Multiple effects of Escherichia coli Nissle 1917 on growth, biofilm formation, and inflammation cytokines profile of Clostridium perfringens type A strain CP4

Abstract

Clostridium perfringens is an important Gram-positive pathogen responsible for food poisoning, necrotic enteritis, gas gangrene, and even death. Escherichia coli Nissle 1917 (EcN) is a well-characterized probiotic strain with demonstrated benefits. In this study, we evaluated the effects of EcN on growth, toxin production, biofilm formation, and inflammatory cytokine responses of C. perfringens. In vitro co-culture experiments demonstrated that EcN inhibited growth, gas production, and toxin production (α-toxin and NetB) of C. perfringens in a dose-dependent manner. The growth inhibition effect was not observed when C. perfringens was incubated with EcN cell-free supernatants (CFS_E), suggesting that growth inhibition was caused by nutrition competition during co-incubation. In vitro studies demonstrated that pre-incubation with EcN did not inhibit C. perfringens attachment to Caco-2 cells, but did reduce C. perfringens total number, toxin production, and cytotoxicity after 24 h. The similar growth inhibition results were also observed during the formation of C. perfringens biofilm. Finally, pre-incubation of EcN with RAW264.7 cells significantly decreased the production of inflammatory cytokines caused by the introduction of C. perfringens. Our results indicate that EcN can inhibit many of the pathological effects of C. perfringens in vitro conditions.

New soldier: Presence of E. coli Nissle strain dramatically represses C. perfringens growth, toxins production, biofilm formation, as well as the release of inflammatory cytokines.

E. coli Nissle 1917, C. perfringens, inhibition, toxin, biofilm, cytokine

Introduction

Probiotics are live microorganisms which provide beneficial effects when ingested. Although the underlying mechanisms remain poorly understood, a number of studies have demonstrated that probiotics can efficiently inhibit the impact of pathogens in the gut either directly by growth competition or indirectly via production of inhibitory substances such as bacteriocins (Sanders, 2011). Typical probiotics including lactic acid bacteria (LAB), bifidobacteria, certain yeasts, and bacilli have been well studied for decades to treat antibiotic-associated diarrhea (D'Souza et al., 2002), lactose intolerance (Sanders, 2000), and colon cancer (Brady et al., 2000). The ability of probiotics to improve host immune function (Ouwehand et al., 2002; Reid et al., 2003), modulate inflammatory and hypersensitivity responses (Reid et al., 2003) has also been documented.

As a nonpathogenic representative of the human intestinal flora, Escherichia coli Nissle 1917 (EcN), the active ingredient of the probiotic drug Mutaflor^®, was isolated in 1916 by the German physician Alfred Nissle and has been used as a probiotic agent in human and animal medicine to treat chronic inflammatory and infectious diseases of the human and animal intestine (Kamada et al., 2005). The EcN strain can reduce bacterial invasion and modulate cytokine expression of Caco-2 cells infected with Crohn's disease-associated E. coli LF82 (Huebner et al., 2011). EcN has also been engineered to express the autoinducer molecule cholera autoinducer 1 (CAI-1). The recombinant strain significantly inhibited Vibrio cholerae virulence gene expression and colonization in an infant mouse model (Duan & March, 2010). A recent study demonstrates that EcN reduces Salmonella typhimurium intestinal colonization and expression of proinflammatory cytokines by competing for iron (Deriu et al., 2013).

Clostridium perfringens is a Gram-positive anaerobic spore-forming bacterium, able to produce more than ten kinds of toxins and enzymes responsible for wound contamination, anaerobic cellulitis, and gas gangrene (Songer, 1996). It is also the third most common cause of food poisoning in the United States (Scallan et al., 2011) mediated primarily by its enterotoxin, CPE (Songer, 1996), and is frequently responsible for non- C. difficile cases of antibiotic-associated diarrhea (Modi & Wilcox, 2001) . Clostridium perfringens strains are classified into five toxinogenic types (A, B, C, D, and E), based on the production of four major toxins (α, β, ι, and ε) (Petit et al., 1999). Type A isolates are the most widespread in the intestines of warm-blooded animals and in the environment, causing enteric disease in a number of domestic animals, including chickens, horses, pigs, and sheep (Songer, 1996). The use of probiotics to control C. perfringens infection has drawn more attention recently. A number of probiotic bacteria such as Bacillus subtilis (Teo & Tan, 2005), Lactobacillus salivarius (Kizerwetter-Swida & Binek, 2009) and Lactobacillus fermentum (Allaart et al., 2011) have been shown to either inhibit C. perfringens growth or repress virulence factor production.

Biofilm formation, which can enhance persistence and increase resistance to stress, has been demonstrated to be involved in an estimated 80% of all infections (Biel, 2010). After colonization, the cells are able to communicate via quorum sensing using products such as N-acyl homoserine lactone to form biofilm, consisting of polysaccharide, extracellular DNA, environment material such as minerals and blood components (Murga et al., 2001; Sutherland, 2001; Moscoso et al., 2006; Nadell et al., 2008). It has been reported that all sequenced C. perfringens strains can form biofilms that protect them from oxidative and antibiotic stress (Varga et al., 2008). Because the probiotics EcN can outcompete a number of different pathogenic E. coli during biofilm formation (Hancock et al., 2010), it would be meaningful to determine the interaction between EcN and C. perfringens during biofilm formation.

