Interactions between the gut microbiome, associated metabolites and the manifestation and progression of heart failure with preserved ejection fraction in ZSF1 rats

Abstract

Background: Heart failure with preserved ejection fraction (HFpEF) is associated with systemic inflammation, obesity, metabolic syndrome, and gut microbiome changes. Increased trimethylamine-N-oxide (TMAO) levels are predictive for mortality in HFpEF. The TMAO precursor trimethylamine (TMA) is synthesized by the intestinal microbiome, crosses the intestinal barrier and is metabolized to TMAO by hepatic flavin-containing monooxygenases (FMO). The intricate interactions of microbiome alterations and TMAO in relation to HFpEF manifestation and progression are analyzed here.

Methods: Healthy lean (L-ZSF1, n = 12) and obese ZSF1 rats with HFpEF (O-ZSF1, n = 12) were studied. HFpEF was confirmed by transthoracic echocardiography, invasive hemodynamic measurements, and detection of N-terminal pro-brain natriuretic peptide (NT-proBNP). TMAO, carnitine, symmetric dimethylarginine (SDMA), and amino acids were measured using mass-spectrometry. The intestinal epithelial barrier was analyzed by immunohistochemistry, in-vitro impedance measurements and determination of plasma lipopolysaccharide via ELISA. Hepatic FMO3 quantity was determined by Western blot. The fecal microbiome at the age of 8, 13 and 20 weeks was assessed using 16s rRNA amplicon sequencing.

Results: Increased levels of TMAO (+ 54%), carnitine (+ 46%) and the cardiac stress marker NT-proBNP (+ 25%) as well as a pronounced amino acid imbalance were observed in obese rats with HFpEF. SDMA levels in O-ZSF1 were comparable to L-ZSF1, indicating stable kidney function. Anatomy and zonula occludens protein density in the intestinal epithelium remained unchanged, but both impedance measurements and increased levels of LPS indicated an impaired epithelial barrier function. FMO3 was decreased (- 20%) in the enlarged, but histologically normal livers of O-ZSF1. Alpha diversity, as indicated by the Shannon diversity index, was comparable at 8 weeks of age, but decreased by 13 weeks of age, when HFpEF manifests in O-ZSF1. Bray-Curtis dissimilarity (Beta-Diversity) was shown to be effective in differentiating L-ZSF1 from O-ZSF1 at 20 weeks of age. Members of the microbial families Lactobacillaceae, Ruminococcaceae, Erysipelotrichaceae and Lachnospiraceae were significantly differentially abundant in O-ZSF1 and L-ZSF1 rats.

Conclusions: In the ZSF1 HFpEF rat model, increased dietary intake is associated with alterations in gut microbiome composition and bacterial metabolites, an impaired intestinal barrier, and changes in pro-inflammatory and health-predictive metabolic profiles. HFpEF as well as its most common comorbidities obesity and metabolic syndrome and the alterations described here evolve in parallel and are likely to be interrelated and mutually reinforcing. Dietary adaption may have a positive impact on all entities.

Keywords: FMO3; HFpEF; Inflammation; Intestinal barrier; Intestinal microbiome; TMAO; ZSF1-rats.

Fig. 1

TMAO, TMA-processing flavin containing dimethylaniline monoxygenase 3 (FMO3), NT-proBNP, carnitine, symmetric dimethylarginine (SDMA) and lipopolysaccharide (LPS) quantities in blood/liver of obese (black circles) ZSF1 rats with HFpEF and lean control rats (white circles) (experimental groups n = 12, except FMO3 5 vs. 5 and LPS 10 vs. 11 due to limited sample availability). Lines indicate the median

Fig. 2

Normalized cellular impedance of colon epithelial cells from obese (black circles and boxes) ZSF1 rats with HFpEF and lean control rats (white circles and boxes) that were exposed to 100 µM histamine (circles) or 1% ethanol (squares) to test the epithelial barrier function. Non-linear fitting curves were added (blue = obese animals, red = lean animals). They differ significantly between both experimental groups (p < 0.0001 for histamine and EtOH)

Fig. 3

Average Height of Mucosa in cross-sections of colon samples from obese (black cicrles) ZSF1 rats and lean control rats (white circles). Average Mean Gray Value per mm² in areas of interest in colon samples stained with ZO1-antibodies

Fig. 4

Phylum-level core microbiome composition averaged over time in lean and obese animals at eight, 13 and 20 weeks using relative abundance of phylae. Taxa were aggregated on the phylum-level using a detection threshold of 0.001 and a prevalence threshold of 0.1. Family-level core microbiome composition averaged over time in lean and obese animals at eight, 13 and 20 weeks using relative abundance of families and portraying the top five most prevalent families. Taxa were aggregated on the family-level using a detection threshold of 0.01 and a prevalence threshold of 0.5

Fig. 5

Top: Alpha-diversity represented by Shannon-Index and richness in obese and lean ZSF1 rats at 8, 13 and 20 weeks of age. Lines indicate the median. Bottom: Beta-Diversity in lean versus obese ZSF1-rats at baseline (p = 0.200), 13 weeks (p = 0.001) und 20 weeks (p = 0.007). Non-metric multidimensional scaling (NMDS) was used to illustrate dissimilarities between samples. Ellipses used to show significant grouping. Diversity parameters were analyzed after rarefaction