|Home | About | Journals | Submit | Contact Us | Français|
Inflammatory bowel disease is a disease that reflects a disequilibrium in host commensal homeostasis. T-bet-/- × RAG2-/- deficient mice develop a spontaneous juvenile ulcerative colitis resulting from a pro-inflammatory response to the commensal microbiota that is dendritic cell and TNF-α driven. The TRUC (T-bet-/- RAG2-/- ulcerative colitis) model is discussed in the broader context of the adaptive and innate immune mechanisms that regulate host-commensal relationships within the intestine.
Upwards of 4 million people are afflicted by inflammatory bowel disease (IBD), a spectrum of chronic illnesses driven by inflammation that can occur in the small and large intestine. Several recent genome wide association studies of human populations and multiple spontaneous and inducible mouse models have shed light on the molecular pathogenesis of these diseases and as a result are improving the therapeutic armamentarium available to treat them(1, 2). Over 6 billion people populate this world and 99.94% of them are spared from IBD. The relatively low incidence of these diseases is in many ways astonishing considering that the human intestine is home to approximately 100 trillion bacterial organisms. The host immune system has evolved elaborate offensive and defensive strategies to protect itself from bacterial invaders. The intestine presents a different challenge to the immune system—a challenge of peaceful co-existence driven by metabolic needs that confer clear fitness and survival advantages. Bacteria function in efficient energy extraction of nutrients for the host, and bacterial metabolism also generates metabolic products essential for host physiology. Multiple mechanisms regulate host-commensal relations in the intestine. The three major players in this complex balancing act are the microbiota, intestinal epithelial cells, and immune cells. Our studies in T-bet-/- × RAG2-/- mice, which develop a commensal dependent ulcerative colitis driven by TNF-α, have provided us with some insight into the interactions amongst some of these cell sub-types(3).
A key factor responsible for this host-commensal co-existence is the epithelial cell barrier that separates the vast numbers of the luminal microbes from the immune cellular networks embedded across the intestinal wall (reviewed in (4)). This barrier has multiple layers of fortification and is much more than a constantly self-renewing single layer of epithelial cells attached via tight junctions. A mucus layer also functions as an important part of this boundary. Intestinal mucus is arranged in a bilayer comprised of gel-forming mucins that are assembled into a firm inner layer and a looser outer layer (5). The functional importance of this element of the barrier is evidenced by the spontaneous colitis observed in mice deficient in the mucin-2 gene, a critical component of intestinal mucus (6).
The mucus layer is not a simple glycocalyx, embedded within it are a number of antimicrobial molecules. Epithelial cells, Paneth cells, neutrophils, and colonic plasma cells elaborate diverse functional classes of bactericides including: defensins, cathelicidins, and C-type lectins (7). The defensins are divided into three groups: α, β, and θ-- functional θ defensins have not been found in mice or humans. These small cysteine-rich proteins possess a dense cationic charge and share both hydrophilic and lipophilic moieties making them adept pore formers. α vs β designation is based on the position of cysteine linkages, however, the secreting cell type and tissue distribution often differ between these two groups (8). The defensins exert their antimicrobial activity by binding to and embedding themselves within bacteria cell membranes generating numerous small pores that compromise cellular function and lead to bacterial cell death. The cathelicidins are lysosomal polypeptides that share some structural homology to the cathepsin cysteine protease inhibitors. They have a strongly cationic pore-forming antimicrobial C-terminal domain that is activated upon cleavage from the N-terminal cathelin domain, which also has antimicrobial activity (9). C-type lectins and in particular the RegIII family of C-type lectins are an important group of bactericidal proteins that shape commensal microbial communities and protect against intestinal infections from food-borne pathogens(10-12). Thus the mucus layer not only blocks pathogens from the epithelium by its intrinsic gel-like structure but provides a matrix that sequesters numerous proteins and peptides with bactericidal activity that are secreted into the lumen.
TRUC mice have altered barrier function, however, this dysfunction does not result from either overt junctional abnormalities in the epithelium or gross aberrations in functional components of mucus. This loss of barrier is not intestinal epithelial cells intrinsic as neither T-bet or RAG2 are expressed in intestinal epithelial cells. T-bet is expressed in T cells, B cells, NK cells, NKT cells, dendritic cells, and the developing mouse vomeronasal organ. Notably, T-bet is not expressed in macrophages (Glimcher unpublished results) or mast cells (13). High levels of mucosal TNF-α elaborated from lamina propria dendritic cells appear to drive epithelial cell death leading to holes in the epithelium and a leaky barrier. We have shown that neutralization of TNF-α with anti-TNF-α antibodies reversed the epithelial death and subsequent holes/ulcers observed. In addition, we have generated TRUC mice deficient in TNFR1 and these mice are protected from colitis as well. However, we have not determined yet whether expression of TNFR1 is necessary on hematopoietic or non-hematopoietic derived lineages or both for the development of disease. Ex vivo cultured TRUC epithelial cell populations have higher levels of cell death in response to TNF-α treatment as compared to RAG2-/- epithelial cells. Both a higher percentage and higher levels of TNF-R expression have been observed in TRUC epithelial cells as compared to RAG2-/- (Garrett and Glimcher unpublished observation). Dying and dead intestinal epithelial cells could understandably result in a state of local and adjacent mucodepletion further exacerbating a barrier breach. In aggregate, our experiments suggest that barrier dysfunction in TRUC is primarily linked to a dysregulation of the cytokine TNF-α.
