Transient internalization of Campylobacter jejuni in Amoebae enhances subsequent invasion of human cells.

The ubiquitous unicellular eukaryote, Acanthamoeba, is known to play a role in the survival and dissemination of Campylobacter jejuni. C. jejuni is the leading cause of bacterial foodborne gastroenteritis world-wide and is a major public health problem. The ability of C. jejuni to interact and potentially invade epithelial cells is thought to be key for disease development in humans. We examined C. jejuni grown under standard laboratory conditions, 11168HCBA with that harvested from within Acanthamoeba castellanii (11168HAC/CBA) or Acanthamoeba polyphaga (11168HAP/CBA), and compared their ability to invade different cell lines. C. jejuni harvested from within amoebae had a ~3.7-fold increase in invasiveness into T84 human epithelial cells and a striking ~11-fold increase for re-entry into A. castellanii cells. We also investigated the invasiveness and survivability of six diverse representative C. jejuni strains within Acanthamoeba spp., our results confirm that invasion and survivability is likely host-cell-dependent. Our survival assay data led us to conclude that Acanthamoeba spp. are a transient host for C. jejuni and that survival within amoebae pre-adapts C. jejuni and enhances subsequent cell invasion. This study provides new insight into C. jejuni interactions with amoebae and its increased invasiveness potential in mammalian hosts.

Introduction

is the leading cause of bacterial food-borne gastroenteritis worldwide [1]. However, it is puzzling that this microaerophile bacterium that is incapable of growing under atmospheric conditions [2] can be so prevalent in the environment and be responsible for such widespread disease in humans. It is still unclear how this pathogen survives and thrives in the environment outside its warm-blooded avian and mammalian hosts. Several studies have reported survival of within free-living protozoa, such as amoebae, as a mode of survival and persistence in the environment [3-5].

Free-living amoebae are widely distributed in the environment and have been isolated from a range of sources including freshwater, seawater, soil, dust and food sources [6, 7]. Amoebae, including Acanthamoeba spp., have long been investigated for their role to phagocytose bacteria as prey, to serve as a vector or a host to pathogenic bacteria, including spp., spp., spp. and spp. [5, 8, 9]. Acanthamoeba spp. as a vector and/or a host include aiding in bacterial survival with or without multiplication. Growth and multiplication of bacteria can lead to subsequent lysis of the amoebae and release of bacteria [9]. This ‘Trojan horse’ principal for bacterial pathogens has been linked to disease outbreaks in contaminated water [10] and food sources [11].

The phenomenon of C. jejuni survival within amoebae has been previously studied without any definitive insight into its role in human disease. Here, we show that survival within amoebae pre-adapts and enhances subsequent invasion to mammalian cells, which could lead to increased disease in mammalian hosts including humans.

Methods

Strains and cultures

Bacteria were stored using Protect bacterial preservers (Technical Service Consultants, Heywood, UK) at − 80 °C. strains were streaked on blood agar (BA) plates containing Columbia agar base (Oxoid) supplemented with 7 % (v/v) horse blood (TCS Microbiology, UK) and Campylobacter Selective Supplement (Oxoid), and grown at 37 °C in a microaerobic chamber (Don Whitley Scientific, UK), containing 85 % N2, 10 % CO2, and 5 % O2 for 48 h. strains were grown on CBA plates for a further 16 h at 37 °C prior to use.

Acanthamoeba castellanii CCAP 1501/10 and Acanthamoeba polyphaga CCAP 1501/14 [Culture collection of Algae and protozoa (Scottish Marine Institute)] were grown to confluence at 25 °C in 75 cm2 tissue culture flasks containing peptone yeast and glucose (PYG) media [12]. Amoebae were harvested by scraping the cells into suspension, and viability was determined by staining with trypan blue and counting by a haemocytometer using light microscopy.

