Small molecules, big effects: microbial metabolites in intestinal immunity

Abstract

The mammalian intestine is host to a vast number of microbial organisms. The immune system must balance tolerance with innate and adaptive defense mechanisms to maintain homeostasis with the microbial community. Interestingly, microbial metabolites have been shown to play a role in shaping the host immune response, thus assisting with adaptations that have significant implications for human health and disease. New investigations have uncovered roles for metabolites in modulating almost every aspect of the immune system. In this minireview, we survey these recent findings, which taken together reveal nuanced interactions that we are just beginning to understand.

INTRODUCTION

The intestinal barrier separates the inside of the body from the vast community of microorganisms and dietary molecules that reside in the gastrointestinal tract. It is the largest interface between the host and the external environment. The intestinal immune system faces the unique challenge of balancing tolerance to commensal microorganisms with inflammatory immune defenses, serving both exclusionary and permissive roles. Whereas microbial pattern recognition by the innate immune system has been firmly established as an efficient means by which the host surveys microbial presence, recent work has uncovered a central role for bacterial metabolites in the fine-tuning of the host immune response. Microbial metabolites are defined as small molecules that are produced, modified, or degraded by the microbiota and are ideal messengers for relaying information to the host regarding microbiome composition, pathogen presence, or environmental challenges.

In this minireview, we turn our attention to recent developments in our understanding of the molecular relationship between microbiota, metabolites, and the host immune system and highlight prominent examples of how metabolites contribute to health and disease. We will focus on investigations examining the role of intestinal metabolites that are generated by the microbiome and how they reconfigure the host immune system. These data collectively elucidate the mechanisms behind these immunological changes and highlight the symbiotic and dynamic nature of the host-microbiome relationship. They also pave the way for potential new therapeutic mechanisms of modulating immune system function by adjusting levels of bioavailable metabolites. We will discuss common themes unifying the currently known microbiota-metabolite-immune interactions by focusing on the specific cellular components of the immune system.

T CELLS

T cells are an essential component of the mammalian adaptive immune system and are responsible for recognizing self and foreign antigens presented by antigen-presenting cells. Upon activation, T cells differentiate into a number of different lineages, each with a different role in modulating the host response to microorganisms. One of the most explored roles for microbial metabolites in dictating host immune responses is the polarization of T cells. It has been well established that butyrate, a microbially produced short-chain fatty acid (SCFA), increases naive T cell differentiation into T regulatory cells (T_regs) in vitro (Fig. 1) (1, 7). Recently, this phenomenon has also been demonstrated in vivo. Sodium butyrate supplementation in drinking water selectively expands T_regs in the intestine as well as the bone marrow (11, 27). The exquisite regulatory sensitivity of this response was further demonstrated by a dose-response assay in which lower doses of butyrate selectively induced T_regs in a transforming growth factor-β (TGFβ)-dependent manner, whereas higher doses induced IFN-γ-producing T_regs or conventional T cells (11). This SCFA was also shown to induce expansion of intestinal and bone marrow-derived T_regs after administration of the commensal Lactobacillus rhamnosus GG in a mouse model, demonstrating an immunological impact remote from the intestinal lumen (17).

Fig. 1.

Modulation of immune cell function by microbial metabolites. Examples of recent discoveries in the field regarding the role of microbial metabolites in the modulation of various cellular components of the immune system. I3A, indole-3-lactic acid; ILC2, innate lymphoid cell 2; SCFAs, short-chain fatty acids; T_reg, regulatory T cell.

In a model of type 1 diabetes, nonobese diabetic (NOD) mice fed diets high in acetate or butyrate show a protected phenotype. High-acetate diets limited the frequency of autoreactive T cells, whereas diets high in butyrate increased the proportion of T_regs, suggesting that these metabolites effect changes in phenotype via different mechanisms. These data support a role for the modulation of autoimmune phenomena by microbiome-associated metabolites (16).

