OUP user menu

Interaction of Listeria monocytogenes with the intestinal epithelium

Justin J.D. Daniels , Ingo B. Autenrieth , Werner Goebel
DOI: http://dx.doi.org/10.1111/j.1574-6968.2000.tb09306.x 323-328 First published online: 1 September 2000


Listeria monocytogenes is a food-borne pathogen that must cross the intestinal epithelial barrier to reach its target organs. We have investigated the importance of M cells in translocation using an experimental mouse model and a novel, recently described in vitro co-culture system that mimics the follicle-associated epithelium (FAE). Our data demonstrate that L. monocytogenes does not require, nor specifically use, M cells of the FAE to cross the gut. We also show that bacterial translocation is rapid and L. monocytogenes can attach very efficiently to exposed basal lamina of the small intestine indicating an important role for extracellular matrix proteins.

  • Listeria monocytogenes
  • M cell
  • Peyer's patch
  • Attachment
  • Invasion
  • Pathogenicity

1 Introduction

Bacterial pathogens that enter the human host through the consumption of contaminated food must somehow translocate the epithelial barrier of the intestine to cause a systemic disease. Bacteria have developed different mechanisms to cross this barrier, some of which have been well characterised in animal models, e.g. Salmonella typhimurium[1,2], Shigella flexneri[3] and Yersinia enterocolitica[4]. Interestingly, these studies have revealed that the aforementioned bacterial pathogens have a specific preference for attachment to and invasion of M cells. Intestinal M cells are only found overlying the lymphoid follicles of the gut-associated lymphoid tissue which are distributed throughout the small and large intestine. When numerous follicles are aggregated together they are collectively termed Peyer's patches.

Listeria monocytogenes, a food-borne pathogen, localises with Peyer's patch tissue after oral infection of mice [5], raising the question of whether L. monocytogenes preferentially interacts with M cells. However, microscopic analysis of the invasion of L. monocytogenes into gut epithelial cells in vivo has not been prolifically studied. Recently, a novel M cell in vitro cell culture system has been established that allows the production of a follicle-associated epithelium (FAE)-like epithelium [6]. By this means, interaction of enteric pathogens with intestinal epithelial cells, including M cells, can be studied in detail.

This study primarily seeks to clarify the importance of M cells in a listeria infection using the standard mouse model and the aforementioned in vitro M cell culture method.

2 Materials and methods

2.1 Bacterial strains and cultivation

L. monocytogenes EGD is a commonly used virulent laboratory strain, was passaged in mice before infection studies and is part of the strain collection of the Lehrstuhl für Mikrobiologie, Universität Würzburg. Y. enterocolitica cured of its virulence plasmid is part of the strain collection of the Max von Pettenkofer-Institut der Ludwig Maximilians Universität, Munich. For all infection studies strains were grown overnight in brain heart infusion broth (BHI), centrifuged (∼3500×g for 10 min) and resuspended in 20% glycerol/BHI, aliquotted into sterile Eppendorf caps and frozen at −80°C. Shortly before use, a bacterial glycerol stock was defrosted, washed with phosphate-buffered saline (PBS) and diluted with PBS to the desired inoculum size.

2.2 Animals studied

Female BALB/c mice 6–8 weeks old were obtained from Charles River, Germany, and housed in specific pathogen-free conditions. Mice were fed with food and water ad libitum unless stated otherwise. At least five mice were used per bacterial strain and per time point. Before orogastric infection mice were starved for 6–8 h to clear the contents of the bowel. Mice were then fed 100 μl containing 1×109–2.5×109 CFU of L. monocytogenes via an orogastric feeding tube. The ligated ileal loop method has already been described [7]. At the appropriate time after infection mice were killed by cervical dislocation after which the relevant tissues (e.g. Peyer's patches and non-Peyer's patch intestinal tissue) were aseptically removed and processed for microscopical analysis.

