Bile acid coordinates microbiota homeostasis and systemic immunometabolism in cardiometabolic diseases

2.1. Homeostatic regulation of BAs axis

BAs are de novo synthesized in the liver and are regulated by a very precising and comprehensive feedback system to maintain homeostasis. The main two BA synthesis enzymes are cholesterol 7 alpha-hydroxylase (CYP7A1) and sterol 27-hydroxylase (CYP27A1), deficient of which disturb BA homeostasis¹⁴. FXR, the first receptor identified for BAs, has been extensively studied for its role in metabolic regulation of lipids, glucose, amino acids, and also the feedback regulation of BAs per se. Hepatic FXR via its effect on transcription of small heterodimer partner (SHP) suppresses the expression of sterol 12α-hydroxylase (CYP8B1) and to a lesser extent, CYP7A1, while intestinal FXR activation results in the production of fibroblast growth factor (FGF) 19 (15 in rodents) that upon binding to hepatic FGFR4/β-Klotho activates c-Jun N-terminal protein kinase/extracellular signal-regulated kinase (JNK/ERK) signaling and suppresses both hepatic CYP7A1 and CYP8B1 expression¹⁵. FXR promotes BAs secretion by upregulating bile salt export pump (BSEP, the principal BA efflux transporter in the liver) and multidrug related protein 2. BAs are reabsorbed from the gut back to the liver via the enterohepatic circulation by several active transporters on ileal enterocytes [apical sodium-dependent bile salt transporter (ASBT) and organic solute transporter alpha and beta (OSTα/β)] and hepatocytes [sodium taurocholate cotransporting polypeptide (NTCP) and organic anion transporting polypeptide (OATP)]. FXR activation inhibit BAs reabsorption by downregulating ASBT and NTCP. FXR can be activated by specific BA metabolites, including chenodeoxycholic acid (CDCA), cholic acid (CA), deoxycholic acid (DCA) and lithocholic acid (LCA), and their glycine and taurine-conjugated metabolites. The potency of various BA species to activate FXR is CDCA > DCA = LCA > CA¹⁶. DCA, LCA, and their derivatives are major components of the recirculating BAs pool¹⁷. Excessive accumulation of BAs in both the liver and the intestine can activate FXR in a feedback mode to repress de novo hepatic synthesis of BAs. Of interest, FXR can also be antagonized by several BA species such as T-βMCA, Gly-MCA and ursodeoxycholic acid (UDCA). Therefore, the co-existence of agonists and antagonists of FXR in both the liver and intestine may constitute a precising regulatory machinery for tightly controlling the homeostasis of BAs.

2.2. Role of BAs axis in orchestrating lipid and glucose metabolism

BAs, as a class of cholesterol metabolites per se, play fundamental roles in regulation of lipid homeostasis, and also more recently are connected with metabolic regulation of glucose and amino acids. BAs synthesis and excretion are major pathways of cholesterol catabolism. As polar derivatives of cholesterol, BAs are classically considered to be detergents in the small intestine, directly incorporating dietary lipids, cholesterol and fat-soluble vitamins into mixed micellar solutions, which was also called lipid emulsification¹⁸. Thus, this amphipathic nature of BAs is pivotal to lipid absorption and systemic energy balance.

The metabolic regulatory effects of BAs in lipids and glucose metabolism and insulin sensitivity have been supposed to be dependent of its receptor FXR and TGR5 in the liver, intestine, adipose tissue and pancreas^19,20. Mechanistically, FXR is essential for the cholesterol catabolism, transport, lipogenesis, and triglyceride (TG) metabolism. Firstly, FXR both takes part in cholesterol transportation and reverses cholesterol transportation (RCT) via acting on high-density lipoprotein (HDL) metabolism²¹. FXR agonist obeticholic acid (OCA) increased fecal cholesterol excretion and macrophage RCT, and targeting hepatic FXR and/or BAs may be useful for boosting RCT and preventing the development of AS²². FXR–SHP pathway plays a key role in lipogenesis and de novo cholesterol synthesis via the interference of sterol regulatory element-binding protein-1c expression²³. FXR takes part in fatty acids (FA) oxidation though peroxisome proliferator-activated receptor alpha (PPARα) and its target genes²⁴. FXR activation by both natural agonist CDCA or taurocholic acid suppresses hepatic Apo CIII expression^25,26. Thus, FXR activity is effective against FAs induced TG accumulation and the prevention of the FA-induced lipotoxicity^27,28. However, another panel of researchers found FXR overexpression in adipose tissue impaired lipid storage capacity, leading to elevated plasma free fatty acids (FFAs) and ectopic fat deposition in the liver and muscles as well as the whole-body insulin resistance (IR)²⁹. Meanwhile, FXR signalling in ileum biopsies of humans positively correlated with body mass index, and high-affinity FXR antagonists can prevent, or reverse, high fat diet (HFD)-induced and genetic obesity, IR and hepatic steatosis in mice^30,31. Furthermore, mouse models of metabolic diseases demonstrated that the inhibition of intestinal FXR signalling by TβMCA reduces obesity, IR and NAFLD by modulation of hepatic and gut bacteria-mediated BA metabolism, and intestinal ceramide synthesis³². The discrepancy of FXR modulation in mediating lipid metabolism and related pathophysiological processes could be due to its different effects in different tissues, and other confounding factors including dietary patterns and therapeutic conditions. In addition to the typical BAs receptor FXR, TGR5, another important BAs receptor, has been also found playing key roles in metabolic regulation of lipids. LCA was identified as the strongest BA species in activating TGR5, followed by DCA, CDCA and CA. Activated TGR5 promotes mitochondrial fission and beige remodeling of white adipose tissue as well as provokes FFAs release through lipolysis, hence inducing β-oxidation and thermogenic activity in adipocytes³³. In rodents, TGR5-cyclic adenosine monophosphate (cAMP)-enzyme type 2 iodothyronine deiodinase (D2) signaling pathway in brown adipose tissues is activated by BAs, which prevents obesity and resistance to insulin³⁴. In addition, TGR5 activation contributed to skeletal muscle function as well as skeletal muscle activity in mice³⁵. Similarly, in humans, administration of CDCA promotes expansion of brown adipocytes and increases energy expenditure via a TGR5-D2 dependent mechanism³⁶ (Fig. 2).

