between 90 and 120 minutes of the hyperglycemic clamp, were not different between groups (p=0.24) (Figure 2E–F). The interaction between time and obesity status on insulin during the hyperglycemic clamp was non-significant (p=0.19), and the area under the curve of the summary data was not different between groups (p=0.18) (Figure 2G–H), suggesting similar beta cell responses to IV glucose. The ratio of steady state GIR to insulin approximates peripheral tissue sensitivity to insulin (13, 17); this measure trended lower in obesity (p=0.06). In summary, most clamp-derived measures of whole-body glucose metabolism and insulin secretion and sensitivity were not significantly different between participants with and without obesity.
Figure 2. Metabolic characterization of lean participants and those with obesity at baseline (fasting) and during acute hyperglycemia.
(A-C) Fasting plasma glucose (p=0.07) (A), insulin (p=0.17) (B), and glucagon (p=0.16) (C) compared by independent samples t-tests. (D) Plasma glucose during the hyperglycemic clamp, target 180 mg/dL (p=0.80 for obesity status, p>0.83 for two-way interaction between time and obesity status, mixed effects analysis). (E) Glucose infusion rate (GIR) during the hyperglycemic clamp (p=0.10 for two-way interaction between time and obesity status, mixed effects analysis). (F) Steady state GIR expressed as the average GIR between 90 and 120 minutes of the hyperglycemic clamp (p=0.24, independent samples t-test). (G) Plasma insulin during the hyperglycemic clamp (p=0.19 for two-way interaction between time and obesity status, mixed effects analysis). (H) Area under the curve (AUC) of the insulin time course in (G) (p=0.18 by Welch’s t-test). (I) Ratio of steady state GIR to steady state insulin averaged between 90 and 120 minutes of the hyperglycemic clamp (p=0.06, independent samples t-test). All data are expressed as mean ± standard deviation.
Plasma amino acids and relationship with brain glucose during hyperglycemia
Alterations in amino acid signatures have been associated with obesity and type 2 diabetes. We utilized liquid chromatography-tandem mass spectrometry to measure plasma amino acids at baseline and at multiple time points during acute hyperglycemia (Figure 3). At baseline, there were few differences in amino acids. Glutamic acid was higher in obesity (154 ± 34 vs. 109 ± 37 μM, p=0.03), and glycine was lower in obesity (201 ± 37 vs. 261 ± 56 μM, p=0.02) (Table 2). There was a statistically significant positive correlation between glutamic acid and BMI (r=0.54, p=0.03) and statistically significant negative correlations between glycine and BMI (r=-0.55, p=0.03), and glycine and insulin (r=-0.55, p=0.03). There were no correlations between either glutamic acid or glycine and fasting glucose or glucagon.
Figure 3. Amino acids measured by liquid chromatography-tandem mass spectrometry at baseline and during acute hyperglycemia in participants with and without obesity.
Individual amino acid concentrations at 60 minutes, 120 minutes, and averaged between 60 and 120 minutes are displayed as a percentage of baseline concentration. Asterisks denote statistically significant comparisons of individual amino acids between the two BMI groups at the indicated time points by one-way ANCOVA with baseline concentration as a covariate.
Table 2. Statistically significant differences in individual plasma amino acids compared between lean participants and those with obesity exposed to acute hyperglycemia.
Variables are expressed as unadjusted mean (μM) ± standard deviation. Significant comparisons at baseline by independent samples t-tests and at all other time points by one-way ANCOVA with baseline concentration as a covariate are shown. The duration of hyperglycemia for one participant in each group was less than 120 minutes, therefore only 7 participants per group are included in calculations incorporating measurements at 120 minutes. AAA, aromatic amino acids (phenylalanine, tryptophan, and tyrosine).