2.3 M cell co-culture assay

The induction of M cells from differentiated Caco-2 cells was performed as described by Kernéis et al. [6]. Briefly, a sub-clone of the Caco-2 cell line, which builds a particularly pronounced brush border similar to human enterocytes in vivo, was seeded (3×105 cells) onto the lower face of tissue culture-treated Transwell filters (3 μm pore size, 6.5 mm diameter; Costar, Cambridge, MA, USA) overnight. The filters were then transferred to the Transwell tissue culture trays and incubated until the cells were fully differentiated (14 days). Then, lymphocytes isolated from Peyer's patches of BALB/c mice were placed in the upper portion of the filter (106 cells) which represents the basolateral side of the Caco-2 cells. These co-cultures were incubated for a further 4–7 days until FITC-conjugated latex bead (0.02 μm diameter, carboxylate-modified fluorospheres, Molecular Probes, Eugene, OR, USA) transport occurred, indicating the presence of M cell-like cells. Transepithelial resistance was measured to confirm a tight monolayer (data not shown). For infection studies, 107 CFU L. monocytogenes was placed as a 100-μl drop on the lower face of the filters (apical side of cells) for 1 h at 37°C. The monolayers were then washed in PBS and processed for microscopy or lysed in ice-cold H2O and plated onto BHI for CFU determination.

2.4 Scanning electron microscopy

Samples were washed in ice-cold PBS and placed in 2% glutaraldehyde in Sorensen's buffer (pH 7.4) overnight at 4°C. The following day samples were dehydrated through a series of acetone dilutions (35–100%), and finally dehydrated with a critical point dryer (CPD 030; BAL-TEC, Walluf, Germany). Samples were gold-sputtered (SCD 005; BAL-TEC) to a thickness of 30 nm, and examined using a Zeiss Digital Scanning Microscope (DSM 962; Zeiss, Oberkochen, Germany).

2.5 Immunostaining and confocal laser scanning microscopy

Immunostaining of the filters was performed with standard methods [8]. Briefly, cells were fixed with 4% paraformaldehyde with 120 mM sucrose in PBS for 15 min at room temperature (RT), washed with PBS and quenched with 50 mM NH4Cl in PBS. Cells were permeabilised for 15 min with 0.2% Triton X-100 in PBS. The appropriate first probes were added (1:40 for UEA-1, TRITC-conjugated [6], from Sigma; 1:50 for anti-L. monocytogenes or Y. enterocolitica) and incubated at RT for 60 min, washed and the second antibody added for 60 min at RT. All samples were then analysed with a confocal laser scanning microscope (Leica Lasertechnik, Heidelberg, Germany). The digitally acquired images were further processed on an Apple Power Macintosh computer using the public domain NIH Image v1.62 program (developed at the U.S. National Institutes of Health and available on the Internet at http://rsb.info.nih.gov/nih-image/) and GraphicConverter v3.8 (Lemke Software, Germany).

3 Results

3.1 Invasion of the small intestine by L. monocytogenes and subsequent dissemination to deeper tissues

We first analysed the dissemination of L. monocytogenes in BALB/c mice after oral infection and obtained very similar results to MacDonald and Carter [5], confirming the ability of L. monocytogenes to cross the intestinal barrier and infect deeper tissues (data not shown). We also saw a greater bacterial load in Peyer's patches as compared to non-lymphoid small intestinal tissue (data not shown). Furthermore, we saw large numbers of L. monocytogenes reach the blood and deeper organs 30 min post ileal loop infection (data not shown) confirming the data of Pron et al. [9]. Together, these results led us to study the interaction of L. monocytogenes with M cells.

