Full text of DOI:10.1007/PL00000693

CMLS, Cell. Mol. Life Sci. 57 (2000) 323–332
1420-682X/00/020323-10 $ 1.50+0.20/0
© Birkha¨user Verlag, Basel, 2000
Review
The role of M cells in mucosal immunity
L. J. Hathaway and J.-P. Kraehenbuhl*
Swiss Institute for Experimental Cancer Research and Institute of Biochemistry, University of Lausanne,
Chemin des Boveresses 155, CH-1066 Epalinges (Switzerland), Fax +41 21 652 69 33,
e-mail: jean-pierre.kraehenbuhl@isrec.unil.ch
Received 29 September 1999; received after revision 4 November 1999; accepted 5 November 1999
Abstract. Mucosa-associated lymphoid tissue in the res- Binding of particles to the apical M cell membrane
piratory and digestive tracts are covered by a special- may be nonspecific or due to specific interaction be-
ized epithelium, the follicle-associated epithelium, which tween molecules such as integrins and lectins. Exploit-
includes M cells, which are specialized for the uptake ing the specific binding to M cells is an aim for
and transcytosis of macromolecules and microorgan- mucosal vaccination, for example to increase the effi-
isms. Following transcytosis, antigens are released to ciency of uptake of an oral vaccine by its conjugation
cells of the immune system in lymphoid aggregates to an M-cell-specific molecule. Alternatively, an M-
beneath the epithelium where antigen processing and cell-specific live vector, such as attenuated Salmonella
presentation and stimulation of specific B and T bacteria, may be used to deliver epitopes of other or-
lymphocytes are achieved. Circulation of the lymphoid ganisms. Mucosal vaccination efficiency may also be
cells enables their homing to their original, and other, enhanced by a temporary increase in the number of M
mucosal sites where they exert the effector function. cells.
Such a response may be dominated by secretory im- Therefore, investigation of the properties and ontogeny
munoglobulin A release and may include cytotoxic T of M cells must be pursued to allow the development of
lymphocyte action.
better mucosal vaccines for the future.
Key words. M cell; mucosa; immunity; vaccination; ontogeny.
moist nature, mucosal surfaces are more susceptible
Introduction
than skin to invasion by pathogens and so have a
dedicated branch of the immune system for their protec-
tion. The mucosal epithelium of the oral cavity, phar-
ynx, oesophagus, urethra and vagina are composed of a
stratified squamous or stratified columnar epithelium,
and antigen sampling at these sites is believed to be
performed by migrating dendritic cells. However, the
vast majority of mucosal epithelia are simple, consisting
of a single cell layer. In particular, the gut and respira-
tory tracts have simple mucosal epithelia within which
are some cells specialized for antigen uptake. In the gut
these specialized cells are above aggregations of
The human body is separated from the external envi-
ronment by its covering of skin which protects from
invasion by pathogens and allows the maintenance of
internal homeostatic conditions. However, exchange of
materials with the environment is crucial for mainte-
nance of this homeostasis, for example respiratory gases
in the lungs and nutrients in the gut. The areas where
transfer occurs, the mucosae, are necessarily thin and
tend to have an epithelial membrane that produces
mucus at its free border. By virtue of their thin and
* Corresponding author.

324
L. J. Hathaway and J.-P. Kraehenbuhl
The role of M cells in mucosal immunity
lymphoid tissue known as Peyer’s patches and are carried to the draining lymph node. Eventually, the
lymph drains into the blood, and the cells home to the
mucosal site from which they originated and to others.
The homing of the cells to different mucosal sites gives
rise to the common mucosal immune system (CMIS)
[5]. The homing occurs by the specific binding of
molecules on the surface of the lymphocytes to
molecules on the cells of high endothelial venules
(HEVs) and small flat venules in mucosal tissue. Human
and mouse HEV cells in the gut express mucosal ad-
dressin cell adhesion molecule-1 (MAdCAM-1), which
binds the integrin h4i7 on the lymphocytes [6]. There
appears to be some compartmentalization within the
mucosal immune system, which may be due to different
homing receptors and their ligands at different sites as
known as M cells due to the microfolds on their apical
surface. They are of great interest for the design of
future vaccines which may be targeted to the M cells for
enhanced efficiency of uptake. It is these M cells, and
their role in mucosal immunity, which are the subject of
this review.
Innate and adaptive immunity at mucosal surfaces
Mucosal surfaces are protected by both innate and
adaptive immune systems. Innate mechanisms include
the trapping of pathogens in mucus, which is enhanced
by the binding of mucin to lectins on bacteria. Move-
ment of the mucus by beating cilia in the lungs and the
peristaltic movement of the gut helps to prevent attach-
ment of pathogens to the mucosal surfaces. Other in-
nate mechanisms to protect mucosal surfaces include
the destruction of pathogens by the low pH in the
stomach and by digestive enzymes in the gastrointesti-
nal tract, and competition for nutrients from commen-
sal organisms. Other mucosal secretions, such as tears,
damage bacterial cell walls due to the action of the
enzyme lysozyme which they contain. Epithelial cells
are also considered to play a role in innate immunity
since they can be induced to produce proinflammatory
cytokines such as interleukin (IL)-8 and GROalpha in
response to stimuli such as invasion by pathogens [1–3].
Such an inflammatory response may lead to the elimi-
nation of the pathogen and therefore constitute a first
line of defence.
Protection at mucosal sites due to adaptive immune
responses is achieved to a large extent by immunoglob-
ulin A (IgA), the predominant antibody at these loca-
tions and by far the most abundant antibody in the
body [4]. Since antibody responses to mucosally admin-
istered antigens can be detected at mucosal sites and
systemically it appears that the mucosal immune system
is separate from, but linked to, the systemic immune
system. Systemic immunization is generally not effective
in inducing immune responses at mucosal sites, but
immunization via mucosal routes can induce systemic
and mucosal immunity.
discussed in
a
recent review [7]. The effector
lymphocytes such as T helper (Th) cells, cytotoxic T
lymphocytes (CTLs) and IgA-producing B cells there-
fore find their way to mucosal effector tissue in the
gastrointestinal, respiratory and genital tracts and secre-
tory glands. Lymphoid tissue at the effector sites tends
to be more diffuse than at the inductive sites.
