| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Unité de Pathogénie Microbienne Moléculaire, Institut National de la Santé et de la Recherche Médicale U389, Institut Pasteur, Paris, France
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
B system through Nod 1 (9), a peptidoglycan sensor (10, 11), thereby promoting massive expression of IL-8 (12). The predominance of IL-8 is confirmed by global transcriptional profiling of Shigella-infected human colonic Caco-2 cells, showing that the amount of IL-8 mRNAs is increased by 300-fold following 2 h of infection by an invasive strain, compared with infection with a noninvasive strain, the IL-8 gene being indeed, by far, the most transcribed (13). In the mouse, the genes encoding IL-8 and its receptor CXCR1 have been deleted, probably when a gene duplication event occurred in a common ancestor of rat and mouse, because these genes are lacking in the muroid lineage (14). However, the murine MIP-2, or the murine keratinocyte-derived protein chemokine KC seem to be the functional homologues of IL-8. They share the extracellular loop reactive+ motif (Glu-Leu-Arg) for the chemoattractant properties of IL-8 family members (15, 16). They also share LPS inducibility (14). Despite the lack of a gene coding for IL-8, the mouse species possesses a receptor homologous to human CXCR2 that is able to mediate neutrophil chemotaxis in response to human IL-8, in addition to MIP-2 and KC (16, 17). Therefore, we postulated that the absence of IL-8 production in mice may prevent Shigella-induced colitis. This hypothesis has been tested by adding exogenous recombinant human IL-8 (rhIL-8) in the course of an experimental infection by Shigella, and by comparing its effects to the addition of other neutrophil chemoattractants, such as the murine CXC chemokines recombinant murine (rm)MIP-2 and rmKC (17, 18).
Thus, we present in this study a new adult model of colonic infection by Shigella in which a double intraluminal addition of rhIL-8, concomitant to invasive S. flexneri, induces mucosal lesions that, by all standards, appear similar to the human disease. This study confirms that the lack of IL-8 expression is either the major or one of the major factors explaining the refractory state of mice to develop acute invasive colitis in the presence of Shigella.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Male BALB/c mice (Centre d’Elevage R. Janvier, France), aged 5–6 wk, were fasted for 24 h with water ad libitum. The next day, they were anesthetized with xylazine 12% (20 mg/kg) ketamine 500 (45 mg/kg) (both from Sigma-Aldrich, St. Louis, MO), and groups of five mice were challenged by intracolonic (i.c.) route, inserting a cannula of 40 mm in the rectum, up to the left angle of the colon and instillating either saline (endotoxin-free 0.9% NaCl), or rhIL-8 (AbCys, Paris, France), or rmKC or rmMIP-2 (both from AbCys), or a suspension of Shigella. rhIL-8 was used at doses of 1, 2, and 4 µg, then at the fixed dose of 2 µg, as well as for rmMIP-2 and rmKC, and adjusted to 50 µl in saline (19, 20). The i.c. route was chosen to overcome the drawbacks of the oral route (6, 21), and to deliver a calibrated dose of Shigella directly in contact with the colonic epithelium. Two stains of Shigella were used: either the noninvasive mutant mxiD that has lost the capacity to express the type III secretory apparatus that allows Shigella to deliver its effector proteins of invasion to the intestinal epithelium (22), or the invasive strain M90T (serotype 5) containing the 213-kb virulence plasmid that encodes the invasive phenotype (23). The inoculum of bacteria was obtained by overnight culture, and an aliquot was grown in 400 ml of Tripticase Soy Broth BBL medium (BD Biosciences, Le pont de Claix, France), up to midexponential growth phase (OD600 = 0.8). Bacteria were then concentrated by centrifugation, and kept at 37°C until i.c. instillation (24), and their number was determined by OD and by plating dilutions on Tripticase Soy Broth BBL agar, to perform inoculums at 109 up to 1011 CFU/ml. OD = 1 corresponded to 5.108 CFU, with a SD of 3%. Groups of five mice were inoculated once, either with 0.1 ml of the culture medium as control, or with 0.1 ml of a suspension of 107 up to 1011 CFU/ml invasive Shigella, or of the noninvasive strain mxiD. Following the selection of inocula of 109 and 1010 CFU/mouse, experiments were conducted with or without rhIL-8 instilled together with strain M90T or control mxiD. Similar experiments were performed with rmKC or rmMIP-2 (17, 18). A second intraluminal administration of rhIL-8 (or rmKC, or rmMIP-2), concomitant to inoculation of S. flexneri, was also performed on groups of five mice, 18 h following the first infection (25). Animals were then sacrificed at 24 h and their colons were collected. The kinetics of Shigella infection of the colon were performed using an inoculum of 109 CFU/mouse, by collecting samples of colon at 1, 3, 6, and 24 h following the initial infection. The whole colon was observed directly for the presence of blood, then its length was measured, and a first portion (5 mm) was cut at 40 mm from the anal orifice, a distance corresponding to the left angle, opened longitudinally, and observed to provide semiquantitative evaluation of the lesions expressed as scores; the following 5 mm of the descending colon were also retained for histology and scores after paraffin inclusion, and finally, another aliquot was kept for myeloperoxidase (MPO) and RT-PCR.
