CHAPTER 2 Mucosal Immunity
Mucosal immunity refers to immune responses that occur at mucosal sites. The demands on the mucosal immune system are distinct from their systemic counterparts. At mucosal sites, the outside world is typically separated from the inner world by a single layer of epithelium. The mucosal immune system exists at a number of sites, including the respiratory tract (especially the upper respiratory tract), the urogenital tract, the mammary glands, and even the eye and ear. Regardless of the site, unique lymphoid and other cell populations are required to handle a wide array of environmental stimuli. Together, these sites are called mucosa-associated lymphoid tissue, or MALT.1–5 However, the site that is most often associated with mucosal immunity is the intestinal tract.
The intestinal tract is unique in several aspects. In contrast to other mucosal sites, it is the least sterile, containing billions to trillions of microorganisms, mainly bacteria. These organisms, along with ingested food, represent an enormous antigenic load that must be tolerated to maintain the status quo in the intestine. This unusual environment and the demands associated with it have resulted in the development of a distinct immune system designated as gut-associated lymphoid tissue (GALT).
The specific characteristics and peculiarities of GALT reflect the unique milieu in which it needs to function. To maintain homeostasis in the intestine, one of the most important tasks of GALT is to distinguish potentially harmful antigens, such as pathogenic bacteria or toxins, from ones that may benefit the body, such as those derived from food or commensal bacteria. To achieve homeostasis, unusual cell types, immunoglobulins (Igs), and secreted mediators need to function in a coordinated fashion. In contrast to the systemic immune system, whose focus is to act quickly within seconds of encountering a foreign antigen, the GALT is poised to respond but is predominantly tolerant, rejecting harmful antigens but allowing beneficial or harmless ones to persist without evoking immune responses such as allergic reactions or inflammation. The unique ways in which GALT performs in its demanding environment are the focus of this chapter, along with the consequences of abnormal GALT function that result in intestinal allergic or inflammatory diseases (discussed in other chapters).
As noted, the hallmark of mucosal, in contrast to systemic, immunity is suppression, exemplified by two linked phenomena, controlled or physiologic inflammation and oral tolerance. These processes are mediated by a unique anatomy, distinct resident cell populations, and selective antibody isotypes.
Within the lamina propria exist billions of activated plasma cells, memory T cells, memory B cells, macrophages, and dendritic cells.6,7 In fact, given the large surface area of the gastrointestinal (GI) tract and the resident cell populations that inhabit this space, the gut is the largest lymphoid organ in the body. Still, in contrast to activated lymphocytes in the peripheral immune system, significant inflammation is not present in the intestine. This phenomenon has been termed controlled/physiologic inflammation (Fig. 2-1).
Figure 2-1. Mechanisms for damping the mucosal immune response. The intestine uses a number of distinct mechanisms to dampen mucosal immune responses. The major source of antigen in the intestine is the commensal bacterial flora, but both innate and adaptive responses control local responses. Physical barriers such as mucins secreted by goblet cells and tight junctions between epithelial cells prevent invasion by luminal flora. Defensins such as HBD-2, -3, and -4 are thought to maintain the sterility of the crypt, whereas SIgA produced by local plasma cells prevents attachment and invasion by luminal bacteria, thereby reducing the antigenic load. Even in the face of antigenic challenge, the lymphocytes, macrophages, and dendritic cells in the intestine are programmed not to respond as a consequence of decreased expression of pattern recognition receptors (e.g., Toll-like receptors) and a decrease in the ability of lymphocytes to be activated through their antigen receptor. DC, dendritic cell; HBD, human β-defensin; IELs, intraepithelial lymphocytes; LPMC, lamina propria mononuclear cells; MAdCAM, mucosal addressin cell adhesion molecule; SIgA, secretory immunoglobulin A, a dimer with a connecting J chain; Treg, T regulatory cells (formerly known as suppressor T cells).
The entry and activation of the cells into the lamina propria is antigen-driven. Germ-free mice have few cells in their lamina propria. However, within hours to days following colonization with normal intestinal flora (without pathogens), there is a massive influx and activation of cells.8–11 Despite the persistence of an antigen drive (luminal bacteria), the cells fail to develop into aggressive inflammation-producing lymphocytes and macrophages. Bacteria or their products play a role in this persistent state of activation12 and likely contribute to the controlled inflammatory process as well.
