Murine Norovirus Transcytosis across an In Vitro Polarized Murine Intestinal Epithelial Monolayer Is Mediated by M-Like Cells

Abstract

Noroviruses (NoVs) are the causative agent of the vast majority of nonbacterial gastroenteritis worldwide. Due to the inability to culture human NoVs and the inability to orally infect a small animal model, little is known about the initial steps of viral entry. One particular step that is not understood is how NoVs breach the intestinal epithelial barrier. Murine NoV (MNV) is the only NoV that can be propagated in vitro by infecting murine macrophages and dendritic cells, making this virus an attractive model for studies of different aspects of NoV biology. Polarized murine intestinal epithelial mIC_cl2 cells were used to investigate how MNV interacts with and crosses the intestinal epithelium. In this in vitro model of the follicle-associated epithelium (FAE), MNV is transported across the polarized cell monolayer in the absence of viral replication or disruption of tight junctions by a distinct epithelial cell with microfold (M) cell properties. In addition to transporting MNV, these M-like cells also transcytose microbeads and express an IgA receptor. Interestingly, B myeloma cells cultured in the basolateral compartment underlying the epithelial monolayer did not alter the number of M-like cells but increased their transcytotic activity. Our data demonstrate that MNV can cross an intact intestinal epithelial monolayer in vitro by hijacking the M-like cells' intrinsic transcytotic pathway and suggest a potential mechanism for MNV entry into the host.

INTRODUCTION

Human noroviruses (HuNoVs) are genetically diverse, environmentally stable, highly infectious viruses that infect their host via the fecal-oral route and aerosolization (1, 2). They are the causative agents of most nonbacterial infectious gastroenteritis worldwide (3–5). HuNoV infections spread rapidly, and outbreaks often take place in closed or semiclosed settings where communities gather (e.g., nursing homes, schools, hospitals, restaurants, and cruise ships) (6–8). Annually, HuNoVs cause an estimate of 21 million cases of acute gastroenteritis and 800 deaths in the United States alone (9, 10). Despite being a major public health concern, the inability to culture HuNoVs in vitro (11, 12) and lack of a small animal model for oral infection (13) have limited our progress in understanding NoV biology. Nevertheless, the discovery of the first murine-specific NoV (MNV), which is highly homologous to its human counterpart and can efficiently replicate in cell culture and in a small animal, provides the means to study NoV biology in detail (14–16).

The early events during viral infection are essential for a productive replication in the host, but little is known about this step during NoV infection. Particularly, how NoVs cross the epithelial barrier to reach their susceptible target cells remains unclear. Since MNV efficiently replicates in macrophages and dendritic cells in vitro (15) and in mice (14), the goal of this study was to understand how MNV interacts with the intestinal epithelium. MNV strains have high sequence similarity (>75%) but differ in their biological phenotypes (17, 18). For example, the fecally isolated MNV strains S99 and CR3 persist in wild-type mice for at least 35 days (17, 19). In contrast, MNV-1 causes acute infections in mice, and virus is not detectable in fecal contents after 7 days postinfection (dpi) (17). Persistence and colonic tropism mapped to a single amino acid residue within the nonstructural protein NS1/2 (20). Further differences between virus strains are observed in culture with respect to carbohydrate interaction. MNV-1 and S99 binding to murine macrophages is dependent on terminal sialic acid residues of the ganglioside GD1a, N-linked, or O-linked glycoproteins, while CR3 binding requires only N-linked glycoproteins (21, 22). Although multiple studies have elucidated aspects of the multistep process by which MNV enters permissive macrophages (21–25), how the virus crosses the intestinal epithelial barrier to reach susceptible macrophages and dendritic cells in the first place is unknown.

