 |
INTRODUCTION |
The Gram-positive bacterium Clostridium perfringens
uses its potent arsenal of 14 toxins to cause enteric and histotoxic
infections in humans and domestic animals (1, 2). Among the most
biomedically important of those toxins is C. perfringens
enterotoxin (CPE),1 which is
responsible (3) for the diarrheal and cramping symptoms of C. perfringens type A food poisoning, the third most commonly identified foodborne disease in the United States (4). CPE plays a
similarly important role when C. perfringens causes
nonfoodborne human gastrointestinal (GI) diseases, such as
antibiotic-associated diarrhea (3, 5), and may also contribute to the
pathogenesis of some veterinary diarrheas (6).
Previous studies demonstrated that CPE can kill enterocytes and other
sensitive mammalian cells (7-9). More recently, this enterotoxin was
also shown (10-12) to affect epithelial tight junctions (TJs). The
mechanism(s) by which CPE causes TJ effects remains unclear, as does
the relationship (if any) between those TJ effects and CPE-induced
cytotoxicity or the contribution of those TJ effects to CPE-mediated GI disease.
The cytotoxic effect of CPE has been extensively studied. This
multistep pathophysiologic process begins when CPE binds to one or more
protein receptors on sensitive mammalian cells (13-20). Upon binding,
CPE remains in the plasma membrane, where it quickly localizes in an
SDS-sensitive small complex of ~90 kDa (21, 22). The small complex,
which contains (at a minimum) the enterotoxin and an ~50-kDa
eucaryotic protein, then interacts with additional eucaryotic proteins
to form SDS-resistant, larger CPE complexes in plasma membranes (18,
21-25). Recent studies (11) using CaCo-2 human intestinal carcinoma
cells CPE-treated in suspension resolved those larger CPE complexes
into at least three distinct species as follows: (i) an ~155-kDa
complex, which was the predominant CPE species formed in those CaCo-2
cells; (ii) an ~200-kDa complex; and (iii) an ~135-kDa complex,
which may serve as an intermediate for formation of the ~155- and
~200-kDa complexes. The eucaryotic protein composition of the CPE
complexes remains to be fully elucidated, but the ~200-kDa complex
formed when CaCo-2 cells are CPE-treated in suspension was shown to
contain the TJ protein occludin (11). Some evidence (18) suggests that
certain members of the claudin family of TJ proteins may also be
present in at least one of the larger CPE complexes.
Substantial evidence indicates that formation of one (or more) of the
three larger CPE-containing complexes is responsible for the plasma
membrane permeability alterations that disrupt the normal
colloid-osmotic equilibrium (26, 27) and thereby cause death of the
CPE-treated cell from either lysis or metabolic disturbances (28). The
earliest evidence supporting the importance of larger complexes for
cytotoxicity was provided by the observation that CPE induces neither
membrane permeability alterations nor cytotoxicity at 4 °C, a
temperature where the toxin can bind and form small complex but cannot
form larger CPE complexes (24). Later, studies characterizing several
CPE deletion fragments (25) demonstrated that the cytotoxic activity of
those fragments closely correlates with their ability to form larger
CPE complexes. Finally, recent studies established that several CPE
point mutants that fail to kill mammalian cells are also specifically
blocked for larger CPE complex formation (22).
Most information regarding the cytotoxic action of CPE has been gleaned
from cell culture studies. Early CPE mechanism of action studies
primarily utilized Vero (African green monkey kidney) cells. However,
recent CPE action studies have increasingly employed CaCo-2 cells, with
the rationale that those polarized intestinal epithelial cells
represent a better model for the natural CPE target, i.e.
enterocytes. Despite the widespread use of Vero and CaCo-2 cells for
dissecting CPE action, the responsiveness of those two cell lines to
CPE treatment has not yet been directly compared. When such comparative
analyses were performed in the current study, a number of important new
insights into CPE action were obtained.
| |
EXPERIMENTAL PROCEDURES |
Materials
CPE was purified to homogeneity from C. perfringens
strain NCTC 8239. The biological activity of that purified toxin was
assayed, as described previously (29). Aliquots (2 mg) of that purified CPE were radiolabeled by a published protocol (11, 30) using lactoperoxidase-glucose oxidase (Bio-Rad) and 2 mCi of
Na125I (17 mci/mg; ICN Radiochemicals). The biological
activity of the 125I-CPE was then confirmed using published
protocols (11, 30).
Cell Culture
Vero cells were cultured in medium M199 (ICN) containing 5%
newborn calf serum, 100 units/ml penicillin, and 100 µg/ml
streptomycin. CaCo-2 cells were routinely maintained as described
previously (11). MDCK cells were grown in Dulbecco's modified Eagle's
medium (Life Technologies, Inc.) supplemented with 3% fetal calf
serum, 100 units/ml penicillin, and 100 µg/ml streptomycin. Before
seeding for experiments, each cell line was grown to confluency in
150-cm2 flasks (Corning Glass), which were incubated at
37 °C in a 5% CO2 atmosphere.
Analyses of CPE-induced Cytotoxicity
Monolayer Cultures--
Vero cells and CaCo-2 cells were seeded
at 50,000 cells/well in 24-well plates (Corning Glass) and grown to
confluency (~4-6 days). Those cultures were radiolabeled for 2 h at 37 °C in Hanks' balanced salt solutions (HBSS, Sigma)
containing 4 µCi/well of 86RbCl (2.2 mCi/mg; PerkinElmer
Life Sciences). After radiolabeling, each culture was washed twice with
HBSS and then incubated at 37 °C for 15 min in 2 ml of warm HBSS,
with or without, 1 µg/ml of native CPE.
Following that 15-min incubation, the supernatant from each culture
well was collected, and its radioactivity was determined using a
Beckman gamma counter. As described previously (11, 30), the obtained
counts/min values were used to calculate the percentage of maximal
86Rb release induced by CPE treatment: % of maximal
release = 100 × (86Rb release in CPE-treated
well 
spontaneous 86Rb release)/(maximal 86Rb
release spontaneous 86Rb release). Spontaneous
release represents the background radioactivity released from a
radiolabeled control (no CPE treatment) culture well during a 15-min
incubation at 37 °C. Maximal release corresponds to the total amount
of cytoplasmic 86Rb radioactivity that had been present in
radiolabeled cultures at the time of CPE challenge. Maximal release was
determined, as described (11, 30), by inducing plasma membrane lysis
with the addition of 1 M citric acid and buffer containing
0.5% saponin.
Transwell® Cultures--
Vero, CaCo-2, or MDCK
cells seeded into 12-mm2 diameter Transwells® (Corning
Costar) at a density of 50,000 cells/well were grown to complete
confluency (~4-6 days). Those cultures were then washed twice with
warm HBSS and radiolabeled for 2 h at 37 °C with warm HBSS
containing 4 µCi/well of 86RbCl.
Following radiolabeling, 0.6-ml aliquots of HBSS (with or without 1 µg/ml CPE, as specified) were added to the top and bottom chambers of
each washed Transwell® culture. After 15 min of incubation at
37 °C, the supernatants in the top and bottom chambers of each Transwell® culture were individually removed, and radioactivity present in those supernatants was then determined with a gamma counter.
The resultant values were then used to calculate the percent of maximal
release induced by CPE treatment, as described above for monolayer cultures.
Analysis of 125I-CPE-specific Binding to Vero Cells
and CaCo-2 Cells
Monolayer Cultures--
Vero cells or CaCo-2 cells inoculated
into 60-mm2 tissue culture dishes (Falcon) at a density of
106 cells/well were grown to confluency (~3-5 days).
