|Home | About | Journals | Submit | Contact Us | Français|
Colonic inflammation in Clostridium difficile infection is mediated by released toxins A and B. We investigated responses to C. difficile toxins A and B by isolated primary human colonic myofibroblasts, which represent a distinct subpopulation of mucosal cells that are normally located below the intestinal epithelium. Following incubation with either purified toxin A or B, there was a change in myofibroblast morphology to stellate cells with processes that were immunoreactive for alpha-smooth muscle actin. Most of the myofibroblasts remained viable, with persistence of stellate morphology, despite exposure to high concentrations (up to 10 μg/ml) of toxin A for 72 h. In contrast, a majority of the toxin B-exposed myofibroblasts lost their processes prior to cell death over 24 to 72 h. At low concentrations, toxin A provided protection against toxin B-induced cell death. Within 4 h, myofibroblasts exposed to either toxin A or toxin B lost expression of the nonglucosylated form of Rac1, and there was also a loss of the active form of RhoA. Despite preexposure to high concentrations of toxin A for 3 h, colonic myofibroblasts were able to recover their morphology and proliferative capacity during prolonged culture in medium. However, toxin B-preexposed myofibroblasts were not able to recover. In conclusion, primary human colonic mucosal myofibroblasts are resistant to toxin A (but not toxin B)-induced cell death. Responses by colonic myofibroblasts may play an important role in mucosal protection, repair, and regeneration in colitis due to C. difficile infection.
Clostridium difficile induces colonic inflammation via two released toxins, A and B, and histologically, the colonic disease is often characterized by focal areas of epithelial ulceration and inflammatory exudates (38, 46). The first host cells that the toxins interact with in the colon are likely to be surface epithelial cells, which normally provide an essential barrier against luminal microorganisms and their products. In intact monolayers of carcinoma-derived intestinal epithelial cell lines in vitro, apically applied toxin A induces injury and loss of barrier function (43, 45). Other early responses by intestinal epithelial cells exposed to toxin A include changes in the cytoskeleton (10, 15) and secretion of cytokines, such as transforming growth factor beta (TGF-β) (20) and interleukin-8 (5, 13, 26), before the cells undergo programmed cell death (6, 26). Our previous studies have shown that primary human intestinal epithelial cells, monocytes, and intestinal macrophages are more sensitive than lymphocytes to C. difficile toxin A-induced cell death (25, 26, 42).
In the intestinal mucosa, myofibroblasts represent a distinct subpopulation of cells that are located below the intestinal epithelium, separated by a porous basement membrane (37). Intestinal myofibroblasts have morphological and other features of smooth muscle cells and fibroblasts, as illustrated by strong expression of alpha-smooth muscle actin (a characteristic of smooth muscle cells) and vimentin (also expressed by fibroblasts). In addition to controlling the deposition of extracellular matrix (24, 29), intestinal myofibroblasts may regulate epithelial functions, such as electrolyte transport (3), restitution (28), and barrier function (3), and also functions of stem cells (18). In view of their capacity to secrete a number of cytokines and to interact with other cell populations (36), intestinal myofibroblasts are also believed to be key players in inflammatory responses in the intestine (1, 9, 12, 22). They have been shown to provide protection to intestinal epithelial monolayers against loss of barrier function in response to low concentrations of toxin A, via secretion of TGF-β (20). Following injury and loss of toxin-exposed epithelial cells in vivo, myofibroblasts, which are capable of migrating onto the surface of the basement membrane (24), would be expected to be exposed to C. difficile toxins. However, responses of primary human intestinal myofibroblasts to C. difficile toxins remain to be characterized.
A majority of the biological effects of C. difficile toxins A and B are believed to be mediated via the ability of their N termini to glucosylate and inactivate Rho GTPases such as RhoA, -B, and C, Rac1-3, RhoG, and Cdc42 (2, 19). Rho GTPases are key regulators of the actin cytoskeleton and influence many cellular processes, such as cell polarity, migration, vesicle trafficking, and cytokinesis (7, 14).
