Altered amino acid levels in young hypopituitarism: impact of NAFLD and insulin resistance Authors: Yuwen Zhang, Jiting Qiu, Shouyue Sun, Xuqian Fang Published: 2024 ABSTRACT Elevated concentrations of amino acids (AAs) are commonly observed in patients with nonalcoholic fatty liver disease (NAFLD). Individuals with hypopituitarism (HP) are at a heightened risk of developing NAFLD due to factors such as visceral obesity, increased insulin resistance (IR), and disturbances in lipid metabolism. However, the changes in AAs concentrations associated with HP remain poorly understood. Therefore, our study aimed to investigate whether individuals with HP, who were not receiving growth hormone replacement therapy (GHRT), exhibited altered AAs compared to controls (CTs), and whether these AAs were associated with IR, the presence of NAFLD, and the Metabolic Syndrome (MetS) score. The AAs profiles of 133 young males with HP (age: 24.5 ± 5.9; 57 with NAFLD and 76 without NAFLD) and 90 age and BMI-matched CTs were analyzed using untargeted metabolomics. The results revealed that most AAs were found to be elevated in subjects with HPs compared to CTs. Glutamate, glutamine, norleucine, and branched-chain amino acids (BCAAs) (leucine and valine) were correlated with the homeostasis model assessment of insulin resistance (HOMA-IR), with glutamate and norleucine showing independent linkage. Glutamate and proline levels were specifically associated with MetS score, while alanine and proline linked to NAFLD. Given that elevated glutamate and BCAAs levels have higher prevalence of NAFLD, we hypothesized that the changes in AAs observed in HPs may be attributed to the impact of NAFLD and IR.Supplementary InformationThe online version contains supplementary material available at 10.1007/s00726-024-03429-y. Introduction Hypopituitarism (HP), a rare endocrine disorder characterized by pituitary hormone deficiencies, has garnered attention in recent years due to its potential link with nonalcoholic fatty liver disease (NAFLD), a prevalent liver condition associated with metabolic syndrome (Prodam et al. 2021). The interplay between HP and NAFLD remains a topic of interest, as emerging evidence suggests a possible association between these two conditions (Huang et al. 2022; Kang et al. 2021). Understanding the relationship between HP and NAFLD could offer insights into shared pathophysiological mechanisms and inform clinical management strategies for affected individuals. Recent research has shed light on the role of amino acid (AA) metabolism in the pathogenesis of NAFLD. The alternation of AA profiles in patients with NAFLD is a critical aspect of the metabolic dysregulation observed in these individuals. Specifically, alterations in the levels of branched-chain amino acids (BCAAs), aromatic amino acids (AAAs), and other essential amino acids have been implicated in NAFLD (Gaggini et al. 2018).The BCAAs (i.e., leucine, isoleucine, and valine) are often found increased and associated with peripheral IR in patients with diabetes or obesity (McCormack et al. 2013; Kolanu et al. 2017). The circulating BCAAs were also found increased with increased severity of liver diseases (Lake et al. 2015) and contribute to inflammation (Zhenyukh et al. 2017). In addition to BCAAs and AAAs, other AAs, such as glutamate, glutamine, and alanine also have been found to be positively associated with IR and increased cardio-metabolic risk in NAFLD (Bianchi et al. 2003). Prevalence of NAFLD and MetS are increased in patients with HP (Miljic and Popovic 2018; Kang et al. 2021), potentially leading to an imbalance in AA profiles in these patients. Features of obesity, visceral adiposity, dyslipidemia, and IR are common in patients with HP, highlighting the complex interplay of AA metabolism in this condition. Therefore, understanding this interplay is crucial for the development of targeted therapies and lifestyle interventions aimed at halting or reversing the progression of this disease. It has been suggested that growth hormone (GH) deficiency may be responsible for the adverse body composition and metabolic profile associated with HP. Consequently, the first aim of our study was to evaluate if AAs were increased/decreased compared to BMI-matched controls (CTs) in patients with HP without growth hormone replacement therapy (GHRT). Our