- Email: [email protected]
Preferential interaction of Salmonella typhimurium with mouse Peyer's patch M cells
© INSTITUT PASTEUR/ELSEVIER
Res. Microbiol.
Pads 1994
1994, 145, 543-552
Preferential interaction of Salmonella typhimurium with mouse Peyer's patch M cells M.A. Clark (*), M.A. Jepson, N.L. Simmons and B.H. Hirst
Gastroimestinal Drug Delivery Research Centre and Department of Physiological Sciences, University of Newcastle upon Tyne, Medical School, Newcastle upon Tyne, NE2 4HH (England)
SUMMARY
We have used a mouse Peyer's patch gut loop model to investigate the role of the intestinal membranous epithelial (M) cePs in the pathogenesis of Salmone//a tlc~imurium. These specialized antigen sampling cells are located in the follicle-associated epithelium (FAE) overlying the isolated and aggregated lymphoid follicles in the small and large intestines. Our studies have demonstrated that S. typhimurium adheres more frequently to the Peyer's patch FAE cells than to the villous enterocytes and that, wilt,,-. the FAE, this bacterium preferentially interacts with the I~, cells. Quantitative lig~.~ microscopic studies, using the lectin Ulex europaeus 1 (UEA1) to identify M ceils, revealed that 34-fold more bacteria bound per unit area of M cells than per unit area of enterocyte. ~r~hin a 30-min incubation period, some M cells had dearly been invaded by the Salmonella. We therefore propose that M cells are a major route by which S. typh[murium penetrates the intestinal epithelial barrier. Bacterial adhesion to M ceils occurred in a non-uniform pattern, suggesting the existence of M-ce.__~ _The__ interaction of S. typhimurium with mouse Peyer's patch M ceils was accompanied by membrane ruffle formation and polymerized actin redistribution similar to that observed in cultured cell lines infected by this bacterium. This study emphasizes the suitability of Salmonella as an oral vaccine delivery system since, by proferenfially interacting w~h the M cells, these bacteria are targeted to sites where cells of the immune system are concentrated.
Key-words: Gut, Salmonella, Adhesion, Invasion, M cell, Follicle-associated epithelium, Peyer's patch, Salmonella typhimurium; Ulex europaeus 1, Membrane ruffles, Actin, Mouse, Oral vaccination.
INTRODUCTION
Salmonella infection are the ileal Peyer's patches
Pathogens transmitted by the oral route must cross the intestinal epithelial barrier prior to the induction of syst¢mic disease. The initial sites of
and possibly also the caecal lymphoid patches (Carter and Collins, 1974). The follicle-associated epithelium (FAE) overlying these lymphoid tissues is characterized by the presence of
Submitted March 3, 1994, accepted April 25, 1994. (*) Corresponding author.
M.A. CLARK ET AL
the specialized antigen sampling membranous epitheb~al or microfold (M) cells. These cells are e x ~ o ~ed by a number of systemically active pathogens to gain access to the host tissues (recent M cell reviews include Tile,,, 1991 ; A m ~ rongen ~,.d l.angermann, 1992; Neutra and Kraebenbuki, 199.2). Aa electron microscopic study perfo~aed on a m o u ~ ligated ileal loop model dedrr~strated that M cells are the primary site of S. ~phi adherence (Kohbata et at, 1986), but at present it is unclear whether other Salmonella species -~-vhlady target to these cells. Cultured mammalian cell lines have been used as in vitro m e d e l s to study the m e c h a n i s m s ~-eslxmsible for the interaction of Salmonella with intestinal epithelial cells. Many of these studies have been based on S. ~phimurium, the causative agent o f mouse typhoid fever. In vitro studies have demonswated that Salmonella adhesion and invasion are independently regulated events. Invasion is associated with degeneration of the microvilli and the formation o f cell membrane protuberances termed membrane ruffles. These changes are accompanied by the rearrangement o f cellular actin (reviewed by Bliska et al., 1993). Membrane ruffling and actin redistribution are crucial for S. typhimurium invasion of cultured cell lines (Fi_n!ay et aL, !99! ; Ga!a'n eta!., 1992a; Ginocchio et a!., 1992; Francis et al., 1992, 1993). Membrane distortions similarly accompany the in vivo infection of gastrointestinal cells with Salmonella. For example, S. typhitour/urn infection of guinea pig (Takeuchi, 1967) and rabbit (Wallis et al., 1986) villous enterocytes is accompanied by degeneration of the naicrovilli, and the interaction of S. typhi with mouse Peyer's patch M cells is associated with loss of the microvilli, extrusion of the M cell cytoplasm and M cell destruction (Kohbata et al., 1986).
CLSM CSA EGF EPEC FAE Frrc LB
= = = = = = =
confocal laser scanning microscopy. common structural antigen. epidermal growth factor. emeropathogenic Escherichia coli. follicle-associated epithefium. fluorescein isothiocyanate. Luria-Bertani (medium).
In the present study, we have used a gut loop model to investigate the interaction of S. typhimur~um with mouse Peyer's patch FAE. Recently, we demonstrated that the lectin Ulex europaeus 1 (UEA1 ; specific for (x-L-fucose residues) is a positive marker of mouse Peyer's patch M cells (Clark et al., 1993). This has enabled us to quantify S. typhimurium adhesion to M cells at the light microscopic level. Our studies demonstrate that S. ~.phimurium preferentially interacts with these cells. Bacterial adhesion to M cells is nonuniform, suggesting the existence of M cell subtypes. The interaction of S. typhimurium with M cells was accompanied by membrane ruffling activity and cellular actin redistribution similar to that which accompanies S. typhimurium invasion of cultured cells.
MATERIALS
AND METHODS
Animals Adult female BALB/c mice housed under specific pathogen-free (SPF) conditions were fasted overnight prior to bacterial inoculation. IDltllgl~lr'liltl
S. typhimurium strain SL1344 is a mouse-virulent strain which was obtained from C.L. Francis, Stanford University, California. The bacterium was grown in a manner similar to that described by Francis et al. (1992) to enrich for invasive bacteria. Briefly, a single colony grown on LB agar was inoculated into 2 ml LB broth and incubated with agitation at 37°C for 7 h. From this starter culture, 10 3 bacteria were inoculated into 5 ml LB broth (in a 6-ml vial) and grown overnight (16 h) at 37°C without agitation. The bacteria were pelleted, washed twice and resuspended in phosphate-buffered saline
M PBS SEM SFB SPF TRITC UEA1
= = = = = = =
membranous epithelial or microfoid (cell). phosphate-buffered saline (pH 7.4). scanning electron microscopy. segmented filamentous bacterium. specific pathogen-free. tetramethylrhodamineisothiocyanate.
Ulex europaeus !.
S. TYPHIMURIUM AND MOUSE PEYER'S PATCH M CELLS pH 7.4 (PBS) to give a total bacterial count of 3 x 10 9 bacteria/ml (1.8-2.0 x 10 9 colony-forming units/ml).
Experimental protocol Mice were anaesthetized by intraperitoneal injection of 0.3 ml hypnorm/midazolam The abdomens were incised and gut loops (jejunal/ileal) containing one or more Peyer's patches created by the application of appropriate ligatures. Care was taken to minimize surgical trauma and to maintain an adequate blood supply to the ligated gut segments. The bacterial suspensions or PBS (controls) were inoculated into the loops using 25 gauge needles and the mice maintained on heating pads. The gut loops were harvested after 30 min and the mice culled by cervical dislocation and exsanguination. Harvested tissues were rinsed thoroughly in PBS. Tissues destined for examination by confocal laser scanning microscopy (CLSM) were fixed for 60 min either in methanol (- 20°C) or in a 2 % solution (in PBS) of freshly depolymerized paraformaldehyde (4°C). Paraformaldehyde-fixed tissues were subsequently permeabilized in 0.1% TritonX-100 (in PBS) for 20 rain. Tissues destined for examination by scanning electron microscopy (SEM) were fixed in 2 % glutaraldehyde (in 100 mM sodium phosphate buffer pH 7.3) at 4°C for 120 rain. After fixation, the tissues were rinsed in PBS and the villi microdissected away from the domes to aid subsequent examination of the EA~. After methanol or paraformaldehyde fixation, bound bacteria were localized by staining with 10 pg/ml goat anti-Salmonella common structural antigen (CSA-1) (Kirkegaard and Perry Laboratories Inc. ; Gaithersburg, USA) for 30 rain followed by fluorescein isothiocyanate (FITC)-conjugated rabbit anti-goat IgG (Sigma; Poole, UK) at a dilution of 1:100 for a further 30 rain. M cells in methanol-fixed tissues were then stained by immersion in 67 pg/ml tetramethylrhodamine isothiocyanate (TRITC)-conjugated UEA1 (Sigma) for 60 min. Polymerized actin was localized in paraformaldehyde-fixed tissues by immersion in 2.5 gg/ml TRITC-phalloidin (Sigma) for 60 rnJn. Stained tissues were mounted on slides in "Vectashield" mounting medium (Vector Laboratories; Peterborough, UK) and examined using a "BioRad MRC-600" confocal laser scanning imaging system equipped with a krypton/argon mixed gas laser. Glutaraldehyde-fixed tissues were processed for SEM using standard techniques (Clark et aL, 1993), mounted on stubs, coated with 15 am gold using a "'Polaron" sputter coater and examined with a "Cambridge $240" SEM.
