Association of Plasma Branched-Chain Amino Acid With Biomarkers of Inflammation and Lipid Metabolism in Women

Abstract

Background: Branched-chain amino acids (BCAAs; isoleucine, leucine, and valine) correlate with insulin resistance and poor glucose control, which may in part explain associations between type 2 diabetes and cardiovascular disease. However, the relationships of BCAAs with other cardiometabolic pathways, including inflammation and dyslipidemia, are unclear. We hypothesized that plasma BCAAs would correlate with multiple pathways of cardiometabolic dysfunction.

Methods: We conducted a cross-sectional analysis among 19 472 participants (mean age=54.9 years, SD=7.2 years) in the Women's Health Study without a history of type 2 diabetes, cardiovascular disease, or cancer. We quantified the concentrations of individual biomarkers of inflammation and lipids, across quartiles of BCAAs, adjusting for age, smoking, body mass index, physical activity, and other established cardiovascular disease risk factors at blood draw.

Results: Women in the highest versus lowest quartiles of plasma BCAAs had higher inflammatory markers including high-sensitivity C-reactive protein (multivariable-adjusted means: 1.96 versus 1.43 mg/L), fibrinogen (367 versus 362 mg/dL), soluble intercellular cell adhesion molecule-1 (361 versus 353 ng/mL), and glycoprotein acetylation (407 versus 371 µmol/L; P trend=0.0002 for fibrinogen; P<0.0001 for others). Similarly for lipids, women with higher BCAAs had lower HDL-C (high-density lipoprotein cholesterol; 49.0 versus 55.0 mg/dL), and higher triglycerides (143 versus 114 mg/dL), LDL-C (low-density lipoprotein cholesterol; 133 versus 124 mg/dL), and lipoprotein insulin resistance score (52.6 versus 37.3; all: P<0.0001). Similar associations with these biomarkers were observed in isoleucine, leucine, and valine, respectively.

Conclusions: Higher circulating BCAA concentrations are associated with adverse profiles of biomarkers of inflammation and dyslipidemia independent of established cardiovascular disease risk factors, and thus, may reflect poorer cardiometabolic health through multiple pathways. Registration: URL: https://www.clinicaltrials.gov; Unique identifier: NCT00000479.

Keywords: amino acids; cardiovascular diseases; dyslipidemia; inflammation; lipid metabolism.

Conflict of interest statement

Disclosures: S Mora received institutional research grant support from Atherotech Diagnostics for research outside the current work, served as a consultant and member of the scientific advisory board to Quest Diagnostics, served as a consultant to Pfizer for work outside the current study, and has a patent regarding the use of GlycA in relation to colorectal cancer risk. P M Ridker received unrelated research grants from Kowa, Novartis, Amarin, and NHLBI during the period of this study and has served as a research consultant to Novartis, AstraZeneca, Jansen, Agepha, Inflazome, IQVIA, Uppton, Agepha, CiviBiopharm, Corvidia, and Flame for work unrelated to this manuscript. The other authors had no conflicts of interests related to this manuscript.

Figures

Figure 1:

Correlation network between BCAA, inflammatory biomarkers, lipid biomarkers, and HbA1c. Correlation network between BCAA, hsCRP, fibrinogen, GlycA, sICAM-1, triglyceride, HDL-c, LDL-c, LPIR, and HbA1c is shown. Spearman correlation coefficients (ρ) were calculated in each pair of biomarkers and only associations with ρ>0.10 are visualized as lines. Thicker lines indicate stronger correlations and green and red lines represent positive and negative correlations, respectively. Abbreviations: BCAA, branched chain amino acid; hsCRP, high-sensitivity C-reactive protein; sICAM-1, soluble intercellular cell adhesion molecule-1; HDL-c, HDL cholesterol; LDL-c, LDL cholesterol; LPIR, lipoprotein insulin resistance score; HbA1c, hemoglobin A1C.

Figure 2:

Standardized differences of cardiometabolic biomarkers per SD changes of BCAA levels. Linear regressions of standardized biomarkers constructed by standardized continuous total BCAA levels and covariates [age at randomization (continuous), assignment to ASA group, assignment to vitamin E group, race (white or non-white), family history of diabetes, smoking history (none, ever, current), menopausal status (premenopausal, postmenopausal [natural], postmenopausal [non-natural], unsure), use of menopausal hormone therapy (never, past, current), parity as number of pregnancies lasting ≥6 months (nulliparous, 0, 1, 2, ≥3), exercise as total MET-hour/week (quintiles), aHEI-2010 (quintiles), alcohol consumption (none,