| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Swiss Institute for Experimental Cancer Research, Epalinges, Switzerland; and
Division of Immunology and Allergy, Centre Hospitalier Universitaire Vaudois, Lausanne, Switzerland
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Binding of the toxins to cell surface exposed receptors constitutes the
first step for enterotoxicity (10, 11). There have been a
number of studies attempting to identify the cell receptors specific
for each toxin (12, 13, 14). The trisaccharide
Gal
3Galß4GlcNacß appears to be the minimum receptor structure
for binding toxin A (15), whereas no receptor for toxin B
has yet been identified. The substance P receptor neurokinin-1 is
required for toxin A-mediated acute inflammatory responses in mouse
intestine (16). After binding to their putative receptors,
toxins A and B enter the cell (17) and cause alterations
to the actin microfilament network through glycosylation of the Rho
protein family (18, 19, 20). Inactivation of Rho proteins
alters epithelial permeability by loosening the tight junctions without
producing cell death. The differing glycosyltransferase activity
of toxins A and B appears to be the main determinant contributing to
the difference in cytotoxic potency (6, 21).
Therapies other than antibiotics have been used in refractory or recurrent C. difficile infections. Anion-exchange resins binding to the toxins or competition by orally administered lactobacilli or nonpathogenic yeasts prove to be promising in the clinic (22, 23, 24). A number of clinical studies indicate that anti-toxin Ab responses in both serum and intestinal secretions may be protective, whereas an inadequate response predisposes to recurrent infection (25). Active immunizations using the C-terminal portion of toxin A (26, 27) and passive transfer of Abs to toxins A and B protect animals and prevent relapse of infection (28, 29, 30), yet a better molecular understanding of the molecular and cellular partners involved will help to design novel approaches for treatment and prevention. Epitopes have been mapped on the two toxins (31), but neutralizing epitopes have only been found associated with toxin A (32, 33), whereas IgG Ab binding toxin B epitopes was unable to inhibit its cytotoxicity in cell culture (34).
In this paper, we have examined the capacity of three molecular forms (IgG, IgAm, and IgAd/p)4 of a mAb specific for toxin A to neutralize the toxins applied either alone or together on the lumenal (apical) or serosal (basolateral) side of cultured human intestinal T84 epithelial monolayers. The toxin A-specific Ab PCG-4 (IgG2a) was able to neutralize the combined action of the two toxins in the apical compartment only. This suggests that opening of the tight junctions by toxin A is a prerequisite for toxin B action. The neutralization properties of PCG-4 IgG2a were lost at 10 h of apical exposure and 6 h when the Ab was added basolaterally. Conversion of PCG-4 IgG2a into IgAd/p with the same Fv domains extended protection for at least 24 h in both the apical and basolateral compartments. By comparing different Ig concentrations and kinetics of neutralization, we demonstrated that toxin A-specific IgAd/p was at least four times more potent than IgG and IgAm. The neutralizing capacity of IgAm resembled that of IgG2a, thus showing that the avidity of IgAd/p contributes to the increased biological activity of the Ab. In addition to better defining the mode of action of toxin B, the data provide solid molecular evidence that an efficacious vaccine should favor production of IgAd/p on both sides of the mucosal epithelium.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The PCG-4 hybridoma cells (35) were provided by OraVax (Cambridge, MA). Human colon carcinoma T84 cells (CCL 248) and Chinese hamster ovary (CHO) dhfr- cells (CRL 9096) were purchased from American Type Culture Collection (Manassas, VA). Purified toxins A and B from the C. difficile strain VPI 10463 was provided by OraVax. Purified toxins were diluted in PBS to a final concentration of 0.51 mg/ml for toxin A and 0.31 mg/ml for toxin B and stored at -70°C. Production and purification of toxin A-specific PCG-4 IgG2a and recombinant chimeric IgAm and IgAd/p are described below.
