Sequence meets function—microbiota and cardiovascular disease

Abstract

The discovery that gut-microbiota plays a profound role in human health has opened a new avenue of basic and clinical research. Application of ecological approaches where the bacterial 16S rRNA gene is queried has provided a number of candidate bacteria associated with coronary artery disease and hypertension. We examine the associations between gut microbiota and a variety of cardiovascular disease (CVD) including atherosclerosis, coronary artery disease, and blood pressure. These approaches are associative in nature and there is now increasing interest in identifying the mechanisms underlying these associations. We discuss three potential mechanisms including: gut permeability and endotoxemia, increased immune system activation, and microbial derived metabolites. In addition to discussing these potential mechanisms we highlight current studies manipulating the gut microbiota or microbial metabolites to move beyond sequence-based association studies. The goal of these mechanistic studies is to determine the mode of action by which the gut microbiota may affect disease susceptibility and severity. Importantly, the gut microbiota appears to have a significant effect on host metabolism and CVD by producing metabolites entering the host circulatory system such as short-chain fatty acids and trimethylamine N-Oxide. Therefore, the intersection of metabolomics and microbiota research may yield novel targets to reduce disease susceptibility. Finally, we discuss approaches to demonstrate causality such as specific diet changes, inhibition of microbial pathways, and fecal microbiota transplant.

1. Introduction

We are beginning to appreciate the role of commensal microbiota in cardiovascular disease (CVD) risk^1,2 and these microbiota are located in a variety of niches within the body including the skin, oral, and gut, and are composed of bacteria, viruses, and fungi. There has been a particular focus on the gut as it is heavily colonized with microbes. It is estimated that more than 70% of all the microbes in the human body are present in the colon alone³ and trillions of commensal microbes including bacteria, bacteriophage, fungi, archaea, and unicellular eukaryotes live in the large intestine that forms a highly diverse dynamic interdependent complex ecological community known as the intestinal microbiota.⁴ The healthy gut microbiota depends on the host for their energy need and provides a range of health benefits by providing nutrients, improved barrier function, shaping host immune system, and preventing diseases including CVD.⁵

The gut microbiota is not the only commensal community studied in relation to CVD risk. There is also evidence that the oral microbiota, which contains more than 10 000 bacterial species from 22 phyla⁶ is also associated with CVD. Data from large epidemiological studies associated poor oral hygiene with increased risk of CVD,⁷ and several oral bacteria including Porphyromonas gingivalis have been detected in atherosclerotic plaques.^8,9 What is not clear is if the association between oral microbiota and CVD indicates a specific mechanism or metabolic profile leading to increased disease or is a marker of other important factors to consider such as access to healthcare. Additional work remains to determine the mechanism(s) underlying the oral microbe–CVD associations and thus the majority of the review focuses on the gut microbiota and CVD.

The combined complexity of the microbiota is only beginning to be appreciated. Maternal transmission,¹⁰ diet,¹¹ host genetics,¹² aging,¹³ and sex¹⁴ all impact the composition of the microbiota. Recent large-scale sequencing efforts have identified tremendous interindividual variation in the microbiota which totals in excess of 45 million genes in the oral and gut niches. Moreover, nearly 50% of all genes were ‘unique to a single metagenomic sample’¹⁵ and only ∼550 000, less than 2%, overlapped between the oral and gut niches. This heterogeneity combined with the possibility of copy number variants within individual bacterial species highlight just how much of the microbiota remains to be characterized.

In this review, we focus on the association between microbiota and CVD, specifically coronary artery disease (CAD) and hypertension. We examine the steps to establishing causality and then discuss the underlying mechanisms by which gut microbiota may affect CVD. Finally, we examine potential therapeutics for CVD using gut microbiota and microbial metabolites.

2. Microorganisms and CVD risk

For over 50 years, the potential role of infectious microorganisms, including bacteria and viruses, as potential risk factors for CVD has been appreciated through a number of epidemiological studies.^16–18 Infection and the subsequent inflammatory processes are thought to induce the onset, progression, and rupture of atherosclerotic plaques.¹⁹ The mechanisms by which infection aggravates atherosclerosis is direct cell invasion that accelerates plaque growth through local effects, or indirect systemic production of inflammatory cytokines that promote development of atherosclerosis.²⁰ Representative infectious agents associated with atherosclerosis include Chlamydia pneumonia, Helicobacter pylori, P. gingivalis, and Influenza A virus.²¹

Although there are a number of pathogenic bacteria associated with CVD, C. pneumonia, a Gram-negative and intracellular bacteria, is an exemplar. C. pneumonia was the first proposed bacteria of CVD etiology responsible for the induction of inflammation in the vascular wall of CVD patients.²² In addition, antibiotics targeting C. pneumonia are thought to have anti-inflammatory effects, and may contribute to atherosclerotic plaque stability.²¹ For example, C. pneumoniae infection in rabbits accelerated the thickness of the tunica intimal wall and atherosclerosis, and treatment with antibiotics reduced the extent of atherosclerosis.²³ This effect is not universal as studies in mice have not shown a similar reduction in lesion size by antibiotics after infection with C. pneumoniae.²⁴ Nevertheless, there are examples of pathogenic bacteria potentially having direct effects on CVD. We now discuss how commensal bacteria of the microbiota may also affect CVD risk.