Evaluation of the lesions by macroscopic and histopathologic examination: Wallace and Ameho scores
The macroscopic lesions of the colon considered encompassed thickness and length of the colon, presence of blood, and edema. Stereomicroscopic examination of fresh colon longitudinally opened confirmed the previous observations and allowed an immediate evaluation of scores according to the Wallace criteria (26), based on the following data: hyperemia, edema, increase in thickness, number, and size of ulcerations, and presence of inflammation or purulent material.
Colons were dissected and prepared for histology: after washing in saline, they were inflated with 50% OCT (Sakura Finetck, Torrance, CA), and frozen, then sectioned in 5-µm slices to perform immunocytochemistry (ICC). For paraffin inclusion, colons were soaked in 10% formaldehyde in PBS overnight at 4°C, processed in paraffin wax, then positioned for longitudinal or ring section. Five-micrometer sections were stained with H&E, or Diff Quick (Baxter Merz & Dade, Düdingen, Switzerland), for identification of recruited cells in the colon.
Histopathological evaluation of colonic sections encompassed the degree of erosion of the epithelium alterations, of crypts, nature of the inflammatory infiltrate, and state of the lamina propria. This allowed determination of a score according to the gradation of Ameho (27).
Quantification of tissue MPO by Western blot
An aliquot of colon was homogenized with an Ultraturrax (T25 Janke and Kunkel; IKAR-Labortechnik, Staufen, Germany) for 1 min at 4°C in cold PBS containing Nonidet, and antiproteases (both from Sigma-Aldrich). After centrifugation, MPO was determined by Western blot (5, 28). The protein content was measured by OD at 205, 235, and 280 nm for each sample, then 50–200 µg of each was analyzed by SDS-PAGE (50 µg without precipitation, or 200 µg following ethanol precipitation). The transfer was performed to PVDF and blotting was performed overnight at 4°C, first with a goat anti-MPO (C-16) polyclonal serum (Santa Cruz Biotechnology, Santa Cruz, CA), then with a second polyclonal anti-goat HRP-conjugated serum (Sigma-Aldrich). Revealation was done by ECL-Plus (Amersham Bioscience, Buckinghamshire, U.K.) allowing quantification of chemiluminescence of the 70-kDa band using a Storm apparatus, and the ImageQuant program (Molecular Dynamics, Sunnyvale, CA).
Detection of S. flexneri invasion by ICC
Amplified ICC using first a biotin-conjugated Ab specific for LPS of S. flexneri (which recognizes invasive and noninvasive strains, a kind gift of Dr. A. Phalipon, Institut Pasteur, Paris, France), then a second streptavidin-peroxidase-conjugated Ab (DakoCytomation, Carpinteria, CA) was performed, the latter being revealed by 3-amino-9-ethylcarbazol (Sigma-Aldrich), as described (28, 29).
Cytokine and chemokine transcriptional profiling by quantitative RT-PCR
Portions of colons were isolated and thoroughly washed with saline. Dispersion was performed with an Ultraturrax (T25 Janke and Kunkel, IKAR-Labortechnik) for 40 s in the RNA Total Lysis buffer from the RNeasy Mini kit (Quiagen, Hilden, Germany) used for RNAs extraction. RT-PCR was performed for colon, using specific primers for TNF-
, 5'-ACT GAA CTT CGG GGT GAT CGG TCC; reverse, 5'-GTG GGT GAG GAG CAC GTA GTC G; IL-1
, 5'-TTg ACg gAC CCC AAA AgA Tg; reverse, 5'-AgA Agg TgC TCA TgT CCT CAT; IL-18, 5'-ACT gTA CAA CCg CAg TAA TAC gg; reverse, 5'-gAg TgA ACA TTA CAg ATT TAT CCC; then those for IFN-
(Th-1), IL-4 (30); then MIP-2, 5'-CAg AAT TCA CTT CAg CCT AgC gCC AT; reverse, 5'-gCT CTA gAg TCA gTT AgC CTT gCC TTT g; KC, 5'-CAG CCA CCC GCT CGC TTC TC; reverse, 5'-TCA AGG CAA GCC TCG CGA CCA T; and -actin as a control, as described (28, 29, 30). Standards, when available, were prepared as described (29, 30, 31, 32). The copy number was calculated according to the OD, then the purified DNA was serially diluted to obtain the appropriate standard containing 0–1,000,000 of copies. The copy number of the sample was calculated relatively to the standard, after PCR amplification on the LightCycler System (Roche Molecular Biochemicals, Mannheim, Germany). When standards were not available, quantification was achieved using a Storm and ImageQuant (Molecular Dynamics), as described (28). Results were presented as a ratio of cytokine:-actin copies.
