Clostridium difficile Toxin B Induces Apoptosis in Intestinal Cultured Cells

Abstract

Toxigenic strains of the anaerobic bacterium Clostridium difficile produce at least two large, single-chain protein exotoxins involved in the pathogenesis of antibiotic-associated diarrhea and colitis. Toxin A (CdA) is a cytotoxic enterotoxin, while toxin B (CdB) is a more potent cytotoxin lacking enterotoxic activity. This study dealt with CdB, providing the first evidence that intestinal cells exposed to this toxin exhibit typical features of apoptosis in that a significant proportion of the treated cells displayed nuclear fragmentation and chromatin condensation. In keeping with ultrastructural data, CdB-treated cells showed the typical flow cytometric hallmark of apoptosis consisting of a distinct sub-G₁ peak. The CdB-induced apoptotic response was dose and time dependent and not simply due to the actin-disrupting effect of the toxin or to the subsequent impairment of cell anchorage. Rather, the inhibition of proteins belonging to the Rho family due to CdB seems to play a role in the induction of apoptosis in intestinal cells. The origin of cells and the growth rate may also be cofactors relevant to such a response.

Toxigenic strains of the anaerobic bacterium Clostridium difficile produce at least two large, single-chain protein exotoxins involved in the pathogenesis of antibiotic-associated diarrhea and pseudomembranous colitis. Toxin A (CdA) is a cytotoxic enterotoxin, while toxin B (CdB) is a more potent cytotoxin lacking enterotoxic activity (22, 33). Both have to be internalized into cells by endocytosis (33) to exert their potent cytotoxicity, which results in vitro from the ability to induce disaggregation of the actin cytoskeleton, leading to rounding up (6). It has recently been reported that CdA and CdB are monoglucosyltransferases which catalyze the incorporation of glucose into Thr-37 of RhoA (15, 16), a small GTP-binding protein of the Rho family, which is involved in the regulation of actin assembly (13). Three subfamilies belong to the Rho family: the Rho subfamily, which induces the formation of actin stress fibers; Rac, which controls membrane ruffling but also the NADPH-oxidase activity in neutrophils; and Cdc42, which regulates the formation of F-actin filaments in filopodia. All three subfamilies are monoglucosylated by C. difficile toxins. This modification renders the Rho family inactive, and Rho loses its ability to induce the polymerization of actin filaments, thus provoking cell retraction and rounding.

In this study, we found the first evidence that cultured intestinal cells exposed to CdB exhibit the typical morphological features and flow cytometric hallmarks of apoptosis. Apoptosis is a physiological form of cell death which plays an important role in tissue development and homeostasis, maintaining a correct cell number in the body by balancing cell growth and death (21). Its hallmarks are distinct morphological alterations (different from those which characterize necrosis) such as nuclear condensation and fragmentation, cell shrinkage, and the absence of inflammation (18). Apoptosis is a multiphase process characterized by (i) an initiation phase, in which cells receive the death stimulus; (ii) an effector phase, in which several reactions triggering cell death occur; and (iii) a cell degradation process, in which irreversible morphological and molecular markers of apoptosis become evident. The biochemical machinery responsible for apoptotic cell death appears to be constitutively expressed in most, if not all, cells and can be triggered by a variety of external or internal signals (4, 21, 31).

In cultured cells growing in a monolayer, apoptosis may be triggered by inhibition of cell adhesion and of anchorage-dependent cell spreading (32). Interestingly, CdB is able to diminish the anchorage of cells to the substrate and to impair the spreading. The apoptotic response to CdB, however, was not only due to the decreased cell adhesion subsequent to actin depolymerization. Our data indicate that, besides the origin of the cell type used, regulatory proteins of the Rho family may play a pivotal role in the induction of apoptosis.

MATERIALS AND METHODS

Cell lines.

