Abstract
Aims
Hyperglucagonaemia contributes to the pathophysiology in type 2 diabetes (T2D), but the mechanisms behind the inappropriate glucagon secretion are not fully understood. Glucagon and amino acids are regulated in a feedback loop referred to as the liver–α cell axis. Individuals with non-alcoholic fatty liver disease (NAFLD) appear to be glucagon resistant, disrupting the liver–α cell axis resulting in hyperglucagonaemia and hyperaminoacidaemia. We investigated the associations between circulating glucagon, amino acids, and liver fat content in a cohort of individuals with T2D.
Methods
We included 110 individuals with T2D in this cross-sectional study. Liver fat content was quantified using 1H magnetic resonance spectroscopy (MRS). Associations between liver fat content and plasma glucagon and amino acids, respectively, were estimated in multivariate linear regression analyses.
Results
Individuals with NAFLD (n = 52) had higher plasma glucagon concentrations than individuals without NAFLD (n = 58). The positive association between plasma glucagon concentrations and liver fat content was confirmed in the multivariable regression analyses. Plasma concentrations of isoleucine and glutamate were increased, and glycine and serine concentrations were decreased in individuals with NAFLD. Concentrations of other amino acids were similar between individuals with and without NAFLD, and no clear association was seen between liver fat content and amino acids in the regression analyses.
Conclusion
MRS-diagnosed NAFLD in T2D is associated with hyperglucagonaemia and elevated plasma concentrations of isoleucine and glutamate and low plasma concentrations of glycine and serine. Whether NAFLD and glucagon resistance per se induce these changes remains to be elucidated.
Introduction
Hyperglucagonaemia contributes to the hyperglycaemia characterizing type 2 diabetes (T2D) (1, 2). Nevertheless, high glucagon concentrations are observed in some, but not all individuals with T2D, indicating that hyperglucagonaemia drives hyperglycaemia in only a subgroup of patients with T2D. The mechanisms behind T2D hyperglucagonaemia are not fully established (3, 4). Glucagon secretion is regulated by a complex interplay between endocrine and paracrine mechanisms, substrate availability, and the autonomic nervous system. Potent regulators of glucagon secretion include hypoglycaemia, insulin, and certain amino acids (4, 5).
A feedback loop between the liver and the pancreas, the liver–α-cell axis, controls amino acid concentrations by modulating hepatic uptake and catabolism of amino acids, while prevailing plasma amino acid concentrations, in turn, regulate glucagon secretion from the pancreatic α cells (6, 7, 8, 9). It has been proposed that hepatic fat deposition impairs glucagon-induced turnover of amino acids in the liver, resulting in increased plasma concentrations of amino acids (10, 11). This may increase glucagon secretion, compensating for the hepatic glucagon resistance and thus restoring amino acid turnover. The price for this is, however, that glucose production increases, as glucagon action on glucose production seems to be preserved in non-alcoholic fatty liver disease (NAFLD) (10, 12). The mechanisms underlying this hepatic glucagon resistance at the level of amino acid turnover may involve downregulation of hepatic amino acid transporters and key enzymes in the ureagenesis (10, 13, 14, 15).
The association between glucagon, amino acids, and liver fat has previously been investigated in healthy and individuals with prediabetes (16), in individuals with severe obesity (17), in a small group of individuals with T2D (18), and in diet and/or metformin-treated individuals with T2D (19). To our knowledge, associations between plasma glucagon and amino acid concentrations and liver fat content as assessed by reference standard methodology (1H magnetic resonance spectroscopy (MRS)) have not been studied in larger T2D cohorts, including individuals with and without NAFLD.
Here, we investigated the association between MRS-assessed liver fat content and plasma concentrations of glucagon and amino acids in a cohort of individuals with T2D. We hypothesized that increased liver fat content, i.e. NAFLD, attenuating the effect of glucagon on amino acid turnover in the liver, would associate with increased plasma concentrations of amino acids and hyperglucagonaemia.
Materials and methods
Cohort of participants with T2D
The present analyses were conducted on data and material obtained from the baseline visit in the randomized clinical trial ‘Effect of the mineralocorticoid receptor antagonist eplerenone on liver fat and metabolism in patients with type 2 diabetes (MIRAD trial)’ (20). In the primary study, 140 participants were recruited from the diabetes outpatient clinic at Herlev and Gentofte Hospital – University of Copenhagen, Herlev, Denmark, based on criteria that included diagnosis of T2D according to WHO criteria of >3 months duration, stable treatment with glucose-lowering and antihypertensive therapy according to guidelines, a history of cardiovascular disease (defined as previous myocardial infarction, significant coronary stenosis, peripheral artery disease, and/or previous stroke) or high risk of cardiovascular disease (identified by N-terminal pro-brain natriuretic peptide (Nt-proBNP) ≥70 ng/L or albuminuria) (20). The study was approved by the Ethical Committee of the Capital Region of Denmark and the Data Protection Agency and registered at ClinicalTrials.gov (NCT02809963) and conducted in accordance with the Declaration of Helsinki. For the present secondary explorative analyses, we excluded 20 MIRAD trial participants without MRS measurements of liver fat content and 10 MIRAD trial participants for whom frozen plasma samples for glucagon analyses were not available. In total, 110 MIRAD trial participants with complete data on liver fat content, glucagon, and amino acids were included in this study.
1H MRS
Liver fat content was measured in a region of interest by 1H MRS 3T Achieva MRI system (Philips Medical Systems) as described previously (20, 21). NAFLD was defined according to guidelines as liver fat content ≥5.6% (22).
Biochemical analyses
Blood samples were collected after an overnight fast of a minimum of 8 h and after 30 min of rest. Plasma was stored at −80°C until batch analyses. Plasma glucagon was measured from EDTA plasma using a sandwich ELISA (Mercodia, Uppsala, Sweden) (23) with an intra-assay and inter-assay coefficient of variation (CV) of <10% (24). Individual amino acids were measured from lithium-heparinized plasma using liquid chromatography–tandem mass spectrometry (LS-MS/MS) (TSQ Quantiva, Thermo Fisher Scientific) (25). The remaining biochemical analyses presented here were described previously (20).