BAs activated FXR and TGR5 signaling is also important for glucose metabolism. The downstream target of BA-activated FXR, FGF15/19, is pivotal for glucose/insulin homeostasis. Activated FXR in a SHP dependent manner also enhanced the expression of gluconeogenic genes^37,38. Meanwhile, intestine-restricted activation of FXR induces TGR5 to stimulate glucagon-like peptide 1 (GLP-1) secretion to improve insulin sensitivity and hepatic metabolism³⁹. In diabetes, binding and activating FXR in the liver and intestine as well as TGR5, in enteroendocrine cells and pancreatic β-cells were proven to be useful in establishing glycemic control. Agonism of intestinal FXR activates TGR5/GLP-1 signaling to improve hepatic glucose and insulin sensitivity and increase adipose tissue browning³⁹. FXR positively regulates hepatic glucose production, which integrates the glucagon/cAMP signal and the forkhead box protein O1/forkhead box A2 signal, via post-translational modification and engagement of protein–protein interactions, respectively³⁸. In the skeletal muscle, activated TGR5 induces muscle hypertrophy and further increases glucose utilization by activating glycolytic flux, contributing to ameliorated glucose metabolism dysfunction due to diet-induced obesity and aging⁴⁰. Alternative mechanism by which TGR5 activation may influence glucose metabolism involves protein kinase A (PKA)-dependent stimulation of insulin secretion in both alpha and beta cells^41,42. FXR expressed in beta cell may stimulate insulin secretion via K_ATP channel inhibition and activation of TGR5 modulates pancreatic islet alpha-cells to promote glucose homeostasis⁴¹.

2.3. Mutual regulation between BAs and gut microbiota

The gut is lined by a large barrier surface, which is a major mediator of crosstalk between microbes in the lumen and host cells, including immune cells in the lamina propria. Gut microbiota influences the immune homeostasis by regulating through its composition and its metabolites and are concomitantly shaped by the immune system. Indeed, the gut and its inhabited microbiota have now been widely acknowledged as a central hub in dictating the systemic immunometabolism, via the host-microbiota co-metabolites which can transport and distribute across the whole body.

The bidirectional crosstalk between the host gut system and microbiota has now been believed is mostly mediated by the co-metabolites. BAs, because of their distinctive features, represent a class of important mediator in microbiota–host crosstalk. Gut microbiota play important roles in expanding the chemical diversity, composition, and maintaining the intrinsic homeostasis of BAs. The microbial enzyme bile salt hydrolases (BSH) play a key role in deconjugation of BAs, acting as an essential upstream step of BA metabolism⁴³. BSH activity has already been identified in several microbial genera, presenting in all major bacterial divisions and archaeal species in the human gut including members of Lactobacilli, Bifidobacteria, Clostridium and Bacteroides. Recently, the analysis of BSH distribution patterns also provides a novel insight of a reasonable relationship between BA metabolism and CMD⁴⁴. Apart from deconjugation, BAs undergo epimerization and oxidation of specific hydroxyl groups (C3, C7, C12) by hydroxysteroid dehydrogenases, another microbial catalytic enzyme, leading to the formation of secondary BAs (e.g., DCA and LCA)⁴⁵. More recently, a metabolic pathway for BAs dihydroxylation by the gut microbiota has been identified, and established a set of six enzymes for the conversion of CA to DCA⁴⁶. Gut microbiota perturbations not only influence the metabolic composition of BAs, but may also induce impairment in the ileal absorption via regulating the apical-sodium BA transporter, resulting in decreased expression of FXR and FGF19 and an imbalance of BAs, notably characterized by an increase in colonic primary conjugated BAs⁴⁷. Thus, the levels and composition of BA metabolites are markedly influenced by gut bacteria, which produce enzymes that carry out dehydroxylation and deconjugation of BAs produced in the liver⁴⁸.

Conversely, BAs can also modulate gut microbial communities and functions directly or indirectly. Studies showed that when blocking the bile flow into the gut, microbial overgrew and translocated into the circulating system, which can be reversed with administration of conjugated BAs⁴⁹. BAs and particularly the unconjugated BAs such as DCA and CDCA are direct antibacterial agents, changing or destroying the bacterial cell membrane⁵⁰. The bactericidal activities of BAs are likely to be linked to the hydrophobic properties, which increased the adhesion of BA molecules to the phospholipid of membrane inducing cellular damage⁵¹. Interestingly, variations in BA compositions may have the potential to protect against pathogens such as Clostridium difficile⁵². Moreover, BAs may regulate gut microbiota via the activation of FXR, which protect against bacterial proliferation through triggering expression of antimicrobial agents, such as inducible nitric oxide synthase and interleukin (IL)-18.