Another element of the barrier that mediates homeostatsis between the host and the commensals and protects from invasive organisms is secretory IgA (SIgA). Secreted at 3-5g/L/day into the intestinal lumen of humans, SIgA exerts a massive force on the commensal microbiota and by abundance is the dominant antibody isotype in humans (14). SIgAs are versatile molecules of both high and low-affinity functioning through multiple mechanisms: 1) trapping invading microbes in mucus 2) preventing attachment of bacteria to the epithelium by binding bacterial attachment-adhering structures 3) altering expression of pro-inflammatory bacterial molecules by cell-surface binding and 4) transporting microbes within the epithelium back to the lumen and facilitating their clearance (15). SIgA shapes the membership of microbial communities as well. An ability to generate a repertoire of both low and high affinitiy SIgA is a key feature of SIgA's ability to effectively regulate the commensal microbiota. Mice deficient in activation-induced cytidine deaminase (AID), which plays important roles in somatic hypermutation, show a 100-fold expansion of anaerobic bacteria in the small intestine (16). Studies of germ-free RAG1-/- bearing a Bacteroides specific capsular SIgA hybridoma and monoassociated with Bacteroides thetaiotaomicron established the peacekeeping role of SIgA by demonstrating that the presence of the SIgA both down-regulated pro-inflammatory and metabolic bacterial genes and dampened pro-inflammatory host responses (17). Thus abnormalities in SIgA production and function could underlie aberrant host inflammatory responses and promote alterations in bacterial gene expression that reinforce pro-inflammatory host programs. As TRUC mice are deficient in RAG2-/-, they lack mature lymphocytes because VDJ recombination cannot be initiated and thus have no SIgA as B cells are absent. While the RAG2-/- mice in our colony do not have intestinal inflammation, the lack of SIgA in TRUC mice may be one of multiple factors that shape a permissive environment for outgrowth of opportunistic commensals and contribute to the development of a host with increased inflammatory tone.
T-bet is a T-box transcription factor first identified as the master-regulator of the CD4 T helper type 1 lineage of T cells. Over the past decade our understanding of the cellular expression patterns of T-bet and the T-bet transcriptionally-regulated universe of genes has expanded. T-bet controls the expression of numerous cytokines, chemokines, and chemokines receptors (18) and depending upon the expressing cell type, T-bet can function as either a repressor or activator of a given gene. In myeloid dendritic cells, we have demonstrated that T-bet binds regions within both the human and mouse TNF- α 5′ untranslated region using chromatin immunoprecipitation (3), and we envision that T-bet exerts its repressive effects as part of a larger multimeric complex. Our group has previously shown that T-bet represses IL-2 in developing Th1 cells by dimerizing with RelA and thus interfering with RelA binding to the IL-2 promoter (19). In vitro, T-bet deficient bone marrow derived DCs (BMDCs) produce more TNF-α in response to LPS as compared to WT DCs. However, in vivo, loss of T-bet only induced spontaneous dysregulated TNF- α production from lamina propria myeloid DCs in the absence of adaptive immunity. This latter observation demonstrates the important role for adaptive immunity in regulating intestinal dendritic cell responses and the multiple layers of regulation within the immune system necessitated by the complex environment of the intestine.
DCs have emerged as key effectors in the pathogenesis of IBD (reviewed in (20)). Data from numerous mouse models are providing insights into how mucosal DCs function in intestinal health and disease. DCs function as linchpins of pathogenic demonstrated in the dextran sulfate sodium model of colitis, wherein deletion of dendritic cells using CD11c diphtheria toxin transgenic mice ameliorated colitis (21). In CD40-induced colitis, Helicobacter hepaticus infection of RAG2-/- induced colitis and in the naive CD4+ CD45RBhi T cell adoptive transfer model of colitis in immunodeficient mice, DCs are pre-eminent producers of IL-23, a central pro-inflammatory cytokine in both innate and adaptive driven intestinal inflammation (22-24). IL-23 is a member of the IL-12 family of cytokines. It is a heterodimer consisting of IL-23p19 and IL-12p40, which is common to IL-12. IL-23 plays a pathogenic role in a number of disease states in addition to IBD, including multiple sclerosis, rheumatoid arthritis, and psoriasis by unleashing pro-inflammatory programs in both T cells and innate immune cells (25). IL-23R polymorphisms have been linked to the two major sub-types of IBD, Crohn's Disease and ulcerative colitis and a distinct polymorphism may, in contrast, confer protection from IBD (26). While IL-23 has a clear role in Th17 stabilization and maintenance, the function of IL-23 is more complex than just promoting the Th17 subsets within the intestine. IL-23 also appears to regulate the induction of Foxp3+ regulatory T cell as well (27).