Those cultures were washed twice with warm HBSS and then treated with 3 ml of HBSS containing 0.5 µg/ml 125I-CPE, with (to
determine nonspecific binding) or without (to determine total binding)
a 100-fold excess of native CPE. After incubation for 15 min at room
temperature (RT) with gentle shaking, the cultures were washed twice
with HBSS to remove unbound 125I-CPE. Adherent cells were
scraped from each dish into 2 ml of HBSS and then combined with
nonadherent cells from that same dish, which had been collected during
washing. The number of cells and amount of radiation present in each
combined sample were determined using a hemocytometer and gamma
counter, respectively. After the counts/min detected in each sample
were converted to nanograms of 125I-CPE, specifically bound
125I-CPE was calculated as described previously (11, 30):
nanograms of 125I-CPE in total binding sample radioactivity in nonspecific binding sample.
Transwell® Cultures--
Vero cells or CaCo-2 cells seeded in
6-well Transwell® plates (24 mm2 diameter) at a density of
106 cells/well were grown to complete confluency (~3-5
days). After two washes with HBSS, a 1.5-ml aliquot of HBSS containing
0.5 µg/ml 125I-CPE, with or without a 100-fold excess of
native CPE, was added to the top or bottom chamber of each washed
Transwell® culture (the other chamber received 1.5 ml of HBSS alone).
Following 15 min of incubation at RT with gentle shaking, the cultures
were washed to remove unbound radiation. Adherent cells were scraped from each Transwell® and combined with the nonadherent cells from that
same culture, which had been collected during washing. The number of
cells and amount of radiation present in each combined sample were then
determined to calculate the nanograms of 125I-CPE
specifically bound/106 cells, as described above for
monolayer cultures.
Large Complex Formation Analyses
Formation of Large Complexes in Isolated Cells--
A
150-cm2 flask containing a confluent (4-6-day-old) culture
of either Vero cells or CaCo-2 cells was gently scraped with a rubber
policeman. Those isolated cells were then resuspended in HBSS with, or
without, CPE (1 µg/ml) and incubated in suspension for 20 min at
37 °C. After two washes with HBSS, the isolated, CPE-treated cells
were resuspended in 400 µl of PBS containing CaCl2 and
MgCl2.
Formation of Large Complexes in Monolayer Cultures--
Another
150-cm2 confluent flask containing CaCo-2 cells or Vero
cells was washed twice with warm HBSS (no scraping). Those cultures
were then treated with 1 µg/ml CPE for 20 min at 37 °C. After that
incubation, the CPE-treated cultures were washed twice with HBSS,
gently scraped into 4 ml of PBS, and centrifuged. The resultant cell
pellet was then resuspended in 400 µl of PBS containing CaCl2 and MgCl2.
Formation of Large Complexes in Transwell® Cultures--
Highly
confluent (4-6 day) cultures of Vero cells or CaCo-2 cells grown in
6-well Transwell® plates were washed twice with HBSS added to both the
upper or lower Transwell® chambers. HBSS (1.5 ml), with or without 1 µg/ml CPE, was then added (as specified) to either the top or bottom
Transwell® chamber; the other Transwell® chamber received 1.5 ml of
HBSS alone. After incubation for 20, 40, 60, or 120 min at 37 °C,
those cultures were gently washed twice with HBSS added to both the top
and bottom chambers. Cells remaining attached to the Transwell®
inserts were then gently scraped into 2 ml of PBS and combined with the
detached cells from that same Transwell® culture, which had been
collected during washing. After centrifugation of the combined sample,
the resultant cell pellet was resuspended in PBS containing
CaCl2 and MgCl2 to bring the final cell
suspension volume to 400 µl.
Western Blot Detection of Large Complexes--
The PBS
suspensions containing Vero cells or CaCo-2 cells CPE-treated in
suspension (i.e. as isolated cells), in monolayers, or in
Transwells® (as described above) were counted using a hemocytometer. Equal numbers (~2 × 106) of cells from each sample
were then resuspended in 15 µl of DNase (1 µg/µl) and 35 µl of
PBS, followed by the addition of 50 µl of 2× SDS sample buffer (2%
SDS, 10% glycerol, 0.02% bromphenol blue, 60 mM Tris pH
6.8; no 
-mercaptoethanol). After 10 min of incubation at RT, the
mixtures were electrophoresed (without sample boiling) overnight at 4 mA on a 4% polyacrylamide gel containing 0.1% SDS, as described
previously (11). Separated proteins were then electrotransferred onto
nitrocellulose (Millipore) and subjected to Western immunoblotting with
either anti-CPE rabbit polyclonal antibodies or affinity-purified
polyclonal rabbit antibodies raised against the C-terminal region of
human occludin (Zymed Laboratories Inc., Inc.), as
described previously (11).
Western Blot Analysis of Occludin Expression in Vero Cells
and CaCo-2 Cells
Vero cells and CaCo-2 cells grown to confluency (4-6 days) in
monolayer cultures were washed twice with PBS containing
CaCl2 and MgCl2. The washed monolayers were
then detached with Versene for 15 min at 37 °C and counted with a
hemocytometer. Aliquots containing 3.5 × 105 detached
cells of each cell line were resuspended in 50 µl of PBS and treated
with 50 µl of 2× SDS sample buffer (containing -mercaptoethanol).
That mixture was boiled for 5 min and then electrophoresed at 30 mA on
10% polyacrylamide gels containing SDS. As described (11), the
separated proteins were transferred onto a nitrocellulose membrane and
subjected to Western immunoblot analysis using affinity-purified
polyclonal rabbit antibodies raised against the C-terminal region of
human occludin (Zymed Laboratories Inc., Inc.).
Immunocytochemistry
Vero cells (106 cells/well) plated on glass
coverslips and CaCo-2 cells (106 cells/well) plated in
Transwell® chambers were grown to confluency. For CPE experiments,
CaCo-2 cells were treated on either their apical or basal surface with
1 µg/ml CPE in HBSS for 60 or 120 min at 37 °C in the presence of
5% CO2. Following that incubation, the CaCo-2 cells were
washed once with warm HBSS and fixed in 3% paraformaldehyde for 20 min. After permeabilization with 0.2% Triton X-100 for 15 min and
preincubation in blocking solution (1% bovine serum albumin plus 2%
normal donkey serum) for 30 min, the Transwell® filters were excised
and incubated for 1 h in blocking solution containing mouse
monoclonal anti-occludin (diluted 1:200, Zymed Laboratories
Inc., Inc.) and rabbit polyclonal anti-ZO-1 (diluted 1:200,
Zymed Laboratories Inc., Inc.) antibodies. For immunolabeling of Vero cells grown on coverslips, the cells were washed
once with warm HBSS and fixed for 20 min in either 3% paraformaldehyde or methanol, which gave better occludin visualization in Vero cells but
interfered with actin visualization in Vero cells (data not shown).
Therefore, to visualize actin in Vero cells, a second culture was
separately fixed with 3% paraformaldehyde. Subsequent processing of
methanol-fixed Vero cell cultures was performed as described above for
CaCo-2 cells; processing of paraformaldehyde-fixed Vero cells is
described below.
