Skip to main content
NCBI home page
Annals of Surgery logoLink to Annals of Surgery
. 2003 Nov;238(5):754–764. doi: 10.1097/01.sla.0000094551.88143.f8

Pseudomonas aeruginosa Expresses a Lethal Virulence Determinant, the PA-I Lectin/Adhesin, in the Intestinal Tract of a Stressed Host

The Role of Epithelia Cell Contact and Molecules of the Quorum Sensing Signaling System

Licheng Wu 1, Christopher Holbrook 1, Olga Zaborina 1, Emelia Ploplys 1, Flavio Rocha 1, Daniel Pelham 1, Eugene Chang 1, Mark Musch 1, John Alverdy 1
PMCID: PMC1356156  PMID: 14578740

Abstract

Objective:

We have previously demonstrated that P. aeruginosa can have profound effects on the intestinal epithelial barrier via one of its virulence factors, the PA-I lectin/adhesin. The aims of the present study were to further characterize the interaction of P. aeruginosa and the intestinal epithelium using both in vitro and in vivo approaches

Methods:

In vitro assays examining the effect of bacterial growth phase, epithelial cell contact, and butanoyl homoserine lactone (C4-HSL), a quorum sensing signaling molecule know to affect various extracellular virulence factors in P. aeruginosa, on PA-I expression in P. aeruginosa were performed. In vivo studies were carried out by modeling catabolic stress in mice using a 30% surgical hepatectomy and direct introduction of P. aeruginosa and various virulence components into the cecum. The effect of this model on PA-I expression in P. aeruginosa was determined

Results:

Results demonstrated that PA-I expression in P. aeruginosa is affected by its phase of growth, its contact to the intestinal epithelium, and its exposure to the quorum sensing molecule, C4-HSL. Furthermore, data from the present study suggest that the PA-I lectin/adhesin of P. aeruginosa may be increased in vivo by local factors within the cecum of mice in response to surgical stress

Conclusions:

These data indicate that multiple factors present in the intestinal microenvironment of a stressed host may induce certain opportunistic pathogens to express key virulence factors leading to a state of lethal gut-derived sepsis.

The presence of Pseudomonas aeruginosa in the intestinal tract of a stressed host is associated with a high mortality rate in both humans and animals. Using a mouse model of lethal gut-derived sepsis due to this organism, the authors demonstrate that local factors within the intestinal tract of a stressed host enhance the virulence of this feared pathogen by activating elements within bacteria that dysregulate the intestinal epithelial barrier.

Gut-derived sepsis is a term used to describe a state of systemic inflammation with organ dysfunction hypothesized to be initiated and perpetuated by the intestinal tract microflora. This process is observed to be a life-threatening problem following clinical disorders such as burn injury,1 neonatal enterocolitis,2 severe neutropenia,3 inflammatory bowel disease, and following acute rejection of intestinal graft transplantation.4 We have developed a unique model of gut-derived sepsis whereby Pseudomonas aeruginosa is introduced directly into the cecum of surgically stressed mice.5 P. aeruginosa was chosen to model gut-derived sepsis based on the observation that hospitalized patients have a high prevalence of this organism in their stool following exposure to antibiotics and by the observation that the persistence of P. aeruginosa in the feces of catabolically stressed patients is associated with a high mortality rate.6 We induced catabolic stress in mice by performing a 30% surgical hepatectomy. At the time of the hepatectomy, P. aeruginosa was introduced into the intestinal tract by direct puncture of the cecum. This model resulted in a mortality rate of nearly 100%, whereas sham-operated controls injected with identical strains of P. aeruginosa completely recovered. In this model there was equal translocation and bacteremia between control and hepatectomy mice. This finding, coupled with the observation that systemic injection of P. aeruginosa resulted in no mortality, suggested that the lethal effect of this organism in the intestinal tract might require in vivo virulence activation locally to effectuate a systemic response.5,7,8

Continuous work in our laboratory with P. aeruginosa has identified that a key virulence determinant, the PA-I lectin/adhesin, facilitates the adherence of P. aeruginosa to the intestinal epithelium and causes a major defect in the intestinal epithelial barrier to 1 of its potent cytotoxins, exotoxin A.5,9 The PA-I lectin/adhesin may be up-regulated in response to local environmental factors that are unique to the intestinal tract of a stressed host, thereby explaining why control animals survive whereas stressed animal do not. Local factors within the intestinal tract of a stressed host, which affect bacterial growth rates, colony pattern formation, exposure to the intestinal epithelium, etc, might play an important role in virulence transformation among opportunistic pathogens such as P. aeruginosa. Virulence gene expression in bacteria is a complex process that is dictated by multiple factors present within the local microenvironment.10 Bacterial contact with the intestinal epithelium, a result of stress-induced mucosal immune dysfunction, itself could induce virulence gene expression in intestinal bacteria. In addition, the expression of multiple virulence genes in P. aeruginosa could be further enhanced by activation of its quorum sensing signaling system, a system by which individual bacteria sense changes in their population density and respond when a critical mass is present, presumably that amount necessary to overcome the host.11 Therefore, understanding the response of certain virulence genes to changes in bacterial growth rate, colony formation patterns, exposure to quorum sensing molecules, and contact to their target cells, such as the intestinal epithelium, should provide important insight into aspects of in vivo virulence transformation for the PA-I lectin of P. aeruginosa that have not been previously addressed.