The mutual regulation of BAs and gut microbiota plays a critical role in the establishment of immunological memory. For instance, previous infection promoted the generation of taurine derived from BAs, leading to long-term functional remodeling of host microbiota, which, in turn, stimulate subsequent antimicrobial defenses⁵³. Acting as active signaling molecules for multiple types of immune cells contained in intestine mucosa, BAs are also involved in mucosal immune regulation⁵⁴. Dietary lard and primary BA-fed mice were characterized by shifts in dominant gut bacterial communities, including decreased relative abundances of Lachnospiraceae and increased occurrence of Desulfovibrionaceae and the species Clostridium lactatifermentans and Flintibacter butyricus⁵⁵. Importantly, BAs also regulate other gut microbiota-derived metabolites such as short-chain fatty acids (SCFAs) and TMAO, some of which have key roles in immunological efficacy. FXR signaling is associated with TMAO generation; it has been shown that flavin-containing monooxygenase 3 (FMO3) generating proatherogenic TMAO was under the regulation of FXR^56,57. Treated with natural (CA) or synthetic (GSK2324 or GW4064) FXR ligands exhibited induction of FMO3 expression and increased TMAO levels⁵⁸ in wild-type mice but not FXR-deficient mice⁵⁸. Furthermore, they showed that hepatic FMO3 expression is regulated by hepatic FXR^56,58. Of interest, TMAO may also regulate BAs signals via differential ways. Inhibition of gut microbial choline trimethylamine lyase was found to be of benefit to the homeostasis of host cholesterol and BA metabolism⁵⁹. TMAO aggravated liver steatosis by increasing BA synthesis and shifting hepatic BA toward FXR-antagonistic activity in NAFLD^60,61. A diet supplemented with carnitine or TMAO-induced AS was associated with BA dysmetabolism, characterized with the prominent increase of tauromuricholic acid, DCA and CA⁶². SCFAs (mainly butyrate, propionate, and acetate) are end products of bacterial fermentation of dietary fiber and present in high concentrations (20–140 mmol/L per day depending on the diet) in the human cecum and the ascending and transverse colon. Increasing evidence indicate that the homeostasis between SCFAs and BAs is reciprocally connected in the interface of gut microbiota and the host intestinal environment. The butyrate-producing bacteria Clostridium butyricum decreased BA-biotransforming bacteria⁶³, and butyrate supplementation reversed dysregulated BA synthesis including hepatic β-muricholic acid and bacteria-generated DCA, and its associated hepatitis in the FXR^–/– mice⁶⁴. Prebiotic fiber, the source of SCFAs, modulated gut microbiota, such as Bifidobacterium and various phyla of Lactobacillus, as well as the intestinal pH, which led to increased BA degradation^65,66. Reciprocally, FXR activation by BAs and its synthetic agonists decreased L-cell GLP-1 secretion in response to inulin-derived SCFA by reducing FFAR2 expression and signaling⁶⁷. In view that BAs possess strong activities in regulation of gut microbiota and relevant host-microbiota co-metabolism as those for the illustrated cases of TMAO and SCFAs, it is reasonable to predict that BAs may play key roles in shaping the metabolic system of gut microbiota, which warrants future research to explore how BAs is connected with other host and microbiota co-metabolites.

3.1. Perturbed BAs axis in CMD

BAs axis, as a master regulator in metabolism homeostasis and immunological responses, should be tightly controlled. It is thus not a surprise to observe that BAs axis is largely disrupted in conditions of CMD, including obesity, NAFLD, nonalcoholic steatohepatitis (NASH), diabetes, and CVD^47,68, 69, 70, 71. BA pool and the total concentrations of circulating BAs were also found increased in obese patients, and were positively correlated with the body mass index. In comparison with non-diabetic controls, the total concentrations of BAs were found increased in most of the previous studies. The levels of CDCA and CA, and to a lesser extent DCA were positively associated to IR in obese and T2D patients. Novel evidence about the perturbed hyocholic acid (HCA) profile was associated with diabetes, and assessed as the future risk of developing metabolic abnormalities⁷².

The de novo synthesis, catabolism, transport, and reabsorption of BAs involve a very complex system that is constituted by various nuclear receptors, enzymes, transporters, and also the gut microbiota. Therefore, the disrupted homeostasis of BA axis observed in conditions of CMD should be expected to be mechanistically related to this complex system. In NAFLD, the hepatic gene expression pattern and the gut microbiome composition consistently support a pattern of elevated BA production, characterized with increased mRNA expression of CYP7A1 but not SHP and BSEP, and reduced level of serum FGF19 hinting to a repressed negative feedback inhibition of BAs synthesis⁷³. Consistently, the BA synthesizing enzyme CYP7A1 was found significantly up-regulated in obese patients⁷⁴. Notably, the transporters of BAs, such as BSEP and ASBT, were found decreased in NAFLD and NASH, hinting to decreased efflux of BAs and thereby the increased retention of BAs in the liver and circulating system⁷⁵. As a further support, OSTα/β, which facilitates the recirculation of BAs from the gut to the liver, was found upregulated in the patients with NASH, obesity and diabetes, leading to the increased accumulation of circulating BAs.

