Oxygen restriction increases the infective potential of Listeria monocytogenes in vitro in Caco-2 cells and in vivo in guinea pigs

Background Listeria monocytogenes has been implicated in several food borne outbreaks as well as sporadic cases of disease. Increased understanding of the biology of this organism is important in the prevention of food borne listeriosis. The infectivity of Listeria monocytogenes ScottA, cultivated with and without oxygen restriction, was compared in vitro and in vivo. Fluorescent protein labels were applied to allow certain identification of Listeria cells from untagged bacteria in in vivo samples, and to distinguish between cells grown under different conditions in mixed infection experiments. Results Infection of Caco-2 cells revealed that Listeria cultivated under oxygen-restricted conditions were approximately 100 fold more invasive than similar cultures grown without oxygen restriction. This was observed for exponentially growing bacteria, as well as for stationary-phase cultures. Oral dosage of guinea pigs with Listeria resulted in a significantly higher prevalence (p < 0.05) of these bacteria in jejunum, liver and spleen four and seven days after challenge, when the bacterial cultures had been grown under oxygen-restricted conditions prior to dosage. Additionally, a 10–100 fold higher concentration of Listeria in fecal samples was observed after dosage with oxygen-restricted bacteria. These differences were seen after challenge with single Listeria cultures, as well as with a mixture of two cultures grown with and without oxygen restriction. Conclusion Our results show for the first time that the environmental conditions to which L. monocytogenes is exposed prior to ingestion are decisive for its in vivo infective potential in the gastrointestinal tract after passage of the gastric barrier. This is highly relevant for safety assessment of this organism in food.


Background
During the last two decades, Listeria monocytogenes has been implicated in several food borne outbreaks and sporadic cases of disease [1,2]. Foods implicated in foodborne listeriosis are generally highly processed foods with prolonged shelflives, supportive of growth of the organism [3]. Several attempts have been made to establish quantitative microbiological criteria for the presence of L. monocytogenes in foods [4,5], by use of dose-response predictions. However, there are indications that the number of bacteria ingested is not the only important determinant for the development of illness. For many enteric pathogens also including Listeria, it is well known that given environmental conditions induce the expression of identified virulence genes and/or contribute to their invasive potential in in vitro models [6][7][8][9][10][11][12][13][14][15]. However, to our knowledge, no reports describe a direct effect of such conditions on the virulence phenotype of these bacteria in vivo. It has thus not been shown whether an induced expression of virulence factors in intestinal pathogens at the time of ingestion affects their virulence in vivo after passage of the gastric barrier.
The objective of the present study was to investigate whether the physiological state of L. monocytogenes prior to ingestion (i.e. determined by the food environment), here exemplified by oxygen availability, could influence its ability to cause infection.
Recent reports point out that mice and rats are not suitable as animal models for human listerial infectivity, since the bacterium does not interact with the epithelial receptor of these animals [16,17]. A guinea pig model [BB Roldgaard, JB Andersen, TR Licht and BB Christensen, submitted] was therefore used for in vivo studies of infectivity in the present study, while Caco-2 cells were applied to assess the infective potential in vitro. In any animal model, variation between individuals is unavoidable. In order to circumvent this variation in the comparative study, challenge of the animals with a mixture of L. monocytogenes ScottA cells cultivated under conditions of different oxygen availability, was included in the investigation. For this purpose, a recently developed fluorescence labeling system [18] was applied, making it possible to distinguish between otherwise isogenic Listeria cells originating from cultures grown with and without oxygen-restriction.
The obtained results reveal that oxygen-restriction clearly increases the infective potential of L. monocytogenes in vitro and in vivo.

Invasiveness of oxygen-restricted and un-restricted L. monocytogenes in Caco-2 cells
Caco-2 cell assays revealed that cultures of L. monocytogenes-CFP, grown to selected densities under oxygen-restricted conditions were approximately 100 fold more invasive than corresponding cultures grown without oxygen restriction (Figure 1). Similar results were obtained with L. monocytogenes-YFP, as well as with stationaryphase overnight cultures grown with and without oxygen restriction (data not shown).