Therefore, the aims of the present study were to build on our previous work on the effect of P. aeruginosa on the intestinal epithelial barrier by first determining whether bacterial-epithelial responses could be reproduced in additional human cultured intestinal epithelial cells not previously reported. We next sought to characterize the interaction between P. aeruginosa and the intestinal epithelium by determining the effects of bacterial growth phase, epithelial cell contact, and the exposure of butanoyl homoserine lactone (C4-HSL), a quorum sensing signaling molecule known to affect various extracellular virulence factors in P. aeruginosa, on PA-I expression in P. aeruginosa. In vivo experiments were then designed to determine whether additional soluble virulence factors in P. aeruginosa could be identified that might also mediate the lethal effect of P. aeruginosa in the intestinal tract of mice. We also determined whether surgical stress itself up-regulated the virulence of intestinal P. aeruginosa by analyzing isolated strains from the cecum of stressed mice for up-regulated PA-I mRNA. Finally, we analyzed various environmental and clinical strains of P. aeruginosa for the presence of the PA-I gene to determine the clinical relevance of PA-I in critically ill patients.

MATERIALS AND METHODS

Experimental Protocol I: The Effect of P. aeruginosa on the Barrier Function of Cultured Human Intestinal Epithelial Cells

Previous studies in our laboratory demonstrated that the clinical strain of P. aeruginosa ATCC 27853 decreased the tight junctional barrier function of the human intestinal epithelial cell line Caco-2 by >50%. To determine whether this effect could also be demonstrated in other human intestinal epithelial cell lines, T-84 cells, a human colon cancer cell line, were apically inoculated with varying doses of live P. aeruginosa (ATCC 27853), and dose and time responses determined. Furthermore, to determine, as previously observed in Caco-2 cells, whether P. aeruginosa mediated effects on T-84 cells involve the paracellular pathway, cumulative mannitol flux analysis was performed. Finally to determine the adhesin(s) responsible for P. aeruginosa-induced effects on cultured human intestinal epithelial cells, various oligosaccharides known to bind to specific bacterial adhesins were used in screening inhibition assays in which Caco-2 cells were exposed to live P. aeruginosa. Briefly, 1% to 3% v/wk solutions of heparin sulfate (glycosaminoglycan-dependent effect), d-fucose (a specific binder of PA-II), d-mannose (a specific binder of type I pili), dextran (MW 4000) (nonspecific inhibitor of bacterial adhesion), and GalNAc (N-acetyl d-galactosamine, a specific binder of PA-I) were added to a 200 μL suspension of 1 × 107 CFU/mL of live P. aeruginosa. Suspensions were then inoculated onto confluent monolayers of Caco-2 cells and maximal fall in resistance at 4 hours recorded. Similarly, bacteria suspended in the various sugars were tested for their ability to inhibit binding to the Caco-2 cells using our previously described bacterial adhesion assay.

Experimental Protocol II: The Effect of Growth Phase, Quorum Sensing Signaling Molecule Exposure, and Epithelial Cell Contact on PA-I Expression in P. aeruginosa

As virulence gene expression in bacteria can be affected by growth phase, population density, and cell contact, we sought to determine the expression of the PA-I lectin during various conditions of growth and cell contact. To determine the effect of growth conditions and population density on PA-I expression, bacteria were grown to late stationary phase in soy tryptic broth for 30 hours and samples taken at various points for PA-I mRNA and protein assay. To determine whether the density of bacterial colonies of P. aeruginosa grown on solid agar affects PA-I expression, bacteria were plated on solid agar overnight at inocula of 104 CFU/mL (separated), 105 CFU/mL (subconfluent), and 106 CFU/mL (confluent) and PA-I expression assessed. Bacterial colonies from the various growth patterns were then harvested and PA-I mRNA relative to total 16s rRNA determined. To assess PA-I expression in ATCC 27853 in response to molecules secreted by the quorum sensing signaling system, bacteria were grown in the presence of varying concentrations (0–100 μmol/L) of the diffusible autoinducer molecule N-butanoyl-L-homoserine lactone (C4-HSL). Dose response curves for PA-I mRNA and protein were then determined. Finally, to determine the effect of epithelial cell contact on PA-I expression, bacteria were apically inoculated onto Caco-2 cells, grown to confluence in tissue culture wells, and harvested at various time points for PA-I mRNA measurements in the presence and absence of GalNAc, an oligosaccharide previously shown to inhibit bacterial contact to Caco-2 cells.5 To assess the specific cell component responsible for cell-associated PA-I expression, Caco-2 cells were processed into nuclear, membrane, and cytosolic fractions and added to suspensions of live P. aeruginosa. PA-I mRNA was then measured at baseline and following exposure to each cell fraction with and without GalNAc (3% wt/vol).

Experimental Protocol III: The Effect of P. aeruginosa and Its Components on Mortality in Mice Following Surgical Stress