545
RESULTS Distribution of S. typh/mur/um interaction with mouse Peyer's patch FAE cells Mouse Peyer's patch gdt loops incubated with S. typhimurium strain SL1344 for 30 min were stained with FITC-conjugated antibodies to localize the bacteria and examined by CLSM. Very large numbers of bacteria were trapped in mucus which was predominantly located between the villi but which was also found overlying some regions of both the villi and the domes. Very few bacteria were adherent to the villous enteroeytes (both Peyer's patch and non-Peyer's patch villi), 1 /"lki~ but larger numbers were adherent to the ~,,t: cells. The number of bacteria adherent to individual domes varied, but the majority of adherent bacteria were located towards the periphery of the domes. FITC-stained bacteria were never observed in control loops inoculated with PBS. We have previously demonstrated that the leetin UEAI binds selectively to mouse Peyer's patch M cells (Clark et al., 1993). Staining with this lectin revealed that the majority of FAEbound S. typhimurium bacteria were associated with UEAl-stained M cells (fig. 1). To quantify ba~t~al binding to the FAE, 50 regions of FAE representing 20 domes in 6 Peyer's patches obtained from 3 mice were selected at random and
imaged by CI.~M. The areas occupied by UEA1stained/unstained cells (representing M cells/enterocytes and goblet cells respectively) were determined and the bacteria associated with each cell type (either adherent to the cell surface or located immediately beneath the surface) enumerated (table I). Based on total bacterial counts, 7-fold more bacteria were associated with M cells than enterocytes/goblet cells. Adjustment of these figures to allow for the relative surface mr,as ~ a pied by these cell types revealed that 34-fold more bacteria bound per unit area of M cell than per unit area enterocyte/goblet cell. Examination of the FAE by CLSM revealed that adherent S. typh/mur/um were not distributed evenly amongst the M cells. Some M cells totally lacked adherent bacteria, while M cells associated with large numbers of Salmonella were ....~ groups .~,+~r+,.~l of-ten locaL,m_ in sm~n . . . ,~. ame~gst
M.A. CLARK ET AL.
Fig. 1. C L S M images of mouse Peyer's patch F A E after 30-rain incubation with S. typhimurium. Adherent Salmonella have been visualized by indirect staining with F1TC-antibodies (a, c and e) and M cells have been localized with T R I T C - U E A I (b, d and f). Panels a-b = projected Z series of a single region of FAE, demonslrating that FITC-stained Salmonella (a) are associated with U E A I stained M cells Ca). Panels c-f = high-power confocal optical sections of a single M cell: c and d are surface views; e and f are at 2 p m below the surface. The latter images show a single bacterium (e) which has invaded the M cell and is surrounded by a ring of UEA1 staining (f, arrow). Scale bars = 10pro.
Table I. Quantitative evaluation of S. ~phimurium strain SL 1344 binding to mouse Peyer's patch FAE cells. Surface area/ptm 2 IIX2)
No. adherent bacteria °l
No. adherent bacteria/ i,000 ~tm ~ a~
Total 2.9 × 105
E+G 2.4 × 105
M cells 5.3 x 104
Total 1,238
E+G 148
M cells 1,090
E+G 0.6
M cells 20.6
~l~ E + G = enterocytes and goblet cells. ~2~The surface areas were determined by computer analysis of CLSM images of UEAI-stained tissues.
re~ons of FAE containing M cells bearing few or no adherent bacteria. Thus, it appears that S. ~'phimurium selectively adheres to a subset of M cells. M cells bearing adherent bacteria were typically associated with a group of bacteria rather than a single bacterium (fig. la and b). Some bacteria had clearly invaded the M cells (fig. ]c-f). Thus, bacteria were able to invade M cells within the 30-min incubation period used in these studies. Adherent and invading bacteria were surrounded by regions o f intense UEA1 staining which appeared as rings in confocal optical sections (fig. le and f, arrows).
M e m b r a n e ruffle f o r m a t i o n
Mouse Peyer's patch FAE is predominantly composed of enterocytes, M cells and goblet cells (fig. 2a). These three cell types are readily identifiable by SEM: enterocytes are the predominant cell type and possess very regular, d e n s e l y p a c k e d m i c r o v i l l i ; M cells possess broader, shorter microvilli which are less regular and less densely packed; goblet cells possess even fewer microvilli, have bulges in their cytoplasmic membranes and are often associated with blebs of mucus.
S. TYPHIMURIUM AND MOUSE PEYER'S PATCH M CELLS S E M revealed that incubation of S. typhhnurium in mouse Peyer's patch gut loops resulted in a dramatic alteration in M-cell surface morphology (fig. 2b, c and d). Affected cells typically occurred in groups scattered a m o n g s t areas o f FAE which contained M cells displaying normal surface characteristics (fig. 2b). This distribution was consistent with the distribution o f bacteria observed by C L S M . Affected M cells displayed a
.
.
.
.
.
.
-
547
range o f surface m o r p h o l o g i e s . Infrequently, adherent bacteria were associated with a localized thickening and elongation o f the M-cell microvilli (fig. 2c). These c h a n g e s p r o b a b l y represented an early stage in the interaction o f S. ~pi~imurium with M cells. More typically, the normal M cell microvillus sla~cture was totally lost and r e p l a c e d by p r o m i n e n ! b u l g e s in the a p i c a l m e m b r a n e s (fig. 2d). S m a l l e r l e s i o n s w e r e
.~
[ W L . - . L - - - . m
Fig. 2. SEM images of mouse Peyer's patch FAE after 30-min incubation with PBS (a) or S. t),,phimurium (b, c and d). Panel a: mouse Peyer's patch FAE cells incubated with PBS display a normal morphology; no adherent bacteria or membrane ruffles are observed. Panel b: S. typhimurium binds to a subset of M cells which display membrane ruffling activity (lower right of image); M cells unassociated with adherent Sahnonella lack membrane ruffles and are morphologically normal (upper left of image). Panels c and d: high-power views of the morphological changes induced by S. typhimurium in mouse Peyer's patch M cells; microvillus elongation and thickening (c) at the site of bacterial attachment probably precedes membrane ruffle formation (d); membrane ruffles are typically doughnut- (d, lower left-hand les;on) or petal-shaped (d, right-hand lesion) and may have either rounded or sharp edges. Scale bars in a and b = 20 lam, in c = 2 lam, in d = 5 lam.
548
M.A. CLARK ET AL.
located at either the centre or the edge of the cell and consisted of a circular protrusion which frequently possessed a central depression (i.e. doughnut-shaped). Some cells possessed two or three such lesions, Larger protrusions occupied the entire cell surface and were more petal-like in appearance. S o m e o f the d o u g h n u t - s h a p e d lesions possessed rounded edges and were morphologically indistinguishable from the membrane ruffles ~ by S. typhimurium in polarized M D C K cells (Galen et al., 1992a; Ginocchio et al., 1992). The petal-shaped protrusions and also a p r o l x ~ o n of the doughnut-shaped lesions possessed sha,~, edges visualized by SEM as a double line. i a e s e sharp-edged protntsions were ~logically distinct from previously identified membrane ruffles. Bacteria were frequently, although not invariably, associated with the affected M cells and were often located in the central depression. Failure to observe bacteria on the remaining distorted M cells was probably due to concealment of the bacteria in the central depression, bacterial invasion or loss of adherent bacteria during processing for SEM. Bacteria were very rarely associated with either morphologically normal M cells or with enterocytes. Extremely few, if any, bacteria were observed adherent to the FAE of control loops inoculated ~ith
PR~
and
r n ~ m h r ~ n e ~ ,-~,f-t'h~ ~ * r *
~l~o.,~,
from these tissues (fig. 2a).
Actin redistribution
Examination of phalloidin-stained mouse P e y e r ' s patch FAE by C L S M revealed that S. ~.'phimuriam induced marked rearrangements in the distribution of polymerized actin (fig. 3). Since M cells are the primary target of these bacteria, these changes largely reflected changes in the actin distribution in M cells. This was confirmed by dual staining with TRITC-phalloidin and FITC-UEA 1 (fig. 3c and d). In affected cells the normal, relatively homogeneous distribution of brush-border-associated polymerized actin was replaced by dense actin accumulations which were frequently in the form of distinct rings (fig. 3b and c). The size and shape of these actin accumulations displayed some variation, but they
were consistent with the m e m b r a n e ruffles observed by SEM. Smaller actin rings were located at either the centre or the edge of the cell, some cells possessing two or three such rings. Larger rings consisted of a thin band of densely stained actin lying just internal to the cell boundaries. Bacteria were frequently associated with the actin rings (fig. 3a and b), though the position of the bacteria relative to the rings varied. Some cells possessing actin rings lacked adherent bacteria: these bacteria had presumably been lost during processing. Actin rings similar to those associated with adherent Salmonella were generally absent from control gut loops inoculated with PBS (fig. 3e). Much smaller actin rings were occasionally observed in both test and control FAE cells. These rings may represent attachment sites of commensal organisms as has been shown for segmented filamentous bacteria (SFB) in the mouse ileum (Jepson et al., 1993).