Paraformaldehyde-fixed CaCo-2 cells and Vero cells or methanol-fixed
Vero cells were washed four times with PBS and incubated for 1 h
in blocking solution containing species cross-absorbed Cy2-conjugated
donkey anti-mouse IgG (1:200, Jackson ImmunoResearch). Also
included in this incubation mix for CaCo-2 cells was 0.1 µg/ml
TRITC-conjugated phalloidin (Sigma) to visualize the actin cytoskeleton. Similar phalloidin staining was performed separately on
the paraformaldehyde-fixed Vero cells. Cells were then washed four
times in PBS, with the first wash containing 0.5 µg/ml
4,6-diamidino-2-phenylindole (DAPI, Sigma) to visualize nuclei. Filters
were mounted under coverslips in Mowoil containing 1%
n-propyl gallate (Sigma) and hardened overnight at 4 °C
before viewing. Microscopy was then performed with a Nikon Eclipse E800
microscope using a 60× PlanApo lens, and images were captured on an
ORCA black and white cooled CCD camera (Hammamatsu). Images were then
processed using OpenLab2 (Improvision) and Adobe Photoshop 5.0 software.
| |
RESULTS |
Larger CPE Complex Formation by Isolated CaCo-2 Cells and Vero
Cells--
As reported recently (11), CPE Western immunoblotting
detected three SDS-resistant CPE complexes when isolated CaCo-2 cells, i.e. cells harvested by gentle scraping, were CPE-treated in
suspension for 20 min (Fig. 1, left
panel). As also observed previously (11), the ~155-kDa complex
was the predominant CPE species present in those isolated CaCo-2 cells,
although appreciable amounts of the ~135- and ~200-kDa
CPE-containing complexes were also present. When those CPE immunoblots
were stripped and re-probed with occludin antibodies, immunoreactivity
was observed only with the ~200-kDa CPE complex (Fig. 1, right
panel), as also reported previously (11).
View larger version (38K):
[in this window]
[in a new window]
|
Fig. 1.
Western immunoblot analysis of large CPE
complex formation in isolated CaCo-2 cells and Vero cells. Equal
numbers (2 × 106) of isolated CaCo-2 and Vero cells,
i.e. cell harvested by gentle scraping, were treated in
suspension with 1 µg/ml CPE for 20 min at 37 °C. After two washes
to remove unbound CPE, those cells were lysed with SDS sample buffer
for 10 min at RT. The resultant lysates were electrophoresed overnight
on a 4% polyacrylamide gel containing SDS. The separated proteins were
then electrotransferred onto a nitrocellulose membrane and
Western-blotted using rabbit polyclonal anti-CPE antibodies (left
panel). After the CPE antibodies were stripped off the blot using
IgG elution buffer (Pierce), the blot was reprobed with rabbit
polyclonal anti-occludin antibody raised against the C terminus of
human occludin (right panel). The arrows on the
left indicate the migration of the three species of CPE
larger complexes (~200, ~155, and ~135
kDa). The two arrows on the right
indicate the migration of myosin (220 kDa) and
-galactosidase (122 kDa), as protein markers.
|
|
Isolated Vero cells similarly harvested and CPE-treated in suspension
for 20 min also formed three SDS-resistant CPE-containing complexes
(Fig. 1, left panel), which co-migrated with the CPE complexes formed in isolated CaCo-2 cells. The predominant larger CPE
complex in the isolated Vero cells was ~155-kDa in size and that
species was present at similar levels as the ~155-kDa CPE complex in
isolated CaCo-2 cells. In contrast, much lower amounts of the
~200-kDa complex were detected in equivalent numbers of isolated Vero
cells versus isolated CaCo-2 cells subjected to similar CPE
treatment (Fig. 1, left panel). The ~200-kDa complex was
the only larger CPE complex present in isolated Vero cells that reacted
with occludin antibodies (Fig. 1, right panel).
Fig. 1 immunoreactivity is attributable to CPE treatment since similar
high Mr immunoreactive bands were not detected after Western immunoblotting of control Vero cells or CaCo-2 cells (data not shown).
Comparison of Larger CPE Complex Formation by Monolayer Cultures of
CaCo-2 Cells and Vero Cells--
Similar CPE Western immunoblotting
experiments demonstrated that 20 min of CPE treatment of monolayer Vero
cell cultures also induces formation of a predominant ~155-kDa CPE
complex, along with small amounts of the ~200-kDa CPE complex
(Fig. 2). However, those CPE complexes
were both present at lower levels than detected in equivalent numbers
of isolated Vero cells subjected, in suspension, to similar CPE
treatment (Fig. 2).
View larger version (38K):
[in this window]
[in a new window]
|
Fig. 2.
Comparative CPE Western immunoblot analysis
of larger complex formation in CaCo-2 cells and Vero cells treated with
CPE as confluent monolayers versus isolated
cells. CaCo-2 cells and Vero cells were treated, as monolayers or
as isolated cells (see Fig. 1 legend), with 1 µg/ml CPE for 20 min at
37 °C. Those cultures were washed twice, scraped gently, and 2 × 106 cells of each sample were then lysed with SDS sample
buffer. The resultant cell lysates were electrophoresed overnight on
4% acrylamide gels containing SDS. After electrotransfer to
nitrocellulose, the separated proteins were subjected to CPE Western
immunoblot analysis, including detection with a chemiluminescent
substrate. The leftmost lane represents the monolayer CaCo-2
sample exposed to x-ray film for 5 min; all other lanes were exposed to
x-ray film for less time. The rightmost lane shows the
migration of purified CPE. Arrows on the left of
the blot indicate the migration of myosin (220 kDa)
and -galactosidase (122 kDa), as marker proteins. The
arrows on the right indicate the location of the
~200-, ~155-, and ~135-kDa CPE complexes,
respectively, if present.
|
|
By using our standard chemiluminescent CPE Western immunoblot detection
conditions, only trace amounts of larger CPE complexes were detected
following 20 min of CPE treatment of CaCo-2 monolayer cultures (Fig.
2). However, longer exposure of those blots to x-ray film revealed the
clear presence of larger CPE complex material (Fig. 2). Therefore,
under identical CPE treatment conditions, less CPE larger complex
material forms in monolayer cultures of CaCo-2 cells than in equivalent
numbers of isolated CaCo-2 cells, isolated Vero cells, or Vero cells
treated with CPE in monolayers. Additionally, those CPE Western blot
analyses demonstrated that an ~155-kDa species is the predominant
larger CPE complex formed by monolayer CaCo-2 cultures, although trace
amounts of an ~200-kDa complex were also detectable.
CPE Cytotoxicity and Binding in Monolayer Cultures of CaCo-2 Cells
and Vero Cells--
Since formation of some (if not all) CPE larger
complexes is necessary for CPE-induced cytotoxicity (see Introduction),
experiments were performed to assess whether the different amounts of
larger CPE complexes formed by CPE-treated monolayer cultures of CaCo-2 cells versus Vero cells (Fig. 2) reflect differences in the
CPE sensitivities of those cultures. Using a standard 86Rb
release assay for detecting CPE-induced membrane permeability alterations and cytotoxicity (Fig.
3A), our standard 1 µg/ml
CPE dose induced 2-fold more 86Rb release from monolayer
cultures of Vero cells compared with monolayer cultures of CaCo-2
cells. All monolayers remained intact during this 15-min CPE
treatment.
View larger version (32K):
[in this window]
[in a new window]
|
Fig. 3.