In addition to butyrate, the vitamin A metabolite all-trans retinoic acid (atRA) has also been shown in vitro to promote the differentiation of T_regs and to be relevant in modulating the immune system in animal models of human disease (15, 17, 20). In a mouse model of colorectal adenocarcinoma, atRA supplementation decreased tumor burden in a cytotoxic T cell-dependent manner. Conversely, increased CYP26A1 (catabolizes atRA) and decreased CD8+ T cell populations were correlated with worse prognosis in a tissue library of sporadic human colorectal cancer (CRC) tumors (4). The link between diet and alterations to the microbiome is well established, and similar connections are being found between diet and intestinal metabolites. Upon administration of a high-salt diet, mice showed reduced fecal indoles, specifically indole-3-lactic acid because of the depletion of the tryptophan-metabolizing species Lactobacillus murinus. Strikingly, mice treated with L. murinus were protected from high-salt diet aggravated experimental autoimmune encephalitis and showed a reduced T_H17 T cell population in the lamina propria. Indole-3-lactic acid reduced T_H17 polarization in vitro, suggesting a possible mechanism by which this commensal affects immunological change. Compellingly, a pilot study in humans demonstrated that a high-salt diet led to decreased Lactobacillus abundance and an increased T_H17 T cell population (28).

In addition to playing a role in the differentiation of T cells, recently, microbial metabolites have also been found to play an essential role in CD8 T cell memory responses. When activated T cells were transferred into germ-free mice, these cells failed to expand in comparison with those transferred into specific pathogen-free (SPF) mice. Exogenous administration of SCFAs in the form of a high-fiber diet ameliorated this defect in memory response. Furthermore, treatment of antigen-activated CD8+ T cells with butyrate before transfer also promoted memory responses in vivo, suggesting that this metabolite is necessary for immunological recall (2).

B CELLS/ANTIBODIES

Antibody production by B cells is an essential component of adaptive immunity, controlling infection by promoting opsonization, activating the complement system, and directly neutralizing pathogens via inhibitory binding. Relatively little is currently known about how microbial-derived metabolites affect B cell function and antibody secretion; however, a recent study demonstrated that mice fed low-fiber diets exhibit reduced cecal IgA and reduced serum IgG compared with mice supplemented with dietary fiber (12). SCFA administration or a diet high in fiber resulted in increased numbers of IgA⁺ cells in the intestine. B cell differentiation and antibody production were also enhanced by SCFAs. Increased metabolism with SCFAs was also noted via a mammalian target of rapamycin (mTOR)-dependent mechanism. After being infected with Citrobacter rodentium, mice fed low-fiber diets had a higher pathogenic burden than mice fed with high fiber or low fiber supplemented with SCFAs.

A high-fiber diet has also been shown to lead to specifically increased acetate and butyrate, leading to increased intestinal IgA (25). A corresponding increase in T follicular helper (TFH) and T follicular regulatory (TFR) cells as well as an increase in germinal center response was observed. Further analysis of Peyer’s Patches revealed a greater proportion of IgA-producing plasma cells. To further pinpoint a mechanism, the SCFA acetate has been shown to play a role in mediating intestinal IgA response (29). The mammalian G protein-coupled receptor GPR43 was implicated, as a knockout mouse failed to exhibit an acetate-dependent intestinal IgA response. Dendritic cells from these mice did not promote B cell IgA class switching in the presence of acetate. This pathway was ultimately demonstrated to be dependent on retinoic acid signaling (itself a vitamin A metabolite), as blockade of this pathway abrogated acetate-induced B cell IgA production. Taken together, these data suggest a role for metabolites in promoting increased antibody production.

MACROPHAGES

Macrophages are a critical part of the mammalian host’s immune system and essential for phagocytizing pathogens, presenting antigens to adaptive immune players and producing essential cytokines. According to a historical classification scheme, macrophages can be classically activated and produce cytokines such as interleukin-12 (IL-12), interferon-γ (IFNγ), and tumor necrosis factor-α (TNFα) and are implicated in the clearance of intracellular pathogens. Alternatively, activated macrophages produce IL-10, arginase-1, and TGFβ and are essential for tissue repair and the control of extracellular pathogens (8). In the past few years, a number of studies have investigated the role of SCFAs on alternative activation of macrophages; however, some results are conflicting. One group showed that in vitro, treatment of IL-4-differentiated murine macrophages with butyrate suppressed alternative activation, as determined by arginase 1 (Arg1) and chintinase-3-like-protein (Ym1) expression. This effect was specific to butyrate and not seen with treatment with other SCFAs such as propionate and acetate. Furthermore, upon treatment with lipopolysaccharide (LPS), macrophages treated with butyrate exhibited defects in nitric oxide production (6). However, other groups have seen alternative effects. One study showed that IL-4-differentiated bone marrow-derived macrophages treated with butyrate expressed higher levels of Arg1, Ym1, and Found in inflammatory zone 1 (Fizz1) (10). These differences may be due to considerable variability in metabolite dose, especially as noted above, and butyrate has a dose-dependent effect on T_reg differentiation. Further studies are required to fully elucidate the role of butyrate in alternative macrophage activation.