IEC-6 (normal rat small intestine; ATCC CRL 1592), Int407 (human embryonic intestine; ATCC CCL 6), HT-29 (human colon adenocarcinoma; ATCC HTB 38), and A431 (human epidermoid carcinoma, ATCC CRL 1555) cells were cultured at 37°C in the appropriate medium supplemented with 10% fetal calf serum (Flow Laboratories, Irvine, United Kingdom), 1% nonessential amino acids, 5 mM l-glutamine, penicillin (100 U/ml), and streptomycin (100 μg/ml). The media used consisted of (i) Dulbecco’s modified Eagle’s medium plus insulin at 10 μg/ml for IEC-6 cells, (ii) Dulbecco’s modified Eagle’s medium for A431 cells, (iii) RPMI medium for HT-29 cells, and (iv) basal Eagle medium for Int407 cells.

Toxins and chemicals.

CdB was kindly provided by Christoph von Eichel-Streiber, Johannes Gutenberg-Universitat, Mainz, Germany, and purified as previously described (35). Horse anti-C. difficile toxin B immunoglobulin G (IgG) was a generous gift from Ingo Just, Institut für Pharmakologie und Toxikologie, Albert-Ludwigs-Universität, Freiburg, Germany. C. spiroforme iotalike toxin (29), chimeric toxin C3B (1), and LT from C. sordellii (28) were kindly provided by P. Boquet, University of Nice, Nice, France. Cytochalasins B and D, cycloheximide (CHX), actinomycin D (AcD), propidium iodide (PI), and tetramethyl rhodamine isothiocyanate-conjugated anti-mouse IgG were from Sigma Chemical Co., St. Louis, Mo.

Cell treatments.

Twenty-four hours after being seeded on glass coverslips in 24-well plates (initial inoculum, 4 × 10⁴ cells/ml), cells were treated with CdB, which was added directly to the culture medium, for 6, 18, 48, 72, and 96 h at 37°C. The concentrations used for CdB ranged from 0.18 to 192 ng/ml (twofold dilutions). For all experiments, we used 3 ng of the toxin per ml because this is the lowest concentration that induces apoptosis in 15 to 20% of IEC-6 cells within 18 h. Denaturation of CdB was obtained by treating the toxin at 95°C for 15 min before addition to the cells. To block the toxin activity, horse anti-CdB IgG (1.5 mg/ml) was incubated with an equal volume of CdB (0.2 μg/ml) for 20 min on ice. The mixture was then incubated with cells as described above. The antibody alone was used as a control. For the other toxins, the concentrations used were 20 μg/ml for C. spiroforme iotalike toxin, 1 μg/ml for C. sordelli LT, and 10⁻⁹ M for C3B. The doses used for CHX, an inhibitor of protein synthesis, and AcD, which inhibits mRNA, were 50 and 0.1 μg/ml, respectively. The concentrations of cytochalasins B and D used were 5 and 2 μg/ml, respectively.

Fluorescence microscopy.

Cells were fixed with 3.7% formaldehyde in phosphate-buffered saline (PBS; pH 7.4) for 10 min at room temperature. After being washed in the same buffer, the cells were permeabilized with 0.5% Triton X-100 (Sigma) in PBS (pH 7.4) for 10 min at room temperature. Cells were stained with Hoechst 33258 (Sigma Chemical Co.). After 30 min at 37°C, cells were washed and coverslips were mounted with glycerol-PBS (2:1) and analyzed with a Nikon Optiphot fluorescence microscope.

SEM.

Cells were fixed with 2.5% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.4) at room temperature for 20 min. Following postfixation in 1% OsO₄ for 30 min, cells were dehydrated through graded ethanol solutions. For scanning electron microscopy (SEM), cells were critical point dried in CO₂ and gold coated by sputtering, and the samples were examined with a Cambridge 360 scanning electron microscope. For transmission electron microscopy, cells were embedded with Agar 100 and the samples were examined with a Zeiss 10C transmission electron microscope.