Notably, conditions of CMD are characterized with chronic inflammation and nuclear receptors including FXR, PPARs, PXR, and VDR were found vulnerable to be disrupted by inflammation. Nuclear factor-κB (NF-κB), that is activated in inflammation, binds directly to the promoter of FXR and inhibit its transcription, promoting the loss of BA transporters, increased biosynthesis of BAs, and ultimately accumulation of BAs in the liver. The downstream signal of NF-κB, myeloid differentiation primary-response gene 88 in the hepatocytes also affect BAs profile and FXR expression⁷⁶. PPARs and mammalian target of rapamycin (mTOR), important modulators in metaflammation involved in conditions of CMD, can also perturbed BAs homeostasis. PPARα modulated numerous genes associated with lipid homeostasis and BA biosynthesis via inhibiting CYP7A1 transcription. Inflammatory stress disrupted PPAR–LXR–CYP7A1/ATP-binding cassette transporter A1-mediated BA synthesis and cholesterol efflux. During the acute phase response, cytokines tumor necrosis factor and IL-1 decreased PPARα/β in liver cells, and thereby disrupting BA homeostasis. In CMD conditions, cholesterol-induced lysosomal stress feed-forward may activate transcriptional factor EB-FGF15/19 regulatory loop, acting via mTOR/ERK signaling and transcriptional factor EB phosphorylation, and thereby controlling hepatic cholesterol and BA homeostasis⁷⁷. Hepatocyte nuclear receptor (HNF)-1α, usually disrupted in CMD condition, is also an essential regulator of BA and plasma cholesterol⁷⁸. Transforming growth factor-beta released by hepatic stellate cells during chronic liver injury plays a critical role in liver inflammation, fibrogenesis, and BAs homeostasis. Prostaglandin E2, an important lipid mediator of inflammation, promotes hepatic BA synthesis by an E prostanoid receptor 3-mediated HNF-4α/CYP7A1 pathway in mice⁷⁹. Hepatic E prostanoid receptor 3 deficiency decreased CYP7A1 expression by elevating PKA-dependent Ser143 phosphorylation of HNF4α. IL-1β inhibits CYP8B1 gene transcription via a mitogen-activated protein kinase/JNK pathway that inhibited HNF4α gene expression and its DNA-binding ability. Glucagon and cAMP inhibit CYP7A1 gene expression in human hepatocytes via PKA phosphorylation of HNF4α, unveiled a discordant regulation of BA synthesis and gluconeogenesis by glucagon in human livers during fasting⁸⁰. FXR transcriptional activity is regulated by the glucagon/PKA and the forkhead box A2 signaling pathways, which act on FXR through phosphorylation and protein–protein interactions, respectively, to increase hepatic glucose synthesis³⁸. Because BA homeostasis involves a very complex regulatory system, previous reports about the dysregulation BA axis in conditions of CMD were sometimes controversial. It is important to analyze and understand BA axis in a context dependent manner, and more intensive studies to delineate the compositions, rather than just the levels, of BAs, together with the multiple signals involved, are warranted for deeper understanding of BA axis in conditions of CMD.

3.2. BAs and relevant metabolites in remodeling immune system in CMD

CMD is characterized with the disruption of metabolic homeostasis and the resultant dysregulated systemic immunometabolism, which may represent the most dominant causes for triggering pathological development of ultimate cardiovascular events. Panels of signaling metabolites, such as BAs, SCFAs, and various lipid species, derived from the host and the microbiota system have been validated for their effects in modulating the immune system. The homeostasis of such metabolites was found largely disturbed in conditions of CMD resulting in chronic metaflammation. Therefore, it is important to delineate how the disturbed metabolites remodel the immune system for exploiting better targeting strategy to the immunometabolism system.

Accumulating evidence support that BAs have very comprehensive effects in both innate and adaptive immune system, and different species may exhibit differential and even converse effects. B and T cells are most common members of adaptive immune system. Various efforts have been made to identify immune-modulating therapies targeting either T or B cells with a potential anti-atherosclerotic impact81, 82, 83. 24-nor-UDCA affected the surface expression level of class II major histocompatibility complex of macrophages and DCs as well as the activation/proliferation of T-lymphocyte⁸⁴. Meanwhile, CDCA, as interferon, activated protein kinase C and PKA, resulting in the induction of class one major histocompatibility complex expression⁸⁵. More recently, it has been found that BAs pool may modulate gut retinoid-related orphan receptor (ROR) ƴ⁺ regulatory Treg cell homeostasis via the BA–VDR axis⁸⁶. 3-oxoLCA inhibited differentiation of Th17 cells by directly binding to the key transcription factor ROR-γt (RORγt), while isoalloLCA increased the differentiation of Treg cells which was enhanced by forkhead box protein 3 and mitoROS⁸⁷. Similarly, 3β-hydroxydeoxycholic acid (isoDCA) increase forkhead box protein 3 induction by acting on DCs to diminish their immunostimulatory properties to potentiate peripheral Treg cell differentiation⁸⁸. Notably, Tregs in the heart acted as a critical factor to protect the heart against myocardial infarction⁸⁹. It is thus of interest to determine whether and how the BAs shaped Tregs are involved in the pathological development of CMD.

In addition to adaptive immunity, innate immune system which initiates sterile inflammation via sensing metabolic disorder and DAMPs play essential roles in the pathological development of CMD. DCA may serve as an initiator to activate macrophages, dose-dependently promoting M1 macrophage polarization and pro-inflammatory cytokines production, at least partially through TLR2 transactivated by M2 muscarinic acetylcholine receptor (M2-mAchR)/Src pathway⁹⁰. Moreover, NF-κB/ERK/JNK signaling downstream of TLR2 was involved in the DCA-induced macrophage polarization⁹⁰, while UDCA alleviated the macrophages-induced inflammation in obese mice⁹¹. Taurochenodeoxycholic acid induced monocyte differentiation toward IL-12 hypo-producing DCs via a TGR5-dependent pathway⁹².