Effects of CPE on monolayer cultures of
CaCo-2 cells and Vero cells. A, the cytotoxic effect of
CPE on 86Rb-labeled monolayer cultures. Data shown are the
CPE-induced percent of maximal 86Rb release induced from
CaCo-2 cell or Vero cell monolayers treated with 1 µg/ml CPE for 15 min at 37 °C. Each bar depicts the mean of three separate
experiments, using duplicate samples in each experiment. Error
bars indicate the S.E. B, specific binding of
125I-CPE to monolayer cultures. 125I-CPE (0.5 µg/ml), in the presence or absence of a 100-fold excess of native
CPE, was added for 15 min at RT to monolayer cultures of Vero cells or
CaCo-2 cells, in order to determine nanograms of 125I-CPE
specifically bound/106 cells (see "Experimental
Procedures"). Each bar depicts the mean of three
experiments, with duplicate samples used in each experiment.
Error bars indicate the S.E.
|
|
125I-CPE binding studies were then performed to evaluate
whether the greater CPE sensitivity of monolayer cultures of Vero cells versus CaCo-2 cells result from differences in the CPE
binding capabilities of those cultures. As shown in Fig. 3B,
those binding studies detected nearly 2-fold more 125I-CPE
specific binding to monolayer cultures of Vero cells compared with
monolayer cultures of CaCo-2 cells.
CPE Cytotoxicity in Transwell® Cultures of CaCo-2 Cells and Vero
Cells--
Since differences have been reported between the
sensitivities of apical versus basal mammalian cell surfaces
to other clostridial toxins (31, 32), studies were conducted to compare
the CPE sensitivity of CaCo-2 cells versus Vero cells grown
to confluency in Transwells®. Consistent with the Fig. 3A
monolayer culture cytotoxicity results, which also largely reflect CPE
treatment of the apical cell surface, addition of CPE to upper
Transwell® chambers affected Vero cells more than CaCo-2 cells (Fig.
4A), i.e.
application of our standard 1 µg/ml CPE dose to the apical surface of
Vero cells grown in Transwell® cultures induced nearly complete
release of intracellular 86Rb label within 15 min, whereas
15 min of application of that same CPE dose to the apical surface of
CaCo-2 cells grown in Transwell® cultures released only ~33% of the
intracellular 86Rb label in those cultures. However, CaCo-2
cells were much more sensitive when that same CPE dose was added for 15 min to the bottom chamber of Transwell® cultures (Fig. 4A).
Significant 86Rb release was also observed when that CPE
dose was added for 15 min to the bottom chamber of Transwell® cultures
of Vero cells (Fig. 4A). All monolayers remained intact
during this 15-min CPE treatment (data not shown).
View larger version (32K):
[in this window]
[in a new window]
|
Fig. 4.
CPE treatment of confluent Transwell®
cultures of CaCo-2 cells, MDCK cells, and Vero cells.
A, cytotoxic effects of 1 µg/ml CPE added for 15 min at
37 °C to the top or bottom chambers (as indicated on the
x axis) of confluent, 86Rb-labeled Transwell®
cultures of CaCo-2 cells, MDCK cells, or Vero cells. After completion
of CPE treatment, the supernatants in the top (white
bars) and bottom (black cross-hatched bars)
Transwell® chambers were individually collected, and their
radioactivity was determined with a gamma counter. After correction for
background 86Rb release, the amount of radioactivity
detected in each supernatant is indicated on the y axis as
the percent of maximal 86Rb release (see "Experimental
Procedures"). Each bar on the graph depicts the mean of
three separate experiments, using duplicate samples in each experiment.
Error bars indicate the S.E. B, specific binding
of 125I-CPE to highly confluent, Transwell® cultures
of CaCo-2 cells or Vero cells. 125I-CPE (0.5 µg/ml), in
the presence or absence of a 100-fold excess of native CPE, was added
to the top or bottom chamber (as indicated on x axis) of
confluent Transwell® cultures of CaCo-2 cells or Vero cells for 15 min
at RT in order to determine nanograms of 125I-CPE
specifically bound/106 cells, as described under
"Experimental Procedures." Each bar on the graph depicts
the mean of three separate experiments, using duplicate samples in each
experiment. Error bars indicate the S.E. Bars
without error bars had errors too small to depict.
|
|
In addition to facilitating comparisons between the CPE sensitivity of
the basal versus apical surfaces of CaCo-2 cells and Vero
cells, the Fig. 4A Transwell® culture experiments also
revealed significant differences in the directionality of CPE-induced
86Rb release from Transwell® cultures of CaCo-2 cells
versus Vero cells. CPE treatment released significant
amounts of radioactivity into both the top and bottom chambers of
Transwell® Vero cell cultures, regardless of which chamber had
actually received CPE. In contrast, CPE induced a unidirectional
86Rb release from Transwell® cultures of CaCo-2 cells,
i.e. almost all radioactivity released by the addition of
CPE to the upper chamber of CaCo-2 Transwell® cultures went into the
supernatant of the upper Transwell® chamber, whereas nearly all
radioactivity released when CPE was added to the bottom chamber of
Transwell® CaCo-2 cell cultures went into the supernatant in the lower
Transwell® chamber.
125I-CPE Binding to Transwell® Cultures--
To
investigate whether the Fig. 4A sensitivity differences
detected in Transwell® cultures of CaCo-2 cells and Vero cells CPE-treated on their basolateral versus apical surfaces
reflect CPE binding differences, 125I-CPE binding analyses
were performed. Those studies (Fig. 4B) detected more
125I-CPE-specific binding when the radiolabeled toxin was
added to the bottom (versus top) chamber of Transwell®
cultures of CaCo-2 cells. In contrast, more
125I-CPE-specific binding was detected when the
radiolabeled toxin was added to the top (versus bottom)
chamber of Transwell® cultures of Vero cells.
Formation of Larger CPE Complexes in Transwell® Cultures--
CPE
Western immunoblot studies were then performed to determine which
larger CPE complex species form in Transwell® cultures of Vero and
CaCo-2 cells treated with our standard 1 µg/ml CPE dose for 20 min on
their apical versus basolateral surfaces. Consistent with
Fig. 4B binding results, considerably more CPE binding and ~155-kDa complex formation were detected when that CPE dose was added
for 20 min to the top (versus bottom) chamber of Transwell® cultures of Vero cells (Fig. 5).
Regardless of whether that CPE dose had been added for 20 min to the
top or bottom Transwell® chamber, nearly all CPE bound to those Vero
cells had localized (Fig. 5) in a complex that co-migrates with the
~155-kDa complex formed by CPE treatment of isolated Vero cells or
CaCo-2 cells (Fig. 1). All monolayers remained intact after this 20-min
CPE treatment (data not shown).
View larger version (26K):
[in this window]
[in a new window]
|
Fig. 5.
Formation of larger CPE complexes in
Transwell® cultures of CaCo-2 cells and Vero cells treated
with CPE for 20 min. Confluent Transwell® cultures of CaCo-2
cells and Vero cells were treated with 1 µg/ml CPE, which was added
to the top (T), bottom (B), or both the top and
bottom (T+B) Transwell® chambers for 20 min at 37 °C.