| BMI<25 | BMI>30 | BMI<25 vs. >30 | |
|---|---|---|---|
| 0′ (n=8/group) | |||
| Glutamic Acid | 109.4 ± 37.1 | 154.2 ± 34.4 | p=0.03 |
| Glycine | 260.9 ± 55.6 | 201.2 ± 37.0 | p=0.02 |
| 60′ (n=8/group) | |||
| Tryptophan | 49.4 ± 8.7 | 56.3 ±7.6 | p=0.03 |
| AAAs (Summed) | 134.7 ± 26.5 | 166.0 ± 33.5 | p=0.02 |
| 120′ (n=7/group) | |||
| Cysteine | 15.8 ± 2.5 | 23.3 ± 5.5 | p=0.01 |
| Cystine | 0.05 ± 0.01 | 0.16 ± 0.09 | p<0.01 |
| Glutamic Acid | 78.0 ± 14.7 | 122.4 ± 23.1 | p<0.01 |
| Glutamine | 477.2 ± 97.1 | 364.9 ± 82.2 | p=0.02 |
| Average 60–120′ (n=7/group) | |||
| Cysteine | 16.7 ± 2.5 | 22.5 ± 4.8 | p=0.03 |
| Cystine | 0.06 ± 0.02 | 0.18 ± 0.12 | p=0.03 |
| Glutamic Acid | 83.6 ± 20.0 | 121.5 ± 20.4 | p=0.03 |
| Methionine | 16.2 ± 2.5 | 20.6 ± 6.2 | p=0.02 |
| Tryptophan | 47.3 ± 6.7 | 51.4 ± 6.5 | p=0.02 |
| Tyrosine | 38.0 ± 6.9 | 49.3 ± 17.0 | p=0.03 |
| AAAs (Summed) | 129.0 ± 19.0 | 149.0 ± 29.9 | p<0.01 |
We then examined plasma amino acids during acute hyperglycemia mid-way through the hyperglycemic clamp at 60 minutes and at the conclusion of the hyperglycemic clamp at 120 minutes. Statistically significant differences between groups at each of these individual time points and averaged between the two time points after adjustment for baseline concentrations are summarized in Table 2. The baseline elevation in glutamic acid in obesity persisted throughout hyperglycemia. Cysteine, cystine, glutamine, methionine, tryptophan, and tyrosine emerged as additional amino acids with differential responses to hyperglycemia in obesity. Of particular note, we observed changes in summed aromatic amino acids during acute hyperglycemia. In contrast, we did not observe changes between groups in BCAA either individually or summed.
We then examined relationships between individual and summed aromatic amino acids and the ratio of brain glucose transport to metabolism (Tmax/CMRgl) measured by 13C MRS during acute hyperglycemia. In our prior study, the brain glucose metabolic or utilization rate (CMRgl) was not different between obesity and lean controls, therefore differences in brain glucose were attributed to differences in the maximum rate of brain glucose transport (Tmax) (4). We observed significant negative correlations between concentrations of each individual aromatic amino acid as well as the summed concentration of aromatic amino acids with Tmax/CMRgl, such that higher aromatic amino acid concentrations were associated with lower Tmax/CMRgl (Figure 4). Similarly, higher cystine and methionine concentrations were associated with lower Tmax/CMRgl (cystine at t=120′: ρ=-0.64, p=0.02; methionine averaged between t=60′ and t=120′: ρ=-0.71, p<0.01). Other correlations between cysteine, cystine, glutamic acid, and glutamine with Tmax/CMRgl during hyperglycemia were not significant. In addition, for amino acids that were not different between groups, we also observed some significant correlations with Tmax/CMRgl at multiple timepoints: (1) t=60′: alanine (ρ=-0.68, p<0.01), arginine (ρ=-0.63, p=0.01), histidine (ρ=-0.68, p<0.01), isoleucine (ρ=-0.65, p<0.01), methionine (ρ=-0.69, p<0.01), proline (ρ=-0.76, p<0.001), and threonine (t=60′: ρ=-0.55, p=0.03); (2) t=120′: proline (ρ=-0.63, p=0.02); and (3) t=60–120′ averaged: arginine (ρ=-0.60, 0.03), histidine (ρ=-0.62, p=0.02); isoleucine (ρ=-0.61, p=0.03), leucine (ρ=-0.57, p=0.04), methionine (ρ=-0.71, p<0.01), and proline (ρ=-0.69, p<0.01).
Figure 4. Relationships between individual and summed aromatic amino acids with the ratio of brain glucose transport to metabolism (Tmax/CMRgl) during acute hyperglycemia.
