This paper describes a novel method for oral infection of mice using Listeria monocytogenes-contaminated food. The protocol can readily be adapted for use with other food borne bacterial pathogens.
L. monocytogenes are facultative intracellular bacterial pathogens that cause food borne infections in humans. Very little is known about the gastrointestinal phase of listeriosis due to the lack of a small animal model that closely mimics human disease. This paper describes a novel mouse model for oral transmission of L. monocytogenes. Using this model, mice fed L. monocytogenes-contaminated bread have a discrete phase of gastrointestinal infection, followed by varying degrees of systemic spread in susceptible (BALB/c/By/J) or resistant (C57BL/6) mouse strains. During the later stages of the infection, dissemination to the gall bladder and brain is observed. The food borne model of listeriosis is highly reproducible, does not require specialized skills, and can be used with a wide variety of bacterial isolates and laboratory mouse strains. As such, it is the ideal model to study both virulence strategies used by L. monocytogenes to promote intestinal colonization, as well as the host response to invasive food borne bacterial infection.
Listeria monocytogenes are facultative intracellular bacterial pathogens that cause food borne infections in humans. The bacteria are resistant to many of the processes used to protect our food supply, such as drying, salting, or refrigeration 1,2 and infections are typically linked to processed, “ready-to-eat” foods that are not heated prior to consumption. In several previous outbreaks, the source of L. monocytogenes-contaminated food was identified, and the restricted group of exposed individuals was monitored closely 3-6. In those examples, clinical disease in otherwise healthy individuals varied from a mild, self-limiting gastroenteritis to more severe intestinal and systemic infections that required hospitalization. L. monocytogenes infections in immune compromised individuals, including both neonates and the elderly, have been associated with a high mortality rate (25-30%), even with antibiotic treatment 7,8. Little is known about the infectious dose for L. monocytogenes, or the host factors that govern susceptibility to infection after oral transmission, primarily due to the lack of a small animal model that recapitulates this wide range of infection outcomes.
The most widely used model of listeriosis is intravenous (i.v.) inoculation of mice. The i.v. model is highly reproducible, and has been extremely useful for studying both naïve and memory T cell responses during infection 9,10. The drawback of the i.v. model is that it completely bypasses the intestinal phase of infection. After food borne transmission, the gut mucosa provides a barrier that presumably slows and limits the number of bacteria that can continually disseminate to peripheral tissues. In contrast, the entire inoculum can be found in the spleen and liver within minutes of i.v. administration, and this large bolus of organisms may overwhelm innate immune defenses in these tissues. Oral infection of mice by gavage is less commonly used, because large doses (109-1011 CFU) are typically required to achieve intestinal colonization 10. Also, intragastric (i.g.) inoculation with a feeding needle does not generate a reproducible period of gastrointestinal infection prior to systemic spread. Some labs have reported that L. monocytogenes reach the spleen and liver within 4-12 hr post infection (hpi), while others showed no systemic spread until 48 hpi 11-15. This lab-to-lab variation may be a consequence of the invasive nature of i.g. inoculation, which can result in minor trauma to the lining of the esophagus, and promote direct bloodstream invasion of the bacteria.
We recently developed a novel mouse model of oral L. monocytogenes infection that closely mimics all phases of human disease 16. Infection occurs when the mice ingest pieces of contaminated food, a process that is non-traumatic, and does not require specialized skills by laboratory investigators. For a discrete period of time (36-48 hr), L. monocytogenes reproducibly colonize only the gastrointestinal tract, thus allowing investigation of the mechanisms used by pathogenic Listeria to translocate across the gut mucosa and disseminate to peripheral tissues. Importantly, the model can be used to study differences in host innate resistance to infection. In a previous study, we showed that BALB/c/By/J mice were highly susceptible to food borne listeriosis, with exponential replication of L. monocytogenes occurring in the gut, spleen, liver, and gall bladder 16. In contrast, C57BL/6 mice were resistant to food borne infection, with only transient colonization of L. monocytogenes occurring in each of these tissues. An additional feature of the food borne model that closely mimics human disease is that natural dissemination to the brain occurred during the later stages of the infection (5-7 dpi).