invasion and survival assay

11168 h, a derivative of the original sequence strain NCTC 11168 was used in this study. cells were either grown on CBA agar (11168HCBA) as described above or harvested after intracellular survival in A. castellanii (11168HAC/CBA) or A. polyphaga (11168HAP/CBA) (Table 1), before invasion of epithelial cells and re-invasion of amoebae. Briefly, a large-scale invasion of Acanthamoeba spp. was carried out in a 150 cm2 tissue culture flask (Falcon), a monolayer of approximately 106 amoebae was infected with at a m.o.i. of 200 : 1 for 3 h at 25 °C in PYG media (although it is possible that density can impact amoebae predation through density-dependent inhibition, the m.o.i. of 200 : 1 was chosen to allow maximal internalization of bacteria). The monolayer was washed 3× with 25 ml of PYG media and incubated for 2 h in 25 ml of PYG media containing 100 µg ml−1 of gentamicin. cells were harvested by scraping the amoebae into suspension and centrifuged for 10 min at 350 to pellet the bacteria and amoebae. Supernatant was discarded and the pellet was suspended in 1 ml of distilled water containing 0.1 % (v/v) Triton X-100 for 10 min at room temperature to lyse the amoebae and release bacteria cells [13], lysis was confirmed by light microscopy. The suspension was then centrifuged for a further 10 min at 4000 , the resultant pellet was resuspended in 1 ml PBS and 200 µl aliquots of this suspension were plated on CBA plates and incubated microaerobically for 48 h at 37 °C to ensure recovery of enough bacteria. For invasion assay, the experiment was performed as described above, with an additional step of serial dilutions and plating out bacteria for c.f.u. invasion of Acanthamoeba spp. was confirmed after 3 h infection with 11168HGFP strain Fig. 1. Invasion assay was performed as described above using strain 11168HGFP that expresses a green fluorescent protein (GFP) that is under the control of porA promoter as described previously [14]. Laser scanning microscopy (Axion) was used to visualize internalized bacteria at objective 63×/1.4 (oil).

Table 1.

Representative strains used in this study

Strain	Description	Multi locus sequence type	Reference
11168 h	A hyper-motile derivative of the original sequence strain NCTC 11168 that shows higher levels of caecal colonization in a chick colonization model	ST-21	Karlyshev et al. [38], Jones et al. [39],
81–176	Highly virulent and widely studied laboratory strain of C. jejuni . MLST	ST-42	Korlath et al. [40],
12 912	Ox liver portion isolate	HS Type-50	Gundogdu et al. [41],
M1	A rarely documented case of direct transmission of C. jejuni from chicken to a person, resulting in enteritis	ST-45	Friis et al. [42],
81116	Genetically stable strain which remains infective in avian models	ST-283	Wassenaar et al. [43],
RM1221	A chicken isolate with unique lipooligosaccharide and ability to colonize chicken skin	ST-354	Fouts et al. [44],
11168H CBA	Grown under standard laboratory conditions		This study
11168H AC/CBA	Strain 11168 h harvested after survival in Acanthamoeba castellanii		This study
11168H AP/CBA	Strain 11168 h harvested after survival in Acanthamoeba polyphaga		This study
11168H GFP	Strain 11168 h expressing green fluorescent protein		Jervis et al. [14],

Fig. 1.

Laser scanning confocal microscopy (LSM). Observation of GFP labelled strain 11168GFP within (a) A. polyphaga and (b) A. castellanii after 3 h infection. is observed as green fluorescent within the amoebae cells. A monolayer of amoebae (106) in a 35 mm imaging dish (Ibidi) were infected with to a m.o.i. of 200 and incubated for 3 h aerobically at 25 °C, cells were washed 3× before imaging. Laser scanning microscopy (Axion) was used to image the cells at objective 63×/1.4 oil. 11168GFP strain was constructed as described previously [14].

Highly virulent and widely studied laboratory strain of . MLST

A rarely documented case of direct transmission of from chicken to a person, resulting in enteritis

For survival assay, the experiment was performed as described above with the following modifications; cells were incubated in the respective media containing 10 µg ml−1 of gentamicin, at each indicated timepoint, the cells were washed three times with PBS to remove residual antibiotics and lysed to plate for enumeration as described above.