Plasma concentrations of phenylalanine (A), tryptophan (B), tyrosine (C), and summed aromatic amino acids (D) at 60 minutes during the hyperglycemic clamp correlated with Tmax/CMRgl. Brain glucose data from one participant with obesity with BMI>35 did not meet quality control standards, therefore n=7 in the obesity group. All correlations were statistically significant by Spearman’s correlation. Closed dots represent lean controls, and open dots represent participants with obesity.
Plasma glucagon and relationship with brain glucose during acute hyperglycemia
Amino acid metabolism is regulated in part by glucagon, which promotes proteolysis and amino acid uptake in the liver in times of low energy availability. We therefore measured glucagon in lean participants and those with obesity and observed a similar pattern of response as described for aromatic amino acids: Plasma glucagon was higher in obesity compared to lean participants during acute hyperglycemia (4.43 ± 2.38 pg/mL in obesity vs. 2.15 ± 1.30 pg/mL in lean participants, p=0.047) (Figure 5A). Higher glucagon levels were associated with lower Tmax/CMRgl (Figure 5B). Of the aromatic amino acids, only tryptophan showed a significant correlation with glucagon (ρ=0.543, p=0.045).
Figure 5. Plasma glucagon and relationship between plasma glucagon and ratio of brain glucose transport to metabolism (Tmax/CMRgl) during acute hyperglycemia.
(A) Plasma glucagon averaged between 60 and 120 minutes of hyperglycemic clamp (p=0.046, independent samples t-test). Data are expressed as mean ± standard deviation. (B) Plasma glucagon averaged between 60 and 120 minutes of the hyperglycemic clamp correlated with Tmax/CMRgl by Spearman’s correlation. Closed dots represent lean controls, and open dots represent participants with obesity.
Discussion
Capturing cohorts of individuals with obesity who are early on the pathway to metabolic dysfunction can shed light on the stepwise deterioration of metabolic health. Here, we have performed peripheral metabolic phenotyping in a subset of participants from our prior study that used 13C MRS to measure brain glucose transport and metabolism during acute hyperglycemia in young, healthy adults with and without obesity (4). We have also related these peripheral changes to changes in brain glucose. Young adults with obesity had higher BMI, systolic blood pressure, and insulinogenic index calculated from OGTT data but still met existing criteria for metabolically healthy obesity (18). Fasting glucose, insulin, and glucagon were not significantly different between groups. We found minimal baseline amino acid differences in obesity, which is in contrast to studies describing increased aromatic amino acids and BCAA and decreased glycine in adults with obesity (6–8) (19), presumably reflective of greater age and/or more advanced obesity pathophysiology in these cohorts. We also investigated hyperglycemic clamp-derived assessments of metabolic health and did not find differences in steady state GIR, suggestive of similar glucose tolerance between groups – a conclusion further supported by equivalent two-hour OGTT plasma glucose concentrations. Together, data acquired at baseline and across hyperglycemic clamp-derived metabolic variables suggest that we have captured a cohort of individuals with obesity who may be at an early step in progression to metabolic dysfunction.
Our findings are consistent with other works suggesting that hyperaminoacidemia and hyperglucagonemia are early, subtle signs of insulin resistance (20, 21). Despite few significant baseline differences, the group with obesity showed higher plasma concentrations of cysteine, cystine, glutamic acid, glutamine, methionine, and aromatic amino acids during exposure to acute hyperglycemia, which may be early manifestations of insulin resistance. Of note, we did not observe any parallel differences in BCAA during hyperglycemia. While both sexes were represented in both groups, higher levels of BCAA and their derived metabolites have been found in men with obesity compared to women with obesity (22). Our cohorts were small and predominantly female, which could underestimate population-level differences in BCAA in obesity. Aromatic amino acids have received less attention in the literature than BCAA, but several potential mechanisms of insulin resistance have been identified, including insulin receptor modifications by phenylalanine (23) and tryptophan (24). While we cannot draw conclusions about whether diminished suppression of amino acids during acute hyperglycemia is a sign of developing insulin resistance or a contributor to worsening insulin resistance, this change is likely one occurring early in the development of metabolic dysfunction.