The intestinal tract comprises multiple types of intestinal cells, including epithelial cells and microfold (M) cells. M cells are specialized epithelial cells usually associated with the follicle-associated epithelium (FAE) overlaying the Peyer's patches where mucosa-associated lymphoid tissues are organized. These cells routinely sample diverse antigens along the entire mucosal surface for immune surveillance, including microorganisms and inert particles (e.g., latex beads) (26–28). Over the years, researchers have taken advantage of established in vitro FAE models for gaining a better understanding of the mechanisms required for enteric pathogen entry into or across the intestinal epithelium. A fraction of the in vitro polarized intestinal epithelial cells acquire characteristics that resemble those of M cells (i.e., uptake of particulate antigens) and show increased uptake of fluorescently labeled polystyrene latex beads after coculture with B cells or Peyer's patch-derived lymphocytes (29–31). Thus, pathogen interaction with M-like cells can also be studied in these polarized intestinal epithelial monolayers (29–33). For example, poliovirus translocates from the apical to the basolateral compartment in a temperature-dependent manner when polarized Caco-2 cell monolayers are cocultured with Peyer's patch lymphocytes to induce M-like cells (34). Another study demonstrated that a human immunodeficiency virus type 1 (HIV-1) strain tropic for the chemokine receptor CXCR4 (but not for CCR5) infects and is transported across polarized Caco-2 monolayers cocultured with B cells in a receptor-dependent manner (35). In addition, human T cell leukemia virus type 1 (HTLV-1) crosses polarized Caco-2 cell monolayers without disruption of tight junctions or infection of the epithelium to productively infect dendritic cells in the basolateral compartment (36).

The current study focused on the interaction of the murine enteric virus MNV with polarized murine intestinal epithelial cells (mIC_cl2 cells) as an in vitro FAE model system to determine whether MNV invades and/or crosses a polarized intestinal epithelium. The mIC_cl2 cell line, when grown on permeable filters, forms polarized cells with tight junctions and conserves main features of small intestine crypt cells (29, 30). Here we demonstrate that MNV traffics across the polarized cell monolayer using M-like cells without replicating or disrupting tight junctions. Addition of B myeloma cells in the cultures did not alter the numbers of M-like cells but instead increased the transcytotic activity of M-like cells. These results suggest that M cells may be a gateway for MNV invasion of the host.

MATERIALS AND METHODS

Cell culture.

The mIC_cl2 cells were generously provided by A. Vandewalle, INSERM, Paris, France (29). Cells were maintained in 75-cm² flasks (Falcon; BD Labware) as described previously (29) and used from passage ∼70 to 90. The Ag8.653 cell line was generously provided by S. Lundy, University of Michigan, Ann Arbor, MI, USA, and was maintained as described previously (37). The murine macrophage RAW 264.7 cell line was maintained as described previously and used for plaque assays (15).

Virus stocks and plaque assays.

The plaque-purified MNV-1 clone (GV/MNV1/2002/USA) MNV-1.CW3 and the fecal isolates CR3 (GV/CR3/2005/USA) and S99 (GV/S99/Berlin/2006/DE) were used at passage 6 for all experiments (17, 19). Viral titers were quantified by plaque assay after visualizing plaques by staining cells with a 0.01% neutral red solution in phosphate-buffered saline (PBS) for 1 to 3 h as previously described (15, 38).

Transcytosis experiments.

mIC_cl2 cells were plated at a density of 10⁶ cells/well on a polyester membrane filter of a 12-well transwell permeable support (3-μm pore size; Costar). Cells were cultured for 10 to 14 days as described previously (29, 30) until the transepithelial resistance (TER) was >250 Ω × cm² using a volt-ohm meter device (World Precision Instruments, FL) (Fig. 1A). For cocultures, the mouse myeloma B cell line Ag8.653 (ATCC, Manassas, VA) was added at 10⁶ cells/well on the bottom of the transwell in mIC_cl2 medium on day 10 of the mIC_cl2 culture. In this experimental setup, Ag8.653 cells did not come into direct contact with polarized mIC_cl2 monolayers during culture, and monolayers were transferred to new wells before performing experiments (Fig. 2A). For transcytosis experiments, monolayers were washed three times with PBS before adding MNV-1, S99, or CR3 to the apical side for the indicated times. All compartments (apical medium, membrane with cells, and basolateral medium) were harvested and freeze-thawed once. Viral titers were determined by plaque assay.

Fig 1.

MNV-1 does not affect epithelial integrity. (A) Schematic of the in vitro intestinal epithelial cell studies. Murine intestinal epithelial cells (mIC_cl2) were plated in a transwell and allowed to polarize for 10 to 14 days before addition of MNV-1 to the apical surface at 37°C for 4 h to perform follow-up assays. (B) MNV-1, S99, or CR3 was added to the apical surface of polarized monolayers at an MOI of 10 or 50 PFU/cell and left for 4 h, and the results were compared to those for a mock lysate and a no-cell control. Transepithelial electrical resistance (TER) was measured after incubation for 4 h. (C to E) Representative ZO-1 immunostaining of polarized monolayers following 4 h of incubation with mock lysate (C) or MNV-1 (MOI of 10 PFU/cell) (D) or after 20 mM methyl-beta-cyclodextrin (MBCD) treatment for 1 h (E) by confocal microscopy. (F) Polarized monolayers were incubated with the indicated combinations of MNV-1 (MOI of 10 PFU/cell), 20 mM MBCD, and mock lysate. Following a 4-h incubation period, lucifer yellow dye was added to the apical side and left for 15 min, and absorbance in the basolateral medium was measured. The dotted line represents the limit of detection. Data are expressed as mean ± SEM for three independent experiments in duplicates.