Human cell lines and culture conditions

T84 (human carcinoma cell line) and Caco-2 (human colorectal adenocarcinoma cells) were grown in Dulbecco’s modified Eagle’s medium and Ham’s F-12 (DMEM/F-12) supplemented with 10 % FBS and 1 % non-essential amino acid. The monolayers, ~105, were seeded in a 24-well tissue culture plates and were grown up to ~106 in a 5 % CO2 atmosphere and were then infected with 11168HCBA, 11168HAC/CBA or 11168HAP/CBA at m.o.i. of 200 : 1 for 3 h as described previously [15]. The monolayers were then washed three times with PBS, incubated in DMEM containing gentamicin (100 µg ml−1) for 2 h at 37 °C to kill extracellular bacteria, the cells were then washed 3× with PBS and then lysed with 0.1 % (v/v) Triton X-100. The cell lysates were serially diluted and plated onto blood agar plates and incubated for 48 h before colonies were enumerated. Experiments were performed in triplicates of three biological replicates. To normalize the numbers of intracellular bacteria recovered, and to confirm c.f.u. consistency of cells, the initial bacterial inoculum was always plated.

Galleria mellonella infection model

G.mellonella larvae (LiveFoods) were kept on wood chips at room temperature. Experiments were performed as previously described [16]. Briefly, 11168HCBA, 11168HAS/CBA or 11168HAP/CBA C. jejuni were suspended in PBS to give OD600nm 0.1 and 10 µl volume of this suspension (~106) was injected into the right foremost leg of the G. mellonella larvae by microinjection (Hamilton) and incubated at 37 °C. For each strain, ten larvae with similar weight were used per replicate. Mortality was observed at 24 h intervals for 72 h.

Statistical analysis

All experiments presented are at least three biological replicates. All data were analysed using Prism statistical software (version 9, GraphPad Software). Values were presented as standard deviation and variables were compared for significance using Student's t-tests to obtain P-values unless otherwise stated.

Results and discussion

Survival of in Acanthamoeba increases subsequent invasion

cells that had survived within Acanthamoeba spp. were tested for their ability to invade T84, Caco-2, A. castellanii and A. polyphaga cells, Fig. 2. Invasion of all cells (human cell lines and amoebae cells) increased significantly (P<0.05) with bacteria harvested from amoebae, 11168HAC/CBA or 11168HAP/CBA, more than with 11168HCBA .

Fig. 2.

11168 h harvested from Acanthamoeba spp. (a) invasion of T84 cells; (b) Caco-2 cells; (c) A. castellanii; and (d) A. polyphaga. Invasion of strain 11168HCBA, 11168HAC/CBA or 11168HAP/CBA were determined by infection of the cell lines for 3 h and enumerated after 100 µg ml−1 gentamicin treatment and lysis of cell. Data is presented as percentage of the original inoculum (~108). Error bars represent sd from three independent experiments. *P <0.05, **P<0.01, ***P<0.001.

We observed that bacteria passaged through amoebae had an increased capacity to invade human epithelial cells compared to non-passaged 11168HCBA with a relative increase in invasion of ~3.7-fold for 1168HAC/CBA and ~2.8-fold for 11168HAP/CBA in T84-cells (Fig. 2a). In Caco-2 cells, we also observed a significant increase in relative invasion, although to a lesser extent, with ~2.4-fold for 11168HAC/CBA and ~1.6-fold for 11168HAP/CBA(Fig. 2b). The levels of invasion varied dependent on the cell type, however, it was always significantly higher than that of 11168HCBA bacteria. These results indicated that cells that survived intracellularly in amoebae undergo priming and are more invasive, this increased invasiveness could lead to a more severe disease in humans. This priming adaptive response has been reported in bacteria [17], including adaptive tolerance to low pH [18]. To our knowledge, this is the first study that shows survival within amoebae enhances subsequent invasion of human epithelial cells. This provides novel insight into the interactions of with protists.