Fasting and post-prandial hyperglucagonemia are well-described features of obesity and type 2 diabetes and parallel insulin resistance due to the role of insulin in regulation of glucagon secretion (25). We observed a trending but nonsignificant increase in fasting plasma glucagon and significantly higher levels of plasma glucagon during acute hyperglycemia generated by hyperglycemic clamp, which is potentially explained by stronger suppression of glucagon by IV than oral glucose (26). Alternatively, CNS control of glucagon secretion by glucose-sensing neurons in the hypothalamus may be impaired in obesity due to obesity-associated hypothalamic inflammation.
In the periphery, glucagon secreted by pancreatic α-cells promotes amino acid transport into the liver, decreasing circulating amino acids and therefore decreasing glucagon – a feedback circuit termed the liver-α cell axis (27–29). The effect of visceral adiposity of metabolic dysfunction-associated steatotic liver disease (MASLD) on the liver-α cell axis and glucagon resistance remains uncertain. Several studies concluded that MASLD is associated with glucagon resistance and hyperaminoacidemia (30, 31). In another study, a short-term hypercaloric diet induced glucagon resistance and hyperaminoacidemia in young, lean men (32). In contrast, other studies have demonstrated equivalent amino acid responses to glucagon infusion irrespective of MASLD, obesity, and type 2 diabetes (33, 34). We did not quantify hepatic fat content in our study, but concurrent elevations in amino acids and glucagon during hyperglycemia could be consistent with glucagon resistance.
Available animal data involving genetic manipulation of the glucagon receptor suggest that glucagon-mediated effects may be specific to individual amino acids. In one study comparing changes in serum amino acids between wild type and glucagon receptor (Gcgr) knockout (Gcgr−/−) mice, phenylalanine and tryptophan showed the smallest, statistically significant difference between groups compared to many other amino acids (35). In another study, serum amino acids were compared between wild type and Gcgr−/− mice with the addition of two wild type groups treated with glucagon receptor antagonists; phenylalanine and tryptophan were unchanged in each of these cases (36). These animal data suggest that glucagon receptor signaling has less effect on aromatic amino acids compared to other amino acids. Therefore, the changes we see in aromatic amino acids during acute hyperglycemia in obesity may not be secondary to higher glucagon levels and glucagon resistance.
In the present study, we also observed negative correlations of cystine, methionine, aromatic amino acids, and glucagon with Tmax/CMRgl during acute hyperglycemia, such that higher circulating amino acid and glucagon concentrations were associated with lower brain glucose. This observation suggests that central and peripheral changes exist contemporaneously in early obesity pathophysiology. Cystine and glutamic acid share a transport system, system xc−, which is expressed in the brain and transports cystine intracellularly to support glutathione production and redox balance (37). The altered relationship between cystine and brain glucose during acute hyperglycemia could have implications for both neuroprotection and neurotoxicity. Aromatic amino acids are also highly important for brain function; tryptophan serves as a precursor for serotonin, kynurenine, and kynurenic acid, and tyrosine serves as a precursor for dopamine and catecholamines. Both aromatic amino acids and BCAA belong to a family of amino acids termed large neutral amino acids (LNAAs) that share a common transporter across the BBB. Therefore, alterations in circulating amino acid levels may result in altered availability of neurotransmitter precursors in the CNS (11).
Amino acid catabolism can supply tricarboxylic acid (TCA) cycle intermediates to re-supply those consumed via anabolic processes, a concept termed anaplerosis (38). Regarding aromatic amino acids in particular, phenylalanine and tyrosine are catabolized to intermediates that enter the TCA cycle directly via fumarate and indirectly via acetoacetate, which can be converted to acetyl-CoA. Tryptophan metabolism can generate acetyl-CoA as well as pyruvate via conversion to alanine. It is not known whether aromatic amino acid elevation in early obesity pathophysiology reflects ongoing proteolysis or impairment in amino acid catabolism. While speculative, ongoing proteolysis and flux through amino acid catabolic pathways could robustly replenish TCA cycle intermediates, lessening the demand for glucose entry into the brain. However, effects of hyperglycemia and developing insulin resistance on cellular metabolism in the brain are likely cell type-specific and represent an area ripe for investigation.