Previous studies with this mouse model demonstrated that P. aeruginosa injection into the cecum of mice undergoing a surgical hepatectomy resulted in 100% mortality while injection into the cecum of sham-operated control mice resulted in 100% survival.5 Using a combination of purified PA-I and the potent cytotoxin exotoxin A, we previously demonstrated that mortality could be induced with coadministration of both components, but not either alone. To extend these previous findings to include additional known extracellular virulence factors of P. aeruginosa, which might induce lethality in this model, the effect of live P. aeruginosa introduced into the cecum of mice undergoing hepatectomy was compared with the combination of PA-I and Pseudomonas elastase similarly introduced in the cecum. Mice underwent a 30% surgical hepatectomy followed by direct cecal injection of a 200 μL suspension of 1 × 107 CFU/mL of live P. aeruginosa in the presence and absence of GalNAc (15% v/wk) as previously described.5 Animals in the sham-operated group underwent liver manipulation without hepatectomy followed by cecal injection of an identical quantity of P. aeruginosa. To account for the possibility that bacterial growth through the course of the experiment might exceed the available amount of binding GalNAc, at the time of bacterial injection, 1 mL of free GalNAc was injected antegrade into the ileum through the same puncture site in the cecum as previously described.5 To determine the lethality of Pseudomonas elastase, alone or in combination with PA-I, 200 μL of 1.33 mg/mL solution diluted in 1:2 phosphate-buffered saline of PA-I, alone or in combination, with Pseudomonas elastase (200 μL of a 30 mg/mL suspension) (Nagase Biochemical, Japan) was injected into the cecum of both sham-operated and hepatectomy mice in the absence and presence of GalNAc (15% wt/vol). Similar to the technique of injection of live P. aeruginosa, 1 mL of GalNAc was simultaneously injected retrograde into the ileum to provide a reservoir of GalNAc available to bind PA-I through the course of the experiment. Mortality was determined at 48 hours. Animals that appeared moribund were killed prior to death.

Experimental Protocol IV: The Effect of Surgical Hepatectomy on In Vivo PA-I Expression in the Mouse Cecum and Bacterial Dissemination

Previous work with this model has demonstrated that the cecum of mice undergoing hepatectomy is a site of extreme environmental stress.9 As these findings could serve as environmental cues for cecal P. aeruginosa to express virulence, PA-I expression was determined in strains harvested from the feces and cecal mucosa by subculturing samples on Pseudomonas isolation agar after 36 hours (prior to death) following the introduction of P. aeruginosa into the cecum of mice undergoing either hepatectomy or sham operation. Finally, to determine the effect of hepatectomy on bacterial adherence to the intestinal mucosa and systemic dissemination (translocation), quantitative culture of feces, washed cecal mucosa, liver, and blood was determined between groups by plating samples on Pseudomonas isolation agar.

Experimental Protocol V: Detection of the PA-I in Clinical Isolates of P. aeruginosa From Critically Ill Patients

Environmental isolates were obtained from various sources and were analyzed for the PA-I gene by PCR using standard techniques. Clinical isolates from critically ill, septic patients were obtained from fecal samples obtained under University of Chicago IRB protocol # 10933A: Virulence expression of P. aeruginosa in septic patients. Briefly, septic ICU patients were identified and fecal samples obtained from discarded stool. Eighteen P. aeruginosa strains were used (Table 1). Individual colonies grown on Pseudomonas isolation agar, were identified and incubated overnight in Luria-Bertani broth. Clinical P. aeruginosa strains were isolated from stool of patients by using Pseudomonas isolation agar; 2 mL of overnight cultures was pelleted by centrifuging at 5000 rpm for 5 minutes. The pellet was resuspended in 200 μL phosphate-buffered saline, and total DNA was isolated by Easy-DNA kit (Invitrogen). The DNA was used as a template to amplify the PA-I gene with forward 5′ CGA TGT CAT TAC CAT CGT CG 3′ and reverse 5′ ACC CTG GAC ATT ATT GGG TG 3′ primers to get the final PCR product of 208 bp.

TABLE 1. The Presence of the PA-I Gene in Environmental and Clinical Isolates of P. aeruginosa

graphic file with name 18TT1.jpg

Open in a new tab

Cell culture and bacterial strains

Caco-2 cell were a generous gift of Dr. Mark Moosker (Yale university, New haven, CT). Caco-2 cells were grown in DMEM supplemented with 10% fetal bovine serum (Gibco BRL, Grand Island, NY), 100,000 IU/L penicillin, 100 mg/L streptomycin, and 30 mg/mL human transferrin (Gibco BRL). T-84 cell were cultured in DMEM/F12 medium containing 1% glutamine, 10% NCS, 10 mmol/L HEPES. Before all the experiments, cells were cultured on antibiotic free medium for 24 hours. P. aeruginosa strain ATCC 27853 is a well-characterized nonmucoid clinical isolate, which at baseline expresses a low level of PA-I. Bacteria were stored at -80°C in sterile vials and freshly cultured for each experiment as previously described.5,9

Mouse experiments

All experiments were approved by the Animal Care and Use Committee at the University of Chicago. Inbred Balb/c mice weighing 20 g and 25 g were used for all experiments. Mice were kept in individual wire bottom cages to avoid coprophagia during the entire experimental period.

Surgical model of hepatectomy

Surgical excision of 30% of the liver in mice is a reproducible model of catabolic surgical stress and has been described in detail previously.5,9 In a volume of 200 μL (∼107 CFU), P. aeruginosa 27853 was introduced into the cecum via direct puncture following cautery excision of the left lobe of the liver. Animals were allowed water ad libitum only for the remainder of the study period. This model results in 100% mortality within 48 hours after the procedure. Animals die a typical septic death with features of chromodacryorrhea, lethargy, scant diarrhea, and ruffled fur. Sham-operated control animals injected with identical doses of P. aeruginosa have 100% survival rates and do not appear ill.