To determine whether this increase in invasion would also be observed in amoebae, we performed re-infection of the Acanthamoeba spp. with 11168HCBA, 11168HAC/CBA or 11168HAP/CBA . Interestingly, re-infection of Acanthamoeba spp. showed a significant increase in invasion with bacteria passaged through amoebae compared to non-passaged bacteria, 11168HCBA. A. castellanii invasion showed significant increase of ~9.2-fold with 11168HAC/CBA and ~11-fold with 11168HAP/CBA C. jejuni (Fig. 2c). The same trend was observed for A. polyphaga re-infection, ~5.0-fold increase in invasion with 11168HAC/CBA and ~7.0-fold with 11168HAP/CBA C. jejuni (Fig. 2d). To ensure that the observed increased in invasion was not caused by resistance to gentamicin; sensitivity tests were performed, and revealed no significant (P<0.05) differences between 11168HCBA and amoebae recovered (data not shown).

The dramatic invasion of 11168HAC/CBA and 11168HAP/CBA within amoebae may be attributed to the higher background of non-specific uptake of bacteria by amoebae including non-pathogenic bacteria. However, given that increase in invasion was observed across all the cell types used in this study, it is more likely that amoebae recovered bacteria have pre-adapted to subsequently invade and survive at a greater rate than that of 11168HCBA bacteria. This finding may be unsurprising, since the symbiotic relationship between amoebae and bacteria has been thought to pre-adapt intracellular microorganisms to survive in other cells including human macrophages [19]. This eco-evo hypothesis [20] is observed with spp. and Legionella pneumophilia, which use similar strategies to interact with various hosts cells and most probably evolved over millions of years during bacterial interactions with primitive unicellular eukaryotes [19, 21, 22]. This greater invasiveness of within amoebae could facilitate longer survival which may lead to an increased ability of to survive and subsequently transmit to humans from environmental sources.

Acanthamoeba spp. are a transient host for

There have been conflicting reports on intracellular multiplication of within amoebae [5, 23]. Whilst some have reported is capable of multiplying within Acanthamoeba spp. [3, 12, 23] others have been unable to observe intracellular multiplication [4, 24]. In our model, we were unable to detect intra-amoebae multiplication of 11168 h and 81–176 . strains (Fig. 3) in different media, PYG (Fig. 3a and b) or in brucella broth (a highly nutritious media used to enrich growth and can sustain Acanthamoeba spp.) (Fig. 3c and d), aerobically at 37 °C as previously described [12, 23]. We found that cells were undetected at 72 h post-infection. These findings led us to conclude that amoebae, at least, the Acanthamoeba spp. used in this study are a transient host for . A previous study presented a hypothetical model suggesting that in the environment multiplies within amoebae, potentially bursting out and re-invading neighbouring amoebic cells [12]. The experimental data presented here partly confirms their hypothesis, however, since we did not observe intracellular multiplication of , we propose that it is more likely that in the environment outside the host, would invade Acanthamoeba cells briefly, but frequently, with increased invasion efficiency. This strategy increases the chances of this bacterium to be transmitted to warm-blooded avian and mammalian hosts.

Fig. 3.

Long-term survival of with Acanthamoeba spp. strains 11168 h or 81–176 survival in (a) A. polyphaga; (b) A. castellanii in PYG media, and (c) A. polyphaga; (d) A. castellanii in brucella broth at 37 °C in aerobic conditions. Amoebae were lysed for enumeration of live bacteria at 24 h interval for 72 h after 10 µg ml−1 gentamicin treatment. Data is presented as percentage of the original inoculum (~108). Error bars represent sd from three independent experiments. Two-way ANOVA multiple comparison was used to test for significance; **P<0.01, ***P<0.001, ****P <0.0001. ND=no bacteria detected.

survival in amoebae is not cytotoxic to G. mellonella larvae infection model

Increased invasion of the different cell lines prompted us to examine whether passaged through amoebae would be more cytotoxic for G. mellonella larvae compared to 11168HCBA bacteria. Using this surrogate infection model, we did not observe any significant cytotoxic differences between 11168HCBA, 11168HAC/CBA or 11168HAP/CBA C. jejuni towards G. mellonella larvae Fig. 4.