3. Microbiota and CVD associations

CVD includes a number of pathologies and diseases such as heart failure, stroke, peripheral artery disease, aortic valve disease, atherosclerosis, and hypertension but we limit our discussion to CAD and hypertension. In a number of case–control studies, distinct shifts of microbial composition have been identified in fecal samples from patients with CVD (Table 1). It is not entirely clear whether these microbial communities are causal to CVD or if they are simply affected by other environmental factors contributing to the development of CVD. In the following section, we provide a summary of a number of recently reported associations between gut microbiota and CAD or hypertension.

Table 1.

Reported alterations in the gut microbiota in various CVD cohorts

Condition	Ethnicity	Technique	Associated microbiota changes	References
Myocardial infarction	European (Swedish)	Metagenomic sequencing	Increased: Collinsella	36
			Decreased: Eubacterium, Roseburia
CAD/T2D	European (Spanish)	16S rRNA V2-V3 region	Increased: Enterobacteriaceae, Streptococcus, and Desulfovibrio	30
			Decreased: Faecalibacterium prausnitzii and Bacteroides fragilis
CVD	European (American)	16S rRNA V4 region	Increased: Prevotella and Tyzzerella	27
			Decreased: Alloprevotella, Catenibacterium
CVD	Asian (Indian)	16S rRNA & Metagenomic sequencing	Increased: Proteobacteria, Actinobacteria, Propionibacterium phages, Pseudomonas phages, Rhizobium phages, Lymphocystis virus, and Torque Teno viruses	28
CAD	Asian (Japanese)	16S rRNA V3-V4 region	Increased: Firmicutes/bateriodetes ratio and Lactobacillales	26
			Decreased: Bacteroides and Prevotella
CAD	Asian (Chinese)	Metagenomic sequencing	Increased: Streptococcus sp. M334 and M143, and Clostridium sp. HGF2	33
CVD	Asian (Chinese)	Metagenomic sequencing	Increased: Escherichia coli, Klebsiella spp, Enterobacter aerogenes, Streptococcus sp, Lactobacillus salivarius, Solobacterium moorei, Atropobium parvulum, Ruminococcus gnavus, and Eggerthella lenta	31
			Decreased: Roseburia intestinalis, Faecalibacterium cf. prausnitzii, Bacteriodes spp, Prevotella copri, and Alistipes shahii
Hypertension	Asian (Chinese)	Metagenomic sequencing	Increased: Prevotella, Klebsiella, Porphyromonas, and Actinomyces	42
			Decreased: Faecalibacterium, Oscillibacter, Roseburia, Bifidobacterium, Coprococcus, and Butyrivibrio
Hypertension	Asian (Chinese)	Metagenomic sequencing	Increased: Klebsiella, Clostridium, Streptococcus, Parabacteroides, Eggerthella, and Salmonella	32
			Decreased: Faecalibacterium, Roseburia, and Synergistetes
Hypertension	European (Finns)	Metagenomic sequencing	Increased: Blautia, Cellulomonas, Collinsella, and Desulfovibrio, Dielma, Eisenbergiella, Holdemania, Megasphaera, Phascolarctobacterium, Ruthenibacterium, Sutterella, and Turicibacter	45
			Decreased: Lactobacillusa, and Citrobacter, Coprobacillus
Hypertension	European/African (American)	16S rRNA V3-V4 region	Increased: Anaerovorax, Clostridium IV, Oscillibacter, and Sporobacter	43
			Decreased: Akkermansia, Ruminococcus, Anaerovorax, Sporobacter, and Asaccharobacter

3.1 Gut microbiota and CAD

Dysbiosis refers to an imbalance in the microbial community in the human body. Alterations in the composition of the gut microbiota associated with the risk of CVD have been the focus of the majority of microbiota studies. Indeed, identifying patterns of specific composition of microbiota associated with CVD susceptibility contribute greatly to understanding the pathogenesis of diseases. The composition of the microbiota associated with aortic lesion size is distinct from the oral and gut microbiota.²⁵ Frequently, studies utilize a case–control design to identify the microbial differences among subjects with CVD compared to controls. Foremost, there have been several consistent results reported from a number of cohorts (Table 1).^26–28 For example, Strepcococcus spp., exhibiting pathogenic expansion in intestinal dysbiosis,²⁹ showed increased abundance in stool samples from various CVD patients.^30–33 In addition, the abundance of Faecalibacterium spp. and Roseburia spp., which are known to have anti-inflammatory effects and reduce atherosclerotic events in mice and humans,^34,35 were relatively depleted in CAD patients.^30,31 Finally, metagenome studies have shown depletion of butyrate-producing bacteria including Faecalibacterium prausnitzii and Roseburia intestinalis, and a high abundance of Enterobacteriaceae and Streptococcus in atherosclerotic patients.³¹ Deeper sequencing approaches have also identified alterations in the metagenomic profile of the gut microbiome of CVD patients. These data demonstrate enrichment of genes encoding peptidoglycan biosynthesis, while healthy cohorts are enriched in phytoene dehydrogenase genes.³⁶ Thus, not only are the composition of the microbiota altered in CVD but the underlying functional capacity of the microbiota may be altered.