Cell culture conditions
T84 cells were cultured and passaged in a 1:1 mixture of DMEM
(1g/l glucose) and Ham’s F12 medium supplemented with 10% FCS (Oxoid
AG, Basel, Switzerland), 10 mM HEPES (pH 7.0), 1 mM sodium pyruvate, 2
mM glutamine, and 100 µg/ml each of penicillin/streptomycin at 37°C
in a water-saturated atmosphere of 95% air and 5%
CO2. To permit cell polarization and
establishment of tight monolayers, cells were seeded onto 24-mm
polycarbonate membrane filters (0.4-µm pore Transwell; Corning
Costar, Cambridge, MA) at a density of 1 x
106 cells/filter or alternatively onto 12-mm
polycarbonate membrane filters (0.4 µm Snapwell; Corning Costar) at a
density of 4 x 105 cells/filter. The
culture medium was replaced freshly every 2–3 days as described
(36). PCG-4 hybridoma cells were grown in 2-liter roller
bottles in RPMI 1640 medium complemented with 10% FCS (Myoclone;
Life-Technologies, Zurich, Switzerland), 25 µM folic acid, 2 mM
glutamine, 1 mM sodium pyruvate, and 100 µg/ml each of
streptomycin-penicillin. CHO cells were cultured in -MEM
supplemented with 10% FCS (dialyzed grade; HyClone, Logan, UT), 10 mM
HEPES (pH 7.0), and 50 µg/ml gentamicin at 37°C in a
water-saturated atmosphere of 95% air and 5%
CO2.
Transepithelial resistance measurements
Transepithelial resistance (TER) was monitored using a
resistance system (Millicell-ERS; Millipore, Bedford, MA) for
electrophysiological readings of filter cups. Cells with a number of
passages ranging from 12 to 21 were grown for 12–20 days on Transwell
or Snapwell membrane filters led to the establishment of TER values
ranging from 1250 to 5000 
.cm2. Such
resistance variation is attributable to monolayer to monolayer
variation in tight junction permeability (37).
Purification of the PCG-4 Ab
The IgG2a Ab was purified from the hybridoma supernatant by protein A-Sepharose (Amersham Pharmacia Biotech, Uppsala, Sweden) following the manufacturer’s instructions. The eluate in 0.1 M glycine (pH 2.75) was immediately brought to pH 7.0 using 1.0 M Tris-HCl (pH 8.0). The elution buffer was exchanged for PBS by three subsequent washes in a Centricon filtration unit (cutoff, 50 kDa; Amicon, Berverly, MA), and the protein concentration was measured using the bicinchoninic acid assay (Pierce, Rockford, IL). Aliquots containing 25.3 ± 2.2 mg/ml of purified Ab were stored at 4°C. The Ag-binding activity of the purified Ab was assessed by ELISA using 100 ng of toxin A as coating Ag, 2.53 µg/ml purified Ab, biotinylated sheep anti-mouse Ig (Amersham Pharmacia Biotech) diluted 1:500, and streptavidin-HRP (Amersham Pharmacia Biotech) diluted 1:1000. Under these assay conditions, purified PCG-4 IgG2a yielded 1.998 + 0.002 OD unit at 492 nm.
Generation of toxin A-specific recombinant IgA Abs
Rearranged genomic variable regions coding for the light
(European Molecular Biology Laboratory access number, MMIGVKPCG, 835
bp) and heavy (European Molecular Biology Laboratory access number,
MMIGVHPCG, 563 bp) chains from hybridoma PCG-4 were isolated
(38) and cloned into expression vectors containing the
genes for either the human 2m (1) or 
constant
regions (39). IgAm and
IgAd/p were produced in CHO cells cotransfected
with expression vectors coding for chimeric mouse-human light and heavy
chains, and the human J chain (pcDNAHygro:Jchain) as described
(39). Following G-418 and hygromycin B selection for 2 wk,
a clone (B2) was isolated and cultured in 175-cm2
T-flasks to permit production of recombinant IgA. Crude supernatants
were sterilized by filtration through 0.22-µm Millipore filters
(Millipore), and stored at -20°C in the form of 45-ml aliquots.
Quantitation of murine-human chimeric Abs in crude supernatant was
performed using ELISA (39). The specific Ag-binding
activity of the chimeric Ab was assessed by ELISA using 100 ng of toxin
A as coating Ag, 0.27 µg/ml of Ab, biotinylated rabbit anti-human
-chain IgG (Dako, Copenhagen, Denmark) diluted 1:1000, and
streptavidin-HRP (Amersham Pharmacia Biotech) diluted 1:1000. Under
these assay conditions, the chimeric IgA yielded 1.865 ± 0.009 OD
unit at 492 nm.