Calculations
The homeostatic model assessment of insulin resistance (HOMA-IR) was used as an estimate of hepatic insulin resistance and calculated as fasting plasma glucose (mmol/L) × fasting serum insulin (µIU/mL)/22.5 (26). The Matsuda index was used to estimate whole-body insulin sensitivity and calculated from four-point measurements of plasma glucose and serum insulin from a 2-h 75 g oral glucose tolerance test (27). Estimated glomerular filtration rate (eGFR) was calculated from the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation (derived from serum creatinine, gender, and age) (28). To assess liver fibrosis, the Fibrosis-4 (FIB-4) score (derived from age, alanine transaminase, aspartate transaminase, and platelets) (29) and NAFLD fibrosis score (NFS) (derived from age, BMI, impaired fasting glucose, and presence of diabetes) (30) were calculated. The total amino acid concentration was calculated as the sum of individual amino acids.
Statistical analyses
All variables were inspected graphically to evaluate the pattern of distribution. Normally distributed variables are reported as mean ± s.d., non-normally distributed variables as median with interquartile range (IQR), and categorical variables as n and %. The cohort was grouped according to liver fat content < or ≥5.6%. Characteristics between groups were compared using unpaired t-test, Wilcoxon test, or chi–squared test, as appropriate. Multivariate linear regression analyses were used to evaluate the association between glucagon (dependent variable) and liver fat content (independent variable). In a stepwise approach, we adjusted for preselected confounders with the potential of affecting the dependent variable. First, data were adjusted for gender, age, BMI, and eGFR (model glucagon 1). Second, data were adjusted for fasting plasma glucose (model glucagon 2). Last, data were adjusted for diabetes medications (model glucagon 3). Similarly, the associations between amino acids (dependent variables) and liver fat content (independent variable) were evaluated by multivariate linear regression analyses adjusted for gender, age, BMI, and eGFR (model amino acids 1). Data were further adjusted for fasting plasma glucagon concentrations (model amino acids 2). Estimates from all regression analyses are displayed as a crude β coefficient with 95% CI. Liver fat content and glucagon were log-transformed to meet the assumption of normality of the residuals. P < 0.05 was considered statistically significant in all tests. Statistical software R version 4.1.0 was used for statistical evaluation and graphical presentations.
Results
Cohort characteristics
Baseline characteristics are displayed in Table 1. Overall, the cohort was overweight with a mean BMI of 30 ± 3.8 kg/m2 and a mean HbA1c of 60 ± 14 mmol/mol. The group was divided into T2D without NAFLD and T2D with NAFLD according to the NAFLD criteria (Table 1). The median liver fat content was 12% (IQR 9–17) in T2D with NAFLD and 1% (IQR 1–3) in T2D without NAFLD (P < 0.001). Individuals with T2D with NAFLD were younger compared to T2D − NAFLD (60 ± 8 years vs 66 ± 9 years, P < 0.001) and had a shorter duration of diabetes (10 ± 6 years vs 13 ± 8 years, respectively, P = 0.017). Markers of glucose homeostasis and insulin resistance were more impaired in the T2D with NAFLD group with higher C-peptide and insulin concentrations, higher HOMA-IR, and lower Matsuda index (all P ≤ 0.005). Liver enzymes (alanine transaminase and aspartate transaminase) were higher in the T2D with NAFLD group compared to T2D without NAFLD (all P ≤ 0.001). There were no statistically significant differences in fibrosis scores (NFS and FIB-4) or diabetes medications between groups. One participant in the T2D without NAFLD group qualified as an extreme outlier with glucagon concentration of 81 pmol/L, which was extraordinarily high compared with other participants who ranged from 3 to 44 pmol/L). All analyses were performed with and without this outlier, with no differences in significance values or overall conclusions. Therefore, for all data presented, the outlier is included.
Participant characteristics presented as the entire cohort and divided into groups based on liver fat content.
| All | T2D without NAFLD | T2D with NAFLD | P | |
|---|---|---|---|---|
| Number, n (%) | 110 | 58 (53%) | 52 (47%) | |
| Gender, n (%) | 0.61 | |||
| Female | 30 (27.3%) | 17 (29.3%) | 13 (25.0%) | |
| Male | 80 (72.7%) | 41 (70.7%) | 39 (75.0%) | |
| Age (years), mean ± s.d. | 63.5 ± 9.3 | 66.3 ± 9.2 | 60.3 ± 8.3 | <0.001 |
| Diabetes duration (years) , mean ± s.d. | 11.7 ± 7.4 | 13.2 ± 7.9 | 9.8 ± 6.4 | 0.017 |
| BMI (kg/m2), mean ± s.d. | 30.1 ± 3.8 | 29.6 ± 4.2 | 31.0 ± 3.2 | 0.060 |
| HbA1c (mmol/mol), mean ± s.d. | 59.5 ± 14.2 | 59.8 ± 15.6 | 59.2 ± 12.6 | 0.82 |
| HbA1c (%), mean ± s.d. | 7.6 ± 1.3 | 7.6 ± 1.4 | 7.6 ±1.2 | 0.82 |
| Fasting plasma glucose (mmol/L), median (IQR) | 7.8 (6.8–9.4) | 7.7 (6.6–9.2) | 8.0 (6.9–9.4) | 0.33 |
| Glucagon (pmol/L), median (IQR) | 12.0 (9.0–17.0) | 11.0 (7.0–15.0) | 14.5 (10.0–21.2) | 0.003 |
| C-peptide (pmol/L), mean ± s.d. | 867 ± 550 | 694 ± 449 | 1062 ± 599 | <0.001 |
| Insulin (pmol/L), median (IQR) | 14.5 (10.6–24.3) | 11.9 (8.8–20.0) | 17.0 (12.8–28.8) | 0.003 |
