Beth A. McCormick^1,2, Samuel I. Miller³, and James L. Madara^1,2.

¹ Departments of Pathology, Brigham and Women's Hospital, and Harvard Medical School, Boston, MA 02115

² The Harvard Digestive Diseases Center

³ Departments of Medicine and Microbiology, University of Washington, Seattle, WA 98195.

Received 06/19/96; Accepted 07/09/96; On-line 08/01/96

4. INFLAMMATORY RESPONSES ELICITED BY SALMONELLA-INTESTINAL EPITHELIAL INTERACTIONS

4.1 Salmonella coordinates mucosal inflammatory responses

Despite progress made in understanding the mechanisms of secretory diarrhea produced by bacterial toxins, such as cholera toxin, how bacterial pathogens such as salmonellae cause gastroenteritis is poorly understood. It has long been recognized that attachment of non-typhoidal Salmonella serotypes such as S. typhimurium to the intestinal epithelium provoke an intense intestinal inflammatory response, consisting largely of neutrophil (polymorphonuclear leukocyte (PMN)) migration across the epithelial lining of the intestine (7). This inflammatory event manifests itself as epithelial dysfunction, namely, diarrhea (7-10, 64, 89). The details of how such S. typhimurium-intestinal epithelial contacts evoke the classical histological legion of neutrophil transepithelial migration are not well characterized (7, 9-10). However, while it is clear that transepithelial migration of neutrophils occurs early after Salmonella-epithelial contact (7), and well after the epithelium loses its structural integrity (8), the mechanisms and cell types responsible for coordinating mucosal inflammatory responses to such pathogens remain largely obscure. Evidence is emerging, however, that bacterial binding to eucaryotic cells can influence the program of transcriptional regulation for synthesis of biologically important eucaryotic products. For example, Interleukin-6 (IL-6) production is stimulated by the binding of adherent E. coli to bladder or kidney epithelial cells (90), and LPS (lipopolysacchride) has been shown to stimulate tumor necrosis factor (TNF), Interleukin-1 (IL-1), IL-6, and Interleukin-8 (IL-8) production in a host of cell types including monocytes, fibroblasts, and endothelial cells (91-93). IL-8 is of particular importance since unlike the other cytokines listed, it is a potent PMN chemotaxin when present in a gradient (94). Moreover, a recent report (95) modeling urinary tract infections (UTIs) also provided direct evidence that urinary epithelia cells exposed to E. coli secrete IL-8, and such observations fit well with studies indicating that IL-8 can be recovered in the urine of patients suffering with UTI. Taken together, such observations imply that contact between the bacterial outer membrane and the cell apical membrane results in the generation of a signal(s) which may be important for the initiation and amplification of the mucosal inflammatory response.

4.2 Salmonella contact with the intestinal epithelium generates signals important for the initiation and amplification of the mucosal inflammatory response

How might such transepithelial signaling of underlying inflammatory responses occur? Recent studies show that S. typhimurium contact with the apical pole of intestinal epithelial cells generates signal(s) which may be responsible for directing the trafficking of neutrophils across the intestinal epithelium (29, 96-99). The transepithelial migration of neutrophils in response to luminal pathogens necessarily involves movement through several anatomic compartments, each with their own complexities: (a) the well recognized rolling, firm adhesion, and subsequent emigration of neutrophils from the microvasculature (100-107); (b) subsequent migration of neutrophils across the lamina propria and into a subepithelial position; and (c) transepithelial migration. While the mechanisms driving these responses have only recently attracted attention, it is clear that bacterial binding to epithelial cells can influence the production of important regulators of inflammation. For example, McCormick et al. (99), has previously demonstrated using in vitro models of intestinal inflammation, that apical attachment of S. typhimurium to intestinal epithelial monolayers specifically stimulates physiologically directed neutrophil transepithelial migration. The signals responsible for orchestration of this response do not utilize the neutrophil n-formyl peptide receptor directed migration - the best understood receptor-mediated pathway from directing neutrophils to a bacterial target (99). However, among the events stimulated by such pathogen and host interactions is the release of chemotaxins which might guide neutrophils to the site of bacterial-epithelial contact (96-99). For example, S. typhimurium-intestinal epithelial cell interactions induce the epithelial synthesis and basolateral release of the potent neutrophil chemotactic peptide IL-8 (96, 98-99). The mechanisms responsible for such basolateral IL-8 secretion are not well characterized. Eckmann et al., (96), however, has suggested that IL-8 secretion elicited by S. typhimurium contact with the epithelial cells may require signaling associated with cell entry. In contrast, recent evidence indicates that such IL-8 stimulated induction by Salmonella may be more complex. For example, current evidence indicates Salmonella strains or serotype-related differences in the ability of salmonellae to induce diffuse enteritis in human does not correlate well with the ability of these organisms to be internalized by intestinal epithelial cells. Yet, the ability of Salmonella interactions with the apical pole of intestinal epithelial cells to elicit transepithelial signaling to neutrophils correlates well with Salmonella serotyes which elicit diffuse enteritis in humans (29). Thus, such evidence strongly suggests that the ability of Salmonellae to elicit transepithelial signaling to neutrophils is a key virulence mechanism underlying Salmonella-elicited enteritis.

