CD is a form of IBD that can affect any part of the gastrointestinal tract, but predominantly targets the terminal ileum and colon. Ulcerative colitis (UC) is another form of IBD which only targets the colon. Despite the fact that the incidence and prevalence of IBD are increasing worldwide and are appearing earlier in life, the etiology and pathogenesis of CD and UC remain ill-defined (95). Although there is an important underlying genetic component (96), disease development requires an environmental trigger (97). Many studies have provided evidence for the loss of gut microbial tolerance in human IBD (98), including the development of B and T cell responses against gut microbial antigens and autoantigens (99–106).

Whereas CD has been generally associated with increased Th17- and Th1-type responses, UC is primarily associated with Th17- and Th2-type responses (107). In CD patients, Th1-associated transcription factors such as STAT4 and T-bet, and cytokine receptors such as IL-12Rβ2 are highly expressed in the lamina propria of the inflamed gut (108). Likewise, the development of CD-like ileitis in SAMP1/YitFc and TnfΔARE mice has been associated with Th1-driven inflammation (109, 110). Indeed, Th1-type cells appear to be necessary for gut inflammation since immune cells from Ifng^–/–, Tbx21^−/−, and Stat4^−/− donors cannot transfer intestinal inflammation into immunocompromised hosts (111–113).

There is also strong evidence for the contribution of Th17-type responses to human IBD. Mucosal biopsies from both CD and UC patients contain increased levels of Th17 cell-derived cytokines (114, 115). In addition, human IBD is associated with genetic polymorphisms at loci encoding Th17 pathway components (e.g., IL6ST, JAK2, STAT3, RORC, IL23R, CCR6) (116). In agreement with these observations, Stat3^–/– and Rorc^–/– mice are resistant to experimental colitis (117, 118). However, since Th17-type cells are known to contribute to the maintenance of normal gut microbial homeostasis, it seems likely that the pathogenic Th17-like cells that contribute to IBD are a different subset (119, 120). Th17-type cells, like other Th cell subsets, are plastic, and the associations between Th17-type cells and both IBD and other autoimmune disorders appear to be mediated by Th17-like cells co-expressing IFNγ and IL-17A (121, 122). In mice, pathogenic Th17-type cells express high levels of the IL-23 receptor (123, 124), and the IL23R gene is strongly associated with human IBD (116). Interestingly, a subset of intestinal human memory CCR6⁺CXCR3⁺ T cells co-expressing Th17 and Th1 markers from CD patients express the Multidrug Resistant Mutation MDR1, a plasma membrane drug efflux pump (125) that is encoded in a gene strongly associated with IBD (ABCB1) (126).

Although the role of Th2-type cells in the pathogenesis of IBD remains unclear, studies in both humans and mice support their involvement. For example, biopsies of UC patients contain increased levels of IL-4 (127) and sera from both UC patients and mice with oxazolone-induced colitis contain elevated levels of IgE, an IL-4-regulated immunoglobulin isotype (128, 129). Likewise, development of ileitis in SAMP1/YitFc mice, and colitis in the TNBS-induced model are associated with Th2-type responses (130, 131).

The onset of IBD has been linked to both microbial dysbiosis and disruption of gut epithelial permeability (132). Decreased abundance on Firmicutes bacteria belonging to the Ruminococcaceae spp., Lachnospiraceae spp, F. prausnitzii and Roseburia spp. families is a signature of microbial dysbiosis in IBD (62, 133). Since these bacteria are butyrate producers, this association may be driven by altered (reduced) induction of iTreg cells in the gut (134, 135). In agreement with this, polymorphisms in genes coding for the immunoregulatory cytokine IL-10 (IL10) or subunits of the IL-10 receptor (IL10RA, IL10RB) are strongly associated with human IBD, particularly with early onset forms of colitis (136). Increased prevalence of pro-inflammatory commensals is yet another mechanism through which dysbiosis may contribute to the pathogenesis of IBD. For example, gut inflammation in patients and murine models has been associated with increased prevalence of R. gnavus (67, 68), which triggers the production of inflammatory cytokines (e.g. TNFα) by DCs via polysaccharide signaling (66). The role of other commensals is less clear. For example, although B. fragilis metabolites promote barrier integrity, the prevalence of this bacteria has been associated with disease exacerbation in CD patients (69), suggesting context-dependent effects.