| Matsuda index, median (IQR) | 0.8 (0.6–1.2) | 0.9 (0.7–1.3) | 0.7 (0.4–0.9) | 0.002 |
| HOMA-IR index, median (IQR) | 5.2 (3.6–8.2) | 4.6 (3.2–6.8) | 7.0 (4.1–10.1) | 0.005 |
| eGFR (mL/min/1,73 m2), mean ± s.d. | 85.4 ± 18.8 | 82.7 ± 19.3 | 88.4 ± 17.9 | 0.12 |
| Liver fat content (%), median (IQR) | 5.2 (1.0–11.6) | 1.0 (1.0–3.2) | 11.8 (8.8–17.2) | < 0.001 |
| Alanine aminotransferase (U/L), median (IQR) | 31 (25–41) | 29.0 (24.3–35.0) | 36.5 (27.8–47.2) | 0.001 |
| Aspartate aminotransferase (U/L), median (IQR) | 28 (24–35) | 26.0 (22.0–30.0) | 31.5 (28.0–37.0) | <0.001 |
| NAFLD fibrosis score, mean ± s.d. | −0.08 ± 0.9 | −0.02 ± 1.00 | −0.15 ± 0.80 | 0.45 |
| FIB-4 score, median (IQR) | 1.4 (1.0–1.7) | 1.3 (1.0–1.8) | 1.4 (1.1–1.7) | 0.68 |
| Total amino acids (µmol/L), median (IQR) | 2796 (2604–3012) | 2775 (2608–2991) | 2803 (2597–3021) | 0.91 |
| Diabetes medication, n (%) | ||||
| Metformin | 91 (82.7%) | 46 (79.3%) | 45 (86.5%) | 0.32 |
| SGLT-2 inhibitor | 18 (16.4%) | 7 (12.1%) | 11 (21.2%) | 0.20 |
| Sulphonylurea | 11 (10.0%) | 4 (6.9%) | 7 (13.5%) | 0.25 |
| DDP-4 inhibitor | 24 (21.8%) | 9 (15.5%) | 15 (28.8%) | 0.09 |
| GLP-1RA | 36 (32.7%) | 22 (37.9%) | 14 (26.9%) | 0.22 |
| Insulin | 39 (35.5%) | 23 (39.7%) | 16 (30.8%) | 0.33 |
Normally distributed data are presented as mean ± s.d., non-normally distributed data as median with IQR in parentheses, and categorical data as n with percent in parentheses. Group comparison with t-test, Kruskal–Wallis test, or chi–squared test, respectively.
DDP-4, dipeptidyl peptidase 4; eGFR, estimated glomerular filtration rate; GLP-1RA, glucagon-like peptide 1 receptor agonist; HOMA-IR, homeostatic model assessment for insulin resistance; IQR, interquartile range; SGLT-2, sodium glucose transporter 2.
Association between glucagon and liver fat content
The T2D with NAFLD group had higher plasma glucagon concentrations compared to the T2D without NAFLD group (15 pmol/L (IQR 10–21) vs11 pmol/L (IQR 7–15), 95% CI −6 to −1, P = 0.003) (Table 1 and Fig. 1A). When adjusting for age, gender, BMI, and eGFR in the multivariate linear regression analysis (model glucagon 1), there was still a positive association between liver fat content and plasma glucagon concentration (Fig. 1B, C and Supplementary Table 1, see section on supplementary materials given at the end of this article). The model estimated that a doubling in liver fat content would associate with an increase in glucagon concentration of 8% (95% CI 1–16, P = 0.039). Further adjustment for the variance in fasting glucose concentrations (model glucagon 2) did not change the estimates. Finally, we adjusted for the variation in diabetes medication (model glucagon 3), resulting in an only minor change in the estimates (Fig. 1B and Supplementary Table 1). According to the models, none of the diabetes medications nor eGFR were significantly associated with plasma glucagon concentration (Fig. 1B and Supplementary Table 1).
Association between plasma glucagon concentrations and liver fat content. (A) Fasting plasma glucagon concentrations in T2D without NAFLD (n = 58) and T2D with NAFLD (n = 52). Boxes represent interquartile range with solid line representing the median. Whiskers depict 1.5× interquartile range. Group comparison by Kruskal–Wallis test. **P ≤ 0.001. (B) Forest plot of the estimated difference in glucagon (log) by a unit increase in the independent variables (multiple linear regression). Light blue colour is model glucagon 1 with adjustment for age, gender, BMI, and eGFR; middle blue colour is model glucagon 2 with further adjustment for fasting glucose; and dark blue colour is model glucagon 3 with further adjustment for diabetes medication. Symbols represent β and whiskers 95% CIs. Variables with a 95% CI not including zero are considered significantly associated with glucagon. (C) Association between liver fat content and fasting plasma glucagon concentration. β/regression line (solid line) with CI (grey area) from linear regression analysis with glucagon (log) as the dependent variable and liver fat content (log) as the independent variable adjusted for age, gender, BMI, and eGFR. n = 110. eGFR, estimated glomerular filtration rate; DDP-4, dipeptidyl peptidase 4; GLP-1RA, glucagon-like peptide 1 receptor agonist; NAFLD, non-alcoholic fatty liver disease; SGLT-2, sodium glucose transporter 2; SU, sulphonylurea; T2D, type 2 diabetes.
Citation: Endocrine Connections 13, 1; 10.1530/EC-23-0161
Association between amino acids and liver fat content
There was no difference in total amino acids between groups (Table 1 and Fig. 2C). T2D with NAFLD had a significantly higher average concentration of glutamate and isoleucine and a significantly lower average concentration of glycine and serine compared to T2D without NAFLD (Fig. 2A and Table 2). All other amino acids were not statistically different between groups. No amino acids associated with liver fat content in the linear regression model after adjusting for age, gender, BMI, and eGFR (model amino acids 1). Adjusting for variance in glucagon concentrations (model amino acids 2) did not change the associations between liver fat content and amino acid concentrations (Fig. 2B and Supplementary Table 2).