Purification of toxin A-specific chimeric polymeric IgA Abs
Crude supernatants (5 x 45 ml) were concentrated 22.5 times using an Amicon Ultrafilter unit PM 10 (Amicon) and purified from serum proteins by chromatography on a Sephacryl S-300 column (1 m x 2.6 cm; Amersham Pharmacia Biotech) equilibrated and run in PBS-0.02% sodium azide. Column fractions were analyzed by immunoblot of SDS/polyacrylamide gels run under nonreducing conditions, and those containing separated IgAd/p and IgAm were pooled. IgA Ab was concentrated by ammonium sulfate precipitation (final concentration, 33 g/l), resuspended in 1 ml of PBS, and desalted by filtration over a prepacked 2.5-ml PD 10 column (Amersham Pharmacia Biotech) equilibrated and run in PBS. Finally, the IgA solutions were sterilized by 0.22-µm filtration, aliquoted in siliconized 1.5-ml plastic tubes, and stored at 4°C.
Exposure of T84 monolayers to toxins and Abs
To evaluate the effect of toxin A or B on TER of T84 cell
monolayers (results shown in Fig. 1
), different concentrations of toxin
A (range, 0.32–3.2 x 10-10 M) or toxin B
(range, 3.7–37 x 10-10 M) were added to
the upper or lower compartments of Transwell and incubated for 2.5
h. Dilutions of the toxins were conducted in complete cell culture
medium. TER values were measured as indicated above using three
independent replicate filters per experiment.
|
), T84-coated filters were
first incubated for 1 h with 0.64 x
10-10 M of toxin A before apical addition of
37 x 10-10 M of toxin B. TER values were
measured after 2.5 and 5 h using five independent replicate
filters per experiment. Control experiments included apical and
basolateral addition of toxins A or B alone. The effect of blocking
toxin A with 3.75 x 10-10 M PCG-4 IgG2a
before combination with toxin B was examined 5 h after addition to
either the apical or basolateral compartments of the Ab-toxin
mixtures.
|
and 6), a constant
concentration of toxin A (3.2 x 10-10 M)
was preincubated with different concentrations of Ab at room
temperature for 30 min in a final volume of 30 µl. For the
neutralization experiment shown in Fig. 4, PCG-4 IgG2a and recombinant
IgAd/p were used at a concentration of 7.5 and
15 x 10-10 M. For the data presented in
Fig. 6, IgG2a, IgAm, and
IgAd/p Ab concentrations ranged from 0.23 x
10-10–7.5 x 10-10
M. Following apical or basolateral addition to the filters, TER values
were measured at time 0 and after various periods of incubation using
three independent replicate filters per experiment. Control experiments
were conducted in the absence of toxin or in the presence of the Ab
preparations only.
|
|
The cell monolayers cultured on filters were fixed for 1 h at 4°C with 3% paraformaldehyde in PBS. One half of each filter was embedded in methacrylate and 1-mm-thick sections were prepared and stained with toluidine blue. The other half was permeabilized with 2% Triton X-100 (Pierce) in PBS for 4 min, and cells were then incubated with rabbit anti-ZO-1 (40) (Zymed, South San Francisco, CA) in a humidified atmosphere. After washing, cells or sections were incubated with 10 mg/ml of fluorescein-conjugated anti-rabbit IgG. Preparations were examined under a Axioskop microscope (Zeiss, Seltbach, Switzerland) equipped for epifluorescence or by confocal laser scanning microscopy (Zeiss).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In contrast to colonic cells in Ussing chambers, T84 cells allow
continuous observation for up to 48 h and are sensitive to lower
toxin doses (5). Thus, we selected this human cell line to
conduct the series of experiments reported here. Because the effect of
the two C. difficile toxins on T84 cell monolayers grown on
permeable filters has been examined upon apical delivery exclusively,
we first determined the polarity of their activity in the same set of
experiments. In cultured T84 cell monolayers, previous results indicate
that TER changes in responses to toxin A are a more sensitive indicator
of barrier function than mannitol flux (8). We therefore
measured the TER of the monolayer after continuous exposure for
2.5 h of increasing concentrations of toxin A (0.32–3.2 x
10-10 M) or B (3.7 x
10-10–37 x 10-10
M) to the apical or basolateral compartment (Fig. 1
). TER was already affected with the
lowest concentration of toxin A used, and substantially reduced when
the T84 monolayer was exposed to 1.9 x
10-10 M of toxin A. In controls, TER values
remained above 1000 .cm2. In contrast, toxin B
was without effect from the apical side, whereas when basolaterally
applied, a marked drop in TER was observed with the lowest
concentration tested. This first series of data suggest a so far
unreported polarized mode of action for toxin B.
