Investigation of motility and biofilm formation by intestinal Campylobacter concisus strains

Eight C. concisus strains that had been previously isolated from patients with a range of intestinal diseases [5] were included in this study. C. concisus strains UNSWCD, UNSW2, UNSW3 (Crohn’s disease), UNSW1 (chronic gastroenteritis), ATCC 51561 (healthy subject), ATCC 51562, UNSWCS and BAA 1457 (acute gastroenteritis) were grown on Horse Blood Agar (HBA) plates [Blood Agar Base No. 2 supplemented with 6% defibrinated horse blood (Oxoid; Heidelberg West, VIC, Australia)], and incubated at 37°C under microaerobic conditions with H₂ [generated using Campylobacter Gas Generating Kits (Cat. #. BR0056A, Oxoid)] for 24 h.

To evaluate motility following growth in liquid medium, the eight C. concisus strains were first grown on HBA plates for 24 h, harvested and then transferred to individual 10 ml Brain Heart Infusion (BHI) broths (Oxoid) containing 10% Foetal Bovine Serum (FBS) (Interpath; Heidelberg West, VIC, Australia), and where relevant, a known concentration of carboxy-methyl-cellulose (CMC) (Sigma-Aldrich; Castle Hill, NSW, Australia) that corresponded to a particular viscosity ([20], Additional file 1). The broths were then incubated for 24 h under microaerobic conditions at 37°C.

C. concisus cultures were centrifuged at 3,619 × g for 5 min and the cell pellets resuspended in 500 μl of PBS. The OD was then measured at 595 nm, and the OD adjusted to 0.5 (optimisation of the OD value is presented in Additional file 2). To conduct the motility assay, semi-solid serum plates (20 ml) [28 g Brucella broth (BD), 3.5 g Bacteriological Agar No. 1 (Oxoid) and 10% FBS] were inoculated with bacteria (3.8 μl), and then incubated at 37°C under microaerobic conditions for 72 h. After 72 h of incubation, a zone of motility was observed around the inoculation point, which represented the distance that the bacteria had migrated.

Two cell lines were used in this study, the human intestinal epithelial cell line Caco-2 (American Type Culture Collection; HTB-37) and the human mucin producing intestinal cell line LS174T (American Type Culture Collection; CL-188).

Caco-2 cells were grown in 10 ml cell culture medium comprised of Minimum Essential Medium (MEM), (Invitrogen; Mulgrave, VIC, Australia) supplemented with 10% FBS, 1 mM sodium pyruvate, 0.1 mM non-essential amino acids, 2.25 mg 1^-1 sodium bicarbonate and 100 μg ml^-1 penicillin and streptomycin (Invitrogen) in 25 cm² tissue culture flasks (In Vitro Technologies; Noble Park, VIC, Australia) at 37°C with 5% CO₂. Cells were seeded at a concentration of 5 × 10⁵ cells ml^-1 into 24-well plates and kept for 2 days at 37°C with 5% CO₂ in order to form a confluent monolayer (confirmed visually) for the adherence and invasion assays. Prior to seeding, the wells were coated with 1 ml collagen (0.338 mg ml^-1) and incubated for 20 min at 37°C with 5% CO₂.

LS174T cells were grown in 10 ml cell culture medium comprising Roswell Park Memorial Institute (RPMI)-1640 medium (Invitrogen) supplemented with 10% FBS and 100 μg ml^-1 penicillin and streptomycin in 25 cm² tissue culture flasks at 37°C with 5% CO₂. Cells were seeded at a concentration of 5 × 10⁵ cells ml^-1 into 24-well plates and kept for 2 days to form a confluent monolayer (confirmed visually) (0-day time point). The confluent monolayer was incubated at 37°C with 5% CO₂ for an extra 2 days to allow the development of a mucin layer (confirmed visually) for the adherence and invasion assays (2-day time point). The medium was changed daily until the development of a mucin layer.

Following incubation, the seeded cells in each well were washed twice with 1 ml of the relevant antibiotic-free medium, after which 1 ml of antibiotic-free medium with CMC at a concentration of 0 centipoise (cp), 20 cp or 74 cp (Caco-2 cells only) were aliquoted into the seeded wells. Monolayers were inoculated with C. concisus UNSWCD at a multiplicity of infection (MOI) of 200. Infected monolayers were then co-incubated with the bacteria for 6 h at 37°C with 5% CO₂ to allow adherence and invasion to occur. Invasion and adherence assays were then performed as previously described by Kaakoush et al.[5]. Bacterial adherence was calculated by subtracting the internalized bacteria determined using the gentamicin protection assay from the bacterial counts obtained using the adherence assay, and expressed as a relative percentage of inoculated bacteria.

