|Home | About | Journals | Submit | Contact Us | Français|
This review will highlight recent advances functionally linking the gut microbiome with mucosal and systemic immune cell activation potentially underlying autoimmunity.
Dynamic interactions between the gut microbiome and environmental cues (including diet and medicines) shape the effector potential of the microbial organ. Key bacteria and viruses have emerged, that, in defined microenvironments, play a critical role in regulating effector lymphocyte functions. The coordinated interactions between these different microbial kingdoms—including bacteria, helminths, and viruses (termed transkingdom interactions)—play a critical role in shaping immunity. Emerging strategies to identify immunologically-relevant microbes with the potential to regulate immune cell functions both at mucosal sites and systemically will likely define key diagnostic and therapeutic targets.
The microbiome constitutes a critical microbial organ with coordinated interactions that shape host immunity.
The intestinal microbiota has emerged as a microbial organ, shaped by host genotype, developmental needs, and environmental exposures. Pioneering advances in using culture-independent methods to identify the components of the human microbiome by 16S rRNA sequencing composition revealed that two bacterial phyla, Firmicutes and Bacteroidetes, constitute >90% of the mammalian genome. Despite substantial interpersonal variation in the microbiome in healthy individuals [1,2], a core minimal metagenome exists with 3.3 million non-redundant bacterial genes (~150 times the human genome). These genes are 93% conserved at the enzyme level and play a critical role in secondary metabolism of carbohydrates and sugars for energy extraction. To understand the relative contribution of host genotype versus environmental conditions in determining this variation in the human microbiota, adult twin-twin and twin-mother comparisons have been conducted . Notably, the results revealed equivalent differences in dizygotic and monozygotic twin-twin comparisons, both of which were more dissimilar than self-self comparison but more similar than twin-mother comparison, suggesting a key role for environmental cues in shaping the microbiome. Recent studies have revealed how key environmental and microbial exposures functionally shape the mucosal immune system with more pervasive effects on systemic immunity. Mechanistic insights provided by these studies will help identify novel biomarkers and therapeutic strategies for autoimmune diseases.
Dietary and xenobiotic (e.g. foreign to host) exposures are key regulators of the gut microbiome. For example, a high-fat, high-sugar diet stereotypically altered the gut microbiota in outbred mouse strains as well as numerous inbred mouse strains deficient for immune-linked genes . Notably, alterations occurred within several days. Pilot studies in humans have confirmed these findings . In particular, ten healthy individuals on a stable diet were switched to either plant- or animal-based diets ad libitum for five days. Similar to mouse studies, significant changes were seen in the composition and diversity within several days of dietary changes. These changes resulted in changes in production of short chain fatty acids critical for promoting barrier health as well shifts in the enzymes involved in nutrient extraction. Correlative findings from the microbiota of obese individuals revealed stereotypic reductions in the Bacteroidetes/Firmicutes ratio. When transferred to gnotobiotic (germ-free) mice, these microbiota resulted in an obese phenotype with increased caloric energy extraction from food .
Dietary changes associated with geographic and cultural factors produce characteristic differences in the microbial communities. For example, substantial differences between the distinct communities from the Amazonas of Venezuela, rural Malawi and US metropolitan areas exists including a stereotypical Bacteroides/Prevotella “trade-off” in the composition of the intestinal microbiome . Within communities, pronounced differences are seen in infancy (<3 years old) compared with adults, reflecting age-associated changes in vitamin and nutrient extraction requirements—e.g. increased abundance of enzymes involved in de novo tetrahydrofolate (THF) synthesis in breast-fed babies compared to an increased abundance of enzymes involved in THF extraction in adults and formula-fed infants. Illustrating the effect of culturally-driven dietary influences on shaping these differences, routine consumption of non-caloric artificial sweeteners led to key shifts in the microbiome, which, similar to the obese microbiome, can transfer a phenotype of inflammatory metabolic syndrome to germ-free mouse hosts .