1. Preparation of Selective Agar Media (BHI/L+G) to Inhibit Intestinal Microbiota
2. Preparation of the Inoculum
3. Infection of Mice
4. Monitoring the Level of Bacteria Shed in Feces
5. Processing of Infected Intestinal Tissues
6. Processing of Infected Mesenteric Lymph Nodes
7. Processing of Infected Spleens, Livers, & Gall Bladders
8. Processing of Infected Brains
9. Supplemental Procedure: Fractionation of Intestinal Tissues
9.1 Isolation of Bacteria in the Mucus Fraction
9.2 Isolation of Bacteria in the Epithelial Cell (EC) Fraction
9.3 Isolation of Bacteria in the Lamina Propria (LP) Fraction
L. monocytogenes colonies will be visible on BHI/L+G plates after 36-48 hr incubation at 37 °C. The colonies have a smooth, dome-shaped creamy white appearance (Figure 1A). Growth will be inhibited for the majority of the intestinal microbiota, but it is common to see some colonies that are not L. monocytogenes, particularly when plating small intestine or colon directly without significant dilution (Figure 1B). Suspect colonies can be confirmed by plating on ChromAgarTM Listeria plates. L. monocytogenes appear as blue colonies surrounded by a white halo on these plates (Figure 1C). The limit of detection for L. monocytogenes in each tissue is dependent on the volume of water used to homogenize the tissue. The sample volumes shown in Table 1 represent the minimal volume needed to effectively process each tissue. Homogenates of spleen, gall bladder, brain, or fractionated intestinal tissue can be stored at 4 °C for 1-2 days and re-plated if necessary. Colon, small intestine or liver homogenates can inhibit some L. monocytogenes growth unless sample is diluted sufficiently (Figure 1D), and these homogenates do not yield similar CFU counts if re-plated after storage at 4 °C.
We previously showed that BALB/c/By/J (BALB) mice were significantly more susceptible to food borne listeriosis than C57BL/6 (B6) mice 16. An inoculum of 107 CFU is sufficient to establish intestinal infection in BALB mice, but at least 108 CFU is needed to colonize the gastrointestinal tract of B6 mice (Figure 2). As shown in Figure 2, the bacterial load in the intestines, spleen, liver and gall bladder will be proportional to the challenge dose given to the mice. Preliminary studies suggest that 5 x 109 CFU is the approximate LD50 for BALB/c/By/J mice using this model 16.
Figure 1. Representative colony growth on selective agar plates. A) L. monocytogenes colonies after 48 hr growth on BHI/L+G agar. B) BHI/L+G agar inhibits the growth of most intestinal microbiota, but some non-Listeria colonies (arrow) may be observed at low dilutions. The agar plate shown here contains 100 μl of a 10-1 dilution on half of the plate, and 50 μl each of the 10-2 and 10-3 dilutions of a colon homogenate. C) L. monocytogenes colonies can be confirmed by growth on CHROM agar Listeria plates. L. monocytogenes appear blue with a white halo surrounding the colony. D) It is not uncommon to see an inhibition of L. monocytogenes growth, resulting in colonies of varying sizes, in the lowest dilution of either intestinal or liver homogenates plates on BHI/L+G agar.
Figure 2. Dose response of food borne infection in BALB/c/By/J and C57BL/6 mice. Female BALB/c/By/J (BALB) and C57BL/6 (B6) mice were infected with 109(n=7), 108(n=8), 107(n=6) and 106(n=4) Lm InlAm and the number of CFU in each tissue was determined 5 dpi. Mean values +/- SD are shown. Pooled data from two different experiments were analyzed. Dashed lines indicate the limit of detection in each organ.
Tissue | Sample volume (ml)a | Limit of detection (CFU) |
Small intestine | 2.0 | 50 |
Cecum | 2.0 | 50 |
Colon | 2.0 | 50 |
Mesenteric Lymph Nodes | 1.5 | 15 |
Spleen | 2.5 | 50 |
Liver | 2.5 | 50 |
Gall bladder | 0.5 | 10 |
Brain | 1.5 | 15 |
Table 1. Limit of detection for L. monocytogenes in tissue homogenates. a Average total sample volume consisting of sterile water plus the homogenized tissue.
Inbred mice are not uniformly receptive to feeding at all times of the day, and their willingness to eat the contaminated bread will depend both on the strain type and the age of the mice 17. In our experience, 6-9 week old B6 mice are receptive to feeding at any time of day, but BALB mice will not consistently eat the bread piece unless it is offered during their dark cycle. The light cycle of the room used to house the animals can be altered so the dark phase coincides with the normal working day for laboratory personnel. However, the mice should be given at least two weeks to acclimate to the altered light cycle prior to infection. During the infection, only red lamps should be used in either the room itself or in the laminar flow hood.