RNA isolation and Northern blot analysis of PA-I mRNA; PCR

Total RNA of P. aeruginosa was isolated by the hot phenol chloroform procedure as described by Oelmuller.12 The RNA pellet is resuspended in Rnase free water and quantitated spectrophotometrically. RNA (10 μg) is separated by gel electrophoresis and transferred to Hybond nylon membranes (Amersham). After cross linking, membranes are prehybridized in xotch solution. Specific cDNA probe both for PA-I and 16s is radiolabeled with α32P-dCTP using a random primer labeling kit (DECA primer II-Ambion). The probe was generated by PCR using PA-I and 16s primer and cloned into PCR2.1 vector (Invitrogen). The inserts were sequenced and matched to PA-I and 16s respectively. PA-I primers: F (ACCCTGGACATTATTGGGTG) R (CGATGTCATTACCATCGTCG); 16S primers: F (GGACGGGTGAGTAATGCCTA) R (CGTAAGGGCCATGATGACTT). Membranes are then incubated with labeled probes for 18 hours. Membranes are washed twice in 2 × SSC/0.1% SDS for 15 minutes and then 3 times in 0.2 × SSC/0.1% SDS at 65°C.

Preparation of Caco-2 cell membranes and cytosolic cell fractions

Isolation of Caco-2 cell membranes was prepared using a modification of the methods described by Bookstein et al.13 After Caco-2 cells were grown to confluence as described above, they were scraped and resuspended into iced hypotonic buffer (2 mmol/L MgCl2,2 mmol/L EGTA, 2 mmol/L EDTA, 10 mmol/L Tris-Cl, pH 7.4) and a protein inhibitor cocktail tablet (Roche, Germany). Cells were disrupted with low speed Ultra Turax for 30 seconds and centrifuged for 3 minutes at 500 g to remove nuclei and cell debris. The supernatant was then centrifuged at 100,000g at 4°C for 1 hour to pellet the membranes. The membrane fraction was assayed for total protein content using the Micro-BCA method.

Western blot analysis for the PA-I lectin/adhesin

Monoclonal antibody specific to the PA-I lectin/adhesin of P. aeruginosa was generated using standard techniques. Briefly, purified PA-I lectin/adhesin from P. aeruginosa was purchased from Sigma (St. Louis, MO) and was further purified as described by Lerrer and Gilboa-Garber.14 Splenocytes were taken from mice with high antibody titers and fused with the myeloma cell line SP2/0. Hybridomas were selected in hypoxanthine-aminopterin-thymidine media. The hybridomas were screened by ELISA, and then the clone of interest was secondarily screened by Western blot analysis. Western blot analysis for PA-I was carried out according to the method described by Winzer et al.15

Statistical Analysis

Data were loaded onto the SigmaStat (Jandel Corporation, San Rafael, CA) program and tested for significance using a one-way ANOVA and Neuman-Keuls post hoc testing where appropriate. For nonparametric data involving percent (incidence) of mortality, the Fisher exact test was used. P values of <0.05 were accepted for statistical significance.

RESULTS

The effect of P. aeruginosa and its components on the barrier function of cultured human intestinal epithelial cells

P. aeruginosa strain ATCC 27853 caused a significant decrease in the TEER of cultured T-84 cells that was both dose and time dependent (Fig. 1). Similar to previous studies,5 > 106 CFU/mL were required in order for a significant fall in TEER to occur. Cumulative mannitol flux analysis revealed the fall in TEER, in response to 107 CFU/mL, to be likely caused by a defect in the paracellular pathway, as this sugar primarily traverses this pathway. In Caco-2 cells, preexposure of P. aeruginosa to GalNAc or dextran attenuated the fall in TEER and inhibited bacterial adherence in accordance with our previous report (Fig. 2).5 All cells were viable based on LDH release assays (data not shown).

graphic file with name 18FF1.jpg

Open in a new tab

FIGURE 1. Time and dose response curves for the effect of P. aeruginosa on T-84 monolayer transepithelial electrical resistance (TEER). ATCC 27853 (P. aeruginosa) was apically inoculated onto cells and TEER determined at the specified intervals. Cumulative mannitol flux across monolayers is also displayed at the specified time interval. *P < 0.05 compared with baseline value. Data are mean ± standard error of the mean of triplicate cultures (n = 7) for each group. Data demonstrate a profound and rapid decrease in TEER following apical exposure to P. aeruginosa that appears to be due to a paracellular permeability defect based on the increase in the paracellular probe mannitol.

graphic file with name 18FF2.jpg

Open in a new tab

FIGURE 2. Effect of various sugars on P. aeruginosa-induced adherence and decrease of the TEER of cultured Caco-2 cells. Percentages reflect experiments performed in triplicate with the various sugars. As previously reported in Caco-2 cells, both GalNAc, an oligosaccharide that specifically binds to PA-I, and dextran, a nonspecific inhibitor of bacterial adherence, attenuated the effects of P. aeruginosa on Caco-2 cells.

The effect of growth phase, quorum sensing signaling molecule exposure, and epithelial cell contact on PA-I expression in P. aeruginosa

PA-I expression from P. aeruginosa harvested at various time points during growth is summarized in Figure 3. Data demonstrate that PA-I mRNA and protein concentration (13 kDa) appear late in the growth cycle of this organism. This pattern of virulence induction appearing in early to late stationary phase is typical of many opportunistic organisms. The effect of bacterial colony density on PA-I mRNA from P. aeruginosa dispersed at different dilutions is also summarized in Figure 3. As it is well established that bacteria grown in dense clusters are able to intercommunicate via local quorum sensing molecule dispersion, examining the levels of PA-I expression between the different conditions could suggest a role for quorum sensing. Data demonstrate that the denser the colonies (ie, the more confluent and intimate they become), the greater is the expression of PA-I. Dose response effects of the quorum sensing signaling molecule C4 butanoyl homoserine lactone (C4-HSL) on PA-I mRNA and protein expression are summarized in Figure 4. The effect of epithelial cell contact on PA-I mRNA is summarized in Figure 5. The cell fraction responsible for this effect appears to be the cell membranes as is summarized in Figure 5. That these effects are attenuated in the presence of GalNAc suggests a specific role for PA-I in this response.

