Introduction

Campylobacter jejuni asymptomatically colonizes chicken gut, but it is one of the prevalent foodborne pathogens in the developed countries. Chicken is the main vector to transmit C. jejuni to host humans, and the bacterium was detected in 20–56% of chicken breast meat from various brands in 20131. More than 12.2 campylobacteriosis cases per 100,000 population were recorded in 2017 in the USA. That number represented a 4.3% increase compared to 2016 and was higher than the combined incidences by the following 8 bacterial pathogens2. A total of 1.3 million individuals are afflicted by the disease, resulting in 76 deaths every year3. Furthermore, C. jejuni infection causes severe post-infectious complications, including arthritis4, the neurodegenerative disorder Guillain-Barré Syndrome5, Irritable Bowel Syndrome6, and Inflammatory Bowel Diseases (IBD)7,8.

Although reducing chicken meat Campylobacter counts by 2 logs is estimated to decrease a 30-fold in human campylobacteriosis9, current strategies to reduce campylobacteriosis are not effective enough10. This was evidenced by the relative consistent rate of campylobacteriosis incidences in the Morbidity and Mortality Weekly Report at CDC infectious disease database from January, 1996 to August, 201611. Interestingly, it has been largely overlooked whether any important events happen when C. jejuni is transmitted from vector chickens to host humans. Foodborne bacterial pathogen contamination is mainly determined by enumerating the pathogens using bacterial culture or PCR method without assessing the possibility of bacterial infection ability alteration after the bacterial animal colonization12. Campylobacter jejuni is present in chicken farms for many years, and the bacterium is transmitted through many batches of chickens13. Consequently, it remains elusive whether C. jejuni chicken transmission impacts its subsequent colonization and campylobacteriosis induction.

Il10−/− mouse infection model has recently been successfully developed to mimic human campylobacteriosis in various labs14,15,16,17,18,19. Specific pathogen free (SPF) Il10−/− mice resisted against C. jejuni-induced colitis, while the mice were susceptible to campylobacteriosis after treated with anaerobe-killing antibiotic clindamycin20. Using HPLC/MS analysis, we found that clindamycin depleted all secondary bile acids, particularly deoxycholic acid (DCA)20. Furthermore, anaerobe metabolite, DCA, was able to prevent and treat C. jejuni-induced colitis in ex-GF mice20. We also found that DCA resisted against chicken colonization of C. jejuni human clinical isolate 81–176 and chicken isolate AR10121. The microbiota composition at the cecal of the infected birds transplanted with DCA-modulated microbiota was assessed at phylum level using real-time PCR, and the results showed that the microbiota compositions were different21. Campylobacter jejuni motility and adherence to cells aren’t changed in the presence of DCA, although DCA induced virulence genes ciaB, cmeABC, dccR, and tlyA22. However, it is unclear whether DCA regulates C. jejuni chicken transmission and subsequent induction of campylobacteriosis.

In this study, we found that chicken-transmitted C. jejuni (passage 1, Cj-P1) induced more severe intestinal inflammation in Il10−/− mice compared to the non-transmitted bacterium (passage 0, Cj-P0), while Cj-P1-DCA-Anaero induced less colitis, bacterial invasion, and inflammatory gene expression in Il10−/− mice compared to Cj-P1. Thus, C. jejuni transmitted from birds raised in various conditions could behave differently in inducing enteritis. The outcome of this study will provide key information about the interplay between chicken microbiome, C. jejuni transmission, and host susceptibility and response, and could help the development of new prevention strategies against foodborne pathogens.

Material and methods

Campylobacter jejuni strains isolated from infected birds

Campylobacter jejuni strain AR101 (Cj-P0) was isolated from experimental chickens at Dr. Billy Hargis’s laboratory at the University of Arkansas at Fayetteville and the bacterium was used in our recent report21. In the report, C. jejuni in the cecal digesta of the 28 days of age birds infected with Cj-P0 was cultured on C. jejuni selective blood plates with five antibiotics (cefoperazone, cycloheximide, trimethoprim, vancomycin and polymyxin B) for 48 h at 42 °C using the GasPak system (BD), and the isolated C. jejuni was named as Cj-P1 (C. jejuni passage 1). In the report, C. jejuni in the cecal digesta of 28 days of age birds fed DCA and infected with Cj-P0 was cultured on the selective C. jejuni plates and was named as Cj-P1-DCA. Cecal digesta from uninfected 28 days of age birds fed with DCA diet was collected and cultured on Brain Heart Infusion (BHI) plates under anaerobic or aerobic conditions, and the isolated bacteria were named anaerobic-microbiota (DCA-Anaero) or aerobic-microbiota (DCA-Aero), respectively. Campylobacter jejuni in cecal digesta of 28 days of age birds colonized with DCA-Anaero or DCA-Aero and infected with Cj-P0 was cultured on the selective C. jejuni plates and was named as Cj-P1-DCA-Anaero or Cj-P1-DCA-Aero, respectively. The microbiota at the cecal of those birds was assessed at phylum level using real-time PCR, and the results showed that the microbiota compositions were different21. The AR101 strains were routinely grown on the selective C. jejuni plates and examined under microscopy for size, morphology and motility”.