After two washes, those cultures were harvested and counted with a
hemocytometer. Aliquots of 2 × 106 cells from each
sample were lysed with SDS, and the resultant cell lysates were
electrophoresed overnight on a 4% acrylamide gel containing SDS. After
electrotransfer of the separated proteins onto nitrocellulose, those
blots were subjected to CPE Western immunoblot analysis, including
detection with a chemiluminescent substrate. For comparison, the
effects of similar CPE treatment on an equal number of isolated CaCo-2
cells and Vero cells are shown on the leftmost and
rightmost lanes, respectively. Also shown is the
migration of purified CPE (CPE). The arrows at
the right indicate the migration of myosin (220 kDa) and -galactosidase (122 kDa), as marker
proteins. The three arrows at left (marked ~200,
~155, and ~135 kDa) indicate the migration of those
larger CPE complexes, respectively, if present.
|
|
Also consistent with our Fig. 4B binding results, these CPE
immunoblots revealed considerably more CPE binding and ~155-kDa complex formation in samples where CPE had been added for 20 min to the
bottom (versus top) chamber of Transwell® CaCo-2 cell
cultures (Fig. 5). However, whether CPE had been added to the top or
bottom Transwell® chamber, nearly all CPE bound to those CaCo-2 cell cultures was localized in a complex that co-migrates with the ~155-kDa complex formed by similar CPE treatment of isolated CaCo-2 or Vero cells (Fig. 5).
Since, during in vivo GI disease, enterocytes might be
exposed to CPE for longer than the 20 min used in the Fig. 4
experiments, another CPE Western immunoblot was performed where our
standard 1 µg/ml CPE dose was added to highly confluent CaCo-2
Transwell® cultures for 40, 60, or 120 min. After 40 min, the
~155-kDa complex remained the only detectable larger CPE complex
(data not shown), regardless of whether CPE had been added to the top
or bottom Transwell® chamber. However, Transwell® cultures of CaCo-2
cells treated for 60 min with the same CPE dose did form the ~200-kDa complex (Fig. 6A) when CPE was
added to the bottom, but not top, Transwell® chamber. By 120 min, even
those cultures receiving apical CPE treatment had formed the ~200-kDa
complex (Fig. 6B), perhaps along with some ~135-kDa
complex. In those experiments, the Transwell® CaCo-2 cell cultures
treated on their basal surface with a 1 µg/ml CPE dose were
exhibiting significant disruption of their monolayer integrity by 60 min; in contrast, apical application of this CPE dose only disrupted
monolayer integrity after 120 min.
View larger version (40K):
[in this window]
[in a new window]
|
Fig. 6.
The effects of longer CPE treatment on larger
CPE complex formation by Transwell® cultures of CaCo-2
cells. A, confluent Transwell® cultures of CaCo-2
cells were treated for 60 min at 37 °C with 1 µg/ml CPE, which was
added to either the top or bottom Transwell®
chamber. After incubation the cells were harvested, washed, and lysed
with SDS sample buffer for 10 min at RT. The lysates were
electrophoresed overnight on a 4% polyacrylamide gel containing SDS.
After electrotransfer, larger CPE complexes were detected by CPE
Western blot analysis using rabbit polyclonal anti-CPE antibody. The
CPE lane indicates the migration of free CPE in this gel.
The three arrows on the left indicate the
migration of the three large CPE complexes, if present, and the
two arrows on the right indicate the location of
two protein markers, myosin (220 kDa) and
-galactosidase (122 kDa). B, confluent
Transwell® cultures of CaCo-2 cells were treated with 1 µg/ml CPE,
which was added to the top Transwell® chamber for 20 or 120 min at
37 °C. Aliquots containing 2 × 106 cells were
lysed, electrophoresed, and subjected to CPE Western immunoblotting.
Also shown is the migration of purified CPE (CPE) on this
blot. The arrows at the left indicate the
migration of myosin (220 kDa) and -galactosidase
(122 kDa), as marker proteins. The three arrows
at the right (marked ~200, ~155, and
~135 kDa) show the migration of the three larger CPE
complexes, respectively, if present.
|
|
It is also possible that, during GI disease, the apical surface of
enterocytes is exposed to higher CPE concentrations than the 1 µg/ml
CPE dose used in Figs. 1-6 experiments. Therefore, the effect of
adding 10-fold higher CPE doses to the upper Transwell® chamber was
assessed by CPE Western immunoblots. Under that CPE treatment
condition, the ~200-kDa CPE complex formed by 20 min, along with
considerable monolayer damage and cell detachment (data not shown).
Morphologic Comparisons of Control CaCo-2 Cells and Vero
Cells--
A series of immunocytochemistry experiments assessed
whether there might be a morphologic basis for the CPE responsiveness differences noted in Fig. 4 between Transwell® cultures of CaCo-2 cells versus Vero cells. The first of those experiments
showed Transwell® cultures of control Vero cells possess the elongated shape and actin cytoskeleton typical of a fibroblastic cell (Fig. 7A). Those control Vero cells
also showed punctate reactivity with antibodies raised against ZO-1,
which is known to localize at cadherin-based cell-cell contacts in
nonepithelial cells (33, 34). The punctate ZO-1 antibody
reactivity was concentrated in short, discontinuous cell-cell contacts
that also exhibited punctate reactivity with antibodies raised against
the epithelial tight junction protein occludin (Fig. 7A).
View larger version (67K):
[in this window]
[in a new window]
|
Fig. 7.
Analysis of control Vero cell
morphology. A, immunocytochemistry; Vero cells were
grown to confluency in Transwell® cultures, fixed with methanol (for
occludin, ZO-1, and nuclei visualization) or 3% paraformaldehyde (for
actin visualization). The fixed cells were then reacted with occludin,
ZO-1, actin antibodies, or DAPI staining (for nuclei visualization).
The samples were then visualized by immunofluorescence microscopy (see
"Experimental Procedures" for detail). Because methanol staining
interfered with actin visualization in Vero cells, the actin panel
shown is from a separate culture of Vero cells fixed with
paraformaldehyde. Magnification shown is (× 60). B,
expression of occludin in CaCo-2 and Vero cells; 3-5-day-old confluent
CaCo-2 and Vero cells were harvested with versene solution and washed
twice with PBS. Aliquots (3.5 × 105) of both cell
types were lysed with SDS sample buffer (containing
-mercaptoethanol) and boiled for 5 min. The extracts were
electrophoresed on a 10% polyacrylamide gel containing SDS. The
separated proteins on that gel were then electrotransferred onto
nitrocellulose membrane and subjected to Western immunoblotting using
rabbit polyclonal anti-occludin antibodies. The arrow at
right indicates the migration of occludin protein in both
cell types (marker is not shown).
|
|
Given the observed reactivity of Vero cells with occludin antibodies
(Figs. 1 and 7A), Western immunoblot experiments were performed (Fig. 7B) to rigorously evaluate whether Vero
cells express occludin. Those immunoblot studies confirmed that the Vero cell protein reacting with occludin antibodies matches the expected ~65-kDa size of occludin and co-migrates with occludin made
by CaCo-2 cells. Collectively, the Figs. 1 and 7 results indicate that
Vero cells do produce occludin (although less than CaCo-2 cells), but
their occludin localizes in discontinuous cell-cell contacts, rather
than tight junctions.
In contrast to the Fig. 7A immunocytochemical results with
control Vero cells, similar analyses showed that confluent 5-6-day-old Transwell® cultures of control CaCo-2 cells exhibit the
characteristics of polarized epithelial cells producing TJs (Fig.
8). For example, labeling for occludin
revealed a continuous band of immunoreactivity encircling each cell at
the apical end of the lateral membrane (Fig. 8a), where it
aids in creating the TJ barrier. ZO-1 is also concentrated in the same
focal plane as occludin and outlines the periphery of each cell (Fig.
8b).
View larger version (156K):
[in this window]
[in a new window]
|
Fig. 8.