Infection of guinea pigs with monocultures of oxygenrestricted and un-restricted L. monocytogenes
Exponentially growing cultures (OD 600 = 1.0) were used for oral dosage of guinea pigs. L. monocytogenes-CFP was recovered from the liver and jejunum of half of the 12 animals dosed twice with the un-restricted bacterial cultures, and was found in the spleen of a single animal. The occurrence of Listeria in organs of animals dosed with oxygenrestricted bacteria was significantly higher (p < 0.05); the pathogen was recovered from jejunum of all of the 12 animals, and was found in the liver and spleen of ten and seven of the guinea pigs, respectively. The average concentrations of Listeria found in positively infected organs were not different in the two groups (Table 1).
Throughout the experiment, fecal counts of Listeria remained significantly higher in the animals dosed with oxygen-restricted L. monocytogenes-CFP than in those dosed with the un-restricted, identical strain. The concentration of Listeria in feces reached a level of 10 6 -10 7 per In vitro infectivity Figure 1 In vitro infectivity. Numbers of invaded un-restricted (grey) and oxygen-restricted (white) L. monocytogenes-CFP per well of Caco-2 cells grown to selected densities. Counts were normalized to a concentration of 10 7 L. monocytogenes per ml in the bacterial cultures prior to infection. Each bar represents an average of three different experiments. Error bars designate standard deviations. OD600 refers to optical density at 600 nm.  Figure  2A).

Competitive infection of guinea pigs with oxygenrestricted and un-restricted L. monocytogenes
Oral dosage of 24 guinea-pigs with a mixture of oxygenrestricted and un-restricted, exponentially growing L. monocytogenes revealed that also in a mixed challenge experiment, the occurrence of the oxygen-restricted strain in internal organs was significantly higher (p < 0.05) than observed for the un-restricted strain. Listeria, which had been grown under un-restricted conditions, was detected in the liver and jejunum of 2 animals, and in the spleen of 4 animals, while Listeria grown under oxygen-restricted conditions prior to dosage was recovered from liver, spleen and jejunum of 18, 12, and 14 animals, respectively. It had no influence on the infectivity (p = 0.165) whether the L. monocytogenes cells were labeled with CFP or YFP. The average concentrations of Listeria found in positively infected organs were similar for oxygenrestricted and un-restricted bacteria ( Table 2).
In all animals, fecal concentrations of Listeria grown under oxygen-restricted conditions prior to dosage were higher than concentrations of cells grown without oxygen restriction ( Figure 2B). This was observed independently of which fluorescent label was used to identify the bacteria ( Figure 2C and 2D). The kinetics of the pathogen occurrence in feces observed in the mixed infection experiment were quite similar to what was observed after dosage with monocultures ( Figure 2).

Discussion
We have shown that oxygen-restriction prior to ingestion significantly (p < 0.05) enhances the in vivo infective potential of L. monocytogenes ScottA (Table 1 and 2, Figure   2). This has to our knowledge not previously been shown, and signifies that gene expression occurring in Listeria before intake influences its infective potential even after passage of the oral/gastric barrier. The trend in the results is that the prevalence of Listeria in internal organs   Numbers of guinea pigs where L. monocytogenes-CFP was recovered from liver and spleen after oral dosage with un-restricted or oxygen-restricted L. monocytogenes-CFP, respectively. Mean log (CFU/g) designate the mean levels found in animals positive for Listeria in the given organs, followed by standard deviations.