Based on these observations, it is becoming increasing clear that interactions between intestinal epithelial cells and S. typhimurium may play a key role in orchestrating the inflammatory response. This is further exemplified by studies performed by Jung et al. (97) who demonstrate that in response to bacterial invasion of a variety of human colon epithelial cell lines, a specific array of four proinflammatory cytokines (IL-8, monocyte chemotactic protein-1(MCP-1), TNF alpha, and GM-CSF), was found to be coordinately expressed and upregulated in human colon epithelial cell ines (97). The coordinate expression of these proinflammatory cytokines seems to be a general property of human colon epithelial cells since freshly isolated human colon epithelial cells had identical responses (55), and suggests that such cytokine production may play an essential role in intercellular communication by delivering signals which influence the target cells upon which they act.

While each of these cytokines plays a critical role in the initiation and amplification of the inflammatory response, only IL-8 acts as a potent neutrophil chemoattractant. Thus, since neutrophils must initially emigrate from the microvasculature to the subepithelial compartment (47, 104, 107) it was originally hypothesized (22, 99) that a potential role for such basolateral IL-8 secretion may be in the recruitment of neutrophils through the epithelial matrix to the subepithelial space, rather than in directing the final movement of neutrophils across the intestinal epithelium. One unique aspect of the intestinal mucosa which might require tandem signaling events for this process is the presence of a vascular countercurrent arrangement in the subepithelial compartment (108). As happens for absorbed solutes, this countercurrent phenomenon may distort transepithelial solute gradients. For example, perfusion of mammalian intestinal loops in vivo with solutions containing fMLP was previously found to induce neutrophil attachment to endothelial cells and structurally defined endothelial activation, but failed to elicit directed migration across the lamina propria (Madara, unpublished observations), suggesting that directed migration may require a more stable gradient than that afforded by the usual soluble signals. For example, once present in inflamed tissue, IL-8 is likely to retain its biological activity for several hours, as shown by local intradermal administration in animals and humans (109-111). In contrast to IL-8, chemokines such as fMLP or LTB4 are degraded rapidly by oxidation or hydrolysis (112). Thus, gradients of IL-8 formed across the lamina propria matrix would likely be relatively resistant to the distorting effects of the complex solvent flow patterns which exist in this microenvironment and could serve to bring PMN into the subepithelial space. Most importantly for sites like the subepithelial matrix of intestinal mucosa where volume flow is extremely high, IL-8, due to its highly cationic nature, binds avidly to glycosaminoglycans of the tissue matrix (113), thus making such bound IL-8 gradients particularly resistant to sweeping away affects of fluid flow.