The aims of our studies were to characterize responses of primary human colonic myofibroblasts (isolated from healthy large intestinal mucosal samples) to purified C. difficile toxins A and B.
Toxin A was purified as previously described (21). C. difficile VPI strain 10463 was cultured anaerobically in brain heart infusion broth (Oxoid, United Kingdom) in dialysis culture flasks. After centrifugation, supernatant samples were loaded onto a bovine thyroglobulin affinity column, and eluted toxin A-containing fractions were then applied to Q Sepharose FF and Mono Q columns (GE Healthcare, Sweden). During the purification steps, toxin A-containing fractions were identified by dot blot analysis using the anti-toxin A antibody PCG-4 (23).
The purification protocol for toxin B was adapted from previous studies (30, 35, 44). A broth culture of C. difficile VPI strain 10463 was centrifuged to remove bacteria from the toxin-containing supernatant. The latter was made 70% saturated by the addition of solid ammonium sulfate. After 1 h, the precipitate was collected by centrifugation and resuspended in 50 mM Tris-HCl buffer (pH 7.4). The process of ammonium sulfate precipitation was repeated to make the solution 50% saturated. Following centrifugation, the pellet was resuspended in 50 mM Tris-HCl buffer (pH 7.4) and then dialyzed for 24 h in 50 mM Tris-HCl buffer (pH 7.4). The dialyzed sample was applied to a DEAE-Sepharose column (GE Healthcare, Sweden) and eluted initially with a linear NaCl gradient in 50 mM Tris-HCl buffer (0.05 to 0.25 M NaCl), followed by 150 ml of 50 mM Tris-HCl buffer (pH 7.4) containing 0.3 M NaCl. Subsequent elution was with a second linear gradient of NaCl (0.3 to 1 M) in 50 mM Tris-HCl buffer (pH 7.4). Toxin B-containing fractions were identified by dot blot analysis using anti-toxin B antibody (Biodesign International), and pooled fractions were dialyzed for 24 h against 10 mM Tris-HCl buffer (pH 8.0) containing 50 mM CaCl2 before application to a Mono Q anion-exchange column (GE Healthcare, Sweden). Purified toxin B was eluted from the Mono Q column by use of a linear NaCl gradient (0 to 1 M) in 10 mM Tris-HCl buffer (pH 8.0) containing 50 mM CaCl2.
Fresh, histologically normal colonic mucosal samples (>5 cm from cancer, with tissue surplus to clinical requirements) were obtained (after informed consent) from operation resection specimens. The median age of tissue donors was 69 years (range, 47 to 87 yrs; 6 females and 12 males). Ethical committee approval was provided by the Nottingham Research Ethics Committee (REC Q1020310).
Intestinal myofibroblasts were isolated and established in pure culture as previously described (24). In brief, surface epithelial cells were detached from mucosal samples by three sequential treatments with 1 mmol EDTA, and the tissue samples, denuded of epithelial cells, were cultured (at 37°C and 5% CO2) in 10% fetal calf serum (FCS)-RPMI (Gibco-Invitrogen, United Kingdom). During culture of mucosal samples devoid of epithelial cells, myofibroblasts migrate out of the subepithelial regions of the lamina propria via basement membrane pores to establish colonies in tissue culture dishes (24). After removal of the tissue samples, the myofibroblasts proliferate to establish monolayers that can be maintained over many passages. Our previous studies have shown that the isolated intestinal myofibroblasts retain their phenotype and functional characteristics over many passages (18, 24, 27).
Isolated colonic myofibroblasts, which expressed alpha-smooth muscle actin and vimentin (but not desmin), were studied at passages 3 to 7 and were maintained in culture in Dulbecco's modified Eagle's medium (Sigma) supplemented with 10% FCS and 1% nonessential amino acids (Gibco-Invitrogen, United Kingdom). Monolayers of the cells were exposed to various concentrations (0 to 10 μg/ml) of purified C. difficile toxin A or B for various intervals (3 h to 72 h). In some experiments, morphological recovery of myofibroblasts was studied after preexposure to toxin A or B (at a concentration of 1 μg/ml or 10 μg/ml) for 3 h to 48 h.