Statistical analysis
All results are presented as means ± SD. Levels of significance were calculated using one-way ANOVA, followed by Scheffe’s test, using the SPSS 6.1 software (SPSS, Chicago, IL) (*, significance between data with a threshold of p < 0.05).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Colonic infection by invasive S. flexneri following single i.c. inoculation: graduation by Wallace and Ameho scores and histology
To evaluate the degree of tissue alteration caused by invasive Shigella, we established scores of inflammation according to the generally used criteria of Wallace and Ameho (26, 27). We present in this study the scores and histopathological features observed following 24 h of infection, and only the salient features of other time points.
Control mice infected with the noninvasive strain mxiD, at different time points after infection (i.e., 1, 2, 4, 6, and 24 h) with inocula of 109 or 1010 CFU, displayed normal macroscopic (no thickening of the colon, no apparent blood) and stereomicroscopic aspects on longitudinally opened colons (no thickening of the colonic wall, no blood, no ulceration, and no pus) similar to noninfected colon. The histological aspect was also normal with an intact epithelial lining, regular crypts of an usual size, lack of inflammatory infiltrate, absence of edema and of significant thickening of the lamina propria or muscularis, and normal and regular basal membrane (Fig. 1, A and C). Altogether, this results in low scores according to the Wallace and Ameho criteria (Table I, line 1; scores were of 1 and 2, respectively).
|
|
30% of the samples) following inoculation with 1010 CFU. This was confirmed by stereomicroscopic observation showing no or minimal thickening of the colon wall, hyperemia when blood traces were observed, and some scarce ulcerations of small size following inocula of 109 CFU, and more frequent following inocula of 1010 CFU; the other parameters described above being similar to the controls, as reflected by the scores of Wallace at 24 h (Table I, lines 2–4, column W; scores were of 2, 3, and 5, for 108, 109, and 1010 CFU, respectively, highlighting a dose-dependent response). Histopathological analysis showed a colonic mucosa with moderate alterations, characterized by scarce superficial and localized erosions of the epithelial lining in general associated with bacterial invasion (Fig. 1, B, D, and F, right, at 109 CFU/mouse). In a few mice, more severe alterations could be observed as soon as 3 h after infection (data not shown), with more extended epithelial erosions and even with some degree of crypt cells destruction, particularly with the highest inoculum of 1010 CFU. These erosions were associated with an inflammatory infiltrate of the lamina propria, composed only of lymphocytes and monocytes/macrophages (Fig. 1, D and F, right). Strikingly, PMN did not appear in the inflammatory infiltrate, even following 24, 48, or 72 h of infection. These observations were graded by Ameho scores 24 h after infection with the invasive strain M90T (Table I, lines 2–4, column A). Scores were of 3, 4, and 5, for 108, 109, and 1010 CFU, respectively, highlighting again a dose-dependent response, as also confirmed by cumulated scores (Wallace and Ameho, W+A), which were of 5, 7, and 10, respectively. Labeling for Shigella-LPS and kinetics of bacterial invasion of the tissue
Labeling for Shigella-specific LPS by immunohistochemistry on tissue sections infected by the noninvasive strain mxiD failed to show bacteria in mucosal tissues, neither in the colonic epithelial lining, nor in subepithelial tissues, confirming that no significant bacterial invasion had occurred between 1 and 24 h (Fig. 1E, left). In contrast, 1 h after infection with invasive Shigella, some bacteria could already be observed in submucosal tissues, and at 90 min, a few of them were present at the bottom of crypts in the lamina propria. Following 6 h of infection, bacteria were present in large numbers in eroded areas of the epithelium lining (Fig. 1F, right) and of crypts, where infection promoted tissue alteration and destruction by plages. By contrast, following 24 h of infection, bacteria were no more observed in the epithelium, but were concentrated at the basis of crypts, in submucosal areas, as further described.