The failure to produce gastrointestinal pathology, despite the activation state of intestinal lymphocytes, is probably the consequence of regulatory mechanisms. The failure of lamina propria lymphocytes (LPLs) to generate “normal” antigen-receptor mediated responses is an important factor in controlled inflammation (e.g., lack of expansion, despite a state of activation). LPLs respond poorly when activated via their T cell receptor (TCR), failing to proliferate although they still can produce cytokines.13,14 This is key to the maintenance of controlled inflammation.
The most recognized phenomenon equated with mucosal immunity and associated with suppression is oral tolerance.15–20 Oral tolerance can be defined as the active, antigen-specific nonresponse to antigens administered orally.18,21,22 Disruption of oral tolerance may result in food allergies (see Chapter 9) and food intolerances such as celiac disease (see Chapter 104). Part of the explanation for oral tolerance relates to the properties of digestion per se, where large potentially antigenic macromolecules are degraded so that potentially immunogenic substances are rendered nonimmunogenic. However, approximately 2% of dietary proteins enter the draining enteric vasculature intact.23 How does the body regulate the response to these potential antigens that have bypassed complete digestion? This is achieved by oral tolerance. Factors affecting the induction of oral tolerance are the age of the host, genetic factors, the nature of antigen, the tolerogen’s form, and dose. In addition, the state of the intestinal barrier affects oral tolerance and, when barrier function is reduced, oral tolerance decreases as well.
Oral tolerance is difficult to achieve in the neonate, probably because of the rather permeable intestinal barrier that exists in the newborn, as well as the immaturity of the mucosal immune system. Within 3 months of age (in the mouse), oral tolerance can be induced and many previous antibody responses to food antigens are suppressed. The limited diet in the newborn may further serve to protect the infant from generating a vigorous response to food antigens. Furthermore, the intestinal flora has been demonstrated to affect the development of oral tolerance. Probiotics such as Lactobacillus GG given to mothers before delivery and during lactation have provided protection against the development of atopic eczema in their offspring.24 The effects of probiotics on oral tolerance are probably mediated through modulation of cytokine responses,25 a positive effect on intestinal barrier function and restitution of tight junctions,26–27 suppression of intestinal inflammation via down-regulation of Toll-like receptor (TLR) expression,28,29 and secretion of metabolites that may inhibit inflammatory cytokine production by mononuclear cells. The role of genetic factors in oral tolerance has been suggested in murine models in which certain strains develop tolerance more easily than others.30
The nature and form of the antigen also play a significant role in tolerance induction. Protein antigens are the most tolerogenic whereas carbohydrates and lipids are much less effective at inducing tolerance.31 The way the antigen is delivered is also critical. For example, a protein delivered in soluble form (e.g., ovalbumin) is tolerogenic, whereas aggregation of this protein reduces its potential to induce tolerance. This phenomenon may be associated with an alteration in the sites of antigen sampling.6 Exposure or prior sensitization to an antigen through an extraintestinal route also affects the development of tolerance responses. Finally, the dose of antigen administered is critical to the form of oral tolerance generated. In mouse models, high doses of antigen are associated with clonal deletion or anergy.32 In this setting, tolerance is not transferable; transfer of T cells from tolerized animals does not lead to the transfer of tolerance. The mechanism underlying T cell deletion is possibly Fas-mediated apoptosis.33 On the other hand, low doses of antigen activate regulatory-suppressor T cells.34,35 Increasing numbers of such T cells occur, both in CD4 and CD8 lineages. Th3 cells were the initial regulatory-suppressor cells described as mediators of oral tolerance.35–37 These cells appear to be activated in Peyer’s patches and secrete transforming growth factor-β (TGF-β). This cytokine plays a dual role in mucosal immunity; it is the most potent suppressor of T and B cell responses while also promoting the production of IgA (TGF-β is the IgA switch factor).38–41 The production of TGF-β by Th3 cells elicited by low-dose antigen administration helps explain an associated phenomenon of oral tolerance, namely, bystander suppression. Whereas oral tolerance is antigen specific, the effector arm is antigen non-specific. If an irrelevant antigen is coadministered systemically with the tolerogen, suppression of T and B cell responses to that irrelevant antigen will also occur (hence, bystander suppression). Secreted TGF-β suppresses the response to the coadministered antigen.