Morphological changes to the cells were studied by phase-contrast microscopy (changes assessed in 10 random high-power fields) and scanning electron microscopy. The nuclear morphologies of control and toxin-exposed myofibroblasts were studied by fluorescence microscopy after staining with Hoechst 33342 dye (42). Immunohistochemical studies were undertaken using monoclonal antibodies to alpha-smooth muscle actin, desmin, and vimentin (all from Sigma). Effects on cell viability were investigated using an assay for mitochondrial dehydrogenase activity. Changes to myofibroblast DNA were studied by flow cytometry following propidium iodide staining.
Control and toxin-exposed myofibroblasts cultured on glass coverslips were fixed in acetone, and immunohistochemical studies were undertaken using a Vectastain ABC peroxidase kit as previously described (24). In brief, the cells were incubated with antibodies to alpha-smooth muscle actin, desmin, and vimentin (all from Sigma) for 30 min, and after being washed, the myofibroblasts were exposed to biotinylated horse anti-mouse antibody, followed by avidin-biotinylated horseradish peroxidase complex. Peroxidase activity was developed with diaminobenzidine, followed by nuclear staining using hematoxylin. Controls included use of buffer instead of primary antibody and an irrelevant primary antibody (anti-cytokeratin).
Metabolism by mitochondrial dehydrogenase of the yellow tetrazolium salt 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) to the purple formazan reaction product can be quantified spectrophotometrically and can be used as an assay for cell proliferation or cell death (33). MTT assays were undertaken as previously described (26). In brief, cells (40 × 103 applied per well) were grown to confluence in 96-well tissue culture plates (Nunc), and the assays were performed in triplicate. After exposure of cells to toxin A or B (or control medium) for various periods, MTT (Sigma Chemical, St. Louis, MO) was added to each well (to a final concentration of 0.5 mg/ml), and incubation was continued for 4 h. The cells were then incubated overnight in solubilization solution (50% sodium dodecyl sulfate in 0.1 mmol/liter HCl). The spectrophotometric absorbance of the samples was subsequently measured with a microtiter enzyme-linked immunosorbent assay (ELISA) plate reader using a 570-nm filter.
Propidium iodide staining was performed as previously described (25, 34). Control and toxin-exposed myofibroblasts were detached using 0.1% (wt/vol) trypsin-0.2% (wt/vol) EDTA. Following centrifugation (400 × g for 10 min), the myofibroblasts were fixed and permeabilized with ice-cold 70% ethanol for 60 min. After being washed with phosphate-buffered saline (PBS), the cells were incubated with propidium iodide (100 μg/ml; Sigma) at room temperature in the dark for 15 min. Fluorescence emission of propidium iodide was analyzed with a Beckman Coulter Altra flow cytometer (Beckman Coulter) at 617 nm after excitation with a blue laser at 488 nm. Events that fell within the marked hypodiploid region (apoptotic events) were enumerated by the statistics function of WinMDI V2.8 software, and the results are expressed as percentages of total events.
Lysates of control and toxin A- or toxin B-exposed myofibroblasts were separated by SDS-polyacrylamide gel electrophoresis on a 12.5% acrylamide resolving gel (using Bio-Rad Protean II slab gel electrophoresis equipment). After transfer to polyvinylidene difluoride membranes, immunostaining was performed using a Vectastain Elite ABC kit (Vector Laboratories) and antibodies to beta-actin (Abcam Plc, United Kingdom), the nonglucosylated form of Rac1 (11) (clone 102; BD Transduction Laboratories), and RhoA (clone 26C4; Santa Cruz Biotechnology Inc.) (RhoA detection was not affected by glucosylation).