3.2 Gut microbiota and hypertension

Similar to CAD there is increasing interest in how gut microbiota may affect blood pressure and development of hypertension. Studies of angiotensin II infused mice, spontaneous hypertensive rats, and hypertensive patients have identified a distinct microbial composition of gut microbiota compared to controls including decreased microbial abundance, diversity, and low intestinal epithelial integrity.^37–39 Broad-based antibiotic treatment in mice reduced intestinal dysbiosis and attenuated hypertension in angiotensin II-infused mice.⁴⁰ Furthermore, angiotensin II-induced hypertension is alleviated in germ-free mice compared to conventionally raised mice, suggesting that the gut microbiota is important for the development of hypertension.⁴¹ Studies in humans have clearly demonstrated that there is an inverse association between measures of α-diversity and hypertension.⁶³ In addition to reduced diversity, there is an increased abundance of opportunistic pathogenic taxa such as Klebsiella, and Streptococcus in subjects with hypertension,^32,42,43 while several taxa are associated with high blood pressure (see Table 1) and may affect intestinal cell inflammation.⁴⁴ In a longitudinal cohort, CARDIA, containing both Caucasians and African American subjects, identified a slightly different profile of bacteria positively associated with hypertension: Anaerovorax, Clostridium IV, Oscillibacter, and Sporobacter.⁴³ Conversely, a number of studies have identified Bacteroides thetaiotaomicron as associated with reduced blood pressure in both hypertensive and healthy subjects.^43–46 Although these differences between hypertensive subjects and healthy cohorts do not indicate causality, they do suggest that the composition and diversity of the gut microbiota are associated with clinical features of hypertension. Importantly, having the microbiota assessed in longitudinal cohorts such as CARDIA allows for epidemiologists to assess the risk of developing hypertension (or CAD) based on the current microbiota composition and thereby allowing for the assessment of the clinical utility of these data for predicting disease.

4. Limitations of association-based microbiota studies and CVD risk

A majority of microbiome-based studies utilize analysis of the bacterial 16S rRNA gene. For details on the methods and analytical techniques needed for 16S analysis, we direct the reader to several excellent reviews.^47–50 16S rRNA studies have provided tremendous insight into the specific bacteria associated with CVD and undeniability have transformed our understanding of disease risk but there are limitations. Commonly used 16S rRNA gene sequencing techniques often do not reach species or sub-species level resolution, and the analysis often excludes less abundant microbial taxa. Thus, studies focused primarily on abundance neglect the functional pathways of microbes that contribute to disease risk, although the disease can be driven by microbes that can represent a small fraction of the microbial community. Additionally, many well-described cohorts designed to longitudinally assess CVD risk have not routinely collected fecal samples. Thus, many of the current reports are cross-sectional in nature and therefore precludes assessment of the relative risk of developing CVD on current composition of the microbiota which is critical to developing therapeutic strategies.

We now appreciate some of the biases and intrinsic issues associated with 16S analysis. These include a number of items and several have recently been discussed in the context of hypertension.⁵¹ First, the quality of DNA obtained from samples can be varied by differences in sample collection method (sample type, collection time, or processing method), storage and processing techniques (DNA extraction, library construction, sequencing depth, or quality filters). Second, from a technical point of view, amplification bias, improper or no internal sequencing control, contamination, or non-standard databases for mapping can lead to alteration in microbial composition independent of actual microbial changes. Finally, the results of these analyzes are usually expressed as a proportion rather than an absolute number of the microbes per gram of sample, and the proportion of specific microbial taxa in a sample may not be related to the risk of disease. Numerous large-scale efforts are underway to standardize microbiome methods and protocols in larger cohorts in an attempt to address some of these issues.^52–54 More recently, there is interest in utilizing whole-genome sequencing of microbiota samples termed ‘metagenomics’, to provide much better taxonomic resolution down to species or strain level. As metagenomics results in sequences of genes contained in the sample, it provides an opportunity for functional profiling of the metabolic pathways present in community⁵⁵ while also providing the taxonomic details available in 16S studies. While metagenomics provides a complete view of the microbial communities, it is expensive and the analytical approaches are computationally complex and time-consuming.

5. Towards function and mechanistic understanding of the microbiota in CVD

Although the microbiota is complex, there are a number of approaches to establishing causality and infer mechanism(s) of microbiota and specific bacteria for CVD (Figure 1). In particular, studies utilizing microbial transplantation in model organism can demonstrate the direct role of the gut microbial taxa in the risk of CVD. Specifically, the microbiota can be ablated from model organisms such as mice using antibiotic cocktails or germ-free animals. This provides a naïve state by which individual or complex communities can be added and then examined for hypertension or CVD.

Figure 1.

Towards Function and Mechanistic Understanding of the Microbiota in CVD. Examples of ways to move from association to causation. (A) Examples of microbiome approaches include where conventional mice are provided an agent to perturb the microbiota (either expand specific taxa or inhibit a pathway). (B) Examples of forward microbiome approaches include a comparison of the gut microbiome between controls and CVD subjects from human cohorts. These fecal microbiota samples are then transplanted into germ-free or antibiotic-treated animals to determine the gut microbiota community effect on BP. (C) Diet, specific bacterial enzyme inhibitor, or engineered microbiota can be introduced to animal model to establish a causal relation.