Campylobacter jejuni motility assay

Cj-P0, Cj-P1, Cj-P1- DCA, Cj-P1-DCA-Anaero or Cj-P1-DCA-Aero was grown on the selective plates, collected, and diluted to an optical density at 600 nm (OD600) of 1. Each bacterium of 1 μl was then stabbed into a 0.4% agar Brain Heart Infusion (BHI) plate without antibiotic cocktail. The less dense agar facilitated C. jejuni to easier move inside the agar and the bacterium formed a halo of growth around the inoculation point. Following microaerobic growth at 42 °C for 24 h, the radius of the ring was calculated relative to that of Cj-P0. Cj-P0 was grown on each plate to control plate-to-plate variation. Experiments were performed in triplicate and repeated three times.

Mouse experiment

Animal experiments were performed in accordance with the Animal Research: Reporting of In Vivo Experiments (https://www.nc3rs.org.uk/arrive-guidelines). The experiments were approved by the Institutional Animal Care and Use Committee of the University of Arkansas. For C. jejuni infection experiments, cohorts of 5 to 9 SPF C57BL/6 Il10 −/− mice were orally gavaged daily with antibiotic clindamycin (Sigma-Aldrich) at 67 mg/kg body weight (BW) for 7 days. One day after the last antibiotic treatment, the mice were gavaged with a single dose (109 CFU/mouse) of Cj-P0 (5 mice), Cj-P1 (9 mice), Cj-P1-DCA-Anaero (5 mice), Cj-P1-DCA (9 mice), Cj-P1-DCA-Aero (5 mice), respectively as described above. Mice were followed clinically for evidence of diarrhea, failure to thrive, and mortality. At the end of experiments at 8 days post infection, tissue and stool samples from mouse colon were collected for protein, RNA, histology, and culture assay. For live C. jejuni counting, MLN and spleen were aseptically resected. Colon tissue was opened, resected, and washed three times in sterile PBS. Colonic luminal content (stool) was also collected. The freshly collected tissues and stool were weighed, homogenized in PBS, serially diluted, and cultured on selective C. jejuni plates supplemented with 5 antibiotics cocktail (cefoperazone, cycloheximide, trimethoprim, vancomycin, and polymyxin B) for 48 h at 37 °C using the GasPak system (BD Biosciences) as described before21. Campylobacter jejuni colonies were counted, and data were presented as CFU per gram tissue or stool. Histopathological images were acquired using a Nikon TS2 fluorescent microscope23. Intestinal inflammation was scored using a scoring system from 0–4 as showed before20,24.

Fluorescence in situ hybridization (FISH)

C. jejuni at intestinal tissue sections was visualized using FISH assay as previously described24. Briefly, tissue sections were deparaffinized and hybridized with the FISH probe for up to 48 h. The tissue sections were then washed, and the mammalian nuclei were visualized by staining with DAPI. The tissue sections were then imaged using the Nikon TS2 fluorescent Microscope system.

Immunohistochemistry (IHC)

Neutrophils in intestinal tissues were detected using anti-myeloperoxidase (MPO) IHC analysis as described previously25. Briefly, intestinal tissue sections were deparaffinized, blocked, and incubated with an anti-MPO antibody (1:400; Thermo Scientific) overnight at 4 °C. After incubation with anti-rabbit biotinylated antibody and three times washing, the tissue sections were incubated with avidin/biotin complex (Vectastain ABC Elite Kit, Vector Laboratories). After three times washing, the tissue sections were added with diaminobenzidine (Dako) within 2 min. The mammalian nuclei were visualized by staining with hematoxylin (Fisher Scientific). The tissue sections were then imaged using the Nikon TS2 Microscope system.

Real-time RT-PCR

Total RNA from colonic tissue was extracted using TRIzol (Invitrogen). cDNA was prepared using M-MLV (Invitrogen). mRNA levels of proinflammatory genes were determined using the SYBR Green PCR Master Mix (Bio-Rad) on a Bio-Rad 384-well Real-Time PCR System and normalized to Gapdh. The primer sequences of the genes Gapdh, 17a, Il1β, and Cxcl1 were reported before20.