The majority of IgA released from plasma cells at
mucosal effector sites is in the form of a dimer. The two
IgA molecules are linked at the constant regions of their
heavy chains by a peptide known as the J chain, which
is also produced by the plasma cell. Mucosal epithelial
cells have polymeric immunoglobulin receptors (pIgR)
on their basolateral membranes [8]. The dimeric IgA
binds to the pIgR, triggering its internalization and
transport through the epithelial cell. The dimeric IgA is
expressed on the luminal side of the epithelial cell where
the pIgR is cleaved to release the antibody. Part of the
pIgR remains attached to the antibody and is known as
the secretory component; it may protect the antibody
from enzymatic cleavage.
Secretory IgA may protect mucosae by several different
mechanisms, including immune exclusion where IgA
binds pathogens, preventing their attachment to epithe-
lial cells [8]. IgA may also have a role in intracellular
neutralization of virus in epithelial cells [9], and the
binding of IgA to pathogens in the lamina propria may
cause them to be exported out into the lumen as im-
mune complexes [10]. Phagocytosis may be promoted as
IgA binds to Fch receptors on various cells, and this
antibody may also enhance the sticking of some bacte-
ria to mucus, neutralize toxins and interfere with
growth factors required by pathogens. Of the two
known subclasses of IgA discovered in humans, IgA1
predominates in the serum and upper aerodigestive
tract, whereas IgA2 predominates in the normal large
bowel mucosa [11].
The mucosal immune system may be divided into induc-
tive sites, where antigen is taken up to initiate an
immune response, and effector sites, where the immune
response is expressed. M cells, which are found in the
epithelium overlying inductive sites, take up antigen
and pass it to the cells of the immune system beneath.
Following uptake by antigen-presenting cells (APCs),
processing occurs and then presentation to T cells,
which in turn stimulate the development of IgA-com-
Cell-mediated cytotoxicity, antibody-dependent cyto-
mitted antigen-specific B cells. The activated antigen- toxicity involving IgA and natural killer cell activity as
specific B and T cells leave the inductive site and are well as functional CTLs have all been found in mucosa-

CMLS, Cell. Mol. Life Sci. Vol. 57, 2000
Review Article
325
associated tissue and may therefore have roles in mu- degree of compartmentalization leads, for example, to
cosa effector sites [12, 13]. the majority of lymphocytes stimulated in GALT to
Once an antigen has been taken up across a mucosal migrate back to gut lamina propria [11]. GALT in-
surface, rather than inducing a protective immune re- cludes PPs, the appendix and solitary lymphoid nodules
sponse, it may induce ‘oral tolerance’. This is where [12]. PPs, first described by J. C. Peyer in 1677 [20], are
mucosal administration of an antigen induces periph- lymphoid follicles in the wall of the small intestine
eral tolerance to subsequent systemic exposure to the which project into the lumen, forming dome shapes.
same antigen. Without this the body would mount an The follicles contain B and T lymphocytes as well as the
immune response to every foreign antigen taken up specialized APCs, dendritic cells and macrophages. The
across the mucosa, of which there will be many, partic- epithelium overlying the PP dome is called the follicle-
ularly in the gut. M cells are specialized for the uptake associated epithelium (FAE) and contains the M cells
of particulate antigens which they then pass to APCs at specialized for the uptake of antigen. M cells have also
the mucosal inductive sites; but soluble antigen may been observed in the epithelium of other MALTs with a
enter the intestinal enterocytes and not reach an induc- morphology similar to that of GALT M cells, for exam-
tive site. Thus, the result of uptake of a soluble antigen ple in the palatine tonsils of rabbits [21].
may be tolerance, whereas a particulate antigen may
induce an immune response [14].
Characteristics of M cells
The morphology of human M cells was first described
Inductive sites of the mucosal immune system
by Owen and Jones [22], who found that rather than
At areas where antigen is taken up by mucosal tissue having the closely-packed microvilli characteristic of
are inductive sites specialized for the first steps of an small intestinal epithelial cells, some cells in the epithe-
immune response. These consist of organized aggregates lium over PPs have microfolds. The M cells also have
of lymphoid cells known as organized mucosa-associ- prominent Golgi complexes and many mitochondria
ated lymphoid tissue or O-MALT. Mucosae of different surrounded by rough endoplasmic reticulum. They are
regions of the body have their own version of MALT, linked to adjacent cells via desmosomes and interdigita-
such as gut-associated lymphoid tissue (GALT), tions. The lack of a brush border of microvilli in these
bronchial-associated lymphoid tissue (BALT) and cells correlates with the redistribution of villin from the
nasal-associated lymphoid tissue (NALT) (table 1). The apical membrane to the cytoplasm and reflects the reor-
various MALTs have a similar, but not identical, cellu- ganization of the F-actin network [23]. In addition,
lar composition: NALT has a greater proportion of T changes in cytoskeleton in these M cells compared with
cells than Peyer’s patches (PPs), and of these a greater enterocytes were shown by the fact that pig M cells, but
percentage are CD4+ as compared with CD8+ T not enterocytes, were labelled intensely with anti-cytok-
cells. NALT T-cell populations resemble those of the eratin 18 antibody, but cytokeratin 19 was more abun-
spleen rather than the PP, but NALT and PP, contain dant in enterocytes than M cells [24]. M cells are
similar numbers of mature IgM+ IgA− B cells [15]. A surrounded by microvilli-bearing cells and contain vesi-
Table 1. A comparison of GALT, NALT and BALT (in mice unless otherwise stated).
Mucosa-associated lymphoid tissue
Location
Cellular composition
Presence of M cell in epithelium
yes
Peyer’s patches
(GALT) [15]
Intestine wall
28% T cells
CD4:CD8 ratio
3.7:1
70% B cells
2% macrophages
40% T cells
CD4:CD8 ratio
10:1
55% B cells
4.5%
macrophages
T cells
CD4\CD8
B cells
NALT [15]
BALT [17]
Entrance of
nasopharyngeal duct
yes in rats [16]
At junctions between
bronchi and bronchioli
or between the bronchus
and an artery
controversial
in rabbits [18, 19]
macrophages

326
L. J. Hathaway and J.-P. Kraehenbuhl
The role of M cells in mucosal immunity
Table 2. Characteristics of M cells from various species and tosis of macromolecules [27]. A further difference be-
locations.
tween M cells and absorptive enterocytes is the lectins
which they can bind to their apical surfaces; for exam-
ple the h-L-fucose-specific lectin Ulex europaeus agglu-
M cell characteristics
Species/location
general [22]
general [22]
pig [24]
Lack of brush border of microvilli,
presence of microfolds
Large invagination of basolateral
membrane containing lymphocytes
Increase in cytokeratin 18, reduction in
cytokeratin 19 compared with
enterocytes
Apical expression of digestive enzymes
(e.g. alkaline phosphatase and sucrase
isomaltase) reduced
Lack of expression of polymeric
immunoglobulin receptor (pIgR)
Lower membrane potential than
enterocytes
tinin 1 (UEA-1) binds specifically to mouse PP M cells
[28]. However, the identification of M cells by their
lectin-binding pattern is complicated by the fact that
within an individual, the lectins bound by M cells vary
along the length of the intestine. Clark et al. [29] found
a difference in surface glycoconjugate expression be-
tween caecal and PP M cells in the mouse with staining
by UEA-1 of PP M cells but not caecal patch FAE.