A drop in TER values is indicative of a loss of epithelial integrity.
Thus, we analyzed the morphology of T84 monolayers exposed for 6 or
24 h to 3.2 x 10-10 M of toxin A
added to the upper chamber of Transwell filters using an Axioskop
device. At 6 h, toxin A caused a slight enlargement of the space
between the epithelial cells without significant cell
alterations (compare A and B of Fig. 2), whereas at 24 h, the monolayer
was completely disorganized as reflected by swollen and vacuolated
cells (Fig. 2C). Toxin A or B added to the basolateral side
triggered changes similar to those of toxin A added apically, with a
more rapid and pronounced effect for basolateral toxin B (data not
shown). We next examined the organization of the tight junction network
using a ZO-1-specific Ab and confocal microscopy. Tight junctions form
a continuous network that seals the apex of the cells (Fig. 2D). After 6 h of apical treatment with toxin A, the
tight junction network was minimally disrupted (Fig. 2E),
whereas exposure for 24 h caused a complete disorganization with
disappearance of ZO-1 labeling (Fig. 2F). When applied
basolaterally, toxin A completely disrupted the tight junction network
after 6 h (Fig. 2G), and there was no further change
with prolonged treatment (Fig. 2H). No alteration was seen
with toxin B applied apically for 6 or 24 h, whereas basolaterally
the lesions are indistinguishable from those seen with toxin A,
although seen more rapidly (data not shown).
|
Toxin A makes toxin B delivered in the apical surface cytotoxic
Consistent with this hypothesis, it is conceivable that the
opening of the tight junctions by toxin A will allow toxin B to reach
the basolateral membrane where its cytotoxic effect will take place and
potentiate the action of toxin A on the damaged monolayer. The
concentration of toxin A (0.64 x 10-10 M)
was adjusted so that the reduction of TER was only partial and thus
permitted to detect subtle changes due to the addition of 37 x
10-10 M toxin B in the apical compartment. In
the presence of the two toxins, there was a more pronounced reduction
of TER after 2.5 h of incubation compared with toxin A alone (Fig. 3A, lanes A/- and
A + B/-). The more marked combinatory effect of both toxins
was still significant after 5 h (p >
0.05), yet toxin A begins to exhibit a cytotoxic effect by itself. In
agreement with data in Fig. 1, toxin B alone did not have any effect on
the TER (Fig. 3A, compare lane B/- with
lane -/-). Conversely, the addition of toxin A, toxin B,
or a combination of the two toxins to the basolateral compartment (Fig. 3B, lanes A/-, B/-, and A +
B/-) induced a drastic drop in TER values. Therefore, our finding
that apical coapplication of toxin B and toxin A synergistically affect
the T84 monolayer TER confirms the possibility that toxin B exerts its
action to the basolateral surface after the initial disruption of tight
junctions by toxin A.
We thus postulated that an Ab able to neutralize toxin A should protect
the epithelial cell monolayers when the two toxins are added together
apically. We tested the protection afforded by PCG-4 IgG2a Ab, which
recognizes two epitopes within the repeating units of toxin A
(35) and is known to neutralize the toxin’s enterotoxic
activity in rabbit ileal loops (41). As shown in Fig. 3A (lanes A + B/+), PCG-4 mAb added with
toxins A and B to the apical compartment protected the monolayers to
the same extent as when mixed with toxin A alone (Fig. 3A,
lane A/+). This argues in favor of the necessary role of
toxin A in making apical toxin B active against the epithelial cell
monolayer. Consistently, PCG-4 IgG2a was inefficient in preventing the
cytotoxic effect of toxin B when the latter was added to the
basolateral surface (Fig. 3B, lanes B/+ and A +
B/+). Further dissection of the combined role of toxins A and B
will await the availability of neutralizing Abs for toxin B.
