Development of the Intestinal Mucosal Barrier

In this tribute to the professional contributions of Professor John Walker-Smith to the field of pediatric gastroenterology, I will review our understanding of the importance of the intestinal mucosal barrier in intestinal host defense, an important research concept that has evolved during the past 25 years. Our laboratory has had a longstanding research interest in the development of the mucosal barrier as it pertains to the pathogenesis of age-related gastrointestinal diseases in infancy. Ironically, the important observations that frame our current understanding of barrier function have occurred during the same period that John Walker-Smith has contributed significantly with his observations of the developmental function of the small intestine. This review will cover important landmarks in our understanding of mucosal barrier function. Initially, the epithelial lining of the intestinal surface was considered simply a physical barrier to foreign antigens and microorganisms in the intestinal lumen. Subsequently, several seminal observations suggested that the gut epithelium was an active participant in the mucosal immune response and provided communication between luminal stimulants and underlying lymphoid elements in the lamina propria. In response to interaction with microbes in the lumen, the epithelium can regulate the transcription of inflammatory cytokines for excretion into the lumen and interstitium and can up-regulate surface molecules, such as class-II antigens and the polymeric immunoglobulin-A (poly-IgA) receptor, for control in handling foreign antigens. This interaction with luminal pathogens and commensal flora with the gut is termed microbial–epithelial “cross talk” and represents an important contributor to intestinal barrier function. Finally, we have evidence that components of the intestinal mucosal barrier are underdeveloped at birth and respond inappropriately to luminal stimuli. As a result of this immature, inappropriate response to microbial stimuli, certain infectious diseases occur in greater frequency during infancy. GUT AS A PHYSICAL BARRIER As the field of mucosal immunology began, the intestinal epithelium was considered a physical barrier to the uptake of foreign antigens and the penetration of pathogens (1–3). By maintaining the integrity of the epithelium with its tight junctions, controlled quantities of antigen could be taken up by a pinocytotic mechanism and this uptake could result in a physiologic pIgA response. The primary function of pIgA, produced by programmed plasma cells in the lamina propria underlying the epithelium of antigen entry, was to agglutinate antigen by forming complexes that prevented excessive uptake and adverse systemic immunologic responses (3). Polymeric IgA binds to a basolateral receptor on the enterocyte and is transported across the epithelium onto the luminal microvillus (4,5) surface. A component of the receptor, known as secretory component, remains with the immunoglobulin to protect it against intraluminal proteolysis. In its polymeric form, pIgA can more efficiently agglutinate antigen on the luminal surface. In addition to a unique immunoglobulin response to antigen/pathogen penetration, several nonspecific intestinal functions contribute to a decreased exposure to luminal antigens. These nonspecific intestinal factors include gastric acidity, mucus, digestive enzymes, and peristalsis. Figure 1 depicts the intestinal defense barriers (1,3). To prevent colonization of the upper small intestine by bacteria ingested in contaminated foods, the stomach maintains an acid milieu to destroy ingested pathogens. The goblet cell is a major epithelial cell lineage throughout the small and large intestine. Goblet cells increase in number along a distal axis from the small to large intestine. Preformed mucus can be released in response to luminal stimulation to provide an added protective coat on the epithelial surface, further preventing excessive uptake of luminal antigens and pathogens (1,3). The intestinal lumen is replete with digestive enzymes released in pancreatic secretions and epithelial cells that facilitate the breakdown of ingested foods and facilitate absorption of breakdown products. These same digestive enzymes breakdown antigens to minimize the antigen load to the epithelium. Finally, peristaltic movement of intestinal smooth muscle in response to stimulation can facilitate the rapid expulsion of antigen–antibody complexes and excessive luminal antigens and microorganisms (1).FIG. 1.: Barriers to antigenic absorption in the intestine. Antigen entry is limited by nonimmunologic and immunologic mechanisms and by the structure of the epithelium. If potential antigens cross into the enterocyte, as they are being degraded and presented as antigens, signals are produced, causing T-cell activation and subsequent additional cytokine production and release. Reproduced with permission from a review by Insoft et al. (13).Initially, it was assumed that these nonspecific intestinal barrier features operated independently from immune barrier functions. However, subsequent studies have demonstrated that the various defense barriers within the intestine work synergistically to provide optimal protection of the epithelial surface (6). For example, we reported that formation of pIgA-antigen complexes on the intestinal surface maintained the antigens in contact with luminal digestive enzymes for longer periods of time and thus facilitated