White blood cell isolation and migration assay

Blood was collected from Il10−/− mice and the red blood cells were lysed in the buffer of 8.3 g/l NH4Cl in 0.01 M Tris–HCl buffer of pH 7.5. The collected white blood cells were resuspended in 1% FBS RPMI 1640 medium. Cells were plated at 104 per insert in 24‐well Transwells (Corning) with 3 μm pores and incubated at 37 °C and 5% CO2. Cj-P0, Cj-P1, Cj-P1-DCA, Cj-P1-DCA-Aero, and Cj-P1-DCA-Anaero at 105 CFU/well were inoculated into the bottom wells. White blood cells migrated into the bottom well were imaged and counted one hour later using the Nikon TS2 Microscope system ,similar to previously described25. Cells in six fields per well were counted.

Statistical analysis

Values were displayed as mean ± standard error of the mean as reported before20. Differences between groups were analyzed using the nonparametric Mann–Whitney U test with Prism 7.0 software. Experiments were considered statistically significant if P value was < 0.05.

Ethics approval and consent to participate

All animal protocols were approved by the Institutional Animal Care and Use Committee of the University of Arkansas at Fayetteville.

Results

Chicken-transmitted C. jejuni induced more severe intestinal inflammation

To address whether the asymptomatic C. jejuni transmission in chickens influenced its induction of intestinal inflammation in susceptible hosts such as human or Il10−/− mice, we cultured infected bird cecal content, isolated C. jejuni, and labeled it as passage 1 or Cj-P1. We reasoned that C. jejuni transmitting through chickens altered virulence and would induce worse intestinal inflammation. To examine this hypothesis, SPF Il10−/− mice were orally gavaged daily with antibiotic clindamycin for 7 days. The mice were then infected with Cj-P0 or Cj-P1 with a single oral gavage dose of 109 CFU/mouse. Interestingly, after 6 days post-infection, Il10−/− mice infected with chicken-transmitted Cj-P1 showed clinical sign of enteritis. After 8 days post infection, Il10−/− mice infected with Cj-P1 demonstrated severe morbidity and bloody diarrhea and the experiment had to be terminated. At cellular level, Cj-P0 induced mild intestinal inflammation in the Il10−/− mice, shown as crypt hyperplasia, mild goblet cell depletion, and mild immune cell infiltration into lamina propria (Fig. 1A). Remarkably, Cj-P1 induced more severe intestinal inflammation and higher histopathological score compared to Cj-P0 ( 2.8 vs. 0.8, P < 0.05) , shown as crypt abscesses, extensive immune cell infiltration and massive goblet depletion (Fig. 1A,B). These results suggest that C. jejuni transmission through chickens enhances infection capacity and induces more severe intestinal inflammation.

Figure 1
figure 1

Chicken-transmitted C. jejuni (Cj-P1) induced more severe colitis compared to AR101 (Cj-P0). Cohorts of 5–9 SPF Il10−/− mice were orally gavaged daily with clindamycin for 7 days. The mice were then gavaged with a single dose of 109 CFU/mouse C. jejuni of Cj-P0 (5 mice) and Cj-P1 (9 mice) and were euthanized 8 days post-infection. (A) H&E staining showing representative intestinal histology of C. jejuni-induced colitis in Il10−/− mice. (B) Quantification of histological intestinal damage score. *, P < 0.05. Data represent means ± SEM. Scale bar is 200 μm. Results are representative of 3 independent experiments.

Full size image

Chicken-transmitted Cj-P1 aggressively invaded colon tissue

Campylobacter jejuni colonization and invasion are essential for its successful induction of campylobacteriosis26. To investigate the mechanism of how Cj-P1 induced more intestinal inflammation compared to Cj-P0, we next evaluated C. jejuni colonization and invasion in colon. Colon content and tissue were weighed, homogenized, serially diluted, and cultured on selective plates. Notably, the luminal colonization level of C. jejuni Cj-P1 was significantly denser in mouse colon compared to that of mice infected with Cj-P0 (3.24 × 107 vs. 6.22 × 106 CFU/g stool, P = 0.049) (Fig. 2A). We then examined C. jejuni invasion into colon tissue. In consistent with luminal C. jejuni colonization levels, Cj-P1 significantly invaded into colon tissue compared to Cj-P0 (6.47 × 105 vs. 8.12 × 104 CFU/g tissue, P = 0.016) (Fig. 2B). To further detect the C. jejuni spatial distribution in colon tissue, we visualized C. jejuni DNA using fluorescence in situ hybridization (FISH) and fluorescence microscopy imaging. Notably, while Cj-P1 DNA was detected widely in the inflamed crypts and the lamina propria section of the mouse intestine, Cj-P0 was seldomly detectable in the mouse colon (Fig. 2C). These results indicate that Cj-P1 gains virulence ability to invade more aggressively into colonic crypts.