T regulatory 1 (Treg1, or Tr1) cells may also participate in bystander suppression and oral tolerance by producing interleukin-10 (IL-10), another potent immunosuppressive cytokine.42–44 Evidence for the activation of CD4+, CD25+ regulatory T cells during oral tolerance also exists, although their exact role in this process is still being investigated.45–49 Tolerance studies carried out in mice depleted of CD4+, CD25+ T cells coupled with neutralization of TGF-β have demonstrated that CD4+, CD25+ T cells and TGF-β together are involved in the induction of oral tolerance, partly through regulation of the expansion of antigen-specific CD4+ T cells.50 The ability to identify regulatory CD4+, CD25+ T cell subpopulations was enhanced by the recognition that these cells express the transcription factor FoxP3. Because not every cell within the CD4+, CD25+ population is a naturally occurring Treg cell, the ability to use FoxP3 as a marker of these Treg cells has been a major breakthrough in our ability to study these cells.51,52 Importantly, the absence of CD4+ T regulatory cell activity in mice results in inflammatory bowel disease (IBD), although this has not been demonstrated in humans.53–56
Preliminary data also support a role for antigen-specific CD8+ T cells in oral tolerance,57–61 as well as in the regulation of mucosal immune responses. Specifically, in vitro activation of human CD8+ peripheral blood T cells by normal intestinal epithelial cells (IECs) results in the expansion of CD8+, CD28− T cells with regulatory activity. Moreover, in the lamina propria of IBD patients, such CD8+, CD28− cells were significantly reduced, supporting a role for these epithelial-induced T regulatory (TrE) cells in the control of intestinal inflammation.62
Another factor affecting tolerance induction is the state of the intestinal barrier. In addition to the failure to generate tolerance in the neonatal period (when intestinal permeability is higher), several other states of barrier dysfunction are associated with aggressive inflammation and a lack of tolerance. During anaphylaxis, increased intestinal permeability caused by the disruption of tight junctions allows luminal antigens to pass through paracellular spaces.63–65 Treatment with interferon-γ (IFN-γ) can disrupt the mucosal barrier in mice and they fail to develop tolerance in response to ovalbumin feeding. Even more interesting is the failure of N-cadherin–dominant negative mice to suppress mucosal inflammation (loss of controlled inflammation),66 possibly because of the enormous antigenic exposure resulting from the leaky barrier in these mice.
Lastly, oral tolerance may also be associated with the cells serving as the antigen-presenting cell (APC; see later), as well as the site of antigen uptake. Orally administered reovirus type III is taken up in mice by microfold (M) cells expressing reovirus type III–specific receptors (Fig. 2-2).67 This epithelial uptake by M cells induces an active IgA response to the virus. Reovirus type I, on the other hand, infects intestinal epithelial cells (IECs) adjacent to M cells and this uptake induces tolerance to the virus. Thus, the route of entry (M cell versus IEC) of a specific antigen may dictate the type of immune response generated (IgA versus tolerance). Interestingly, poliovirus vaccine, one of the few oral vaccines effective in humans, binds to M cells, which may account for its ability to stimulate active immunity in the gut.68
Figure 2-2. M cell. This transmission electron micrograph from the noncolumnar region of the Peyer’s patch epithelium shows a cross-sectional view of a microfold (M) cell, as well as associated microvillus-covered epithelial cells and at least three lymphoid cells (L). Note the attenuated cytoplasm of the M cell (between arrows) that bridges the surface between microvillus-covered epithelial cells, forming tight junctions with them and producing a barrier between the lymphoid cells and the intestinal lumen (×9600). B, B cell; E, epithelial cell.
(From Owen RL, Jones AL: Epithelial cell specialization within human Peyer’s patches: An ultrastructural study of intestinal lymphoid follicles. Gastroenterology 1974; 66:189-203.)