DISCUSSION
The present study demonstrates that, in ligated mouse ileal loops, S. typhimurium strain SL1344 primarily adheres to and invades Peyer's patch M cells. The closed intestinal loop was chosen as the model for study in order to achieve a high concentration of bacteria in the gut lumen. Since the incubation period was short (30 min), it is reasonable to assume that the intestinal mucosa was not significantly d a m a g e d during these experiments. We therefore propose that M cells are a major route by which S. t y p h i m u r i u m invades the intact host. This proposal supports the concept that, in addition to acting as antigen-sampling cells, M cells act as a portal for the invasion of systemically active pathogens. The intense UEAI staining surrounding M-cell-associated Salmonella suggested that these bacteria were endocytosed in membrane-bound vacuoles. This is consistent with M-cell uptake of other microorganisms and macromolecules (reviewed by Trier, 1991). This putative M-cell route is not the sole route of Salmonella penetration, since uptake via villous enterocytes has also been demonstrated (Takeuchi, 1967; Wallis et al., 1986).
S. TYPHIMURIUM
AND MOUSE
PEYER’S
PATCH
M CELLS
WI
Fig. 3. CLSM images of mouse Peyer’s patch FAE after 30-min incubation with S. typhimurium
(a-d) or PBS (e). Adherent Salmonella have been visualized with FlTC-antibodies (a), polymerized actin has been localized with TlUTC-phalloidin (b, c and e) and M cells have been localized with FlTC-UEAI (d). Panels a and b: projected Z series, demonstrating that FlTC-stained bacteria (a) are associated with cells containing prominent actin rings (b, arrows). Salmonella-associated phalloidin-stained actin rings (c) are located in UEAI-stained M cells (d). Tissues incubated with PBS lack actin-stained rings (e). Scale bars = 10 pm.
Recently, we demonstrated that mouse Peyer’s patch M cells may be identified by the selective binding of the lectin UEAl to a-L-fucose residues in the M-cell glycocalyx (Clark et al., 1993). In view of the fact that many bacteria bind to cell surface glycoconjugates (reviewed by Sharon and Lis, 1989), we proposed that M-cellspecific surface glycoconjugates may be responsible for the targeting of microorganisms to M cells. In the present study, we have demonstrated that S. typhimurium preferentially interacts with only a subset of M cells which were not distinguished by their UEAl-binding characteristics. It therefore seems unlikely that expression of a-Lfucose residues is the primary factor determining the pattern of S. typhimurium adherence. However, tie possibility that a-L-fucose residues are involved in S. typhimurium adhesion cannot be excluded, since additional factors such as glycocalyx structure and/or luminal contents such as mucus or food particles might limit bacterial access to the M-cell surface receptors. Variations in the M-cell apical membranes may also affect the distribution of Salmonella adherence. Previous investigations have demonstrated that M cells exhibit a range of ultrastructural characteristics (Bye et al,; 1984) and possibly also a range of
alkaline phosphatase activities (Smith et al., 1988). It may be argued, therefore, that M cells are not J single, homogeneous population but exist as a range of subtypes expressing different morphological and functional characteristics. The interaction of S. typhim~ri~
with mouse Peyer’s patch M cells resulted in dramatic changes in cell surface morphology and actin redistribution. These changes were analogous to those observed in cultured cell lines such as Caco2, Henle-407, HEp-2, MDCK and Swiss 3T3 fibrobtasts i&c&d with this bacterium (Fiiay et d., 1991; Gal&~ et aZ., 1992~1; Ginocchio et al., 1992 ; Francis et al., 1992, 1993 ; Jones et aZ., 1993). Changes in cell surface morphology are also observed in other in vivo models of Salmonella infection (Takeuchi, 1967 ; Kohbata et d, 1986; Wallis et al., 1986). We therefore propose that changes in cell surface morphology and actin redistribution are generalized phenomena which accompany S. typhimurium interaction with epithelial cells both in vitro and in vivo. Based on these criteria, it would appear that cultured cell lines are valid models for the study of S. typbrium interaction with gastrointestinal epithelial cells. Siice studies performed in vitrv hive dem-
550
M.A. CLARK ET AL.
oastrated that membrane ruffle formation and cellular actin redistribution correlate with SalmoneUa invasion (reviewed by Bliska et aL, 1993), we that these phenomena are similarly associatod with Salmonella invasion in vivo. In the present study, membrane ruffles induced by the interaction of S. ~phanurium with mouse Peyer's tmtr,h M cells more closely resembled the membrane ruffles associated with S. typhimurium invasion of cultured cell lines such as Caco-2, HEp-2 and M D C K (e.g. Finlay and Falkow, 1990; Francis et al., 1992) than the microvillus disruption accompanying Salmonella infection of villous entexocytes (I'akeuchi, 1967: Wallis et al., 1986). Since the Salmonella-induced changes were more extensive in cells possessing a less organized brush border and apical cytoskeleton (i.e. M cells and cultured cell lines), it is likely that the state of cell differentiation determines the extent of the morphological changes induced by Salmonella. Studies performed on cultured cell lines have demonstrated that membrane ruffles induced by Salmonella a~ae preferential sites for the interaction of subsequent bacteria (Francis et al., 1992, 1993). This may explain why we frequently observed groups of bacteria binding to single M cells. We occasionally observed bacteria adherent to M cells in the absence of membrane ruffle formarion or cellular actin redistribution. These bacteria may have only recently made contact with the M-cell surface or they may have been of a phenotype which was unable to elicit membrane ruffle formation or actin redistribution. Bacterial growth conditions and growth phase may affect the ability of Sahnonella to invade cultured cells and consequently their ability to induce membrane ruffles, anaerobically grown or logarithmic phase cultures being more invasive than aerobically grown or stationary phase cultures respectively (Ernst et al., 1990; Lee and Falkow, 1990 ; Francis et al., 1992). In the present study, the bacteria were grown in a manner which had previously been found to enrich for the invasive phenotype (Lee and Falkow, 1990; Francis et al., 1992). However, a recent study has demonstrated that the capacity of Salmonella to adhere to and invade cultured cells is related to the multiplicity of infection and is independent of bacterial growth phase (Kusters et al., 1993). Conse-
quently, the importance of the bacterial growth phase in the current in vivo model of infection is unclear. Membrane ruffling (accompanied by actin redistribution) is not limited to Salmonella invasion but is a generalized phenomenon associated with the activation of several different cell surface receptors including the epidermal growth factor (EGF) receptor (e.g. Brunk et al., 1976; Kadowaki et al., 1986). Ruffle formation is accompanied by a temporarily increased pinocytic activity which results in the passive entry of adherent, otherwise non-invasive species (Francis et al., 1993). There is, however, no evidence to suggest that the antigen-sampling capacity of M cells is based solely on this phenomenon. While microorganisms other than Salmonella induce changes in M-cell surface morphology including localized loss 6f microvilli and the formation of pseudopod-like cytoplasmic extensions (e.g. Fujimura, 1986; Owen et al., 1986), the formation of distinct membrane ruffles is apparently unique to Salmonella infection. The mechanisms by which Salmonella induce membrane ruffle formation in cultured cell lines are unclear (reviewed by Bliska et al., 1993). It wan
plol,~J~r.,-u U l a t UlC I : ~ U r
lC~ptol
lll~l.y p l a y
a
role in S. typhimurium invasion of cultured cells and in the induction of Salmonella-associated membrane ruffles (Galdn et al., 1992b ; Pace et al., 1993). However, the EGF receptor is not essential for S. typhimurium invasion or the induction of Salmonella-associated membrane ruffling acti, ity (Francis et al., 1993; Jones et al., 1993). In conclusion, we have demonstrated that, when incubated in mouse Peyer's patch gut loops, S. typhimurium strain SL1344 preferentially adheres to and invades a subset of M cells. Assuming these findings represent the situation in the intact host, M cells represent a major route by which S. typhimurium penetrates the gut. Previous studies have demonstrated that attenuated Salmonella c~.,Tying heterologous antigens are potentially very important as oral vaccine delivery systems (reviewed by Dougan, 1994). The present study emphasizes the suitability of Salmonella as agents for oral vaccine delivery since, by preferentially interacting with M cells, they
S. T Y P H I M U R I U M A N D M O U S E P E Y E R ' S PATCH M CELLS are tm'geted to sites where cells o f the i m m u n e system are concentrated.
551
M o t s - c l d s : lntestin, Salmonella, Adherence, Invasion, Cellule M, Epith61ium folliculaire, Plaque de Peyer, S a l m o n e l l a t y p h i m u r i u m : Ulex europaeus 1, Plissements membranaires, At:dneo Souris, Vaccination per os.