For example, studies utilizing germ-free animals have been critical to further our understanding of the functional role of the microbiota in disease susceptibility.⁵⁶ Germ-free mice can also receive fecal microbiota transfers from samples collected in case/control studies. For example, blood pressure of germ-free mice increased when they received a transplant of stool from hypertensive patients as compared to mice receiving samples from healthy donors. Blood pressure elevations after FMT have also been observed in conventional mouse models.⁵⁷ Germ-free mice administered stool samples from hypertensive patients or hypertensive rats have an elevation of blood pressure.^42,58 These studies provide evidence that gut microbiota can affect the development of hypertension. Studies have been performed using two strains of mice that differ in atherosclerosis susceptibility and report that a portion of atherosclerosis can be attributed to the microbiota.⁵⁹ It is important to note that the complexity of the microbiota can produce unexpected results such as increased blood pressure in Dahl salt-sensitive rats receiving fecal transplants from salt-resistant normotensive rats.⁶⁰ The latter indicates that the relationship between microbiota and host biology is more complex than simply classifying microbiota as ‘good’ or ‘bad’.

In addition to studies examining the transplantation of fecal samples, there have been attempts at understanding the role of specific bacteria in CVD. For example, colonization of germ-free ApoE-deficient (ApoE^−/−) mice with Roseburia intestinalis isolated from humans reduced the levels of inflammatory markers and atherosclerosis providing evidence of the causative role of Roseburia intestinalis in CVD.³⁵ As the literature expands with similar studies, novel pathways and therapeutic targets may be identified. In the following section, we provide an overview of studies utilizing these functional approaches to provide potential mechanisms for how the microbiota affects CVD risk. In particular, we focus on host:microbe effects on the immune system, gut permeability and microbial derived metabolites. It is important to note that by its nature the gut microbiota is a diverse community and thus it is likely that there are multiple mechanisms for its effects on CVD.

5.1 Effect of gut microbiota on cardiovascular health via modulation of immune function

Gut microbiota is a strong modulator of host immunity and host immune response plays a key role in a wide range of pathology including CVD. For example, atherosclerosis which underlies many forms of CAD is considered a chronic inflammatory disease with the involvement of both innate and adaptive immunity.⁶¹ Several gut bacteria have been reported to influence distinct immune cells and there is an indication that T-cells are important in these processes.^62–65 How these systemic alterations in adaptive immunity impact specific microbiota and CAD or hypertension remains to be fully elucidated.

In mice, there is growing evidence suggesting that the gut microbiota can affect atherosclerosis and inflammatory pathways. For example, microbiota transfer from mice prone to inflammatory dysregulation (Caspase 1^−/−) into atherosclerosis prone low-density lipoprotein receptor deficient (LDLR^−/−) mice does in fact increase atherosclerotic plaque formation compared to those who received microbiota from LDLR^−/− mice.⁶⁶ One possible mechanism of this effect is that gut microbiota elevated systemic inflammatory cytokines interleukin (IL)-1β, IL-2, and interferon (IFN)-γ. On the other hand, transplants with Bacteroides vulgatus and Bacteroides dorei attenuated atherosclerotic lesions and decreased plasma tumor necrosis factor-α (TNFα) level.⁶⁷ Similarly, Lactobacillus plantarum ATCC 14917 supplementation inhibited atherosclerotic lesion formation by decreasing serum oxidized LDL (OxLDL), TNFα, and IL-1β production in the aorta.⁶⁸ Higher abundance of the Roseburia and Blautia among others was associated with a decreased atherosclerotic lesion in mice and reduced plasma total cholesterol, TNFα, and IL-1β concentration.⁶⁹

Similarly, hypertension is associated with inflammation and a number of alterations in immune system (reviewed in Norlander et al.⁷⁰). In hypertensive humans, decreased relative abundances of butyrate-producing Roseburia and Faecalibacterium was observed along with increased TNFα:IFN-γ ratio, TNFα and IL-6 production in the isolated peripheral blood mononuclear cells compared to normotensive people. Additionally, high-salt diet-induced hypertension is associated with depleted gut Lactobacillus abundance, increased IFN-γ⁺ CD4 T cells and serum IFN-γ level, and decreased TGF-β1⁺ CD4 T-cells and serum TGF-β1 concentration.⁷¹ These findings do suggest that therapies that modulate the gut microbiota may simultaneously affect inflammatory processes and subsequently CVD.