Quantification of MPO indicates no neutrophil recruitment following infection with M90T alone
The amount of MPO protein was measured by quantitative Western blot performed upon homogenized colonic tissue samples. It showed no detectable level, regardless of experimental conditions: uninfected tissues, tissues infected with the noninvasive strain mxiD, as well as tissue infected by the invasive strain M90T at all time points (Fig. 2). These results strikingly confirmed that PMN were barely recruited in the inflammatory infiltrate, even in the presence of invasive Shigella. As MPO is accurately correlated with neutrophil accumulation in tissues (5, 28), this marker was subsequently considered of primary importance to monitor PMN infiltration, and select optimal experimental conditions.
|
Having described that the mouse colonic mucosa was permissive to a certain degree of Shigella invasion and able to elicit a mononuclear infiltrate, we established a transcriptional profile of this particular inflammatory response. We selected a battery of cytokines and chemokines genes whose profile of expression is characteristic of the inflammatory profile observed in the mouse lung infection assay (33), as well as in the rectal tissue of human experiencing shigellosis (34).
Our profile included: 1) IL-18 and IFN-, two factors essential to protect mice against Shigella infection (35, 36, 37); 2) IL-1 and TNF-, two proinflammatory cytokines known to promote rupture of the epithelial barrier and bacterial dissemination (35, 36, 38); 3) the cytokine IL-4, reflecting the orientation of the immune innate reaction by Shigella; and 4) MIP-2 and KC, two chemokines acting as chemoattractants for PMN that are thought to compensate for the lack of IL-8 in the mouse. Evolution of the transcriptional pattern of these cytokines and chemokines was studied according to time, although only selected time points are reported in this study in Figs. 3 and 4.
|
|
(Fig. 3IIA).
In contrast, transcripts for TNF- and IL-1 were not expressed by colonic tissues, either not infected, or infected by the noninvasive mutant mxiD. Following infection with the invasive strain M90T, transcripts could already be detected at 1 h and continued to increase at 2 h (TNF-, Fig. 3IIIA; IL-1, Fig. 3IVA). IL-1 transcripts then subsequently decreased, whereas TNF- transcripts kept increasing until 24 h. IL-4 transcripts were strongly induced at 6–24 h after infection with the invasive strain M90T (Fig. 3VA), compared with the noninvasive strain mxiD. Transcripts for MIP-2 and KC were not detected in noninfected tissues or tissues infected by the noninvasive strain mxiD. In contrast, MIP-2 and KC transcripts started to increase 1 h following infection with the invasive strain M90T and kept increasing to peak at 24 h (Fig. 4, IA and IIA, respectively).
Strategy retained to study the effects of rhIL-8 added to Shigella
To study whether the concomitant administration of IL-8 and Shigella was able to induce a profile of infection mimicking the disease observed in humans, rhIL-8 was delivered intraluminally on the epithelial surface of the colon. In one group of mice, rhIL-8 and the Shigella inoculum were instilled concomitantly (invasive strain or noninvasive control). In other groups of mice, a second administration was performed 18 h following the first administration, either of rhIL-8 alone, or of rhIL-8 and Shigella. Colons were collected following the sacrifice of animals at 24 h. MPO was retained as a first criterion for quantifying neutrophil recruitment and deciding upon the best protocol of instillation of bacteria and rhIL-8.
Optimization of the protocol of administration of Shigella and rhIL-8
The amount of MPO expressed in infected tissues was selected as a read out to evaluate the number of PMN in the inflammatory infiltrate. Fig. 2 (column 5) shows that the amount of MPO was not significantly increased in colon extracts at 24 h following rhIL-8 plus M90T instillation, compared with M90T alone (column 4), and to the control associating mxiD and rhIL-8 (columns 1–3). However, when the first i.c. administration of M90T plus IL-8 was followed, 18 h later, by a second instillation of rhIL-8 alone, the amount of MPO measured appeared significantly increased (Fig. 2, column 7), and increased even more when both M90T and rhIL-8 were readministered (Fig. 2, column 9), compared with the amount of MPO in mice receiving the corresponding control mxiD plus rhIL-8 (columns 6 and 8, respectively). Thus, the latter condition encompassing two administrations of M90T plus rhIL-8, 18 h apart, appeared optimal and was selected for additional experiments.
Combination of invasive Shigella and rhIL-8 cause neutrophilic inflammation in infected colons.