Acknowledgements We are grateful to Dr. T.A. Booth, Biomedical Electron Microscopy Unit, University of Newcastle upon Tyne, for assistance with the scanning electron microscopy. This work was supported under the LINK programme in Selective Drug Delivery and Targeting. funded by SERC/MRC/DTI and industry (SERC grant GR/F 09747). MAJ was also supported by a Wellcome Postdoctoral Research Fellowship (039684/Z/93/Z). Additional support was provided by the Royal Society for equipment.
Interaction pr~f4rentieHe de Salmonella typhimurium avec les cellules M des plaques de Peyer chez la souris Nous avons utilis6 un mod~le murin de plaques de Peyer pour 6tudier le r61e des cellules 6pith6!iales M dans la pathogen~se de l'infection due ~ Salmonella typhimurium. Ces cellules sp6cialis6es dans la collecte antig6nique sent localis6es dans l'6pith61ium folliculaire (FAE) de l'intestin grSle et .21__ uu colon. Nous avons d6montr6 que S. typhimurium adhere plus fr6quemment au FAE des plaques de Peyer q u ' a u x ent6rocytes villeux et que, dans le FAE, cette bact6de interagit pr6f6rentiellement avec les cellules M. Une analyse quantitative microscopique, utilisant la lectine UEA1 (Ulex europeus) pour identifier les cellules M, a r6v616 que le hombre des bact6ries fix6es sur les cellules M est 34 fois sup6rieur au hombre not6 sur les ent6rocytes. Une incubation de 30 wan pe..rmet de noter que des cellules M sent nettement envahies par Salmonella. C ' e s t pourquoi nous pensons que les cellules M repr6sentent une voie pr6f6rentielle d'invasion par .5". typhimurium ~ travers la harfi~re 6pith61iale intestihale. L'adh6rence bact6fienne aux cellules M n'est pas homog~ne, ce qui sugg~re l'existence de soustypes de cellules M. L'interaction de S. typhimurium a v e c les c e l l u l e s M des p l a q u e s de P e y e r s'accompagne d'ondulations membranaires et d'une redistribution d'actine polym6ris~e similaire h c e qui s ' o b s e r v e darts les cultures de lign6es cellulaires infect6es par cette bact6de. Nos observations font ressortir l'opportunit~ de la vaccination par voie orale puisque ces bact6ries, intetagissant pr6f6rentiellement avec les cellules M, peuvent ~tre dirig6es sur les sites o f / l e s cellules du syst~me immunitaire sent particuli~rement concentr~es.
References Amerongen, H.M. & Langermann, S. (1992), Antigen transport by intestinal M cells. Curt. Opin. Gastroenterology, 8, 983-987. Bliska, J.B., Gahin, J.E. & Falkow, S. (1993), Signal transduction in the mammalian cell during bacterial attachment and entry. Cell, 73, 903-920. Brunk, U., Schellens, J. & Westermark, B. (1976), Influence of epidermal growth factor (EGF) on ruffling activity, pinocytosis and proliferation of cultivated human glia cells. Exp. Cell Res., 103, 295-302. Bye, W.A., Allan, C.H. & Trier, J.S. (1984), Strucam~, distribution, and origin of M cells in Peyer's patches of mouse ileum. Gastroenterology, 86, 789-801. Carter, P.B. & Collins, F.M. (1974), The route of enteric infection in normal mice. J. Exp. Med., 139, 11891203. Clark, M.A., Jepson, M.A., Simmons, N.L., Booth, T.A. & Hirst, B.H. (1993), Differential expression of lectin binding sites defines mouse intestinal M cells. J. H/stochem. Cytochem., 41, 1679-1687. Dougan, G. (1994), The molecular basis for the virulence of bacterial pathogens: implications for oral vaccine development_ Mirrnhlnl, !6~_~2lS-9~4_ Ernst, R.K., Dombroski, D.M. & Merrick, J.M. (1990), Anaerobiosis, type I f'unbriae, and growth_ phase are factors that affect invasion of HEp-2 cells by Sa/monella typhimurium. Infect, lmmun., 58, 2014-2016. Finlay, B.B. & Falkow, S. (1990), Sa/monel/a interactions with polarised human intestinal CaCo-2 epithelial cells. J. Infect. Dis., I62, 1096-1106. Finlay, B.B., Ruschkowski, S. & Dedhar, S. (1991), Cytoskeletal rearrangements accompanying Sa/monella entry into epithelial cells. J. Cell Sci., 99, 283296. Francis, C.L., Ryan, T.A., Jones, B.D., Smith, SJ. & Falkow, S. (1993), Ruffles induced by Salmonella and other stimuli direct macropinocytosis of bacteria. Natu,.'e (Lend.), 364, 639-642. Francis, C.L., Starnbach, M.N. & Falkow, S. (1992), Morphological and cytoskeletal changes in epithelial cells occur immediately upon interaction with SoJmonella typhimurium grown under low-oxygen conditions. Molecular MicrobioL, 6, 3077-3087. Fujimura, Y. (1986), Functional morphology of microfold cells (M cells) in Peyer's patches. Phagocytosis and transport of BCG by M cells into rabbit Peyer's patches. Gastroenterologia Japonica, 21,325-335. Gahin, J.E., Ginocchio, C. & Costeas, P. (1992a), Molecular and functional characterisation of the Salmonella invasion gene InvA: homology of lnvA to members of a new protein family. J. Bacteriol., 174, 43384349.
552
M.A. C L A R K E T AL.
G ~ , . J.E., Pace, J. & Hayman, MJ. (1992b), Involvemeat of the ep~.mml growth factor receptor in the invasion of cultured mammalian cells by Salmonella O~mUmur/um.Nature (Lond.), 357, 588-589. Gim3cchio, C, Pace, J. & Gal,'m, J.E. (1992), Identification and molecular chatactedsafion of a Sa/monella typhinmr/um gene involved in triggering the internalisalion of salmonellae into cultured epithelial cells. Proc. Natl. AcaR Sc£ USA, 89, 5976-5980. Jepson, M.A., Clark, M.A., Sitnmons, N.L. & Hirst, B.H. (1993), Actin accumulation at sites of attachment of indigenous apathogenic segmented filamentous bacLeria to mouse ileal epithelial cells. Infect. immun., 61.4001-4004. Jones, B.D., Paterson, H.F., Hall, A. & Falkow, S. (1993), Sa/mone//a typh/mm~'um induces membrane ruffling by a growth factor-n~tor-independent mechanism. Proc. Nat/. Acad. Sc/. USA, 90, 10390-10394. Kadowaki, T., Koyasu, S., Nishida, E., Sakai, H., Takaku, F., Yahara, I. & Kasuga, M. (1986), Insulin-like growth factors, insulin, and epidermal growth factor cause rapid cytoskeletal reorganisation in KB cells. J. Biol. Chef&, 261, 16141-16147. K~l~ta, S., Yokoyama, H. & Yabuuchi, E. (1986), Cytopathogenic effect of Salmonella typhi GIFU 10007 on M cells of murine ileal Peyer's patches in ligated ileal loops : an ultrastructural study. Microbiol. /mmmlol., 30, 1225-1237. Kustcts, J.G., Muldets-lCremets, G.A.W.M., Van Doomik, C.E.M. & Van der Zeijsk B.A.M. (1993), Effects of multiplicity of infection, bacterial protein synthesis, and growth phase on adhesion to and invasion of human cell lines by Salmonella typhimurium` Infect. Immun., 61, 5013-5020.
Lee, C.A. & Falkow, S. (1990), The ability of Salmonella to enter mammafian cells is affected by bacterial growth state. Proc. Natl. Acad. Sci. USA, 87, 43044308. Neutra, M.R. & Kraehenbuh!, J.P. (1992), Transepithelial transport and mucosal defense. - - I. The role of M cells. Trends Cell Biol., 2, 134-138. Owen, R.L., Pierce, N.F., Apple, R.T. & Cray, W.C. (1986), M cell transport of Vibrio cholerae from the intestinal lumen into the payer's patches: a mechanism for antigen sampling and for microbial transepithelial migration. J. Infect. Dis., 153, 11081118. Pace, J., Hayman, M.J. & Gal~ln, J.E. (1993), Signal transduction and invasion of epithelial cells by S. typhimurium. Cell, 72, 505-514. Sharon, N. & l_is, H. (1989), Lectins as cell recognition molecules. Science, 246, 227-234. Smith, M.W., James, P.S., Tivey, D.R. & Brown, D. (1988), Automated histochemical analysis of cell populations in the intact follicle-associated epithelium of the mouse Peyer's patch. Histochem` J., 20, 443-448. Takeuchi, A. (1967), Electron microscope studies of experimental Salmonella infection. - - I. Penetration into the intestinal epithelium by Salmonella typhimurium` Am. J. Pathol., 50, 109-136. Trier, J.S. (1991), Structure and function of intestinal M cells. Gastroenterol. Clin. N. Amer., 20, 531-547. Wallis, T.S., Starkey, W.G., Stephen, J., Haddon, S.J., Osborne, M.P. & Candy, D.C.A. (1986), The nature and role of mucosal damage in relation to Salmonella typhimurium-induced fluid secretion in the rabbit ileum. Y. Med. Microbiol., 22, 39-49.