Fig 2.

MNV-1 crosses a polarized intestinal epithelial monolayer in a saturable, temperature-dependent manner. (A) Murine intestinal epithelial cells (mIC_cl2) were placed in a transwell and allowed to differentiate for 10 to 14 days before addition of Ag8.653 B myeloma cells in the bottom of the transwell. After 3 days of coculture, the transwell was moved to a new well before adding MNV-1 to the apical surface at 37°C for 4 h to perform follow-up assays. (B) MNV-1 is transcytosed across intestinal epithelial monolayers in a saturable manner, remains infectious, and does not replicate in these cells. MNV-1 (MOI of 10 PFU/cell) was added to the apical side of polarized monolayers and incubated at 37°C for the times shown. Viral titers in each compartment were quantified by plaque assay. (C) MNV-1 does not replicate in polarized intestinal monolayers. Neutral red (NR) light-sensitive MNV-1 was added to the apical side of cocultured polarized monolayers and incubated at 37°C for the times shown. Viral titers in the basolateral compartment were quantified by plaque assay in the dark to measure total virus or with light exposure to measure replicated virus. (D) TERs from cultures with B myeloma cells (Ag8.653) were measured before and after MNV-1 addition at the indicated time points. (E) Polarized mIC_cl2 monolayers following coculture with B myeloma cells (Ag8.653) were incubated with MNV-1 at 4°C for the indicated times. Viral titers in each compartment were quantified by plaque assay. (F) Polarized monolayers were incubated at 4°C for the indicated times before adding lucifer yellow to the apical side for 15 min and measuring absorbance in the basolateral media. The dotted lines represent the limit of detection for each assay. Data are expressed as mean ± SEM for at least three independent experiments from duplicate wells.

To measure viral replication, a neutral red (NR) light-sensitive virus was generated and used as described previously (24). Briefly, polarized mICcl2 monolayers cocultured with B myeloma cells were incubated with NR light-sensitive MNV-1 in the dark for 4 and 24 h before harvesting all compartments (apical medium, membrane with cells, and basolateral medium) in the dark. Samples were freeze-thawed once, and duplicate plaque assays were performed either in the dark or in the light following a 10-min light exposure to measure total or replicated virus, respectively.

To test for tight junction integrity, lucifer yellow dye (Life Technologies) with a molecular mass of 457.24 Da was used as described previously (32). In the case of mIC_cl2 cocultures, the transwell insert was transferred to a new 12-well plate without B myeloma cells before performing the assay.

To measure the transcytotic activity of polarized monolayers cultured with or without Ag8.653 cells, monolayers were transferred to a new well and incubated with 2.5 × 10¹⁰ beads/ml of 200-nm fluorescently green polystyrene latex nanoparticles (Fluoresbrite, YG; Polysciences) for 0, 0.5, 1, 2, 4, and 8 h. The basolateral medium was collected for each time point, and fluorescence was measured by flow cytometry as described previously (39).

To measure the temperature dependence, polarized mIC_cl2 monolayers cultured with or without Ag8.653 B myeloma cells were incubated with MNV-1 for 0, 15, 30, 45, and 60 min at 4°C, and viral titers were measured by plaque assay.

Immunofluorescence analysis.

Immunostaining of the tight junction-associated protein ZO-1 was performed on 0.1% saponin-permeabilized mIC_cl2 cells as described previously (32). To enumerate M-like cells in polarized mIC_cl2 monolayers, 200-nm fluorescent polystyrene latex nanoparticles (i.e., microbeads) (Fluoresbrite; Polysciences) and IgA isolated from human colostrum (0.5 mg/ml; Sigma) were incubated on the apical side of the monolayer at 37°C for 30 min. Following a PBS wash, cells were fixed in 4% paraformaldehyde in PBS for 15 min and permeabilized with 0.1% Triton X-100 for 15 min. Monolayers were washed and incubated for 1 h with a 1:500 anti-human IgA fluorescein isothiocyanate (FITC)-conjugated antibody (F-9637; Sigma) in PBS containing 1 μg/ml DAPI (4′,6′-diamidino-2-phenylindole) at room temperature. After three washes with PBS, membranes were dissected from transwells, mounted with ProLong Gold antifade reagent containing DAPI (Invitrogen, Grand Island, NY), and processed for confocal microscopy as detailed previously (32).