Association between plasma amino acid concentrations and liver fat content. (A) Fasting plasma amino acid concentrations in T2D without NAFLD (n = 58) and T2D with NAFLD (n = 52). Boxes represent interquartile range with solid lines representing the median. Whiskers depict 1.5× interquartile range. Group comparison by Kruskal–Wallis test; *P ≤ 0.05, **P ≤ 0.001. (B) Forest plot of the estimated difference in plasma amino acid concentration (μmol/L) by a doubling in liver fat content (%) (multiple linear regression). Light blue colour is model amino acids 1 with adjustment for age, gender, BMI, and eGFR and dark blue colour is model amino acids 2 with further adjustment for glucagon. Filled circles represent β/regression coefficient and whiskers 95% CIs. Variables with a 95% CI not including zero are considered significantly associated with the liver fat content. n = 110. (C) Association between liver fat content and fasting plasma total amino acid concentration. n = 110. eGFR, estimated glomerular filtration rate; NAFLD, non-alcoholic fatty liver disease; T2D, type 2 diabetes.
Citation: Endocrine Connections 13, 1; 10.1530/EC-23-0161
Concentrations of individual amino acids.
| Median (IQR) (µmol/L) | T2D without NAFLD | T2D with NAFLD | P | CI |
|---|---|---|---|---|
| Alanine | 399 (362–494) | 439 (360–517) | 0.53 | −50–30 |
| Arginine | 73 (64–81) | 74 (62–83) | 0.85 | −8.0–6.0 |
| Asparagine | 43 (37–49) | 40 (34–48) | 0.083 | −0.0–6.0 |
| Aspartate | 5.6 (4.0–7.5) | 6.3 (5.0–8.2) | 0.084 | −1.7–0.1 |
| Glutamate | 88 (64–116) | 112 (91–156) | 0.003 | −45 to −10 |
| Glutamine | 585 (496–633) | 556 (470–647) | 0.33 | −24–67 |
| Glycine | 207 (181–235) | 190 (169–217) | 0.042 | 1.0–32 |
| Histidine | 77 (74–83) | 76 (69–86) | 0.43 | −2.0–5.0 |
| Isoleucine | 73 (66–79) | 79 (67–90) | 0.047 | −13 to −0.0 |
| Leucine | 130 (120–145) | 134 (115–160) | 0.41 | −15–6.0 |
| Lysine | 187 (171–213) | 188 (173–209) | 0.82 | −10–13 |
| Methionine | 22 (19–25) | 22 (19–25) | 0.61 | −1.0–2.0 |
| Phenylalanine | 57 (53–64) | 57 (52–62) | 0.37 | −2.0–5.0 |
| Proline | 203 (168–241) | 207 (178–259) | 0.33 | −33–11 |
| Serine | 105 (90–117) | 95 (81–106) | 0.020 | 2.0–16 |
| Threonine | 110 (99–124) | 109 (89–126) | 0.52 | −6.0–13 |
| Tryptophan | 55 (47–58) | 58 (50–61) | 0.14 | −6.0–1.0 |
| Tyrosine | 57 (49–64) | 54 (48–62) | 0.25 | −2.0–6.0 |
| Valine | 246 (221–264) | 267 (230–293) | 0.13 | −33–4.0 |
Fasting plasma concentrations of individual amino acids. Data are presented as median with IQR in parentheses. Group comparison between T2D without NAFLD (n = 58) and T2D with NAFLD (n = 52) with Kruskal–Wallis test.
IQR, interquartile range.
Association between amino acids and glucagon
From the linear regression models with amino acids as the dependent variables and liver fat content as the independent variable, we could further estimate the associations between glucagon and amino acids (model amino acids 2), taking the variance of liver fat content, age, gender, BMI, and eGFR into account (Fig. 3A and Supplementary Table 2). Glucagon associated positively with isoleucine, leucine, valine, methionine, and phenylalanine and negatively with glycine and glutamine (Fig. 3B, C, D, E, F, and G and Supplementary Table 2).
Associations between plasma amino acids and plasma glucagon. (A) Forest plot of the estimated difference in plasma amino acid concentration (μmol/L) by a doubling in plasma glucagon concentration (multiple linear regression). Filled circles represent β/regression coefficient and whiskers 95% CIs. Variables with a 95% CI not including zero are considered associated with the liver fat content. n = 110. Association between fasting plasma glucagon concentration and fasting plasma concentration of valine (B), leucine (C), isoleucine (D), phenylalanine (E), glycine (F) and glutamine (G). Beta/regression line (solid line) with CI (grey area) from linear regression analysis with plasma amino acid concentration as the dependent variable and plasma glucagon concentration (log) as the independent variable adjusted for liver fat content (log), age, gender, BMI and eGFR. n = 110. eGFR, estimated glomerular filtration rate.
Citation: Endocrine Connections 13, 1; 10.1530/EC-23-0161
Discussion
We hypothesized that hepatic fat deposition attenuates the effect of glucagon on amino acid turnover in the liver, resulting in increased plasma concentrations of amino acids and hyperglucagonaemia. In the present cohort of 110 individuals with T2D with established or at high risk of cardiovascular disease and MRS-determined liver fat content of varying degree (n = 52 with NAFLD, n = 58 without NAFLD), we found that the T2D with NAFLD group was characterized by elevated fasting plasma glucagon concentration. Furthermore, the T2D with NAFLD group had significantly higher plasma concentrations of glutamate and isoleucine, whereas glycine and serine concentrations were lower compared to T2D without NAFLD. The concentrations of other amino acids were similar between individuals with and without NAFLD.
A positive association between glucagon and liver fat content has previously been reported in a small cohort of individuals with T2D and BMI-matched non-diabetic controls (18) and in individuals with normoglycaemia or prediabetes (16). In contrast, there was no association in a cohort of diet and/or metformin-treated individuals with T2D (19) or in individuals with severe obesity (17). The abovementioned studies used liver biopsy or MRS to assess liver fat content. In the present study, the higher average plasma glucagon concentration in the T2D with NAFLD group was supported by a positive association between glucagon and liver fat content in the regression analysis after adjustment for age, gender, BMI, and eGFR. However, there was a substantial variation in glucagon concentrations within both groups, and liver fat content only contributed little to the overall variation of glucagon, while none of the other included covariates associated significantly with glucagon (Fig. 1B and Supplementary Table 1).