Res. Microbiol.
Pads 1994
1994, 145, 543-552
Preferential interaction of Salmonella typhimurium with mouse Peyer's patch M cells M.A. Clark (*), M.A. Jepson, N.L. Simmons and B.H. Hirst
Gastroimestinal Drug Delivery Research Centre and Department of Physiological Sciences, University of Newcastle upon Tyne, Medical School, Newcastle upon Tyne, NE2 4HH (England)
SUMMARY
We have used a mouse Peyer's patch gut loop model to investigate the role of the intestinal membranous epithelial (M) cePs in the pathogenesis of Salmone//a tlc~imurium. These specialized antigen sampling cells are located in the follicle-associated epithelium (FAE) overlying the isolated and aggregated lymphoid follicles in the small and large intestines. Our studies have demonstrated that S. typhimurium adheres more frequently to the Peyer's patch FAE cells than to the villous enterocytes and that, wilt,,-. the FAE, this bacterium preferentially interacts with the I~, cells. Quantitative lig~.~ microscopic studies, using the lectin Ulex europaeus 1 (UEA1) to identify M ceils, revealed that 34-fold more bacteria bound per unit area of M cells than per unit area of enterocyte. ~r~hin a 30-min incubation period, some M cells had dearly been invaded by the Salmonella. We therefore propose that M cells are a major route by which S. typh[murium penetrates the intestinal epithelial barrier. Bacterial adhesion to M ceils occurred in a non-uniform pattern, suggesting the existence of M-ce.__~ _The__ interaction of S. typhimurium with mouse Peyer's patch M ceils was accompanied by membrane ruffle formation and polymerized actin redistribution similar to that observed in cultured cell lines infected by this bacterium. This study emphasizes the suitability of Salmonella as an oral vaccine delivery system since, by proferenfially interacting w~h the M cells, these bacteria are targeted to sites where cells of the immune system are concentrated.
Key-words: Gut, Salmonella, Adhesion, Invasion, M cell, Follicle-associated epithelium, Peyer's patch, Salmonella typhimurium; Ulex europaeus 1, Membrane ruffles, Actin, Mouse, Oral vaccination.
INTRODUCTION
Salmonella infection are the ileal Peyer's patches
Pathogens transmitted by the oral route must cross the intestinal epithelial barrier prior to the induction of syst¢mic disease. The initial sites of
and possibly also the caecal lymphoid patches (Carter and Collins, 1974). The follicle-associated epithelium (FAE) overlying these lymphoid tissues is characterized by the presence of
Submitted March 3, 1994, accepted April 25, 1994. (*) Corresponding author.
M.A. CLARK ET AL
the specialized antigen sampling membranous epitheb~al or microfold (M) cells. These cells are e x ~ o ~ed by a number of systemically active pathogens to gain access to the host tissues (recent M cell reviews include Tile,,, 1991 ; A m ~ rongen ~,.d l.angermann, 1992; Neutra and Kraebenbuki, 199.2). Aa electron microscopic study perfo~aed on a m o u ~ ligated ileal loop model dedrr~strated that M cells are the primary site of S. ~phi adherence (Kohbata et at, 1986), but at present it is unclear whether other Salmonella species -~-vhlady target to these cells. Cultured mammalian cell lines have been used as in vitro m e d e l s to study the m e c h a n i s m s ~-eslxmsible for the interaction of Salmonella with intestinal epithelial cells. Many of these studies have been based on S. ~phimurium, the causative agent o f mouse typhoid fever. In vitro studies have demonswated that Salmonella adhesion and invasion are independently regulated events. Invasion is associated with degeneration of the microvilli and the formation o f cell membrane protuberances termed membrane ruffles. These changes are accompanied by the rearrangement o f cellular actin (reviewed by Bliska et al., 1993). Membrane ruffling and actin redistribution are crucial for S. typhimurium invasion of cultured cell lines (Fi_n!ay et aL, !99! ; Ga!a'n eta!., 1992a; Ginocchio et a!., 1992; Francis et al., 1992, 1993). Membrane distortions similarly accompany the in vivo infection of gastrointestinal cells with Salmonella. For example, S. typhitour/urn infection of guinea pig (Takeuchi, 1967) and rabbit (Wallis et al., 1986) villous enterocytes is accompanied by degeneration of the naicrovilli, and the interaction of S. typhi with mouse Peyer's patch M cells is associated with loss of the microvilli, extrusion of the M cell cytoplasm and M cell destruction (Kohbata et al., 1986).
CLSM CSA EGF EPEC FAE Frrc LB
= = = = = = =
confocal laser scanning microscopy. common structural antigen. epidermal growth factor. emeropathogenic Escherichia coli. follicle-associated epithefium. fluorescein isothiocyanate. Luria-Bertani (medium).
In the present study, we have used a gut loop model to investigate the interaction of S. typhimur~um with mouse Peyer's patch FAE. Recently, we demonstrated that the lectin Ulex europaeus 1 (UEA1 ; specific for (x-L-fucose residues) is a positive marker of mouse Peyer's patch M cells (Clark et al., 1993). This has enabled us to quantify S. typhimurium adhesion to M cells at the light microscopic level. Our studies demonstrate that S. ~.phimurium preferentially interacts with these cells. Bacterial adhesion to M cells is nonuniform, suggesting the existence of M cell subtypes. The interaction of S. typhimurium with M cells was accompanied by membrane ruffling activity and cellular actin redistribution similar to that which accompanies S. typhimurium invasion of cultured cells.
MATERIALS
AND METHODS
Animals Adult female BALB/c mice housed under specific pathogen-free (SPF) conditions were fasted overnight prior to bacterial inoculation. IDltllgl~lr'liltl
S. typhimurium strain SL1344 is a mouse-virulent strain which was obtained from C.L. Francis, Stanford University, California. The bacterium was grown in a manner similar to that described by Francis et al. (1992) to enrich for invasive bacteria. Briefly, a single colony grown on LB agar was inoculated into 2 ml LB broth and incubated with agitation at 37°C for 7 h. From this starter culture, 10 3 bacteria were inoculated into 5 ml LB broth (in a 6-ml vial) and grown overnight (16 h) at 37°C without agitation. The bacteria were pelleted, washed twice and resuspended in phosphate-buffered saline
M PBS SEM SFB SPF TRITC UEA1
= = = = = = =
membranous epithelial or microfoid (cell). phosphate-buffered saline (pH 7.4). scanning electron microscopy. segmented filamentous bacterium. specific pathogen-free. tetramethylrhodamineisothiocyanate.
Ulex europaeus !.
S. TYPHIMURIUM AND MOUSE PEYER'S PATCH M CELLS pH 7.4 (PBS) to give a total bacterial count of 3 x 10 9 bacteria/ml (1.8-2.0 x 10 9 colony-forming units/ml).
Experimental protocol Mice were anaesthetized by intraperitoneal injection of 0.3 ml hypnorm/midazolam The abdomens were incised and gut loops (jejunal/ileal) containing one or more Peyer's patches created by the application of appropriate ligatures. Care was taken to minimize surgical trauma and to maintain an adequate blood supply to the ligated gut segments. The bacterial suspensions or PBS (controls) were inoculated into the loops using 25 gauge needles and the mice maintained on heating pads. The gut loops were harvested after 30 min and the mice culled by cervical dislocation and exsanguination. Harvested tissues were rinsed thoroughly in PBS. Tissues destined for examination by confocal laser scanning microscopy (CLSM) were fixed for 60 min either in methanol (- 20°C) or in a 2 % solution (in PBS) of freshly depolymerized paraformaldehyde (4°C). Paraformaldehyde-fixed tissues were subsequently permeabilized in 0.1% TritonX-100 (in PBS) for 20 rain. Tissues destined for examination by scanning electron microscopy (SEM) were fixed in 2 % glutaraldehyde (in 100 mM sodium phosphate buffer pH 7.3) at 4°C for 120 rain. After fixation, the tissues were rinsed in PBS and the villi microdissected away from the domes to aid subsequent examination of the EA~. After methanol or paraformaldehyde fixation, bound bacteria were localized by staining with 10 pg/ml goat anti-Salmonella common structural antigen (CSA-1) (Kirkegaard and Perry Laboratories Inc. ; Gaithersburg, USA) for 30 rain followed by fluorescein isothiocyanate (FITC)-conjugated rabbit anti-goat IgG (Sigma; Poole, UK) at a dilution of 1:100 for a further 30 rain. M cells in methanol-fixed tissues were then stained by immersion in 67 pg/ml tetramethylrhodamine isothiocyanate (TRITC)-conjugated UEA1 (Sigma) for 60 min. Polymerized actin was localized in paraformaldehyde-fixed tissues by immersion in 2.5 gg/ml TRITC-phalloidin (Sigma) for 60 rnJn. Stained tissues were mounted on slides in "Vectashield" mounting medium (Vector Laboratories; Peterborough, UK) and examined using a "BioRad MRC-600" confocal laser scanning imaging system equipped with a krypton/argon mixed gas laser. Glutaraldehyde-fixed tissues were processed for SEM using standard techniques (Clark et aL, 1993), mounted on stubs, coated with 15 am gold using a "'Polaron" sputter coater and examined with a "Cambridge $240" SEM.