The aim of this study was to investigate the impact of EcN on C. perfringens type A strain. Our results show that EcN inhibits both growth and toxin production by C. perfringens in vitro, as well as biofilm formation, indicating that the use of EcN may be a useful, alternative approach to controlling the pathogenic effects of C. perfringens type A infection.

Materials and methods

Bacterial strains and growth conditions

Strains and plasmids used in this study are listed in Table 1. Clostridium perfringens strain CP4 (CP4) is a virulent type A strain isolated from a necrotic enteritis (NE) case in Ontario (Thompson et al., 2006). Plasmid pJIR750 (Bannam & Rood, 1993) was used to transform CP4 as described (Scott & Rood, 1989) to yield CP4(pJIR750) and provided a chloramphenicol resistance (Cm⁺) marker to distinguish CP4 from EcN whenever co-culture experiments were performed. Plasmid pWSK129 (Wang & Kushner, 1991) was used to transform EcN yielding EcN(pWSK129) with kanamycin resistance (Kan⁺) marker whenever needed.

Table 1

Strains and plasmids used in this study

Strains and plasmid	Relevant serotype or genotype	Source or reference
EcN	E.coli Nissle 1917, O6:K5:H1	Ulrich Sonnenborn GmbH, Germany
C. perfringens
CP4	Type A, field isolate, α-toxin, NetB toxin positive	J.F. Prescott Guelph, Canada
pJIR750	Chloramphenicol-resistant, shuttle vector	Julian Rood Monash University, Australia
pWSK129	pSC101 replicon, kanamycin resistance	Lab stock (Wang & Kushner, 1991)

Stability of both plasmids was evaluated as described before (Kang et al., 2002; Brenneman et al., 2012) with the exception that strains were grown in the absence of chloramphenicol or kanamycin. Plasmids pWSK129 and pJIR750 were found to be stable in the absence of antibiotics for more than 50 generations. Strains containing each plasmid grew at a similar rate in the absence of antibiotics as wild-type strains did. All experiments were performed under anaerobic conditions using the BD GasPak EZ Anaerobe Container System.

All strains were routinely cultured at 37 °C in trypticase–peptone–glucose (TPG) broth consisting of 5% (w/v) Bacto-Tryptone (Difco Laboratories, Detroit), 0.5% (w/v) proteose peptone (Difco), 0.4% (w/v) glucose, and 0.1% (w/v) sodium thioglycolate (Leslie et al., 1989) unless otherwise specified. CP4 and EcN strains grow well in TPG broth under anaerobic condition. Tryptone sulfite cycloserine agar (TSC) plates (Merck) with 5 μg mL⁻¹ chloramphenicol and Luria–Bertani (LB) agar plates with 50 μg mL⁻¹ kanamycin were used for the enumeration of CP4(pJIR750) and EcN(pWSK129) from mixed cultures separately. Tryptic soy broth (TSB) with 10 mM lactose was used in the C. perfringens biofilm study. All centrifugations were performed at room temperature unless otherwise specified.

Tissue culture cells and growth conditions

Human epithelial colorectal adenocarcinoma cells (Caco-2) cells and mouse macrophage cells (RAW264.7) were grown in Dulbecco's modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin–streptomycin (pen/strep) and maintained in a humidified environment with an atmosphere of 5% CO₂ at 37 °C.

Co-culture experiments with C. perfringens and EcN

CP4(pJIR750) and EcN(pWSK129) strains were cultured individually under anaerobic conditions in 3 mL TPG broth with chloramphenicol or kanamycin, respectively, overnight to reach a stationary phase. Both cultures were inoculated into fresh 3 mL TPG broth (1 : 20) with appropriate antibiotics and grown to an optical density at 600 nm (OD₆₀₀) of 0.8. The cultures were then centrifuged at 16 000 g for 5 min and the pellet was washed once with phosphate-buffered saline (PBS) and resuspended in the same volume of fresh TPG broth. A preliminary assay we found that at equal OD₆₀₀ values (0.8), the EcN culture contained about twice as many CFU as the CP4 culture. For all co-culture experiments, a total final volume of 3.2 mL was used and 0.1 mL of resuspended CP4(pJIR750) was inoculated into each tube. Then, we added 0.05, 0.1, or 0.2 mL of resuspended EcN(pWSK129) to tubes designated Co-1 (Cp:EcN≈1 : 1), Co-2 (Cp:EcN≈1 : 2), and Co-4 (Cp:EcN≈1 : 4) to achieve the indicated ratios of Cp to EcN. Negative control cultures contained CP4(pJIR750) only (C.p.) and EcN(pWSK129) only. The EcN controls EcN1, EcN2, and EcN4 were inoculated with 0.05, 0.1, and 0.2 mL EcN, respectively. Fresh TPG medium was added to all tubes to achieve the final volume of 3.2 mL. All experiments were performed in triplicate. Cultures were incubated anaerobically for 20 h, and in some experiments, OD₆₀₀ and pH measurements were taken every 2 h until 8 h and CFU of CP4(pJIR750) and EcN(pWSK129) were determined on TSC agar plates (5 μg mL⁻¹ Cm) and LB agar plates (50 μg mL⁻¹ Kan), respectively. Supernatants obtained from 20-h cultures were collected by centrifugation (16 000 g, 10 min) and filtered through a 0.22-μm-pore-size filter (Millipore), followed by Western blot as described below.