Ueki et al. [30] found differences between M cells from
PPs of the small intestine and colon in humans: small
intestinal M cells were human leukocyte antigen-D re-
lated (HLA-DR)-positive, ICAM-1-negative, whereas
colonic M cells were HLA-DR-negative and ICAM-1-
positive. ICAM-1 was found predominantly at the api-
cal membrane but was not found on adjacent cells of
the FAE. Also, there is variation between species; for
example caecal rabbit M cells differ from those of other
species in their lectin binding [31], and mouse M cells
are exclusively labelled by fucose-specific lectins. In
humans the M-cell glycosylation patterns are distinct
from other species and preferentially display the sialyl
Lewis A antigen [32]. However, this carbohydrate epi-
tope is not a unique marker for M cells, because it is
also expressed on some enterocytes of the FAE and in
goblet-cell mucins. In addition, FAE enterocytes and
goblet cells have lectin-binding characteristics which
differ from those of the villus enterocytes and goblet
cells, particularly in humans [33]. As well as M cells in
the gut, lectin-binding properties of M cells in rabbit
tonsil crypts have been described [34]. M cells selectively
bound UEA-1, Lotus tetragonolobus (LTA), Helix po-
matia (HPA) and Vicia 6illosa (VVA) lectins, whereas
the other epithelial cells bound Ricinus communis
(RCA-1) and Arachis hypogaea (PNA). Both gut and
tonsillar M cells in the rabbit may also be identified by
antibody to vimentin [35], but vimentin does not appear
to be present in mouse M cells [23]. Some characteristics
of M cells are summarized in table 2.
general [25]
human [26]
mouse [27]
mouse [28]
Binding to Ulex europaeus agglutinin 1
(UEA-1) lectin
PP but not caecal
FAE [29]
small intestine [30]
colon [30]
HLA-DR+ ICAM-1−
HLA-DR− ICAM-1+
Expression of sialyl Lewis A antigen (also human [32]
on some FAE enterocytes and in goblet
cell mucins)
Binding to UEA-1, Lotus tetragonolobus
(LTA), Helix pomatia (HPA), Vicia
6illosa (VVA)
rabbit tonsil [34]
Vimentin
rabbit gut and
tonsils [35]
not in mouse [23]
cles within their cytoplasm. M cells have a large invagi-
nation of the basal membrane forming a pocket in
which lymphoid cells are found. A narrow layer (0.3 mm
thick) of the M cell cytoplasm is therefore all that
separates the lymphoid cells from the lumen, whilst the
integrity of the epithelium is preserved by the tight
junctions between M cells and adjacent cells [22].
Mouse M cells have a shape comparable to that of
human M cells and contain fewer lysosomes and have
fewer and shorter microvilli than adjacent absorptive
cells. M cells outside the gut, such as in rabbit palatine
tonsil epithelium, also have this characteristic shape
with an invagination of the basolateral membrane form-
ing a pocket containing lymphocytes. Gebert [21] found
that the region of the basolateral membrane forming
the pocket had a different lectin-binding pattern than
the rest of the basolateral membrane, suggesting that
the compostion of the membrane is not uniform. Apical
alkaline phosphatase is less abundant, but esterase is
more abundant in M cells than that of adjacent absorp-
tive cells [25]. In contrast to villus enterocytes, human
FAE does not express the pIgR and so does not appear
to transport secretory IgA to the gut lumen [26]. In
addition, M cells have a lower membrane potential than
This lectin-binding ability of M cells may reflect the
situation in vivo as lectin-bearing microorganisms such
as Escherichia coli RDEC-1 [36], Salmonella ty-
phimurium [37] and Yersinia enterocolitica [38] bind spe-
cifically to M cells. Yersinia pseudotuberculosis may
target M cells by the binding of its invasin to i1-inte-
grin, which can be expressed on the apical membrane of
M cells, but which on other epithelial cells is restricted
to the basolateral membrane [39]. Lectins bound to M
cells may be rapidly endocytosed and transcytosed [28],
which demonstrates a mechanism by which pathogenic
microorganisms may gain entry across the epithelium
differentiated enterocytes, which may aid their endocy- by binding to M cells via lectins.

CMLS, Cell. Mol. Life Sci. Vol. 57, 2000
Review Article
327
M-cell-dependent immune response induction
ligand on the vibrio surface by such treatments or due
to the reduction of their motility. The ability of M cells
to take up particles may, in part, be determined by their
lack of a terminal web of microfilaments as well as their
lack of an organized brush border of microvilli and
glycocalyx. In addition, Owen et al. found that there
was a paucity of mucus over M cells relative to that
over absorptive enterocytes, thus increasing the accessi-
bility of M cells to lumen antigens.
Antigens tend to be transcytosed through M cells and
released into the pocket beneath intact [46], and so M
cells are not thought to have a major role in the
processing and presentation of antigen. However, acidic
lysosomal-like vesicles in the cytoplasm and major his-
tocompatibility complex (MHC) class II expression on
the basolateral membrane was found in a study of rat
intestinal M cells which would be consistent with a role
in antigen processing and presentation [47]. The aspar-
tic proteinase cathepsin E has been localized to M cells
of humans and rats, which may enable M cells to
contribute to the processing of macromolecules and
microorganisms being transported by them to the
lymphoid cells beneath the epithelium [48].
