Gut Microbiota-Generated Phenylacetylglutamine and Heart Failure

Abstract

Background: The gut microbiota-dependent metabolite phenylacetylgutamine (PAGln) is both associated with atherothrombotic heart disease in humans, and mechanistically linked to cardiovascular disease pathogenesis in animal models via modulation of adrenergic receptor signaling.

Methods: Here we examined both clinical and mechanistic relationships between PAGln and heart failure (HF). First, we examined associations among plasma levels of PAGln and HF, left ventricular ejection fraction, and N-terminal pro-B-type natriuretic peptide in 2 independent clinical cohorts of subjects undergoing coronary angiography in tertiary referral centers (an initial discovery US Cohort, n=3256; and a validation European Cohort, n=829). Then, the impact of PAGln on cardiovascular phenotypes relevant to HF in cultured cardiomyoblasts, and in vivo were also examined.

Results: Circulating PAGln levels were dose-dependently associated with HF presence and indices of severity (reduced ventricular ejection fraction, elevated N-terminal pro-B-type natriuretic peptide) independent of traditional risk factors and renal function in both cohorts. Beyond these clinical associations, mechanistic studies showed both PAGln and its murine counterpart, phenylacetylglycine, directly fostered HF-relevant phenotypes, including decreased cardiomyocyte sarcomere contraction, and B-type natriuretic peptide gene expression in both cultured cardiomyoblasts and murine atrial tissue.

Conclusions: The present study reveals the gut microbial metabolite PAGln is clinically and mechanistically linked to HF presence and severity. Modulating the gut microbiome, in general, and PAGln production, in particular, may represent a potential therapeutic target for modulating HF.

Registration: URL: https://clinicaltrials.gov/; Unique identifier: NCT00590200 and URL: https://drks.de/drks_web/; Unique identifier: DRKS00020915.

Keywords: gut microbiome; heart failure; metabolism; phenylacetylglutamine.

Fig. 1.. The association of PAGln with heart failure (HF).

Box whisker plots of circulating PAGln levels in (A) US Cohort subjects (n=3256) or (C) European Cohort subjects (n=829) without (n=2544 and n=276, respectively) and with (n=712 and n=553, respectively) HF. Data are represented as boxplots: middle line is the median, the lower and upper hinges are the first and third quartiles, the whiskers represent 10^th and 90^th percentile; P values were calculated using Wilcoxon rank sum test. (B/D) Risk of HF among all test subjects according to PAGln quartile levels using a multivariable logistic regression models. Unadjusted odd ratio (blue), adjusted model 1 [age, sex, smoking status, SBP, LDL, HDL, TG, hs-CRP, diabetes and obesity (BMI≥30 kg/m²), black]; and adjusted model 2 [model 1 plus indices of renal function (eGFR≥60, or <60 mL/min per 1.73 m²), red]. Symbols represent odds ratios, and the 5%–95% confidence intervals (CI) are indicated by the line length.

Fig. 2.. Box-whisker plots showing circulating PAGln levels stratified by heart failure status in the merged US and European Cohort (n=4085).

Whiskers represent 5^th and 95^th percentile, P values were calculated using Kruskal - Walls (K.W.) test with Dunn’s test. The classification of heart failure (HF), according to left ventricle ejection fraction (LVEF): HF with preserved ejection fraction, (HFpEF), LVEF≥50; HF with mildly reduced ejection fraction, (HFmEF), 40<LVEF<50; or HF with reduced ejection fraction, HFrEF, LVEF≤40).

Fig. 3.. PAGln tracks with reduced ejection fraction and increased circulating levels of NT-proBNP.

Left ventricle ejection fraction (%; LVEF) broken down by PAGln quartile in (A) US Cohort subjects (n=3256) or (B) European Cohort subjects (n=829). Circulating NT-proBNP (pg/ml) levels broken down by PAGln quartile in (C) US Cohort subjects (n=3256) and (D) Europe Cohort subjects (n=829). Data are represented as boxplots: middle line is the median, the lower and upper hinges are the first and third quartiles, the whiskers represent 10^th and 90^th percentile; P values were calculated using Kruskal-Wallis (K.W.) test.

Fig. 4.. PAG rapidly induces BNP gene expression in vitro and in vivo.

(A) Fold-change of BNP (Nppb) gene expression in H9c2 rat cardiomyoblasts, normalized to GAPDH, after 4 h exposure to 100 μM PAGln or PAGly measured in the indicated number of replicates (n=10–12 as indicated). (B) Fold-change of BNP (Nppb) gene expression in isolated left atria, normalized to GAPDH, 15 min after IP injection of PAGln (50 mg/kg), PAGly (50 mg/kg), or vehicle control measured in the indicated number of animals (n=14–16 as indicated). Circulating plasma levels of PAGln and PAGly (mean ± SEM μM) are indicated below the graph. One-way ANOVA followed by Tukey’s multiple comparison test listed. All bars represented the mean ± SEM and dots indicate single data points.

Fig. 5.. PAGln elicits negative ionotropic effect.

(A, C) Myograph of sarcomere contraction over ~1 second interval. Sarcomere contraction tracings of vehicle, epinephrine (Epi), PAGln, PAGly, Epi+PAGln, and Epi+PAGly were recorded over time. All tracings were trimmed so contractions were overlaid. Expanded single sarcomere contraction cycle. (B, D) Relative contraction distance represented as sarcomere shortening, was calculated considering five different scanning windows of a single sarcomere contraction cycles shown in panels A and C (n=5). Data shown represents mean ± SEM of the average maximal shortening per contraction. Nonparametric Mann-Whitney (M.W.) test was used for non-pairwise comparisons and Kruskal-Wallis (K.W.) test for multiple comparisons.

Fig. 6.. Scheme showing relationship between gut microbial PAGln pathway and heart failure.

Following dietary ingestion of protein, phenylalanine is metabolized by the gut microbiome to produce phenylacetic acid. Upon absorption into the host and further conjugation, phenylacetyl glutamine (PAGln) is produced. An obligate gut-microbial metabolite, PAGln is shown to be clinically associated with heart failure presence, severity, and numerous heart failure associated phenotypes, including decreased cardiomyocyte contraction via adrenergic receptors.