Fig. 4.

The effect of 11168HCBA, 11168HAC/CBA and 11168HAP/CBA in the G. mellonella infection model. G. mellonella larvae were injected with a 10 µl inoculum of 106 c.f.u. by microinjection in the right foremost leg. PBS was used as control. Larvae were incubated at 37 °C, with survival and appearance recorded after 72 h. For each experiment, ten G. mellonella larvae were infected, and experiments were repeated in triplicate. Error bars represent sd. Cytotoxicity was monitored at 24 h intervals for 72 h.

Although this infection model has been previously used to determine cytotoxicity [15, 16, 25], we cannot rule out that significant differences may be observed in avian and mammalian host cells. A previous study by Snelling et al. showed increased chicken colonization after 7 days post-infection with intra-amoebae [26]. It would be worth studying cytotoxicity in chicken colonization/infection models.

A. castellanii supports greater survival of

In the environment, would encounter multiple species of amoebae. We examined the invasiveness and survivability of six strains; 11168 h, 81–176, 12912, M1, 81116 and RM1221 within A. castellanii and A. polyphaga Fig. 5. These strains were selected because of their diverse genetic backgrounds and sources, thus our observations are more representative of the species (Table 1).

Fig. 5.

Quantification of strains' survival within Acanthamoeba species. strains 11168 h, 81–176, 12 912, M1, 81 116 and RM1221 in A. polyphaga and A. castellanii. Quantification of intracellular bacteria was determined by viable counts at 3 h and 24 h after amoebae infection at 25 °C in aerobic conditions. Data is presented as percentage of the original inoculum (~108). Error bars represent sd from three independent experiments. Two-way ANOVA multiple comparison was used to test for significance; *P <0.05, **P<0.01, ***P<0.001, ****P <0.0001. ND=no bacteria detected.

We observed a general trend of a greater survival rate of strains within A. castellanii compared to A. polyphaga in PYG at 25 °C under aerobic conditions. Our results show differences in invasiveness and survival capabilities between the range of the strains tested. Interestingly, these results are similar to previous studies that correlated invasiveness and survivability of is both bacterial strain and host-cell-dependent [27-30]. In our amoebae model, the greater survivability of within A. castellanii seems to be a host susceptibility factor rather than being bacterial induced. This phenomenon was also reported in other bacteria, where greater levels of invasion and intracellular survivability of was observed in A. castellanii [31] compared to other Acanthamoeba spp. A recent review proposed that invasion and intracellular occurrences of microbes within amoebae is dependent on the genotype of the host [32]. Based on 18S RNA sequence studies, A. castellanii is from the T4 genotype [33] whilst A. polyphaga is from the T2 genotype [34], although how host genotype plays a role in intracellular survival remains unknown.

The mechanisms of survival within amoebae has been compared to that of macrophages, and a review by Vieira et al. [5], predicted various factors that could utilize to invade and survive within amoebae cells. Although survivability of in our model is most likely host-dependent, it would be intriguing to determine the bacterial factors involved. Unlike other enteric pathogens, the genome is relatively small at 1.6 Mb [35, 36], and this bacterium would plausibly use the same factors to invade and survive within amoebae as it would for avian and mammalian host cells. Therefore, elucidating these factors may give new insights into infection, which compared to other enteric pathogens is poorly understood. Future studies could use molecular tools such as genome-wide transposon mutant libraries of multiple strains like those used by de Vries et al. [37], to help improve our understanding of infection. This study communicates our observations that sets the scene for future work to uncover the mechanisms of interactions with amoebae and may provide new insights into the persistence of this problematic pathogen.

In conclusion, given the diverse selection of strains tested, this consistent data provides insight into a natural phenomenon that will be important in survival, transmission and infection. Additionally, the comparison between the two species of amoebae gives valuable information for future work. We propose that Acanthamoeba spp. are a transient host for and survival within this ‘Trojan horse’ environment subsequently increases invasiveness.