This plasticity of the microbiome driven by dietary factors may therefore modify an individual’s risk for systemic inflammatory disease. Seminal work from Eugene Chang’s group illustrated the importance of saturated fat in providing a sulfur-rich environment to support the expansion of the Deltaproteobacteria B. wadsorthia. Coupled with genetic susceptibility for colitis, this expansion of B. wadsorthia resulted in more severe colitis in mouse models . A recent study illustrated the similar potential of commonly-used dietary emulsifiers, carboxymethylcellulose (CMC) and polysorbate 80 (P80), that are detergent-like molecules used as food additives to affect product viscosity. CMC and P80 compromise barrier functions and permeability of intestinal mucosa, leading to microbial population changes that may cause colitis and metabolic syndrome in genetically susceptible animals. . We recently described the expansion of Prevotella copri in patients with new onset rheumatoid arthritis (NORA) . Notably, Prevotella abundance in these NORA patients inversely correlated with genetic susceptibility conferred by the shared epitope HLA-DRB1, further suggesting a potential role for the microbiome in driving an inflammatory phenotype. A Prevotella-predominant microbiome may also confer susceptibility to inflammatory cardiovascular disease in patients with meat-based diets high in L-carnitine. Metabolism of dietary L-carnitine by the intestinal microbiota results in higher serum trimethylamine-N-oxide (TMAO) which correlates with atherosclerosis in humans and accelerates atherosclerosis in mouse models . A further understanding of these potential biomarkers may guide rational therapeutic use of diet in modulating inflammatory disease.
Additionally, the specificity of the microbiome may regulate the efficacy of xenobiotic or therapeutic drug intervention for the treatment of autoimmunity. The conversion of digoxin to inactive derivatives by the Actinobacterium Eggerthella lenta provides a classic example of this effect . Transcriptional profiling and comparative genomics using high-throughput sequencing allowed the recent discovery of a cytochrome-encoding operon (called cardiac glycoside reductase (cgr)), which is inhibited by arginine in E. lenta and regulates digoxin metabolism . Similarly, in NORA, functional analysis of the Prevotella-dominated metagenome reveals a significant decrease in purine metabolic pathways, including tetrahydrofolate reductase, which may have implications for the therapeutic efficacy of methotrexate . Collectively, these recent studies offer insight into the regulation of drug metabolism by the host’s microbial organ.
While plasticity exists in response to dietary alteration, emerging data suggest that a window of opportunity exists in imprinting the early childhood microbiota with critical effects on the metabolic function of the gut(?) microbial organ(?). Recent analysis of malnourished twins in both Bangladesh and Malawi revealed that therapeutic intervention with ready-to-use therapeutic food (RUTF) engendered only transient restoration in metabolic function and the gut microbiome of malnourished children, both of which regressed when RUTF was stopped [16,17]. Gnotobiotic mice colonized with fecal communities from malnourished donors lost significantly more weight than mice colonized with fecal communities from the non-malnourished twin illustrating the potential microbial-dependence of severe malnutrition (Kwashiorkor). In contrast, alterations in the microbiome can support enhanced caloric energy extraction and growth. Indeed, farmers have historically capitalized on the ability to alter long-term metabolic outcomes by feeding low-dose antibiotics to livestock for growth promotion. Recent studies in mice using low-dose penicillin (LDP) delivered from birth (compared to after weaning) revealed a critical window in development for imprinting a durable microbiome with the propensity for enhanced energy extraction and growth . Thus, in combination with environmental triggers, such as high-fat diets typical of western society, early-life exposure may result in durable effects on the microbiome which predispose towards metabolic and inflammatory disease.
In addition to modulating metabolic functions, commensal microbiota within the gut ‘microbial organ’ maintain the delicate balance of immune effectors, which must remain tolerant to innocuous microbial antigen yet poised to protect the host against invasive pathogens. While a thick mucus layer coating the mucosal surface provides a “demilitarized zone” to promote segregation , we have proposed two main immune mechanisms for maintaining this balance: homeostatic inhibition and homeostatic induction afforded by mucosal immune cells  (Figure 1).
Recent evidence revealed the contribution of intestinal mononuclear phagocytes (MNPs) to homeostatic inhibition. Two developmentally distinct classes of CD11b+ CD11c+ MHCII+ MNPs exist in the lamina propria—CD103+ conventional DCs (cDCs) and the CX3CR1+ MNPs. Historically, CX3CR1+ MNPs were thought not to migrate, but recent data from our lab suggest that microbiota actively limit the trafficking of CX3CR1+ MNPs . In the context of antibiotic-induced dysbiosis, CX3CR1+ MNPs can migrate to the mesenteric LN and traffic luminal, non-invasive bacteria, resulting in aberrant immunity to commensals. This pathway may contribute to commensal reactivity seen in IBD as well as epitope spreading resulting in aberrant immunity to benign commensal and/or self-antigens. Detailed analysis of these cell populations in response to commensals during steady state as well as inflammation will likely identify targetable pathways in inflammatory disease.