Consistent with this notion, recent evidence indicates that biophysically confluent T84 cell monolayers apically colonized by S. typhimurium resulted in the imprinting of a neutrophil chemotactic signal on the underlying epithelial-derived matrix which was primarily due to the basolateral secretion of the C-X-C- family member, IL-8 (98). Such studies provided evidence to substantiate the notion that basolateral IL-8 secretion may act to guide neutrophils through the matrix to the subepithelial space. For example, recent investigations (98) indicate that when underlying matrices were isolated from biophysically confluent polarized monolayers of the human intestinal epithelial cell line T84, they failed to support substantial transmatrix migration of PMN unless an exogenous chemotactic gradient was imposed. However, such matrices isolated from confluent monolayers apically colonized with S. typhimurium, supported spontaneous transmatrix migration of PMN. Such chemotactic imprinting of underlying matrices was resistant to volume wash and as also paralleled by secretion of the known matrix binding chemokine IL-8. Moreover, such chemotactic imprinting of the matrix underlying S. Typhimurium colonized monolayers was found to be independent on epithelial protein synthesis, was directional implying the existence of a basolateral-driven gradient, and was neutralized by antibodies either to IL-8 or to the IL-8 receptor on PMN. Even an avirulent S. typhimurium strain, PhoP^c, which attaches to epithelial cells as efficiently as wild-type S. typhimurium, but fails to induce basolateral secretion of IL-8, failed to imprint matrices. Together, these observations clearly indicate that the epithelial surface can respond to the presence of a lumenal pathogen and subsequently imprint the subepithelial matrix with retained IL-8 gradients sufficient to resist washout effects of the volume flow which normally traverse this compartment. The impact of such events may have substantial importance in assisting movement of neutrophils to the subepithelial space. Additionally, it has been noted in patients with cystic fibrosis that neutrophil elastase can induce IL-8 gene expression in respiratory epithelia (114), and inhibition of IL-8 gene expression by aerosolized secretory leukoprotease inhibitor suppresses both IL-8 secretion and neutrophil infiltration of this epithelium in this disorder (115). Such data further support the notion that the primary role for basolateral secretion of IL-8 by the intestinal, and likely other epithelia, is recruitment of PMN through the matrix to the subepithelial space, rather than directing the final movement of PMN across the epithelium. It is now speculated that IL-8 may act in concert with a transcellular chemotactic factor which directs neutrophil migration across the intestinal epithelium (98). Figure 2 illustrates the current paradigm of neutrophil emigration in response to apical epithelial attachment of S. typhimurium. While there is now substantial evidence that the epithelial cells themselves may play a proactive role in organizing/initiating such inflammatory responses (95-97, 99), a recent precedent has also been established that indicates that products, other than n-formyl peptides, from enteric bacteria might play a more complex role in regulating the activity of the mucosal immune system. Products of E. coli were shown to regulate lymphocyte activation and cytokine production, and suggest that these products may have important influences in modifying gastrointestinal immune responses to enteric bacterial infection (116). Additionally, cell bound components of Helicobacter pylori (gastric inflammatory response) were shown to release factors that are chemotactic for neutrophils and/or monocytes (117-122). One study showed that the N-terminal end of the large subunit of H. pylori urease was chemotactic for neutrophils (120), and other, as yet identified, soluble chemotactic factors of H. pylori have also been described (119, 122-123).

Figure 2. Multistep model of transepithelial of neutrophils across the intestinal mucosa in response to apically attached S. typhimurium. Such neutrophil transmigration in response to luminal pathogens involves movement through several anatomic compartments, each with their own complexities. Step 1: Initially, neutrophils must migrate from the microvasculature to the subepithelial lamina propria. This process is best understood in terms of the molecular interactions that correspond to the initial tethering (selectin mediated) and subsequent firm attachment (ß2 integrin mediated) of neutrophils to the endothelial surface. Step 2: IL-8 is released from the basolateral aspect of epithelial cells in response to adherent S. typhimurium. IL-8 imprints the subepithelial matrix with retained haptotactic gradients sufficient to resist washout effects of the volume that normally traverses this compartment. Thus, the primary role for basolateral secretion of IL-8 is the recruitment of neutrophils through the matrix to the subepithelial space, rather than directing the final movement of neutrophils across the epithelium. Step 3: Recent evidence indicates that an apically secreted soluble factor with physical characteristics unlike IL-8 (B. A. McCormick and J. L. Madara), acts in concert with IL-8, and appears responsible for driving this final in neutrophil transepithelial migration (modified from reference 99).