graphic file with name 18FF3.jpg

Open in a new tab

FIGURE 3. The effect of growth phase and bacterial cell crowding on PA-I mRNA and protein from P. aeruginosa harvested at various time points during growth and grown on solid agar under separated, subconfluent, and confluent growth conditions. Data shown are representative of 3 separate experiments. PA-I production and mRNA were observed when bacterial colony counts exceeded >1010 CFU/mL. The effect of bacterial colony dispersion grown on solid agar on PA-I expression. Bacterial colonies were scraped off the agar plates and PA-I mRNA and normalized to 16s rRNA to ensure equal mRNA loading among groups. Cell colony confluence played a major role in the expression of PA-I mRNA.

graphic file with name 18FF4.jpg

Open in a new tab

FIGURE 4. Dose response effects of the quorum sensing signaling molecule C4-HSL on PA-I mRNA and protein content. Data are representative blots of 3 separate experiments. C4-HSL had a dose-dependent effect of both PA-I mRNA and protein content in P. aeruginosa.

graphic file with name 18FF5.jpg

Open in a new tab

FIGURE 5. The effect of epithelial cell contact on PA-I mRNA. Blots are representative samples of 3 separate experiments. Bar graph represent % increase in PA-I mRNA above control (mean ± SEM, n = 7) for each group. Overnight culture of P. aeruginosa with Caco-2 cells demonstrated a markedly increased abundance of PA-I mRNA. Incubation of P. aeruginosa with the various cell fractions demonstrated cell membranes to be the likely cell fraction responsible for the induction of PA-I. Bar graph represents the effect of epithelial cell contact on PA-I mRNA at 2 hours (*P < 0.05 vs. control). The small image of blots within the graph is representative of a typical experiment with the 3 groups. Epithelial cell contact had a rapid (∼ 2 hours) and significant effect on PA-I mRNA, an effect that was attenuated in the presence of GalNAc, a specific binder of PA-I.

The effect of P. aeruginosa and its components on mortality in mice following surgical stress

The effect of P. aeruginosa and its various components injected into the cecum of mice following 30% hepatectomy compared with sham-operated controls is summarized in Figure 6. As in our previous studies, P. aeruginosa induced mortality only in mice undergoing surgical hepatectomy, an effect that was abrogated when P. aeruginosa was preexposed to GalNAc.5 The combination of PA-I and elastase-induced mortality in both sham-operated controls and hepatectomy groups while neither component alone had any effect on mortality (Fig. 6). These results are similar to our previous data using a combination of PA-I and exotoxin A. That these effects are abrogated by preincubation of components with GalNAc suggests a putative role for PA-I in this response.

graphic file with name 18FF6.jpg

Open in a new tab

FIGURE 6. The effect of cecal injection of P. aeruginosa, PA-I, and Pseudomonas elastase, alone or in combination, in mice subjected to sham laparotomy or 30% hepatectomy. Data are means ± standard error of means for 7 animals in each group. As previously reported, P. aeruginosa was lethal only in mice undergoing hepatectomy. The combination of PA-I and elastase was lethal to both sham-operated controls and mice following hepatectomy. GalNAc protected against the lethal effect of both live P. aeruginosa and the combination of PA-I and elastase (*P < 0.05 vs. control).

The effect of hepatectomy on quantitative bacterial cultures of the stool, washed cecal mucosa, liver, and blood are summarized in Figure 7. There was significant dissemination of P. aeruginosa from the cecum in both groups of mice, although this effect was greater in mice following hepatectomy.

graphic file with name 18FF7.jpg

Open in a new tab

FIGURE 7. Quantitative bacterial counts in the cecum, liver, and blood of mice undergoing sham laparotomy or 30% hepatectomy followed by P. aeruginosa injection into the cecum. Cultures were performed 36 hours following bacterial injection. A significant increase in adherence of P. aeruginosa to the cecal mucosa was observed in mice subjected to 30% hepatectomy compared with sham-operated controls. Bacterial dissemination to the liver and blood was significant in both groups yet greater in mice following hepatectomy. *P < 0.05.

The effect of surgical hepatectomy on in vivo PA-I expression in the mouse cecum

Results of Northern blot analysis of PA-I mRNA from P. aeruginosa strains harvested from the cecum of mice following cecal injection of P. aeruginosa subjected to either 30% hepatectomy or sham laparotomy are summarized in Figure 8. A greater abundance of PA-I mRNA was observed in both feces and cecal tissues from mice subjected to hepatectomy compared with sham-operated controls.

graphic file with name 18FF8.jpg

Open in a new tab

FIGURE 8. Northern blot analysis of PA-I mRNA in P. aeruginosa strains retrieved from the mouse cecum at 36 hours prior to death in mice subjected to sham laparotomy or 30% surgical hepatectomy and injected into the cecum with P. aeruginosa. Blots are representative images of 4 mice. PA-I mRNA abundance is increased in the feces and cecal tissues of recovered strains of P. aeruginosa.

The presence of the PA-I gene in environmental and clinical isolates

As can be seen in Table 1, all environmental and clinical isolates tested were positive by polymerase chain reaction for the PA-I gene.

