Exploring the Role of Glycine Metabolism in Coronary Artery Disease: Insights from Human Genetics and Mouse Models

Abstract

Background: Circulating glycine levels have been associated with reduced risk of coronary artery disease (CAD) in humans but these associations have not been observed in all studies. We evaluated whether the relationship between glycine levels and atherosclerosis was causal using genetic analyses in humans and feeding studies in mice. Methods: Serum glycine levels were evaluated for association with risk of CAD in the UK Biobank. Genetic determinants of glycine levels were identified through a genome-wide association study (GWAS) and used to evaluate the causal relationship between glycine and risk of CAD by Mendelian randomization (MR). A dietary supplementation study was carried out with atherosclerosis-prone apolipoprotein E deficient (ApoE^-/-) mice to determine the effects of increased circulating glycine levels on cardiometabolic traits and aortic lesion formation. Results: Among 105,718 UK Biobank subjects, elevated serum glycine levels were associated with significantly reduced risk of prevalent CAD (Quintile 5 vs. Quintile 1 OR = 0.76, 95% CI 0.67-0.87; p < 0.0001) and incident CAD (Quintile 5 vs. Quintile 1 HR = 0.70, 95% CI 0.65-0.77; p < 0.0001) after adjustment for age, sex, ethnicity, anti-hypertensive and lipid-lowering medications, blood pressure, kidney function, and diabetes. A GWAS meta-analysis with 230,947 subjects identified 61 loci for glycine levels, of which 26 were novel. MR analyses provided modest evidence that genetically elevated glycine levels were causally associated with reduced systolic blood pressure and risk of type 2 diabetes, but did not provide significant evidence for an association with decreased risk of CAD. Glycine supplementation in mice had no effects on cardiometabolic traits or atherosclerotic lesion development. Conclusions: While expanding the genetic architecture of glycine metabolism, MR analyses and in vivo feeding studies did not provide evidence that the clinical association of this amino acid with atherosclerosis represents a causal relationship.

Keywords: Mendelian randomization; atherosclerosis; coronary artery disease; dietary supplementation; genome-wide association study; glycine; mice.

Figure 1

Association of Serum Glycine Levels with Risk of CAD in the UK Biobank. Individuals in the highest quintile of glycine levels had significantly reduced risk of prevalent CAD (OR = 0.76, 95% CI 0.67–0.87; p < 0.0001) (A) and incident CAD (HR = 0.70, 95% CI 0.65–0.77; p < 0.0001) (B) compared to individuals in the first quintile. p-value for trend for association with risk of prevalent CAD (A) and log-rank p-value for association with incident risk of CAD (B) across quintiles are also shown, including the number of subjects at risk of incident CAD in each quintile at baseline and the indicated follow-up time (bottom panel in B).

Figure 1

Association of Serum Glycine Levels with Risk of CAD in the UK Biobank. Individuals in the highest quintile of glycine levels had significantly reduced risk of prevalent CAD (OR = 0.76, 95% CI 0.67–0.87; p < 0.0001) (A) and incident CAD (HR = 0.70, 95% CI 0.65–0.77; p < 0.0001) (B) compared to individuals in the first quintile. p-value for trend for association with risk of prevalent CAD (A) and log-rank p-value for association with incident risk of CAD (B) across quintiles are also shown, including the number of subjects at risk of incident CAD in each quintile at baseline and the indicated follow-up time (bottom panel in B).

Figure 2

Multi-ancestry GWAS Meta-analysis and Genetic Risk Score (GRS) Analysis for Circulating Glycine Levels. (A) Manhattan plot shows 61 loci significantly associated with circulating glycine levels in 230,947 subjects. Novel (26) and known (35) loci are indicated by red and green dots, respectively. The seven non-pleiotropic loci are indicated by yellow dots, of which two loci on chromosomes 3 and 16 were also novel (purple arrow heads). Genome-wide thresholds for significant (p = 5.0 × 10⁻⁸) and suggestive (p = 5.0 × 10⁻⁶) association are indicated by dashed gray lines. p-values are truncated at −log₁₀ (p-value) = 40. (B) Serum glycine levels are increased as a function of quintiles of a weighted GRS constructed with the number of glycine-raising alleles carried by individuals in the UK Biobank for the 61 loci identified in the meta-analysis (n = 23,283/quintile; total n = 116,412). Mean glycine levels are shown for each quintile (p-trend = 1.6 × 10⁻⁴).

Figure 3

Results of MR Analyses to Evaluate Causal Association of Circulating Glycine Levels with Risk of CAD. Effect sizes of lead variants for circulating glycine levels identified in the meta-analysis (x-axis) are plotted against effect sizes for risk of CAD based on previously published summary statistics (y-axis). Slopes of the regressions are represented by the colored lines and derived from tests of MR by inverse variance weighted (light blue), weighted median (dark blue), or MR Egger (green) methods for all 61 glycine-associated loci (A), the 7 non-pleiotropic loci (B), and the 3 non-pleiotropic loci harboring genes involved in the glycine cleavage system (C).

Figure 4

Effect of Glycine Supplementation on Plasma Cardiometabolic Traits and Atherosclerosis Development in ApoE^−/− Mice. Compared to the 0.3% glycine diet (Control), mice fed the 2% glycine content diet (Glycine) had significantly higher fasting (A) and non-fasting (B) glycine levels, particularly among male mice. After 16 weeks of glycine supplementation, there were no differences in body weight (C) or fasting plasma levels of total cholesterol (D) and LDL/VLDL (E), whereas fasting levels of HDL cholesterol (F) and triglycerides (G) were decreased. There were also no differences with respect to metabolic traits, including fasting glucose (H) and insulin (I) levels, or HOMA-IR (J). Glycine supplementation did not affect atherosclerotic lesion formation assessed through serial cryosections at the aortic arch (K) or along the entire aorta by en face analysis (L). Representative sections of aortic lesions (K) and en face aortas stained for lipid content (L) are shown from female mice in the control and glycine-fed groups. Data are represented as mean ± SE. p-values are derived from t-tests carried out between control and glycine fed groups separately in males (square symbols) and females (circle symbols). * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001. n = 25–36 for data in panels (A,C–J); n = 9–14 for data in panel (B); n = 16–27 for data in panel (K); n = 8–9 for data in panel (L).