Changes in levels of the active, GTP-bound form of RhoA in myofibroblasts were studied using an ELISA-based Rho activation assay (RhoA G-LISA activation assay; Cytoskeleton Inc.). Lysates (in buffer containing protease inhibitor) were obtained from monolayers of isolated colonic myofibroblasts that had been exposed for 4 h to control medium (Dulbecco's modified Eagle's medium supplemented with 10% FCS and 1% nonessential amino acids), 1 μg/ml toxin A, or 1 μg/ml toxin B. The lysates were applied to wells coated with Rho-GTP-binding protein. In this assay, active, GTP-bound RhoA in cell lysates binds to the wells, while inactive, GDP-bound Rho is removed during washing steps. The bound active RhoA was detected using a RhoA-specific antibody, followed by a secondary antibody conjugated to horseradish peroxidase. The spectrophotometric absorbance was subsequently measured using a 490-nm filter.
Data are expressed as means with standard errors of the means (SEM) and were analyzed by one-way analysis of variance and an unpaired t test.
Morphological responses to different concentrations (0.1, 1, and 10 μg/ml) of purified toxins A and B were studied by phase-contrast microscopy. Monolayers of isolated intestinal myofibroblasts maintained in control medium spread to occupy a large surface area. In the presence of the highest concentration (10 μg/ml) of either toxin, approximately 50% of cells were rounded (with visible processes) at 2 h, and all cells were rounded at 4 h. When cells were exposed to lower concentrations of the toxins (1 and 0.1 μg/ml), myofibroblast cell rounding occurred a few hours later in toxin A-exposed cells than in those cultured with toxin B.
Myofibroblasts exposed to toxin A changed in morphology from flat cells to those with a stellate appearance. Over 72 h, the cell cytoplasm around the nucleus appeared to decrease further in size, with maintenance of prominent processes. Morphological changes in toxin B-exposed myofibroblasts were initially similar to those in toxin A-exposed cells, but at later time points (48 to 72 h), a majority of the myofibroblasts lost their processes and many appeared nonviable (Fig. 1E and J). However, some toxin B-exposed myofibroblasts remained viable at 72 h, with persistence of processes (Fig. (Fig.1J1J).
Scanning electron microscopy of myofibroblasts exposed to toxin A or toxin B (1 μg/ml) for 24 h showed that in addition to numerous processes of various thicknesses, distinct globular structures were seen either close to the cell body or at a distance and were associated with processes (Fig. (Fig.22).
Immunohistochemical studies (at 6 h in response to 1 μg/ml toxin A) showed that the processes in both toxin A- and toxin B-exposed myofibroblasts were strongly immunoreactive for alpha-smooth muscle actin (Fig. (Fig.3).3). At 48 h and 72 h, and as observed by phase-contrast microscopy (see above), a majority of the myofibroblasts exposed to toxin B (1 μg/ml) had lost processes (Fig. (Fig.3),3), but in toxin A-exposed (1 μg/ml toxin) cells, there was persistence of processes that were immunoreactive for alpha-smooth muscle actin. At 72 h, many toxin B-exposed cells appeared to have lost the integrity of their cell membranes, but this did not occur in toxin A-exposed cells.
In studies undertaken using myofibroblasts isolated from 8 different donors (cells from 3 to 5 donors were used for each experimental condition), morphological recovery of the cells was studied by phase-contrast microscopy after preexposure to toxin A for 3 h, 24 h, or 48 h. Following 3 h of preexposure to toxin A (cells were incubated with 1 and 10 μg/ml toxin A for 3 h, followed by washing and culture in medium only), the colonic myofibroblasts returned to their normal morphology over 28 to 63 days, with the monolayers becoming fully confluent. Following trypsinization and reculture (passage at a 1:3 split ratio), the 3-h toxin A-preexposed myofibroblasts proliferated to confluence. Following preexposure to 1 μg/ml toxin A for 24 h or 48 h, myofibroblast recovery was slower, but all the cells showed normal morphology after culture in medium for 28 days.