5.2 Gut permeability and intestinal barrier dysfunction

The cells of the intestine serve as a critical barrier to the bacteria in the gut. This function is maintained by tight junctions between epithelial cells, mucus production, and mucosal immunity. When the intestinal barrier is compromised, lipopolysaccharides (LPS) derived from Gram-negative bacteria can enter the host circulation resulting in endotoxemia. Endotoxemia constitutes a strong risk factor of early atherogenesis⁷² and circulating LPS can bind to the host's Toll-like receptor (TLR) resulting in an inflammatory response in the host.⁷³ Patients with CVD have higher levels of endotoxin in the blood compared to normal individuals.^74,75 Translocation of LPS from the intestine is supported by a higher concentration of endotoxin in the hepatic vein compared to blood drawn directly from the ventricle.⁷⁶ There is evidence that specific Bacteria may alter gut permeability and endotoxemia. Administration of live A. muciniphila reduced intestinal permeability, circulating endotoxin and aortic atherosclerosis in ApoE^−/− mice.⁷⁷ A recent proof-of-concept study in humans found that administration of pasteurized A. muciniphila for 3 months reduced circulating LPS in obese patients.⁷⁸

5.3 Microbial metabolites and CVD

Seminal studies by Hazen and colleagues⁷⁹ have demonstrated that gut microbiota may in fact generate metabolites that affect overall health or CVD pathogenesis. We are beginning to appreciate that many gut-derived metabolites can act on organs such as the liver via portal circulation and be metabolized by host enzymes. Microbial metabolites also help us understand the underlying mechanism by which gut bacterial taxa may influence host biology and thus health and disease (Table 2; Figure 2). We will briefly discuss the association between trimethylamine N-Oxide (TMAO) and CVD, and focus on other microbial metabolites such as short-chain fatty acids (SCFAs), tryptophan metabolites, and bile acid metabolites in this review.

Table 2.

Mechanism of microbiota-related metabolites on CVD

Metabolites	Foods rich in metabolites	Bacteria producing metabolites	Mechanism	Study group	References
TMAO	Red meat, eggs, fish, poultry	Anaerococcus, Clostridium, Desulfovibrio, Edwardsiella, Proteus, Providencia, and others161	Cholesterol accumulation	ApoE−/− mice	162
			Foam cell formation	ApoE−/− mice	162
			Platelet hyper-reactivity	Human and germ-free mice	79
			Vascular endothelial dysfunction	Human and germ-free mice	79
			Fibrosis and remodeling	Human and germ-free mice	79
			Activation of PERK-FoxO1 signaling	ob/ob mice, In vitro studies (primary rat hepatocytes and HEK239T cells)	163
SCFAs	Fermented foods (cheese, butter, yoghurt)	Anaerostipes, Blautia, Coprococcus, Eubacterium, Faecalibacterium, Marvinbryantia, Megasphaera, Roseburia, Ruminococcus, and others164,165	Lowering blood pressure by binding SCFA binding G protein-coupled receptor (GPR41, GPR43, GPR109A)	GPR41−/− mice, GPR43/GPR109A−/− mice, and germ-free mice	97 , 99
Tryptophan metabolites
Indole, ILA, IPA, IAA	Red meat, eggs, fish, poultry	Bacteroides, Bifidobacterium, Clostridium, Lactobacillus, Peptostreptococcus, Ruminococcus, Ruminiclostridium, and others105,106	Act on AHR found in intestinal immune cells and thereby alter innate and adaptive immune responses	GF and conventional mouse	110
Indoxyl sulfate			Induces expression of proinflamatory cytokine IL-6	Hypertensive rats	113
			Induces expression of monocyte chemo-attractant protein-1	Cell culture	166
			Caused endothelial cell senescence,	Cell culture	167
			Proliferation and migration of vascular smooth muscle cells	Cell culture	168

TMAO, Trimethylamine N-oxide; SCFA, short-chain fatty acid; IPA, indolepropionic acid; ILA, indolelactic acid; IAA, indoleacetic acid.

Figure 2.

Overview of potential mechanisms of microbiota–host interactions and CVD. 1- Pathogenic bacteria or bacterial components (LPS) can enter the host circulation resulting in endotoxemia and an inflammatory response. 2- local and systemic inflammatory processes can affect CVD development 3- Metabolites such as SCFAs and TMA can either bind host receptors or be further metabolized to pro- or anti- CVD molecules. SCFAs, short-chain fatty acids; GPR, G protein-coupled receptor; TLR, toll-like receptor; TMA, trimethylamine; TMAO, trimethylamine N-oxide; FMOs, flavin-containing monooxygenases; IL, interleukin; IFN, interferon; TGF-B, transforming growth factor beta.

5.3.1 Trimethylamine-oxide

Many studies have been conducted to investigate metabolites associated with the pathogenesis of CVD, and one of the most studied and reviewed metabolites to date is TMAO. We briefly report the association between TMAO and CVD since there are several well written published reviews which provide detailed reviews of TMAO.⁸⁰ Dietary choline and L-carnitine are known to be converted to trimethylamine (TMA) by microbial enzymes (TMA lyase) contained in the genomes of specific gut microbiota. TMA is then absorbed in the intestine, delivered to the liver via portal vein, and then converted to TMAO by hepatic flavin monooxygenase 3 (FMO3).⁸¹ In subsequent studies combining data from more than 4000 subjects who underwent coronary angiography, TMAO elevation is associated with death, MI, and stroke over a 3-year period.^82,83 These prognostic effects also have been evaluated in patients with a history of diabetes,⁸⁴ chronic kidney disease, heart failure, MI, and peripheral arterial disease.⁸⁰

Although the association between TMAO levels and various CAD events has been reproduced in many studies, there are studies where TMAO is not associated with CAD. For example, a recent study showed that TMAO levels were not associated with atherosclerosis in the Framingham Heart Study Offspring cohort (1215 individuals) or supporting animal studies,⁸⁵ or CARDIA.⁸⁶ One small clinical trial examined FMT of low TMAO producing vegan donors into subjects at risk of CVD but failed to observe a significant reduction in TMAO levels.⁸⁷ Understanding what role, if any differences in the microbiota or diet, play in these disparate results remains to be determined.