second aim was to evaluate if alterations in circulating AAs are linked to insulin resistance (IR), the presence of NAFLD, and the Metabolic Syndrome (MetS) score. Patients and methods To standardize the influence of age and gender on metabolic parameters, this study exclusively enrolled young adult male’s patients and gender-matched healthy controls within the same age bracket. The term “young adult” is defined as individuals between the ages of 18 to 30 years old. A total of 133 patients with HP and 90 healthy controls were included in the study, with their characteristics summarized in Table 1. The HPs group comprised individuals with both congenital (n = 83) and acquired HPs (n = 50). Congenital HP is associated with pituitary stalk interruption syndrome, while acquired HP is a result of surgical or radiation injury during childhood. ALL CTs exhibited normal liver enzymes and normal liver findings on ultrasound examination. Patients with HP were divided into two groups: those with NAFLD (NAFLD-YES, n = 57) and those without NAFLD (NAFLD-NO, n = 76) based on ultrasound scans. All participants provided informed consent before participating in the study, which was approved by the ethics committee of Ruijin Hospital and conducted in accordance with the Helsinki Declaration. Table 1Clinical characteristics of study subjectsCTsn = 90HPsn = 133NAFLD-YESn = 57NAFLD-NOn = 76P ValueHPs Vs. CTsANOVA of NAFLD-YES, NO and CTsP ValueP valueNAFLD-YES Vs. NOP valueNAFLD-YES Vs. CTsP valueNAFLD-NO Vs. CTsSex (M/F)90/0133/057/076/0—————Age (years)24.1 ± 3.324.5 ± 5.923.3 ± 5.925.4 ± 5.80.9000.0370.0340.5750.195BMI (kg/m2)23.8 ± 1.924.0 ± 4.125.7 ± 4.222.4 ± 3.40.970< 0.001< 0.0010.0010.015TG (mmol/L)1.3 ± 0.82.1 ± 1.62.6 ± 1.81.7 ± 1.2< 0.001< 0.001< 0.001< 0.0010.102CHOL (mmol/L)4.6 ± 0.85.1 ± 1.35.4 ± 1.44.8 ± 1.10.007< 0.0010.003< 0.0010.352HDL-C (mmol/L)1.3 ± 0.31.2 ± 1.01.1 ± 0.31.3 ± 1.20.9100.4070.4560.4550.999LDL-C (mmol/L)4.0 ± 0.53.4 ± 1.03.7 ± 1.23.2 ± 0.8< 0.001< 0.0010.0040.053< 0.001FBG (mmol/L)4.5 ± 0.45.0 ± 1.34.8 (4.6, 5.5)4.6 (4.4, 6.0)0.0020.0030.002< 0.0010.233Insulin (µU/mL)10.7 ± 4.013.6 ± 9.516.8 (9.5, 26.5)8.2 (4.9, 11.0)0.020< 0.001< 0.001< 0.0010.889HOMA-IR2.1 ± 0.83.2 ± 2.83.5 (2, 6)1.7 (0.9, 2.5)0.004< 0.001< 0.001< 0.0010.968ALT (U/L)21 (14, 32)25 (15, 46)31 (17, 59)23 (15, 34.5)0.1100.0200.0930.0170.795AST (U/L)21 (17, 26)28 (22, 40)31 (22, 47)26 (22, 32)< 0.001< 0.0010.099< 0.0010.006GGT (U/L)25 (17,42)23 (15, 43)27 (18, 62)20 (14, 28)0.2900.0650.0920.0920.996ALP (U/L)74 (62, 88)116 (92, 182)125 (92, 196)114 (77, 163)< 0.001< 0.0010.468< 0.001< 0.001Variables were presented as mean ± SE for normally distributed data, or as median (IQR) for non-normally distributed data. Abbreviations: CT, control; HP, hypopituitarism; NAFLD, nonalcoholic fatty liver disease; BMI, body mass index; TG, triglycerides; CHOL, total cholesterol; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; FBG, fasting blood glucose; HOMA-IR, homeostasis model assessment of insulin resistance; ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, gamma-glutamyltransferase; ALP, alkaline phosphatase Table 1Clinical characteristics of study subjectsCTsn = 90HPsn = 133NAFLD-YESn = 57NAFLD-NOn = 76P ValueHPs Vs. CTsANOVA of NAFLD-YES, NO and CTsP ValueP valueNAFLD-YES Vs. NOP valueNAFLD-YES Vs. CTsP valueNAFLD-NO Vs. CTsSex (M/F)90/0133/057/076/0—————Age (years)24.1 ± 3.324.5 ± 5.923.3 ± 5.925.4 ± 5.80.9000.0370.0340.5750.195BMI (kg/m2)23.8 ± 1.924.0 ± 4.125.7 ± 4.222.4 ± 3.40.970< 0.001< 0.0010.0010.015TG (mmol/L)1.3 ± 0.82.1 ± 1.62.6 ± 1.81.7 ± 1.2< 0.001< 0.001< 0.001< 0.0010.102CHOL (mmol/L)4.6 ± 0.85.1 ± 1.35.4 ± 1.44.8 ± 1.10.007< 0.0010.003< 0.0010.352HDL-C (mmol/L)1.3 ± 0.31.2 ± 1.01.1 ± 0.31.3 ± 1.20.9100.4070.4560.4550.999LDL-C (mmol/L)4.0 ± 0.53.4 ± 1.03.7 ± 1.23.2 ± 0.8< 0.001< 0.0010.0040.053< 0.001FBG (mmol/L)4.5 ± 0.45.0 ± 1.34.8 (4.6, 5.5)4.6 (4.4, 6.0)0.0020.0030.002< 0.0010.233Insulin (µU/mL)10.7 ± 4.013.6 ± 9.516.8 (9.5, 26.5)8.2 (4.9, 11.0)0.020< 0.001< 0.001< 0.0010.889HOMA-IR2.1 ± 0.83.2 ± 2.83.5 (2, 6)1.7 (0.9, 2.5)0.004< 0.001< 0.001< 0.0010.968ALT (U/L)21 (14, 32)25 (15, 46)31 (17, 59)23 (15, 34.5)0.1100.0200.0930.0170.795AST (U/L)21 (17, 26)28 (22, 40)31 (22, 47)26 (22, 32)< 0.001< 0.0010.099< 0.0010.006GGT (U/L)25 (17,42)23 (15, 43)27 (18, 62)20 (14, 28)0.2900.0650.0920.0920.996ALP (U/L)74 (62, 88)116 (92, 182)125 (92, 196)114 (77, 163)< 0.001< 0.0010.468< 0.001< 0.001Variables were presented as mean ± SE for normally distributed data, or as median (IQR) for non-normally distributed data. Abbreviations: CT, control; HP, hypopituitarism; NAFLD, nonalcoholic fatty liver disease; BMI, body mass index; TG, triglycerides; CHOL, total cholesterol; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; FBG, fasting blood glucose; HOMA-IR, homeostasis model assessment of insulin resistance; ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, gamma-glutamyltransferase; ALP, alkaline phosphatase Clinical characteristics of study subjects Variables were presented as mean ± SE for normally distributed data, or as median (IQR) for non-normally distributed data. Abbreviations: CT, control; HP, hypopituitarism; NAFLD, nonalcoholic fatty liver disease; BMI, body mass index; TG, triglycerides; CHOL, total cholesterol; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; FBG, fasting blood glucose; HOMA-IR, homeostasis model assessment of insulin resistance; ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, gamma-glutamyltransferase; ALP, alkaline phosphatase Hormone replacement plan of patients with hypopituitarism Following diagnosis of HP, patients are initiated on replacement therapy with physiological doses of glucocorticoids and thyroid hormones. A subset of patients had undergone a phase of growth hormone replacement during childhood to facilitate height increase. However, at the time of blood collection for our assays, GHRT had been discontinued for at least 24 months. Measurements and calculations Serum total cholesterol (CHOL), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), triglycerides (TG), and liver enzymes were quantified using an automated enzymatic method on an autoanalyzer (Beckman Coulter, California, USA). Plasma glucose levels were determined through the glucose oxidase method on the same autoanalyzer. Serum insulin levels were measured using a commercially available RIA kit (Diagnostic Systems Laboratories, Minnesota, USA). Insulin resistance (IR) was assessed by calculating the homeostasis model assessment of insulin resistance (HOMA-IR) index [HOMA-IR = insulin (µU/mL) × glucose (mmol/L) / 22.5]. The AA plasma profile was evaluated in fasting samples using the Agilent 1290 Infinity LC system (Agilent Technologies, California, USA) coupled with an AB SCIEX Triple TOF6600 system (AB SCIEX, California, USA). In brief, 80 µL of an ethylene diamine tetra acetic acid plasma sample was deproteinized with 1 mL of methanol and subsequently purified through ion exchange columns. Variable importance in projection (VIP) values for each variable in the orthogonal partial least squares-discriminant analysis (OPLS-DA) model were calculated to identify metabolites contributing to classification. Variables with P values < 0.05 and VIP values > 1 were considered statistically significant. Our method allows for a relative quantitative analysis of the plasma AA profile through untargeted metabolomics based on peak areas. The identity of each amino acid was confirmed by matching the retention time and the MRM transition with those of authenticated standards. With this approach, levels of glutamate, glutamine, leucine, norleucine, valine, alanine, lysine, arginine, phenylalanine, pyroglutamic, proline, tryptophan, and histidine were relatively quantified (Table 2). The levels of AAs were normalized by dividing the peak area of the metabolite by 10,000. Table 2Circulated amino acid levelsCTsn = 90HPsn = 133NAFLD-YESn = 57NAFLD-NOn = 76q valueHPs Vs. CTsANOVA of NAFLD-YES, NO and CTsP ValueP valueNAFLD-YES Vs. NOP valueNAFLD-YES Vs. CTsP valueNAFLD-NO Vs. CTsGlutamate*4.1 ± 1.9&6.6 ± 2.48.7 ± 4.17.5 ± 2.6< 0.0001< 0.00010.2720< 0.0001< 0.0001Glutamine*2.4 ± 1.20.9 ± 0.40.8 ± 0.40.9 ± 0.4< 0.0001< 0.00010.8400< 0.0001< 0.0001Leucine#4.6 ± 1.15.3 ± 0.95.6 ± 1.25.2 ± 0.8< 0.0001< 0.00010.0445< 0.00010.0054Norleucine*#2.3 ± 0.80.7 ± 0.30.8 ± 0.30.6 ± 0.2< 0.0001< 0.00010.0386< 0.0001< 0.0001Valine*17.9 ± 5.328.4 ± 6.729.2 ± 7.127.8 ± 6.3< 0.0001< 0.00010.4072< 0.0001< 0.0001Alanine*0.3 ± 0.22.7 ± 1.42.4 ± 1.22.9 ± 1.4< 0.0001< 0.00010.0053< 0.0001< 0.0001Lysine*#23.1 ± 6.97.7 ± 4.48.1 ± 4.37.5 ± 4.5< 0.0001< 0.00010.6786< 0.0001< 0.0001Arginine*107.2 ± 