3.2 Microscopical analysis of L. monocytogenes interaction with intestinal tissue

Given the higher numbers of L. monocytogenes in Peyer's patch tissue after orogastric inoculation, one could expect either that invasion was more efficient in this location, or that survival was better. To help answer this question scanning electron microscopy (SEM) analysis of Peyer's patches and villous tissue was performed after orogastric infection of mice. However, only very rarely rod-shaped bacteria, characteristic of L. monocytogenes, were observed attaching to epithelial cells (data not shown). By far the predominant observation was tissue devoid of L. monocytogenes-shaped bacteria. Segmented filamentous bacteria were often seen colonising the FAE (Fig. 1A) which are part of the normal bacterial flora for rodents [10]. Likewise, in the ligated ileal loop test with a very large inoculum of L. monocytogenes (∼109 per loop) only few bacteria attached to, or invaded into, the epithelium over a 120-min time period. However, rod-shaped bacteria were seen to attach to columnar epithelial cells of the upper villi (Fig. 1B), and to the epithelial cells and M cells of the FAE (Fig. 1C,D). These rod-shaped bacteria were not seen in non-infected tissue (data not shown). Very occasionally, M cells had numerous bacteria associated with the apical membrane causing significant membrane alterations (Fig. 1D). In columnar epithelial cells no such membrane alterations were observed. Additionally, no bacteria were seen within the crypt areas between villi. These observations were confirmed using confocal microscopy (data not shown).


SEM of intestinal tissue after ileal loop procedure: non-infected dome area of Peyer's patch showing some of the normal flora (A); after infection with L. monocytogenes, bacteria were occasionally seen attached to villus epithelium (B) as frequently as bacteria attached to FAE (C). Very occasionally, an M cell apical membrane was seen full with bacteria (D). Arrows=bacteria, M=M cell.

Occasionally, small areas of the domes devoid of FAE were observed which left the underlying basement membrane exposed (Fig. 2A). It was not clear whether these areas of damage were present before infection or were induced by L. monocytogenes, however they were not observed in non-infected control mice. Furthermore, numerous rod-shaped bacteria were observed attaching to the basement membrane (Fig. 2B). No listeria-like bacteria were observed adhering to the surrounding FAE.


SEM of FAE after loop infection with L. monocytogenes. Occasionally small areas of damage could be observed within the FAE, with numerous bacteria attached to the basal lamina. Also, segmented filamentous bacteria can be observed. The white square in image A indicates the enlarged area shown in image B.

3.3 Infection of M cell-like cells grown in culture by L. monocytogenes

The recently reported method of inducing Caco-2 cells of human origin into M cell-like cells in vitro [6] was used to study in more detail the interaction of L. monocytogenes with FAE. Using TRITC-conjugated UEA-1, which stains the apical side of non-M cell-like cells in this system [6], and a polyclonal antibody against the bacteria, the distribution and localisation of bacteria over the monolayer could be visualised using confocal microscopy. As a positive control a strain of Y. enterocolitica was used which targets M cells ([4] and Autenrieth et al., unpublished data). Cells infected with Y. enterocolitica showed, as expected, explicit targetting of M cells (Fig. 3A,B). This was is stark contrast to L. monocytogenes-infected cells, which exhibited random distribution over, and invasion of, the monolayer (Fig. 3C,D). Thus, listeria attached to, and invaded, both enterocytes and M cells, but no preferential targeting of M cells was detected. Furthermore, attachment and invasion per se was rare. Analysis of infected monolayers by transmission electron microscopy showed a comparable picture of attachment and invasion of L. monocytogenes (data not shown). Using a simple gentamicin invasion assay, it could also be shown that co-cultures (i.e. with M cells) were even less permissive to L. monocytogenes than fully differentiated Caco-2 cells alone (data not shown), further supporting the notion that M cells are not crucial for invasion.


Confocal laser scanning microscopic images of a Caco-2 monolayer with induced M cells infected with Y. enterocolitica (A and B) or L. monocytogenes (C and D). Y. enterocolitica is seen preferentially attaching to the M cells in great numbers, whereas L. monocytogenes is seen only sporadically and randomly adhering to the monolayer. Caco-2 enterocytes are stained positively with UEA-1 (M cells are not stained), bacteria were stained with polyclonal antibodies specific to the bacteria.

4 Discussion

MacDonald and Carter [5] first showed that after oral inoculation of mice, L. monocytogenes associated predominantly with Peyer's patches of the small intestine. Since then there has been some debate as to whether L. monocytogenes preferentially attaches to and invades M cells, especially given that other intestinal pathogens invade via M cells, e.g. S. typhimurium[1,2], S. flexneri[3], and Y. enterocolitica[4]. The evidence for L. monocytogenes using M cells is patchy and conflicting. Hence, the main aim of this work was to determine whether M cells were important in the translocation process.