In contrast to our hypothesis, we did not observe hyperaminoacidaemia in individuals with T2D with NAFLD vs individuals with T2D without NAFLD. As mentioned above, the group with T2D with NAFLD had higher plasma concentrations of glutamate and isoleucine and lower plasma concentrations of glycine and serine than T2D without NAFLD, whereas other amino acids and total amino acids were similar between the two groups (Fig. 2A). No associations were found between amino acids and liver fat in the regression analyses. Hypothetically, the lack of associations between liver fat content and other amino acids could be explained by compensatory increased glucagon levels in individuals with higher liver fat content, which would lower amino acids despite differences in glucagon sensitivity in the liver. However, in the model with adjustment for variation in glucagon levels, no association between liver fat content and amino acid concentrations was seen (Fig. 2B and Supplementary Table 2). Lack of associations between plasma amino acid concentrations and liver fat content was reported for individuals with T2D (19). Another study showed that individuals with severe obesity had higher concentrations of amino acids compared to lean controls, but the coexistence of NAFLD did not aggravate the hyperaminoacidaemia (17). In contrast, liver fat content and amino acids were positively associated in healthy individuals and individuals with prediabetes (16) and in a small cohort of obese individuals with and without T2D (18).
The lack of associations between liver fat content and most amino acids in our cohort of T2D patients point towards other factors than steatosis alone contributing to hyperaminoacidaemia. Insulin resistance results in increased muscle protein catabolism and release of amino acids from muscle to circulation (31) and may, accordingly, constitute a major determinant of amino acid levels. A large cross-sectional study recently found that individuals with NAFLD (diagnosed by transient elastography) had higher levels of most amino acids and lower levels of glycine compared to controls without NAFLD (32). However, in a sub-analysis of selected HOMA-IR-matched individuals from the two groups, the significant differences for all amino acids, except for alanine, disappeared (32). When comparing NAFLD with fibrosis to HOMA-IR-matched controls without NAFLD in the same sub-analysis, the only amino acids that remained altered in the group with NAFLD and fibrosis were glutamate (higher) and glycine (lower) (32). This suggests that NAFLD-associated hyperaminoacidaemia may – at least partly − be driven by reduced insulin sensitivity (32). In the present study, most amino acids were similar between groups despite the T2D with NAFLD group appeared more insulin resistant based on HOMA-IR and Matsuda index. This may be attributed to the fact that both groups were indeed insulin resistant despite differences in estimates.
In the present study, the T2D with NAFLD group had a higher plasma concentration of isoleucine and glutamate and a lower concentration of serine and glycine than T2D without NAFLD. Low plasma glycine levels have consistently been reported in obesity and T2D (33, 34) and shown to predict the development of diabetes in healthy individuals (33). The combination of high circulating glutamate and low serine and glycine levels has previously been reported in non-diabetic individuals with NAFLD (diagnosed by MRS or liver biopsy, respectively) (35, 36) and in NAFLD with fibrosis (diagnosed by transient elastography) (32). Interestingly, these three amino acids are all involved in the synthesis of the antioxidant glutathione. Glutathione is produced in response to oxidative stress and reactive oxygen species during liver damage and excessive fat oxidation (35). High plasma levels of glutamate may, in turn, be explained by the accelerated production of glutathione, which is cleaved into glutamate and a glutathione remnant by the enzyme gamma glutamyltransferase when it leaves the hepatocyte, resulting in a net increase of glutamate (36). In support of this notion, individuals with NAFLD had increased hepatic uptake and degradation of glycine, serine and alanine and a net release of glutamate compared to controls in a study measuring splanchnic fluxes in the fasted state (37). These findings led to the proposal of an NAFLD-specific metabolomic index consisting of the ratio between glutamate and the product of serine and glycine (glutamate/(serine + glycine)) (36).
We found a positive association between glucagon and the amino acids valine, leucine, isoleucine, methionine and phenylalanine (the three first being branched chain amino acids (BCAAs)). BCAAs are mainly metabolized in muscles where there are no glucagon receptors (38, 39). Moreover, BCAAs are not suspected to stimulate glucagon secretion (40) and we do not regard this as a causal association but believe that glucagon and BCAAs are increased independently of each other in parallel with general dysmetabolism in T2D. Glucagon upregulates glutaminase, the enzyme that deaminates glutamine into glutamate and ammonium (41), supporting our findings of a negative association between glucagon and glutamine. Further, glucagon increases glutamine uptake in the liver (42, 43). Finally, the observed negative association between glucagon and glycine has previously been explained by glucagon-induced upregulation of the glycine cleavage system (44, 45, 46).
Strengths of the study include using MRS for evaluation of liver fat content, mass spectrometry-validated sandwich ELISA for analysis of glucagon, and targeted mass spectrometry for analysis of individual amino acids. We also believe that the well-characterized cohort of individuals with T2D recruited according to prespecified inclusion criteria, yielding a relatively homogeneous group of individuals with T2D with established or at high risk of cardiovascular disease and different degrees of liver fat content, is a strength for association studies like the present. Further, we included multivariate linear regression analyses aiming to adjust for confounders. However, the disadvantage of linear regression, in this case, is that it simplifies results by assuming a completely linear relationship between the data, which turned out to be a simplification in our study. We omitted including plasma insulin and HOMA-IR in the models as the insulin assay used fails to precisely measure some of the insulin analogues used in the cohort (47), thus questioning the reliability of the insulin concentrations in those participants on insulin treatment. Limitations of the study include that these secondary exploratory analyses were not predefined and that information on dietary protein intake, which may have affected concentration of plasma amino acids, were not available. Finally, it should be emphasized that we did not adjust for multiple comparisons due to the exploratory nature of these secondary analyses. Thus, type I errors may have occurred, although our positive findings correspond well to previous reports.