LS174T cells were seeded at 5 × 10⁵ cells ml^-1 in 24-well plates and incubated at 37°C. After 2 days the cells were lysed and collected using 400 μl radioimmunoprecipitation (RIPA) buffer. A bicinchoninic acid assay was used to determine the protein concentration for each cell lysis sample collected. OD values were measured at 595 nm using the Bio-Rad 8550 Microplate Reader (Bio-Rad; Gladesville, NSW, Australia).

Proteins were then separated on 12% SDS-PAGE gels, and transferred to methanol-treated polyvinylidine difluoride membranes with use of the Trans-blot cell transfer system (Bio-Rad). Membranes were probed in accordance with the Immun-StarWesternC Kit protocol (Bio-Rad). Membranes were immunolabeled with mouse monoclonal antibodies against Mucin-1 (MUC1) (1:200), Mucin-2 (MUC2) (1:200), or β-actin (1:1000) (Santa Cruz Biotechnology Inc.; Santa Cruz, CA, USA). Goat anti-mouse IgG antibodies coupled to HRP (1:2000; Bio-Rad) were used as a secondary antibody. Bands were visualized and quantified using a LAS-3000 (Fujifilm; Brookvale, NSW, Australia).

In order to perform statistical analyses on the data, all experiments were repeated a minimum of three times, with each biological replicate consisting of two technical replicates. Analysis of Variance (ANOVA) was conducted using Minitab 15.1.0.0 (Minitab Inc.; State College, PA, USA) for all statistical studies conducted, with the significance level set at p < 0.05. The data were checked to ensure that it fitted the assumptions of ANOVA including homogeneity of variance and normal distribution. Tukey’s post hoc-multiple comparisons were conducted where significant differences were found. To determine if prior growth on the three different media types affected the motility of the eight C. concisus strains, a two-way ANOVA was conducted. In the statistical model, strain and medium were set as fixed factors, and the interaction of these factors were tested. A one-way ANOVA was performed for all other statistical analyses within this study.

The motility of eight strains of C. concisus following growth on solid agar and liquid media were measured. All C. concisus strains were found to be motile. ANOVA detected significant differences between strains (F_7,73 = 26.5, p < 0.01) (Figure 1). In contrast, the type of growth medium (solid or liquid) had no effect on the motility of any of the eight C. concisus strains tested (Figure 1). In relation to their motility, the eight strains formed two distinct groups; UNSWCD, UNSW1, UNSW2, UNSW3 and UNSWCS having higher overall levels of motility than BAA-1457, ATCC 51562 and ATCC 51561 (Figure 1). Interestingly, C. concisus UNSWCD, UNSW1, UNSW2 and UNSW3 were originally isolated from patients with chronic gastroenteritis, whereas strains UNSWCS, BAA-1457 and ATCC 51562 were isolated from patients with acute gastroenteritis and strain ATCC 51561 from a healthy individual [5]. While this finding suggests that the level of motility may play a role in disease outcome, the fact that the motility level of UNSWCS, isolated from a patient with acute gastroenteritis, was similar to that of strains isolated from patients with chronic gastroenteritis, would suggest that other pathogenicity factors, may also contribute to virulence. From a biological viewpoint, such differences in motility raise the possibility that strains with high motility may be more capable of swimming through the intestinal mucus and reaching the epithelium, as compared with strains with lower motility which potentially would be lost upon mucus turn-over.

The colonic mucus is composed of two semipermeable layers totalling a thickness of 800 μm [21]. The top layer of mucus is less viscous (approximately 20 cp), is much thicker and is normally colonised by commensal bacteria [22]. In comparison, the bottom layer is denser, thinner and much more elastic giving the layer properties that make it impermeable to normal commensal bacteria [22]. Given this, we were interested in determining the effect on bacterial motility following growth in a viscous liquid, as this would mimic to some degree the viscosity encountered in the host, where prior to reaching the intestinal epithelium, the bacteria must move through the intestinal mucus layers that act as a barrier against many pathogens. Due to the lack of homogeneity of the two mucus layers, many in vitro studies have used CMC to estimate bacterial motility through an intestinal mucus-like medium [9, 23]. The advantages of CMC are that it is non-toxic to tissue culture and bacterial cells, is more chemically defined than purified human mucin, and can be easily manipulated to create the desired viscosity; however, it remains an artificial compound that does not mimic the complex environment within the intestinal tract.