Our study is a small secondary analysis, and our observed relationships between amino acids, brain glucose, and obesity are correlative. However, these data underscore the need for ongoing inquiry into metabolic interplay between the brain and the periphery, especially in the context of epidemic metabolic diseases like obesity and type 2 diabetes; more work is needed to define these relationships and associated mechanisms.
In summary, we demonstrate that in a cohort of young, healthy individuals with obesity exposed to acute hyperglycemia, there are higher plasma levels of select amino acids, including aromatic amino acids, and glucagon as well as altered dynamics of metabolites such as glutamate and cystine associated with the system xc− carrier and reductive stress. These changes correlate with diminished uptake of brain glucose and suggest possible mechanisms for parallel development of central and peripheral changes in early obesity pathophysiology.
Acknowledgements
The authors would like to thank all participants of the study; the Yale Hospital Research Unit and Church Street Research Unit staff for assistance conducting the study; Ying Cao for coordination of study sample logistics; and the staff of the UNC Respiratory TRACTS Core Laboratory. FG’s current affiliation is with the Division of Endocrinology, Department of Internal Medicine, UT Southwestern, Dallas, Texas, USA.
Grants
National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases, Grant Numbers: R01DK123227 and R03DK121048 (to JJH), R01DK108283 (to GFM), P30DK124723 (North Carolina Diabetes Research Center, to JJH and CBN), and P30DK045735 (to Yale Diabetes Research Center); National Institute on Aging, Grant Numbers: R21AG073897 (to JJH) and R56AG079086 (to DLR); National Institute of Neurological Disorders and Stroke, Grant Numbers: R01NS087568 (to DLR); National Institute on Alcohol Abuse and Alcoholism, Grant Numbers: R01AA031401 (to GFM); Endocrine Fellows Foundation, Fellows’ Research Grant (to FG). Prior Presentation: These data were presented in part as abstracts at ENDO in June 2024 in Boston, Massachusetts, and the American Diabetes Association’s 85th Scientific Sessions in June 2025 in Chicago, Illinois (Ref).
Footnotes
Disclosures
The authors do not have perceived or potential conflicts of interest to disclose relevant to the present study.
Data Availability
Data presented in the current study are available from the corresponding author upon reasonable request.
REFERENCES
- 1.Ward ZJ, Bleich SN, Cradock AL, Barrett JL, Giles CM, Flax C, et al. Projected U.S. State-Level Prevalence of Adult Obesity and Severe Obesity. N Engl J Med. 2019;381(25):2440–50. [DOI] [PubMed] [Google Scholar]
- 2.Collaborators GBDD. Global, regional, and national burden of diabetes from 1990 to 2021, with projections of prevalence to 2050: a systematic analysis for the Global Burden of Disease Study 2021. Lancet. 2023;402(10397):203–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.O’Brien PD, Hinder LM, Callaghan BC, Feldman EL. Neurological consequences of obesity. Lancet Neurol. 2017;16(6):465–77. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Gunawan F, Matson BC, Coppoli A, Jiang L, Ding Y, Perry R, et al. Deficits in brain glucose transport among younger adults with obesity. Obesity (Silver Spring). 2024;32(7):1329–38. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Riserus U, Arnlov J, Berglund L. Long-term predictors of insulin resistance: role of lifestyle and metabolic factors in middle-aged men. Diabetes Care. 2007;30(11):2928–33. [DOI] [PubMed] [Google Scholar]
- 6.Adibi SA. Influence of dietary deprivations on plasma concentration of free amino acids of man. J Appl Physiol. 1968;25(1):52–7. [DOI] [PubMed] [Google Scholar]