Control mice instilled with rhIL-8 (2 µg) alone, once or twice, did not show any significant histological alterations (Fig. 5A), and rhIL-8 instilled at doses up to 1.5–2 µg, together with the noninvasive strain mxiD did not induce significant alteration either, even when the dose was repeated 18 h later (Fig. 5B). This led to low scores of Wallace and Ameho (Table I, lines 6 and 7; scores were of 3 and 4, and 4 and 4, respectively). In contrast, when coadministration of invasive strain M90T and rhIL-8 was performed twice, at 18-h interval, according to the protocol selected, we observed a dramatic increase in colon wall thickening (doubled, or x4), in extent of thickening (up to 20 mm length, exceptionally more, reflecting pancolitis), edema, and blood associated with extensive ulcerations. This translated in increased Wallace scores (Table I, W, lines 8–10: scores were of 7.5, 10, and 12 for 108, 109, and 1010 CFU, respectively), indicating a marked dose-dependent response. Histopathological observation confirmed a dramatic thickening of the colon wall, the presence of erosions (Fig. 5, C and E, compared with Fig. 5B), destruction of the epithelial lining and of the basal membrane, and deep alteration and destruction of the structure of crypts, such as shortening, rounding (Fig. 5C, middle), and some tissue necrosis. Edema was observed, associated with a marked infiltration by numerous neutrophils in the lamina propria, and at the basis of crypts, in addition to other inflammatory cells such as lymphocytes, monocytes/macrophages (Fig. 5, C and D), and inflammation-associated lymphoid (immune cell) aggregate (Fig. 5C, high part). These massive alterations, now characterized by an inflammatory infiltrate dominated by PMN, translated into much higher Ameho scores (Table I, A, lines 8–10). For 108, 109, and 1010 CFU, cumulated scores (W+A) were of 15.5, 20, and 23, respectively.
|
ICC shows a colocalization of invasive Shigella with PMN in samples treated twice with rhIL-8 and M90T
At 6 h following noninvasive Shigella infection, the control 2 x (mxiD + IL-8) showed no labeling (Fig. 6A). In contrast, mice infected with invasive M90T showed a marked LPS labeling at the site where the lining epithelium was completely eroded, and abscesses were observed (Fig. 6B), which promote necrosis, as observed at the basis of crypts (Fig. 6B, bottom).
|
Cytokine and chemokine transcriptional profiles after infection with twice (rhIL-8 plus M90T)
Intraluminal rhIL-8 instillation, induced dose-dependent transcription of some cytokines, such as IL-18, TNF– (Fig. 3B, for both cytokines), and KC and MIP-2 mRNAs (Fig. 4, IC and IIC). Consequently, the synthesis of TNF- following 2 x (rhL8 + M90T), was markedly increased (at 24 h, Fig. 3IIIB), as well as KC and MIP-2 synthesis (Fig. 4, IB and IIB).
rmKC or rmMIP-2 used in the same conditions as IL-8 failed to induce neutrophil recruitment at the site of infection
The amount of MPO protein measured by quantitative Western blot performed upon whole colon homogenates from mice treated by rmKC or rmMIP-2 (2 µg) and invasive S. flexneri, in the same conditions as these set for rhIL-8 (2 x 2 µg of the chemokine), failed to increase (Fig. 7). Higher doses of the chemokines were not suitable because they induced alterations in the control, and 1 x 4 µg did not induce significant PMN recruitment. Histopathological observation confirmed the lack of neutrophils in these experimental conditions (data not shown).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
To bypass the transit of bacteria through the whole gastrointestinal tract, we have chosen the i.c. route to deliver Shigella directly in contact with the colonic epithelium, thus avoiding unwanted interferences that may severely reduce the bacterial inoculum (6, 21). This model allowed us to demonstrate that in BALB/cJ mice, the colonic mucosa can be invaded by virulent Shigella, indicating that the refractory state does not reside in an inability of bacteria to penetrate epithelial cells. Furthermore, invading bacteria elicit a significant degree of cytotoxicity to the epithelium and recruit an inflammatory infiltrate that, however, remains modest and exclusively composed of mononuclear cells. The striking feature of this infectious pattern is the lack of a PMN infiltrate, thus the lack of massive rupture and subsequent inflammatory destruction of the intestinal epithelium that is characteristic of the experimental lesions observed in the rabbit ileal loops (7) and the clinical lesions observed in humans (39). The inability of the mouse to promote marked PMN inflammation in response to Shigella invasion suggested that the murine defect resided more in the sensory and effector program of the epithelium. Preliminary experiments in which we tried to prime the colonic epithelium with a chemical agent, such as ethanol, were unsuccessful. Thus, we decided to investigate the hypothesis that the lack of IL-8 in mice may largely account for the absence of further development of the infectious process (7, 14).
