Circulating Amino Acid Profiles in Adults with Abnormal Body Mass Index: Associations with Triglycerides, HDL Cholesterol, LDL Cholesterol, and Total Cholesterol in an Exploratory Cross-Sectional Metabolomic Pilot Study

Background/Objectives: Circulating amino acids are not only markers of nutritional status; in experimental and interventional models they have been linked to hepatic lipogenesis, lipoprotein assembly, mitochondrial fatty acid oxidation, and bile acid conjugation, which makes them plausible candidate correlates of obesity-related lipid dysregulation. Despite this, most metabolomic studies of excess adiposity have focused either on a single lipid parameter—typically triglycerides—or only on branched-chain amino acids (BCAAs). This pilot study was designed to generate hypotheses about these associations across the full standard lipid panel using a targeted 19-amino acid liquid chromatography–tandem mass spectrometry (LC-MS/MS) panel in adults spanning the body mass index (BMI) spectrum, with an analytical framework oriented around lipid phenotype variance rather than body weight classification. The design is cross-sectional and the analysis exploratory; no causal or predictive claim is made. Methods: Targeted LC-MS/MS quantification of 19 plasma amino acids was performed in 50 adults grouped as normal weight (n = 20; BMI 22 ± 1.5 kg/m2), overweight (n = 20; BMI 28 ± 1.5 kg/m2), or obese (n = 10; BMI 34 ± 2.0 kg/m2). The analytical framework included: (i) one-way analysis of variance (ANOVA) with Bonferroni correction; (ii) Pearson correlation analysis; (iii) principal component analysis (PCA) performed on the lipid profile itself with amino acid projection vectors; (iv) K-means clustering based on lipid phenotype (K = 3); (v) Ward-linkage hierarchical clustering of the amino acid–lipid correlation matrix; (vi) Random Forest permutation importance for all four lipid outcomes; and (vii) composite lipid risk indices including atherogenic index (TG/HDL-C) and non-HDL cholesterol. Results: A distinct amino acid correlation pattern was observed for each of the four lipid fractions. Triglycerides (TG) correlated most strongly with glutamic acid (r = 0.58) and inversely with glutamine (r = −0.58). High-density lipoprotein cholesterol (HDL-C) correlated most strongly with glutamic acid (r = −0.61) and serine (r = 0.49). The strongest correlates of low-density lipoprotein cholesterol (LDL-C) were phenylalanine (r = 0.56) and leucine (r = 0.56), and those of total cholesterol (TC) were leucine (r = 0.63) and, inversely, glycine (r = −0.54). The atherogenic index (TG/HDL-C) increased 2.9-fold from normal to obese and was most strongly correlated with glutamic acid, isoleucine, and glycine. In the multivariable models the amino acid panel accounted for a modest share of the variance in TG (adjusted R2 = 0.43; F(19,30) = 2.97, p = 0.004) and HDL-C (adjusted R2 = 0.35; F(19,30) = 2.40, p = 0.016). For LDL-C and TC the adjusted R2 values were close to zero (0.06 for both) and the overall models were not statistically significant (both p > 0.33); no interpretable amino acid signal was therefore present for these two fractions, and no predictors are reported for them. Lipid-based K-means clustering identified three lipid-phenotype clusters (Favorable, Intermediate, Adverse lipid profiles) with differences in amino acid z-scores. Conclusions: In this exploratory pilot cohort, lipid dysregulation in abnormal BMI was associated with two partially separable amino acid axes: a glutamic acid–glutamine axis (TG and partially HDL-C), and a glycine–serine putatively protective pattern opposing all atherogenic lipid parameters. These findings extend the established BCAA–insulin-resistance paradigm and suggest that targeted amino acid profiling—particularly for glutamic acid, leucine, glycine, and serine—may serve as a way of discovering candidate biomarkers for further study for dyslipidemia in individuals with abnormal BMI.