Flow cytometry.

For experiments aimed to investigate the cell cycle and apoptotic features of CdB-treated cells, plasma membrane and cytoplasmic proteins were removed by exposing cells to a nuclear isolation medium prepared by adding 0.5% Nonidet P-40 (Sigma) and 1 mM EDTA to Ca²⁺- and Mg²⁺-free PBS. PI (40 mg/ml) was the DNA stain. DNA analysis was performed by acquiring at least 15,000 events with the Lysis II software (Becton Dickinson) and a doublet exclusion gate so that analysis was performed on single nuclei and a potential source of artifacts was eliminated (5). Analysis of DNA content was performed by using both logarithmic and linear scales. Irrespectively of the scale used, the apoptotic values obtained did not differ significantly. However, we show only results of the analysis performed with the logarithmic scale as described by Darzynkiewicz and coworkers (3), who proposed that this type of presentation is the most suitable for hypotonic-buffer-treated (lysed) cells.

Statistical analysis.

The values in Fig. 1, 3, and 4 are the means ± the standard deviations from four separate experiments. The Student t test was used for analysis of statistical significance. A P value of less than 0.05 was considered significant.

FIG. 1.

CdB induces morphological changes typical of apoptosis in IEC-6 cells. (a and b) SEM of control cells (a) and cells treated with CdB for 18 h (b). Arrows indicate cells displaying surface blebbing. (c to e) Fluorescence microscopy of IEC-6 cells after staining with Hoechst 33258. c, control cells; d and e, cells exposed to CdB for 18 h. Cells with chromatin fragmentation and/or condensation are indicated by the arrow and the arrowhead, respectively. Bars: a and b, 30 μm; c to e, 10 μm. (f) Percentage of apoptotic cells upon exposure to CdB as detected by Hoechst staining. Induction of apoptosis increased with the time of treatment and was also dose dependent.

FIG. 3.

(a) Graph showing the relationship between CdB-induced cell rounding (as viewed by phase-contrast microscopy) and apoptosis (as detected by Hoechst staining). (b) Percentage of apoptosis in IEC-6 cells pretreated with both CHX and AcD before exposure to CdB or CD for 18 h. Pretreatment with CdB caused a significant increase in the percentage of apoptotic cells. Control cells were exposed to medium only.

FIG. 4.

(a) Percentages of apoptosis in different cell lines exposed to CdB. A higher percentage of intestinal cell cultures derived from normal tissue (Int407 and IEC-6) undergo apoptosis than do other cell lines of both intestinal and nonintestinal origins. Compared to subconfluent monolayers, cells maintained in a confluent state for several days showed an increase in the percentage of apoptotic cells in response to CdB. (b to d) Percentages of A431 (b), HT29 (c), and Int407 cells undergoing (d) apoptosis upon exposure to CdB as detected by Hoechst staining. Only very high doses and prolonged exposure to CdB caused apoptosis in more than 10% of the cells.

RESULTS

CdB induces morphological changes typical of apoptosis in IEC-6 cells.

The most striking morphological changes caused by CdB in monolayers of mammalian cells are retraction and rounding up of the cell body (6). Such alterations are known to be the consequence of actin derangement due to the monoglucosylation of Rho proteins (15). All of the cell lines tested so far (except for a mutant cell line [see reference 2]) undergo the same modification, including IEC-6 cells. When observed by SEM (Fig. 1a and b), control IEC-6 cells formed a monolayer adhering well to the substrate (Fig. 1a). Exposure to 3 ng of CdB per ml for 18 h caused retraction and rounding up in the whole cell population, with some cells (about 10%) displaying surface blebs (Fig. 1b). Under fluorescence microscopy (Fig. 1 c to e), monolayers of IEC-6 cells stained with Hoechst 33258 showed roundish and regular nuclei (Fig. 1c), whereas after treatment with CdB for 18 h (Fig. 1d and e), a percentage of cells presented chromatin fragmentation and/or condensation. These changes in nuclear morphology are typical of cells undergoing apoptosis (4).