DISCUSSION

P. aeruginosa is becoming recognized to be among the most feared hospital pathogens in immunocompromised hosts. While much information has been concentrated on its pathogenesis in the lung among patients with ventilatory-associated pneumonia and with cystic fibrosis, its effect and behavior in the intestinal tract of a stressed host are poorly understood. The mere presence of P. aeruginosa in the gastrointestinal tract of a critically ill adult or infant portends an extremely poor prognosis independent of whether it is present at sites remote from the intestinal tract reservoir.16,17 Recent surveillance reports on the prevalence and endemicity of this organism in the intestinal tract of critically ill patients suggest that not only are a growing number of hospitalized patients harboring this organism within their feces (> 50%), but routine rectal cultures have identified as many as 30% of these strains as antibiotic resistant.18

We have successfully modeled gut-derived sepsis due to P. aeruginosa using a clinically relevant mouse model of surgical stress. In this model, the PA-I lectin/adhesin of P. aeruginosa appears to play a key role in the pathogenesis of gut-derived sepsis due to this organism. Data from our previous work and the present study suggest that the intestinal environment of a stressed host may be a unique site in which key virulence determinants of P. aeruginosa are activated.9 That P. aeruginosa might be invoked in gut-derived sepsis is especially attractive given the potency of its extracellular virulence factors, its well-described biosensor capacity through the quorum sensing signaling system, and its ability to evade the immune system and antibiotics by forming impenetrable biofilms.19

Data from the present study confirm and extend our previous studies, which demonstrate that P. aeruginosa has a profound effect on the intestinal epithelial barrier. The dose and time dependency of this effect with T-84 human intestinal colon cells are similar to our previous data in Caco-2 cells. The inhibitory effect of GalNAc and dextran on the adherence and TEER of Caco-2 cells, while other sugars had no effect, confirms our previous findings that PA-I may play a key role in the altered epithelial barrier function due to the whole organism.5,9 We and others have shown that neither lipopolysaccharide nor exotoxin A has any effect on the intestinal epithelial barrier when applied apically.20 Yet purified PA-I is able to alter intestinal epithelial barrier function equal to that of the whole organism. These data, coupled with the observation that mutant strains of P. aeruginosa lacking functional PA-I have attenuated effects on intestinal epithelial barrier function, support the notion that PA-I is a key element in the dysregulatory effect of P. aeruginosa on the intestinal epithelium.

Data from the present study also demonstrate that the virulence of P. aeruginosa may require numerous environmental and host induction factors for it to develop into a lethal phenotype within the intestinal tract of a susceptible host. In the ideal growth conditions of laboratory media, transitional growth and late stationary phase growth appeared to induce PA-I lectin/adhesin expression as bacterial counts increased. In this environment, PA-I expression was both density and time dependent, requiring a significant bacterial population density (>1010 CFU/mL) and time (>20 hours). Also, as is typical with virulence gene expression among a variety of pathogens, bacterial virulence is differentially expressed in liquid versus solid growth media. In solid growth media, allowing colonies to grow separated or confluently, appeared to influence PA-I mRNA. The reasons for this observation remain unclear. Bacterial adherence to the intestinal epithelium is often observed to be a random spatial event characterized by bacterial clusters attached at various sites on epithelia.21 Data from the present study suggest that cell density and crowding may confer a specific advantage in Pseudomonas virulence expression, which may benefit the pathogenic effect of the organism against the intestinal mucosal epithelium.

An important element to consider when examining bacterial virulence gene expression in ideal growth media is the absence of mammalian cells, which themselves may serve as potent activators of bacterial gene expression. In the present study, intestinal epithelial cells induced a profound and rapid (∼2 hours) effect on PA-I mRNA. Furthermore, by separating cells into crude fractions, we were able to demonstrate that the membrane fractions of Caco-2 epithelial cells appeared to be most active in inducing PA-I expression. While the dependency of PA-I expression on contact is inferred by its inhibition in presence of GalNAc, it is possible that GalNAc occupies domains on epithelial cells that also inhibit cell contact via mechanisms other than via PA-I. The abundance and rapidity of PA-I mRNA in response to cell contact suggest that this effect may be a major pathway by which P. aeruginosa is induced to alter its virulence phenotype within the complex environment of the intestinal tract. As catabolic stress is known to result in profound alterations in local mucosal immune function, epithelial cell contact-induced expression of the PA-lectin/adhesin of P. aeruginosa may represent an example by which the permissive effects of catabolic stress enhance the virulence phenotype of transient intestinal pathogens.22,23

The finding that PA-I is responsive to molecules involved in the quorum sensing signaling system further underscores the ability of this pathogen to regulate its virulence phenotype in response to changes in its population density and environment. We chose to test the effects of C4-HSL in this study as it is reported to be among the more diffusible homoserine lactone molecules.24 P. aeruginosa has two well-described and separate quorum sensing circuits (termed las and rhl) as well as a sensor-regulator25 that can modulate gene transcription in response to increasing AHL concentrations.26 The responsiveness of PA-I expression in P. aeruginosa to either C4-HSL or epithelial cell contact emphasizes the highly coordinated and opportunistic nature of this pathogen, which can change its virulence phenotype depending on its perception of the presence of a critical mass of bacteria and/or the presence of a vulnerable epithelial surface. In the context of a critically ill host’s intestinal tract, both the proliferation of resistant microbes due to antibiotic use as well as the erosion of epithelial defense (loss of the anaerobic flora, low IgA, decreased surfactant, and peristalsis) have the potential to unmask opportunism among pathogens whose virulence regulation is directly triggered by these signals.