The effect of EcN supernatants on C. perfringens growth and toxin production

EcN was incubated anaerobically in TPG broth overnight at 37 °C. After centrifugation at 16 000 g for 10 min, the supernatant was filtered (0.22 μm, Millipore) to remove any resident cells yielding the EcN cell-free supernatant (CFS_E). Amicon Ultra-15 Centrifugal Filter Units (10 000 NMWL, Millipore) were used to prepare 5× and 10× concentrated EcN supernatants, designated 5CFS_E and 10CFS_E. All CFS_Es were used immediately or stored at −20 °C until use.

Fresh CP4 cultures were prepared as mentioned above. Then, 0.1 mL of resuspended CP4 cells was inoculated into either 3 mL TPG medium (Cp-neg), 2.4 mL TPG plus 0.6 mL 5CFS_E (Cp-5CFS_E), or 2.4 mL TPG plus 0.6 mL 10CFS_E (Cp-10CFS_E). All experiments were performed in triplicate. Cultures were incubated anaerobically at 37 °C. OD₆₀₀ values were determined at 2-h intervals for 8 h, at which time culture supernatants collected and subjected to Western blot analysis.

Caco-2 cell attachment and cytotoxicity assay

Caco-2 cell attachment experiments were performed as previously described (Martin & Smyth, 2010) with some modification. Caco-2 cells were seeded into 24-well plates at 5 × 10⁵ cells per well and cultured for 4 days. EcN cultures were collected by centrifugation at 16 000 g for 10 min and resuspended in DMEM containing 5% FBS to a final concentration of 5 × 10⁷ CFU mL⁻¹. Caco-2 cells were washed once with PBS and replaced in 1 mL fresh DMEM with 5% FBS. Then, 100 μL of EcN (5 × 10⁶ CFU) was added at a multiplicity of infection (MOI) ≈10 and incubated for either 2 h (Cp+, EcNpre2h) or 6 h (Cp+, EcNpre6h). Caco-2 cells without EcN were used as the negative control (Cp+, EcN−). Overnight cultures of CP4(pJIR750) grown in TPG broth (5 μg mL⁻¹ Cm) were collected by centrifugation and resuspended with DMEM containing 5% FBS to 5 × 10⁷ CFU mL⁻¹. Hundred microliters of resuspended CP4(pJIR750) was inoculated into each well (MOI≈10) and incubated at 37 °C under anaerobic conditions.

For bacteria attachment experiments, the culture medium was removed after 1-h incubation with CP4(pJIR750), and the cells were washed twice with sterile PBS. Two hundred microliters 0.05% trypsin was added to each well and incubated for 15 min at room temperature. An additional 800 μL PBS was added to each well and mixed completely by pipetting. A series of 10-fold dilutions were plated onto TSC plates with 5 μg mL⁻¹ chloramphenicol. The bacterial colonies obtained from each well were counted and recorded after 24-h anaerobic incubation at 37 °C.

For cell cytotoxicity experiments, Caco-2 cells were incubated with either CP4(pJIR750) alone (MOI≈10) or together with EcN pre-incubation (MOI≈10) for 2 or 6 h as described above. After incubation at 37 °C for 20 h, the supernatants from each well were collected by centrifugation at 16 000 g for 10 min and passed through a 0.22-μm filter (Millipore). Lactate dehydrogenase (LDH) enzyme activity in each well was determined using the lactose dehydrogenase activity kit (Biovision) following the manufacturer's protocol. Equal volumes of residual supernatants from each well were precipitated with 10% TCA and subjected to Western blot assay. Additional triplicate wells were included in each group, and the CFU numbers of CP4 were determined at the end of the assay.