Time dependence of toxin A neutralization by IgG2a Abs
T84 cell monolayers were treated with toxin A [3.2 x
10-10 M: the highest toxin A concentration
tested in the dose-response curve (Fig. 1)] and purified IgG2a (7.5 or
15 x 10-10 M) for 3, 4, 6, 8 10, and
24 h. On the apical side, protection was efficient for the first
8 h at both Ab concentrations, whereas a drop in TER below the
reference value was observed at 10 and 24 h (Fig. 4A). Consistently, little
alteration in the tight junction network was observed after 6 h;
at 24 h, the morphological consequence were much more pronounced
(Fig. 4, B and C). In contrast, protection on the
basolateral side was efficient for the first 4 h only (Fig. 4D), as reflected by the severe disruption observed at
6 h and disappearance of the ZO-1 signal at 24 h (Fig. 4, E and F). A differential apical and basolateral
expression of toxin receptor(s) might explain variations in kinetics of
cellular damages. In the absence of a clearly identified receptor, we
postulated that the limitation in the duration of protection might well
result from a dynamic exchange of toxin A between the Ab and its
binding sites on the epithelial cells. We addressed this possibility by
comparing the protection capacity of IgAd/p,
IgAm, and IgG2a Abs carrying the same Fv
fragments derived from PCG-4 and hence expressing the same Ag-binding
specificity.
Toxin A-specific IgAd/p Abs enhance efficacy and duration of neutralization
To examine the neutralizing capacity of monomeric and polymeric
Abs, we cloned the genes coding for the variable regions of the PCG-4
Ab into vectors carrying human constant heavy and light chain genes
(39). The constructs were expressed together with a human
J chain expression vector in stably transfected CHO cells.
IgAd/p and IgAm Ab
molecules were recovered from clone B2 and separated by sieving
chromatography (Fig. 5A). The
identity of the Ab molecular forms in column fractions corresponding to
the peaks was confirmed by Western blot analysis of samples subjected
to SDS-PAGE under nonreducing conditions (Fig. 5B).
Postpurification identical amounts of IgAm and
IgAd/p were checked by ELISA for their toxin
A-binding activity (H. Stubbe, unpublished observations). To
specifically assess the role of each Ig as such, we did not combine IgA
with secretory component, as the latter has been shown to bind to toxin
A (42).
|
, G and J). Maintenance of TER values
correlated with the integrity of the tight junction network (Fig. 4, H and I and K and L).
Direct comparison of the ZO-1 pattern at 6 and 24 h following
apical or basolateral application of the Ab-toxin A complex indicated
that, at Ab:Ag molar ratios of either 2.3 or 4.6,
IgAd/p fully protected the epithelial monolayer.
In agreement with our working hypothesis, this suggests that the
Ab-toxin A complex is much more stable when cross-linking is mediated
by the multivalent IgAd/p Ab.
Since IgG and IgAd/p carry the same Fv fragments
and therefore have the same basal affinity for the Ag, we next tested
whether the difference in avidity accounts for the enhanced duration of
protection. Toxin A (3.2 x 10-10 M) was
added with decreasing concentrations of IgG,
IgAm, and IgAd/p in the
apical compartment and TER was monitored after 2.5, 4, and 6 h
(Fig. 6). At 2.5 h,
IgAd/p concentrations as low as 0.46 x
10-10 M neutralized the toxin, whereas either
monomeric Ig at the same low concentration gave TER values below the
range of confidence. A similar capacity to maintain TER values was
obtained with 7.5 x 10-10 M of either
monomeric Ig. This indicates that in this experimental setting,
IgAd/p is 16-fold more efficacious than its
monomeric counterpart, and that 1 mol of IgAd/p
can neutralize 6 mol of toxin A (see Discussion for possible
explanations). At 4 h, 1.9 x 10-10 M
of IgAd/p still fully prevented toxin A-induced
TER drop. This shows that IgAd/p remains superior
to monomeric Ig in neutralizing toxin A after longer exposure time. At
6 h, a slight molar excess of IgAd/p
(3.7 x 10-10 M) over toxin A was still
able to maintain TER to control values, whereas 7.5 x
10-10 M monomeric IgG2a or
IgAm could no longer neutralize toxin in the
assay. This suggests that free, non-neutralized toxin A capable to
exert its destructive effect becomes available to the T84 cell surface
more slowly when the Ag is complexed to IgAd/p.