As an integral part of innate immunity, the inflammasome recognize a wide range of DAMPs and then trigger further metabolic disorders indirectly through the secretion of active cytokines⁹³. By far, accumulating evidence has highlighted that NLRP3 inflammasome activated by BAs is a cardinal feature of systemic chronic low-grade inflammation. Lipopolysaccharide (LPS)-induced down-regulation of hepatic FXR and the resultant reverse transport of BAs, leading to increased levels of BAs, such as DCA and CDCA, in circulating system for the activation of NLRP3 inflammasome in the macrophages via activating both signal 1 and signal 2⁹⁴. Accompanying with the increase of accumulating levels, DCA and CDCA may further elicit mitochondrial permeability transition and thereby activating the non-canonical inflammasome caspase-4 via promoting the assembly of a protein complex apoptotic protease activating factor-1 pyroptosome for the execution of facilitated pyroptosis⁹⁵. Meanwhile, BAs-related DAMPs synergize with other DAMPs, such as adenosine triphosphate, participate in the process of NLRP3 inflammasome activation. DCA may trigger NLRP3 inflammasome via the sphingosine-1-phosphate receptor 2 (S1PR2)–cathepsin B pathway, leading to subsequent IL-1β maturation and release^96,97. Controversially, BAs also exhibit anti-inflammatory actions in part by repressing NLRP3 inflammasome. As endogenous regulators, LCA inhibits NLRP3 inflammasome activation through the TGR5 signaling in bone marrow-derived macrophage⁹⁸. To be specific, LCA exert their inhibitory effects by inducing NLRP3 ubiquitination and degradation, which is associated with the Ser291 phosphorylation driven by PKA.

It seems that different BA species might have differential modulating effects in NLRP3 inflammasome. One possible explanation of the discrepancy is that BAs may have dual regulatory effect on inflammasome activation by different mechanisms, which may depend on BA concentration, specific BA species, receptors involved, and the presence or absence of other DAMPs. The BAs and their receptor TGR5 activation blocked NLRP3 inflammasome-dependent inflammasome, inducing LPS-induced systemic inflammation, alum-induced peritoneal inflammation and type2 diabetes related inflammasome through a TGR5–cAMP–PKA axis⁹⁸. In NASH, TGR5 expression was decreased in liver tissues, of which deficiency facilitated M1 macrophage polarization by promoting NLRP3 inflammasome activation and caspase-1 cleavage⁹⁹. In the development of AS, activated TGR5 may inhibit cytokine production through the cAMP–NF-κB signaling pathway and decreases ox-LDL uptake in macrophages and thereby protecting against on atherosclerotic plaque¹⁰⁰. The net effects of BAs in regulation of NLRP3 inflammasome in conditions of CMD remains unclear. However, in view that in conditions of AS and CMD, the levels of BA species such as CDCA and DCA and their taurine conjugates, which are potential DAMPs, are increased in the circulating system and thus may promote activation, rather than repression, of NLRP3 inflammasome and thereby contributing to chronic low-grade inflammation. Future research is warranted to explore whether targeting BAs metabolism and their signals, via modulating inflammasome activation, would be of an alternative approach to the prevention of therapy for CMD (Fig. 3).

As discussed earlier in this review, BAs axis functions as a nexus in integrating the metabolic homeostasis of various metabolites including lipids, amino acids, and glucose via engaging diverse receptors and the bidirectional interactions with the microbiota. Thus, the disruption of BAs axis in conditions of CMD is likely to initiate consequential metabolic change of other signaling metabolites which are also key players in remodeling the immune system. Systemic immunometabolism initiated by these metabolites have been comprehensively reviewer elsewhere, we thus make only an insightful summary of how BAs axis altered and its relevant metabolites are implicated in immunological regulation in CMD. BAs can determine the homeostasis of FFA101, 102, 103, ceramides¹⁰⁴, and TG, all of which have been validated to play key roles in driving pathogenesis of AS, the major causes of cardiovascular events, via triggering chronic and low-grade inflammation. In AS, systemic signal molecules produced with the help of BAs signaling such as cholesterol-containing lipoproteins accumulated in arterial wall and recruited several types of adaptive immune cells. For example, contained in very low-density lipoprotein and low-density lipoprotein (LDL), apo B-100 contribute to vascular inflammation by triggering the response of T and B cells locally, which may influence the vascular inflammation plaques stability¹⁰⁵. CCs are key component in atherosclerotic lesions which is eliminated by macrophages. As well known, CCs play a key role in AS, via serving as DAMPs in activation of NLRP3 inflammasome^106,107. Additionally, ox-LDL and palmitic acid also contribute to the pro-inflammation of macrophages through TLR2/4 mediated NLRP3 inflammasome activation^108,109. It is importantly to note that BAs are de novo synthesized from cholesterol, and via targeting FXR and other nuclear receptors, regulate the homeostasis of lipids. Thus, in conditions of AS and CMD, it is reasonable to expect that BAs axis play detrimental roles in lipids triggered chronic inflammation via regulating the disposition of various lipid species like cholesterol, ox-LDL, and FFAs, all of which may represent DAMPs in triggering chronic inflammation in CMD.

3.3. Signaling pathways of BAs axis in shaping immune responses in CMD

The different profile and species of BAs shape the immune system in a systemic and/or an organ-associated manner by changing the metabolic microenvironment and/or via their receptors, which is expressed in a wide range of immune cells. However, the molecular mechanisms and targets underlying BAs-remodeling immune system are largely unclear. Some effects were found dependent of their typical receptors like FXR and TGR5 but some others not. Moreover, FXR and TGR5 may also possess BAs independent effects in shaping the immune system.