The unique antibody, secretory IgA (SIgA), is the hallmark of MALT-GALT immune responses (Fig. 2-3). Although IgG is the most abundant isotype in the systemic immune system, IgA is the most abundant antibody in mucosal secretions.69–71 In fact, given the numbers of IgA-positive plasma cells and the size of the MALT system, IgA turns out to be the most abundant antibody in the body. SIgA is a dimeric form of IgA produced by plasma cells in the lamina propria and transported into the lumen by a specialized pathway through the intestinal epithelium (Fig. 2-4). Two IgA monomers are bound together by a J chain (also produced by plasma cells). Subsequently, the dimer binds to secretory component (SC), also known as the polymeric immunoglobulin receptor (pIgR), a highly specialized 55-kd glycoprotein produced by IECs. SC (pIgR) is expressed on the basolateral aspect of the IEC and binds only to dimeric IgA or to IgM (also polymerized with J chain; see later). Once bound to SC, SIgA is actively transported within vesicles to the apical membrane of the IEC. The vesicle fuses with the apical membrane and the SC-IgA complex is released into the lumen. Once in the lumen, SC serves its second function, protection of the SIgA dimer from degradation by luminal proteases and gastric acid. SIgA and SIgM are the only antibodies that can bind SC and therefore withstand the harsh environment of the GI tract.
Figure 2-4. Assembly and secretion of dimeric immunoglobulin A (IgA). IgA and J chain produced by IgA-committed plasma cells (bottom) dimerize to form polymeric IgA, which covalently binds to membrane-bound polymeric Ig receptor produced by epithelial cells (top). This complex is internalized, transported to the apical surface of the epithelial cell, and secreted into the lumen. SC, secretory component.
In addition to its unique form, SIgA is also unique as an immunoglobulin in that it is anti-inflammatory in nature. SIgA does not bind classic complement components but rather binds to luminal antigens, preventing their attachment to the epithelium or promoting agglutination and subsequent removal of the antigen in the mucus layer overlying the epithelium.69,72 This binding of luminal antigens by SIgA reflects immune exclusion, as opposed to the nonspecific mechanisms of exclusion exerted by the epithelium, the mucous barrier, proteolytic digestion, and other mechanisms.
As noted, IgM is the other antibody capable of binding SC (pIgR). Like IgA, IgM uses J chains produced by plasma cells to form polymers—in the case of IgM, a pentamer. SC binds to the Fc portion of the antibody formed during polymerization. The ability of IgM to bind SC may be important in individuals with IgA deficiency in which secretory IgM (SIgM) may compensate for the absence of IgA in the lumen.
Whereas SIgA is the major antibody isotype produced in GALT, IgG has been detected as well.73–74 The neonatal Fc receptor (FcRN), expressed by IECs, might serve as a bidirectional transporter of IgG75,76 and may be important in control of neonatal infections and IgG metabolism. In IBD, marked increases in IgG in the lamina propria and lumen have been observed.77
Even IgE production may play an important role in intestinal diseases in GALT. CD23 (low-affinity IgE Fc receptor) has been reported to be expressed by IEC. One model has suggested that CD23, or FcεRII, may play a role in facilitated antigen uptake and consequent mast cell degranulation in food allergy. In this setting, IgE transcytosis and mast cell degranulation may be associated with fluid and electrolyte loss into the lumen, an event that is intimately associated with allergic reactions in the gut and airways.78,79
The cells, structures, and mediators separating the intestinal lumen from the lamina propria function as a physical barrier. However, this physical barrier is a biologically active structure that constantly interacts with its ever-changing environment. The intestinal barrier changes not only on a daily basis but also over the years. Many barrier mechanisms are not fully developed at birth, and evidence exists to support less restricted antigen transport in neonates compared with adults, specifically in animals.
Physiologic factors operative in the upper GI tract influence the antigenic load that reaches the major sites of GALT in the small and large bowels. Detailed exploration of these factors are beyond the scope of this chapter and are discussed elsewhere, but include proteolysis, gastric acidity, and peristalsis.