Increased antibody responses against various gut microbial antigens have been described in IBD. For example, CD patients have increased serum titers of antibodies against the E. coli membrane porin C (OmpC), yeast S. cerevisiae mannose epitopes (ASCA) and bacterial flagellins (CBir) (105, 106), and both the presence and titers of these antibodies are associated with disease severity (63). CD4+ T cell responses against some of these gut microbial antigens have also been associated with IBD. For example, activated OmpC-specific CD4+ T cells are colitogenic in mice (64), and OmpC-specific CD4+ T-cells have been detected in the peripheral blood of IBD patients (65). There is also evidence for a role of flagellin-specific CD4+ T cells; increased frequencies of flagellin-specific CD4+ T cells with an activated, gut homing phenotype were detected in CD and UC patients versus controls. Furthermore, CD, albeit not UC, is associated with increased serum levels of anti-flagellin IgG and IgA antibodies (137).

Many studies have provided evidence for the contribution of an autoimmune component in the maintenance of chronic intestinal inflammation. Most UC, and to a lesser extent CD patients develop peri-nuclear anti-neutrophil cytoplasmic antibodies (pANCA) (99, 138). These antibodies cross-react with the OmpC protein, suggesting a possible role for B cell autoreactivity and gut microbial antigenic cross-reactivity in the pathogenesis of IBD (139). Likewise, IBD has been associated with autoantibody responses against Glycoprotein 2 (GP2) (100, 101), a receptor for bacterial adhesin FimH that is upregulated in the gut epithelium of patients (100). Of note, high levels of anti-GP2 IgA antibodies have been described in pediatric IBD patients (101). An increased prevalence of autoantibodies against FAM84A, a neuronal sensory protein expressed in the gastrointestinal tract, has also been associated with IBD (102).

The contribution of autoreactive T-cell responses to IBD is much less clear. Some studies have reported the ability of commensal bacteria to activate colitogenic T-cells or autoreactive T cells. A Citrobacter infection in ovalbumin specific TCR transgenic mice triggered the apoptotic cell death of infected colonic epithelial cells, promoting intestinal inflammation via the activation of autoreactive Th17 CD4+ T cells (45). Importantly, there is also evidence for protective autoreactive T cell responses in IBD. Specifically, a Bacteroides integrase epitope was shown to induce the recruitment of a highly prevalent low avidity IGRP_206-214 specific CD8+ T cell subset to the gut, affording the mice protection against experimental colitis (53).

The above observations suggest that the relative contribution of autoimmune vs. non-autoimmune phenomena to UC and CD is different. Thus, whereas autoreactivity against colonic epithelial cells may play a role in UC, immune reactivity against the intestinal flora is primarily a feature of CD. Multiple environmental factors, genetic determinants as well as the specific contribution of commensal bacteria to dysbiosis could bias the inflammatory response and the disease phenotype in each of these two inflammatory bowel diseases.

MS is a CNS-specific autoimmune disease that is largely driven by Th17-type cells and is characterized by CNS inflammation, demyelination, and progressive neurodegeneration (140). Most patients suffer a relapsing-remitting form of disease (RR-MS). Although the etiology of MS is complex and incompletely understood, both genetic and environmental factors clearly play a role (141, 142).

Several studies have provided evidence for associations between dysbiosis and MS, such as increases in the prevalence of Methanobrevibacter (Archaea) and Akkermansia (143) or firmicutes (70), as well as a reduction in the prevalence of Butyricimonas (143).

In experimental autoimmune encephalomyelitis (EAE), a mouse model of MS, induction of autoreactive B cell responses against myelin oligodendrocyte glycoprotein (MOG) requires the presence of the microbiota (144). In addition, induction of EAE in germ-free mice was associated with reduced levels of IL17 and IFNγ in both the intestine and spinal cord as well as increased levels of Treg cells. Interestingly, colonization of these germ-free mice with SFB restored EAE susceptibility, implicating the microbiota on the development of encephalitogenic Th17 responses (71). More recently, two different bacteria from the Erysipelotrichaceae family and L. reuteri have been associated with the severity of EAE via effects on Th17 cells and MOG molecular mimicry, respectively (58). Another study reported defective production of IL-10 by Treg cells from mice colonized with fecal samples from MS patients, suggesting that MS patients harbor a specific repertoire of commensals that favor CNS autoimmunity (48).

Cross-reactive T cell responses against gut microbial antigens have also been described in MS. In a recent study, cerebrospinal fluid (CSF)-infiltrating T cells specific for GDP-L-fucose synthase cross-reacted with Akkermansia antigens (79). Furthermore, GDP-L specific clones recognized the myelin basic protein (MBP) epitope MBP_83-99, suggesting that T cell cross-reactivity between gut microbial and CNS autoantigens could act as a trigger of CNS inflammation.