Fecal densities
increases from Day 4 to Day 7 post challenge with the unrestricted bacteria, while a decrease in prevalence is seen in the same period after dosage with the oxygen-restricted bacteria ( Table 1). The effect of oxygen-restriction prior to dosage is thus larger on Day 4 than on Day 7 post dosage, which is not surprising because the expression patterns present in the cultures at the time of ingestion must be expected to approach each other over time, since the cells are all exposed to similar environmental conditions in the gut.
While oxygen restriction clearly affected the number of animals carrying L. monocytogenes in their internal organs, the concentration of bacteria present in positively infected organs was not affected (Table 1 and 2). This suggest that oxygen restriction increases the initial translocation of Listeria from the gut lumen to internal organs, but does not influence the ability of the bacteria to proliferate inside the investigated organs.
Obviously, an increased ability to survive the gastric barrier will increase the probability of causing an infection [19,20]. It is well known, that genes involved in many kinds of stress-responses are co-regulated, and that exposure to one type of stress therefore improves the ability to survive another type of stressful condition [21,22]. It could thus be speculated, that exposure to oxygen restriction would increase the viability of L. monocytogenes under low pH and/or its resistance to gastric enzymes. This is however not the full explanation for our observations, since also the ability of this strain to infect Caco-2 cells in vitro is significantly increased by oxygen restriction ( Figure  1). In the in vitro studies, the effect of oxygen restriction was seen for exponentially growing cultures as well as for cultures in stationary phase, and the induction of the virulent phenotype was thus expected to be independent of the growth stage of the bacterial cells ( Figure 1). Still, we chose to use exponentially growing cultures for the animal studies in order to eliminate putative contributions to in vivo infectivity from the many genes known to be induced in stationary phase.
We suggest that the observed increased infectivity of L. monocytogenes grown under oxygen-restricted conditions can be attributed to an increased expression of the Intern-alinA (InlA) protein, which is known to be a key factor for virulence of L. monocytogenes [23], on the surface of the bacterial cells. In concordance with this hypothesis, recent reports indicate that anaerobic physiology contributes significantly to an enhanced production of InlA in aro mutants of L. monocytogenes [24,25]. InlA mediates the interaction of Listeria cells with specific receptors (E-cadherin) in the human gut [26], and may therefore be important for bacterial attachment to the epithelial wall. We observed that oxygen-restriction significantly increased the prevalence of L. monocytogenes in the jejunum of guinea-pigs (Table 1 and 2), suggesting that an increased InlA-expression occurring prior to ingestion caused increased attachment of L. monocytogenes to the jejunal mucosa. We speculate that the observed increased translocation to spleen and liver (Tables 1 and 2), as well as the increased invasion of Caco-2 cells (Figure 1) may be attributed to an increased initial InlA-mediated attachment to the epithelial receptors. However, also other genes are reported to be induced under oxygen restricted conditions and to affect attachment of Listeria [27].
The percentage of animals in which Listeria was recovered from internal organs was significantly lower (p < 0.05) in animals infected with mixed cultures, than in animals infected with monocultures (Tables 1 and 2). The sensitivity of the mixed culture infection approach was thus slightly lower than observed for infection with monocultures, probably because the dosed numbers of cells grown under a given condition in the mixed infections were only half of the corresponding numbers dosed as monocultures. Total numbers of cells in each dosage were similar in the two approaches. In spite of this, the observed difference between oxygen-restricted and un-restricted Listeria was more significant in the mixed culture experiment (p liver < 0.0001; p spleen = 0.0143; p jejunum = 0.0002) than in the monoculture infections (p liver = 0.0833; p spleen = 0.0094; p jejunum = 0.0047) due to the higher amount of Numbers of guinea pigs where L. monocytogenes was recovered from liver and spleen after oral dosage with a 1:1 mixture of un-restricted and oxygen-restricted L. monocytogenes, carrying either of the two different fluorescent labels CFP and YFP. Samples were taken either four or seven days post first dosage. Mean log (CFU/g) designate the mean levels found in animals positive for Listeria in the given organs, followed by standard deviations.
data resulting from the mixed-infection approach. This shows that the co-infection model, which has the advantage of elimination of variations attributed to individual animals, additionally had a better discriminatory power than a monoculture approach involving the same number of animals.

Conclusion
Our results are of particular importance for the risk assessment of Listeria in food. For the first time, we have shown that the environmental conditions to which a bacterium is exposed before ingestion can be decisive for its infective potential when it reaches the gut. This means that not only the number of Listeria present in a given food item, but that also the physiological condition of these bacteria is important for food safety. The in vitro and in vivo data suggest that an oxygen-restricted L. monocytogenes cell represents a significantly higher risk than a cell grown without oxygen restriction. This should be taken into account in future quantitative risk profiling and dose-response models.

Strains and media
The clinical isolate, Listeria monocytogenes ScottA, carrying erythromycin resistance, and labeled with either Cyan Fluorescent Protein (CFP) or Yellow Fluorescent Protein (YFP) as previously described [18] was used in all experiments. Bacteria were cultivated on BHI-agar (Oxoid) or in liquid BHI (Oxoid) buffered with 100 mM 3-(N-Morpholino)-propanesulfonic acid (MOPS), pH = 6. When appropriate, Erythromycin (Sigma) was used at a final concentration of 10 ug/mL and Nalidixic acid at a final concentration of 100 ug/mL.