Immunocytochemical analysis of occludin,
ZO-1, actin, and nuclei in confluent cultures of CaCo-2 cells treated
with 1 µg/ml CPE for 60 min. Confluent
cultures of CaCo-2 cells were grown in Transwells® and incubated for
60 min at 37 °C with, or without, 1 µg/ml CPE added to the top or
bottom Transwell® chamber. After washing, those cultures were fixed in
3% paraformaldehyde, permeabilized with 0.2% Triton X-100, and
preincubated in 1% bovine serum albumin and 2% normal donkey serum
for 30 min. Those Transwell® filters were then incubated for 1 h
with mouse monoclonal anti-occludin (Zymed Laboratories
Inc.) and rabbit polyclonal anti-ZO-1 antibodies. For actin
staining the paraformaldehyde-fixed CaCo-2 cells were cross-absorbed
with Cy2-conjugated donkey anti-mouse IgG and treated with 0.1 µg/ml
TRITC-conjugated phalloidin. For nuclei staining, DAPI was used.
Microscopy was performed using a × 60 planApo lens (× 60 total magnification), and images were taken by black and white CCD
camera. a-d show the localization of occludin, ZO-1, actin,
and nuclei respectively, in CPE untreated Cells (CONTROL);
e-h show the localization of occludin, ZO-1, actin and
nuclei, respectively, in cells treated with CPE added to the top
Transwell® chamber (APICAL); and i-l show the
localization of occludin, ZO-1, actin, and nuclei, respectively, in
cells treated with CPE added to the bottom Transwell® chamber
(BASAL).
|
|
CPE Cytotoxic Effects on Other Polarized Epithelial Cells--
The
morphologic results shown in Figs. 7 and 8 indicating that control
CaCo-2 cells, but not Vero cells, form TJs could indicate that the
presence (or absence) of TJs contributes to the differences in CPE
responsiveness detected between those cells in Fig. 3. To test that
hypothesis, 86Rb-release experiments were repeated using
highly confluent, CPE-treated Transwell® cultures of MDCK cells,
another polarized epithelial cell line known to bind CPE and form TJs
(10). Results from those experiments demonstrated that Transwell® MDCK
cell cultures respond to CPE treatment very similarly as Transwell®
cultures of CaCo-2 cells, i.e. Transwell® cultures of MDCK
cells were more sensitive when CPE was applied to their basolateral
(versus apical) surface and they exhibited unidirectional
86Rb release into the Transwell® chamber containing CPE.
Collated with the Fig. 4 CaCo-2 cell results, these MDCK results are
consistent with the hypothesis that TJs affect cellular responsiveness
to CPE treatment.
Immunocytochemical Studies of CPE-treated Transwell® Cultures of
CaCo-2 Cells--
Since confluent cultures of CaCo-2 cells form TJs
containing occludin (Fig. 8) and CPE interacts with occludin via the
~200-kDa CPE complex (Ref. 11 and this study), additional
immunocytochemistry experiments (Figs. 8 and
9) evaluated whether CPE treatment
affects the distribution of occludin and other TJ-associated proteins in confluent Transwell® cultures of CaCo-2 cells. When such CaCo-2 cultures were incubated for 20 min with 1 µg/ml CPE applied to either
the apical or basolateral surfaces, the cultures remained undamaged and
no changes were detected in occludin, actin, or ZO-1 distribution in
CPE-treated versus control cultures (data not shown).
View larger version (140K):
[in this window]
[in a new window]
|
Fig. 9.
Immunocytochemical analysis of occludin,
ZO-1, actin, and nuclei in confluent cultures of CaCo-2 cells treated
with 1 µg/ml CPE for 120 min. Confluent
Transwell® cultures of CaCo-2 cells were incubated for 120 min at
37 °C with, or without, CPE. After washing, the cultures were fixed
in 3% paraformaldehyde, permeabilized with 0.2% Triton X-100, and
preincubated in 1% bovine serum albumin and 2% normal donkey serum
for 30 min. The fixed cultures were then incubated with mouse
monoclonal anti-occludin, rabbit polyclonal anti-ZO-1 antibodies, and
TRITC- conjugated phalloidin stain for the detection of occludin,
ZO-1, and actin cytoskeleton, respectively. For nuclei staining the
same fixed cultures were also incubated with DAPI, followed by PBS
washes. Microscopy was performed as described in the Fig. 8 legend.
a-d show the presence of occludin, ZO-1, actin, and nuclei
respectively, in CPE-untreated cells (CONTROL);
e-h show the distribution of occludin, ZO-1, actin, and
nuclei, respectively, in CaCo-2 cells treated with CPE added to the top
Transwell® chamber (APICAL); i-l show the
presence of occludin, ZO-1, actin, and nuclei, respectively, in CaCo-2
cells treated with CPE added to the bottom Transwell® chamber
(BASAL).
|
|
Similar experiments were then repeated with cultures incubated for 60 min at 37 °C with HBSS alone (control cells) or with HBSS containing
a 1 µg/ml CPE dose applied to either the apical or basolateral cell
surface. As described earlier, occludin (Fig. 8a) and ZO-1
(Fig. 8b) focused at TJs in control CaCo-2 cells, which also
contained F-actin concentrated in an apical perijunctional ring with
microvilli (Fig. 8c) and basal stress fibers (data not shown). Nuclei in the control cells were oblong and rounded. Similar results were observed when CaCo-2 cells were CPE-treated for 60 min on
their apical surface (Fig. 8, e-h).
However, a 60-min application of 1 µg/ml CPE to the basal surface of
Transwell® cultures of CaCo-2 cells caused a partial redistribution of
occludin from the TJ to a centrally located intracellular component
(Fig. 8i). Those cultures exhibited no noticeable change in
their subcellular distribution of ZO-1 (Fig. 8j). The
distribution of F-actin, as visualized by TRITC-phalloidin, also
changed notably after 60 min of basal CPE treatment, with a loss of
visible microvilli and basal stress fibers (data not shown) and a
reduction in the perijunctional actin ring (Fig. 8k). In
those basally CPE-treated CaCo-2 cultures, F-actin was apparently
restricted to a thin ring, in the same focal planes as ZO-1, at the
apical-lateral end of the CaCo-2 cell. Nuclear staining revealed
compacted and angular nuclei (Fig. 8l), as found in dying cells.
Consistent with our Fig. 6 experiments, 60 min of basal CPE treatment
of Transwell® CaCo-2 cell cultures also disrupted monolayer integrity,
as evidenced by small holes visible throughout the monolayer (see Fig.
8, i-l); ZO-1 at tight junctions was sometimes visible at
free edges of the remaining cells in those monolayers (Fig.
8j). However, removal of occludin and actin from TJs was not
restricted to those CaCo-2 cells immediately adjacent to monolayer holes. Collectively, these results indicate that 60 min of CPE exposure
to the basal, but not the apical, surface of CaCo-2 cells damages those
cultures and causes partial internalization of tight junctional
occludin and actin without causing any obvious effect on ZO-1 distribution.
Additional experiments were performed to investigate the effects of
even longer CPE exposure on confluent Transwell® CaCo-2 cell cultures
(Fig. 9). Localization of occludin (Fig. 9a), ZO-1 (Fig.
9b), and actin (Fig. 9c), as well as nuclear
morphology (Fig. 9d), in control cells incubated for 120 min
in HBSS was similar to that described for 60-min control cells.