Following growth in a viscous medium (20.0 cp), the motility of the eight C. concisus strains showed significant changes (F_2,73 = 132.0, p < 0.0001), and this effect varied between strains (F_14,73 = 2.0, p = 0.03). In seven of the eight C. concisus strains motility levels were found to drop significantly as compared with motility levels following growth on solid or liquid media (Figure 2). While the distance travelled for strain UNSW1 slightly decreased following growth in viscous liquid medium, this was not significant (Figure 2). This decrease in motility following growth in viscous medium was unexpected. However, we speculate that this may result from the upregulation of auto-agglutination following the mechanosensing of the viscous environment. This possibility is strengthened by the fact that C. concisus strains have been observed to aggregate in large numbers around mucus [5].

To examine more closely the effect of increasing viscosity on C. concisus motility, further studies were conducted on C. concisus UNSWCD as this strain was shown in the above motility studies to have a high level of motility. Thus, we considered that any differences in motility at differing viscosity levels would be more detectable. Determination of the effect of increasing viscosity levels on C. concisus UNSWCD motility showed an initial decrease in the distance travelled (motility) as viscosity increased from 0 cp to 3.0 cp (Figure 3). Interestingly, between the viscosities of 3.0 cp to 20.0 cp, motility levels increased with increasing viscosity, being significantly higher at 20.0 cp than at 3.0 cp (p < 0.01) (Figure 3). This viscosity (20.0 cp) represents the approximate viscosity level in the outer mucus layer of the colon [9]. Following this peak in motility, levels fell between 20.0 cp and 40.0 cp, after which a further increase in motility was observed at 74.7 cp (Figure 3). Following 74.7 cp, motility levels dropped as viscosity levels increased. Given that the inner mucus layer of the intestine is thinner and more dense, the finding that C. concisus UNSWCD’s motility had a possible second motility peak at higher viscosity (74.7 cp) may suggest that it can adapt its motility to pass through the inner intestinal mucus layer. However, given that the viscosity of the inner intestinal mucus layer remains unknown, no direct comparisons can be made.

Interestingly, two peaks of motility have also been reported in C. jejuni which has been postulated to represent an adaptation that assists C. jejuni to reach the intestinal epithelial surface [23]. Shigematsu et al. suggested that the first increase in velocity was predominantly due to the bacterium’s flagella propelling the motion [23], while the second increase in motility they suggested was due predominantly to the spiral shape of the bacterium. In comparison, this same study showed the motility of the well-established pathogen Salmonella enterica Serovar Typhimurium had only one peak of motility at 1.5 cp after which the velocity decreased dramatically as the viscosity increased [23]. Interestingly, although S. Typhimurium is flagellated, it does not have a spiral body like C. jejuni and C. concisus.

Szymanski et al.[9] have previously shown that as viscosity increases from 0 to 141 cp, adherence to and invasion of C. jejuni to Caco-2 epithelial cells significantly increases (p < 0.05). Thus, we considered it possible that C. concisus may show a similar trend. To test this, adherence and invasion assays were performed with the addition of medium at two different viscosity levels following culture of the Caco-2 cells. The two viscosities chosen, 20.0 cp and 74.7 cp, were based on the two peak motilities shown in our studies examining the effect of viscosity on C. concisus UNSWCD motility (Figure 3). A control containing medium alone (no CMC) was also examined. The results of the adherence assay showed that C. concisus UNSWCD at zero viscosity (0 cp) adhered to Caco-2 cells at a similar level (6.01 ± 0.33%) to that previously reported in a study where centrifugation was used [5]. A finding that would suggest that for UNSWCD centrifugation may not be a necessary step for adherence. At a viscosity of 20.0 cp, C. concisus adherence was reduced (4.48 ± 0.53%) and at 74.7 cp it decreased even more (3.48 ± 0.50%) (F_2,21 = 8.9, p = 0.0015). Similar to the adherence assay, the invasion level of C. concisus into Caco-2 cells in the absence of centrifugation (0.56 ± 0.08%), was again similar to that reported in a previous study where centrifugation was included (0.47 ± 0.04%) [5]. When the viscosity was increased to 20.0 cp and 74.7 cp, invasion of C. concisus into Caco-2 cells remained the same (0.51 ± 0.07% and 0.50 ± 0.09%, respectively) and did not differ from the control (F_2,6 = 0.17, p = 0.84). The observed reduction in C. concisus adherence to Caco-2 cells upon exposure to viscous medium is consistent with the hypothesis that C. concisus may upregulate auto-agglutination upon mechanosensing the viscous environment.