In adhering cells, the percentage of apoptosis was clearly time and dose dependent (Fig. 1f). In fact, the higher percentage of apoptotic cells was reached at 48 h of exposure to the toxin and the higher percentage of cells displaying nuclear condensation and/or fragmentation appeared after treatment with 48 ng of toxin per ml. Augmenting the concentrations of CdB significantly reduced the apoptotic effect (8% with 192 ng of CdB per ml), probably because of the toxic disruption of the cells. In Fig. 1f, only the results obtained with some concentrations of CdB are reported. The trend, however, was the same with all of the doses tested. When detached cells were considered, only a small increase in the percentage of apoptosis was observed, whereas after starvation or growth at confluence for several days, the number of cells displaying morphological features of apoptosis consistently increased (data not shown). No obvious correlation was found between surface blebbing and nuclear changes; blebbing cells displayed both normal and apoptotic nuclei.

All of the effects described above were absent when the toxin was heat inactivated or when a polyclonal antibody against C. difficile toxin B was incubated with the toxin before its addition to cells (Table 1).

TABLE 1.

Prevention of CdB-induced apoptosis in IEC-6 cells by heat treatment or anti-CdB antibodies

Treatment(s)a	Mean % ± SEM of cells showing:
	Rounding	Apoptosis
Control	1.5 ± 0.5	1.5 ± 0.4
CdBb	98 ± 2	17 ± 3.2
CdB + 95°C	2 ± 0.6	1 ± 0.2
Anti-CdB	0.5 ± 0.1	0
CdB + anti-CdB	1.5 ± 0.3	0.5 ± 0.1

^aEach treatment was administered for 18 h. 

^bThe concentration of CdB used was 3 ng/ml. 

CdB-treated IEC-6 cells reveal typical flow cytometric hallmarks of apoptosis.

It has been reported that a typical flow cytometric hallmark of apoptosis is the appearance of a distinct sub-G₁ hypodiploid peak. The localization of dying cells in the sub-G₁ peak is due mainly to the reduced amount of DNA per cell (5). PI-generated fluorescence allowed the identification and quantification of cells in the diverse phases of the cell cycle, as well as detection of the hypodiploid peak (5). When IEC-6 cells were cultured in the presence of CdB, the appearance of a distinct hypodiploid peak was observed (Fig. 2). In accordance with the morphological data, the proportion of apoptotic cells increased with time of exposure to the toxin (Fig. 2b and c), being about 28 and 33% upon 18 and 48 h of exposure to CdB, respectively. The higher percentage of apoptosis obtained with this method with respect to direct fluorescence observation is due to the presence of cell debris, which is generally included in the hypodiploid peak.

FIG. 2.

CdB induces flow cytometric hallmarks of apoptosis in IEC-6 cells. Histograms derived from analysis of IEC-6 cells exposed to medium alone for 18 (a) or 48 (c) h or treated with CdB for 18 (b) or 48 (d) h. Nuclear PI fluorescence area was measured on a log scale. CdB induced the appearance of a distinct hypodiploid peak (arrows) which increased with time. The results of one experiment representative of four are shown.

Role of cell rounding and protein synthesis in the apoptotic response of IEC-6 cells to CdB.

The first microscopically visible sign of CdB intoxication is rounding up. This change already occurs rapidly after 1 to 2 h or even earlier, depending on the dose. As shown in Fig. 3a, the rounding was not noticeably linked to apoptosis, which was detectable within 6 h in only 3 to 6% of the cells. Thus, the apoptotic response clearly follows the rounding up and neither rounding up nor cell detachment is sufficient to trigger cell death.