In addition, normal gut microbiota promote homeostatic inhibition by group 3 innate lymphoid cells (ILC3). ILC3s are major producers of IL-22, a key cytokine that acts on epithelial cells to promote healing both in the steady state and during infection ,. IL-22 also induces antimicrobial peptide production, including the C-type lectins RegIII, by epithelial cells. These antibacterial proteins directly target Gram-positive, but not Gram-negative bacteria, by forming a hexameric membrance-permeabilizing oligomeric pore . Recently, another, albeit indirect, mechanism of ILC3-dependent reguIation of luminal bacteria was discovered: epithelial cell fucosylation. These terminal fucose moieties can be cleaved by bacterial-derived fucosidases and confer a selective survival advantage to bacteria capable of foraging host-derived carbohydrates, such as Bacteroides thetaiotomicron . Thus, ILC3 maintain homeostatic inhibition by wielding the proverbial carrot and stick (fucosylated carbohydrates and RegIII, respectively) .
How are these bacterial signals sensed by ILC3? During inflammation in both mice  and humans , MNPs expand in the lamina propria. Secretion of CXCL16 by the MNPs acts on CXCR6+ ILC3 to promote co-localization . Early-life microbial exposure, particularly in the neonatal period, may help shape barrier immunity by epigenetically imprinting the Cxcl16 locus . We have recently shown that stimulation of MNPs, but not cDCs, induced the production of IL-23, IL-1β, and the TNF superfamily member 15 (TNFSF15), which potently augment IL-22 production by ILC3 . In patients with mild to moderate colitis (both Crohn’s and ulcerative colitis), exposure to the fecal stream increased production of IL-22 by ILC3, implicating the microbiota in this response . Mechanistic understanding of selective and potentially targetable pathways will be critical for therapeutic intervention to promote mucosal healing.
In contrast to innate processes of homeostatic inhibition, commensals are also capable of inducing specific T cell differentiation pathways in the absence of barrier damage, termed homeostatic induction. For example, Clostridium clusters IV and XIVa, isolated from a healthy human donor, induce colonic Foxp3+ regulatory T cells (Tregs) that produce the anti-inflammatory protein IL-10 . Notably, these strains are decreased in patients with IBD  and colonization of germ free mice with these bacteria attenuated disease models of colitis. Mechanistically, Clostridial induction of Tregs requires TGFβ and high luminal concentration of short chain fatty acids—primarily butyrate—that can induce the differentiation of Tregs in vivo and in vitro [33,34]. The therapeutic efficacy of these microbes in regulating mucosal and systemic inflammation in human will be important to evaluate.
Similar to the Clostridia, segmented filamentous bacteria (SFB) promote immunity without inducing overt intestinal inflammation. In particular, SFB penetrate the mucus layer and bind tightly to the epithelial surface of the ileum where they induce CD4+ T helper cells that produce IL-17a, IL-17f, and IL-22 (called Th17 cells) . Colonization of mice with SFB can protect against concurrent infectious colitis . In contrast, SFB colonization can also promote systemic Th17 cell activation that supports inflammatory arthritis in the K/BxN mouse model  and exacerbates experimental autoimmune encephalomyelitis in mice . The dichotomy of Th17 in promoting barrier protection at the mucosal surface and supporting inflammation systemically may underlie the differential effects of anti-IL17 therapy seen clinically in psoriasis compared to inflammatory bowel disease .
The potential effect of luminal microbes on systemic CD4+ T cell responses raises the question of cognate antigen specificity of these cells. While the concept of “molecular mimicry”—cross-activation by sequence similarity in self-antigen—at distal sites of inflammation may be consistent with microbial specificity, an alternate theory is that inflammatory micro-environments regulate the quality of effector lymphocyte differentiation independently of cognate antigen . By cloning the TCRs of Th17 cells from the small intestine of mice colonized with SFB and expressing them in a hybridoma reporter system, we found that most TCRs were indeed responsive to two immunodominant epitopes in SFB. Confirming the cognate antigen specificity of SFB-induced Th17, co-transfer of cells expressing transgenic TCRs specific for SFB or ovalbumin resulted in Th17 polarization of the SFB transgenic TCRs only . Moreover, concomitant exposure of mice to both SFB and Listeria monocytogenes resulted in Th17 polarization of SFB-specific T cells and Th1 polarization of Listeria-specific T cells suggesting specificity of the micro-environment and antigen delivery in guiding T cell polarization. As such, genetic insertion of the immunodominant SFB epitope into Listeria resulted in Th1 polarization of SFB-specific T cells. In contrast, acute infection with Toxoplasma gondii strongly promotes a stereotypic Th1 response and can act in trans to promote non-cognate, long-lived Th1 memory responses to flagellin antigens expressed by the commensal microbiota . These differences may reflect distinct interactions with the mucosal barrier or particular subsets of MNPs and potentially contribute to systemic autoimmunity.