SDS-PAGE and Western blot assay

Total proteins in the supernatants were precipitated by adding 10% (v/v) trichloroacetic acid (TCA) and incubated at 4 °C overnight. Then, the pellets were collected by centrifugation at 16 000 g for 10 min at 4 °C and washed twice with cold acetone containing 0.1% β-mercaptoethanol. The pellets were then dissolved in 100 μL SDS loading buffer. Equal volumes of each sample were separated by discontinuous 12% SDS-PAGE and transferred electrophoretically to nitrocellulose membranes. Membranes were blocked with 5% skim milk in 100 mM Tris containing 0.9% NaCl and 0.1% Tween 20 (pH 7.4). Filters were incubated first with rabbit polyclonal antibodies specific for α-toxin C-terminal (PlcC) (Zekarias et al., 2008) and NetB (laboratory stock, made by Hua Mo). Filters were washed and the secondary antibody, an alkaline phosphatase-conjugated goat anti-rabbit immunoglobulin G (Southern Biotechnology, Birmingham, AL), was added. Immunoreactive bands were detected by the addition of nitroblue tetrazolium (NBT)-5-bromo-4-chloro-3-indolyl phosphate (Sigma). The reaction was stopped after 2–5 min by washing the membranes several times with large volumes of deionized water. The band intensities were analyzed by image j software (Wayne Rasband National Institutes of Health, Bethesda, MD), and the relative expression rate was calculated and normalized to control cultures.

Biofilm assays

Clostridium perfringens biofilm assays were performed as described by Varga et al., (2008) with some modifications. Overnight cultures of CP4(pJIR750) and EcN(pWSK129) were washed once in PBS and resuspended to an OD₆₀₀ of 0.1 in TSB medium supplemented with filter-sterilized lactose (10 mM). Then, 100 μL of resuspended CP4(pJIR750) was inoculated into 96-well polystyrene tissue culture plates. Additional 100 or 200 μL of resuspended EcN(pWSK129) culture was inoculated into each well, designated Cp+EcN1 and Cp+EcN2_, respectively. Wells with 100 μL CP4(pJIR750) or 100 μL EcN(pWSK129) were also included and named Cp-control and EcN-control individually as negative controls. Extra TSB medium (10 mM lactose) was added into wells to achieve a final volume of 300 μL whenever needed. Cultures were incubated anaerobically at 30 °C for 5 days in a sealed container to prevent evaporation. Each plate also contained another series of wells to which 300 μL of TSB medium was added to serve as negative controls. Biofilm formation and bacteria colony numbers in biofilm and floating supernatant were determined.

Biofilm formation was determined using the distribution between planktonic (free-floating) and sessile (biofilm) cells as described (Varga et al., 2008) with minor modifications. After 5 days, the supernatant from each well was transferred to a new 96-well plate and the OD₆₀₀ values were measured. The biofilm remaining in the original wells was then gently washed twice with PBS and incubated with 300 μL of 0.1% crystal violet for 30 min at room temperature. The crystal violet was then removed from the wells carefully, and the resident wells were gently washed twice with PBS. Then, 300 μL of methanol was added into each well and incubated for 30 min at room temperature to extract bound crystal violet. After transferring the extracted crystal violet to a new 96-well plate, the A₅₇₀ values of the methanol-extracted dye were measured in a SpectraMax M2 Multi-Mode Microplate Reader (Molecular Devices, LLC). Wells containing TSB medium were used to subtract nonspecific staining background. The ratio of the A₅₇₀ to OD₆₀₀ was used as a relative measure of biofilm production. The number of CP4(pJIR750) and EcN(pWSK129) in both biofilm and planktonic phase was also determined on selective plates.

Cytokines reaction in the presence of RAW264.7 cells

Murine macrophage RAW264.7 cells were seeded into 24-well cell culture plates at 5×10⁵ cells per well and cultured at 37 °C for 24 h. After washing twice with PBS, 1 mL DMEM containing 5% FBS was added to each well. Overnight EcN cultures were collected by centrifugation (16 000 g for 5 min) and resuspended with the same volume of TPG. Then, 100 μL of EcN was inoculated into each well (MOI≈100) and incubated for 3 h before the infection of CP4. CP4 cells from an overnight culture were collected and resuspended in TPG and diluted appropriately. Hundred microliters of CP4 (MOI≈10 or MOI≈100) was added to each well. Cells incubated with EcN alone (MOI≈100), CP4 alone (MOI≈10 or MOI≈100), and mock-infected cells were used as controls. After a 20-h incubation, the supernatants from each well were collected and centrifuged (16 000 g, 10 min, 4 °C) to remove suspended bacteria. The supernatants were analyzed directly by Bio-Plex Multiplex Cytokine Assay (Bio-Rad) or stored at −20 °C prior to cytokine analysis. Granulocyte colony-stimulating factor (G-CSF), granulocyte–macrophage colony-stimulating factor (GM-CSF), interleukin-1 beta (IL-1β), and interleukin-6 (IL-6) were measured.

Statistical analysis

All statistics were carried out using graphpad prism 6.0 (GraphPad Software, San Diego, CA). Data were expressed as means ± standard errors ( n = 2 for cytokines assay; n = 3 for the rest of assays) and evaluated with two-tailed t-test. Differences were considered significant at P < 0.05.