In conclusion, in the present cohort of well-characterized individuals with T2D, MRS-assessed liver fat content was positively associated with plasma glucagon concentrations. Only a few amino acids differed in plasma concentration between T2D with NAFLD compared to the T2D without NAFLD (higher glutamate and isoleucine and lower glycine and serine). It may be speculated that the hyperaminoacidaemia previously observed in NAFLD is driven by peripheral proteolysis due to insulin resistance independently of coexisting NAFLD. Future mechanistic studies are needed to investigate this and to delineate the relationship between glucagon, individual amino acids, and NAFLD. In addition, larger epidemiologic studies are warranted to investigate if certain amino acids, perhaps in combination with glucagon, may serve as biomarkers for NAFLD risk or established NAFLD.
Supplementary materials
This is linked to the online version of the paper at https://doi.org/10.1530/EC-23-0161.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the study reported.
Funding
This study include the Foundation for the Promotion of Medicine AP Moeller, Herlev Hospital Research Foundation, the Danish Diabetes Academy, the Danish Heart Foundation (grant number: 15-R99-A5855), the Danish Medical Association Research Foundation and the Innovation Fund Denmark.
Data availability
The data supporting the findings of this study are not currently available in a public repository.
Author contribution statement
I.R. contributed to the conceptualization of the present secondary analysis, performed statistical analyses and drafted the manuscript. M.L.J contributed to study conduct and data collection. A.L. and M.P.S contributed to the conceptualization of the present secondary analysis. E.C., G.v.H, N.W.A., and J.J.H provided analyses. C.K. contributed to study conduct and data collection. F.K.K. contributed to the conceptualization of the present secondary analysis and edited the draft manuscript.
Acknowledgements
We thank all study participants, project staff, and the research laboratory at Herlev and Gentofte Hospital – University of Copenhagen, Herlev, Denmark.
References
- 1↑
Unger RH, & Cherrington AD. Glucagonocentric restructuring of diabetes: a pathophysiologic and therapeutic makeover. Journal of Clinical Investigation 2012 122 4–12. (https://doi.org/10.1172/JCI60016)
- 2↑
Pettus JH, D’Alessio D, Frias JP, Vajda EG, Pipkin JD, Rosenstock J, Williamson G, Zangmeister MA, Zhi L, & Marschke KB. Efficacy and safety of the glucagon receptor antagonist RVT-1502 in Type 2 diabetes uncontrolled on metformin monotherapy: a 12-week dose-ranging study. Diabetes Care 2020 43 161–168. (https://doi.org/10.2337/dc19-1328)
- 3↑
Wewer Albrechtsen NJ, Pedersen J, Galsgaard KD, Winther-Sørensen M, Suppli MP, Janah L, Gromada J, Vilstrup H, Knop FK, & Holst JJ. The liver-alpha cell axis and type 2 diabetes. Endocrine Reviews 2019 40 1 353–1 366. (https://doi.org/10.1210/er.2018-00251)
- 4↑
Gromada J, Franklin I, & Wollheim CB. Alpha-cells of the endocrine pancreas: 35 years of research but the enigma remains. Endocrine Reviews 2007 28 84–116. (https://doi.org/10.1210/er.2006-0007)
- 5↑
Rix I, Nexøe-Larsen C, Bergmann NC, Lund A, & Knop FK. Glucagon physiology. In Endotext. Feingold KR, Anawalt B, Boyce A, Chrousos G, de Herder WW, Dhatariya K, Dungan K, Hershman JM, Hofland J, Kalra S, et al. eds. South Dartmouth , MA, USA: MDText.com, Inc., 2019. (available at: https://www.ncbi.nlm.nih.gov/books/NBK279127/)
- 6↑
Solloway MJ, Madjidi A, Gu C, Eastham-Anderson J, Clarke HJ, Kljavin N, Zavala-Solorio J, Kates L, Friedman B, Brauer M, et al.Glucagon couples hepatic amino acid catabolism to mTOR-dependent regulation of α-cell Mass. Cell Reports 2015 12 495–510. (https://doi.org/10.1016/j.celrep.2015.06.034)
- 7↑
Dean ED, Li M, Prasad N, Wisniewski SN, Von Deylen A, Spaeth J, Maddison L, Botros A, Sedgeman LR, Bozadjieva N, et al.Interrupted glucagon signaling reveals hepatic α cell axis and role for L-glutamine in α cell proliferation. Cell Metabolism 2017 25 1362–1373.e5. (https://doi.org/10.1016/j.cmet.2017.05.011)
- 8↑
Kim J, Okamoto H, Huang Z, Anguiano G, Chen S, Liu Q, Cavino K, Xin Y, Na E, Hamid R, et al.Amino acid transporter Slc38a5 controls glucagon receptor inhibition-induced pancreatic α cell hyperplasia in mice. Cell Metabolism 2017 25 1348–1361.e8. (https://doi.org/10.1016/j.cmet.2017.05.006)
- 9↑
Holst JJ, Albrechtsen NJW, Pedersen J, & Knop FK. Glucagon and amino acids are linked in a mutual feedback cycle: the liver–α-cell axis. Diabetes 2017 66 235–240. (https://doi.org/10.2337/db16-0994)
- 10↑
Suppli MP, Bagger JI, Lund A, Demant M, van Hall G, Strandberg C, Kønig MJ, Rigbolt K, Langhoff JL, Wewer Albrechtsen NJ, et al.Glucagon resistance at the level of amino acid turnover in obese subjects with hepatic steatosis. Diabetes 2020 69 1090–1099. (https://doi.org/10.2337/db19-0715)
- 11↑