25.8138.3 ± 22.0135.9 ± 23.9140.2 ± 20.3< 0.0001< 0.00010.5674< 0.0001< 0.0001Phenylalanine#21.7 ± 7.831.0 ± 9.631.6 ± 9.830.6 ± 9.4< 0.0001< 0.00010.8225< 0.0001< 0.0001Pyroglutamic *38.1 ± 10.348.9 ± 11.348.8 ± 12.549.1 ± 10.5< 0.0001< 0.00010.9959< 0.0001< 0.0001Proline*9.5 ± 4.513.9 ± 5.115.0 ± 5.813.0 ± 4.4< 0.0001< 0.00010.0407< 0.0001< 0.0001Tryptophan#7.5 ± 3.911.4 ± 4.611.3 ± 4.911.5 ± 4.3< 0.0001< 0.00010.9452< 0.0001< 0.0001Histidine6.0 ± 2.98.4 ± 3.78.1 ± 3.48.5 ± 3.8< 0.0001< 0.00010.65430.0011< 0.0001&The value is the relative peak area; *glucogenic amino acids; #ketogenic amino acids. Variables were presented as mean ± SE for normally distributed data. Abbreviations: CT, control; HP, hypopituitarism; NAFLD, nonalcoholic fatty liver disease Table 2Circulated amino acid levelsCTsn = 90HPsn = 133NAFLD-YESn = 57NAFLD-NOn = 76q valueHPs Vs. CTsANOVA of NAFLD-YES, NO and CTsP ValueP valueNAFLD-YES Vs. NOP valueNAFLD-YES Vs. CTsP valueNAFLD-NO Vs. CTsGlutamate*4.1 ± 1.9&6.6 ± 2.48.7 ± 4.17.5 ± 2.6< 0.0001< 0.00010.2720< 0.0001< 0.0001Glutamine*2.4 ± 1.20.9 ± 0.40.8 ± 0.40.9 ± 0.4< 0.0001< 0.00010.8400< 0.0001< 0.0001Leucine#4.6 ± 1.15.3 ± 0.95.6 ± 1.25.2 ± 0.8< 0.0001< 0.00010.0445< 0.00010.0054Norleucine*#2.3 ± 0.80.7 ± 0.30.8 ± 0.30.6 ± 0.2< 0.0001< 0.00010.0386< 0.0001< 0.0001Valine*17.9 ± 5.328.4 ± 6.729.2 ± 7.127.8 ± 6.3< 0.0001< 0.00010.4072< 0.0001< 0.0001Alanine*0.3 ± 0.22.7 ± 1.42.4 ± 1.22.9 ± 1.4< 0.0001< 0.00010.0053< 0.0001< 0.0001Lysine*#23.1 ± 6.97.7 ± 4.48.1 ± 4.37.5 ± 4.5< 0.0001< 0.00010.6786< 0.0001< 0.0001Arginine*107.2 ± 25.8138.3 ± 22.0135.9 ± 23.9140.2 ± 20.3< 0.0001< 0.00010.5674< 0.0001< 0.0001Phenylalanine#21.7 ± 7.831.0 ± 9.631.6 ± 9.830.6 ± 9.4< 0.0001< 0.00010.8225< 0.0001< 0.0001Pyroglutamic *38.1 ± 10.348.9 ± 11.348.8 ± 12.549.1 ± 10.5< 0.0001< 0.00010.9959< 0.0001< 0.0001Proline*9.5 ± 4.513.9 ± 5.115.0 ± 5.813.0 ± 4.4< 0.0001< 0.00010.0407< 0.0001< 0.0001Tryptophan#7.5 ± 3.911.4 ± 4.611.3 ± 4.911.5 ± 4.3< 0.0001< 0.00010.9452< 0.0001< 0.0001Histidine6.0 ± 2.98.4 ± 3.78.1 ± 3.48.5 ± 3.8< 0.0001< 0.00010.65430.0011< 0.0001&The value is the relative peak area; *glucogenic amino acids; #ketogenic amino acids. Variables were presented as mean ± SE for normally distributed data. Abbreviations: CT, control; HP, hypopituitarism; NAFLD, nonalcoholic fatty liver disease Circulated amino acid levels &The value is the relative peak area; *glucogenic amino acids; #ketogenic amino acids. Variables were presented as mean ± SE for normally distributed data. Abbreviations: CT, control; HP, hypopituitarism; NAFLD, nonalcoholic fatty liver disease Measurements of metabolic parameters Clinical and anthropometric parameters, such as age, sex, body mass index (BMI), blood pressure (BP), and laboratory results, were retrieved from the medical records. Biochemical analyses were performed after an 8-hour fast within 7 days of abdominal ultrasonography. The MetS score was calculated based on the following 5 parameters: (1) Hypertension: BP ≥ 130/85 mmHg or use of antihypertensive medication; (2) Central obesity: Waist circumference (WC) ≥ 90 cm in men and ≥ 80 cm in women; in cases where WC measurements were unavailable, a BMI ≥ 24 kg/m2 was used to indicate visceral obesity; (3) Elevated TG: TG ≥ 1.7 mmol/L (150 mg/dL); (4) Low HDL-C: HDL-C < 0.9 mmol/L (40 mg/dL) in men; (5) Hyperglycemia: Fasting blood glucose (FBG) level ≥ 5.6 mmol/L (100 mg/dL) or use of hypoglycemic medications. Each criterion met was assigned one point, and the MetS score was calculated as the sum of these points. Patients with HP were stratified into two groups based on their MetS scores: MetS-2 (n = 50) for those with a score ≥ 2 and MetS-1 (n = 83) for those with a score < 2. Statistical analysis Variables were presented as either mean ± SE for normally distributed data or as median (IQR) for non-normally distributed data. Statistically significant variations among the three groups of CTs, NAFLD-YES, and NAFLD-NO groups were evaluated using one-way ANOVA. Correlation analysis employed Pearson’s correlation coefficient for normally distributed variables and Spearman’s rank correlation for non-normally distributed variables. The associations between AA levels and metabolic parameters, including IR, MetS Scores, and NAFLD, were analyzed using multiple linear regression or multiple logistic regression. All