During adipose tissue inflammation, activation of the TGR5–AKT–mTOR complex 1 in macrophage was shown to reduce chemokine expression, hence ameliorating obesity-induced IR and macrophage migration¹¹⁰. In hepatic inflammation status, TGR5 activation suppresses the phosphorylation of NF-κB inhibitor IκBα, the translocation of p65, NF-κB DNA-binding activity, and its transcription activity¹¹¹. In support of this TGR5-dependent role in diminishing inflammatory liver injury, researchers found that macrophage migration, infiltration and pro-inflammatory M1 activation were attenuated and macrophage M2 polarization was facilitated via a TGR5–Cat E-dependent pathway¹¹². TGR5 signaling involves cAMP–PKA-mediated inhibition of NF-κB and directly regulate IL-10 expression by a cAMP/PKA/phosphorylated cAMP response element-binding protein pathway¹¹³. Of interest, HCA was found to promote GLP-1 production and secretion in enteroendocrine cells via simultaneously activating TGR5 while inhibiting FXR¹¹⁴ to improve glucose homeostasis in diabetes. Increasing evidence support that TGR5 may serve as a potential target in developing therapeutics against metaflammation.

The immunological modulating effects of FXR have been extensively studied. However, most of the studies observed from the use of FXR agonists were hard to discriminate the direct immunological effects and the indirect effects from the regulation of metabolic homeostasis. In one hand, upon agonists ligation, FXR may directly repress the transcription of proinflammatory cytokines and thereby the inhibition of inflammatory responses. In the other hand, FXR activation in vivo may restore the metabolic homeostasis of BAs, lipids, and glucose, and also insulin resistance, thereby may constitute a feedforward loop to repress metaflammaiton. In the context of obesity, elevated acetylation of FXR appears to play a causative role in promoting hepatic inflammation and abnormal metabolism, while agonist-activated FXR is SUMOylated and exerts anti-inflammatory response by trans-repressing NF-κB target inflammatory genes¹¹⁵, indicating that the dynamic post-translational modification might represent an important mechanism in determining the differential effects of FXR in different context. Notably, ligand independent non-canonical effects of FXR have been reported to show that FXR is a negative regulator of NLRP3 and NF-κB via a mechanism of protein–protein interaction. Furthermore, because the protein of FXR is liable to be degraded by inflammatory stimulation, the decreased FXR protein levels seen in NASH and other CMD conditions may compromise both the ligands dependent and independent effects of FXR in combating CMD.

VDR, which can be also engaged by BAs, has been reported to have a role of anti-inflammation via regulation of macrophages¹¹⁶, DCs¹¹⁷ or T cells¹¹⁸. A recent study showed that LCA impairs Th1 activation at physiological relevant concentrations via VDR signaling¹¹⁹. Another study indicates that LCA has a significant protective effect on TNF-α-induced injury of intestinal barrier function through the VDR suppressing NF-κB signaling and activating the silent information regulator 1/nuclear factor erythroid2-related factor 2 pathway might be one of the mechanisms underlying the protective effect of LCA¹²⁰. However, there is still a lack of research on the direct effect of LCA–VDR signal on immunometabolism in CMD. Thus, whether physiological or pathological levels of LCA and its derivatives modulate immunometabolism via VDR signal in vivo remains to be determined.

A panel of other receptors CAR, LXRs, PXR, RORγT, and S1PR2 has been shown to be involved in BAs signaling, albeit to a lesser extent of affinity in comparison with their specific ligands. Moreover, it seems that there are still unknown targets of BAs, particularly considering the chemical diversity of BA species, which might be involved in regulation of immunometabolism involved in CMD. It warrants future research to explore the causal link of signaling targets of BAs to the immunometabolic homeostasis in conditions of CMD.

4.1. BA sequestrants and BA derivates

BA sequestrants, including colesevelam, colestimide, cholestyramine, and colestipol, have beneficial effects on lipid metabolism, glycemic control and insulin sensitivity in AS and diabetic rodents^125,126. Cholestyramine is well tolerated and to exert cholesterol-lowering effects as well as to reduce CVD mortality^127,128. A randomized clinical trial demonstrated that together with a fat-modified diet, cholestyramine could improve the course of CVD in men with phenotypic familial hypercholesterolaemia¹²⁹. Recently, an ongoing randomized clinical trial protocol designed to investigate the combined efficacy and safety of Elobixibat and cholestyramine for NAFLD¹³⁰, and a recent meta-analysis of randomized controlled trials unequivocally demonstrated that BA sequestrants also improve glycemic control, all of which are important risk factors for CMD¹³¹. Indeed, numerous evidences demonstrated that specific BAs replacement is helpful for CMD treatment.

UDCA, a critical endogenous agonist for TGR5, is clinically used for treating cholestatic liver disease. Clinical studies indicated that UDCA may promote the M1 to M2 polarization of macrophages in patients post liver IR injury¹¹². In a rodent study of obesity, UDCA is directly involved in the crosstalk of inflammatory and metabolic pathways via the blockade of the NOTCH1 inflammatory response and improvement of macrophage infiltration⁹¹. Other evidence indicates that UDCA exhibits anti-atherogenic activity via inhibiting the disturbed flow- or cholesterol crystals-induced inflammatory responses^132,133. However, a small study that included patients with chronic heart failure failed to find any anti-inflammatory effects of UDCA¹³⁴. By exerting FXR-antagonistic effects, UDCA treatment in NAFLD patients strongly impacts on cholesterol and BA synthesis and induces neutral lipid accumulation in both liver and visceral white adipose tissue¹³⁵. Thus, it remains uncertain about whether UDCA is of benefit to the therapy of CMD and more evidence is warranted to support possible larger clinical trials to validate the potential benefit of UDCA in CMD.