545
RESULTS Distribution of S. typh/mur/um interaction with mouse Peyer's patch FAE cells Mouse Peyer's patch gdt loops incubated with S. typhimurium strain SL1344 for 30 min were stained with FITC-conjugated antibodies to localize the bacteria and examined by CLSM. Very large numbers of bacteria were trapped in mucus which was predominantly located between the villi but which was also found overlying some regions of both the villi and the domes. Very few bacteria were adherent to the villous enteroeytes (both Peyer's patch and non-Peyer's patch villi), 1 /"lki~ but larger numbers were adherent to the ~,,t: cells. The number of bacteria adherent to individual domes varied, but the majority of adherent bacteria were located towards the periphery of the domes. FITC-stained bacteria were never observed in control loops inoculated with PBS. We have previously demonstrated that the leetin UEAI binds selectively to mouse Peyer's patch M cells (Clark et al., 1993). Staining with this lectin revealed that the majority of FAEbound S. typhimurium bacteria were associated with UEAl-stained M cells (fig. 1). To quantify ba~t~al binding to the FAE, 50 regions of FAE representing 20 domes in 6 Peyer's patches obtained from 3 mice were selected at random and
imaged by CI.~M. The areas occupied by UEA1stained/unstained cells (representing M cells/enterocytes and goblet cells respectively) were determined and the bacteria associated with each cell type (either adherent to the cell surface or located immediately beneath the surface) enumerated (table I). Based on total bacterial counts, 7-fold more bacteria were associated with M cells than enterocytes/goblet cells. Adjustment of these figures to allow for the relative surface mr,as ~ a pied by these cell types revealed that 34-fold more bacteria bound per unit area of M cell than per unit area enterocyte/goblet cell. Examination of the FAE by CLSM revealed that adherent S. typh/mur/um were not distributed evenly amongst the M cells. Some M cells totally lacked adherent bacteria, while M cells associated with large numbers of Salmonella were ....~ groups .~,+~r+,.~l of-ten locaL,m_ in sm~n . . . ,~. ame~gst
M.A. CLARK ET AL.
Fig. 1. C L S M images of mouse Peyer's patch F A E after 30-rain incubation with S. typhimurium. Adherent Salmonella have been visualized by indirect staining with F1TC-antibodies (a, c and e) and M cells have been localized with T R I T C - U E A I (b, d and f). Panels a-b = projected Z series of a single region of FAE, demonslrating that FITC-stained Salmonella (a) are associated with U E A I stained M cells Ca). Panels c-f = high-power confocal optical sections of a single M cell: c and d are surface views; e and f are at 2 p m below the surface. The latter images show a single bacterium (e) which has invaded the M cell and is surrounded by a ring of UEA1 staining (f, arrow). Scale bars = 10pro.
Table I. Quantitative evaluation of S. ~phimurium strain SL 1344 binding to mouse Peyer's patch FAE cells. Surface area/ptm 2 IIX2)
No. adherent bacteria °l
No. adherent bacteria/ i,000 ~tm ~ a~
Total 2.9 × 105
E+G 2.4 × 105
M cells 5.3 x 104
Total 1,238
E+G 148
M cells 1,090
E+G 0.6
M cells 20.6
~l~ E + G = enterocytes and goblet cells. ~2~The surface areas were determined by computer analysis of CLSM images of UEAI-stained tissues.
re~ons of FAE containing M cells bearing few or no adherent bacteria. Thus, it appears that S. ~'phimurium selectively adheres to a subset of M cells. M cells bearing adherent bacteria were typically associated with a group of bacteria rather than a single bacterium (fig. la and b). Some bacteria had clearly invaded the M cells (fig. ]c-f). Thus, bacteria were able to invade M cells within the 30-min incubation period used in these studies. Adherent and invading bacteria were surrounded by regions o f intense UEA1 staining which appeared as rings in confocal optical sections (fig. le and f, arrows).
M e m b r a n e ruffle f o r m a t i o n
Mouse Peyer's patch FAE is predominantly composed of enterocytes, M cells and goblet cells (fig. 2a). These three cell types are readily identifiable by SEM: enterocytes are the predominant cell type and possess very regular, d e n s e l y p a c k e d m i c r o v i l l i ; M cells possess broader, shorter microvilli which are less regular and less densely packed; goblet cells possess even fewer microvilli, have bulges in their cytoplasmic membranes and are often associated with blebs of mucus.
S. TYPHIMURIUM AND MOUSE PEYER'S PATCH M CELLS S E M revealed that incubation of S. typhhnurium in mouse Peyer's patch gut loops resulted in a dramatic alteration in M-cell surface morphology (fig. 2b, c and d). Affected cells typically occurred in groups scattered a m o n g s t areas o f FAE which contained M cells displaying normal surface characteristics (fig. 2b). This distribution was consistent with the distribution o f bacteria observed by C L S M . Affected M cells displayed a
.
.
.
.
.
.
-
547
range o f surface m o r p h o l o g i e s . Infrequently, adherent bacteria were associated with a localized thickening and elongation o f the M-cell microvilli (fig. 2c). These c h a n g e s p r o b a b l y represented an early stage in the interaction o f S. ~pi~imurium with M cells. More typically, the normal M cell microvillus sla~cture was totally lost and r e p l a c e d by p r o m i n e n ! b u l g e s in the a p i c a l m e m b r a n e s (fig. 2d). S m a l l e r l e s i o n s w e r e
.~
[ W L . - . L - - - . m
Fig. 2. SEM images of mouse Peyer's patch FAE after 30-min incubation with PBS (a) or S. t),,phimurium (b, c and d). Panel a: mouse Peyer's patch FAE cells incubated with PBS display a normal morphology; no adherent bacteria or membrane ruffles are observed. Panel b: S. typhimurium binds to a subset of M cells which display membrane ruffling activity (lower right of image); M cells unassociated with adherent Sahnonella lack membrane ruffles and are morphologically normal (upper left of image). Panels c and d: high-power views of the morphological changes induced by S. typhimurium in mouse Peyer's patch M cells; microvillus elongation and thickening (c) at the site of bacterial attachment probably precedes membrane ruffle formation (d); membrane ruffles are typically doughnut- (d, lower left-hand les;on) or petal-shaped (d, right-hand lesion) and may have either rounded or sharp edges. Scale bars in a and b = 20 lam, in c = 2 lam, in d = 5 lam.
548
M.A. CLARK ET AL.
located at either the centre or the edge of the cell and consisted of a circular protrusion which frequently possessed a central depression (i.e. doughnut-shaped). Some cells possessed two or three such lesions, Larger protrusions occupied the entire cell surface and were more petal-like in appearance. S o m e o f the d o u g h n u t - s h a p e d lesions possessed rounded edges and were morphologically indistinguishable from the membrane ruffles ~ by S. typhimurium in polarized M D C K cells (Galen et al., 1992a; Ginocchio et al., 1992). The petal-shaped protrusions and also a p r o l x ~ o n of the doughnut-shaped lesions possessed sha,~, edges visualized by SEM as a double line. i a e s e sharp-edged protntsions were ~logically distinct from previously identified membrane ruffles. Bacteria were frequently, although not invariably, associated with the affected M cells and were often located in the central depression. Failure to observe bacteria on the remaining distorted M cells was probably due to concealment of the bacteria in the central depression, bacterial invasion or loss of adherent bacteria during processing for SEM. Bacteria were very rarely associated with either morphologically normal M cells or with enterocytes. Extremely few, if any, bacteria were observed adherent to the FAE of control loops inoculated ~ith
PR~
and
r n ~ m h r ~ n e ~ ,-~,f-t'h~ ~ * r *
~l~o.,~,
from these tissues (fig. 2a).
Actin redistribution
Examination of phalloidin-stained mouse P e y e r ' s patch FAE by C L S M revealed that S. ~.'phimuriam induced marked rearrangements in the distribution of polymerized actin (fig. 3). Since M cells are the primary target of these bacteria, these changes largely reflected changes in the actin distribution in M cells. This was confirmed by dual staining with TRITC-phalloidin and FITC-UEA 1 (fig. 3c and d). In affected cells the normal, relatively homogeneous distribution of brush-border-associated polymerized actin was replaced by dense actin accumulations which were frequently in the form of distinct rings (fig. 3b and c). The size and shape of these actin accumulations displayed some variation, but they
were consistent with the m e m b r a n e ruffles observed by SEM. Smaller actin rings were located at either the centre or the edge of the cell, some cells possessing two or three such rings. Larger rings consisted of a thin band of densely stained actin lying just internal to the cell boundaries. Bacteria were frequently associated with the actin rings (fig. 3a and b), though the position of the bacteria relative to the rings varied. Some cells possessing actin rings lacked adherent bacteria: these bacteria had presumably been lost during processing. Actin rings similar to those associated with adherent Salmonella were generally absent from control gut loops inoculated with PBS (fig. 3e). Much smaller actin rings were occasionally observed in both test and control FAE cells. These rings may represent attachment sites of commensal organisms as has been shown for segmented filamentous bacteria (SFB) in the mouse ileum (Jepson et al., 1993).