Evidence supporting the role of TMAO in the development of hypertension is not yet clear. Preclinical studies have shown that experimental hypertensive rats have higher intestinal permeability and portal blood TMA level in the colon tissue,⁸⁸ and TMAO treatment increased plasma aquaporin-2 concentration which elicits greater water reabsorption, and eventually leads to hypertension.⁸⁹ Apart from these animal studies, evidence has been established in humans through a systematic review that high TMAO plasma levels are associated with high blood pressure risk in 8 studies with 11 750 individuals and 6176 hypertensive cases.⁹⁰ However, the studies included in this systematic review recruited most of the participants were from the United States. Further large-scale prospective cohorts are expected to characterize the association, especially the causality in the general population.

5.3.2 Short-chain fatty acids

The main products of microbial enzyme reaction of dietary fiber are SCFAs, such as acetate, butyrate, and propionate. The main butyrate producers in the human colon are Firmicutes (phylum), with Lachnospiraceae (family), and the Ruminococcaceae (family) the two most abundant groups.^91,92 An additional pathway in lactic acid-utilizing bacteria exists where lactate and acetate are converted to butyrate.⁹¹ A number of other phyla produce butyrate and the efficiency of this production may reflect expression of specific genes such as butyryl-CoA transferase, butyryl-CoA dehydrogenase, and butyrate kinase.⁹³ Two species are well characterized with regard to SCFA production: Bifidobacterium species produce acetate and lactate⁹⁴ and Akkermansia muciniphila produces acetate and propionate.^92,95

It is known that the most direct route through which SFCA modulates the risk of CVD is the regulation of blood pressure. Initial clinical intervention study showed that fiber intake reduced blood pressure and that SCFAs are involved in blood pressure control.⁹⁶ SCFA metabolites lower blood pressure by modulating the SCFA binding G protein-coupled receptor (GPR) 41 or olfactory receptor 78 in vitro and in vivo.^44,97,98 Supplementing the diet with SCFAs protects against the development of hypertension and involves the SCFA receptor GPR43/GPR109A, which also regulates the abundance of Treg cells in mice.^99–101 Thus, SCFAs may regulate cardiovascular homeostasis by activating receptors in the cells of the cardiovascular system. Supporting these mechanistic studies are data from a controlled trial that showed that butyrate (600 mg/day) significantly reduced diastolic blood pressure measured after a 10-min rest period in 15 patients with type 2 diabetes.¹⁰² Although SCFAs are among the most frequently published intestinal microbial-derived metabolites, more clinical trials and mechanistic studies are needed to validate the effects of SCFAs on CVD and its risk factors.

5.3.3 Other microbial metabolites

Identification of microbial derived metabolites that affect host physiology is an active area of investigation. There is particular interest in small molecules derived from dietary tryptophan.^103,104 In the gut, tryptophan is catabolized by a variety of pathways associated with several classes of gut bacteria including Lactobacillus, Bacteroides, Bifidobacterium, and Clostridium to produce a number of metabolites such as tryptamine, indole, indolelactic acid (ILA), indolepropionic acid (IPA), indoleacetic acid (IAA), indolealdehyde (IAld), and metabolite 3-methylindole (skatole).^105–108 These tryptophan metabolites have diverse biological effects including enhancing the intestinal epithelial barrier, stimulating gastrointestinal motility, helping the secretion of gut hormones, exerting anti-inflammatory and anti-oxidative effects in the systemic circulation, and modulating gut microbial composition.¹⁰⁹ Some of these effects may be through the aryl hydrocarbon receptor (AHR) and thereby alter innate and adaptive immune responses in a ligand-specific fashion¹¹⁰ which could modulate CVD risk. Not all tryptophan metabolites are associated with beneficial effects on gut health or CVD risk. In particular, indoxyl sulfate is associated with aortic calcification,¹¹¹ increased carotid intima-media thickness,¹¹² increases expression of cytokines in vascular cells.¹¹³ Further understanding if compositional differences in the microbiota or if specific bacteria modulate the metabolism of tryptophan or other nutrients may further shed light on the underlying mechanisms by which the microbiota affect CVD risk.

6. Therapeutic potential for modulating microbiota on CVD risk

The association between altered composition of intestinal microbiota in CVD patients, production of microbial metabolites, and CVD risk mentioned above suggest that gut microbiota may be a significant modulator of CVD, and their relationship has become a potential target for new therapeutics (Figure 3). We review a number of potential strategies to module the microbiota and CVD risk including diet, inhibition of microbial pathways, and fecal microbiota transplants.

Figure 3.