Winther-Sørensen M, Galsgaard KD, Santos A, Trammell SAJ, Sulek K, Kuhre RE, Pedersen J, Andersen DB, Hassing AS, Dall M, et al.Glucagon acutely regulates hepatic amino acid catabolism and the effect may be disturbed by steatosis. Molecular Metabolism 2020 42 101080. (https://doi.org/10.1016/j.molmet.2020.101080)
- 12↑
Janah L, Kjeldsen S, Galsgaard KD, Winther-Sørensen M, Stojanovska E, Pedersen J, Knop FK, Holst JJ, & Wewer Albrechtsen NJ. Glucagon receptor signaling and glucagon resistance. International Journal of Molecular Sciences 2019 20 E3314. (https://doi.org/10.3390/ijms20133314)
- 13↑
Lake AD, Novak P, Shipkova P, Aranibar N, Robertson DG, Reily MD, Lehman-McKeeman LD, Vaillancourt RR, & Cherrington NJ. Branched chain amino acid metabolism profiles in progressive human nonalcoholic fatty liver disease. Amino Acids 2015 47 603–615. (https://doi.org/10.1007/s00726-014-1894-9)
- 14↑
De Chiara F, Heebøll S, Marrone G, Montoliu C, Hamilton-Dutoit S, Ferrandez A, Andreola F, Rombouts K, Grønbæk H, Felipo V, et al.Urea cycle dysregulation in non-alcoholic fatty liver disease. Journal of Hepatology 2018 69 905–915. (https://doi.org/10.1016/j.jhep.2018.06.023)
- 15↑
Eriksen PL, Vilstrup H, Rigbolt K, Suppli MP, Sørensen M, Heebøll S, Veidal SS, Knop FK, & Thomsen KL. Non-alcoholic fatty liver disease alters expression of genes governing hepatic nitrogen conversion. Liver International 2019 39 2094–2101. (https://doi.org/10.1111/liv.14205)
- 16↑
Gar C, Haschka SJ, Kern-Matschilles S, Rauch B, Sacco V, Prehn C, Adamski J, Seissler J, Wewer Albrechtsen NJ, Holst JJ, et al.The liver-alpha cell axis associates with liver fat and insulin resistance: a validation study in women with non-steatotic liver fat levels. Diabetologia 2021 64 512–520. (https://doi.org/10.1007/s00125-020-05334-x)
- 17↑
Pedersen JS, Rygg MO, Kristiansen VB, Olsen BH, Serizawa RR, Holst JJ, Madsbad S, Gluud LL, Bendtsen F, & Wewer Albrechtsen NJ. Nonalcoholic fatty liver disease impairs the liver–alpha cell axis independent of hepatic inflammation and fibrosis. Hepatology Communications 2020 4 1610–1623. (https://doi.org/10.1002/hep4.1562)
- 18↑
Wewer Albrechtsen NJ, Junker AE, Christensen M, Hædersdal S, Wibrand F, Lund AM, Galsgaard KD, Holst JJ, Knop FK, & Vilsbøll T. Hyperglucagonemia correlates with plasma levels of non-branched-chain amino acids in patients with liver disease independent of type 2 diabetes. American Journal of Physiology 2018 314 G91–G96. (https://doi.org/10.1152/ajpgi.00216.2017)
- 19↑
Otten J, Stomby A, Waling M, Chorell E, Ryberg M, Svensson M, Holst JJ, & Olsson T. The liver-alpha-cell axis after a mixed meal and during weight loss in type 2 diabetes. Endocrine Connections 2021 10 1101–1110. (https://doi.org/10.1530/EC-21-0171)
- 20↑
Johansen ML, Schou M, Rossignol P, Holm MR, Rasmussen J, Brandt N, Frandsen M, Chabanova E, Dela F, Faber J, et al.Effect of the mineralocorticoid receptor antagonist eplerenone on liver fat and metabolism in patients with type 2 diabetes: a randomized, double-blind, placebo-controlled trial (MIRAD trial). Diabetes, Obesity and Metabolism 2019 21 2305–2314. (https://doi.org/10.1111/dom.13809)
- 21↑
Chabanova E, Fonvig CE, Bøjsøe C, Holm JC, & Thomsen HS. 1H MRS assessment of hepatic fat content: comparison between normal- and excess-weight children and adolescents. Academic Radiology 2017 24 982–987. (https://doi.org/10.1016/j.acra.2017.02.010)
- 22↑
European Association for the Study of the Liver (EASL), European Association for the Study of Diabetes (EASD) & European Association for the Study of Obesity (EASO). EASL-EASD-EASO Clinical Practice Guidelines for the management of non-alcoholic fatty liver disease. Journal of Hepatology 2016 64 1388–1402. (https://doi.org/10.1016/j.jhep.2015.11.004)
- 23↑
Wewer Albrechtsen NJ, Hartmann B, Veedfald S, Windeløv JA, Plamboeck A, Bojsen-Møller KN, Idorn T, Feldt-Rasmussen B, Knop FK, Vilsbøll T, et al.Hyperglucagonaemia analysed by glucagon sandwich ELISA: nonspecific interference or truly elevated levels? Diabetologia 2014 57 1919–1926. (https://doi.org/10.1007/s00125-014-3283-z)
- 24↑
Alexiadou K, Cuenco J, Howard J, Wewer Albrechtsen NJ, Ilesanmi I, Kamocka A, Tharakan G, Behary P, Bech PR, Ahmed AR, et al.Proglucagon peptide secretion profiles in type 2 diabetes before and after bariatric surgery: 1-year prospective study. BMJ Open Diabetes Research and Care 2020 8 e001076. (https://doi.org/10.1136/bmjdrc-2019-001076)
- 25↑
Bornø A, & Hall van G. Quantitative amino acid profiling and stable isotopically labeled amino acid tracer enrichment used for in vivo human systemic and tissue kinetics measurements. Journal of Chromatography 2014 951–952 69–77. (https://doi.org/10.1016/j.jchromb.2014.01.019)
- 26↑
Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, & Turner RC. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 1985 28 412–419. (https://doi.org/10.1007/BF00280883)
- 27↑
Matsuda M, & DeFronzo RA. Insulin sensitivity indices obtained from oral glucose tolerance testing: comparison with the euglycemic insulin clamp. Diabetes Care 1999 22 1462–1470. (https://doi.org/10.2337/diacare.22.9.1462)
- 28↑