- 7.Felig P, Marliss E, Cahill GF Jr. Plasma amino acid levels and insulin secretion in obesity. N Engl J Med. 1969;281(15):811–6. [DOI] [PubMed] [Google Scholar]
- 8.Swendseid ME, Umezawa CY, Drenick E. Plasma amino acid levels in obese subjects before, during, and after starvation. Am J Clin Nutr. 1969;22(6):740–3. [DOI] [PubMed] [Google Scholar]
- 9.White PJ, McGarrah RW, Herman MA, Bain JR, Shah SH, Newgard CB. Insulin action, type 2 diabetes, and branched-chain amino acids: A two-way street. Mol Metab. 2021;52:101261. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Amaral AU, Wajner M. Pathophysiology of maple syrup urine disease: Focus on the neurotoxic role of the accumulated branched-chain amino acids and branched-chain alpha-keto acids. Neurochem Int. 2022;157:105360. [DOI] [PubMed] [Google Scholar]
- 11.Coppola A, Wenner BR, Ilkayeva O, Stevens RD, Maggioni M, Slotkin TA, et al. Branched-chain amino acids alter neurobehavioral function in rats. Am J Physiol Endocrinol Metab. 2013;304(4):E405–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Matsuda M, DeFronzo RA. Insulin sensitivity indices obtained from oral glucose tolerance testing: comparison with the euglycemic insulin clamp. Diabetes Care. 1999;22(9):1462–70. [DOI] [PubMed] [Google Scholar]
- 13.DeFronzo RA, Tobin JD, Andres R. Glucose clamp technique: a method for quantifying insulin secretion and resistance. Am J Physiol. 1979;237(3):E214–23. [DOI] [PubMed] [Google Scholar]
- 14.Buschiazzo PM, Terrell EB, Regen DM. Sugar transport across the blood-brain barrier. Am J Physiol. 1970;219(5):1505–13. [DOI] [PubMed] [Google Scholar]
- 15.de Graaf RA, Pan JW, Telang F, Lee JH, Brown P, Novotny EJ, et al. Differentiation of glucose transport in human brain gray and white matter. J Cereb Blood Flow Metab. 2001;21(5):483–92. [DOI] [PubMed] [Google Scholar]
- 16.Gray N, Zia R, King A, Patel VC, Wendon J, McPhail MJ, et al. High-Speed Quantitative UPLC-MS Analysis of Multiple Amines in Human Plasma and Serum via Precolumn Derivatization with 6-Aminoquinolyl-N-hydroxysuccinimidyl Carbamate: Application to Acetaminophen-Induced Liver Failure. Anal Chem. 2017;89(4):2478–87. [DOI] [PubMed] [Google Scholar]
- 17.Sherwin RS, Kramer KJ, Tobin JD, Insel PA, Liljenquist JE, Berman M, et al. A model of the kinetics of insulin in man. J Clin Invest. 1974;53(5):1481–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Bluher M Metabolically Healthy Obesity. Endocr Rev. 2020;41(3). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Newgard CB, An J, Bain JR, Muehlbauer MJ, Stevens RD, Lien LF, et al. A branched-chain amino acid-related metabolic signature that differentiates obese and lean humans and contributes to insulin resistance. Cell Metab. 2009;9(4):311–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Wiklund P, Zhang X, Pekkala S, Autio R, Kong L, Yang Y, et al. Insulin resistance is associated with altered amino acid metabolism and adipose tissue dysfunction in normoglycemic women. Sci Rep. 2016;6:24540. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Ferrannini E, Muscelli E, Natali A, Gabriel R, Mitrakou A, Flyvbjerg A, et al. Association of fasting glucagon and proinsulin concentrations with insulin resistance. Diabetologia. 2007;50(11):2342–7. [DOI] [PubMed] [Google Scholar]