In contrast to the case with toxin A, none of the myofibroblast monolayers preexposed (for 3 h, 24 h, or 48 h) to toxin B (at 1 μg/ml) showed any signs of recovery. Thus, after culture in medium for 28 days, no viable myofibroblasts were seen in toxin B-preexposed monolayers.
MTT assays were undertaken to assess changes in myofibroblast viability in response to toxin A or B. In order to assess the ability of the assay to detect maximal loss of cell viability, monolayers of primary colonic myofibroblasts were cultured in medium for 24 h, in the absence or presence of 0.1% Triton X-100. Compared to the control, there was an 81.3% reduction in the mean absorbance value for the formazan reaction products of myofibroblasts exposed to Triton X-100 (0.353 [0.017] versus 0.066 [0.004]; P < 0.0001).
In studies conducted for up to 72 h, mitochondrial dehydrogenase activities of myofibroblasts exposed to 0.1, 1, and 10 μg/ml toxin A did not differ from those of cells cultured in control medium (Fig. (Fig.4a).4a). In contrast, myofibroblasts exposed to toxin B (especially those exposed to 1 and 10 μg/ml of the toxin) showed a significant loss of mitochondrial enzyme activity (Fig. (Fig.4b).4b). Moreover, the mean (SEM) absorbance values of the formazan reaction products of myofibroblasts exposed to 1 μg/ml and 10 μg/ml toxin B at 24 h (0.391 [0.028] and 0.405 [0.024], respectively), 48 h (0.323 [0.013] and 0.371 [0.021], respectively), and 72 h (0.204 [0.012] and 0.221 [0.007], respectively) were largely similar. This suggests that while the majority of myofibroblasts are susceptible to toxin B (at concentrations of ≥1 μg/ml), a proportion remain viable despite exposure to the highest concentration (10 μg/ml) of toxin B for up to 72 h.
In the presence of toxin A at a concentration of 0.1 μg/ml, myofibroblasts were protected against the effects of toxin B applied at the same concentration (absorbance value for myofibroblasts [expressed as % of value for control cells] after exposure to toxin B only for 72 h, 75.76% [5.83%]; that for cells after exposure to both toxins A and B for 72 h, 101.10% [6.17%]; P < 0.01 [n = 3]). When toxin B was present at a concentration of 1 μg/ml, absorbance values of myofibroblasts exposed to toxin B alone (for 72 h) were not significantly different from those of cells incubated with both toxins A and B (57.31% [2.33%] versus 66.25% [4.58%]).
Events in the hypodiploid region of cellular DNA profiles reflect DNA fragmentation due to apoptotic cell death (25, 34). In myofibroblasts (isolated from 3 donors) exposed to 1 μg/ml toxin B, there was a time-dependent increase in the proportion of events in the hypodiploid region (control versus toxin B-exposed cells at 8 h, 1.13% [0.14%] versus 1.29% [0.06%]; at 24 h, 1.44% [0.041%] versus 2.30% [0.24%] [P < 0.03]; at 48 h, 1.17% [0.21%] versus 13.38% [3.81%] [P < 0.04]; and at 72 h, 0.80% [0.019%] versus 35.15% [1.31%] [P < 0.0001]) (Fig. (Fig.5a).5a). In contrast, myofibroblasts exposed to 1 μg/ml toxin A showed only small (but statistically significant) increases in hypodiploid events at 48 h (for control versus toxin A-exposed cells, 0.58% [0.16%] versus 2.14% [0.29%]; P < 0.01) and 72 h (for control versus toxin A-exposed cells, 0.81% [0.029%] versus 4.22% [0.94%] P < 0.03) (Fig. (Fig.5b).5b). DNA fluorescence profiles of myofibroblasts exposed to 10 μg/ml toxin A showed that the majority of the cells showed normal DNA profiles (Fig. (Fig.5c)5c) that would be expected for viable cells.