It has been suggested that M cells may aid the immune
response induction to the antigen they are transporting
by releasing a costimulatory signal for T and B cell
proliferation in GALT. Pappo and Mahlman [49] iso-
lated M cells from rabbit FAE by flow cytometry using
a monoclonal antibody which recognizes an M cell
pocket-domain-restricted surface molecule. The M cells
secreted IL-1, as assessed by the ability of their culture
supernatant to induce proliferation of a T cell clone,
and this secretion was increased when they were stimu-
lated with lipopolysaccharide (LPS). Antibodies against
IL-1 inhibited the T cell proliferation in response to the
M cell supernatant. Therefore, in vivo as M cells trans-
cytose bacteria from the lumen to the pocket, LPS may
stimulate them to release IL-1 aiding proliferation of
the lymphocytes for the successful induction of the first
step of a mucosal immune response. There are more T
cells at the FAE than the villi epithelium, and the T
cells appear to be clustered adjacent to M cells. Since
the ratio of CD4+ to CD8+ T cells is also greater at
the FAE than the villi epithelium, the FAE may be
associated with induction of helper T cells [50]. These T
cells may help both a protective secretory IgA response
and a local CTL response to antigen taken up by M
cells [51].
M cells are specialized for the uptake of antigens, their
transcytosis and release to cells of the immune system in
the lymphoid tissue of the mucosae. Even inert particles
have been shown to be taken up from the intestinal
lumen specifically by M cells. Pappo and Ermak [40]
showed this when they administered fluorescent 600–
750-nm latex microparticles into intestinal loops of rab-
bits and, by microscopy, followed the passage of the
microparticles with time. No particles were found to be
taken up by the villi but, 10 min after administration of
the microparticles, 95% were at the FAE surface. By 90
min post-administration, some microparticles had
moved through the FAE and been released into the
lymphoid area of the dome. By the use of monoclonal
antibodies, the cells which had taken up the latex mi-
croparticles were identified as M cells.
Cationized ferritin (CF) has also been used to investi-
gate uptake and transport by M cells in comparison
with absorptive enterocytes [41]. Bye et al. [42] allowed
60 min of exposure to CF and then assessed the abun-
dance of CF in endocytic vesicles of absorptive entero-
cytes and mature and immature M cells of mice. Only
15% of absorptive enterocytes had endocytic vesicles
containing CF, whereas 80% of mature M cells and 58%
of immature M cells had. M cells also had greater
quantities of CF internalized, and only mature M cells
transported CF across the cells. Immature M cells lack
a pocket and often have more free ribosomes than
mature M cells and an apical surface which resembles
undifferentiated crypt cells.
Since such inert substances as latex microparticles and
cationized ferritin are taken up by M cells, this indicates
that specific receptor binding is not required for uptake.
However, it appears that some bacteria and viruses do
target M cells, presumably to exploit their transcytotic
activity for invasion of the body. For example,
Salmonella typhi bacteria pass through M cells to cause
a systemic disease [37], reoviruses are transported
through M cells to cause pathogenesis in mice [43] and
human immunodeficiency virus-1 (HIV-1) is endocy-
tosed and transcytosed by M cells of mice and rabbits
[44]. Entry of bacteria generally considered as noninva-
sive, such as Vibrio cholerae, via M cells has been
investigated in rabbits. Owen et al. [45] inoculated
vibrios into the intestinal lumen and observed by trans-
mission electron microscopy that they were phagocy-
tosed by M cells into vesicles which were released from
the basolateral membrane to the underlying
lymphocytes and macrophages of the PPs. Uptake of V.
cholerae by epithelial cells other than M cells was not
observed. No uptake of vibrios killed with acid, heat,
Mucosal vaccination
Since the majority of pathogens enter the body by
formalin or ultraviolet (UV) irradiation was observed, crossing mucosal surfaces, there is great interest in
perhaps due to the alteration of an M cell binding developing vaccines which induce protective immune

328
L. J. Hathaway and J.-P. Kraehenbuhl
The role of M cells in mucosal immunity
responses at mucosae as well as systemically. The most than uncoated liposomes. Zhou et al. [59] combined
successful mucosal vaccine used for humans to date is these methods of enhancing a mucosal immune re-
the oral (Sabin) poliomyelitis vaccine, which consists of sponse by immunizing mice rectally with antigen encap-
three live attenuated strains of the poliomyelitis virus sulated in liposomes coated with IgA and administered
[52]. Another mucosal vaccine which has been used to with cholera toxin as an adjuvant. The IgA on the
prevent disease in humans is the Ty21a strain of Sa- surface of the liposomes increased their uptake into PP
monella typhi bacteria, which protected against typhoid mucosa. Antigen may also be protected for mucosal
when administered orally [53]. Salmonella can be used delivery by its encapsulation in poly(DL-lactide-co-gly-
as a vector to carry foreign antigens [54] and binds colide) microparticles; for example Jones et al. [60] used
selectively to M cells [55]. An alternative approach to this polymer to encapsulate DNA for oral admini-
the use of live vectors for mucosal vaccination is the use stration.
of subunit vaccines which avoid the risk of reversion to Secretory IgA was found by Kato [61] in the glycocalyx
virulence of the delivery vector but which tend to be less and on microfolds of M cells rather than adjacent
immunogenic and may therefore need to be adminis- absorptive cells, and when Weltzin et al. [62] injected
tered with an adjuvant. The B subunit of cholera toxin labelled IgA into intestinal loops, it bound to M cell
(CTB) has been proposed as a protein with unique luminal membranes and was transported across the M
mucosal adjuvant properties which could be exploited cells in vesicles. Therefore, attempts have been made to
to enhance the uptake of mucosally administered anti- exploit the selective adherence of secretory IgA to M
gens. Frey et al. [56] investigated the uptake of CTB cells by using it as a vaccine delivery vector for foreign
coupled to fluorescein isothiocyanate (FITC) represent- epitopes. Corthe´sy et al. [63] inserted a Shigella flexneri
ing a soluble antigen, and CTB coupled to gold parti- IpaB invasin epitope into rabbit secretory component
cles, final diameter 1.13 mm, representing bacteria. CTB (SC) and then reassociated the SC to IgA and used the
coupled to FITC and CTB-coated fluorescent micropar- construct to immunize mice orally. Antibodies to IpaB
ticles were administered to ligated loops in rabbits, and invasin were induced in serum and saliva.