- 22.Patel MJ, Batch BC, Svetkey LP, Bain JR, Turer CB, Haynes C, et al. Race and sex differences in small-molecule metabolites and metabolic hormones in overweight and obese adults. OMICS. 2013;17(12):627–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Zhou Q, Sun WW, Chen JC, Zhang HL, Liu J, Lin Y, et al. Phenylalanine impairs insulin signaling and inhibits glucose uptake through modification of IRbeta. Nat Commun. 2022;13(1):4291. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Sun WX, Zhang KH, Zhou Q, Hu SH, Lin Y, Xu W, et al. Tryptophanylation of insulin receptor by WARS attenuates insulin signaling. Cell Mol Life Sci. 2024;81(1):25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Faerch K, Vistisen D, Pacini G, Torekov SS, Johansen NB, Witte DR, et al. Insulin Resistance Is Accompanied by Increased Fasting Glucagon and Delayed Glucagon Suppression in Individuals With Normal and Impaired Glucose Regulation. Diabetes. 2016;65(11):3473–81. [DOI] [PubMed] [Google Scholar]
- 26.Meier JJ, Deacon CF, Schmidt WE, Holst JJ, Nauck MA. Suppression of glucagon secretion is lower after oral glucose administration than during intravenous glucose administration in human subjects. Diabetologia. 2007;50(4):806–13. [DOI] [PubMed] [Google Scholar]
- 27.Haedersdal S, Andersen A, Knop FK, Vilsboll T. Revisiting the role of glucagon in health, diabetes mellitus and other metabolic diseases. Nat Rev Endocrinol. 2023;19(6):321–35. [DOI] [PubMed] [Google Scholar]
- 28.Wewer Albrechtsen NJ, Holst JJ, Cherrington AD, Finan B, Gluud LL, Dean ED, et al. 100 years of glucagon and 100 more. Diabetologia. 2023;66(8):1378–94. [DOI] [PubMed] [Google Scholar]
- 29.White PJ, Wewer Albrechtsen NJ, Campbell JE. Islet hormones at the intersection of glucose and amino acid metabolism. Nat Rev Endocrinol. 2025. [DOI] [PubMed] [Google Scholar]
- 30.Wewer Albrechtsen NJ, Junker AE, Christensen M, Haedersdal S, Wibrand F, Lund AM, et al. Hyperglucagonemia correlates with plasma levels of non-branched-chain amino acids in patients with liver disease independent of type 2 diabetes. Am J Physiol Gastrointest Liver Physiol. 2018;314(1):G91–G6. [DOI] [PubMed] [Google Scholar]
- 31.Suppli MP, Bagger JI, Lund A, Demant M, van Hall G, Strandberg C, et al. Glucagon Resistance at the Level of Amino Acid Turnover in Obese Subjects With Hepatic Steatosis. Diabetes. 2020;69(6):1090–9. [DOI] [PubMed] [Google Scholar]
- 32.Suppli MP, Hogedal A, Bagger JI, Chabanova E, van Hall G, Forman JL, et al. Signs of Glucagon Resistance After a 2-Week Hypercaloric Diet Intervention. J Clin Endocrinol Metab. 2024;109(4):955–67. [DOI] [PubMed] [Google Scholar]
- 33.Grondahl MFG, Bagger JI, Suppli MP, Van Hall G, Albrechtsen NJW, Holst JJ, et al. The effect of exogenous glucagon on circulating amino acids in individuals with and without type 2 diabetes and obesity. Endocr Connect. 2024;13(3). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Heeboll S, Wegener G, Gronbaek H, Nielsen S. Comparable glucagon-stimulated amino acid suppression in individuals with and without hepatic steatosis or steatohepatitis. Am J Physiol Endocrinol Metab. 2024;327(6):E679–E85. [DOI] [PubMed] [Google Scholar]
- 35.Solloway MJ, Madjidi A, Gu C, Eastham-Anderson J, Clarke HJ, Kljavin N, et al. Glucagon Couples Hepatic Amino Acid Catabolism to mTOR-Dependent Regulation of alpha-Cell Mass. Cell Rep. 2015;12(3):495–510. [DOI] [PubMed] [Google Scholar]
- 36.Dean ED, Li M, Prasad N, Wisniewski SN, Von Deylen A, Spaeth J, et al. Interrupted Glucagon Signaling Reveals Hepatic alpha Cell Axis and Role for L-Glutamine in alpha Cell Proliferation. Cell Metab. 2017;25(6):1362–73 e5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Bridges RJ, Natale NR, Patel SA. System xc(-) cystine/glutamate antiporter: an update on molecular pharmacology and roles within the CNS. Br J Pharmacol. 2012;165(1):20–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Inigo M, Deja S, Burgess SC. Ins and Outs of the TCA Cycle: The Central Role of Anaplerosis. Annu Rev Nutr. 2021;41:19–47. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Data presented in the current study are available from the corresponding author upon reasonable request.