Therapeutic interventions for improving CVD. Current strategies for improving cardiovascular disease by manipulating intestinal microbiota, including bacterial TMA lyase enzyme inhibitors, fecal microbial transplantation, prebiotics, probiotics, and dietary interventions.

6.1 Diet and prebiotics in the CVD–microbiota axis

Diet has been considered to be the most direct driving factor for gut microbiota composition.^114–117 In many cases, diet is a factor modulating CAD, hypertension¹¹⁸ and gut microbiota. Epidemiological, clinical, and experimental studies have demonstrated that diet and nutrition play a central role in the prevention of CVD.¹¹⁹ Based on these data, the American Heart Association has made formal dietary recommendations (https://www.heart.org/en/healthy-living/healthy-eating/eat-smart/nutrition-basics/aha-diet-and-lifestyle-recommendations, last accessed on 02/06/21). Some of these specific recommendations are now known to affect the microbiota.

An example of these are recommendations to reduce salt intake as certain humans and model organisms are sensitive to a high-salt diet.^60,120 We now appreciate that high salt intake influences gut microbial diversity in rodents and humans.^121–123 High salt intake altered microbiota diversity and increased the abundance of Erwinia genus and Corynebacteriaceae family in mice¹²² and depleted the abundance of Lactobacillus murinus in human. Mechanistically, a high salt diet promotes local and systemic tissue inflammation via increases of pro-inflammatory cytokines and increased gut permeability in both human and animal studies.^121,124,125 Some of the effects of dietary salt may be attributed to specific bacteria such as Bacteroides fragilis, which through a variety of intermediate metabolic effects ultimately activates the mineralocorticoid receptor and increases blood pressure.¹²⁶

In addition to specific dietary components that affect CVD risk and the microbiota, there is evidence that dietary patterns also play a role. For example, one of the most studied diets in CVD research is the Mediterranean diet (MeD). Epidemiological studies found that MeD has a cardio-protective effect compared to western diets.^127,128 Several studies^129,130 have demonstrated that MeD elicits beneficial microbiota profiles and microbial metabolite production. For example, one study¹²⁹ found that higher consumption of MeD is associated with increased levels of Prevotella, fiber degrading Firmicutes, fecal SCFAs, and lower urinary TMAO compared to a western diet. Similarly, another study¹³⁰ reported that closer adherence to the Mediterranean dietary pattern and greater consumption of plant-based nutrients such as vegetable proteins and polysaccharides were associated with a lower ratio of Firmicutes: Bacteroidetes and Streptococcus; higher Catenibacterium, Bifidobacterium, and fecal SCFAs. One recent study¹³¹ found that MeD with probiotics containing Bifidobacterium longum and Lactobacillus rhamnosus increased gut microbial diversity and decreased Bacteroidetes-to-Firmicutes ratio along with other health benefits including improved BMI, fasting glucose, and homeostasis, which was not fully observed in MeD alone indicating that at least a part of the health benefit of MeD depends on the gut microbial composition.

Dietary fiber is an important macronutrient in the context of gut microbiota and CVD. Large population-based observational studies¹³² found that a diet containing high dietary fiber was associated with a reduced CVD risk, which was largely mediated via the reduction of LDL-cholesterol. Clinical trials found that ingestion of soluble fiber (2–10 g/day) was associated with a significant 7% LDL-cholesterol reduction¹³³ in a dose-dependent manner.¹³⁴ Several gut bacteria utilize dietary fiber and produce SCFAs which prevent CVD as discussed earlier in this manuscript. In addition to SCFAs production, dietary soluble fibers bind to the bile acids in the gut and block reabsorption of the bile acids, which leads to increased bile acid synthesis in the liver, and ultimately increased LDL clearance from blood.^135,136

Growing evidence showing that prebiotics supplementation reduces CVD, specifically blood pressure and atherosclerosis, by manipulating the gut microbiota, which supports prebiotic interventions to prevent or treat CVD. Prebiotics are the non-digestible food ingredients that beneficially affect the host by selectively stimulating the growth and/or the activity of one or a limited number of bacterial species already resident in the colon.¹³⁷ Inulin, a linear β-2,1 fructosyl-fructose polydisperse carbohydrate material, feeding decreased atherosclerosis in the aortic root of mice and normalized the altered microbial abundance of Bifidobacterium, Lactobacillus, Akkermansia, Allobaculum, and Coprococcus.¹³⁸ These results are supported by another study where inulin supplementation increased Akkermansia and Bifidobacterium abundance, decreased bacterial taxa involved in secondary bile acid metabolism, and reversed endothelial dysfunction.¹³⁹ β-glucan is a glucose polysaccharide that can sequester cholesterol, scavenges reactive oxygen species, and produces SCFAs when digested by gut microbiota.¹⁴⁰ In addition to serving as an energy source for gut bacteria, β-glucans are immunostimulatory through activation of β-glucan receptors, such as dectin-1 or CR3 on the intestinal macrophages.^141,142 Oat β-glucan supplementation increased high-density lipoprotein (HDL)-cholesterol and decreased plasma triglyceride (TG) and atherosclerosis alone with enrichment of the genus Akkermansia in the gut.¹⁴³