In addition, direct comparison between different isotypes with the same
variable domains indicate that IgAd/p provide
more efficacious and prolonged protection on both surfaces of the
epithelial monolayer than monomeric Ig. The biologic implication of
these findings is discussed below.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
and 2). One implication is that the two
toxins, despite structural similarities, recognize distinct epithelial
cell receptors with the expression of the receptor for toxin B
restricted to the basolateral cell surface. An inaccessibility (or
lack) of receptors for toxin B on enterocytes has been suggested to
explain the discrepancy in intestinal activities of toxins A and B
(43). Identification of this putative receptor will
require that polarized cells are used to permit its initial biochemical
characterization. The difference in sensitivity of apical and
basolateral surfaces to pathogens has already been reported
(44) and suggests that cell polarity of intact epithelia
is likely to contribute to mucosal defense. Furthermore, by using short
times of incubation (2.5 and 5 h) and suboptimal concentrations of
toxin A, we have shown that toxin B is rapidly active when delivered at
the apical surface (Fig. 3). This would agree with the strongly
potentiating effect of toxin B on sublethal amounts of toxin A
eliciting enterotoxic effects in rabbits, rats, mice, and hamsters
(4, 9). Thus, the culture model mimics very well the in
vivo situation seen in experimental animals. Moreover, the experiments
performed with anti-toxin A IgG2a imply that neutralization of
toxin A might be sufficient to prevent initial damage and basolateral
delivery of toxin B to its preferential site of action. Because toxin B
has been shown to induce the release of proinflammatory cytokines and
chemokines by lamina propria leukocytes, including mast cells
(45), macrophages (46, 47), and neutrophils
(48, 49), inhibition of its diffusion into the serosal
compartment should reduce tissue damage and propagation of the
bacterium. Another issue challenged by our results is the type of optimal detection which should be conducted on stool specimens subjected to clinical analysis. Many hospital laboratories do not perform assays for both toxins and consequently it is important to determine which C. difficile toxin(s) is(are) prevalent and responsible for pathogenesis (50). In pediatric patients, testing for toxin A would have detected 50% of the infection, whereas testing for toxin B alone would have detected C. difficile in 82% of the cases (51). Given that toxin A at low concentration makes epithelial surfaces sensitive to toxin B, our data suggest that a negative test for toxin A cannot exclude C. difficile infection. Failure to diagnose C. difficile in the early stage of infection may result in worsening of the patient condition and increased costs for subsequent treatment.
Direct comparative studies of local immunity directed against infections of mucosal surfaces have been hindered in the past because of the problem of obtaining sufficient amounts of Ab molecules with the same specificity, but of distinct isotypes. The advantages of recombinant Abs are numerous: their degree of purity and achievable concentration are well above those of Ab purified from body fluids including bile, colostrum, nasal, or intestinal washings; purified Ig samples are devoid of cellular/glandular contaminants such as other Igs and lymphokines known to modulate epithelial function (52); the specificity and isotyping of the Ab are strictly defined, avoiding side effects due to uncontroled mixture of components; and for several biochemical criteria, IgA produced in CHO cells is identical to that found in natural secretions (39). Clearly, the recombinant Ig molecules produced in this study represent the best tool so far to strictly investigate how IgG, IgAm, and IgAd/p carrying the same Fv fragments neutralize a bacterial toxin on the surface of reconstituted monolayers of epithelial T84 cells.
Our data provide strong molecular evidence that the avidity of IgAd/p Ab increases both its intrinsic efficacy as predicted, but also the duration of its neutralizing capacity. Although not expected from in vitro experiments, this observation suggests a mechanism whereby the favored tethering of toxin A on IgAd/p limits the productive contact between the toxin and the epithelial cells leading to G protein activation and subsequent disruption of the epithelial monolayer. Consistent with this mechanism derived from the pure in vitro experimental setting used here, colonic aspirates from healthy individuals frequently contain secretory IgA (S-IgA) Ab molecules to toxin A able to block the binding of toxin A to its intestinal receptor (53). Moreover, colostral S-IgA has been shown to neutralize both toxins A and B (54), which may explain the low frequency of detection of C. difficile toxins in newborns and young infants (51). Interestingly, in the absence of mucus and peristaltic movement, IgAd/p Ab remains much more efficient than monomeric IgG, indicating that immune exclusion believed to prevent diffusion of cross-linked Ags is also operative in vitro.