To visualize MNV-1 within polarized mIC_cl2 monolayers, MNV-1 at a multiplicity of infection (MOI) of 100 PFU/cell was added to the transwells' apical surface together with 200-nm fluorescently labeled polystyrene latex nanoparticles at 37°C for 30 min. Cells were fixed and permeabilized as described above. Cells were blocked with 10% (vol/vol) fetal bovine serum (FBS) and 1% (vol/vol) normal goat serum (NGS) (Gibco) in PBS for 30 to 60 min. MNV-1 was detected by staining with the mouse monoclonal antibody A6.2, recognizing the MNV-1 capsid (15) (0.5 μg/ml in PBS containing 1 μg/ml DAPI), at room temperature for 1 h, followed by another 1 h of incubation with a 1:500 dilution of an Alexa 488-labeled secondary goat anti-mouse antibody (Invitrogen) in PBS containing DAPI. Membranes were then dissected from transwells, mounted, and processed for laser scanning confocal microscopy and z-stacks of 0.5- to 1.0-μm slides were obtained using the LSM software on a Zeiss confocal microscope. Immunofluorescence images were quantified from three to six regions of the monolayer each from three to four independent experiments using the scoring system of intensities by the Metamorph Premier v6.3 image analysis software (Molecular Devices, Downington, PA).

To visualize the transport of microbeads in each culture, monolayers were incubated for 30 min at 4°C as well as for 2 h or 4 h at 37°C, immunostained for ZO-1, and processed for confocal imaging as mentioned above.

Statistical analysis.

Data are presented as mean ± standard error of the mean (SEM). Statistical analysis was performed using Prism software version 5.01 (GraphPad Software, CA). The two-tailed Student t test was used to determine statistical significance.

RESULTS

MNV does not disrupt the integrity of an in vitro polarized intestinal epithelium.

To determine how NoVs overcome the intestinal epithelial barrier, we first investigated the effect of MNV on tight junction integrity in an in vitro model of the polarized intestinal epithelium (Fig. 1A). Toward that end, three MNV strains MNV-1, S99, and CR3, were added to the apical surface of polarized murine intestinal epithelial mIC_cl2 cell monolayers in a transwell system and left for 4 h at 37°C (Fig. 1A). Transepithelial resistance (TER), an indicator of tight junction integrity, was greater than 300 Ω × cm² and remained unaffected by incubation with any MNV strain, even at a multiplicity of infection (MOI) of 50 PFU/cell, compared with mock-treated monolayers (Fig. 1B), suggesting that the integrity of the epithelial barrier is not affected by incubation with MNV. In addition, MNV-1 did not alter localization of the tight junction-associated protein ZO-1 compared to that of the mock-treated control (Fig. 1C and D). As a control, monolayers were treated with methyl-beta-cyclodextrin (MBCD), a drug that at high doses disrupts tight junctions (40), and relocalization of ZO-1 staining to the cytoplasm was readily observed (Fig. 1E). Finally, lucifer yellow, a tight junction- and membrane-impermeable fluorescent dye (32), was added apically, and fluorescence was quantified in the basolateral compartment. No passive diffusion of the dye was observed after incubating monolayers with the three MNV strains or a mock-treated control (Fig. 1F). This was in contrast to wells without epithelial cells or monolayers treated with 20 mM MBCD, where lucifer yellow diffused into the basolateral compartment (Fig. 1F). Taken together, these data indicate that incubation with MNV does not disrupt tight junctions of an in vitro polarized intestinal epithelial monolayer.

Transcytosis of MNV is increased following cocultures with B myeloma cells.