excessive proteolysis, reducing the amount of antigen available to cross the epithelium (7)(Fig. 2). We also reported that formation of complexes on the intestinal surface enhances the release of mucus from goblet cells, thereby thickening the physical barrier to antigen/pathogen uptake (Fig. 3). Other studies have implicated the release of cytokines and other soluble factors from lymphoid cells and enterocytes in activating gastric acidity (8). Activating acid release from gastric cells and the contribution of smooth muscle cells through neurotransmitters enhances peristalsis, thereby broadening the protective function of the gut. These observations suggest that an integrated system of physical barriers exist in respond to noxious stimuli in the gut.FIG. 2.: Schematic representation of the processing of protein antigen at the surface of the gut. Before immunization (left), a small portion of ingested protein escapes intraluminal digestion and is taken up by the enterocyte and transported to the intercellular spaces. After immunization (right), antibodies on the gut surface interact with antigen to form complexes, thereby preventing or decreasing the binding of antigen to, and subsequent pinocytosis of antigen by, intestinal epithelial cells. Pancreatic enzymes absorbed to the gut surface may degrade antigen complexed with antibodies in the mucus coat (glycocalyx). Consequently, less antigen is available to intestinal epithelial cells. Reprinted with permission from Walker et al. (7).FIG. 3.: Cartoon of the relative role of antigen–polymeric immunoglobulin A (pIgA) complexes compared with antigen alone in the release of mucus from goblet cells in the small intestine. Formation of complexes triggers enhances mucus release that increases the physical barrier of the mucus that coats the luminal surface of enterocytes.GUT EPITHELIUM AS AN ACTIVE PARTICIPANT IN MUCOSAL IMMUNITY Almost a decade ago, investigators from several laboratories provided evidence that the intestinal epithelium can actively participate in the mucosal immune response. After antigenic / bacteria–bacterial toxin interaction with the microvillus of the enterocyte, the cell responds with the transcriptional up-regulation of several immunologically relevant molecules (9) (Fig. 4). This interaction, by mostly unknown mechanisms, activates the up-regulation of genes for inflammatory cytokines that are secreted and act as sensors for underlying lymphocytes, which are in turn activated as a result of enterocyte stimulation (10). The up-regulation of the enterocyte transcription of the chemokine, interleukin (IL)–8, which stimulates migration of polymorphonuclear neutrophils to combat bacterial invasion is an example of this interaction. In addition, a variety of luminal stimuli can increase the enterocyte transcription of immunologic surface molecules such as the pIgA receptor, class-II antigens, and adhesions (10). These molecules migrate to the enterocyte luminal or basolateral surfaces to enhance the enterocyte's function as a nonclassical immune cell. Finally, in response to inflammation and inflammatory cell destruction, the enterocyte can increase the transcription of growth factors that results in increased cell turnover and differentiation as a mechanism for accelerated repair of the epithelial barrier.FIG. 4.: Cartoon of the enterocyte response to bacterial and antigenic stimulation. The enterocyte responds with the up-regulation and secretion of cytokines and prostaglandins, the increased expression of polymeric immunoglobulin A (pIgA) receptors, class-II antigens, and cytokine receptors on the luminal and basolateral surfaces, and the release of growth factors and GF-binding proteins to facilitate repair of damaged epithelium. This cartoon underscores the active role of enterocytes in a nonclassical mucosal immune response.Microbial–Epithelial Cross Talk The role that colonizing bacteria play in intestinal mucosal barrier function is a major area of research interest in the field of mucosal immunology (11). Several lines of evidence suggest that bacteria or bacterial toxin attachment to the epithelial surface results in communication between prokaryotic and eukaryotic cells through signal transduction pathways (12). Pathogens can interact with the intestinal surface by using physiologic receptors and their signal transduction pathways for their own purpose, for example, Salmonella typhimurium interacting with the epidermal growth factor receptor or enteropathogenic Escherichia coli inserting their own receptor into the microvillus membrane through a type-III secretion (12). Alternatively, bacteria can use the sparse microvillus and glycocalyx on follicular-associated epithelial cells (M cells) to cross the cellular membrane and infect the intestinal mucosa (for example, Shigella flexneri) (12). In like manner, nonpathogenic bacteria, such as commensal bacteria, play an important role in activating the mucosal immune system. At birth, the newborn leaves a sterile intrauterine environment to enter a contaminated extrauterine world (13). The full-term infant can mount a normal, mature mucosal immune response (Fig. 5) but requires initial intestinal colonization to activate the efferent component of the immunologic mucosal immune defenses. In like manner, the predominant helper T-cell subclass response in utero is a Th2 response (14). With initial colonization, a balance is established between Th2 and Th1 subclasses to prevent the development of allergic food reactions in allergy-prone