Microbial metabolites have also been shown to affect macrophage cytokine production in models other than alternative activation. In a model of nonalcoholic fatty liver disease, macrophages activated with palmitic acid and LPS showed decreased TNF-α, IL-1β, and monocyte chemoattractant protein-1 (MCP-1) transcripts when exposed to indole-3-acetate and tryptamine. Whereas TNF-α and IL-1β are cytokines that are often associated with classically activated macrophages, MCP-1 is a chemoattractant responsible for promoting the infiltration of macrophages to sites of inflammation. These authors further showed that treatment of macrophages with indole-3-acetate and tryptamine inhibited migration of bone marrow-derived macrophages (BMDMs) treated with MCP-1 (13).

In addition to affecting macrophage polarization, microbial-derived metabolites can also affect macrophage function. A recent study demonstrated that human monocytes differentiated into macrophages in the presence of butyrate show increased antimicrobial activity compared with those differentiated in the absence of butyrate. This result was reproduced in vivo. Furthermore, macrophages differentiated in the presence of butyrate showed increased levels of antimicrobial peptides as well as LC3-associated immune defense. This increase in antimicrobial activity was found to be dependent on the histone deacetylase (HDAC) inhibitory effects of butyrate (23).

Collectively, these data strongly support a role for bacterial metabolites in altering the effects of macrophages, which could have significant effects on the host’s ability to respond to infection.

INNATE LYMPHOID CELLS

Type 2 innate lymphoid cells (ILC2 cells) promote control of parasitic helminths via type 2 immunity. The microbial metabolite succinate has recently been demonstrated to promote expansion of ILC2s in the small intestine, the mechanism of which is dependent on tuft cells, a unique epithelial cell type thought to be exclusively responsible for intestinal IL-25 production (26). Administration of exogenous succinate in vivo in a mouse model resulted in increased ILC2 proliferation and IL-13 production in an IL-25-dependent manner. IL-25 production was further shown to be dependent on the activation of the tuft cell succinate receptor SUCNR1. Increased IL-13 production from ILC2s resulted in increased tuft cell differentiation, creating a feed-forward loop (18, 22). Type 3 innate lymphoid cells (ILC3s) have also been shown to be responsive to metabolites. Metabolites of glucosinolates, a group of phytochemicals found in cruciferous vegetables, are ligands of the aryl hydrocarbon receptor (AhR). AhR-dependent production of IL-22 by ILC3 cells and γδ T cells was demonstrated to be crucial in the DNA damage response (DDR), a conserved pathway ultimately responsible for the DNA repair or apoptosis of compromised cells. Mice fed diets virtually devoid of glucosinolates exhibited low levels of ILC3-produced IL-22, and the DDR in intestinal epithelial stem cells was impaired. These findings suggest that the innate immune response is at least in part sensitive to luminal metabolites (9).

REGULATION OF HOST INFLAMMATORY RESPONSE

The production of metabolites by commensal bacteria is beneficial not only to the host but to the microbiome in its entirety via regulation of host inflammatory response. Clostridium sporogenes produces indolepropionic acid (IPA) via phenyllactate dehydratase. IPA is a metabolite derived from dietary tryptophan and has previously been shown to regulate intestinal barrier integrity (27a). Germ-free mice colonized with C. sporogenes with impaired phenyllactate activity showed elevated frequencies of circulating inflammatory cells, including monocytes, neutrophils, and effector/memory T cells as well as an increase in circulating C. sporogenes-specific IgG (5).

Other tryptophan-derived metabolites have also been shown to play a role in suppressing host inflammation. Peripheral blood mononuclear cells (PBMCs) from healthy human donors once treated with indoleacrylic acid and stimulated with LPS showed decreased IL-1β and IL-6 secretion as well as the downregulation of many inflammatory genes compared with untreated cells (19).

In a mouse model of human inflammatory bowel disease (IBD), caspase recruitment domain family member 9 (Card 9)-knockout animals demonstrate increased susceptibility to colitis. Under homeostatic conditions, Card 9 promotes IL-22 production. Microbiota from these mice do not metabolize tryptophan into metabolites that act as AhR ligands. If administered commensal strains that metabolize tryptophan or an AhR agonist, colitis is significantly reduced. Overall, this suggests that Card 9-dependent metabolism of tryptophan into AhR ligands may be a mechanism of reducing intestinal inflammation (14).