IL-8 in humans is produced by numerous cell types, particularly monocytes/macrophages, epithelial cells, neutrophils, fibroblasts, and endothelial cells upon infection by bacteria or stimulation by bacterial products (i.e., pathogen-associated molecular patterns), or stimulated by cytokines such as IL-1 and TNF- (40, 41). IL-8 is mainly active on neutrophils, promoting their recruitment and also their strong activation characterized by activation of the leukotriene pathway (42), by the release of their granular content (i.e., elastase (43), lactoferrin (44), and MPO), and their increased adherence to endothelial cells, as well as NO activation (19, 40, 41, 42, 45). IL-8 is also a chemoattractant for other cell types such as basophils, T lymphocytes, and NK cells, and also enhances permeability of endothelial cells (19, 40, 45). In ulcerative colitis whose histopathological lesions strongly resemble those observed in shigellosis, the level of IL-8 expression correlates with the severity of the disease (46, 47), particularly with neutrophilic inflammation and mucosal destruction.
Interestingly, when rhIL-8 was administered to the mouse, together with invasive Shigella, the disease gained a profile similar to human shigellosis and to the disease observed in animal models such as the rabbit ligated ileal loop model (7). The whole histopathological pattern became characteristic of bacillary dysentery, with a marked rupture of the epithelial barrier, massive bacterial invasion of the epithelium and subjacent tissues, and a marked inflammation again characterized by a dominant PMN infiltrate, but also associated to mononuclear cells and to dramatic tissue injury (deep erosion, crypt alterations, necrosis, edema).
Indeed, bacteria appeared massively present in areas of epithelial destruction in the immediately subjacent lamina propria, but also at the basis of the crypts, 24 h following inoculation. The reason for this localization is unclear, although it has already been observed in guinea pigs (8). One possibility is that bacteria may have penetrated through the crypt epithelium or via M cells (48). Alternatively, Shigella may have been transported by migrating cells from the colonic lumen to the lamina propria, such as CD18-expressing phagocytes (49), or by dendritic cells recruited at the site of epithelial colonization, that may directly perform uptake of bacteria across the epithelial lining. Indeed, these latter cells may open the tight junctions between epithelial cells, send dendrites outside the epithelium to sample bacteria from the lumen, and subsequently carry them as they migrate toward the basal lamina propria on their way to lymphoid organs (50).
In any event, Shigella were also associated in these areas with a PMN infiltrate, which may be due to a chemotactic gradient formed through the crypt epithelium by intraluminal IL-8, or caused by the bacteria themselves, following local activation of MIP-2 or KC production and/or production of microbial chemoattractants for PMN, such as fMLP, formylnorleucyl-leucyl-phenylalanine, or bacterial peptides (51, 52, 53, 54).
This situation is characteristic of a system in which bacteria are likely to make use of PMN to disrupt the epithelial lining and invade the epithelium and subjacent tissues. It agrees with a model of loop amplification (i.e., vicious circle) previously described, according to which bacterial invasion, following strong signaling for IL-8, promotes recruitment of a PMN infiltrate that leads to the rupture of the epithelium, its destruction promoting further Shigella invasion (55). In this model, the PMN infiltrate eventually leads to bacterial elimination through efficient generic antibacterial mechanisms, including production of lactoferrin (44), and more specific mechanisms, such as the degradation of Shigella virulence effectors by released elastase that disarms the pathogen (43). Thus, in the IL-8-mediated loop, massive tissue destruction appears to be the cost of final bacterial eradication, thereby dictating the pathologic profile of shigellosis.