Studies in C. jejuni have shown that adherence to and invasion of intestinal epithelial cells is significantly greater when the epithelium is covered in a mucus-like medium of the same viscosity as intestinal mucus, than in medium of zero viscosity [9]. Thus, the decrease in C. concisus adherence with increased viscosity differs from that previously observed for C. jejuni. To gain a better understanding of the changes in C. concisus pathogenesis in viscous environments that more closely resemble the situation within the host, the mucus-producing intestinal goblet-like cell line LS174T [24] was employed. In this experiment we determined if the presence and thickness of the mucus layer affected the attachment and invasion of C. concisus UNSWCD. To determine changes in the composition of the mucus layer produced by LS174T cells over time, the levels of MUC1 and MUC2, synthesised by LS174T, we analysed any changes in their expression at two different time points (0 and 2 days). Changes in MUC1 and MUC2 expression were compared relative to day 0 by normalising MUC1 or MUC2 chemiluminescence values against the β-actin control. While MUC1 levels were shown to significantly increase from day 0 to day 2 (p = 0.016) (Figure 4A,B), MUC2 expression in LS174T cells did not change from day 0 to day 2 (p = 0.67) (Figure 4A,B). Both visually and through the increase in expression of MUC1, we determined that there was a higher accumulation of mucus at the apical surface of LS174T cells at day 2 as compared with day 0.

Given this, we determined using adherence and invasion assays the ability of C. concisus UNSWCD to adhere to and invade LS174T cells at day 0 and 2. We observed that the percentage of C. concisus UNSWCD adhering to LS174T cells increased from 4.61 ± 0.44% at day 0 to 8.20 ± 0.50% at day 2 (1.8-fold increase, p = 0.0036). Interestingly, the percentage of C. concisus UNSWCD that invaded LS174T cells also increased 1.8-fold from day 0 to day 2, although overall the levels of invasion were significantly lower than those observed for Caco-2 cells (~50-fold lower). One possible reason for the increased adherence with increase in mucus is the chemoattractive properties of mucins. For example, Hugdahl et al. have reported positive chemotaxis by C. jejuni towards mucin [25]. Alternatively, C. concisus may bind to mucins to help facilitate adhesion to LS174T cells. Interestingly, a study by McAuley et al. has shown that C. jejuni binds to oligosaccharide ligands on the surface of mucins using adhesins [26]. Collectively, the increase in adherence of C. concisus to the surface of LS174T cells may be due to the interplay between the chemoattractive property of mucins along with the ability of C. concisus to bind to mucin structures.

While the observed increase in invasion may have resulted from increased adherence, another explanation could be that exposure to mucus also modulates the invasive potential of C. concisus. For example, Tu et al. have reported that exposure of C. jejuni to mucin results in the upregulation of the Campylobacter invasion antigen ciaB that allows the internalisation of C. jejuni into mammalian cells, and that cadF which encodes fibronectin binding protein facilitates binding to host fibronectin and allows adhesion to the surface of host cells [27]. These proteins are also encoded by C. concisus UNSWCD [28], and whilst studies involving the change in gene expression of C. concisus in the presence of mucus have not been conducted, the study in C. jejuni raises the possibility that C. concisus UNSWCD gene regulation may also occur in the presence of mucus.

Previous studies by our group, which examined the attachment of C. concisus strains to Caco-2 cells using ScEM have shown that several C. concisus strains adhere in an aggregative pattern on the surface of Caco-2 cells [5, 12]. This finding led us to postulate that this aggregation may actually represent biofilm formation. To determine the ability of the eight C. concisus strains to form biofilms, the method of Gunther IV et al.[16] was optimised. All C. concisus strains showed the formation of biofilms (Figure 5). There was no significant difference in the amount of biofilm among the eight strains (F_7,20 = 0.30, p = 0.94).

Given that biofilm formation is reported to assist in bacterial survival, colonisation and protection from host immune responses and antibacterial therapies [29], the ability of C. concisus to form biofilms may represent an important virulence mechanism in relation to its pathogenesis and transmission. Interestingly, based on an observation that an increased level of biofilm formation occurs in C. jejuni when incubated under aerobic conditions, Reuter et al. suggested that biofilms may aid the survival of C. jejuni in the environment, and that this adaptation may contribute to its zoonotic lifestyle [30]. While currently it is unknown whether C. concisus is a zoonotic infection, the possibility has been raised by the detection of C. concisus in cats, dogs, chickens and cattle [3]. In relation to survival within the host, our observation of biofilm like aggregations following infection of intestinal cells with C. concisus raises the possibility that like uropathogenic E. coli UTI89, biofilm formation may be a critical factor contributing to a persistent infection in the human host [31]. Given that within the intestinal tract C. concisus is continually subjected to peristalsis and mucus turn-over, the ability to produce biofilms may allow the bacterium to remain within its niche.