analyses were performed with SPSS software version 23.0 (SPSS Inc., Chicago, IL, United States) and GraphPad Prism software version 9.0 (GraphPad Software, Inc). Significance tests were two-tailed, and statistical significance was set at P < 0.05. Results The clinical and biochemical parameters of the study population (HP and CT) are summarized in Table 1. The BMI of HP subjects was similar to that of CTs (25.6 ± 0.5 versus 24.1 ± 0.8 kg/m2, respectively), with no significant difference observed. Compared to the CT group, subjects with HP exhibited elevated levels of TG, CHOL, FBG, insulin and HOMA-IR. Compared to NAFLD-NO group, NAFLD‐YES subjects was significantly higher in BMI (25.7 ± 4.2 versus 22.4 ± 3.4), TG (2.6 ± 1.8 versus 1.7 ± 1.2), alanine aminotransferase (ALT) (median: 31.0 versus 23.0), gamma-glutamyltransferase (GGT) (median: 27 versus 20), insulin (median: 16.8 versus 8.2) and HOMA-IR (median: 3.5 versus 1.7) (P < 0.05), with no significant difference in age, CHOL, HDL-C, and LDL-C. When analyzed as a whole and compared to CTs, subjects with HP had higher circulated levels of alanine, phenylalanine, arginine, glutamate, pyroglutamic acid, leucine, valine, histidine, proline, and tryptophan, while the levels of glutamine, lysine, and norleucine was significantly decreased (all P < 0.0001). Moreover, when compared to NAFLD-NO, NAFLD-YES showed elevated levels of leucine, norleucine, and proline, while the levels of alanine decreased with significant difference (Table 2). Given the highly significant differences in AA levels observed between the CT and HP groups, statistically significant also observed among the three groups of CT, NAFLD-YES, and NAFLD-NO. Insulin-like growth factor-1 (IGF-1) was mainly used to assess the severity of GH deficiency in patients with hypopituitarism. Correlation analysis was performed to calculate the relationship between IGF-1 levels and circulating AA levels. A statistically significant correlation was observed only between glutamine and IGF-1, with a correlation coefficient (r) of 0.209 and a P-value less than 0.05 (Supplementary Table 1). Amino acids and insulin resistance In this cohort, 56.3% (75/133) of them have HOMA-IR values under 2.5, a level typically regarded as indicative of normal IR. Additionally, 15.8% (21/133) of patients had HOMA-IR scores ranging between 2.5 and 3.9, suggesting mild IR. Furthermore, 25.6% (34/133) of patients exhibited HOMA-IR scores between 4.0 and 9.9, indicating moderate IR. Only 2.25% (3/133) of patients recorded HOMA-IR scores exceeding 10, a threshold commonly associated with severe IR. (Fig. 1A). Fig. 1The prevalence of IR (A), MetS (B), and NAFLD (C) in patients with HP The prevalence of IR (A), MetS (B), and NAFLD (C) in patients with HP As expected, HOMA-IR exhibited significant correlations with insulin (r = 0.88, P < 0.001), FBG (r = 0.71, P < 0.001), BMI (r = 0.41, P < 0.001), TG (r = 0.37, P < 0.001), LDL-C (r = 0.33, P < 0.001), and CHOL (r = 0.24, P = 0.007), respectively. Certain AAs showed significant correlations with HOMA-IR (glutamate, r = 0.39, P < 0.001; glutamine, r = -0.22, P < 0.001; valine, r = 0.26, P = 0.002; leucine, r = 0.19, P = 0.025; norleucine, r = 0.19, P = 0.033) (Fig. 2). Fig. 2Pearson correlation between amino acids and biomarkers in carbohydrate, lipid metabolism in HP. The differential expression levels of amino acids and biomarkers in carbohydrate, lipid metabolism were processed using Z-score data normalization Pearson correlation between amino acids and biomarkers in carbohydrate, lipid metabolism in HP. The differential expression levels of amino acids and biomarkers in carbohydrate, lipid metabolism were processed using Z-score data normalization Multivariable linear regression analysis indicated that glutamate (P = 0.001), and norleucine (P = 0.039) were independently linked to HOMA-IR (Table 3). Furthermore, glutamate was found to be correlated with various clinical characteristics including BMI (r = 0.25, P = 0.004), LDL-C (r = 0.22, P = 0.010), GGT (r = 0.66, P < 0.001), FBG (r = 0.50, P < 0.001), and MetS score (r = 0.25, P = 0.004). Table 3Multiple regression analyses of AAs associated with HOMA-IR, MetS and NAFLD in HPsHOMA-IRMetS ScoreNAFLDβ ± SE P β ± SE P OR (95% CI) P Alanine-0.215 ± 0.1980.280-0.108 ± 0.0750.1550.647 (0.415–0.967)0.042Proline0.074 ± 0.0500.1430.043 ± 0.0180.0211.102 (1.009–1.219)0.042Glutamine-0.617 ± 0.7230.396-0.485 ± 0.2610.0660.423 (0.106–1.535)0.202Glutamate0.233 ± 0.0680.0010.073 ± 0.0260.0061.104 (0.962–1.310)0.215Norleucine2.703 ± 1.2920.0390.593 ± 0.5000.23711.55 (1.052–161.4)0.054Abbreviations: AA, amino acid; HOMA-IR, homeostasis model assessment of insulin resistance; MetS, metabolic syndrome; NAFLD, nonalcoholic fatty liver disease; HP, hypopituitarism Table 3Multiple regression analyses of AAs associated with HOMA-IR, MetS and NAFLD in HPsHOMA-IRMetS ScoreNAFLDβ ± SE P β ± SE P OR (95% CI) P Alanine-0.215 ± 0.1980.280-0.108 ± 0.0750.1550.647 (0.415–0.967)0.042Proline0.074 ± 0.0500.1430.043 ± 0.0180.0211.102 (1.009–1.219)0.042Glutamine-0.617 ± 0.7230.396-0.485 ± 0.2610.0660.423 (0.106–1.535)0.202Glutamate0.233 ± 0.0680.0010.073 ± 0.0260.0061.104 (0.962–1.310)0.215Norleucine2.703 ± 1.2920.0390.593 ± 0.5000.23711.55 (1.052–161.4)0.054Abbreviations: AA, amino acid; HOMA-IR, homeostasis model assessment of insulin resistance; MetS, metabolic syndrome; NAFLD, nonalcoholic fatty liver disease; HP, hypopituitarism Multiple regression analyses of AAs associated with HOMA-IR, MetS and NAFLD in HPs Abbreviations: AA, amino acid; HOMA-IR, homeostasis model assessment of insulin resistance; MetS, metabolic syndrome; NAFLD, nonalcoholic fatty liver disease; HP, hypopituitarism Amino acids and mets scores In this cohort, 12.8% (17/133) of patients were found to have MetS scores of 3, while 24.8% (33/133) displayed MetS scores of 2. Furthermore, 30.8% (41/133) of patients had MetS scores of 1, and 31.6% (42/133) had MetS scores of 0 (Fig. 1B). Significant correlations were observed between MetS scores and various clinical factors, including BMI (r = 0.49, P < 0.001), TG (r = 0.54, P < 0.001), HDL-C (r = -0.18, P = 0.034), FBG (r = 0.43, P < 0.001), insulin (r = 0.54, P < 0.001), HOMA-IR (r = 0.59, P < 0.001), and NAFLD (r = 0.40, P < 0.001). Additionally, specific AAs showed significant correlations with MetS scores, such as glutamate (r = 0.25, P = 0.004), glutamine (r = -0.23, P = 0.008), valine (r = 0.19, P = 0.032), and proline (r = 0.22, P = 0.012) (Fig. 2). Among them, glutamate (P = 0.006) and proline (P = 0.021) were identified as independently associated with MetS scores through multivariable linear regression analysis (Table 3). Group 1 consisted of MetS score 1 and 0 (n = 83), while group 2 was defined as MetS score ≥ 2 (n = 50). HOMA-IR exhibited a significant difference between group 2 and group 1 (P < 0.0001). Additionally, the AAs, glutamate, glutamine, valine, and proline, also showed significant differences between these two groups with P values of 0.049, 0.005, 0.047, and 0.0004, respectively (Fig. 3). Fig. 3The distribution of HOMA-IR (A) and amino acids [glutamate (B); glutamine (C); valine (D); proline (E)] depending MetS scores in patients with HP and health controls. Group 2: Mets scores ≥ 2 (n = 50); Group 1: Mets scores < 2 (n = 83). ****P < 0.0001, ***P < 0.001, **P < 0.01, *P < 0.05 The distribution of HOMA-IR (A) and amino acids [glutamate (B); glutamine (C); valine (D); proline (E)] depending MetS scores in patients with HP and health controls. Group 2: Mets scores ≥ 2 (n = 50); Group 1: Mets scores < 2 (n = 83). ****P < 0.0001, ***P < 0.001, **P < 0.01, *P < 0.05 Amino acids and NAFLD In this cohort, 42.8% (57/133) of patients were diagnosed with NAFLD (Fig. 1C). The presence of NAFLD was found to be correlated with clinical factors, including BMI (r = 0.39, P < 0.001), TG (r = 0.28, P = 0.001), CHOL (r = 0.24, P = 0.005), LDL-C (r = 0.23, P = 0.008), insulin (r = 0.41, P < 0.001), and HOMA-IR (r = 0.39, P < 0.001) (Fig. 2). The alterations in AA levels in the presence of NAFLD were investigated. Some AAs showed significant correlations with NAFLD, including leucine (r = 0.21, P = 0.014), norleucine (r = 0.23, P = 0.009), alanine (r = -0.29, P = 0.032), and proline (r = 0.22, P = 0.016). Notably, alanine (P = 0.042) and proline (P = 0.042) was identified as being independently associated with NAFLD based on multivariable logistic regression analysis. Discussion Patients with HP may have an increased risk of developing NAFLD due to the growth hormone deficiency, known to influence lipid metabolism and insulin sensitivity, both of which are key factors in the development of NAFLD (Garmes 2024). Most HP patients have NAFLD present normal liver enzymes, even they have a high risk to develop cirrhosis (Garmes et al. 2021). For this reason, it is important to identify circulating markers for HP patients at risk of NAFLD. In our study, we found AAs of glutamate and norleucine independently correlated with HOMA-IR, while alanine and proline were associated with NAFLD. The MetS score was linked to glutamate and proline. However, neither glutamic acid nor BCAAs alone are useful for distinguish of HP patients with NAFLD. IR, Mets and NAFLD in HP The state of IR on HP still in debate, with discrepant results. Some studies showed that these patients present IR (Garmes et al. 2021; Regal et al. 2001), however most of them showed that IR are similar to controls matched for sex, age and BMI (Guevara-Aguirre et al. 2020). As obesity is indirectly correlated with IR, and so it would be important to pair patients with a control group with the same age and BMI to obtain more reliable results. We selected a cohort of HP without GHRT, to exclude the impact of GH replacement on metabolic parameters, as well as a control group with similar gender, age and BMI. We found subjects with HP had higher levels of TG, CHOL, FBG, insulin, and IR. There are 15.8% (21/133) with mild IR, 25.6% (34/133) with moderate IR, and 2.25% (3/133) with severe IR. When divided according to the presence of NAFLD, NAFLD-YES subjects was significantly higher in BMI, insulin, and HOMA-IR. This indicates that there is a definite increase in the level of IR in young patients with childhood-onset HP, which is independent of BMI. The rise in IR and the increase in BMI, potentially heightens the susceptibility to NAFLD in this patient population. MetS is a very complex disorder that represent risk factors for the development of cardiovascular disease, type 2 diabetes mellitus (T2DM), central fatty tissue and IR (Shin et al. 2018). MetS is a common finding in adult HPs, which is mainly influenced by increased BMI and untreated GH deficiency. Most studies on patients with HP demonstrate that the prevalence of MetS is increased, but with varying results. It shows the overall prevalence rate of MetS in HP without GHRT was 51.8% in the USA and 28.6% in Europe (Nyenwe et al. 2009; van der Klaauw et al. 2007). In our study, the prevalence rate of MetS was relatively lower with 12.8% (17/133), which is likely that our enrolled patients are relatively young, with a mean age of 24.5 years. Although there are relatively few patients who meet the criteria for MetS diagnosis, many patients are in pre-MetS stage with MetS scores of 2 (24.8% (33/133) and 1 (30.8% (41/133). Which indicating that the prevalence of MetS in this age group will increase as their age grow. As expected, NAFLD has a strong correlation with MetS. Circulating glutamate, glutamine and BCAAs in HP AA levels are often found altered in metabolic diseases, including T2DM, NAFLD, and obesity, possibly they are also associated with IR (Lehmann et al. 2013). In our study, metabolomic analysis showed markedly higher levels of alanine, phenylalanine, arginine, glutamate, pyroglutamic acid, leucine, valine, histidine, proline, and tryptophan, while the levels of glutamine, lysine, and norleucine was significantly decreased. Among them, the AAs of glutamate and valine show a positive correlation with IR, while glutamine show negative correlation with IR. This is in agreement with previous studies that have often found glutamate and BCAAs increases in an IR state, such as obesity, T2DM, prediabetes, and/or NAFLD (Floegel et al. 2013). Glutamate, a pivotal player in numerous biochemical reactions, emerges as a product of AA transamination (Yelamanchi et al. 2016). Glutamate serves as a crucial substrate for the synthesis and turnover of intracellular glutathione (GSH), a key antioxidant molecule. A compelling report has illuminated the dysregulation of hepatic enzymes associated with glutamate metabolism in patients afflicted with NASH. Notably, the study revealed that most enzymes implicated in this metabolic pathway, were significantly down-regulated in the liver tissues (Mardinoglu et al. 2014). An increased circulating glutamate and BCAAs were frequently found in diseases related with IR. Kalhan et al. reported (Kalhan et al. 2011) higher levels of glutamate and