Previous studies of pathogen interactions with polarized intestinal epithelial monolayers used cocultures of freshly isolated murine Peyer's patch lymphocytes or human Raji B lymphoblast-like cells with polarized human colonic Caco-2 or mouse small intestinal mIC_cl2 cells (29–31). Thus, to investigate whether MNV replicated in murine intestinal epithelial cells or crossed a polarized intestinal epithelial monolayer, polarized mIC_cl2 monolayers were cultured in the absence or presence of BALB/c Peyer's patch lymphocytes or Ag8.653 murine B myeloma cells in the basolateral compartment (Fig. 1A and 2A). MNV-1 (MOI of 10 PFU/cell) was added to the apical surface of these monolayers and incubated for defined intervals. Infectious particles in the apical and basolateral compartments and the membrane (i.e., cell-associated virus) were then quantified by plaque assay. Three to four log units of MNV-1 was transported across the monolayer alone or following coculture with either Peyer's patch lymphocytes or Ag8.653 cells (Table 1, Fig. 2B, and data not shown). Since mIC_cl2 monolayers cocultured with Ag8.653 cells were more stable and reproducibly increased MNV-1 transcytosis, the mIC_cl2-Ag8.653 system was adopted for the remainder of the studies. MNV-1 titers initially increased quickly in the basolateral compartment but stabilized around 4 h, indicating that transport through the monolayer occurred in a saturable manner (Fig. 2B). Virus that crossed the epithelial barrier remained infectious and was capable of infecting the murine macrophage cell line RAW 264.7 used for plaque assay. No significant increase in total MNV-1 titers occurred over 24 h in mIC_cl2 monolayers cultured with Ag8.653 cells or in Ag8.653 cells alone (Fig. 2B and data not shown), while increases are seen within 12 h in permissive macrophages and dendritic cells (15). To further verify the lack of MNV replication, a neutral red (NR) light-sensitive MNV-1 was used to distinguish input virus from replicated virus (24). Neutral red is a photoactivated dye, which is passively incorporated into viral particles. When particles are exposed to white light, the dye cross-links the viral genome and the capsid protein, rendering the virus noninfectious (41). Monolayers were incubated with the NR light-sensitive MNV-1 for 4 and 24 h. Viral titers in the basolateral compartment were measured in the dark (i.e., total virus) or following light exposure (i.e., replicated virus) and were below the limit of detection after light exposure despite detection of several log units of total virus (Fig. 2C). Taken together, these data demonstrate that MNV-1 does not replicate in mIC_cl2 or in Ag8.653 cells. As observed before, TER was not altered by MNV-1 addition, demonstrating that tight junctions remained intact and suggesting that MNV-1 was trafficked intracellularly (Fig. 2D).

Table 1.

Percentages of MNV-1 and S99 transcytosed to the basolateral compartment after 4 h

Sample	% Transcytosis (mean ± SEM)a
	MNV-1	S99
mICcl2 alone	0.003 ± 0.001	0.002 ± 0.001
Coculture	0.076 ± 0.040 (P = 0.0501)	0.023 ± 0.007 (P = 0.0257)

^aPercentages were calculated based on viral titers with the equation (basolateral/[apical + membrane + basolateral]) × 100. P values compare coculture with mIC_cl2 alone.

Transcytosis is a form of intracellular transport in polarized cells, which is inhibited at 4°C (34, 42). To determine whether MNV-1 transport across polarized monolayers was temperature dependent, MNV-1 was incubated with the apical side of the monolayer at 4°C for 0, 15, 30, 45, and 60 min (Fig. 2E and F). Tight junction integrity was monitored by measuring passive diffusion of lucifer yellow (Fig. 2F). No significant amount of virus was detected in the basolateral compartment at time points when tight junction integrity was intact (i.e., 0 to 45 min) (Fig. 2E and F). Taken together, these data demonstrate that MNV-1 transport across polarized monolayers occurs intracellularly by a temperature-dependent mechanism, suggestive of transcytosis.

To determine whether MNV strains with different persistence phenotypes are transported similarly across this polarized intestinal epithelial monolayer, MNV-1 or S99 was added to the apical side of the cells and incubated for 4 h (Fig. 3). Viral titers in each compartment were quantified by plaque assay. Transport of both MNV strains to the basolateral compartment was observed, but this was significantly increased in the cocultures compared with mIC_cl2 cells alone (Fig. 3 and Table 1). Tight junction integrity remained unaffected throughout the experiment based on monitoring TER (Fig. 3B and D). Taken together, these data indicate that MNV traffics intracellularly through the mIC_cl2 monolayer to reach the basolateral compartment, and this process is enhanced following coculture with Ag8.653 cells.

Fig 3.