infants and to establish a Th3 response, which provides oral tolerance to oral protein antigens and thereby prevents autoimmune or other adverse systemic immunologic responses. Two exciting recent observations have suggested specific advantages to microbial–epithelial cross talk in normal mucosal immune function. Investigators from the Pasteur Institute and the Cancer Institute of Lausanne, Switzerland, have reported (15) that luminal bacteria and lamina propria lymphocytes grown with the human cancer cell line, Caco-2, can cause enterocytes to transdifferentiate into follicular epithelial cells. Another recent report (16) has shown that commensal bacteria can down-regulate the NFκB transcription factor that induces the inflammatory cytokine response to endotoxin stimulation by inhibiting ubiquitination of phosphorylated IκB and thereby preventing a free NFκB from translocating into the nucleus to activate cytokine genes. These observations help explain the antiinflammatory effect of commensal bacteria in the human intestine. These seminal reports underscore the importance of a balanced (pathologic versus commensal) colonizing gut flora in preventing gastrointestinal diseases.FIG. 5.: Ontogeny of the human mucosal immune system. The major components of the mucosal immune system are shown. The respective developmental time points for key components are highlighted. Note germinal centers in Peyer patches, immunoglobulin A plasma cells in the lamina propria, and predominant CD8+ intraepithelial lymphocytes do not appear until the perinatal and neonatal periods. The efferent mucosal immune response requires the initial colonization of the gut after birth. Reproduced with permission from Insoft et al. (13).In the past few years, we have gained a better understanding of how bacteria and their cell-wall components or secreted proteins interact with eukaryotic cells to initiate innate immune responses. Using the lipopolysaccharide cell-wall component of Gram-negative bacteria (endotoxin) as a prototypic activator, extensive studies have shown that LPS must first interact with an agglutinin, lipopolysaccharide-binding protein (LBP), in serum and secretions to form a LPS–LBP complex that facilitates its interaction with a lymphocyte cell-membrane–binding protein, CD14, before an innate immune response, such as activation of inflammatory cytokines or phagocytosis (macrophages), can occur. Until recently, the mechanism of this activation with lymphoid cells was unknown because CD14 lacks an intracellular component and is unable to use a signal transduction pathway to activate the cell. Several recent observations (17,18) have identified toll-like receptors (toll receptors exist on Drosophila to regulate development (embryo) and innate immunity (adult fly)) that interact with the LPS–LBP–CD14 complex and use signal transduction molecules identical to those in the IL-1 cytokine receptor family. This interaction activates the NFκB transcription pathway and up-regulates genes that mediate the innate immune response. With this description in mind, a recent publication (19) has shown that toll-like receptors (toll-like receptors 2 and 4) are constitutively expressed on the microvillus surface of human intestinal cancer cell lines (Caco-2, HT-29, T-84). This observation helps explain the mechanism of bacterial toxin–epithelial cross talk. Because more than 10 different toll-like receptors have been cloned, these receptors may help explain the mechanism of positive (commensal) and negative (pathogen) bacterial interactions with the gut. DEVELOPMENTAL DIFFERENCE IN MUCOSAL BARRIER FUNCTION AND GASTROINTESTINAL DISEASE STATES During the past 10 years, our laboratory has studied the development of mucosal barrier function in an attempt to explain the pathogenesis of several age-related infectious gastrointestinal disease states. This will be discussed in the context of specific diseases. We hypothesized that the basis for these age-related diseases related to a developmental difference in the mucosal host defense. The evidence to support this hypothesis will be reviewed. Pathologic Bacterial Colonization of the Newborn Gut We have developed human intestinal in vitro (primary and conditionally immortalized human fetal cell lines, organ cultures, and the micro-Ussing chamber technique) and in vivo (xenograft transplantation of fetal human intestine) (Table 1) models to study mucosal barrier function in the context of neonatal gastrointestinal diseases (20,21). For bacteria to colonize the gut, they must bind to glycoconjugates on microvillus-membrane glycolipids and proteins. Glycosyltransferase enzymes located in intestinal epithelial cells regulate the expression of glycoconjugates. Specific sugars (mannose, fucose, sialic acid) interact with specific bacteria in a lectinlike manner. We have reported that several key glycosyltransferase enzymes are developmentally regulated in the newborn rat and the fetal human intestine and that the immature gut is more likely to be colonized with pathogens than is the more mature gut (Fig. 6) (22,23) (Dai et al., unpublished data). We believe that this immaturity in the regulation of glycoconjugates may account for the pathologic colonization in the human newborn and the increased incidence of bacterial gastroenteritis in this patient population.TABLE 1: Models of human development of intestinal barrier functionFIG. 