Preparation of L. monocytogenes for infection studies
A fluorescent single colony of L. monocytogenes was inoculated into 10 mL MOPS buffered BHI media supplemented with Erythromycin and incubated at 37°C for 8 hours. The resulting cultures, for which the optical densies (OD 600 ) ranged between 1,5 and 2, were subsequently diluted between 10 4 and 10 5 fold into 2 L Bluecap flasks containing 450 mL MOPS buffered BHI medium supplemented with Erythromycin.
To obtain oxygen-restricted cultures of Listeria monocytogenes ScottA, atmospheric air above the diluted cultures were exchanged with sterile Nitrogen and the lid of the Bluecap flasks was tightened and sealed prior to incubation. To obtain non-restricted cultures, the Bluecap flasks were incubated with the lid loosely tightened allowing free exchange of atmospheric air. All cultures were incubated at 37°C in a rotary shaker set at 200 rpm and samples for in vitro invasion studies and in vivo infection studies were taken after approximately 20 hours of incu-bation, when they reached the optical densities reported below.
Samples for in vitro invasion studies were taken at OD 600 = 0.5 (exponential phase), OD 600 = 1.0 (exponential phase), and from cultures that had been in stationary phase for approximately 10 hours (referred to as over night cultures). The samples were diluted directly into 37°C MEM medium (Gibco) to a concentration of 10 7 bacteria/mL and immediately used for challenge of Caco-2 cells as described below.
Samples for oral challenge of guinea pigs were cultivated until OD 600 = 1, and 400 mL culture per group of 6 animals was harvested by centrifugation (5 minutes, 7000 g, 25°C). The pellet was resuspended in 4 mL of double cream (38 % fat, pH 7). Aliqouts of 0.5 mL, containing either Listeria monocultures, or a mixture of two cultures grown at oxygen restricted and unrestricted conditions, respectively, were subsequently administered to guinea pigs as described below. Samples of the inoculums were diluted and spread onto BHI-agar supplemented with Erythromycin to estimate the numbers of Listeria present in the double cream suspensions. Microscopy revealed that the bacteria were present in the hydric phase.

Caco-2 cell infection experiments
Enterocyte-like Caco-2 cells were cultivated and prepared as previously described [28]. Bacteria were cultivated as described above to the desired optical density and diluted in 37°C MEM medium (Gibco) to a concentration of 10 7 bacteria/mL immediately before the invasion assays. One ml of bacterial culture was applied to each well, resulting in a multiplicity of infection of approximately 25 bacteria per Caco-2 cell. Following 1 hour of invasion and 2 hours of gentamycin treatment to kill extracellularly located bacteria, Caco-2 cells were lysed and the numbers of Listeria present in each well was estimated as described in the section 'Enumeration of L. monocytogenes in samples'. Sampling was done in triplicate, and the experiments were performed twice. A total of 48 guinea pigs were dosed with L. monocytogenes ScottA. Two groups of 12 animals were dosed with monocultures of CFP-labelled bacteria, cultivated either with or without oxygen-restriction. Two other groups of 12 animals were dosed with a 1:1 mixture of oxygen-restricted and un-restricted bacteria carrying the two different fluorescent labels CFP and YFP. In one of these groups, it were the oxygen-restricted Listeria, which carried the CFP label, while in the other group the labels were reversed so that the unrestricted bacteria were labelled with CFP.

Animal experiments
All animals were dosed with 0.5 ml double cream containing 38 % milk fat and approximately 5 × 10 10 L. monocytogenes at Day 0 and again at Day 1. The number of cells in the inoculum was approximately the same in the mono-and mixed cultures.
Dosage was done directly in the oral cavity, between the incisors and the molars. Following dosage the animals had access to food and water ad libitum throughout the duration of the study. Fresh fecal samples were collected every day, avoiding bedding and traces of urin from the cages. Six animals from each of the groups were euthanized on Days 4 and 7, respectively, for investigation of intestinal segments, liver and spleen. Samples of jejunal content from the middle part of the small intestine were squeezed out with tweezers, and samples from spleen and liver of approximately 0.6 g were homogenized prior to investigation as described below.

Ethical aspects
Animal experiments were carried out under the supervision of the Danish National Agency for Protection of Experimental Animals The recently developed guinea-pig model [BB Roldgaard, JB Andersen, TR Licht and BB Christensen, submitted] is much less stressful to the animals than previously published models [17], since it involves no intubation, injection or anesthetizing. Additionally, the approach of treating each animal with a mixture of cultures reduces the number animals needed.

Estimation of L. monocytogenes in samples
Samples of lysed Caco-2 cells and homogenized organs were cultivated on BHI-agar supplemented with Erythromycin. Samples from faeces and jejunum were cultivated on BHI-agar supplemented with Erythromycin and Nalidixic acid. After 48-72 hours of incubation at 37°C, BHI plates were placed on a UV table (excitation at 302 nm), and fluorescent colonies (either CFP or YFP) were enumerated.

Statistics
Statistical analysis was performed using the software package JMP (SAS institute, Copenhagen, Denmark). Pearsons Chi 2 test was used to compare the numbers of infected animals after dosage with either oxygen-restricted or unrestricted L. monocytogenes. The same test was used to compare the results obtained with either mono infection or mixed infection, and with either YFP or CFP labelling. A significance level of 0.05 was used in all cases.