The mucous coat lining the intestinal tract is composed of a mixture of glycoproteins (mucins). The protein core of mucins is enriched in serine, threonine, and proline residues, and carbohydrate moieties are attached via N-acetylgalactosamine residues. At least six different mucin species have been identified,80 each with a distinct carbohydrate and amino acid composition. Mucus protects the intestinal wall by several mechanisms. Its stickiness and competitive binding to glycoprotein receptors decrease the ability of microorganisms to penetrate the intestine.81 It also generates a stream that moves luminal contents away from epithelial cells.
Underneath the mucous layer, the physical barrier that prevents penetration of antigen across the intestinal epithelium consists of the actual epithelial cell (the transcellular route) and the tight intercellular spaces (the paracellular route) regulated by tight junction (TJ) complexes (e.g., zona occludens) and the subjunctional space.82 Of the two structures, TJs have the greater role in preventing macromolecular diffusion across the epithelium, because these junctions exclude almost all molecules present in the lumen (see Chapter 96). The barrier formed by the TJ is a dynamic structure, preserved even when epithelial cells themselves are damaged; this feature might be crucial for the prevention of intestinal inflammation (e.g., as seen in idiopathic IBD).
The epithelial cells themselves serve as a physical barrier in several ways—their microvilli are at a distance of about 25 nm from each other and are negatively charged. Thus, a negatively charged luminal molecule would be repelled from passage even if its diameter were well below 25 nm. However, intact antigens may traverse the epithelium by fluid phase endocytosis and enter the circulation.83
To accomplish the two major goals of the mucosal immune system in the intestine (maintenance of homeostasis and clearance of pathogens), several key features have been identified. Compartmentalization of cells into distinct regions and sites, despite being millimeters away from each other, is a hallmark of the GALT. Cell populations and the immune response in the epithelium, subepithelial region, lamina propria (LP), Peyer’s patches, and mesenteric lymph node (MLN) may differ substantially. The cells residing in these compartments differ not only topographically but also phenotypically and functionally, depending on the anatomic site in GALT. Cells with distinct phenotypes and functions are attracted to specific sites in GALT.
The follicle-associated epithelium (FAE), which contains microfold (M) cells, is a specialized epithelium overlying the only organized lymphoid tissue of GALT, Peyer’s patches (PPs). The M cell, in contrast to the adjacent absorptive epithelial cell, has few microvilli, a limited mucin overlayer, a thin elongated cytoplasm, and a shape that forms a pocket surrounding subepithelial, T, B, macrophages, and dendritic cells (see Fig. 2-2). M cells are capable of taking up large particulate antigens from the lumen and transporting them intact into the subepithelial space.84–86 M cells contain few lysosomes, so little or no processing of antigen occurs.87 M cells are exposed to the lumen, thus having a larger area for contact with luminal contents than adjacent epithelial cells. The M cell expresses several unique lectin-like molecules that help promote binding to specific pathogens—the prototype being poliovirus.88 Antigens that bind to the M cell and are transported to the underlying PP generally elicit a positive (SIgA) response. Successful oral vaccines bind to the M cell and not to the adjacent epithelium. Thus, M cells appear to be critical for the positive aspects of mucosal immunity.
The M cell is a conduit to PPs. Antigens transcytosed across the M cell and into the subepithelial pocket are taken up by macrophages and DCs and carried into PPs. Once in the patch, TGF-β–secreting T cells promote B cell isotype switching to IgA.89 Importantly, there is a clear relationship between M cells and PPs. The induction of M cell differentiation has been shown to be dependent on direct contact between the epithelium and B lymphocytes in PPs.90 M cells do not develop in the absence of PPs. For example, M cells have not been identified in B cell–deficient animals in whom there are no PPs.91 Even though M cells and PPs may be involved in oral tolerance,92–94 PP-deficient mice are capable of developing tolerance after oral administration of soluble antigen.95
After activation in PPs, lymphocytes are induced to express specific integrins (α4β7), which provide a homing signal for mucosal sites (where the ligand is MadCAM-1).96–98 Cells then travel to MLNs and subsequently into the main intestinal lymphatic drainage system, the thoracic duct, and finally into the systemic circulation (Fig. 2-5). There, mucosally activated lymphocytes with their mucosal addressins circulate in the bloodstream to exit in high endothelial venules in various mucosal sites. Those bearing α4β7 molecules exit in the MALT-GALT lamina propria where they undergo terminal differentiation. Chemokines and their receptors (see later) as well as adhesion molecules and ligands may help direct this trafficking pattern.