When IEC-6 cells were pretreated with an inhibitor of protein or mRNA synthesis (CHX or AcD, respectively), subsequent overnight incubation with 3 ng of CdB per ml provoked an increase in the percentage of apoptotic cells (Fig. 3b). This result was obtained irrespectively of the confluent state of the cells. Thus, the induction of apoptosis in CdB-treated cells was apparently prevented by newly synthesized proteins, whose lack (upon challenge of cells with protein synthesis inhibitors) triggers apoptosis in cultures treated with CdB.

Involvement of the small GTPases of the Rho family in CdB-induced apoptosis.

Once inside the cells, CdB first modifies all proteins belonging to the Rho family (15), triggering the observed changes in the actin network which finally induce cell rounding and cell detachment. We have thus investigated the events that happen before rounding up, checking whether the alteration of the actin cytoskeleton induced by CdB via Rho was somehow responsible for the apoptotic effect. IEC-6 cells exposed to different toxic agents known to directly or indirectly modify actin (Table 2) were analyzed by Hoechst staining. This study was performed by using, in addition to cytochalasins, large clostridial cytotoxins, whose mode of action is detailed in reference 34. When IEC-6 cells were treated for 18 h with direct actin-disrupting agents, such as cytochalasins B and D and C. spiroforme iotalike toxin, apoptotic features were poorly, if at all, detectable. Overnight exposure of cells to toxins known to disrupt actin by inactivating its regulatory proteins, such as LT from C. sordellii and C3 from C. botulinum, caused apoptosis like that in cells challenged with CdB, although a difference was detectable in this group of toxins (Table 2). Thus, not impairment of actin assembly but rather inactivation of proteins belonging to the Rho family may be involved in promoting the apoptotic response.

TABLE 2.

Actin-disrupting toxins and apoptosis in IEC-6 cells

Toxin	Mode of action	Mean % of cells apoptotic ± SEM
C. sordellii LT	Glucosylation of Rac, Ras, Rap	24 ± 0.7
C. botulinum C3	ADP-ribosylation of Rho	12 ± 2.6
C. spiroforme iotalike toxin	ADP-ribosylation of G actin	8 ± 2.2
Cytochalasin B	Direct actin depolymerization	2 ± 0.8
Cytochalasin D	Direct actin depolymerization	2 ± 0.5

CdB-induced apoptosis is influenced by the cell type and the confluent state of the monolayer.

Although all of the cell lines tested, which had different origins and growth characteristics, underwent cell retraction and rounding upon exposure to CdB, apoptosis was not a general response to the toxin. In fact, the percentage of cells undergoing this type of cell death differed, depending on the confluence of the monolayer and the cell type. In particular, when sparsely growing cells were considered, only a significant percentage of IEC-6 cells underwent apoptosis (Fig. 4a). On the other hand, maintenance of the cells in a confluent state for several days provoked a general increase in the number of apoptotic cells which was significant, however, only in intestinal cell cultures derived from normal tissue (namely, IEC-6 and Int-407 cells, of which the latter were derived from human embryonal cells). Only very high doses and prolonged exposure to CdB could cause apoptosis in more than 10% of A431, HT-29, and Int407 cells growing in subconfluence (Fig. 4b to d).

DISCUSSION

In this report, we have shown for the first time that CdB can act as an apoptosis inducer in intestinal crypt cells. The blebbing phenomenon, often reported as a typical morphological marker of this type of cell death and previously described as one of the responses of HEp-2 cells to CdB (20), does not reflect apoptosis in IEC-6 cells. The relatively low percentage of CdB-treated cells with recognizable features of apoptosis probably reflects the fact that, given the brevity of the execution phase compared with the condemnation phase, only 20 to 40% of in vitro cultured cells are actually in the execution phase at the peak of the apoptotic process in the best experimental system (4).