Other studies reported protective effects of certain commensal bacteria against CNS autoimmunity. Administration of B. fragilis PSA has been shown to protect mice against CNS autoimmune inflammation by promoting the expansion of Foxp3+ Tregs expressing CD39 (31) and by inducing tolerogenic DCs (32). Likewise, administration of P. histicola resulted in reduced frequencies pro-inflammatory Th1 and Th17 cells and increased frequencies of FoxP3+ Treg cells, tolerogenic DCs and suppressive macrophages (74). Other species, such as Bifidobacterium and Lactobacillus have been shown to protect mice against EAE by promoting Treg cell responses and reducing Th1- and Th17-type responses (75, 76). The E. coli Nissle 1917 strain was also shown to suppress CNS inflammation by promoting the formation of IL-10-producing autoreactive Treg cells (78).

Similar observations have been reported in humans. A study in a small cohort of pediatric MS patients reported a reduced prevalence of Bacteroides (72). In addition, increases in the Bacteroides content of the gut microbiota of RRMS patients with commensal modifying therapies was associated with disease-protective effects (73). Likewise, a reduced prevalence of Prevotella strains has been associated with increased frequencies of Th17 cells and disease activity in MS patients (70), suggesting a potential protective role for these bacteria against CNS inflammation. In addition, gut microbiota from MS patients imprinted defective IL-10 responses in fecal transplanted host mice, promoting the development of spontaneous EAE (48). In a recent human clinical trial, oral delivery of Lactobacillus and Bifidobacterium spp ameliorated MS symptoms (77).

SLE is a systemic (multi-organ) autoimmune disease characterized by development of autoantibody responses against nucleic acids, histones and ribonucleoproteins, leading to the formation and deposition of pathogenic immune complexes in various organs, including the kidney. Th17 polarization and higher frequencies of TFH cells have been described in the peripheral blood of SLE patients (145), consistent with the extensive autoantibody response underlying disease pathogenesis. The etiology of SLE, as is also the case for most other autoimmune diseases, remains unclear. There is an important genetic component that, although necessary, is insufficient for disease development (146). Environmental cues, such as infectious agents, are suspected to play a role as triggers of disease development in individuals at risk.

Recent evidence points to the microbiota as another potential contributing factor to the development of SLE (147, 148). Disruption of the barrier function of the gut, leading to translocation of commensal bacteria and pathobionts is one of the possible mechanisms underlying this association. For example, E. gallinarum was detected in the liver of SLE patients (as well as in the liver of patients with autoimmune hepatitis) but not in the liver of healthy controls (59). In a murine model of SLE, antibiotic treatment reduced mortality and decreased the production of pathogenic autoantibodies and autoreactive T cells, in part by suppressing the growth of E. gallinarum in tissues (59).

The composition of the early gut microbiota in mice also appears to have an impact on the development of anti-nuclear autoantibodies (149). SLE patients develop autoantibodies against the evolutionarily conserved RNA binding protein Ro60 (150, 151). Multiple gut commensals, such as B. thetaiotaomicron, express Ro60 orthologues with high sequence homology to human Ro60. Furthermore, colonization of germ-free mice with B. thetaiotaomicron led to the development of T and B cell reactivity against Ro60, as well as to glomerular immune complex deposition mimicking lupus nephritis (85). The same study reported that bacterial Ro60-specific T cell clones isolated from SLE patients cross-reacted with the human orthologue. Together, these data provided evidence for gut microbial molecular mimicry as a potential contributor to the development of SLE.

T1D is a multifactorial autoimmune disorder characterized by immune-mediated destruction of the pancreatic β-cells, in which numerous genetic elements and putative environmental triggers play a role. Several different alterations of gut microbial health have been associated with T1D in both animal models and humans (152–154), including alterations of intestinal permeability (155, 156), as well as loss of gut microbial diversity before the onset of disease (157). Pro-diabetogenic, oral antibiotic-induced gut dysbiosis in NOD mice has been associated with impaired enteric Th17/Treg responses (158). More recently, R. gnavus has been suggested to protect mice against streptozotocin (STZ)-induced diabetes, as well as to promote the development of anti-diabetogenic CD8+CD122+ Treg cells in T1D patients (83).

Gut microbial molecular mimicry has also been implicated as a possible mechanism of autoreactive T-cell activation in the pathogenesis of T1D. A protein from L. goodfellowii was suggested to function as a structural mimic of the murine diabetogenic IGRP_206‐214 epitope, as it could promote the activation of cognate TCR-transgenic CD8+ T-cells in vitro (159). In another study, however, metagenomic sequencing of the gut microbiota failed to verify the presence of L. goodfellowii in the gut microbiota of both mice and patients (53). Most importantly, the latter study identified the Bacteroides integrase, an abundant gut microbial antigen, as a true structural and functional mimic of IGRP_206‐214 (53). However, experiments in mono-colonized germ-free mice indicated that this gut microbial epitope promotes the recruitment and activation of anti-colitogenic, low avidity IGRP_206-214-specific CD8+ T-cells, rather than the activation of their diabetogenic high-avidity counterparts (53).