Coculture with B myeloma cells increases MNV transcytosis. Cocultures were established as outlined for Fig. 2A. (A and C) Cocultures increase viral titers in the basolateral side. Cultures in the presence or absence of B myeloma cells were incubated with either MNV-1 (A) or S99 (C) (MOI of 10 PFU/cell) at 37°C for 4 h. Viral titers were quantified by plaque assay. (B and D) TERs were measured in the presence or absence of B myeloma cells (Ag8.653) before and 4 h after MNV-1 (B) or S99 (D) incubation. Results represent duplicates of at least three independent experiments. Data are expressed as mean ± SEM. *, P < 0.05; **, P < 0.01.

M-like cell numbers are similar in mIC_cl2 monolayers, but cocultured monolayers transcytose more efficiently.

The previously described presence of M-like cells in polarized monolayers following coculture (29–31, 43) prompted us to test whether the increase in MNV transcytosis following Ag8.653 coculture is due to an increase in M-like cell numbers within the mIC_cl2 monolayer. M cells are known to selectively bind and endocytose IgA or secretory IgA in its natural form or when it is added exogenously to mouse Peyer's patches (44). Also, M cells have the innate ability to take up fluorescently labeled beads (45). Therefore, cells cultured in the absence or presence of Ag8.653 B myeloma cells were analyzed for these two M cell-associated properties, i.e., the abilities to bind exogenous IgA and to take up fluorescently labeled beads (microbeads). Monolayers were incubated with fluorescently labeled microbeads or IgA isolated from human colostrum followed by a fluorescently labeled anti-IgA secondary antibody and analyzed by confocal microscopy (Fig. 4A to F). The number of cells positive for IgA and microbeads alone or positive for both was not significantly different in cultures with or without B myeloma cells (Fig. 4G). Approximately 2% of total cells were single positive for either IgA or microbeads, while ∼10% of total cells in each monolayer were double positive for both (Fig. 4G). These results suggest that IgA- and microbead-positive cells are present within polarized mIC_cl2 monolayers and that addition of B myeloma cells to the basolateral compartment does not increase the number of these cells in the monolayer. Taken together, the data demonstrate that the increase in MNV transcytosis following coculture with Ag8.653 is not due to an increase in M-like cell numbers within the mIC_cl2 monolayer.

Fig 4.

Polarized mIC_cl2 monolayers alone or from cocultures have similar numbers of M-like cells but increased transcytosis of particles. (A to F) Representative confocal images of cell monolayers from cultures without (mIC_cl2 alone) or with (cocultures) B myeloma cells stained for IgA (green) and microbeads (red), as well as the merge of both. DAPI was used to stain the nuclei. Arrowheads indicate individual double-positive cells. A 10-μm scale bar is shown in the upper right corner of each image. (G) Quantification of images. Images from three or four independent experiments each of three to six fields of view within independent monolayers were quantified using Metamorph. (H) Microbeads were added to the apical side of the cultures and left for the times shown, and basolateral medium was analyzed by flow cytometry. Results represent duplicates of at least nine independent experiments. Data are expressed as mean ± SEM. *, P < 0.05 **, P < 0.01.

An alternative hypothesis to changes in cell numbers that could lead to an increase in virus titers in the basolateral compartment following coculture is an increase in the rate of particle transport. To investigate this possibility and verify that mono- and cocultured polarized monolayers could transcytose particulate antigens, microbead uptake was measured over time in monolayers cultured with or without Ag8.653 cells by first quantifying microbeads in the basolateral compartment by flow cytometry (Fig. 4H). Microbeads were being transcytosed under both conditions. However, cocultures showed significantly increased numbers of microbeads in the basolateral chamber compared to mIC_cl2 cells alone.

The transport of fluorescent microbeads across polarized monolayers cultured with or without Ag8.653 B myeloma cells was further verified by confocal microscopy (Fig. 5 and 6). As anticipated, microbeads incubated with the apical side for 30 min at 4°C remained near the apical surface of the cells under both culturing conditions (Fig. 5A and 6A) because transcytosis is inhibited at this temperature. In contrast, microbeads were located near the transwell membrane in the basolateral portion of cells after 2 to 4 h of incubation at 37°C (Fig. 5B and C and 6B and C).

Fig 5.

Polarized mIC_cl2 monolayers alone transcytose microbeads in a time- and temperature-dependent manner. Polarized mIC_cl2 monolayers were incubated with microbeads (in red) for 30 min at 4°C (A) as well as for 2 h (B) or 4 h (C) at 37°C. ZO-1 (in green) was used to identify tight junctions. Dashed lines in the y-z and x-z projections correspond to the relative location of the membrane in the transwell insert. Representative z-stack images of 0.5-μm slices are shown on the left, while the corresponding z-stack sections are shown on the right. A white square indicates the x-y view shown on the left. In each gallery of images, the first image is near the apical (top) side, while the last image is near the bottom (membrane) side.