Interestingly, a prospective analysis of serum tryptophan, IL-22, and stool metabolites in IBD patients revealed a negative correlation between serum tryptophan levels and disease activity. Levels of the tryptophan metabolite quinolinic acid and serum IL-22 positively correlated with active disease, suggesting a high degree of tryptophan metabolism in these patients. It is currently not known whether these findings represent increased metabolism or overall deficiency and where this metabolite contribution fits into the pathogenesis of disease activity (19). Collectively, these studies highlight the important role of microbial-derived metabolites in modulating host inflammatory responses.

NEUROIMMUNE AXIS

Given that many metabolites are present in high concentrations circulating in the periphery, it is of relatively recent interest to establish whether metabolites can modulate immunity through their roles as neurotransmitters. In fact, multiple bacterial strains have been found to produce a variety of neurotransmitters, including dopamine, norepinephrine serotonin, γ-aminobutyric acid (GABA), acetylcholine, and histamine (24). Therefore, by definition, these common neurotransmitters are microbial-derived metabolites.

Although the role of many of these microbial-derived neurotransmitters in host immunity is relatively unknown, a few recent key studies have started to link microbial metabolites and host immune regulation. Recently, it has been shown that dopamine production in the periphery inhibits IL-4 and IFNγ production by invariant natural killer T (iNKT) cells in the liver, resulting in ameliorated autoimmune hepatitis in mice (30). Additionally, treatment of mice with antibiotics was found to reduce levels of tyrosine hydroxylase (an enzyme involved in dopamine biosynthesis) specifically in the small intestine, resulting in exacerbated liver damage in a model of hepatitis. These results strongly suggest that dopamine production by the microbiome is essential for controlling iNKT cell-mediated inflammation in the liver.

In addition to dopamine, histamine has also been found to modulate host inflammation. It has been shown that administration of a histamine-secreting strain of Escherichia coli resulted in ameliorated inflammation in a murine model of ovalbumin respiratory inflammation. Treatment of mice with a histamine-secreting strain of E. coli resulted in reduced pulmonary eosinophilia and reduced airway cytokine production. This effect was found to be dependent on histamine decarboxylase (3).

Additionally, other metabolites not traditionally thought to have neurotransmitter effects have been found to influence neuroinflammation. It has recently been shown that tryptophan metabolites limit central nervous system (CNS) inflammation. Tryptophan metabolites have been shown to act through AhR to increase type I interferon production and thereby limit CNS inflammation in a model of experimental autoimmune encephalitis (EAE) (21). These recent studies highlight the role that microbial metabolites can play in modulating host immunity through neurotransmitter functionality.

CONCLUDING REMARKS

Over the last three years there has been an abundance of compelling evidence elucidating the collective impact of microbiome-derived metabolites on the host immune system (Fig. 1). These contribute substantially to the growing body of literature piecing together the nature of the host-microbiome relationship and its role in human health and disease.

The diversity of investigations, however, means that direct extrapolation may be unwise and that seemingly contradictory results may, in fact, be the result of procedural variation. Use of specific mouse lines, varying metabolite doses, housing conditions, age at the time of experiment, natural variations in normal microbiota all may underlie seemingly incongruous findings. Care must be taken to appreciate these disparities.

Moving forward, it is tempting to imagine targeted therapeutics designed to alter or supplement the metabolome. The impact of vascular, autoimmune, or neurological diseases may be dramatically altered in the long term not by direct manipulation of the microbiome but possibly by molding the metabolome. We eagerly anticipate new data expanding the potential of microbial metabolites in human disease.

GRANTS

M. Levy is a 2019 Searle Scholar and supported by the NIH Director’s New Innovator Award (DP2AG067511), the University of Pennsylvania Institute for Immunology, Center for Molecular Studies in Digestive and Liver Diseases, and the Institute on Aging and Alzheimer’s Disease. L. Glotfelty is supported by grant number 5T32DK007066-45 from the NIH.

DISCLAIMERS

The contents of this publication are solely the responsibility of the authors and do not necessarily represent the official views of the National Institutes of Health.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

L.G.G. and A.C.W. prepared figures; L.G.G., A.C.W., and M.L. drafted manuscript; L.G.G., A.C.W., and M.L. edited and revised manuscript; M.L. approved final version of manuscript.