DCA administration activates TGR5 signaling pathway, contributing to decreasing inflammation and ameliorating heart post-infarction¹³⁶. HCA species improve glucose homeostasis through a distinct TGR5 and FXR signaling mechanism¹¹⁴. Hyodeoxycholic acid, a natural secondary BA, improved HDL function and inhibited AS formation by increasing the expression of genes involved in cholesterol efflux, such as ATP-binding cassette transporter A1, ABCG1, and apolipoprotein E¹³⁷. Gly-MCA is a selective high-affinity FXR antagonist and thereby reducing the biosynthesis of intestinal-derived ceramides, representing a candidate for the treatment of metabolic disorder³⁰. Notably, TβMCA in mice and glycoursodeoxycholic acid in humans are primary BAs produced in the liver as FXR antagonists that can be deconjugated by BSH expressed by certain species of gut bacteria^138,139. Supplementation of glycoursodeoxycholic acid could inhibit endoplasmic reticulum stress to treat the diet-induced metabolic disorders, including insulin resistance and hepatic steatosis¹⁴⁰. In wild-type mice, shifted microbial community composition with lowered BSH activity increased intestinal taurine-conjugated BA levels, which resulted in markedly elevated intestinal TβMCA levels and inhibition of FXR signaling for combating obesity and IR¹³⁹. A novel BA analog, A17, ameliorated NASH in high-fat diet-fed hamsters¹⁴¹.

4.2. FXR/TGR5 synthetic agonists

For BAs receptors, FXR and TGR5 are the most two extensively studied receptors and have been exploited as potential targets for developing therapeutics against CMD. The activation of FXR expressed in cardiomyocytes, endothelial cells and vascular smooth muscle cells may elicit beneficial effects against pathogenesis of CMD via repressing endothelin 1 transcription, NF-κB activation, and angiotensin II-induced ERK activation and growth proliferation. Conversely, FXR deficiency might attenuate AS development, most likely resulting from reduced oxLDL-cholesterol uptake by macrophages¹⁴². As a support, a more recent finding indicated that suppression of intestinal FXR leads to improved AS in ApoE^–/– mice fed a high-cholesterol diet¹⁴³. The TGR5-selective agonist INT-777, a widely used as an experimental drug, protected against high glucose-induced cardiomyocyte inflammation by suppressing the expression of proinflammatory cytokines and NF-κB¹⁴⁴. INT-777 may have anti-inflammatory effects in pancreatic acinar cells, shown by reduction of reactive oxygen species production and inhibition of NLRP3 inflammasome activation¹⁴⁵. INT-767, a dual FXR/TGR5 agonist, directly down-regulates the expression of Ly6C on bone marrow-derived monocytes and decreases production of proinflammatory cytokines by macrophages¹⁴⁶. INT-767 also attenuates the pro-inflammatory responses by suppression of NF-κB signaling pathway, hence exerting protective effects against NASH and AS^147,148. A study showed FXR and TGR5 dual deficiency exacerbated the development of AS and the anti-atherogenic effect of INT-767 requires the anti-inflammatory effect but not the lipid-lowering effect through the simultaneous activation of FXR and TGR5¹⁴⁸. TGR5 agonist BAR501 (a selective TGR5 ligand) and BAR502 (a dual FXR and TGR5 ligand) both from BAR Pharmaceuticals attenuated liver fat deposition and fibrosis in mice fed a high-fat diet149, 150, 151. A study showed BAR501 reversed liver and vascular damage, highlighting a potential role for TGR5 agonism in treating the liver and vascular component of NASH¹⁴⁹. In addition, BAR501 also regulated the M1/M2 phenotype of intestinal macrophage from murine colitis¹¹³. RDX8940, a potent, selective, minimally systemic oral TGR5 agonist, improves liver steatosis and insulin sensitivity in a mouse model of NAFLD and does not inhibit gallbladder emptying in mice¹⁵². Klemens and colleges¹⁵³ identified isonicotinamides exemplified by compound 3 as nonsteroidal TGR5 agonists, reduced the production of proinflammatory cytokines and stabilized the alterative macrophage phenotype.

FXR has been extensively validated to be a possible target to developing therapeutics against NASH, for which CVD represents the most dominant cause for the death of patients. Therefore, it is important to test whether FXR agonists, currently developed as possible therapeutics against NASH, can also be evaluated for their potential benefit to CMD. Fexaramine, an intestine-FXR restricted agonist, resulted in decreased HFD-induced metabolic phenotypes in mice^39,154. Mechanistically, fexaramine increased FGF15 synthesis, which is delivered to the liver where it decreases expression of the hepatic BA synthesis enzyme CYP7A1 and increases adipose browning. Fexaramine alters the gut microbiota composition by increasing Bacteroides and Acetatifactor, which are the predominant bacteria that convert CDCA and UDCA to LCA, a secondary BA identified as a TGR5 agonist, resulting in increased secretion of GLP-1 that improves insulin sensitivity and decreases obesity through white adipose browning. Hence, the beneficial metabolic effects of fexaramine are accompanied by activation of both intestinal FXR signaling and TGR5 signaling. These pronounced metabolic improvements suggest tissue-restricted FXR activation as a new approach in the treatment of obesity and metabolic syndrome in animal models¹⁵⁴.