6.2 Targeting the microbiota to affect CVD using small molecules

The approach of manipulating intestinal microbial communities and their metabolic pathways has not yet reached clinical practice, but some studies show promising results. Targeting microbial enzymes (TMA lyases) that convert nutrients such as choline or carnitine into TMA regulates TMAO levels and experimental studies have found several chemical compounds that can trigger the modulatory effects on CVD. One example is 3,3-dimethyl-1-butanol (DMB), which is a TMA lyase inhibitor found in natural products such as olive oil, which can decrease plasma TMAO levels without perturbing microbial cell viability in in-vivo mouse models.^144,145 Mice receiving DMB have been shown to have reduced atherosclerotic lesion, attenuated foam cell formation, and alleviated progression of CVD.¹⁴⁵ Chemically synthesized compounds have been used to inhibit TMA lyase in mice and these molecules also showed selective targeting and sustained inhibition of TMA-lyase in the host.¹⁴⁶

In addition to TMAO, there are a number of other microbial pathways whose modulation or inhibition may affect CVD risk. One possible pathway to target is the microbial cholesterol dehydrogenase enzyme (ismA gene), which converts cholesterol to the sterol coprostanol.¹⁴⁷ Coprostanol is not absorbed as efficiently as cholesterol in the gastrointestinal tract and thus this pathway contributes to lowering blood cholesterol levels.¹⁴⁸ Individuals carrying coprostanol-forming microorganisms have significantly lower cholesterol levels in stools and lower plasma total cholesterol, with effects comparable to those attributed to variations in human genes involved in lipid homeostasis. Therefore, altering abundance of the bacteria-containing ismA or increasing its expression could reduce CVD risk by lowering intestinal and serum cholesterol levels.

An alternative approach is to use small molecules, such as cyclic d, l-α-peptides, which modulate growth of specific bacteria. Initial experiments have been promising as these compounds remodel the microbiota of high-fat diet-fed Ldlr^−/− mice to resemble a low-fat diet gut microbiota and inhibited atherosclerosis development.¹⁴⁹ The effects of this modulation are broad and include decreased plasma cholesterol, suppressed pro-inflammatory cytokines such as IL-6, TNFα, and IL-1B, and altered levels of SCFAs and bile acids in the feces and plasma. Identification of similar compounds that affect host phenotype by targeting microbial processes is an exciting and active area of research.

6.3 Fecal microbiota transplantation

While most often used for mechanistic studies, fecal microbiota transplantation (FMT) has also been used as a therapy for patients with Clostridium difficile infection¹⁵⁰ and ulcerative colitis.¹⁵¹ This process includes the collection of stools collected from healthy donors or the recipients themselves (self FMT) prior to administration into the intestine of patients suffering from disease or related dysbiosis. These studies are complex and to date have not been extensively studied with regards to CVD endpoints. While the clinical utility of such approach is under debate for CVD,¹⁵² the therapeutic effect of FMT has been reported effective against insulin resistance and small intestinal permeability.^153,154

7. Concluding remarks and future perspectives

Many studies have confirmed the link between gut microbiota and CVD and we are beginning to understand the underlying mechanisms of these associations. One exciting aspect of this work is related to metabolomics. In some sense, the gut microbiota is an intermediate trait between diet (environmental factor) and CVD risk (clinical trait) that produces metabolites, some of which play an important role in the pathogenesis of CVD. Recent studies have confirmed that specific microbial taxa are associated with CVD, and that various gut-derived metabolites, including TMAO or SCFAs, can promote or attenuate CVD. However, much still remains to be investigated. For example, only bacterial communities are being studied extensively and other members in gut microbiota such as virus, fungal, or archaea are not widely studied and thus their roles in human disease remain underappreciated. Therefore, combining these members in the analysis of the microbiota–CVD association might open us a new door for potential therapeutics.

Perhaps the most important questions remain—how will these exciting results be applied clinically (if at all)? Studies utilizing the LifeLines-DEEP population cohort have identified that candidate SNPs explain 3–7% of the variation in HDL and triglycerides while adding 16S microbial diversity explains another 4–6% of the variation in HDL and triglycerides.¹⁵⁵ Interestingly, when genetic and microbiome data are combined they explain significantly more variation in HDL and triglycerides. Thus, there is an indication that obtaining microbiota data may further characterize patients as we move towards a model of precision medicine. For example, application of machine learning to datasets containing gut microbiota, genetics, and diet together¹⁵⁶ predicted better postprandial plasma TG, glucose, and insulin responses. In addition to direct causal pathways, it is likely that specific bacteria or microbiota components will be casually implicated or simply act as moderating variables for both genetic studies^157,158 and drug therapies.^159,160 Thus, the microbiota remains a critical part of the movement towards personalized medicine. Ultimately utilizing microbiota data that has been vetted to be clinically relevant may in fact refine our risk predictions and therapeutic interventions.

Authors’ contributions

B.J.B. supervised all portion of the review process, interpreted the results, and mentored manuscript writing. M.N.H. and M.K. conducted the literature search, extracting the information and drafting the manuscript. M.N.H. and M.K. also addressed co-authors comments and concerns. B.J.B., M.N.H., and M.K. critically revised the manuscript. B.J.B. had primary responsibility for final content. All authors read and approved the final manuscript.