Enteric viruses are not only a frequent causative agent of human GI disease, but active participants in regulating the outcome of commensalism, including host immunity and microbial homeostasis. Recent results using a positive-strand RNA Calicivirus endemic to mouse facilities (called mouse norovirus or MNV) revealed the ability of a single virus to restore the altered intestinal pathology and lymphocyte function associated with either germ-free or antibiotic treated mice . While the dependence of these findings on the host’s innate response to IFNα production reflects a potential key role for co-evolution of eukaryotic viruses with the intestinal immune system of mammals, immune cell activation induced by MNV in the context of genetic susceptibility (for example, mutations in ATG16L associated with IBD) results in aberrant Paneth cell function and increased disease susceptibility.
Similarly reflecting potential co-evolution, recent reports revealed the importance of “transkingdom interactions” between bacteria and viruses, which have developed systems (both direct and indirect) to regulate each other. Seminal work from Skip Virgin’s lab has identified the importance IFN lamba, another type I IFN, which is induced by commensal microbiota  and controls the persistence of MNV . The effect of these signals in the intestine at the steady state or during systemic inflammatory disease will need to be assessed. Reciprocally, viruses likely help shape the bacterial microbiome. Analysis of dsDNA virus-like particles in IBD patient cohorts revealed a significant expansion of Caudovirales bacteriophages, which correlate with significant changes in the bacterial microbiome . Animal models to test the ability of these prokaryotic viruses to regulate the bacterial microbiome in a “predator-prey” fashion during physiology and therapy will be an important area for future investigation.
In addition to bacteria and viruses, helminth colonization of the intestine is frequent, particularly within the developing world. Notable correlations with the ability to control pathogen infection (including M. tuberculosis, HIV, and Plasmodium) have been established, but the mechanisms by which they modulate intestinal and systemic immunity is the subject of ongoing research. Recent work identified a key mechanistic role for helminth-mediated induction (Heligomosomoides polygrus and Schistosoma mansoni) of type 2 immunity (predominantly IL-4) to allow for a permissive environment for viral re-activation . Similar results hold using another helminth, Trichinella, in blunting the CD8+ T cell response to coinfection with a mouse norovirus . Notably, these effects depend on the immunomodulatory function of the helminth-derived molecule Ym1 and STAT6-dependent alternative activation of macrophages, but are independent of the microbiota. Finally, in addition to helminths, fungi and fungal-derived molecules play a critical role in immunomodulation .
Sequencing technology has revolutionized our understanding of microbiota composition, both in health and disease. The next phase of this revolution is to understand the functional elements of the microbial organ and the ability to target key markers diagnostically and therapeutically. From an autoimmune perspective, it will be important to understand which of the luminal microbiota are recognized by the host immune system. Analysis of the microbiome recognized by IgA has revealed an enrichment in key microbes with dominant immunological effects in mice, including SFB discussed above . Sentinel species have been identified from IBD patients  and malnourished individuals  that can dominantly transfer susceptibility to intestinal inflammation in mouse models. Further work is needed to apply these strategies to define the microbes mediating systemic autoimmunity.
The gut microbiome constitutes a critical microbial organ with coordinated interactions that shape host immunity. Expanding evidence suggests that a dynamic interaction between microbes and environmental cues (including diet and medicines) exist within the microbial organ that shape mucosal and systemic immunity. The clinical relevance of a critical developmental window in microbiome development with long-lasting implications for metabolic health will need to be explored in autoimmunity. Key bacteria and viruses provide critical regulation of mucosal and systemic immunity in mouse models, but the identification of immune-relevant bacteria in human autoimmunity will help focus our investigation on the microbial signals driving disease. Emerging research in this field will yield new insights into both diagnostic and therapeutic targets of autoimmunity.
We would like to thank Gretchen Diehl and Jim Castellanos for their critiques.
Financial support and sponsorship
The work was supported by the NIH (DK083256-02), the American Gastroenterological Association (R.S.L.) and the Howard Hughes Medical Institute (D.R.L.).
Conflicts of interest
Randy S. Longman, Jill Roberts Institute for IBD Research, Department of Medicine, Weill Cornell Medical College, Belfer Research Building, 413 E. 69th Street, Room 514, New York, NY 10021.
Dan R. Littman, Howard Hughes Medical Institute, Department of Pathology, New York University School of Medicine, Skirball Institute, 540 First Avenue, Lab 2-17, New York, NY.