Results

EcN represses growth of C. perfringens during co-culture

In preliminary co-culture experiments with C. perfringens strain CP4, we found that EcN interestingly decreased gas production by CP4. At an inoculation ratio of 1 : 1 (Co-1), gas production by CP4(pJIR750) decreased slightly compared to the negative control (C.p.). When the ratio of CP4(pJIR750) to EcN(pWSK129) was increased to 1 : 2 (Co-2) and 1 : 4 (Co-4), gas production was dramatically reduced (Fig. 1a). These results indicate that EcN(pWSK129) inhibits CP4(pJIR750) gas production in a dose-dependent manner. The major pathogenic effects of C. perfringens are primarily due to the toxins they produce. CP4 is a poultry isolate that causes necrotic enteritis and is known to produce and secrete two toxins important for that disease, α-toxin and NetB toxin (Chalmers et al., 2008; Keyburn et al., 2008). Compared to the C.p. control, the amount of α-toxin and NetB toxin present in the culture supernatants after 20-h incubation was reduced as the amount of EcN(pWSK129) present in the initial inoculation increased from 1 : 1 to 1 : 2 and 1 : 4 (Fig. 1b). We note that the band on the Western blot below the α-toxin is a cross-reactive band present in the EcN(pWSK129) culture supernatant and not a degradation product of α-toxin, because the band could also be observed in EcN(pWSK129) supernatant control (indicated by an arrow in Fig. 1b). The intensity of Western blot bands were calculated using image j software and the relative expression ratio of toxins were described as the percentage normalized to control cultures (Fig. 1c). Presence of EcN(pWSK129) significantly repressed the production of both α-toxin and NetB toxin in a dose-dependent manner. We also determined the bacteria number of CP4(pJIR750) and EcN(pWSK129) in each group after 20-h incubation. The results indicated that the final CP4(pJIR750) number decreased dramatically as the inoculation ratio of EcN(pWSK129) increased from 1 : 1 to 1 : 2 and 1 : 4 (Fig. 1d). Besides, there appeared to be a growth competition effect between EcN(pWSK129) and CP4(pJIR750) because the presence of CP4(pJIR750) also inhibited the growth of EcN(pWSK129), while this kind of inhibition effects was decreased as the original inoculation dose of EcN(pWSK129) increased from 1 : 1 to 1 : 4 (Fig. 1e).

Fig. 1

Growth curves, pH, and toxin production during co-culture experiments. CP4(pJIR750) and EcN(pWSK129) were inoculated into TPG medium at the ratio of 1 : 1 (Co-1), 1 : 2 (Co-2), and 1 : 4 (Co-4) based on CFU and cultured at 37 °C for 20 h. Negative controls included CP4(pJIR750) negative control (C.p.) and EcN(pWSK129) only culture, named EcN1, EcN2, and EcN4, respectively. (a) Gas production in control culture (C.p.) and CP4-EcN co-culture tubes Co-1, Co-2, and Co-4 at 8 h. (b) Western blot analysis of α-toxin and NetB expression. Cell-free supernatants after 20-h co-culture were precipitated with 10% TCA and analyzed by Western blot. Blots were probed with anti-PlcC (upper) and anti-NetB (lower) antibody. Equal volumes were loaded in each lane. A representative Western blot from three independent repeats is shown. Arrows indicate the nonspecific bands in EcN cultures. Lines to the left indicate the molecular masses of the protein standards in kDa. (c) Semi-quantitative analysis of α-toxin and NetB production. The band intensities of alpha-toxin and NetB in Western blot were analyzed by image j software and normalized to control group (C.p.) shown as percentages (%). (d) CFU of CP4(pJIR750) in C.p., Co-1, Co-2, and Co-4 cultures were determined at 20 h. (e) CFU of EcN(pWSK129) in Co-1, Co-2, Co-4, as well as control cultures EcN1, EcN2, and EcN4, were determined at 20 h. (f) Samples from co-culture Co-1, negative control C.p. and EcN1 were taken and OD₆₀₀ values were determined at 0, 2, 4, 6, 8, and 20 h. (g) CFU of EcN(pWSK129) in Co-1 and EcN1, CFU of CP4(pJIR750) in Co-1 and C.p. cultures were determined at 0, 2, 4, 6, 8, and 20 h. (h) pH values of Co-1, EcN1, and C.p. cultures at each time point. Data are presented as mean ± SD ( n = 3). Ns, no significant difference; * P < 0.05; ** P < 0.01; *** P < 0.001; **** P < 0.0001.

To gain some insight into the mechanism driving the above results, we used the cultures containing equal inoculation dose of EcN(pWSK129) and CP4(pJIR750) (Co-1), as well as negative control cultures C.p. and EcN1, to explore the inhibition effect in more detail. Samples were taken at 2, 4, 6, 8, and 20 h to measure the OD₆₀₀, CFU, and pH values. The OD₆₀₀ values of both control cultures were similar, although the final values of EcN1 were a little higher than C.p. culture (Fig. 1f). However, the actual CFU numbers of EcN(pWSK129) increased much faster compared to CP4(pJIR750) (Fig. 1g). As we mentioned before, the growth competition effects were observed for both strains, because the CFU numbers of each bacteria in co-culture condition were lower than the individual control groups (Fig. 1g). While the pH in all cultures declined steadily during the first 6 h of growth, although the pH decreased faster and reached a lower minimum pH, around 5.7, when EcN was present (Fig. 1h). We also noticed that the pH declined much lower as long as the inoculation dose of EcN increased (data not shown). The lower pH in the mixed culture may be at least partially responsible for lower titer achieved by CP4 in the presence of EcN (Fig. 1g). All the results above indicate that the presence of EcN(pWSK129) could inhibit the growth of CP4(pJIR750) in a dose-dependent manner.