Data from mouse experiments in the present study are in accordance with our previous observations demonstrating the importance of PA-I in P. aeruginosa gut-derived sepsis.5 The observation that the combination of PA-I and elastase is lethal to both sham-operated and hepatectomy mice compliments our previous observations with the lethal effect of the combination of PA-I and exotoxin A in this model. While both exotoxin A and elastase are potent cytotoxins of P. aeruginosa, Pseudomonas elastase can be particularly destructive to mammalian tissues. Despite this, an extremely high dose of elastase directly injected into the mouse cecum had no effect on mortality. We have applied Pseudomonas elastase to the apical surface of Caco-2 cells and have not observed any effect on epithelial barrier function or cell membrane integrity (unpublished observations). This is similar to our observations with exotoxin A. Yet, in the present study, PA-I coinjected with elastase into the mouse cecum was lethal, emphasizing the importance of the ability to dysregulate the intraepithelial tight junction to a potential cytotoxin on P. aeruginosa pathogenicity in this model.

We attempted to determine whether PA-I expression was up-regulated by the unique environment of the cecum in mice undergoing hepatectomy. Unfortunately, this process required subculturing of the stool on Pseudomonas isolation agar prior to analysis, which itself is problematic, given that the very environmental pressures of interest are eliminated during the subculturing process. Nonetheless, a difference in PA-I mRNA was observed from cecal tissues and feces in the few mice tested, which suggests a possible environmental effect of hepatectomy on virulence expression in the cecum. Future experiments using in vivo expression technology are underway to confirm this finding.27

Culture results from the mouse experiments involving injection of live P. aeruginosa into the cecum demonstrate that despite dramatic differences in mortality, both groups of animals developed bacterial dissemination to the liver and blood. Although there was a statistically significant increase in translocation to the liver and blood in the hepatectomy group, this quantity of bacteria is still insufficient to induce mortality alone, based on our previous studies.5 While it is possible that more virulent clones of P. aeruginosa proliferate and translocate in mice following hepatectomy, and thereby account for the higher mortality in this group, lethal virulence traits of P. aeruginosa are rarely expressed in bacteria at these concentrations. Yet data from the present study would suggest that cell contact can elicit P. aeruginosa virulence gene expression at bacterial concentrations far less than seen in culture alone. Therefore, whether translocated bacteria are transformed into lethal phenotypes in the bloodstream at the concentrations observed in this study cannot be answered by examining simple quantitative cultures or by reproducing systemic infection by intravenous bacterial injection, as both these methods provide inadequate answers. This point has been made by others who have developed and employed signature-tagged mutagenesis and other in vivo expression technologies to formally determine the in vivo mechanism of virulence in pathogenic bacteria.28 Further studies will be necessary to fully elucidate the mechanisms of lethality of P. aeruginosa in this model.

In summary, herein we demonstrate that the expression of the PA-I lectin/adhesin, a key virulence determinant of experimental P. aeruginosa gut-derived sepsis, may be induced by intestinal epithelial cell contact, activation of the quorum sensing signaling system, bacterial growth phase, and bacterial population density. The finding that PA-I is prevalent among clinical strains from critically ill patients’ fecal flora emphasizes the importance of understanding the behavior of this pathogen in the complex environment of the intestinal tract.

Footnotes

Supported by NIH grant RO1 GM62344–01 (to J.A.).

Reprints: John Alverdy, MD, FACS, University of Chicago Pritzker School of Medicine, 5841 S. Maryland MC 6090, Chicago, IL 60037. E-mail: jalverdy@surgery.bsd.uchicago.edu.