Levey AS, Coresh J, Greene T, Stevens LA, Zhang YL, Hendriksen S, Kusek JW, Van Lente F & Chronic Kidney Disease Epidemiology Collaboration. Using standardized serum creatinine values in the modification of diet in renal disease study equation for estimating glomerular filtration rate. Annals of Internal Medicine 2006 145 247–254. (https://doi.org/10.7326/0003-4819-145-4-200608150-00004)
- 29↑
Shah AG, Lydecker A, Murray K, Tetri BN, Contos MJ, & Sanyal AJ. Use of the FIB4 index for non-invasive evaluation of fibrosis in nonalcoholic fatty liver disease. Clinical Gastroenterology and Hepatology 2009 1104–1112. (https://doi.org/10.1016/j.cgh.2009.05.033)
- 30↑
Angulo P, Hui JM, Marchesini G, Bugianesi E, George J, Farrell GC, Enders F, Saksena S, Burt AD, Bida JP, et al.The NAFLD fibrosis score: a noninvasive system that identifies liver fibrosis in patients with NAFLD. Hepatology 2007 45 846–854. (https://doi.org/10.1002/hep.21496)
- 31↑
Lynch CJ, & Adams SH. Branched-chain amino acids in metabolic signalling and insulin resistance. Nature Reviews. Endocrinology 2014 10 723–736. (https://doi.org/10.1038/nrendo.2014.171)
- 32↑
Hasegawa T, Iino C, Endo T, Mikami K, Kimura M, Sawada N, Nakaji S, & Fukuda S. Changed amino acids in NAFLD and liver fibrosis: a large cross-sectional study without influence of insulin resistance. Nutrients 2020 1450. (https://doi.org/10.3390/nu12051450)
- 33↑
Guasch-Ferré M, Hruby A, Toledo E, Clish CB, Martínez-González MA, Salas-Salvadó J, & Hu FB. Metabolomics in prediabetes and diabetes: a systematic review and meta-analysis. Diabetes Care 2016 39 833–846. (https://doi.org/10.2337/dc15-2251)
- 34↑
Newgard CB. Metabolomics and metabolic diseases: where do we stand? Cell Metabolism 2017 25 43–56. (https://doi.org/10.1016/j.cmet.2016.09.018)
- 35↑
Mardinoglu A, Bjornson E, Zhang C, Klevstig M, Söderlund S, Ståhlman M, Adiels M, Hakkarainen A, Lundbom N, Kilicarslan M, et al.Personal model-assisted identification of NAD+ and glutathione metabolism as intervention target in NAFLD. Molecular Systems Biology 2017 13 916. (https://doi.org/10.15252/msb.20167422)
- 36↑
Gaggini M, Carli F, Rosso C, Buzzigoli E, Marietti M, Della Latta V, Ciociaro D, Abate ML, Gambino R, Cassader M, et al.Altered amino acid concentrations in NAFLD: impact of obesity and insulin resistance. Hepatology 2018 67 145–158. (https://doi.org/10.1002/hep.29465)
- 37↑
Hyötyläinen T, Jerby L, Petäjä EM, Mattila I, Jäntti S, Auvinen P, Gastaldelli A, Yki-Järvinen H, Ruppin E, & Orešič M. Genome-scale study reveals reduced metabolic adaptability in patients with non-alcoholic fatty liver disease. Nature Communications 2016 7 8994. (https://doi.org/10.1038/ncomms9994)
- 38↑
Boden G, Tappy L, Jadali F, Hoeldtke RD, Rezvani I, & Owen OE. Role of glucagon in disposal of an amino acid load. American Journal of Physiology 1990 259 E225–E232. (https://doi.org/10.1152/ajpendo.1990.259.2.E225)
- 39↑
Felig P. Amino acid metabolism in man. Annual Review of Biochemistry 1975 44 933–955. (https://doi.org/10.1146/annurev.bi.44.070175.004441)
- 40↑
Galsgaard KD, Jepsen SL, Kjeldsen SAS, Pedersen J, Wewer Albrechtsen NJ, & Holst JJ. Alanine, arginine, cysteine, and proline, but not glutamine, are substrates for, and acute mediators of, the liver-α-cell axis in female mice. American Journal of Physiology-Endocrinology and Metabolism 2020 318 E920–E929. (https://doi.org/10.1152/ajpendo.00459.2019)
- 41↑
Miller RA, Shi Y, Lu W, Pirman DA, Jatkar A, Blatnik M, Wu H, Cárdenas C, Wan M, Foskett JK, et al.Targeting hepatic glutaminase activity to ameliorate hyperglycemia. Nature Medicine 2018 24 518–524. (https://doi.org/10.1038/nm.4514)
- 42↑
Kraft G, Coate KC, Winnick JJ, Dardevet D, Donahue EP, Cherrington AD, Williams PE, & Moore MC. Glucagon’s effect on liver protein metabolism in vivo. American Journal of Physiology 2017 313 E263–E272. (https://doi.org/10.1152/ajpendo.00045.2017)
- 43↑
Pedersen C, Kraft G, Edgerton DS, Scott M, Farmer B, Smith M, Laneve DC, Williams PE, Moore LM, & Cherrington AD. The kinetics of glucagon action on the liver during insulin-induced hypoglycemia. American Journal of Physiology 2020 318 E779–E790. (https://doi.org/10.1152/ajpendo.00466.2019)
- 44↑
Alves A, Bassot A, Bulteau AL, Pirola L, & Morio B. Glycine metabolism and its alterations in obesity and metabolic diseases. Nutrients 1356 2019. (https://doi.org/10.3390/nu11061356)
- 45↑
Jois M, Hall B, Fewer K, & Brosnan JT. Regulation of hepatic glycine catabolism by glucagon. Journal of Biological Chemistry 1989 264 3347–3351. (https://doi.org/10.1016/S0021-9258(1894072-X)
- 46↑
Jog R, Chen G, Wang J, & Leff T. Hormonal regulation of glycine decarboxylase and its relationship to oxidative stress. Physiological Reports 2021 9 e14991. (https://doi.org/10.14814/phy2.14991)
- 47↑
Parfitt C, Church D, Armston A, Couchman L, Evans C, Wark G, & McDonald TJ. Commercial insulin immunoassays fail to detect commonly prescribed insulin analogues. Clinical Biochemistry 2015 48 1354–1357. (https://doi.org/10.1016/j.clinbiochem.2015.07.017)