In addition to mimicking the histopathological profile of shigellosis, addition of IL-8 also generated a relevant pattern of cytokine and chemokine transcription that agrees with the known mechanisms of innate response against Shigella. Interestingly, Shigella appeared to quickly down-regulate the expression of IL-18 and IFN-, which are critical in mediating the protection of mice and the growth of Shigella in tissues at the initial stage of the innate immune response (35, 36, 37, 56). Absence of IL-18 has been shown to severely impair the capacity of infected mice to eradicate a Shigella or an Escherichia coli inoculum (38, 57), and mice devoid of the IFN- receptor appear highly susceptible to Shigella infection (37). Therefore, it would make sense that Shigella, quickly following epithelial invasion, turn down IL-18 expression, thereby down-regulating expression of IFN-, particularly by NK cells. This is consistent with the observation that, in the rectal mucosa of Shigella-infected patients, IFN--producing cells appear dramatically reduced in number (34), thus adding relevance to the current mouse model. Conversely, Shigella invasion in mice led to the development of a clear proinflammatory profile characterized by the induction of genes coding for IL-1 and TNF-. Both cytokines participate in the previously described amplification loop through their ability to induce disruption of the epithelial barrier (35, 36, 38, 51, 58, 59). In particular, these cytokines promote inflammation via the expression of adhesion molecules and chemokines favoring recruitment of cells, such as PMN and NK cells (60), thereby enhancing tissue injury. Finally, the induction of IL-4 transcription may reflect a quickly established tip in the Th1/Th2 balance of the mucosal response, in agreement with the dominance of the humoral adaptative protection against Shigella infection (61). MIP-2 and KC, the IL-8 homologues in the mouse, were also induced following Shigella infection, but were still unable to promote significant PMN recruitment to a point allowing mice to develop colitis. The reasons for this are unclear. Perhaps the invasive process by Shigella is not sufficient in the mouse epithelium to elicit expression of MIP-2 and KC at the critical level required for activation of CXCR2 and signal transduction, allowing to recruit a sufficient PMN infiltrate. Alternatively, MIP-2 and KC may be not as potent chemoattractants as IL-8 in mouse (62, 63). Finally, expression of the three chemokines in conjunction might be required for CXCR2 activation then effective PMN recruitment, as suggested using transgenic IL-8 mice and anti-MIP-2 or anti-KC Abs (64). Thus, IL-8 is likely to remain the predominant effector, because ectopic addition of either rmMIP-2 or rmKC, both chemotactic for PMN in vitro and in vivo (65), but with various efficiencies (65, 66), failed to promote PMN recruitment in the same conditions of those defined for rhIL-8.
In conclusion, the model seems to adequately reflect the human disease, with some limitations, among which the necessity to administer two doses of a high bacterial inoculum (i.e., 109 CFU compared with the 102-103 CFU required to cause disease in humans), in addition to IL-8. This possibly reflects the short half-life of the rhIL-8 signal in the colonic lumen, providing a very short-lived luminal-mucosal gradient, insufficient to promote PMN recruitment. This might also reflect weak efficiency of IL-8, possibly due to a lower affinity of the ligand IL-8-CXCR2 receptor system from nonhomologous species (human-mouse), or to the lack of CXCR1-receptor. It is likely that more sustained expression of IL-8 in the presence of Shigella invasion of the colonic epithelium would achieve a more prolonged gradient, allowing use of a lower Shigella inoculum.
In addition, the experimental procedure might also shortcut a crucial signaling function of the epithelial barrier. For instance, in the case of Salmonella infection, it has been shown that a chemoattractant gradient (i.e., pathogen-elicited epithelial chemoattractant activity) was created apico basally following apical interaction of the invasive microorganism with intestinal epithelial cells (67). This gradient has been shown to account for the basoapical transmigration of the PMN, themselves recruited to the basal side of the epithelium by the basomucosal gradient formed by IL-8 produced by epithelial cells (67, 68). If Shigella were unable to cause production of an equivalent pathogen-elicited epithelial chemoattractant activity gradient in mice, compared with humans, it may explain the partial effect observed in mice by adding IL-8 apically. A more sustained level of IL-8 expression with the generation of a gradient allowing PMN recruitment at the site of infection is needed, suggesting use of transgenic mice whose IL-8 expression would be induced following Shigella infection, in the colonic epithelium. Despite these limitations, the current model constitutes a major step forward in the development of a murine model of bacillary dysentery.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
1 Address correspondence and reprint requests to Dr. Monique Singer, Unité de Pathogénie Microbienne Moléculaire, Institut National de la Santé et de la Recherche Médicale U389, Institut Pasteur, 25 rue du Dr Roux, 75015, Paris, France. E-mail address: msinger{at}pasteur.fr 
2 P.J.S. is a Howard Hughes Medical Institute Scholar.
3 Abbreviations used in this paper: PMN, polymorphonuclear leukocytes; MPO, myeloperoxidase; ICC, immunocytochemistry; i.c., intracolonic; rh, recombinant human; KC, keratinocyte-derived cytokine; rm, recombinant murine.