Remarkably, IgAd/p was still operative at 24
h, whereas the IgG molecule was no longer able to prevent the toxin A
effects. Thus, neutralization by IgAd/p and IgG
are quantitatively different phenomena. The nature of this difference
was examined by preparing IgAm, which behavior
was found to be close to that of the IgG counterpart. Therefore, the
polymeric nature of IgAd/p, along with its
concomitant increased avidity, explains the difference in biological
function. One can argue that the receptor on epithelial cells can 1)
act as scavenger for the Ab molecules, and thus preclude their
neutralizing activity, or 2) favor cellular entrance of Ab-Ag
complexes, thus potentiating the action of the toxin. First, we were
unable to detect FcRn (55) on the surface of T84 cells (B.
Corthésy, unpublished observations), implying that the
concentration of IgG Ab remains stable and that IgG-Ag complexes cannot
exert any intracellular activity. Second, the very weak expression of
pIgR at the basolateral surface of T84 does not reduce the amount of
IgAd/p capable of neutralizing toxin A, nor does
it transcytose Ab-Ag complexes to a detectable level. Finally, the
expression of the Fc receptor (CD89) is limited to cells of the
myeloid lineage (56) and does not intervene in the process
under study here.
The data in Fig. 5 indicate that dimeric IgA-containing fractions also
contain a significant amount of higher molecular
weight IgA polymers. This can explain why we observed that 1 mol of
recombinant IgAd/p could neutralize 6 mol of
toxin A (Fig. 4), the valence of the Ig molecules in the assay being on
average higher than 4. Carbohydrate moieties on IgA could also
contribute nonspecific binding sites for toxin A (57), and
thus enhance toxin neutralization as already reported for
Escherichia coli type 1 fimbriae lectin Ag
(58). In intestinal secretions,
IgAd/p molecules are associated with the
secretory component, which is heavily glycosylated (59).
Given that the secretory component is a potent binder of toxin A
(42), this may help to tether even more toxin A on the Ab
molecule. In addition to blocking bacterial attachment
(60), this might represent another mechanism whereby
sugars on Ig molecules can exert their biological function.
Based on our observations, it appears that an efficient C. difficile vaccine should elicit Abs that neutralize toxin A in the gut lumen and inactivate both toxins A and B in the lamina propria. Neutralization of pathogens or toxins in the gut lumen is best accomplished by S-IgA, which are produced as IgAd/p and transported in secretions by receptor-mediated transcytosis (61). Local IgA production and protective secretory immune responses in mucosal tissues require mucosal vaccination (62, 63), yet this concept appears to be challenged in the case of C. difficile vaccination. Indeed, Formalin-inactivated culture filtrates from a highly toxigenic C. difficile strain induced high S-IgA responses, but low protection, when administered rectally and intragastrically, whereas intranasal, i.p., and s.c. routes were 100% protective against death but only partially protective against diarrhea (64). Although a combination of nasal and peritoneal immunization fully protected hamsters from both death and diarrhea (64), the same group also reported that rectal immunization in conjunction with i.m. vaccination provided full protection; protection was associated with high levels of toxin-neutralizing Abs in serum (65). Rabbits inoculated orally with an attenuated Vibrio cholerae strain expressing a truncated form of toxin A induced both a systemic and mucosal protective immunity against this toxin (27). Together, these experiments indicate that optimal protection against C. difficile infection requires Abs in both the lumenal and serosal compartments, consistent with our in vitro findings. Significantly, in patients with diarrhea, IgAm and IgAd/p in the serum, not IgG, have been implicated in C. difficile neutralization (66). This is consistent with the observation that IgA-associated immunity to mucosal Ags can complement that of S-IgA in the defense against pathogens (67).
In conclusion, we demonstrate that an Ab specific for C. difficile toxin A protects human colonic epithelial T84 cell monolayers against both toxins A and B when the Ab acts in the lumenal but not in the serosal compartment. We provide evidence that increasing the avidity of the Ab enhances both efficacy and duration of protection. Although suggested in previous studies (68), our data represent the first demonstration that the neutralizing capacity of model Ab molecules with the same Fv fragments is dependent on their isotype and degree of polymerization. The good parallel observed between in vivo data and the results reported here makes it possible to draw this conclusion, and further suggests that a potent vaccine against C. difficile should elicit a strong mucosal, polymeric IgA-mediated immune response.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Current address: Galli-Valerio Institute, Rue César-Roux 37, CH-1014 Lausanne, Switzerland.