the CTB-gold particles were applied to mucosal ex- Uptake of fluorescent polystyrene microspheres by M
plants ex vivo. CTB can bind ganglioside GM1 glyco- cells was increased at least threefold when the micro-
lipid on the apical membrane of all intestinal epithelial spheres were conjugated to an anti-M-cell monoclonal
cells, and the soluble antigen CTB-FITC was found to antibody and administered into intestinal loops in rab-
bind to absorptive enterocytes and M cells. In contrast, bits [64]. In mice, coating of latex microspheres with
the CTB-gold conjugates adhered only to M cells, Ulex europaeus 1 lectin enhanced their binding to M
whereas the CTB-coated microparticles did not adhere cells by 100-fold while not affecting binding to entero-
to any of the epithelial cells. Frey et al. explained this cytes [65]. If such an M cell-specific factor for human M
by the fact that M cells have a relatively thin glycocalyx cells could be coated onto particles, vaccine uptake
which allows soluble antigens and 28.8-nm-diameter efficiency might be greatly increased.
particles to cross. The thicker glycocalyx of absorptive
enterocytes appears to allow soluble antigens to cross
but not particles of 28.8-nm diameter. However, M cells
seemed to have a thick enough layer to prevent particles
Ontogeny of M cells
greater than 1 mm from binding to them. Therefore, In addition to targeting vaccines to M cells, future
attaching CTB to a vaccine would only be useful if the approaches to mucosal vaccination may include induc-
antigen was below a certain size. To protect antigens tion of a temporary increase in the number of M cells in
from digestion by gut enzymes, they may be encapsu- order to increase vaccine uptake. Differentiation into M
lated within liposomes, thus also making them particu- cells is incompletely understood, whereas differentiation
late. Cholera toxin (CT) and its B subunit have been into absorptive enterocytes has been more extensively
coupled to liposomes to combine the advantages of studied [66]. Intestinal epithelial cells originate in the
antigen protection with increased uptake. Harokopakis crypts from undifferentiated stem cells. An individual
et al. [57] have described a method for this coupling crypt situated between a villus and a PP dome may
which maintains the biological activity and immuno- contribute epithelial cells to the villus from the villus
genicity of the CTB which was used to immunize rats side of the crypt and to the FAE of the dome from the
intragastrically. Liposomes have also been coated with dome side of the crypt. As cells migrate from the crypt
antigen-antibody complexes, irrelevant to the antigen, up the villus, they acquire the phenotype of differenti-
within the liposomes to exploit the ability of M cells to ated enterocytes. This differentiation is achieved by the
recognize antigen-antibody complexes. Velez et al. [58] transcription of genes encoding proteins required for
administered such liposomes to mice intrajejunally and the functions of absorptive enterocytes, including diges-
found that they induced greater mucosal IgA responses tive enzymes such as sucrase-isomaltase and trans-

CMLS, Cell. Mol. Life Sci. Vol. 57, 2000
Review Article
329
porters of nutrients. The pathway of differentiation of
M cells is less clear. FAE cells are also derived from the
crypts, but whether the M cells differentiate from ab-
sorptive enterocytes on the dome or are a separate cell
lineage coming directly from the crypt is controversial.
Gebert et al. [67] favoured the latter hypothesis in their
recent paper describing two different types of crypts:
one specialized to produce epithelial cells for domes
which appeared to produce M cells, and ordinary crypts
which did not produce M cells. M cell precursors were
found on the sides of dome-associated crypts by la-
belling lectin, and cells with M cell morphology were
found in strips leading from the crypt onto the dome.
Gebert et al. concluded that M cells represent a separate
cell line deriving directly from undifferentiated crypt
stem cells. However, other studies have indicated that
differentiated absorptive enterocytes can be induced to
Figure 1. Transmission electron microscopy view of M cell in-
duced by coculture of PP-derived lymphocytes with a monolayer
of the Caco-2 epithelial cell in vitro (L, lymphocytes; Fil, filter).
Reprinted with permission from Kerneis S., Bogdanova A., Krae-
henbuhl J.-P. and Pringault E. (1997) Science 277: 949–952. ©
1997 American Association for the Advancement of Science.
convert into
M
cells following contact with
lymphocytes. FAE forms above aggregates of lymphoid
tissue, and lymphocytes, as well as professional APCs,
nestle into the pocket formed in the basolateral mem-
brane of M cells. Evidence for the role of lymphocytes
in M cell formation came from the in vitro model used
by Kerneis et al. [68] in which Caco-2 epithelial cells
were grown on a membrane with micropores until they
formed a tight monolayer. PP-derived lymphocytes
were then applied to the basolateral side of the mono-
layer. The lymphocytes crossed the membrane and be-
came closely associated with the epithelial cells which
acquired an M-cell-like morphology (fig. 1) and the
ability to transcytose fluorescently labelled latex beads.
In addition, when lymphocytes from PPs were injected
into the intestinal wall of mice, M cells appeared in the
epithelium over the lymphoid cells [68] (fig. 2). How-
ever, Gebert et al. [67] found that holes in the dome
basal membrane, corresponding to sites where
lymphocytes had crossed, were evenly distributed over
the dome, in contrast to the striped distribution of M
cells. They took this to indicate that it was not interac-
tion with the lymphocytes which caused formation of
the M cells. Nonetheless, the M cell phenotype does
appear to be extremely plastic since a continuum of
phenotypes between differentiated enterocytes and M
cells has been observed [69], and phenotypically mature
M cells are found predominantly in a ring on the dome
with very few at the dome apex. Since apoptosis of cells
from the dome occurs almost exclusively at its apex,
and in chickens and mice M cells exhibited no signs of
apoptosis, then it appears that in vivo the M cells
(re)gain a differentiated enterocyte phenotype on their
journey to the apex [70, and Sierro et al., unpublished].
Figure 2. Transmission electron microscopy view of M cell in- The necessity of a signal, which may come from
duced by injection of PP lymphocytes into duodenal mucosa of a
lymphocytes, is also suggested by studies with im-
mouse. Reprinted with permission from Kerneis S., Bogdanova
munodeficient mice [71, and Debard et al., unpub-
A., Kraehenbuhl J.-P. and Pringault E. (1997) Science 277: 949–
lished]. In RAG-deficient mice no PPs develop, but in
B- or T-cell-deficient mice a few small PPs with FAE do
952. © 1997 American Association for the Advancement of Sci-
ence.