Interestingly, when 1 µg/ml CPE was applied for 120 min to the apical
surface of CaCo-2 cells, significant morphologic alterations were
detected, with small holes visible throughout the monolayer (Fig. 9,
e-h). No change was noted in ZO-1 distribution, and nuclei
still appeared relatively normal in these cells (Fig. 9, h
and f). However, occludin labeling in the remaining adherent
cells was more diffuse, with some redistribution of occludin to
intracellular puncta (Fig. 9e). Phalloidin labeling revealed
a loss of the diffuse perijunctional actin, with the remaining
F-actin concentrated in the plane of the junctional complex (Fig.
9g). These changes in occludin and actin distribution were
not restricted to cells at the edge of the monolayer holes (Fig. 9,
e and g).
Application of CPE to the basal surface of CaCo-2 cells for 120 min
released most cells from the Transwell® filter insert, with only small
cell patches (as shown on the far right of Fig. 9) remaining
for analysis. Occludin was almost undetectable in the periphery of
those remaining cells, whether or not those cells were at the edge of a
monolayer hole, with most of that protein now concentrated in a central
intracellular compartment (Fig. 9i). However, ZO-1
maintained its reticular TJ appearance in the remaining cells (Fig.
9j). As shown in Fig. 9k, the F-actin normally present in apical microvilli or encircling the cell periphery in a wide
band was no longer visible. Instead, nearly all F-actin in the
remaining CaCo-2 cells was focused at the TJ, in a pattern identical to
ZO-1. Nuclei showed similar morphologic alterations as noted after 60 min of basal CPE treatment (Fig. 8).
Collectively, the Figs. 8 and 9 results indicate the following: (i) CPE
can induce the removal of occludin and actin, but not ZO-1, from TJs of
CaCo-2 cells; (ii) those TJ effects develop concurrently with monolayer
damage but also occur in cells not directly adjacent to monolayer
holes, and (iii) those TJ effects occur more rapidly following CPE
treatment of the basal versus apical surface of CaCo-2 cells.
| |
DISCUSSION |
By comparing the CPE responsiveness of CaCo-2 cells and Vero cells
under several experimental conditions, the current study provides a
number of new insights into CPE action and CPE-mediated GI disease. In
particular, this study helps elucidate the relationships between, and
functions of, the larger CPE complexes. For example, previous kinetic
studies of larger CPE complex formation in isolated CaCo-2 cells (11)
had suggested the ~135-kDa CPE complex might be a precursor for
~155- and ~200-kDa CPE complex formation. However, that possibility
appears inconsistent with our current results indicating the following:
(i) confluent monolayers of CaCo-2 cells or Vero cells treated with CPE
for 20 min contain the ~155-kDa CPE complex but not the ~135-kDa
(CPE complex), and (ii) ~135-kDa CPE complex formation, if any,
occurs in CPE-treated monolayers only after >60 min of CPE treatment.
The composition and contribution to CPE action (if any) of the
~135-kDa CPE complex requires further study.
The formation and possible role of the ~200-kDa CPE complex in CPE
action is also addressed by our study. For example, confluent Transwell® monolayers of CaCo-2 cells (and Vero cells) formed the
~200-kDa CPE complex, but only after they became damaged by CPE-induced cytotoxicity. Therefore, it appears that occludin, which is
present in the ~200-kDa CPE complex (see Ref. 1 and this study), is
initially inaccessible to CPE applied to either the basal or apical
surface of confluent Transwell® CaCo-2 cell monolayers, but later
contacts CPE as a result of cytotoxicity-induced damage. How cellular
damage enhances CPE-occludin interactions is not clear. One possibility
might be that CPE preferentially interacts with nonphosphorylated
occludin, which (unlike apically oriented phosphorylated occludin) is
mostly present on the lateral epithelial cell surface (35) and thus
should be inaccessible to CPE in the absence of cell damage.
Our study also found that ~200-kDa complex formation correlates with
removal of occludin (and actin), but not ZO-1, from TJs to
intracellular spaces in CPE-treated Transwell® CaCo-2 cell cultures.
That correlation is consistent with the ~200-kDa CPE complex
formation, which results (11) from interactions between CPE and
occludin (and perhaps other TJ proteins), triggering an endocytosis
event that removes specific TJ proteins from the cell surface. The
putative removal of occludin from TJs, via endocytosis of the
~200-kDa CPE complex, could be functionally important since occludin
is a structural component of TJs (36). Therefore, the CPE-mediated
removal of occludin from TJs could explain why previous studies
detected TJ structural damage more quickly in hepatocytes treated with
native CPE (12), which can react with occludin, compared with MDCK
cells treated with a C-terminal CPE fragment that can remove claudins
from TJs (10) but cannot form larger CPE complexes (18, 21) and should
not interact with occludin (11).
Interestingly, even the slow-developing TJ structural damage caused by
that C-terminal CPE fragment produced paracellular permeability changes
in MDCK cells (10). Since native CPE (i) interacts with even more TJ
proteins (see Ref. 11 and this study) than the CPE fragment, (ii)
causes nearly complete removal of occludin (and some actin) from TJs
(this study), and (iii) induces structural damage to TJ fibrils (12),
native CPE may produce paracellular permeability alterations in the
intestines. If verified by ongoing studies, such CPE-induced
paracellular permeability effects could contribute significantly to the
diarrheal symptoms of CPE-mediated GI disease.
However, the current data also provide new support for the emerging
view (37) that CPE effects on TJs and paracellular permeability are not
the initial, primary intestinal effects of CPE. For example, the
current study observed that TJ effects develop only slowly in confluent
Transwell® CaCo-2 cell cultures exposed to moderate CPE doses on their
apical surface, which is the epithelial surface initially exposed to
CPE during in vivo GI disease (4). Furthermore, formation of
the ~200-kDa CPE complex containing occludin and removal of occludin
(or actin) from TJs were only detected in confluent Transwell® CaCo-2
cell cultures already exhibiting damage from CPE-induced cytotoxicity.
Collectively, these findings suggest CPE-induced TJ effects in
confluent monolayers of polarized cells are a secondary consequence of
cytotoxicity, which is fully consistent with previous animal model
studies reporting that histopathologic damage (caused by CPE-induced
cytotoxicity) is required to obtain fluid and electrolyte secretion in
the CPE-treated intestines (38-40).
Given this new evidence for cytotoxicity as the initiator of the
intestinal effects of CPEs, the single most important contribution of
our study to understanding CPE action may be to establish a specific
linkage between the ~155-kDa complex and CPE-induced cytotoxicity.
Previous studies had shown that formation of some larger CPE complex
material is necessary to obtain the CPE-induced cytotoxic response (11,
22, 24, 25), but it had been unclear which of the three larger CPE
complex species is required for cytotoxicity. By demonstrating strong
86Rb release from Transwell® cultures of CaCo-2 or Vero
cells containing only the ~155-kDa CPE complex, our current results
now indicate that ~155-kDa CPE complex formation is sufficient to
obtain a cytotoxic response.