In developing this model, bread was chosen as the food source to transmit L. monocytogenes largely because bread is absorbent. Thus, it was easy to saturate small pieces with the bacterial inoculum and ensure that each mouse ingested the same dose. However, this model could readily be adapted to use other food sources. Indeed, this is the only mouse model that can be used to directly test the effect of food composition or storage conditions on bacterial infectivity. L. monocytogenes can easily adapt to growth in high salt, low pH, or cold temperatures 1,2,18, but it is not known if these growth conditions alter the ability of the bacteria to establish intestinal infection.
Approximately 10 min is needed to harvest all of the infected organs from a single mouse. The gall bladder is easily ruptured, so it is best to remove it first. The mesenteric lymph nodes are easiest to spot when the intestines have not been disturbed, so they should be removed next. The intestines can be harvested as a single tissue by cutting just below the stomach and just above the appendix, and then separated into duodenum, jejunum, ileum, cecum, and colon just prior to flushing and homogenizing. The spleen and liver are typically removed next, and the brain is harvested after retrieving organs from the peritoneal cavity, as it is necessary to turn the mouse over to access the skull. When working with multiple mice, all tissues should be stored on ice until processed. For most tissues, one agar plate was used to determine the total number of CFU present in the tissue. The plate was divided into thirds, and a 50-100 μl sample of three separate dilutions of the tissue homogenate was plated on each third.
The methods described here will identify the total number of L. monocytogenes in each mouse tissue, not just the intracellular organisms. The unique intracellular life style of L. monocytogenes is thought to be a key virulence determinant during infection 18,19. However, L. monocytogenes are facultative, not obligate, intracellular pathogens, and there are no definitive published reports to indicate the percentage of Listeria that actually reside within cells in vivo. Most previous studies using oral L. monocytogenes infection relied on gentamicin treatment of gut tissues to inhibit the growth of intestinal microbiota. Pre-treatment with gentamicin has the potential to eliminate extracellular L. monocytogenes, allowing recovery of only the intracellular bacteria, although it is not clear to what extent gentamicin is able to penetrate whole intestinal tissues in vitro. The supplemental protocol described here allows for determination of both extracellular and intracellular bacteria in the mucus layer, the epithelium, and the lamina propria compartment of infected intestinal tissues. In this procedure, gentamicin is used to treat single cell suspensions, ensuring that all recovered bacteria were inside cells within the tissue.
Three washes with NAC typically removed the majority of the mucus from intestinal tissues, and the pooled washes did not contain any eukaryotic cells (viable or dead), as determined by trypan blue staining. Additional washes did not yield visible mucus as determined by Diff-Quik staining of cytospin slide preparations. The cellularity of the EC and LP fractions can also be confirmed using Diff-Quik or Giemsa staining. The EC fraction should consist of primarily epithelial cells, which can readily be distinguished from the few intraepithelial lymphocytes present in the intestinal epithelium. The LP fraction is more diverse, containing a mixture of mononuclear cells that changes composition after infection when inflammatory monocytes infiltrate the tissue.
The food borne model of listeriosis can be used with a wide variety of L. monocytogenes isolates, including the mouse-adapted InlAm-expressing strain 15 described here, wild type EGDe, and deletion mutant derivatives of these strains 16. In a previous study, we showed that the level of intestinal colonization did not vary significantly among these strains, however, isolates that did not express a high affinity ligand for murine E-cadherin did have a slight defect in dissemination to the mesenteric lymph nodes and spleen 16. Likewise, any mouse strain can be used for infection, making this an attractive model system to study both bacterial pathogenesis and host responses to infection. Finally, although we have not yet tested this directly, we propose that the basic procedures used here should be widely applicable to other food borne bacterial pathogens such as Salmonella, Yersinia, Escherichia, Campylobacter and Citrobacter species.
The authors have nothing to disclose.
This work was supported by grants from the National Institutes of Health (AI079442 and AI091918) awarded to S.E.F.D.
Name of Reagent/Material | Company | Catalog Number | Yorumlar |
Brain Heart Infusion Agar | Difco | BD-241830 | |
Lithium chloride | Sigma | L9650 | |
Glycine | Omnipur | 4840 | |
EDTA | Gibco | 15575-038 | |
DTT | Sigma | D5545 | |
Collagenase, type IV | Worthington | LS004089 | |
DNAse I | Worthington | LS002007 | |
Diff-Quik | Dade-Behring | B4132-1A | |
PowerGen 1000 homogenizer | Fisher | 14-261-06 | |
stainless steel type 304 mesh #80 | Small Parts, Inc. | CX-0080-C | |
Cytospin | Statspin | M801-22 |