330
L. J. Hathaway and J.-P. Kraehenbuhl
The role of M cells in mucosal immunity
develop. It is also known that exposure to some bacte-
ria such as Streptococcus pneumoniae R36a [72] and
inflammation [73] can increase the number of M cells,
showing that this phenotype is inducible. Oral infection
of germ-free mice with Salmonella typhimurium aro
A− bacteria increased the number of M cells by two- to
threefold [74]. Therefore, M cells may derive only from
specialized crypts with only certain stem cells able to
give rise to M cell precursors, but this formation of M
cell precursors may require a signal which can be pro-
vided by lymphocytes. Perhaps undifferentiated Caco-2
cells are equivalent to stem cells from dome-associated
8
9
Mestecky J. and McGhee J. (1987) Immunoglobulin A (IgA):
molecular and cellular interactions involved in IgA biosynthe-
sis and immune response. Adv. Immunol. 40: 153–245
Mazanec M. B., Kaetzel C. S., Lamm M. E., Fletcher D. and
Nedrud J. G. (1992) Intracellular neutralization of virus by
immunoglobulin A antibodies. Proc. Natl. Acad. Sci. USA
89: 6901–6905
10 Kaetzel C. S., Robinson J. K. and Lamm M. E. (1994)
Epithelial transcytosis of monomeric IgA and IgG cross-
linked through antigen to polymeric IgA. J. Immunol. 152:
72–76
11 Brandtzaeg P. (1995) Basic mechanisms of mucosal immunity
– a major adaptive defense system. Immunologist 3: 89–96
12 Staats H. F. and McGhee J. R. (1996) Application of basic
principles of mucosal immunity to vaccine development. In:
Mucosal Vaccines, pp. 17–39, Kiyono H., Ogra P. L. and
McGhee J. R. (eds), Academic Press, San Diego
crypts which give rise to
M cells. Without a
lymphocyte-derived signal, the Caco-2 cells differentiate 13 Klavinskis L. S., Bergmeier L. A., Gao L., Mitchell E., Ward
R. G., Layton G. et al. (1996) Mucosal or targeted lymph
into absorptive enterocytes, but when cocultured with
node immunization of macaques with a particulate SIVp27
lymphocytes, they receive the necessary signal to fulfil
protein elicits virus-specific CTL in the genito-rectal mucosa
their destiny as M cells.
and draining lymph nodes. J. Immunol. 157: 2521–2527
14 Czerkinsky C., Sun J.-B. and Holmgren J. (1999) Oral toler-
ance and anti-pathological vaccines. In: Defense of Mucosal
Surfaces: Pathogenesis, Immunity and Vaccines, pp. 79–91,
Kraehenbuhl J.-P. and Neutra M. R. (eds), Springer, Berlin
15 Heritage P. L., Underdown B. J., Arsenault A. L., Snider D.
P. and McDermott M. R. (1997) Comparison of murine
nasal-associated lymphoid tissue and Peyer’s patches. Am. J.
Respir. Crit. Care Med. 156: 1256–1262
The mechanisms by which M cells are formed are
currently being actively investigated. If their differentia-
tion were fully understood, it might be possible in the
future to manipulate the number of M cells to increase
mucosal vaccine efficiency.
16 Spit B. J., Hendriksen E. G., Bruijntjes J. P. and Kuper C. F.
(1989) Nasal lymphoid tissue in the rat. Cell Tissue Res. 255:
193–198
Conclusion
17 Breel M., Van Der Ende M., Sminia T. and Kraal G. (1988)
Subpopulations of lymphoid and non-lymphoid cells in
bronchus-associated lymphoid tissue (BALT) of the mouse.
Immunology 63: 657–662
18 Gebert A. and Hach G. (1992) Vimentin antibodies stain
membranous epithelial cells in the rabbit bronchus-associated
lymphoid tissue (BALT). Histochemistry 98: 271–273
19 Alonso L. M., Cortes A., Barrutia M. G., Roma T., Varas A.
and Zapata A. G. (1997) Bronchial epithelium associated to
lymphoid tissue does not selectively express vimentin. Histol.
Histopathol. 12: 931–935
20 Peyer H. C. (1677) Exercitatio anatomico-medica de glandulis
intestinorum, earumque usu et affectionibus. Schafhausen
Switzerland, Onophrius et Waldkirch
21 Gebert A. (1997) M cells in the rabbit palatine tonsil: the
distribution, spatial arrangement and membrane subdomains
as defined by confocal lectin histochemistry. Anat. Embryol.
195: 353–358
22 Owen R. L. and Jones A. L. (1974) Epithelial cell specializa-
tion within human Peyer’s patches: an ultrastructural study of
intestinal lymphoid follicles. Gastroenterology 66: 189–203
23 Kerneis S., Bogdanova A., Colucci-Guyon E., Kraehenbuhl
J.-P. and Pringault E. (1996) Cytosolic distribution of villin in
M cells from mouse Peyer’s patches correlates with the ab-
sence of a brush border. Gastroenterology 110: 515–521
24 Gebert A., Rothkotter H. J. and Pabst R. (1994) Cytokeratin
18 is an M-cell marker in porcine Peyer’s patches. Cell Tissue
Res. 276: 213–221
25 Owen R. L. and Bhalla D. K. (1983) Cytochemical analysis of
alkaline phosphatase and esterase activities and of lectin-bind-
ing and anionic sites in rat and mouse Peyer’s patch M cells.
Am. J. Anat. 168: 199–212
26 Brandtzaeg P. and Bjerke K. (1990) Immunomorphological
characteristics of human Peyer’s patches. Digestion 46(Suppl):
262–273
27 Cremaschi D., James P. S., Rossetti C. and Smith M. W.
(1990) Chloride conductance and intracellular chloride accu-
mulation in mouse Peyer’s patch enterocytes. J. Physiol. 427:
71–80
As the understanding of M cell features, functions and
induction increases, we move closer to being able to
manipulate the uptake of mucosally administered vac-
cines and drugs by these cells. The identification of an
M-cell-specific marker would aid their study and if
apically expressed may be a target for mucosal vaccines.
1
2
Jung H. C., Eckmann L., Yang S. K., Panja A., Fierer J.,
Morzycka-Wroblewska E. et al. (1995) A distinct array of
proinflammatory cytokines is expressed in human colon ep-
ithelial cells in response to bacterial invasion. J. Clin. Invest.