REFERENCES

  • 1.Gang RK, Bang RL, Sanyal SC, et al. Pseudomonas aeruginosa septicaemia in burns. Burns, 1999;25:611–616. [DOI] [PubMed] [Google Scholar]
  • 2.Tsai MJ, Teng CJ, Teng RJ, et al. Necrotizing bowel lesions complicated by Pseudomonas septicaemia in previously healthy infants. Eur J Pediatr. 1996;155:216–218. [DOI] [PubMed] [Google Scholar]
  • 3.Todeschini G, Franchini M, Tecchio C, et al. Improved prognosis of Pseudomonas aeruginosa bacteremia in 127 consecutive neutropenic patients with hematologic malignancies. Int J Infect Dis. 1998;3:99–104. [DOI] [PubMed] [Google Scholar]
  • 4.Rubin RH. Gastrointestinal infectious disease complications following transplantation and their differentiation from immunosuppressant-induced gastrointestinal toxicities. Clin Transplant. 2001;15(suppl 4):11–22. [DOI] [PubMed] [Google Scholar]
  • 5.Laughlin RS, Musch MW, Hollbrook CJ, et al. The key role of Pseudomonas aeruginosa PA-I lectin on experimental gut-derived sepsis. Ann Surg. 2000;232:133–142. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Marshall JC, Christou NV, Meakins JL. The gastrointestinal tract: the ‘undrained abscess’ of multiple organ failure. Ann Surg. 1993;218:111–119. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Schook LB, Carrick L Jr, Berk RS. Murine gastrointestinal tract as a portal of entry in experimental Pseudomonas aeruginosa infections. Infect Immun. 1976;14:564–570. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Kurahashi K, Kajikawa O, Sawa T, et al. Pathogenesis of septic shock in Pseudomonas aeruginosa pneumonia. J Clin Invest. 1999;104:743–750. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Alverdy J, Holbrook C, Rocha F, et al. Gut-derived sepsis occurs when the right pathogen with the right virulence genes meets the right host: evidence for in vivo virulence expression in Pseudomonas aeruginosa. Ann Surg. 2000;232:480–489. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Mekalanos JJ. Environmental signals controlling expression of virulence determinants in bacteria. J Bacteriol. 1992;174:1–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Parsek MR, Greenberg EP. Acyl-homoserine lactone quorum sensing in Gram-negative bacteria: a signaling mechanism involved in associations with higher organisms. Proc Natl Acad Sci USA. 2000;7:8789–8793. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Oelmuller UN, Kruger A, Steinbuchel A, et al. Isolation of prokaryotic RNA and detection of specific mRNA with biotinylated probes. J Microbiol Methods. 1990;11:73–81. [Google Scholar]
  • 13.Bookstein C, Xie Y, Rabenau K, et al. Tissue distribution of Na+/H+ exchanger isoforms NHE2 and NHE4 in rat intestine and kidney. Am J Physiol. 1997;273:C1496–1505. [DOI] [PubMed] [Google Scholar]
  • 14.Lerrer B, Gilboa-Garber N. Differential staining of western blots of avian egg white glycoproteins using diverse lectins. Electrophoresis. 2002;23:8–14. [DOI] [PubMed] [Google Scholar]
  • 15.Winzer K, Falconer C, Garber NC, et al. The Pseudomonas aeruginosa lectins PA-IL and PA-IIL are controlled by quorum sensing and by RpoS. J Bacteriol. 2000;182:6401–6411. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Bertrand X, Thouverez M, Talon D, et al. Endemicity, molecular diversity and colonisation routes of Pseudomonas aeruginosa in intensive care units. Intensive Care Med. 2001;27:1263–1268. [DOI] [PubMed] [Google Scholar]
  • 17.Pierro A, van Saene HK, Jones MO, et al. Clinical impact of abnormal gut flora in infants receiving parenteral nutrition. Ann Surg. 1998;227:547–552. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Arbo MD, Hariharan R, Nathan C, et al. Utility of serial rectal swab cultures for detection of ceftazidime- and imipenem-resistant gram-negative bacilli from patients in the intensive care unit. Eur J Clin Microbiol Infect Dis. 1998;17:727–730. [DOI] [PubMed] [Google Scholar]
  • 19.Kharazmi A. Mechanisms involved in the evasion of the host defence by Pseudomonas aeruginosa. Immunol Lett. 1991;30:201–205. [DOI] [PubMed] [Google Scholar]
  • 20.Benoit R, Rowe S, Watkins SC, et al. Pure endotoxin does not pass across the intestinal epithelium in vitro. Shock. 1998;10:43–48. [DOI] [PubMed] [Google Scholar]
  • 21.Wells CL, Moore EA, Hoag JA, et al. Inducible expression of Enterococcus faecalis aggregation substance surface protein facilitates bacterial internalization by cultured enterocytes. Infect Immun. 2000;68:7190–7194. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Alverdy JC, Spitz, J, Hecht G, et al. Causes and consequences of bacterial adherence to mucosal epithelia during critical illness. New Horizons Crit Care. 1994;2:264–272. [PubMed] [Google Scholar]
  • 23.Alverdy JC, Aoys E. The effect of glucocorticoid administration on bacterial translocation: evidence for an acquired immunodeficient state. Ann Surg. 1991;214;719–723. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Pearson JP, Van Delden C, Iglewski BH. Active efflux and diffusion are involved in transport of Pseudomonas aeruginosa cell-to-cell signals. J Bacteriol. 1999;181:1203–1210. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Rumbaugh KP, Griswold JA, Hamood AN. The role of quorum sensing in the in vivo virulence of Pseudomonas aeruginosa. Microbes Infect. 2000;2:1721–1731. [DOI] [PubMed] [Google Scholar]
  • 26.Whiteley M, Greenberg EP. Promoter specificity elements in Pseudomonas aeruginosa quorum-sensing-controlled genes. J Bacteriol. 2001;183:5529–5534. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Slauch JM, Camilli A. IVET and RIVET: use of gene fusions to identify bacterial virulence factors specifically induced in host tissues. Methods Enzymol. 2000;326:73–96. [DOI] [PubMed] [Google Scholar]
  • 28.Chiang SL, Mekalanos JJ, Holden DW. In vivo genetic analysis of bacterial virulence. Annu Rev Microbiol. 1999;53:129–154. [DOI] [PubMed] [Google Scholar]
  • 29.Avichezer D, Katcoff DJ, Garber NC, et al. Analysis of the amino acid sequence of the Pseudomonas aeruginosa galactophilic PA-I lectin. J Biol Chem. 1992;267:23023–23027. [PubMed] [Google Scholar]
  • 30.Avichezer D, Gilboa-Garber N, Garber NC, et al. Pseudomonas aeruginosa PA-I lectin gene molecular analysis and expression in Escherichia coli. Biochim Biophys Acta. 1994;1218:11–20. [DOI] [PubMed] [Google Scholar]
  • 31.Zaborina O, Dhiman N, Ling Chen M, et al. Secreted products of a nonmucoid Pseudomonas aeruginosa strain induce two modes of macrophage killing: external-ATP-dependent, P2Z-receptor-mediated necrosis and ATP-independent, caspase-mediated apoptosis. Microbiology. 2000;146:2521–2530. [DOI] [PubMed] [Google Scholar]
  • 32.Zaborina O, Misra N, Kostal J, et al. P2Z-Independent and P2Z receptor-mediated macrophage killing by Pseudomonas aeruginosa isolated from cystic fibrosis patients. Infect Immun. 1999;67:5231–5242. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Stover CK, Pham XQ, Erwin AL, et al. Complete genome sequence of Pseudomonas aeruginosa PA01, an opportunistic pathogen. Nature. 2000;406:959–964. [DOI] [PubMed] [Google Scholar]

Articles from Annals of Surgery are provided here courtesy of Lippincott, Williams, and Wilkins

RESOURCES