EcN supernatants do not inhibit the growth of C. perfringens

The growth inhibition observed when EcN is present could be due to a direct competition for nutrients or secretion of inhibitory compounds by EcN (e.g. bacteriocins), or some combination of the two. To evaluate the possibility that EcN is secreting an inhibitory compound, we prepared cell-free supernatants from EcN cultures and concentrated them either fivefold (5CFS_E) or 10-fold (10CFS_E) (see Materials and methods). . The results showed that the addition of either CFS_E had no apparent effect on the growth of CP4 and toxins production (data not shown). Thus, we conclude that secretion of inhibitory compounds by EcN is not a factor in the observed effects on CP4 growth and toxin production. Therefore, the direct competition during co-culture process should be responsible for the inhibition effect.

EcN represses C. perfringens growth in the presence of Caco-2 cells

Adherence of enteropathogens to intestinal epithelium cells is an important first step in establishing an infection (Reis & Horn, 2010). Some probiotics have shown the ability to compete with pathogens for adherence to host cells (Lu & Walker, 2001; Vine et al., 2004). Thus, we examined the effects of EcN on CP4 adherence to Caco-2 epithelial cells. The results showed that the pre-incubation of Caco-2 cells with EcN(pWSK129) (MOI≈10) for 2 or 6 h had no obvious inhibition effect on C. perfringens cell attachment ability (Fig. 2a). However, the total number of C. perfringens after 20 h was decreased in cells with EcN pretreatment for 2 and 6 h compared to wells in which EcN was not present ( P < 0.01) (Fig. 2b). The production of both α-toxin and NetB toxin was decreased in the presence of EcN ( P < 0.05) (Fig. 2c and d), consistent with co-culture experiments described previously. Finally, the presence of EcN increased the intactness of Caco-2 cells about sixfold ( P < 0.0001) shown by decreased level of cellular LDH (Fig. 2e). The results demonstrated that the inhibition effects of EcN on C. perfringens during co-culture experiments were also observed in the presence of host cells.

Fig. 2

Growth inhibition effects of EcN on CP4 in the presence of Caco-2 cells. Caco-2 cells were exposed to Clostridium perfringens strain CP4(pJIR750) (MOI≈10) alone (Cp+, EcN−), or together with EcN(pWSK129) (MOI≈10) pre-incubation for 2 h (Cp+, EcNpre2h) and 6 h (Cp+, EcNpre6h) separately. (a) CP4 attachment rate relative to original inoculated bacteria number, after 1-h incubation. (b) Total CP4 CFU after 20-h incubation. (c) Western blot analysis of α-toxin (upper) and NetB toxin (lower). After 20-h incubation, the supernatants from each well were collected by centrifugation and filter (0.22 μm), then precipitated by 10% TCA and equal volumes samples were separated by SDS-PAGE and immunoblotted with anti-PlcC and anti-NetB antibody. Lines to the left indicate the molecular masses of the protein standards in kDa. (d) Semi-quantitative analysis of α-toxin and NetB production. The band intensities of alpha-toxin and NetB in Western blot were analyzed by image j software and normalized to control group (Cp+, EcN−) shown as percentages (%). (e) Lactate dehydrogenase (LDH) released from Caco-2 cells after 20-h incubation with CP4(pJIR750) alone or together with EcN(pWSK129) pre-incubation. Data are presented as mean ± SD ( n = 3). * P < 0.05; ** P < 0.01; **** P < 0.0001.

Presence of EcN outcompetes CP4 growth during biofilm formation

The ability to form biofilms is an important strategy for bacterial persistence in host and free-living environments, including C. perfringens (Varga et al., 2008). EcN is a good biofilm former and able to outcompete enteropathogenic, enterotoxigenic, and enterohemorrhagic E. coli strains during biofilm formation (Hancock et al., 2010). Thus, we were interested in determining whether EcN could decrease biofilm formation of C. perfringens. To address this question, we set up a biofilm assay. CP4(pJIR750) cells were inoculated into 96-well plates with and without the addition of EcN(pWSK129) at the ratio (CP4/EcN) of 1 : 1 (Cp+EcN1) or 1 : 2 (Cp+EcN2). After 5 days, biofilm formation was observed in all wells according to ratio of A₅₇₀ to OD₆₀₀, especially in wells containing CP4(pJIR750) alone (Fig. 3a), consistent with previous reports for other C. perfringens strains (Varga et al., 2008). Both mixed cultures (Cp+EcN1 and Cp+EcN2) had significantly less biofilm formation compared to CP4 alone or EcN alone (Fig. 3a), indicating that there were some interference effects between C. perfringens and EcN during biofilm formation. We then determined individual bacteria numbers in biofilm and floating supernatant, separately. The existence of EcN decreased the colony number of CP4 in both biofilm (Fig. 3b) and supernatant (Fig. 3c), especially when the starting inoculation ratio was 1 : 2 (Cp+EcN2). At the same time, the number of EcN in both biofilm and floating supernatant was not obviously affected by CP4 (Fig. 3d and e). We noticed that high-dose inoculation (Cp+EcN2) did not cause more EcN cells in biofilm; instead, significantly more EcN cells were observed in the supernatants (Fig. 3e). The ratio of CP4 cells in the presence of EcN was also calculated using the mean number of CP4 in Cp-control culture as a standard, which demonstrated that the CP4 number decreased about 60% in biofilm and 70% in supernatant in the presence of EcN (Cp+EcN2) (Fig. 3f). Therefore, we concluded that the presence of EcN significantly outcompeted the growth of CP4 during biofilm formation.