95: 55–65
Rasmussen S. J., Eckmann L., Quayle A. J., Shen L., Zhang
Y. X., Anderson D. J. et al. (1997) Secretion of proinflamma-
tory cytokines by epithelial cells in response to Chlamydia
infection suggests a central role for epithelial cells in chlamy-
dial pathogenesis. J. Clin. Invest. 99: 77–87
3
4
Laurent F., Eckmann L., Savidge T. C., Morgan G., Theodos
C., Naciri M. et al. (1997) Cryptosporidium par6um infection
of human intestinal epithelial cells induces the polarized secre-
tion of C-X-C chemokines. Infect. Immun. 65: 5067–5073
Conley M. E. and Delacroix D. L. (1987) Intravascular and
mucosal immunoglobulin A: two separate but related systems
of immune defense? Ann. Int. Med. 106: 892–899
Bienenstock J. and Befus A. D. (1980) Mucosal immunology.
Immunology 41: 249–270
Shyjan A. M., Bertagnolli M., Kenney C. J. and Briskin M. J.
(1996) Human mucosal addressin cell adhesion molecule-1
(MAdCAM-1) demonstrates structural and functional similar-
ities to the h4i7-integrin binding domains of murine MAd-
CAM-1, but extreme divergence of mucin-like sequences. J.
Immunol. 156: 2851–2857
5
6
7
Brandtzaeg P., Farstad I. N. and Haraldsen G. (1999) Re-
gional specialization in the mucosal immune system: primed
cells do not always home along the same track. Immunol.
Today 20: 267–277

CMLS, Cell. Mol. Life Sci. Vol. 57, 2000
Review Article
331
28 Clark M. A., Jepson M. A., Simmons N. L. and Hirst B. H. 48 Finzi G., Cornaggia M., Capella C., Fiocca R., Bosi F.,
(1995) Selective binding and transcytosis of Ulex europaeus 1
lectin by mouse Peyer’s patch M-cells in vivo. Cell Tissue Res.
282: 455–461
Solcia E. et al. (1993) Cathepsin E in follicle associated
epithelium of intestine and tonsils: localization to M cells and
possible role in antigen processing. Histochemistry 99: 201–
211
29 Clark M. A., Jepson M. A., Simmons N. L. and Hirst B. H.
(1994) Differential surface characteristics of M cells from
mouse intestinal Peyer’s and caecal patches. Histochem. J. 26:
271–280
49 Pappo J. and Mahlman R. T. (1993) Follicle epithelial M cells
are a source of interleukin-1 in Peyer’s patches. Immunology
78: 505–507
30 Ueki T., Mizuno M., Uesu T., Kiso T. and Tsuji T. (1995) 50 Bjerke K., Brandtzaeg P. and Fausa O. (1988) T cell distribu-
Expression of ICAM-1 on M cells covering isolated lymphoid
follicles of the human colon. Acta Med. Okayama 49: 145–
151
tion is different in follicle-associated epithelium of human
Peyer’s patches and villous epithelium. Clin. Exp. Immunol.
74: 270–275
31 Gebert A. and Hach G. (1993) Differential binding of lectins 51 London S. D., Rubin D. H. and Cebra J. J. (1987) Gut
to M cells and enterocytes in the rabbit cecum. Gastroenterol-
ogy 105: 1350–1361
mucosal immunization with reovirus serotype 1/L stimulates
virus-specific cytotoxic T cell precursors as well as IgA mem-
ory cells in Peyer’s patches. J. Exp. Med. 165: 830–847
52 Sabin A. B. (1984) Strategies for elimination of poliomyelitis
in different parts of the world with use of oral poliovirus
vaccine. Rev. Infect. Dis. 6(Suppl.): S391–396
53 Cryz S. J. Jr, Vanprapar N., Thisyakorn U., Olanratmanee
T., Losonsky G., Levine M. M. et al. (1993) Safety and
immunogenicity of Salmonella typhi Ty21a vaccine in young
Thai children. Infect. Immun. 61: 1149–1151
54 Newton S. M., Kotb M., Poirier T. P., Stocker B. A. and
Beachey E. H. (1991) Expression and immunogenicity of a
streptococcal M protein epitope inserted in Salmonella flag-
ellin. Infect. Immun. 59: 2158–2165
55 Kohbata S., Yokoyama H. and Yabuuchi E. (1986) Cyto-
pathogenic effect of Salmonella typhi GIFU 10007 on M cells
of murine ileal Peyer’s patches in ligated ileal loops: an
ultrastructural study. Microbiol. Immunol. 30: 1225–1237
56 Frey A., Giannasca K. T., Weltzin R., Giannasca P. J.,
Reggio H., Lencer W. I. et al. (1996) Role of the glycocalyx
in regulating access of microparticles to apical plasma mem-
branes of intestinal epithelial cells: implications for microbial
attachment and oral vaccine targeting. J. Exp. Med. 184:
1045–1059
32 Giannasca P. J., Giannasca K. T., Leichtner A. M. and
Neutra M. R. (1999) Human intestinal M cells display the
sialyl Lewis A antigen. Infect. Immun. 67: 946–953
33 Sharma R., van Damme E. J., Peumans W. J., Sarsfield P.
and Schumacher U. (1996) Lectin binding reveals divergent
carbohydrate expression in human and mouse Peyer’s
patches. Histochem. Cell Biol. 105: 459–465
34 Gebert A. (1996) M-cells in the rabbit tonsil exhibit distinctive
glycoconjugates in their apical membranes. J. Histochem.
Cytochem. 44: 1033–1042
35 Gebert A., Willfuhr B. and Pabst R. (1995) The rabbit M-cell
marker vimentin is present in epithelial cells of the tonsil
crypt. Acta Otolaryngol. 115: 697–700
36 Inman L. R. and Cantey J. R. (1983) Specific adherence of
Escherichia coli (strain RDEC-1) to membranous (M) cells of
the Peyer’s patch in Escherichia coli diarrhea in the rabbit. J.
Clin. Invest. 71: 1–8
37 Clark M. A., Jepson M. A., Simmons N. L. and Hirst B. H.
(1994) Preferential interaction of Salmonella typhimurium with
mouse Peyer’s patch M cells. Res. Microbiol. 145: 543–552
38 Autenrieth I. B. and Firsching R. (1996) Penetration of M
cells and destruction of Peyer’s patches by Yersinia enterocol-
itica: an ultrastructural and histological study. J. Med. Micro-
biol. 44: 285–294
57 Harokopakis E., Childers N. K., Michalek S. M., Zhang S. S.
and Tomasi M. (1995) Conjugation of cholera toxin or its B
subunit to liposomes for targeted delivery of antigens. J.