Gut microbial metabolites have also been implicated in the immunopathogenesis of T1D. Increased prevalence of Bacteroides species as well as deficiencies in bacteria that produce SCFAs have been described in T1D patients (84, 160). For example, children with T1D-associated autoantibody seropositivity have a reduction in the abundance of the butyrate producer F. prausnitzii (84). Another multicenter study of 783 children showed that the microbiota of healthy children is enriched in SCFA-producers, without obvious associations with any taxa, suggesting that microbial function rather than composition might contribute to T1D development (161). In agreement with these observations, NOD mice fed with diets promoting gut microbial production of acetate and butyrate were almost completely protected from T1D via SCFA-mediated immunomodulation (162).

RA is an organ-specific autoimmune disease that is characterized by chronic inflammation and progressive destruction of the joint tissues by arthritogenic T cells and autoantibodies. Although the pathogenesis of RA remains incompletely defined, both genetic and environmental factors, including alterations in the gut microbiota, have been implicated in its development. As is the case for the other autoimmune diseases discussed above, alterations in intestinal permeability (163) and gut microbial composition (164) have been found to predate the onset of disease in RA. Commensal bacteria such as Collinsella have been associated with increased gut permeability and disease severity in both an experimental model of arthritis and in human RA. In RA patients, for example, pro-arthritogenic IL-17A responses in a subset of RA patients were associated with an increased prevalence of Collinsella (91). In another study, colonization of germ-free mice with SFB bacteria was sufficient to induce arthritogenic Th17 responses (86).

Other studies have suggested a role for molecular mimicry as a trigger of arthritogenic autoimmune responses in both animal models and patients. An early work reported the presence of immunoreactivity against an E. coli epitope, QKRAA, in the synovial fluid of patients as compared to controls (165). A more recent study found that autoreactive CD4+ T cells against the autoantigens Filamin A (FLNA) and N-acetylglucosamine-6-sulfatase (GNS) cross-react with similar sequences found in Prevotella, Butyricimonas and Parabacteroides species (50). Furthermore, increased prevalence of Prevotella species, such as P. copri were detected in patients with new-onset RA (166). In mice, Prevotella has also been proposed to contribute to RA development, in this case by both, activating autoreactive T cells specific for the arthritis-relevant autoantigen Ribosomal Protein L23a (RPL23A), and by inducing pro-inflammatory Th17 responses (90).

The oral microbiota has also been implicated in RA. Periodontitis induced by P. gingivalis, an established oral pathobiont linked to this condition, has been associated with the exacerbation of autoimmune arthritis, presumably by inducing pathogenic Th17 responses via TLR2- and IL-1-signalling (87). Of interest, P. gingivalis has been found to contribute also to the generation of citrullinated proteins (antigenic targets of RA) in the oral cavity of RA patients (89, 167), suggesting a potential link between immune responses against these post-translationally modified oral proteins and downstream joint inflammation (168). This property has also been documented for another RA-associated oral pathobiont, A. actinomycetemcomitans (60).

Psoriasis is a prevalent autoimmune disease characterized by keratinocyte hyperproliferation and skin inflammation, where both genetic and environmental factors also play a role (169). Psoriatic skin lesions are associated with dermal and epidermal infiltration of leukocytes, triggered and maintained by T lymphocytes (170, 171). Most of the T cells that infiltrate the psoriatic dermis are CD4+, whereas those that infiltrate the epidermis are primarily CD8+ (171, 172). Different clinical phenotypes have been associated with the presence of bacterial skin commensals capable of inducing local pro-inflammatory Th17 responses (173). However, multiple studies have also underscored the importance of the gut-skin axis on cutaneous autoimmunity. Gut dysbiosis induced by oral antibiotic treatment in neonatal mice promoted the development of psoriasis by increasing the frequency of cutaneous IL-22 producing γδ+T cells (174). In addition, induction of experimental psoriasis via imiquimod exposure is blunted in germ free or antibiotic treated mice, in association with a reduction in Th17 cells (175, 176). In particular, Helicobacter pylori infection has been associated with psoriasis (94), potentially via both local and systemic effects of the inflammatory response (177, 178), such as increased permeability of the gastric mucosa to food antigens, among others (92). In addition, the H. pylori enterotoxin binds to the T cell receptor and induces the expression of T cell skin homing receptors (179, 180).

Vitiligo is another T cell-dependent autoimmune disorder of the skin characterized by skin depigmentation due to immune mediated killing of melanocytes (181). Recently, in a murine model of vitiligo harboring tyrosinase-reactive T cells, oral ampicillin treatment decreased disease severity, suggesting that the gut microbiota may also play a role in this disease (182).