DISCUSSION
The present study demonstrates that, in ligated mouse ileal loops, S. typhimurium strain SL1344 primarily adheres to and invades Peyer's patch M cells. The closed intestinal loop was chosen as the model for study in order to achieve a high concentration of bacteria in the gut lumen. Since the incubation period was short (30 min), it is reasonable to assume that the intestinal mucosa was not significantly d a m a g e d during these experiments. We therefore propose that M cells are a major route by which S. t y p h i m u r i u m invades the intact host. This proposal supports the concept that, in addition to acting as antigen-sampling cells, M cells act as a portal for the invasion of systemically active pathogens. The intense UEAI staining surrounding M-cell-associated Salmonella suggested that these bacteria were endocytosed in membrane-bound vacuoles. This is consistent with M-cell uptake of other microorganisms and macromolecules (reviewed by Trier, 1991). This putative M-cell route is not the sole route of Salmonella penetration, since uptake via villous enterocytes has also been demonstrated (Takeuchi, 1967; Wallis et al., 1986).
S. TYPHIMURIUM
AND MOUSE
PEYER’S
PATCH
M CELLS
WI
Fig. 3. CLSM images of mouse Peyer’s patch FAE after 30-min incubation with S. typhimurium
(a-d) or PBS (e). Adherent Salmonella have been visualized with FlTC-antibodies (a), polymerized actin has been localized with TlUTC-phalloidin (b, c and e) and M cells have been localized with FlTC-UEAI (d). Panels a and b: projected Z series, demonstrating that FlTC-stained bacteria (a) are associated with cells containing prominent actin rings (b, arrows). Salmonella-associated phalloidin-stained actin rings (c) are located in UEAI-stained M cells (d). Tissues incubated with PBS lack actin-stained rings (e). Scale bars = 10 pm.
Recently, we demonstrated that mouse Peyer’s patch M cells may be identified by the selective binding of the lectin UEAl to a-L-fucose residues in the M-cell glycocalyx (Clark et al., 1993). In view of the fact that many bacteria bind to cell surface glycoconjugates (reviewed by Sharon and Lis, 1989), we proposed that M-cellspecific surface glycoconjugates may be responsible for the targeting of microorganisms to M cells. In the present study, we have demonstrated that S. typhimurium preferentially interacts with only a subset of M cells which were not distinguished by their UEAl-binding characteristics. It therefore seems unlikely that expression of a-Lfucose residues is the primary factor determining the pattern of S. typhimurium adherence. However, tie possibility that a-L-fucose residues are involved in S. typhimurium adhesion cannot be excluded, since additional factors such as glycocalyx structure and/or luminal contents such as mucus or food particles might limit bacterial access to the M-cell surface receptors. Variations in the M-cell apical membranes may also affect the distribution of Salmonella adherence. Previous investigations have demonstrated that M cells exhibit a range of ultrastructural characteristics (Bye et al,; 1984) and possibly also a range of
alkaline phosphatase activities (Smith et al., 1988). It may be argued, therefore, that M cells are not J single, homogeneous population but exist as a range of subtypes expressing different morphological and functional characteristics. The interaction of S. typhim~ri~
with mouse Peyer’s patch M cells resulted in dramatic changes in cell surface morphology and actin redistribution. These changes were analogous to those observed in cultured cell lines such as Caco2, Henle-407, HEp-2, MDCK and Swiss 3T3 fibrobtasts i&c&d with this bacterium (Fiiay et d., 1991; Gal&~ et aZ., 1992~1; Ginocchio et al., 1992 ; Francis et al., 1992, 1993 ; Jones et aZ., 1993). Changes in cell surface morphology are also observed in other in vivo models of Salmonella infection (Takeuchi, 1967 ; Kohbata et d, 1986; Wallis et al., 1986). We therefore propose that changes in cell surface morphology and actin redistribution are generalized phenomena which accompany S. typhimurium interaction with epithelial cells both in vitro and in vivo. Based on these criteria, it would appear that cultured cell lines are valid models for the study of S. typbrium interaction with gastrointestinal epithelial cells. Siice studies performed in vitrv hive dem-
550
M.A. CLARK ET AL.
oastrated that membrane ruffle formation and cellular actin redistribution correlate with SalmoneUa invasion (reviewed by Bliska et aL, 1993), we that these phenomena are similarly associatod with Salmonella invasion in vivo. In the present study, membrane ruffles induced by the interaction of S. ~phanurium with mouse Peyer's tmtr,h M cells more closely resembled the membrane ruffles associated with S. typhimurium invasion of cultured cell lines such as Caco-2, HEp-2 and M D C K (e.g. Finlay and Falkow, 1990; Francis et al., 1992) than the microvillus disruption accompanying Salmonella infection of villous entexocytes (I'akeuchi, 1967: Wallis et al., 1986). Since the Salmonella-induced changes were more extensive in cells possessing a less organized brush border and apical cytoskeleton (i.e. M cells and cultured cell lines), it is likely that the state of cell differentiation determines the extent of the morphological changes induced by Salmonella. Studies performed on cultured cell lines have demonstrated that membrane ruffles induced by Salmonella a~ae preferential sites for the interaction of subsequent bacteria (Francis et al., 1992, 1993). This may explain why we frequently observed groups of bacteria binding to single M cells. We occasionally observed bacteria adherent to M cells in the absence of membrane ruffle formarion or cellular actin redistribution. These bacteria may have only recently made contact with the M-cell surface or they may have been of a phenotype which was unable to elicit membrane ruffle formation or actin redistribution. Bacterial growth conditions and growth phase may affect the ability of Sahnonella to invade cultured cells and consequently their ability to induce membrane ruffles, anaerobically grown or logarithmic phase cultures being more invasive than aerobically grown or stationary phase cultures respectively (Ernst et al., 1990; Lee and Falkow, 1990 ; Francis et al., 1992). In the present study, the bacteria were grown in a manner which had previously been found to enrich for the invasive phenotype (Lee and Falkow, 1990; Francis et al., 1992). However, a recent study has demonstrated that the capacity of Salmonella to adhere to and invade cultured cells is related to the multiplicity of infection and is independent of bacterial growth phase (Kusters et al., 1993). Conse-
quently, the importance of the bacterial growth phase in the current in vivo model of infection is unclear. Membrane ruffling (accompanied by actin redistribution) is not limited to Salmonella invasion but is a generalized phenomenon associated with the activation of several different cell surface receptors including the epidermal growth factor (EGF) receptor (e.g. Brunk et al., 1976; Kadowaki et al., 1986). Ruffle formation is accompanied by a temporarily increased pinocytic activity which results in the passive entry of adherent, otherwise non-invasive species (Francis et al., 1993). There is, however, no evidence to suggest that the antigen-sampling capacity of M cells is based solely on this phenomenon. While microorganisms other than Salmonella induce changes in M-cell surface morphology including localized loss 6f microvilli and the formation of pseudopod-like cytoplasmic extensions (e.g. Fujimura, 1986; Owen et al., 1986), the formation of distinct membrane ruffles is apparently unique to Salmonella infection. The mechanisms by which Salmonella induce membrane ruffle formation in cultured cell lines are unclear (reviewed by Bliska et al., 1993). It wan
plol,~J~r.,-u U l a t UlC I : ~ U r
lC~ptol
lll~l.y p l a y
a
role in S. typhimurium invasion of cultured cells and in the induction of Salmonella-associated membrane ruffles (Galdn et al., 1992b ; Pace et al., 1993). However, the EGF receptor is not essential for S. typhimurium invasion or the induction of Salmonella-associated membrane ruffling acti, ity (Francis et al., 1993; Jones et al., 1993). In conclusion, we have demonstrated that, when incubated in mouse Peyer's patch gut loops, S. typhimurium strain SL1344 preferentially adheres to and invades a subset of M cells. Assuming these findings represent the situation in the intact host, M cells represent a major route by which S. typhimurium penetrates the gut. Previous studies have demonstrated that attenuated Salmonella c~.,Tying heterologous antigens are potentially very important as oral vaccine delivery systems (reviewed by Dougan, 1994). The present study emphasizes the suitability of Salmonella as agents for oral vaccine delivery since, by preferentially interacting with M cells, they
S. T Y P H I M U R I U M A N D M O U S E P E Y E R ' S PATCH M CELLS are tm'geted to sites where cells o f the i m m u n e system are concentrated.
551
M o t s - c l d s : lntestin, Salmonella, Adherence, Invasion, Cellule M, Epith61ium folliculaire, Plaque de Peyer, S a l m o n e l l a t y p h i m u r i u m : Ulex europaeus 1, Plissements membranaires, At:dneo Souris, Vaccination per os.