Fig 6.

Polarized mIC_cl2 monolayers following cocultures with Ag8.653 B myeloma cells transcytose microbeads in a time- and temperature-dependent manner. Polarized mIC_cl2 monolayers cocultured with Ag8.653 cells were incubated with microbeads (in red) for 30 min at 4°C (A) as well as for 2 h (B) or 4 h (C) at 37°C, and with ZO-1 (in green), similarly to Fig. 5. Dashed lines in the y-z and x-z projections correspond to the relative location of the membrane in the transwell insert. Representative z-stack images of 0.5-μm slices are shown on the left, while the corresponding z-stack sections are shown on the right. A white square indicates the x-y view shown on the left. In each gallery of images, the first image is near the apical (top) side, while the last section is near the bottom (membrane) side.

Taken together, these data demonstrate that a fraction of mIC_cl2 enterocytes exhibit M-like cell properties (i.e., the ability to transport microbeads) in polarized monolayers cultured with or without Ag8.653 B myeloma cells. Microbead transport per se is independent of the presence of Ag8.653 B myeloma cells, but the process is improved by coculturing with the Ag8.653 B cells.

MNV transcytosis is specifically mediated by M-like cells.

To determine whether MNV is transcytosed by M-like cells, mIC_cl2 cell monolayers cocultured with Ag8.653 B myeloma cells were incubated with fluorescently labeled microbeads and either MNV-1 or mock lysate for 30 min at 37°C and analyzed by confocal microscopy (Fig. 7). Virus was observed through the length of the cell alongside microbeads (Fig. 7A to C). As anticipated, mock-inoculated cells stained positive only for microbeads (Fig. 7D to F). Quantification of the immunofluorescence images indicated that the majority of immune-stained cells and approximately 10% of all cells were positive for both microbeads and MNV-1 (Fig. 7G). In contrast, only a few cells (∼2% of total cells) were positive for microbeads or MNV-1 alone (Fig. 7G). Thus, MNV transcytosis is mediated primarily by microbead-positive M-like cells present in the monolayer.

Fig 7.

MNV-1 transcytosis is mediated by M-like cells. Cocultures of mIC_cl2 and B myeloma cells were incubated for 30 min with microbeads and MNV-1 (MOI of 100 PFU/cell) or mock lysate and processed for confocal microscopy. (A to C) Representative z-stack image of 1.0-μm slices showing staining of MNV-1 (green) (A) and microbeads (red) (B) and the merge (yellow) (C). DAPI was used to stain the nuclei. A 10-μm scale bar is shown in the upper left corner of each image. (D to F) Representative z-stack image of 0.5-μm slices incubated with mock lysate and microbeads. (A and D) The dashed line on the y-z and x-z projections indicates the relative location of the membrane from the transwell insert. (G) Quantification of confocal images. Three to six different regions in the monolayer were quantified using Metamorph from three independent experiments. Data are expressed as mean ± SEM. *, P < 0.05.

DISCUSSION

Many enteric viral pathogens have evolved strategies to infect their host by hijacking or circumventing intestinal host defenses. Studies aimed at understanding the initial step of enteric virus infection, i.e., of overcoming the intestinal epithelial barrier, can identify potential targets for intervention. In the case of NoVs, not much is known about how NoVs initiate a productive infection in the host. Here we used the murine enteric virus MNV and a murine intestinal epithelial cell line to point to M-like cells as a cell type mediating MNV transport across the intestinal epithelium in the absence of viral replication.

Our data demonstrated that MNV does not disrupt the epithelial integrity of polarized intestinal epithelial cells in vitro for at least 24 h as measured by staining for ZO-1, monitoring TER, and passive diffusion of lucifer yellow dye (see Fig. 1). In contrast, symptomatic human norovirus infections exhibit reduced expression of tight junction proteins, increased epithelial cell apoptosis, and decreased TER, resulting in epithelial barrier dysfunction in duodenal biopsy specimens from infected individuals (46). The inability of MNV to replicate in intestinal epithelial cells and to disrupt tight junctions in this in vitro model may also explain why MNV-infected wild-type mice show no overt symptoms of diarrhea (16, 17).