In a guinea pig model [11], L. monocytogenes was shown to enter villous epithelium from the apical side, indicating that L. monocytogenes can invade enterocytes, but was not seen to associate with crypt cells. However, this did not exclude the possibility of a more efficient invasion via M cells. In a rat ligated loop model Pron et al. [9] showed that attachment and invasion of the intestinal epithelium was a rare event and saw no specific interaction with M cells of the FAE, as did Marco et al. [12]. However, Jensen et al. [13] showed specific M cell interaction using a high dose in a murine ligated loop system.

Our observation that after oral infection a slightly greater number of L. monocytogenes were found in Peyer's patches, but very few bacteria were seen attaching or invading the FAE suggests a higher rate of invasion or better survival of the bacteria in the Peyer's patch tissue, or possibly both. Even after a high dose in the ileal loop system, few bacteria were seen attaching to or invading the FAE. A similar observation has been described by Pron et al. [9]. The infection of in vitro induced M cell-like cells [6] also showed an identical pattern – despite a high inoculum only few L. monocytogenes were to be seen in or on the monolayer. The non-permissiveness of the intestine suggests that a susceptible host may have certain predisposing factors, which explains why only few people that have consumed contaminated food contract listeriosis and the majority remain healthy [5,14].

Very occasionally, numerous bacteria were observed invading an M cell, causing significant apical membrane rearrangement, but no such alterations were seen in enterocytes. M cell membrane alteration induced by L. monocytogenes has also been reported by Jensen et al. [13]. However, it is not yet clear if this is an integral part of the L. monocytogenes invasion process. Nevertheless, in both our murine and in vitro systems, no preference for M cells was seen.

The brisk translocation (∼30 min) of bacteria into the blood, liver and spleen was unexpected as at least two cells have to be crossed by the bacteria to reach the blood: an enterocyte and an endothelial cell of the underlying blood capillary. However, the bacteria found in livers and spleens were only a very small percentage (≤0.001%) of the bacteria inoculated. Similarly, Pron et al. [9] reported that 15 min after injecting 109 bacteria into a rat ligated loop, between 104.4 and 105.5 bacteria (approx. 0.01% of the inoculum) could be found in the liver and spleen of these animals. It has also been commonly thought that L. monocytogenes invades macrophages and is transported around the body via lymph, reaching its target organs via parasitised phagocytes [15]. Although our results do not exclude this as a possibility, they suggest a quicker alternative route – directly via blood.

Another very interesting finding was that whenever a part of the intestinal epithelium was seen to be damaged, numerous bacteria were attached to the exposed basement membrane. These areas of epithelial damage were seldom seen, raising the question of whether the damage was induced by L. monocytogenes, as suggested by Jensen et al. [13], or was present before inoculation occurred. L. monocytogenes can bind extracellular matrix proteins (e.g. fibronectin [16], and heparan sulfate proteoglycans [17]) which is also a common trait of many other pathogens [18,19]. Therefore, this observation may explain the apparently random nature of human listeriosis in otherwise healthy individuals [14] in that all humans periodically experience slight mucosal damage which is normally quickly repaired [20], and hence a damaged individual may be at more risk of an L. monocytogenes infection.

The data herein provide compelling evidence that, in contrast to other pathogens (for reviews see [21,22]), L. monocytogenes does not need M cells to invade the intestinal epithelium. Moreover, listeria can invade the apical side of intestinal epithelial cells; however, attachment to the apical side of epithelium is rare, or the bacteria are only loosely associated. This is in contrast to listeria's ability to attach to exposed basement membrane.


We thank Sonja Preger for excellent technical assistance, Lars Greiffenberg for printing expertise, and Guido Dietrich, Universität Würzburg, for the rabbit polyclonal anti-listeria antibody.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
View Abstract