In myofibroblast monolayers exposed to toxin B, Hoechst staining confirmed significant levels of apoptosis, as demonstrated by nuclear condensation (due to dense chromatin) and nuclear fragmentation (Fig. (Fig.66).
Western blot analysis using a specific antibody (11) showed loss of the nonglucosylated form of Rac1 from 4 h onwards in myofibroblasts cultured with either toxin A or toxin B (Fig. 7a and b). This implies glucosylation of intracellular Rac1 by both toxins.
In a Western blot analysis conducted over 24 h, levels of expression of RhoA protein (whose detection by the antibody used was not altered by glucosylation) did not change in myofibroblasts exposed to either toxin. However, an ELISA-based Rho activation assay showed marked reductions in levels of active, GTP-bound RhoA in lysates of myofibroblasts exposed to either toxin A or toxin B (Fig. (Fig.88).
Intestinal myofibroblasts represent a distinct subpopulation of cells located immediately below the single monolayer of epithelial cells. Primary colonic epithelial cells are highly susceptible to C. difficile toxin-induced cell death (26, 41). The loss of epithelial cells in vivo would be expected to expose myofibroblasts to C. difficile toxins. By using monolayers of myofibroblasts isolated from colonic mucosal samples from different donors, we showed that the cells consistently responded to toxins A and B by a change in morphology to stellate cells. Moreover, scanning electron microscopy showed small globular structures in cells incubated with the toxins, and further studies are required to characterize the nature of these structures. The cell body and processes in toxin-exposed myofibroblasts were strongly immunoreactive for alpha-smooth muscle actin. Although morphological studies and MTT assays suggested that a proportion of colonic myofibroblasts remained viable after 72 h of incubation with toxin B, all cells eventually underwent cell death, even after short (3 h) preexposure periods. However, it should also be noted that at low concentrations, the presence of toxin A provided protection against a toxin B-induced loss of viability. Such exposure of myofibroblasts to both toxins is likely to occur in the colonic mucosa in vivo.
In contrast to responses to toxin B, a majority of the toxin A-exposed myofibroblasts remained viable, even when exposed to a high (10 μg/ml) concentration of the toxin for up to 72 h. Interestingly, after preexposure (for 3 h to 48 h) to high concentrations (1 and 10 μg/ml) of toxin A (followed by washing and culture in medium only), the myofibroblast morphology gradually reverted to normal. Subsequently, the cells were also able to proliferate after passage. This remarkable characteristic of the myofibroblasts may have important implications for mucosal tissue repair and regeneration in patients with colitis due to C. difficile infection. Our previous studies showed that following 3 h of preexposure to <0.01 μg/ml toxin A, the T84 carcinoma-derived colonic epithelial cell line was able to recover its barrier function during subsequent culture in medium. However, this was not the case if the initial preexposure was to ≥0.01 μg/ml toxin A (20). Moreover, there was a significant loss of viability in T84 and Caco-2 cell monolayers exposed to ≥0.1 μg/ml toxin A (unpublished observations in studies performed using the same batch of toxin as that used in the current series of experiments, together with our previous studies ). Thus, primary human colonic myofibroblasts not only are much more resistant to cell death than these carcinoma-derived epithelial cell lines but also are able to recover after preexposure to high concentrations of toxin A.
Primary human colonic epithelial cells are more sensitive to toxin A-induced cell death than Caco-2 and T84 carcinoma-derived epithelial cell lines (26). Indeed, our previous studies have shown that among the normal human colonic mucosal cells, epithelial cells and macrophages (and also monocytes) are more sensitive than lymphocytes to toxin A-induced cell death (25, 26, 42). Although direct comparisons have not been made, our current studies suggest that colonic myofibroblasts may be more resistant to toxin A-induced cell death than the colonic lamina propria lymphocytes. After exposure to 1 μg/ml toxin A, a large proportion of colonic lamina propria T cells underwent programmed cell death, mainly from 72 h onwards (25). In contrast, most of the colonic myofibroblasts incubated with 10 μg/ml toxin A were viable at 72 h. Reasons for the differences in susceptibility of colonic mucosal cells to toxin A-induced cell death remain to be determined. One possibility is that specific cell types bind and internalize different amounts of toxin A. However, since there is rapid cell rounding and glucosylation of Rac1 (see below), a significant amount of toxin A is likely to be internalized by the myofibroblasts.