3 Address correspondence and reprint requests to Dr. Blaise Corthésy, Division of Immunology and Allergy, Centre Hospitalier Universitaire Vaudois, BH-18-701, Rue du Bugnon, CH-1011 Lausanne, Switzerland. E-mail address:
4 Abbreviations used in this paper: IgAm, monomeric IgA; IgAd/p, polymeric IgA; S-IgA, secretory IgA; CHO, Chinese hamster ovary; Fv, the variable domain of Ig; TER, transepithelial electrical resistance.
Received for publication September 14, 1999. Accepted for publication December 9, 1999.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-galactose receptors and mimics the effects of Clostridium difficile toxin A in rat colon. Gastroenterology 110:1704.[Medline]
-3
Gal-ß-4 GlcNac-ß-binding Clostridium difficile toxin A
human natural anti--galactosyl IgG and the monoclonal antibody
Gal-13: characterization of a binding-active human glycosphingolipid
non-identical with the animal receptor. Glycobiology 6:599.
, and leukotrienes. Infect. Immun. 65:2740.[Abstract]
R). Crit. Rev. Immunol. 16:423.[Medline]
This article has been cited by other articles:
|
|
 |
|
  N. J. Mantis, J. Palaia, A. J. Hessell, S. Mehta, Z. Zhu, B. Corthesy, M. R. Neutra, D. R. Burton, and E. N. Janoff Inhibition of HIV-1 Infectivity and Epithelial Cell Transfer by Human Monoclonal IgG and IgA Antibodies Carrying the b12 V Region J. Immunol., September 1, 2007; 179(5): 3144 - 3152. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. E. Fasching, T. Grossman, B. Corthesy, A. G. Plaut, J. N. Weiser, and E. N. Janoff Impact of the Molecular Form of Immunoglobulin A on Functional Activity in Defense against Streptococcus pneumoniae Infect. Immun., April 1, 2007; 75(4): 1801 - 1810. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Jank, T. Giesemann, and K. Aktories Rho-glucosylating Clostridium difficile toxins A and B: new insights into structure and function Glycobiology, April 1, 2007; 17(4): 15R - 22R. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Perrier, N. Sprenger, and B. Corthesy Glycans on Secretory Component Participate in Innate Protection against Mucosal Pathogens J. Biol. Chem., May 19, 2006; 281(20): 14280 - 14287. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Meckelein, D. Externest, M. A. Schmidt, and A. Frey Contribution of Serum Immunoglobulin Transudate to the Antibody Immune Status of Murine Intestinal Secretions: Influence of Different Sampling Procedures Clin. Vaccine Immunol., September 1, 2003; 10(5): 831 - 834. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Cottet, I. Corthesy-Theulaz, F. Spertini, and B. Corthesy Microaerophilic Conditions Permit to Mimic in Vitro Events Occurring during in Vivo Helicobacter pylori Infection and to Identify Rho/Ras-associated Proteins in Cellular Signaling J. Biol. Chem., September 6, 2002; 277(37): 33978 - 33986. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I. Schwartz-Cornil, Y. Benureau, H. Greenberg, B. A. Hendrickson, and J. Cohen Heterologous Protection Induced by the Inner Capsid Proteins of Rotavirus Requires Transcytosis of Mucosal Immunoglobulins J. Virol., July 17, 2002; 76(16): 8110 - 8117. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Guy Evaluation of Events Occurring at Mucosal Surfaces: Techniques Used To Collect and Analyze Mucosal Secretions and Cells Clin. Vaccine Immunol., July 1, 2002; 9(4): 753 - 762. [Full Text] [PDF]
|
|
|
|
|
|
A. Roos, L. H. Bouwman, D. J. van Gijlswijk-Janssen, M. C. Faber-Krol, G. L. Stahl, and M. R. Daha Human IgA Activates the Complement System Via the Mannan-Binding Lectin Pathway J. Immunol., September 1, 2001; 167(5): 2861 - 2868. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C PHILLIPS Serum antibody responses to Clostridium difficile toxin A: predictive and protective? Gut, August 1, 2001; 49(2): 167 - 168. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) or in the
presence of 7.5 x 10-10 M (
) or 15 x
10-10 M (
) Ab for the indicated time. Cellular
protection afforded by the mAb was assessed by confocal microscopy
after 6 h (B, E, H,
and K) and 24 h (C,
F, I, and L). Sections
were obtained as in Fig. 2
), after 2.5 h (
), 4 h (•), and 6 h
(
).