6.: Cartoon of the crypt–villus axis of epithelial cells in the adult and infant intestine. Because of a developmental regulation of glycosyltransferases and glycoconjugates through the microvillus surface, the infant's microvillus membrane (MVM) is more prone to adherence by pathologic bacteria.Necrotizing Enterocolitis Human premature infants have a high incidence of hemorrhagic necrosis in the distal ileum and proximal colon, leading to perforation and death when nonhuman milk is given as initial oral feedings (a condition known as necrotizing enterocolitis). We hypothesised that this common life-threatening gastrointestinal emergency is caused by an inappropriate initial colonization of the intestine and an accentuated inflammatory response to bacterial stimulation of the immature intestinal epithelium. In addition to the high incidence of pathogenic colonization of the immature gut mentioned above, we have reported recently that fetal human enterocytes from 21-week-old fetuses respond to an exogenous lipopolysaccharide (LPS) and endogenous (IL-1β) inflammatory stimulus with an accentuated (10-to 100-fold) up-regulation of transcription of the enterocyte gene for IL-8 (a polymorphonuclear neutrophil chemokine) and excessive secretion of this cytokine into the lamina propria (24) (Fig. 7). This observation in human in vitro fetal enterocyte cell models suggests that the high incidence of necrotizing enterocolitis in human premature infants is caused by an inappropriate, accentuated inflammatory response to colonizing pathologic enteric flora.FIG. 7.: Depiction of interleukin-8 (IL-8) secretion (A) and IL-8 mRNA induction (B) in fetal and infant intestinal organ cultures in response to LPS (50 μg/mL) or media alone as a control. Secreted IL-8 is expressed as ng/mg of total tissue protein. Interleukin-8 mRNA was quantified by densitometry and normalized to the relative level of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA. The results are given as means + SEM. The number of independent experiments for each data point ranged from three to four. Reproduced with permission from Nanthakumar et al. (24).Toxigenic Diarrhea Several exotoxin-induced secretory diarrheas occur more commonly in young infants (25). To determine whether the increased incidence of these age-related diseases was associated with immaturity in intestinal mucosal barrier function, our laboratory and several others have studied the mechanism of exotoxin-induced secretory diarrhea in developmental animal models and in human fetal intestinal models. In certain instances, the expression of exotoxin receptors is up-regulated in the developing intestine. This is true of diarrhea associated with stable E. coli toxin. Receptors are overexpressed in the immature gut, in which the incidence of diarrhea is highest, and decreased in the mature gut, in which the incidence is decreased considerably (26). We have examined cholera toxin and E. coli labile toxin, toxins that use the same receptor, the GM1 ganglioside. Although receptor number affinity slightly increases in the immature intestine animal models (27,28), the slight increase does not account for the strikingly excessive chloride secretion by immature intestinal epithelial cells exposed to toxin compared with the same dose of toxin in the mature intestine (29). Therefore, we have looked for postreceptor, developmentally regulated steps that may account for this excessive secretory response. We noted that in preweaned compared with weaned rats, adenyl cyclase activity increased in a dose-response manner to cholera toxin (30). Subsequently, we noted that the excessive adenyl cyclase response in the immature intestine was caused by an overexpression of the alpha component of the GTP stimulatory signal transduction molecule, which activates adenyl cyclase in response to ribosylation by the active component of CT after interaction with the receptor (31). In some instances, because of a developmental delay in the expression of toxin receptors (shiga toxin and clostridial A toxin), secretory/inflammatory diarrhea does not occur during the newborn period (32). In these instances, developmental differences in the intestinal response to toxins can be beneficial to the newborn infant. SUMMARY AND CONCLUSIONS In this review, we have discussed the various levels of understanding the mucosal barrier function during the past 25 years. Initially, the intestinal mucosal barrier, particularly the intestinal epithelium, was considered simply a physical barrier to antigens/microbial penetration. Release of polymeric IgA in response to antigenic stimulation was considered the principal immunologic protective function. Nonspecific intestinal functions, such as gastric acidity, mucus release, proteolysis, and peristalsis, were considered adjunctive protective functions separate from immunologic function. However, several seminal observations have shown that these nonspecific intestinal functions are immunologic control and function in a to prevent and systemic immunologic the gut epithelium has been noted to actively participate in the mucosal immune response through increased transcription and release of immunologic surface molecules, and growth factors in response to luminal stimulation. between bacteria and the epithelial surface in response to pathologic and commensal flora has been a major area of Colonization of the intestine is a to normal mucosal immune function. Finally, we have considered and inappropriate epithelial responses to bacteria and their toxins as the basis for age-related gastrointestinal diseases such as necrotizing enterocolitis and