The persistence of stellate morphology and adherence to the culture dish are other distinct features of toxin A-exposed myofibroblasts compared with epithelial cells. Since similar appearances were not seen with toxin A-exposed NIH 3T3 cells (unpublished observations), the stellate morphology of myofibroblasts could be due to the persistence of alpha-smooth muscle actin. It is likely that the changes in the morphology of myofibroblasts incubated with toxin A or B are due to inactivation (following glucosylation) of the Rho subfamily of GTPases, which are major intracellular targets of the toxins (2, 19). Both toxins induced a loss of the nonglucosylated form of Rac1 in colonic myofibroblasts over similar periods (within 4 h). Together with largely similar rates of cell rounding, these studies imply that the uptake of both toxins is similar in myofibroblasts and that the greater sensitivity to cell death in response to toxin B is unlikely to be mediated via Rac1. Previous studies using rat basophilic leukemia (RBL) cells also suggested that glucosylation/inactivation of Rac1 is responsible for cell rounding (cytopathic effect) but not for cell death (cytotoxicity) (17). Studies of other cell types have shown that inactivation of RhoA is required for cell death (4, 16, 32). However, following 4 h of exposure to C. difficile toxin A or B, there was a marked reduction in levels of active, GTP-bound RhoA, implying toxin-induced RhoA inactivation in the colonic myofibroblasts. In contrast, total levels of RhoA did not change in myofibroblasts exposed to either toxin A or B for 24 h. This is in contrast to the case for RBL cells, where high concentrations of toxin B (which led to cell death) induced a loss of RhoA (17). Such a loss may be due to the fact that glucosylated RhoA is degraded more efficiently by the proteasome than the nonmodified form (8, 11), and further studies are required to determine why total levels of RhoA do not change in toxin-exposed myofibroblasts. Since there were marked reductions in levels of active, GTP-bound RhoA in response to both toxins, other factors are likely to be involved in the induction of cell death in myofibroblasts cultured with toxin B.
In addition to functions in the cytoskeleton and cell adhesion, Rho proteins have been found to regulate a wide range of cellular activities, such as cell polarity, endocytosis, vesicle trafficking, and differentiation (7). Peptide regulatory factors such as platelet-derived growth factor (PDGF) and epidermal growth factor (EGF) stimulate Rac, leading to Rho activation (39, 40). Inactivation of Rac and other members of the Rho subfamily may therefore have a significant impact on the normal functions of colonic myofibroblasts, which include regulation of epithelial barrier function (3, 20) and repair of epithelial wounds (created by loss of injured epithelial cells). For the latter, myofibroblasts have been shown, via their contractile properties, to reduce the size of the wound (31) and to also enhance restitution (28), a process by which epithelial cells at the wound edge migrate to cover the mucosal defect. Such protective responses by the myofibroblasts may explain, at least in part, the focal nature of mucosal inflammation in C. difficile-associated colitis. However, once cells are exposed to high concentrations of C. difficile toxins, impairment of myofibroblast functions may lead to severe inflammation, as seen in pseudomembranous colitis. The concentrations of C. difficile toxins used in our studies are likely to be representative of those seen in vivo. Thus, we have previously shown that although the median stool concentration of toxin A in patients with C. difficile infection was 4.3 ng/ml, the levels ranged from 0.6 ng/ml to 19 μg/ml (42).
This study was supported by the Medical Research Council.
We thank Jacqueline Webb for technical support.
Editor: B. A. McCormick
Published ahead of print on 18 January 2011.