Fig. 3

Biofilm formation by Clostridium perfringens strain CP4(pJIR750) in the presence of EcN(pWSK129). Overnight CP4(pJIR750) and EcN(pWSK129) cultures were washed once with PBS and resuspended in TSB medium (10 mM lactose) to an OD₆₀₀ of 0.1, individually. Hundred microliters of resuspended CP4(pJIR750) was inoculated into 96-well polystyrene tissue culture plates. Then, 100 or 200 μL of resuspended EcN(pWSK129) culture was inoculated into each well, designated Cp+EcN1 and Cp+EcN2_, respectively. Wells with 100 μL CP4(pJIR750) or 100 μL EcN(pWSK129) were also included and named Cp-control and EcN-control individually as negative controls. Plates were incubated anaerobically at 30 °C for 5 days. (a) Biofilm formation measured by A₅₇₀/OD₆₀₀. (b) Number of C. perfringens in biofilm. (c) Number of C. perfringens in floating supernatant. (d) Number of EcN in biofilm. (e) Number of EcN in floating supernatant. (f) Decreased percentages of C. perfringens in biofilm and floating supernatant. Data are presented as mean ± SD ( n = 3). Ns, no significant difference; * P < 0.05, ** P < 0.01; *** P < 0.001.

EcN modulates cytokine expression in RAW264.7 cells infected with CP4

Tissue damage and inflammation response are caused by toxins secreted by C. perfringens (Songer, 1997; Wallace et al., 1999; Stevens, 2000; Oda, 2012). Mice injected with α-toxin secreted dramatically increased proinflammatory cytokines such as tumor necrosis factor-alpha (TNF-α), IL-1β, IL-6, and interferon gamma (IFN-γ), which could be important factors responsible for lethality and systemic hemolysis (Oda et al., 2008). Because the co-incubation experiments mentioned above showed that EcN could inhibit CP4 growth and reduced toxins production, therefore it would be interesting to measure the inflammatory cytokines profile caused by C. perfringens with or without the presence of EcN.

Compared to mock-infected cells, RAW264.7 cells infected with CP4 (MOI≈10) caused the release of significantly increased levels of inflammatory molecules, including 4906 ± 256.0 ( n = 2) pg mL⁻¹ IL-1β ( P < 0.01) (Fig. 4a), 408.4 ± 59.90 pg mL⁻¹ IL-6 ( P < 0.05) (Fig. 4b), 57021 ± 375.2 pg mL⁻¹ G-CSF ( P < 0.0001) (Fig. 4c), and 2207 ± 43.52 pg mL⁻¹ GM-CSF ( P < 0.001) (Fig. 4d). Pre-incubation with EcN (MOI≈100) significantly decreased the level of cytokine synthesis compared to CP4 infection alone. Similar results were observed when RAW cells were incubated with CP4 plus EcN at an MOI≈100. The level of cytokine secretion was lower when RAW cells were incubated with a high dose of CP4 (MOI≈100) than at a lower dose (MOI≈10) (Fig. 4a–d), probably due to early cell death caused by the higher dose of CP4.

Fig. 4

Cytokine production of RAW264.7 cells in response to Clostridium perfringens strain CP4(pJIR750) incubation with or without EcN(pWSK129) pre-incubation. RAW264.7 cells were pre-incubated with EcN(pWSK129) (MOI≈100) for 3 h, and then, CP4(pJIR750) were inoculated (MOI≈10 and 100 individually). After 20-h co-incubation, cytokines including (a) IL-1β, (b) IL-6, (c) G-CSF, and (d) GM-CSF were measured using Bio-Plex ELISA kit (Bio-Rad). Cells incubated with CP4(pJIR750) (MOI≈10 and 100) and mock-infected cells (negative) were included as control. Data are presented as mean ± SD ( n = 2). * P < 0.05; ** P < 0.01; *** P < 0.001; **** P < 0.0001.

Discussion

The relationship between commensals and pathogens have drawn more and more attention recently (Kamada et al., 2012). In particular, probiotics may provide an alternative to antibiotics for fighting against infectious diseases. In our study, we observed direct competition effects between EcN and C. perfringens type A strain CP4. The presence of EcN in co-culture experiments inhibited the growth of CP4 in a dose-dependent manner, resulted in reduced gas production, total CP4 numbers, and toxin production (α-toxin and NetB toxin).