Figure 2-5. Mucosal lymphocyte migration. Following antigenic stimulation, T and B lymphocytes migrate from the intestine to the draining mesenteric lymph nodes, where they further differentiate and then reach the systemic circulation via the thoracic duct. Cells bearing the appropriate mucosal addressins then selectively home to mucosal surfaces which constitute the common mucosal immune system, including the intestine.
The epithelium is composed of a single layer of columnar cells. These IECs are derived from the basal crypts and differentiate into absorptive villous or surface epithelium, secretory goblet cells, neuroendocrine cells, or Paneth cells (see Chapter 96). In addition to their function as a physical barrier in GALT discussed earlier, IECs contribute to innate and adaptive immunity in the gut and may play a key role in maintaining intestinal homeostasis.
The ability of intact antigen to cross the lipid bilayer at the surface of the IEC (underneath the microvilli) is limited. However, invagination of apical membranes occurs regularly, allowing macromolecules to be carried into the cell within membrane-bound compartments.
Binding to the surface of the cell depends on the structure of the antigen and the chemical composition of the microvillous membrane. For instance, bovine serum albumin binds less efficiently to the surface of the IEC than bovine milk protein and, as a consequence, is transported less efficiently.99 In addition, structural alterations in an antigen caused by proteolysis might also affect its binding, because this will change the physicochemical characteristics of the molecule.100
Several factors influence the transport of antigens from the apical to the basolateral surface of IECs. The rate of vesicular passage to the basolateral membrane depends on the rate of endocytosis, the proportion of vesicles trafficking to the lysosome, and the speed of travel of membrane-bound compartments. Lysosomally derived enzymes determine the rate of breakdown of products contained in membrane compartments. These include proteases such as cathepsins B and D (found throughout the length of the intestine, particularly in the mid and distal thirds of the small intestine), as well as those enzymes that catalyze carbohydrate breakdown, such as acid phosphatase and mannosidase. It is the degree to which the organellar contents encounter such enzymes (in the lysosome or in endocytic vesicles) that determines the rate of intracellular destruction of macromolecules.101 Although cathepsins are capable of catalyzing antigens, they may not completely digest the protein, which may require further proteolysis by peptidases in the cytoplasm.
Classic APCs in the systemic immune system possess the innate capacity to recognize components of bacteria and viruses, called pathogen-associated molecular patterns (PAMPs). Receptors for these PAMPs are expressed on the cell surface (e.g., TLRs) and inside the cell (e.g., NOD2 [see later]). Despite the fact that IECs live adjacent to large numbers of luminal flora, IECs retain the ability to recognize components of these bacteria. In general, proinflammatory responses are normally down-regulated (i.e., expression of the lipopolysaccharide [LPS] receptor, TLR4, is absent) and expression of some of these pattern recognition receptors is maintained, such as TLR5, which recognizes bacterial flagellin. TLR5 is expressed basolaterally, so it is poised to identify organisms such as Salmonella species that have invaded the epithelial layer.102 Following invasion and engagement of TLR5, the IEC is induced to secrete a broad array of cytokines and chemokines that attract inflammatory cells to the local environment to control the spread of infection. In contrast to invading pathogens, some bacteria are probiotic and induce the IEC to produce anti-inflammatory cytokines (e.g., IL-10) and to increase the expression of peroxisome proliferator-activated receptor-γ (PPAR-γ).103 Furthermore, other bacterial products help promote the barrier and IEC differentiation (e.g., products of Bacteroides thetaiotaomicron).
The significance of the ability of IECs to recognize PAMPs via surface TLRs, such as TLR5, or via intracellular nuclear oligomerization domain 1, 2 (NOD1, 2), has been increasingly recognized over the past decade. The latter ability has been shown to contribute to intestinal inflammation, because about 25% of patients with Crohn’s disease have mutations in the NOD2-CARD15 gene, interfering with their ability to mount an appropriate immune response to bacterial stimuli (discussed further in Chapter 111).104–108