In addition, MNV remained infectious after being transcytosed by M-like cells to the basolateral compartment of in vitro polarized intestinal epithelial monolayers. This process was enhanced by cocultures with Ag8.653 B myeloma cells, although MNV did not replicate in these cells (Fig. 2 and 4 and data not shown). Similar findings were made with HTLV-1, which show that HTLV-1 remains infectious following transcytosis across a polarized Caco-2 monolayer but does not replicate in these cells (36). The average percentage of MNV transcytosis across polarized monolayers was less than 0.1%. While this may reflect an inefficiency of the in vitro system, others have reported a similar level of transcytosis for poliovirus (∼0.2%) across polarized Caco-2 monolayers (34). For enteric bacteria, transcytosis frequency is more variable, with values of between 0.1 and 10% (31, 47). While MNV-1 colocalized with the majority of microbead-positive M-like cells present in the monolayer, some microbead-positive cells without MNV-1 were also observed (Fig. 7G). This result suggested that some of the M-like cells do not internalize virus. Furthermore, the observation of cells single positive for MNV-1 suggested that microbeads may not label all M-like cells in the monolayer or that a small percentage of MNV-1 may be transcytosed by epithelial cells. We favor the former hypothesis that MNV-1 is taken up by M-like cells not stained by microbeads because of reports that, in vivo, certain immune stimuli or a Salmonella enterica serovar Typhimurium virulence factor can cause phenotypic transdifferentiation and lead to the generation of new M cell subsets without cell division (48–50). In this scenario, microbead staining would not occur in all stages of M-like cell differentiation, and other markers may be required to identify those M-like cells.

Using IgA binding and microbead uptake as properties of M-like cells, we demonstrated that the presence of Ag8.653 B myeloma cells in in vitro polarized intestinal epithelial cell monolayers did not significantly alter M-like cell numbers compared to mIC_cl2 cultures without Ag8.653 B myeloma cells. Instead, coculture with Ag8.653 cells increased the transport of particulate antigens across the monolayer (Fig. 4) despite the lack of direct contact between Ag8.653 B myeloma cells and the intestinal epithelial cells. One potential explanation for the different levels of transcytosis might lie in the maturation state of the M-like cells under each culture condition, which may be influenced by secreted factors from Ag8.653 B myeloma cells. In support of this hypothesis are findings that transcytosis is an acquired property of M cells as they mature (51, 52) and that secreted factors influence this process (39, 53). For example, macrophage migration inhibitory factor (MIF) secreted by Raji B cells induces M-like cell conversion in in vitro polarized Caco-2 cell monolayers, and MIF-deficient mice fail to upregulate M cell-mediated antigen sampling upon bacterial challenge (39). Furthermore, CD137 (a tumor necrosis factor [TNF] superfamily receptor member, TNFRSF9) is highly induced in intestinal epithelial monolayers, and CD137-deficient mice show abnormal M cell differentiation and defects in particle transcytosis (53). In addition, M cell differentiation is regulated by the ETS transcription factor Spi-B (51, 54), and ectopic expression of SPIB and EHF (encoding another ETS transcription factor) partially substituted for Raji B cell stimulated signals in differentiated TC7 cells (a Caco-2 cell subclone) (55). Interestingly, when we cocultured mIC_cl2 monolayers in conditioned medium from the Ag8.653 B myeloma cells instead of Ag8.653 cells themselves, a trend (although not statistically significant) of increased MNV-1 transcytosis compared to that for mIC_cl2 cells cultured without conditioned medium was observed (data not shown). Thus, one potential way that coculturing with Ag8.653 cells increases M-like cell transcytosis is via the secretion of cytokines that can lead to TNF/lymphotoxin signaling and/or the expression of ETS transcription factors. Further studies are needed to test this model and identify secreted factors important in this context.

Taken together, our work demonstrates that M-like cells mediate MNV transport across polarized murine intestinal epithelial monolayers but do not support viral replication. This suggests that M cells are a likely gateway for MNV entry into the host. Support for the importance of M cells in vivo during the establishment of a productive MNV infection comes from the observation that mice depleted of M cells have significantly lower MNV titers than isotype control depleted mice (M. B. Gonzalez-Hernandez et al., submitted for publication). Our study further demonstrates that MNV transcytosis is saturable in this in vitro murine intestinal epithelial model, indicating the presence of a receptor, and that addition of B myeloma cells increases the transcytotic activity of the polarized monolayer. Establishment of this in vitro murine FAE model provides a foundation for future studies to identify an M cell receptor for MNV and may reveal critical targets for the development of effective norovirus vaccines because targeting M cells is one approach to elicit effective mucosal immune responses (56).