EcN has been demonstrated to inhibit pathogens adhesion and invasion ability. The mechanisms behind this are still not completely clear, but it is generally accepted that probiotics outcompete pathogens for nutrients and space, often producing inhibitory substances, such as lactic acid and other organic acids, hydrogen peroxide, and bacteriocins (Mital & Garg, 1995; Bogovic-Matijasic et al., 1998; Salminen et al., 2010; Dobson et al., 2012). In this study, pre-incubation of epithelial cells Caco-2 with EcN did not inhibit binding of C. perfringens, as was shown for Salmonella (Schierack et al., 2011) and adherent-invasive E. coli (AIEC) strains (Boudeau et al., 2003), possibly due to a difference in cell receptors targeted by different pathogens. However, pre-incubation with EcN dramatically reduced the overall numbers of CP4 and the level of toxins production (Fig. 2), which is consistent with previous co-culture experiments.

It has been reported that all sequenced C. perfringens can form biofilms that protect them from oxidative and antibiotic stress (Varga et al., 2008). To be consistent, we observed obvious biofilm formation using C. perfringens CP4 strain (Fig. 3a). EcN has been demonstrated to outcompete several pathogenic bacteria such as enterotoxigenic E. coli and enteropathogenic E. coli strains in biofilm formation (Hancock et al., 2010). In our study, the presence of EcN severely decreased the C. perfringens CP4 biofilm production (Fig. 3a), as well as the number of C. perfringens in biofilm (Fig. 3b) and floating supernatant (Fig. 3c). The mechanism underlined is possibly due to the unique yersiniabactin system of EcN, the ability to produce several adhesions and fimbriae, or the combination (Hancock et al., 2010). Our results demonstrated that the competition between EcN and C. perfringens observed in vitro culture condition also existed during the biofilm formation.

Cytokines are a key element in the inflammatory response that characterizes sepsis and septic shock (Oda et al., 2008). Proinflammatory cytokines induced by C. perfringens and its toxins including IL-8, TNF-α, IL-1, and IL-6 (Zhou et al., 2009; Forder et al., 2012; Oda, 2012; Lee et al., 2013; Tuovinen et al., 2013) are likely to be responsible for C. perfringens-induced septic shock and death. Chickens infected with C. perfringens have been shown to produce increased levels of proinflammatory cytokines such as IL-1β and IL-6 (Sugiarto & Yu, 2004), responsible for unregulation of several β-defensins (van Dijk et al., 2007). GM-CSF and G-CSF are induced in in vitro studies by the lipoteichoic acid component of the cell membranes of Gram-positive bacteria (Saba et al., 2002; Seo et al., 2008; Chou & Lu, 2011). Thus, the increased levels of cytokines by RAW cells in our study induced by either the C. perfringens cell membrane and/or secreted toxins were observed. It has been shown that probiotics can ameliorate the inflammation by inhibiting proinflammatory cytokine production (Hegazy & El-Bedewy, 2010; Mencarelli et al., 2011; Ganguli et al., 2013). As a classic probiotic bacteria, EcN has been shown to modulate the cytokine profile in Caco-2 cells infected with Crohn's disease-associated E. coli LF82 (Huebner et al., 2011) and to decrease IL-10 and TNF-α levels induced by Salmonella typhimurium in gnotobiotic pigs (Splichalova et al., 2011). The administration of EcN also results in reductions of intestinal inflammation in S. typhimurium-infected C57BL/6 mice (Deriu et al., 2013). In this study, the inhibitory effects of EcN on C. perfringens growth and toxin production were likely responsible for the observed decrease in the secretion of inflammatory cytokines.

Taken together, our results indicate that EcN can repress the growth of C. perfringens and decrease toxin production both in vitro condition in a dose-dependent manner. EcN also inhibits the proinflammation response caused by C. perfringens. All these results indicate that EcN could possibly provide another strategy to control C. perfringens-related infection.

Acknowledgements

The work was supported by NIH R01 AI60557. We appreciate Dr. Ulrich Sonnenborn (Ardeypharm GmbH, Germany) for generously providing E. coli Nissle 1917 strain. We also thank Dr John F. Prescott (University of Guelph, Ontario, Canada) for providing C. perfringens strain and Dr Julian Rood (Monash University, Clayton, Vic., Australia) for plasmid pJIR750. We thank Ms Hua Mo and Dr Bereket Zekarias for providing anti-α-toxin antibody and anti-NetB antibody.

In this report, the probiotic strain Escherichia coli Nissle 1917 (EcN) is shown to inhibit growth, gas production, and toxin production of Clostridium perfringens, most likely due to nutrition competition during co-incubation. While EcN did not inhibit the attachment to Caco-2 cells, C. perfringens biofilm growth was reduced, as was toxin production and cytotoxicity. Finally, Jiang et al. also demonstrate that EcN inhibits the C. perfringens-induced pro-inflammatory response in macrophages.