BCAAs, but lower plasma GSH and GSH metabolites in NASH compared to healthy controls. Moreover, elevated glutamate levels have been linked to a higher prevalence of T2DM, cardiovascular diseases and NAFLD (Chaouche et al. 2024). In two cohorts free of diabetes study, Cheng et al., (Cheng et al. 2012) found plasma glutamate concentration was strongly associated with IR and circulating TG. Low glutamine and high glutamate concentrations were associated with increased risk of heart failure (Papandreou et al. 2020). In another diabetes study, decreased glycine and increased valine, leucine, and combined glutamine and glutamate was associated to conversion to T2DM (Palmer et al. 2015). Elevated concentrations of BCAAs have been implicated in the pathogenesis of IR, T2DM, and various cardio metabolic diseases (White et al. 2021). In disease of obesity and IR, levels of valine and leucine/isoleucine have been reported to increase by 20% and 14%, respectively. This rise in BCAAs has been linked to the downregulation of multiple enzymes in the BCAA oxidation pathway (She et al. 2007; Herman et al. 2010). The circulating BCAAs increased in IR related diseases, however, they are mere consequences, causative factors, or merely biomarkers of IR, remains to be elucidated. Cahill et al. proposed that the elevated levels of BCAAs and aromatic amino acids observed in individuals with obesity may be a consequence, rather than a cause, of IR (Felig et al. 1969). However, nevertheless, some evidence suggests that BCAAs may independently contribute to IR. Metabolomic studies have indicated that elevated BCAA levels in individuals with normal fasting glycemia are associated with an increased risk of IR and diabetes (White et al. 2021). Recent human genetic studies examining variants affecting insulin sensitivity and lipid traits in relation to BCAA levels (Wang et al. 2021). According to this model, the increases in BCAA observed in pre-diabetic individuals with obesity may primarily result from IR. However, once elevated, BCAAs could potentially play a causal role in the progression from prediabetes to full-blown diabetes (Wang et al. 2017). In a large cohort follow-up study involving 1302 subjects, high levels of BCAAs were found to exacerbate the risk of coronary heart disease and cardiovascular events by influencing the occurrence and progression of NAFLD (Guo et al. 2022). It is conceivable that elevated BCAAs may stem from IR but may subsequently affect lipid and glucose metabolism in the context of HP. However, our study does have several limitations. Firstly, as a cross-sectional one, our study has limitations in establishing causal relationships due to its static nature. Secondly, the patient cohort enrolled in this study consisted of individuals with HP, a condition in which hormonal replacement therapy is known to influence AA metabolism. To mitigate the impact of GH replacement on AA metabolism, all participants ceased GH therapy for over 24 months prior to sample collection in this examination. However, it was not feasible to discontinue replacement therapy with glucocorticoids and/or thyroid hormone, and thus, they received standard dosages as part of their routine treatment. Furthermore, given that these patients were of reproductive age with fertility goals, they also underwent replacement therapy with sex hormones,. Consequently, this study cannot rule out the potential effects of glucocorticoids, thyroid hormone, and/or sex hormones on AA metabolism. Additionally, without data on other medications that patients may be taking, which can also alter AA metabolomics, the study may not accurately reflect the true impact of the factors being studied. In conclusion, most AAs were found to be elevated in HP subjects compared to CTs. Glutamate, glutamine, norleucine and branched-chain amino acids (BCAAs) (leucine and valine) were correlated with the HOMA-IR, with glutamate and norleucine showing independent linkage. Glutamate and proline levels were specifically associated with MetS score, while alanine and proline linked to NAFLD. Given that elevated glutamate and BCAAs levels have higher prevalence of NAFLD, we hypothesized that the changes in AAs observed in HP may be attributed to the impact of NAFLD and IR. Electronic supplementary material Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1