Received for publication March 31, 2004. Accepted for publication July 2, 2004.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
(PPAR) heterodimer: a basis for new therapeutic strategies. J. Exp. Med. 193:827.B through a lipopolysaccharide-dependent innate intracellular response and leads to IL-8 expression in epithelial cells. J. Immunol. 165:903.-mediated oral tolerance. J. Exp. Med. 183:2605. interferon production is associated with acute shigellosis. Infect. Immun. 65:5338.[Abstract]
and IL-18 are essential for Shigella flexneri-induced inflammation. Immunity 12:581.[Medline]
, TNF) in the induction of epithelial cell death and renewal. Eur. J. Immunol. 28:730.[Medline]
. Eur. J. Immunol. 30:3057.[Medline]
. J. Immunol. 161:1204.This article has been cited by other articles:
|
|
 |
|
  D. Borenshtein, K. A. Schlieper, B. H. Rickman, J. M. Chapman, C. W. Schweinfest, J. G. Fox, and D. B. Schauer Decreased Expression of Colonic Slc26a3 and Carbonic Anhydrase IV as a Cause of Fatal Infectious Diarrhea in Mice Infect. Immun., September 1, 2009; 77(9): 3639 - 3650. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Ohkusa, T. Yoshida, N. Sato, S. Watanabe, H. Tajiri, and I. Okayasu Commensal bacteria can enter colonic epithelial cells and induce proinflammatory cytokine secretion: a possible pathogenic mechanism of ulcerative colitis J. Med. Microbiol., May 1, 2009; 58(5): 535 - 545. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. L. Mumy, J. D. Bien, M. A. Pazos, K. Gronert, B. P. Hurley, and B. A. McCormick Distinct Isoforms of Phospholipase A2 Mediate the Ability of Salmonella enterica Serotype Typhimurium and Shigella flexneri To Induce the Transepithelial Migration of Neutrophils Infect. Immun., August 1, 2008; 76(8): 3614 - 3627. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M.-I. Fernandez, B. Regnault, C. Mulet, M. Tanguy, P. Jay, P. J. Sansonetti, and T. Pedron Maturation of Paneth Cells Induces the Refractory State of Newborn Mice to Shigella Infection J. Immunol., April 1, 2008; 180(7): 4924 - 4930. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Chamekh, A. Phalipon, R. Quertainmont, I. Salmon, P. Sansonetti, and A. Allaoui Delivery of Biologically Active Anti-Inflammatory Cytokines IL-10 and IL-1ra In Vivo by the Shigella Type III Secretion Apparatus J. Immunol., March 15, 2008; 180(6): 4292 - 4298. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. L. Mumy, X. Chen, C. P. Kelly, and B. A. McCormick Saccharomyces boulardii interferes with Shigella pathogenesis by postinvasion signaling events Am J Physiol Gastrointest Liver Physiol, March 1, 2008; 294(3): G599 - G609. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. N. Schroeder and H. Hilbi Molecular Pathogenesis of Shigella spp.: Controlling Host Cell Signaling, Invasion, and Death by Type III Secretion Clin. Microbiol. Rev., January 1, 2008; 21(1): 134 - 156. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Shimizu, Y. Kida, and K. Kuwano Mycoplasma pneumoniae-Derived Lipopeptides Induce Acute Inflammatory Responses in the Lungs of Mice Infect. Immun., January 1, 2008; 76(1): 270 - 277. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Simon, D. M. Heithoff, M. J. Mahan, and C. E. Samuel Comparison of Tissue-Selective Proinflammatory Gene Induction in Mice Infected with Wild-Type, DNA Adenine Methylase-Deficient, and Flagellin-Deficient Salmonella enterica Infect. Immun., December 1, 2007; 75(12): 5627 - 5639. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. M. Williams, Y.-T. Chen, T. M. Andermann, J. E. Carter, D. J. McGee, and K. M. Ottemann Helicobacter pylori Chemotaxis Modulates Inflammation and Bacterium-Gastric Epithelium Interactions in Infected Mice Infect. Immun., August 1, 2007; 75(8): 3747 - 3757. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D.-H. Shim, T. Suzuki, S.-Y. Chang, S.-M. Park, P. J. Sansonetti, C. Sasakawa, and M.-N. Kweon New Animal Model of Shigellosis in the Guinea Pig: Its Usefulness for Protective Efficacy Studies J. Immunol., February 15, 2007; 178(4): 2476 - 2482. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. R. Bracke, A. I. D'hulst, T. Maes, K. B. Moerloose, I. K. Demedts, S. Lebecque, G. F. Joos, and G. G. Brusselle Cigarette Smoke-Induced Pulmonary Inflammation and Emphysema Are Attenuated in CCR6-Deficient Mice J. Immunol., October 1, 2006; 177(7): 4350 - 4359. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. A. Burton, A. M. Pendergast, and A. Aballay The Caenorhabditis elegans ABL-1 Tyrosine Kinase Is Required for Shigella flexneri Pathogenesis. Appl. Envir. Microbiol., July 1, 2006; 72(7): 5043 - 5051. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. P. Allen, M. M. Randolph, and J. M. Fleckenstein Importance of Heat-Labile Enterotoxin in Colonization of the Adult Mouse Small Intestine by Human Enterotoxigenic Escherichia coli Strains Infect. Immun., February 1, 2006; 74(2): 869 - 875. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