Interestingly, the current results also indicate that ~155-kDa
complex levels do not perfectly correlate with cellular CPE sensitivity. For example, Transwell® Vero cell cultures treated with
CPE on their basal surface contained less ~155-kDa complex but
released more 86Rb label than similarly CPE-treated
Transwell® CaCo-2 cell cultures. Since nearly all CPE present in those
Transwell® cultures was localized in the ~155-kDa complex, those CPE
sensitivity differences cannot be attributed to the formation of
different CPE larger complex species by Vero versus CaCo-2
cells. At least two factors might explain these sensitivity
differences, as follows: (i) the ~155-kDa CPE complex of Vero cells
might be more efficient than the ~155-kDa CPE complex of CaCo-2 cells
at inducing small molecule permeability alterations, or (ii) formation
of the ~155-kDa complex is necessary for obtaining CPE-induced
cytotoxicity, but additional factors also influence cellular
responsiveness to CPE treatment. Factors potentially affecting cellular
CPE sensitivity might include plasma membrane repair capacity or the
presence (or absence) of membrane transporters, which could
counterbalance CPE-induced small molecule loss and thus help a cell
resist CPE-induced death.
Whether CPE was added to the top or bottom Transwell® chamber, the
current study found nearly all 86Rb released by CPE from
confluent polarized cells went into the same Transwell® chamber
that contained CPE. That finding provides the first clear evidence that
both the apical and basolateral membranes of polarized mammalian cells
can respond to the cytotoxic action of CPE. It is also consistent with
CPE-induced membrane permeability alterations in polarized cells being
restricted to plasma membrane regions in direct contact with CPE, which
supports previous proposals (41, 42) that CPE-induced plasma membrane permeability alterations result from pore formation. In contrast, if
CPE were inducing plasma membrane permeability alterations via global
signal transduction cascades or activation of endogenous enzymes,
86Rb should have been released from both surfaces when
polarized cells were CPE-treated on a single cell surface. Assuming a
pore does cause CPE-induced membrane permeability alterations, our Fig.
5 results identify the ~155-kDa CPE complex, whose presence in
confluent Vero cells or CaCo-2 cells is sufficient to obtain 86Rb release, as a likely candidate for that pore.
In contrast to confluent CaCo-2 cell monolayers, CPE treatment of
confluent Vero cells resulted in substantial 86Rb release
into both Transwell® chambers, regardless of which Vero cell surface
had been CPE-treated. Combining that observation with Fig.
4B results indicating that Vero cells (unlike CaCo-2 cells)
are more sensitive when CPE-challenged on their apical surface, it
seems apparent that both differences and similarities (e.g.
formation of similar sized larger CPE complexes) exist between the
action of CPE on Vero cells versus CaCo-2 cells.
A possible explanation for why Transwell® cultures of CaCo-2, but not
Vero cells, exhibit strong unidirectional 86Rb release is
provided by our Figs. 8 and 9 results, which clearly demonstrate the
presence of TJs in confluent monolayers of CaCo-2 cells but not Vero
cells. The likelihood of TJs affecting CPE responsiveness receives
further support from the similar 86Rb release patterns
observed for both CPE-treated Transwell® monolayers of CaCo-2 cells
and MDCK cells, another polarized cell known to form TJs (10).
Both the barrier and fence functions (43) of TJs may affect the
responsiveness of polarized cells to CPE treatment. The barrier
function of TJs may help explain the unidirectional release of
86Rb observed in CPE-treated confluent Transwell® cultures
of CaCo-2 cells or MDCK cells, i.e. until cellular damage
develops, the presence of TJs could prevent released radiolabel from
moving across CPE-treated polarized cell monolayers. Although Vero cell monolayers exhibit a partial permeability barrier (as evident from the
inability of CPE to pass across Vero cell monolayers in Figs. 4,
A and B, and 5), they do not form complete TJs
(Fig. 7A). Thus, it is possible that CPE treatment of one
surface of confluent Transwell® Vero cell cultures allows, as observed
in Fig. 4A, substantial amounts of 86Rb to pass
across intact Vero cell monolayers into the opposite Transwell®
chamber. The fence function of TJs, which separates the apical and
basolateral plasma membranes, may also influence cellular CPE
responsiveness. For example, if the ~155-kDa CPE complex is a pore,
TJs might inhibit diffusion of that complex from the CPE-challenged
apical CaCo-2 cell surface to their basolateral membrane, thus
preventing permeability alterations from developing on the basolateral
surface. In contrast, the absence of TJs in Vero cells may permit some
diffusion of the ~155-kDa complex to untreated plasma membrane
surfaces, causing those surfaces to exhibit membrane permeability
alterations. These possibilities, fully consistent with our Fig.
4A observations, are currently being explored.
The greater sensitivity of confluent Transwell® CaCo-2 cell cultures
when CPE-challenged on their basal versus apical surface is
attributable, at least in part, to more CPE binding to the basal
(versus apical) surface of CaCo-2 cells. The precise
identity of CPE receptors in intestinal cells (or their derivatives,
such as CaCo-2 cells) remains unknown, but previous studies (18-20) with other cell types demonstrated that certain claudins, a family of
TJ proteins, can serve as CPE receptors to convey cytotoxicity. Furthermore, the presence of CPE-binding claudins has been demonstrated on the basolateral surface of rat small intestinal and colonic cells
(44). Therefore, studies are now underway to investigate whether CaCo-2
cells also produce claudin CPE receptors that are preferentially
distributed on their basolateral surface.
While establishing that the basal surface of CaCo-2 cells is more
CPE-sensitive, our studies also demonstrate that the apical surface of
confluent CaCo-2 cell monolayers responds to CPE. That observation
confirms the usefulness of CaCo-2 cells as an in vitro model
for CPE-mediated GI disease, a situation where CPE is initially released in the intestinal lumen and then encounters the apical surface
of enterocytes (4). An unresolved question is whether the CPE receptors
present on the apical and basolateral surfaces of CaCo-2 cells (or Vero
cells) are the same protein.
Since our experiments with Transwell® CaCo-2 cell cultures mimic the
intestinal epithelium by utilizing highly confluent polarized intestinal cells that produce TJs, a new model can be proposed to
explain CPE action during GI disease. This model envisions CPE as a
bifunctional toxin that, in vivo, interacts initially with
the apical surface of enterocytes (4). That interaction leads to the
primary, cytotoxic effect of CPE, which may result from pore formation
involving the ~155-kDa CPE complex. The resultant cellular damage
provides CPE access to previously hidden TJ proteins, leading to the
second action of CPE, i.e. removal of selected TJ proteins
(including occludin and some claudins) from intestinal TJs. That
effect, perhaps involving endocytosis of the ~200-kDa complex,
induces TJ structural damage and, as observed in other systems (10,
12), paracellular permeability alterations that contribute to the
intestinal secretion of fluids and electrolytes that manifest as
diarrhea during CPE-mediated GI disease. The intestinal damage caused
by CPE-induced cytotoxicity may also allow CPE to access the
basolateral surfaces of other enterocytes. Since CaCo-2 cells are more
sensitive when CPE-treated on their basal surface, such interactions
between CPE and basolateral membranes could enhance cytotoxicity and
intestinal histopathologic damage and thereby further contribute to GI
disease. Thus this model is not only consistent with animal model
studies showing that intestinal histopathologic damage is required for
CPE to induce fluid and electrolytes losses from the intestines
(38-40) but also explains why frank intestinal secretion follows the
onset of histopathologic damage in the CPE-treated intestine (38-40,
45, 46).
Finally, the current study provides an interesting observation
regarding TJ formation and structure, i.e. Vero cells
express significant amounts of occludin but do not to form TJs. Since Vero cells produce claudin mRNA (18, 19), they presumably also
express claudins. Assuming Vero cells produce both occludin and claudin
yet do not form TJs, it would appear that expression of occludin and
claudin (two important structural proteins of TJs) is not necessarily
sufficient for TJ assembly.
,
, and
TOP
ABSTRACT
INTRODUCTION
