Stress, whether physical or psychological, can have a notable effect on host physiology, with the earliest and greatest impact occurring in the gastrointestinal tract . In vitro work has shown deleterious effects of stress on intestinal integrity [16, 18], which may enhance pathogen adherence to the intestinal epithelium. Interestingly, some reports show that multiple exposures to mild stress can induce a cytoprotective effect in intestinal epithelial cells against future, more severe, stressors, likely due to induction of the heat shock response . However, in vitro models have shown that acute stress can decrease transepithelial resistance of epithelial cells [43, 45], increase expression or secretion of proteins such as fibronectin  or heat shock proteins [46, 47], which are targeted as receptors by some enteric pathogens [13, 48, 49]. While in vivo studies with food producing animals have associated stress with increased intestinal colonization and shedding of Salmonella and other enteric pathogens [9, 10], less is known of how stress may influence Salmonella interaction with the human intestinal epithelium, or how probiotic bacteria may mediate this interaction.
Here we report increased binding of S. Typhimurium to Caco-2 cells following 1 h of thermal stress (41°C) (Table 1, Fig 1). Epithelial cells may be subjected to stress in a variety of ways, and it is worth noting that bacterial infection itself may serve as a stressor to host tissues. A natural S. Typhimurium infection can induce fever in humans and animals [50, 51]. Little information exists on the influence of fever on epithelial colonization by Salmonella, and although many systemic host factors are involved in the fever response that cannot be accounted for in a cell culture model, our data suggests that high temperature may influence epithelial susceptibility for infection.
Thermal stress alone elicited a mild increase in LDH release (9.46% cytotoxicity) from Caco-2 cells, agreeing with previous reports of epithelial cell damage induced by temperatures near 41°C . However, transmission electron micrographs showed that this heat treatment alone was not sufficient to cause discernable changes in epithelial structure (Fig 4A, B). However, LDH release was greatest when stressed cells were infected with S. Typhimurium and to a lesser extent following exposure to nonpathogenic E. coli K12 (Table 1). The high level of cytotoxicity observed during infection with S. Typhimurium was likely due to membrane damage elicited by enterotoxins  (Table 1). Exposure of Caco-2 cells to L. rhamnosus or L. gasseri did not induce LDH release, suggesting that Gram-negative bacterial products, absent in Lactobacilli, may have enhanced the cytotoxic effect of high temperature. Interestingly, others have reported potential cytoprotective effects of Lactobacilli on intestinal epithelial cells [54, 55], and exposure to these probiotics may protect the epithelium against negative effects of physiological stress or infection [38, 56].
We also tested the influence of high temperature and Salmonella exposure on cytokine expression in Caco-2 cells, since certain cytokines can alter gut integrity and influence the outcome of infection [32, 39, 40]. As reported by others, Salmonella infection significantly increased expression of IL-6 and IL-8 [41, 42]. However, we found no effect of thermal stress (Fig 5) on cytokine levels, indicating that alteration of epithelial cytokine production is not a likely mechanism by which stress affects intestinal susceptibility to S. Typhimurium colonization.
Initial adhesion to the intestine is the critical first step in establishing colonization or infection of the host . Recent studies have demonstrated the importance of genes encoded on SPI3 in intestinal colonization of S. Typhimurium. SPI3-encoded T5SS (Type 5 Secretion System) pathway members MisL and ShdA were shown to bind to intestinal fibronectin in the mouse and mediate persistent S. Typhimurium colonization [14, 15]. Disturbance of the intact epithelial barrier by stress or disease may increase exposure of basolateral proteins such as fibronectin, and may increase opportunity for pathogen binding. In the current study, we used a ShdA-specific antibody to block the ShdA protein on the surface of S. Typhimurium prior to conducting adhesion assays. The preliminary data showed that blocking ShdA reduced S. Typhimurium adhesion to normal and stressed Caco-2 cells, confirming that ShdA is also important for binding to human intestinal cells. While treating Salmonella with the anti-ShdA antibody did significantly reduce adhesion to thermal-stressed monolayers, adhesion after antibody treatment was still greater than that observed in unstressed cells. This indicates that while fibronectin exposure may play an important role in Salmonella colonization during stress and non-stress conditions, it is not the only factor involved in promoting colonization during epithelial cell stress. Indeed, a variety of pili and adhesion molecules also contribute to Salmonella binding and invasion during normal host condition  and are likely to promote binding when intestinal homeostasis is perturbed .
Previous reports have demonstrated the ability of probiotic bacteria to decrease pathogen binding and ameliorate mucosal damage elicited by infection. We recently showed that Lactobacillus bulgaricus inhibits binding and cytotoxic effect of Clostridium difficile with a Caco-2 cell model . In addition, probiotics Streptococcus thermophilus and Lactobacillus acidophilus limited adhesion and invasion of enteroinvasive E. coli, and increased transepithelial resistance and tight junction integrity during infection . Exposing epithelial cells to Lactobacillus casei prior to infection with adherent-invasive E. coli reduced adhesion of the pathogen by 73% .
In the current study, we examined the influence of L. rhamnosus GG and L. gasseri on Salmonella infection during acute epithelial stress. We chose these organisms because numerous reports indicate their effectiveness as probiotics, by improving epithelial integrity during infection [60, 61], and by limiting pathogen binding through either direct competition or by lactic acid production [38, 62]. Here, we demonstrate that L. rhamnosus GG significantly reduced the cytotoxic effect of Salmonella in thermal-stressed Caco-2 cells, which agrees with other reports of the effectiveness of this probiotic in improving mucosal integrity and epithelial cell health during infection or exposure to toxins [60, 63]. We also observed that L. rhamnosus significantly decreased Salmonella adhesion to stressed Caco-2 cells, but did not alter binding to unstressed cells (Fig 1C). Unlike L. rhamnosus, L. gasseri neither protected Caco-2 cells from the cytotoxic effect of high temperature and S. Typhimurium, nor altered adhesion of Salmonella. In contrast to our data, others found that both L. rhamnosus GG and L. gasseri limited adhesion of Salmonella  and E. coli [38, 64] to unstressed host cells. These discrepancies could be due to differences in the specific strains of L. gasseri or L. rhamnosus used in those studies, or to differences in the dose of probiotic or pathogen applied in the infection studies.