As outlined above DCs are important contributors for the inflammatory response in aberrant chronic inflammatory states in the intestine. However, DCs subsets are usually quite adept at mediating appropriate pro-inflammatory responses against invasive intestinal pathogens and also pro-tolerogenic responses to food-derived antigens and commensal microbes. DCs are systems of cells that are sentinels and adept interpreters of the luminal microbiota they survey—evaluating the microbiota for commensals, opportunists, or intestinal pathogens. T-bet may prove to be one of several key genes that tune dendritic cell signaling pathways that sense intestinal microbiota. Not only do DCs sample the microbiota directly via sending extensions of their plasma membrane through epithelial tight junctions into the lumen (28, 29), DCs also interpret signals via epithelial cells. TSLP is a soluble epithelial cell secreted molecule shapes DC responses to promote T helper type 2 and T regulatory subsets that tip the balance against intestinal inflammation (30, 31). However, other recent data suggest the mucosal phenotype of a T-regulatory promoting CD103+ DC may derive not from intestinal epithelial TSLP but rather from epithelial cell-derived TGF-β and retinoic acid (31).
TRUC mice lack mature T lymphocytes becaue of a RAG2 mutation. Since T-bet-/- mice on an immunosufficient background are free of spontaneous intestinal inflammation in our colony, we reasoned that an adaptive immune subset may actively inhibit colitis. Adoptive transfer of CD4+CD62LhiCD25- cells resulted in extremely moribund mice within 2 weeks of transfer stemming from robust T cell inflammatory responses superimposed on a background of chronic colonic ulceration and inflammation. Transfer of T-regulatory type cells (T-regs) has been shown to block the development of intestinal inflammation in inducible colitis models in immunodeficient mice lacking these subsets and to reverse established colitis (reviewed in (32)). T-regs are not a single cell type but consist of at least three distinct subsets (Foxp3+ T-regs+, Foxp3- type 1 T-regs, and follicular helper T cells). They balance inflammatory responses by regulating other T cell subsets and innate immune cell function (33) via cell-contact dependent mechanisms and via the secretion of TGF-β, IL-10, and IL-35 (reviewed in (34)). Following transfer of CD4+CD62LhiCD25+ cells into TRUC mice at four weeks of age, mice appeared healthy and when colons were removed and examined a month after transfer there was no evidence of inflammation. The large DC infiltrates, which are a prominent feature of TRUC colons, were absent. In unpublished observations, we have seen that the anti-inflammatory effects of T-regs on TRUC colitis do last several months but are not durable beyond that time frame.
TRUC colitis is antibiotic-responsive as are most mouse models of colitis; similarly, patients with IBD often experience symptomatic benefit from treatment with antibiotics such as metronidazole. TRUC mice, interestingly, appear to harbor microbes that are generally colitogenic. Specifically, we have observed the transmission of colitis from TRUC to T-bet sufficient mice in both cross-fostering experiments and experiments wherein adult TRUC mice are co-housed with RAG2-/- or wild type mice. Notably, this transmissible colitis does not appear to be initiated and driven by TNF-α as is observed in TRUC. There are a number of identified microbes that drive and/or enhance intestinal inflammation in rodents, for example, Helicobacter hepaticus, Citrobacter rodentium, Enterotoxigenic Bacteroides fragilis (Btf+ B. fragilis), and certain strains of Bacteroides vulgatus- however, TRUC mice lack these organisms (35-38). There have been numerous observations about correlations between microbes and IBD in patients, and, the general consensus is that IBD may be driven by aberrant pro-inflammatory host responses to the commensal flora (39, 40). Intestinal inflammation can have marked affects on commensal microbes (41) and on microbe gene expression (42). Inflammation itself can have dramatic effects on intestinal microbial communities and perpetuate the outgrowth of select groups which may promote chronic inflammation (43). Hopefully, our ongoing studies of the TRUC microbiota will provide additional insight into the fascinating interactions between host innate immunity and the commensal microbiota.
TRUC colitis is ameliorated by antibiotics, T-regulatory cell infusion, and TNF-α neutralization and thus TRUC mice provide a useful model for interrogating and evaluating the mechanisms by which these modalities ameliorate intestinal inflammation. Our studies of TRUC pathogenesis and the prominent role of DCs therein furnish additional evidence of the importance of dendritic cells in IBD. RAG2-/- deficiency, as discussed above, confers several disadvantages on TRUC mice regarding their ability to maintain a peaceful co-existence with their microbiota, specifically, loss of mature lymphocytes depriving TRUC of the benefits of secretory IgA and T-regulatory subsets. The additional loss of T-bet highlights another control mechanism—a checkpoint on lamina propria myeloid DC TNF-α production. The profound effect of this DC-derived TNF-α on the epithelium highlights the important role of T-bet in DCs for maintenance of the epithelial barrier. The protective role of T-bet in DCs may be harnessed in the future to provide T-bet based therapeutics for IBD.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.