Orally OCA activated FXR and protected mice from LPS-induced inflammatory infiltration¹⁵⁵ and has been shown to have clinical benefits of NASH156, 157, 158, 159, 160, primary biliary cholangitis¹⁶¹, NAFLD and diabetes¹⁶². In 2015 and 2019, two multicenter randomized clinical trial reported that OCA 25 mg significantly improved fibrosis and key components of NASH disease activity among patients with NASH, which showed clinically significant histological improvement with predictable clinical benefit^156,157. Another randomized clinical trial found OCA led to weight loss in up to 44% of patients with NASH, and OCA therapy and weight loss had additive benefits on serum aminotransferases and histology¹⁵⁹. Meanwhile, a double-blind, placebo-controlled, phase 3 trial of OCA reported markedly decreased high-sensitivity CRP level in primary biliary cholangitis patients treated with OCA for 12 months¹⁶¹. In addition, there are also several ongoing trials measured the role of OCA in the patients of NAFLD with raised alanine transaminase (NCT03836937), lipodystrophy (phase 2, NCT02430077), morbidly obesity (phase 1, NCT02532335), NASH with fibrosis (NCT02548351), and compensated cirrhosis cue to NASH (NCT03439254). However, in healthy volunteers and NASH patients, OCA dose-independently decreased HDL-cholesterol and increased LDL-cholesterol^163,164. This LDL-cholesterol increasing phenotype also existed in patients with NASH, which was proven to be safely mitigated with atorvastatin by a randomized phase 2 study¹⁵⁸. And OCA decreased cholesterol solubility in bile by increasing human gallbladder cholesterol saturation and BA hydrophobicity via increasing hepatobiliary levels of FGF19¹⁶⁵. Moreover, pruritus represents a key side effect of OCA, which may explain why FDA had not approved OCA as a NASH therapeutics, in spite of that OCA had reached the clinical endpoint in the phase 3 clinical trial. In addition to OCA, there have been many other FXR agonists with different chemical structures undergoing different stage of clinical trials for the therapy of NASH. Interestingly, in vivo study showed that precise delivery of OCA via nanoapproach triggered natural killer T cells population and thus eliciting anti-tumor effects¹⁶⁶. EDP-305 is a steroidal, non-BA, FXR agonist. In a phase 1 multiple ascending dose study, EDP-305 was shown to activate FXR as demonstrated by increased levels of FGF19 and decreases BA synthesis¹⁶⁷. An ongoing clinical trial has been designed to measure the clinical benefit of EDP-305 in NASH (NCT04378010). Tropifexor (NOVARTIIS), previously known as LJN452, is currently investigated in combination with cenicriviroc in patients with NASH and fibrosis (the TANDEM study)¹⁶⁸. Tropifexor has advanced into phase 2 human clinical trials in patients with NASH and primary biliary cholangitis¹⁶⁹. Results from a stand-alone therapy in patients with NASH, the FLIGHT-FXR Phase 2 study (NCT02855164), has shown efficacy in reducing the liver fat content by 5.4% and 10.7% at 60 and 90 μg doses, respectively¹⁷⁰. Tropifexor administration was associated with a mild increase in LDL and a decrease in HDL and pruritus was reported in 14% and 8% of subjects at the 60-μg and 90-μg doses, respectively, compared to 7% in the placebo group¹⁷⁰. Cilofexor, previously known as GS-9674, is a nonsteroidal FXR agonist, which showed well-tolerated and provided significant reductions in hepatic steatosis, liver biochemistry, and serum BAs in patients with NASH^171,172. In a 12-week randomized clinical trial, MET409, a non-BA FXR agonist with a unique chemical scaffold, lowered liver fat content in patients with NASH and delivered a differentiated pruritus and LDL-C profile at 50 mg, providing the first clinical evidence that the risk–benefit profile of FXR agonists can be enhanced through structural optimization¹⁷³. Px-102 from Phenex, synthetic FXR agonists, induces HDL-mediated transhepatic cholesterol efflux in mice and monkeys and prevent AS in cholesteryl ester transfer protein transgenic LDLR^–/– mice¹⁷⁴, has been advanced to a phase 2a in NAFLD patients from phase 1¹⁷⁵. Notoginsenoside Ft1 acts as a TGR5 agonist but FXR antagonist to alleviate HFD-induced obesity and insulin resistance in mice¹⁷⁶. Accumulating evidence support that FXR is a promising target for developing novel therapeutics against CMD. However, because of the complexity of FXR signaling, future studies are warranted to explore the differential roles of FXR expressed in different organs/tissues, the dynamic features in protein–protein interactions and posttranslational modifications, and the canonical and noncanonical efficacies, to better exploit FXR as a potential target for developing therapeutics against CMD.

4.3. Targeting other signals involved in BAs axis

Besides the extensively exploited receptors FXR and TGR5, BAs have been increasingly found to be engaged with diverse other receptors and signals, which could be also expected and exploited to the development of CMD therapeutics. PPARs are closely associated with BAs axis. PPAR ligands have been extensively validated both preclinically and clinically for the prevention of CVD¹⁷⁷, diabetes and NAFLD¹⁷⁸. Recent researches found the combination of PPAR agonist bezafibrate and UDCA exhibited specific anticholestatic properties¹⁷⁹. Transgenic expression of CYP7A1 in the liver prevented HFD-induced obesity and IR in mice¹⁸⁰. Metformin, which is also relevant to BAs axis regulation, and ASBT inhibitor IMB-1715¹⁸¹, have been tested for their effects in modulating systemic immunometabolism in conditions of CMD¹⁸². The pleiotropic effects of FGF-19, including inhibition of BA synthesis from cholesterol via CYP7A1, and inhibition of insulin-induced hepatic lipogenesis, making FGF-19 agonism an attractive potential therapeutic mechanism¹⁸³. Other receptors including VDR, CAR, LXRs, PXR, RORγT, and S1PR2 can be also engaged by BAs, and their effects in modulating systemic immunometabolism in the process of CMD conditions have been reported, and thus, it is reasonable to expect these BAs engaged targets can be exploited for their potentials in developing therapeutics against CMD.