Acknowledgements We are grateful to Dr. T.A. Booth, Biomedical Electron Microscopy Unit, University of Newcastle upon Tyne, for assistance with the scanning electron microscopy. This work was supported under the LINK programme in Selective Drug Delivery and Targeting. funded by SERC/MRC/DTI and industry (SERC grant GR/F 09747). MAJ was also supported by a Wellcome Postdoctoral Research Fellowship (039684/Z/93/Z). Additional support was provided by the Royal Society for equipment.
Interaction pr~f4rentieHe de Salmonella typhimurium avec les cellules M des plaques de Peyer chez la souris Nous avons utilis6 un mod~le murin de plaques de Peyer pour 6tudier le r61e des cellules 6pith6!iales M dans la pathogen~se de l'infection due ~ Salmonella typhimurium. Ces cellules sp6cialis6es dans la collecte antig6nique sent localis6es dans l'6pith61ium folliculaire (FAE) de l'intestin grSle et .21__ uu colon. Nous avons d6montr6 que S. typhimurium adhere plus fr6quemment au FAE des plaques de Peyer q u ' a u x ent6rocytes villeux et que, dans le FAE, cette bact6de interagit pr6f6rentiellement avec les cellules M. Une analyse quantitative microscopique, utilisant la lectine UEA1 (Ulex europeus) pour identifier les cellules M, a r6v616 que le hombre des bact6ries fix6es sur les cellules M est 34 fois sup6rieur au hombre not6 sur les ent6rocytes. Une incubation de 30 wan pe..rmet de noter que des cellules M sent nettement envahies par Salmonella. C ' e s t pourquoi nous pensons que les cellules M repr6sentent une voie pr6f6rentielle d'invasion par .5". typhimurium ~ travers la harfi~re 6pith61iale intestihale. L'adh6rence bact6fienne aux cellules M n'est pas homog~ne, ce qui sugg~re l'existence de soustypes de cellules M. L'interaction de S. typhimurium a v e c les c e l l u l e s M des p l a q u e s de P e y e r s'accompagne d'ondulations membranaires et d'une redistribution d'actine polym6ris~e similaire h c e qui s ' o b s e r v e darts les cultures de lign6es cellulaires infect6es par cette bact6de. Nos observations font ressortir l'opportunit~ de la vaccination par voie orale puisque ces bact6ries, intetagissant pr6f6rentiellement avec les cellules M, peuvent ~tre dirig6es sur les sites o f / l e s cellules du syst~me immunitaire sent particuli~rement concentr~es.
References Amerongen, H.M. & Langermann, S. (1992), Antigen transport by intestinal M cells. Curt. Opin. Gastroenterology, 8, 983-987. Bliska, J.B., Gahin, J.E. & Falkow, S. (1993), Signal transduction in the mammalian cell during bacterial attachment and entry. Cell, 73, 903-920. Brunk, U., Schellens, J. & Westermark, B. (1976), Influence of epidermal growth factor (EGF) on ruffling activity, pinocytosis and proliferation of cultivated human glia cells. Exp. Cell Res., 103, 295-302. Bye, W.A., Allan, C.H. & Trier, J.S. (1984), Strucam~, distribution, and origin of M cells in Peyer's patches of mouse ileum. Gastroenterology, 86, 789-801. Carter, P.B. & Collins, F.M. (1974), The route of enteric infection in normal mice. J. Exp. Med., 139, 11891203. Clark, M.A., Jepson, M.A., Simmons, N.L., Booth, T.A. & Hirst, B.H. (1993), Differential expression of lectin binding sites defines mouse intestinal M cells. J. H/stochem. Cytochem., 41, 1679-1687. Dougan, G. (1994), The molecular basis for the virulence of bacterial pathogens: implications for oral vaccine development_ Mirrnhlnl, !6~_~2lS-9~4_ Ernst, R.K., Dombroski, D.M. & Merrick, J.M. (1990), Anaerobiosis, type I f'unbriae, and growth_ phase are factors that affect invasion of HEp-2 cells by Sa/monella typhimurium. Infect, lmmun., 58, 2014-2016. Finlay, B.B. & Falkow, S. (1990), Sa/monel/a interactions with polarised human intestinal CaCo-2 epithelial cells. J. Infect. Dis., I62, 1096-1106. Finlay, B.B., Ruschkowski, S. & Dedhar, S. (1991), Cytoskeletal rearrangements accompanying Sa/monella entry into epithelial cells. J. Cell Sci., 99, 283296. Francis, C.L., Ryan, T.A., Jones, B.D., Smith, SJ. & Falkow, S. (1993), Ruffles induced by Salmonella and other stimuli direct macropinocytosis of bacteria. Natu,.'e (Lend.), 364, 639-642. Francis, C.L., Starnbach, M.N. & Falkow, S. (1992), Morphological and cytoskeletal changes in epithelial cells occur immediately upon interaction with SoJmonella typhimurium grown under low-oxygen conditions. Molecular MicrobioL, 6, 3077-3087. Fujimura, Y. (1986), Functional morphology of microfold cells (M cells) in Peyer's patches. Phagocytosis and transport of BCG by M cells into rabbit Peyer's patches. Gastroenterologia Japonica, 21,325-335. Gahin, J.E., Ginocchio, C. & Costeas, P. (1992a), Molecular and functional characterisation of the Salmonella invasion gene InvA: homology of lnvA to members of a new protein family. J. Bacteriol., 174, 43384349.
552
M.A. C L A R K E T AL.
G ~ , . J.E., Pace, J. & Hayman, MJ. (1992b), Involvemeat of the ep~.mml growth factor receptor in the invasion of cultured mammalian cells by Salmonella O~mUmur/um.Nature (Lond.), 357, 588-589. Gim3cchio, C, Pace, J. & Gal,'m, J.E. (1992), Identification and molecular chatactedsafion of a Sa/monella typhinmr/um gene involved in triggering the internalisalion of salmonellae into cultured epithelial cells. Proc. Natl. AcaR Sc£ USA, 89, 5976-5980. Jepson, M.A., Clark, M.A., Sitnmons, N.L. & Hirst, B.H. (1993), Actin accumulation at sites of attachment of indigenous apathogenic segmented filamentous bacLeria to mouse ileal epithelial cells. Infect. immun., 61.4001-4004. Jones, B.D., Paterson, H.F., Hall, A. & Falkow, S. (1993), Sa/mone//a typh/mm~'um induces membrane ruffling by a growth factor-n~tor-independent mechanism. Proc. Nat/. Acad. Sc/. USA, 90, 10390-10394. Kadowaki, T., Koyasu, S., Nishida, E., Sakai, H., Takaku, F., Yahara, I. & Kasuga, M. (1986), Insulin-like growth factors, insulin, and epidermal growth factor cause rapid cytoskeletal reorganisation in KB cells. J. Biol. Chef&, 261, 16141-16147. K~l~ta, S., Yokoyama, H. & Yabuuchi, E. (1986), Cytopathogenic effect of Salmonella typhi GIFU 10007 on M cells of murine ileal Peyer's patches in ligated ileal loops : an ultrastructural study. Microbiol. /mmmlol., 30, 1225-1237. Kustcts, J.G., Muldets-lCremets, G.A.W.M., Van Doomik, C.E.M. & Van der Zeijsk B.A.M. (1993), Effects of multiplicity of infection, bacterial protein synthesis, and growth phase on adhesion to and invasion of human cell lines by Salmonella typhimurium` Infect. Immun., 61, 5013-5020.
Lee, C.A. & Falkow, S. (1990), The ability of Salmonella to enter mammafian cells is affected by bacterial growth state. Proc. Natl. Acad. Sci. USA, 87, 43044308. Neutra, M.R. & Kraehenbuh!, J.P. (1992), Transepithelial transport and mucosal defense. - - I. The role of M cells. Trends Cell Biol., 2, 134-138. Owen, R.L., Pierce, N.F., Apple, R.T. & Cray, W.C. (1986), M cell transport of Vibrio cholerae from the intestinal lumen into the payer's patches: a mechanism for antigen sampling and for microbial transepithelial migration. J. Infect. Dis., 153, 11081118. Pace, J., Hayman, M.J. & Gal~ln, J.E. (1993), Signal transduction and invasion of epithelial cells by S. typhimurium. Cell, 72, 505-514. Sharon, N. & l_is, H. (1989), Lectins as cell recognition molecules. Science, 246, 227-234. Smith, M.W., James, P.S., Tivey, D.R. & Brown, D. (1988), Automated histochemical analysis of cell populations in the intact follicle-associated epithelium of the mouse Peyer's patch. Histochem` J., 20, 443-448. Takeuchi, A. (1967), Electron microscope studies of experimental Salmonella infection. - - I. Penetration into the intestinal epithelium by Salmonella typhimurium` Am. J. Pathol., 50, 109-136. Trier, J.S. (1991), Structure and function of intestinal M cells. Gastroenterol. Clin. N. Amer., 20, 531-547. Wallis, T.S., Starkey, W.G., Stephen, J., Haddon, S.J., Osborne, M.P. & Candy, D.C.A. (1986), The nature and role of mucosal damage in relation to Salmonella typhimurium-induced fluid secretion in the rabbit ileum. Y. Med. Microbiol., 22, 39-49.