Epithelial cells may undergo apoptosis when detached from the substrate (32), whereas adhesion to the extracellular matrix seems to trigger the entry of cells onto the survival pathway (23). A percentage of CdB-treated IEC-6 cells still adhering to the substrate clearly entered apoptosis, and interestingly, detachment per se was not enough to cause apoptosis in the whole floating population. Moreover, apoptosis caused by CdB cannot be a direct consequence of cytoskeletal changes, since a toxin known to directly disrupt actin organization, such as a cytochalasin or C. spiroforme iotalike toxin (29), was unable or only poorly able, respectively, to cause apoptosis.

On the other hand, our data showing that bacterial toxins inhibiting the activity of proteins of the Rho family may induce apoptosis in intestinal epithelial cells are in agreement with reports in the literature containing evidence indicating how these regulatory G proteins may control apoptosis (12, 24, 25). Accordingly, it has very recently been reported that Rho plays a selective role in early thymic development as a critical determinant of proliferation and cell survival signals and that inactivation of Rho by C. botulinum C3 leads to apoptosis (14). Consistent with these findings is also our recent observation that Escherichia coli cytotoxic necrotizing factor 1, a bacterial toxin which activates Rho (8, 11), can protect epithelial cells from UVB-induced apoptosis (9). Interestingly, activation of Rho by cytotoxic necrotizing factor 1 increases the expression of antiapoptotic proteins of the Bcl-2 family (10). However, we also have to stress that the Rho-inhibiting toxins we used did not possess the same ability to induce apoptosis. CdB, which acts on the Rho, Rac, and Cdc42 subfamilies, appeared to be as potent as C3 (which is specific for the Rho subfamily) but maybe less potent than LT (which interacts with Rac but also with Ras) in promoting cell death. Of course, we cannot conclude by extrapolating from our data whether a different role is played by the three subfamilies in controlling the apoptotic pathway, although this might be a subject of further investigation.

In addition to the activity on Rho, other factors seem to play a role in the CdB-induced apoptotic response in intestinal cells which is apparently prevented by new synthesized proteins, as shown upon challenge of cells with protein synthesis inhibitors. By starving cells or maintaining them in a confluent state for several days, it was possible to increase the amount of apoptosis. This was probably due to the reported antiproliferative effect of the toxin (33) and to the slowing down of the passage of the cells through G₁ which follows the inactivation of Rho (26). Together with the confluence state of the monolayer, the origin of cells appeared to be significant for the apoptotic response. Intestinal cells seem to be the most suitable cell model for investigation of apoptosis induced by CdB, in particular, cells derived from the small intestine. In vivo, in fact, apoptosis occurs spontaneously more easily in the small intestine than in the colon, probably because of the absence of Bcl-2 (an antiapoptotic gene product) in the former (30). Accordingly, CdA was also reported to cause apoptosis in IEC-6 cells (7), as well as in other intestinal cells, such as Caco-2 and HT29 (19). However, in the latter case, apoptosis was detected only in cells which had already lost their anchorage to the substrate.

The importance of apoptosis in the pathogenicity of different infectious diseases is being increasingly recognized (27). Some bacterial pathogens have evolved ways to induce eukaryotic cell death, and protein toxins are among the most potent weapons they use. Up to now, only a few bacterial protein toxins have been described as apoptosis inducers (36), most of them in macrophages (17, 37). The effects induced by CdB in intestinal crypt cells, as well as those reported for CdA in polarized intestinal cells (19), appear to be relevant since the gut is the actual target of the toxins in experimental animals (33), although this does not prove that apoptosis plays a significant role in the pathogenicity of at least toxin B in vivo. We can only speculate that, as reported for CdA, intestinal cells undergoing apoptosis upon exposure to CdB can liberate cytokines capable of mediating an inflammatory process (19). In addition, the actin-disrupting effect of CdB could somehow impair the migration of macrophages, which could become unable to phagocytose apoptotic cells before they lyse. Experiments that address these questions are in progress.

In conclusion, our results show that CdB can act as an apoptosis inducer in intestinal cells, adding new evidence about the role ascribed to the Rho protein in directing cells toward survival or death.