Immunol. Methods 185: 31–42
39 Clark M. A., Hirst B. H. and Jepson M. A. (1998) M-cell
surface beta1 integrin expression and invasin-mediated target-
ing of Yersinia pseudotuberculosis to mouse Peyer’s patch M 58 Velez C. N., Tonkonogy S. L., Lichtman S. N. and Cho M.
cells. Infect. Immun. 66: 1237–1243
J. (1997) Sensitized liposomes as an antigen delivery system
for the stimulation of mucosal immunity. J. Drug Target. 5:
15–24
40 Pappo J. and Ermak T. H. (1989) Uptake and translocation
of fluorescent latex particles by rabbit Peyer’s patch follicle
epithelium: a quantitative model for M cell uptake. Clin. Exp.
Immunol. 76: 144–148
59 Zhou F., Kraehenbuhl J.-P. and Neutra M. R. (1995) Mu-
cosal IgA response to rectally administered antigen formu-
lated in IgA-coated liposomes. Vaccine 13: 637–644
60 Jones D. H., Clegg J. C. and Farrar G. H. (1998) Oral
delivery of micro-encapsulated DNA vaccines. Dev. Biol.
Stand. 92: 149–155
41 Neutra M. R., Phillips T. L., Mayer E. L. and Fishkind D. J.
(1987) Transport of membrane-bound macromolecules by M
cells in follicle-associated epithelium of rabbit Peyer’s patch.
Cell Tissue Res. 247: 537–546
42 Bye W. A., Allan C. H. and Trier J. S. (1984) Structure,
distribution and origin of M cells in Peyer’s patches of mouse
ileum. Gastroenterology 86: 789–801
61 Kato T. (1990) A study of secretory immunoglobulin A on
membranous epithelial cells (M cells) and adjacent absorptive
cells of rabbit Peyer’s patches. Gastroenterol. Jpn. 25: 15–23
43 Wolf J. L., Rubin D. H., Finberg R., Kauffman R. S., Sharpe 62 Weltzin R., Lucia-Jandris P., Michetti P., Fields B. N., Krae-
A. H., Trier J. S. et al. (1981) Intestinal M cells: a pathway
for entry of reovirus into the host. Science 212: 471–472
44 Amerongen H. M., Weltzin R., Farnet C. M., Michetti P.,
Haseltine W. A. and Neutra M. R. (1991) Transepithelial
transport of HIV-1 by intestinal M cells: a mechanism for
transmission of AIDS. J. Acquir. Immune Defic. Syndr. 4:
760–765
henbuhl J. P. and Neutra M. R. (1989) Binding and transep-
ithelial transport of immunoglobulins by intestinal M cells:
demonstration using monoclonal IgA antibodies against en-
teric viral proteins. J. Cell Biol. 108: 1673–1685
63 Corthesy B., Kaufmann M., Phalipon A., Peitsch M., Neutra
M. R. and Kraehenbuhl J.-P. (1996) A pathogen-specific
epitope inserted into recombinant secretory immunoglobulin
A is immunogenic by the oral route. J. Biol. Chem. 271:
33670–33677
45 Owen R. L., Pierce N. F., Apple R. T. and Cray W. C. Jr
(1986) M cell transport of Vibrio cholerae from the intestinal
lumen into Peyer’s patches: a mechanism for antigen sampling 64 Pappo J., Ermak T. H. and Steger H. J. (1991) Monoclonal
and for microbial transepithelial migration. J. Infect. Dis. 153:
antibody-directed targeting of fluorescent polystyrene micro-
spheres to Peyer’s patch M cells. Immunology 73: 277–280
1108–1118
46 Neutra M. R., Frey A. and Kraehenbuhl J.-P. (1996) Epithe- 65 Foster N., Clark M. A., Jepson M. A. and Hirst B. H. (1998)
lial M cells: gateways for mucosal infection and immuniza-
tion. Cell 86: 345–348
47 Allan C. H., Mendrick D. L. and Trier J. S. (1993) Rat 66 Kedinger M., Lefebvre O., Duluc I., Freund J. N. and
Ulex europaeus 1 lectin targets microspheres to mouse Peyer’s
patch M-cells in vivo. Vaccine 16: 536–541
intestinal M cells contain acidic endosomal-lysosomal com-
partments and express class II major histocompatibility com-
plex determinants. Gastroenterology 104: 698–708
Simon-Assmann P. (1998) Cellular and molecular partners
involved in gut morphogenesis and differentiation. Philos.
Trans. R. Soc. Lond. B. Biol. Sci. 353: 847–856

332
L. J. Hathaway and J.-P. Kraehenbuhl
The role of M cells in mucosal immunity
67 Gebert A., Fassbender S., Werner K. and Weissferdt A.
(1999) The development of M cells in Peyer’s patches is
restricted to specialized dome-associated crypts. Am. J.
Pathol. 154: 1573–1582
68 Kerneis S., Bogdanova A., Kraehenbuhl J.-P. and Pringault
E. (1997) Conversion by Peyer’s patch lymphocytes of human
enterocytes into M cells that transport bacteria. Science 277:
949–952
69 Kerneis S. and Pringault E. (1999) Plasticity of the gas-
trointestinal epithelium: the M cell paradigm and oppor-
tunism of pathogenic microorganisms. Semin. Immunol. 11:
205–215
70 Takeuchi T., Kitagawa H., Imagawa T. and Uehara M.
(1999) Apoptosis of villous epithelial cells and follicle-associ-
ated epithelial cells in chicken cecum. J. Vet. Med. Sci. 61:
149–154
71 Debard N., Sierro F. and Kraehenbuhl J.-P. (1999) Develop-
ment of Peyer’s patches, follicle-associated epithelium and M
cell: lessons from immunodeficient and knockout mice. Semin.
Immunol. 11: 183–191
72 Meynell H. M., Thomas N. W., James P. S., Holland J.,
Taussig M. J. and Nicoletti C. (1999) Up-regulation of micro-
sphere transport across the follicle-associated epithelium of
Peyer’s patch by exposure to Streptococcus pneumoniae R36a.
FASEB J. 6: 611–619
73 Cuvelier C. A., Quatacker J., Mielants H., De Vos M., Veys
E. and Roels H. J. (1994) M-cells are damaged and increased
in number in inflamed human ileal mucosa. Histopathology
24: 417–426
74 Savidge T. C., Smith M. W., James P. S. and Aldred P. (1991)
Salmonella-induced M-cell formation in germ-free mouse
Peyer’s patch tissue. Am. J. Pathol. 139: 177–184
.