Figure 2
figure 2

Cj-P1 aggressively invaded into colon tissue. Cohorts of mice were infected as in Fig. 1. Samples from colon content and tissue were collected and aseptically cultured. (A) C. jejuni count in colon content. (B) C. jejuni count in colon tissue. (C) Clustered C. jejuni (red dots) in colonic section of infected mice was detected using FISH. Scale bar is 10 μm. *, P < 0.05. Data represent means ± SEM. Results are representative of 3 independent experiments.

Full size image

Transmitted Cj-P1 induced severe crypt abscesses in colon

Since C. jejuni infection induced strong intestinal inflammation in Il10−/− mice, we examined the histopathology of the infected mice at higher magnification to have more detailed assessment. Notably, Cj-P1 depleted the majority of goblet cells and induced massive immune cell infiltration into lamina propria and numerous crypt abscesses compared to Cj-P0 (Fig. 3A). Since crypt abscesses were observed in histopathology slides, we then detected the neutrophils by targeting neutrophil marker myeloperoxidase (MPO) using immunohistopathology (IHC). As showed in Fig. 3B, Cj-P0 induced a few MPO positive cells into crypt and formed fewer crypt abscesses, whereas Cj-P1 showed stronger induction of crypt abscesses compared to Cj-P0. These results indicate that Cj-P1 gains virulence capacity to induce more infiltration of immune cells such as neutrophils.

Figure 3
figure 3

Cj-P1 strongly induced neutrophil infiltration. Cohorts of mice were infected as in Fig. 1. (A) Representative of intestinal high magnification histopathology of C. jejuni-induced colitis in Il10−/− mice. (B) Immunohistochemistry of myeloperoxidase positive neutrophils (brown dots). Yellow arrows indicate neutrophil accumulation in the crypt lumen and formation of crypt abscesses. *P < 0.05. Data represent means ± SEM. Scale bar is 20 μm. Results are representative of 3 independent experiments.

Full size image

Transmitted Cj-P1 induced stronger migration of immune cells in vitro

To understand how Cj-P1 induced more neutrophil infiltration into crypts compared to Cj-P0, we ran an in vitro immune cell migration assay similar to what we have described before25. Furthermore, we have reported that DCA and DCA modulated anaerobes (DCA-Anaero) prevented C. jejuni AR101 cecal colonization in chickens21, so we also isolated C. jejuni strains transmitted through chickens treated with DCA (Cj-P1-DCA), DCA modulated aerobe (Cj-P1-DCA-Aero), and DCA modulated anaerobe (Cj-P1-DCA-Anaero). Mouse white blood cells at 104 cells/well were placed on the top inserts and Cj-P0, Cj-P1, Cj-P1-DCA, Cj-P1-DCA-Aero, and Cj-P1-DCA-Anaero were inoculated in the well bottom. One hour after the inoculation of C. jejuni, white blood cells migrated into the well bottom were imaged and counted under microscope. Interestingly, without C. jejuni infection, only a few immune cells were observed at the bottom of the wells (Fig. 4A). Cj-P1 induced the strongest immune cell migrations. Notably, Cj-P1-DCA-Anaero and Cj-P1-DCA-Aero induced 95% and 71% less immune cell migration compare to Cj-P1 (Fig. 4A,B), while Cj-P1-DCA induced comparable immune cell migration. We also examined the chicken-transmitted C. jejuni motility. Notably, after passaging chicken for one time, Cj-P1 increased motility by 1.36 folds compared to Cj-P0, while dietary DCA increased comparable motility compared to Cj-P1 (Supple Fig. 1). Interestingly, Cj-P1-DCA-Anaero or Cj-P1-DCA-Aero decreased the motility by 38 and 56%, respectively, compared to Cj-P1. The results suggest that C. jejuni transmitted through chickens increases its virulence of motility and attraction of immune cell migration, while the DCA-Anaero and DCA-Aero reduce the virulence.

Figure 4
figure 4

Cj-P1 strongly induced immune cell migration compared to Cj-P1-DCA-Anaero. Peripheral white blood cells were isolated and plated in a Transwell system, and the cells’ migration in response to C. jejuni at the bottom well was visualized and enumerated. Cells in six fields per well were counted. (A) Representative light images of white blood cells migrated into bottom wells. Scale bar is 10 μm. (B) Quantitative measurements of migrated white blood cells. Different letter of a-b means p < 0.05. Data represent means ± SEM. Results are representative of 3 independent experiments.

Full size image

DCA-Anaero attenuated chicken transmitted Cj-P1 induction of campylobacteriosis

We then reasoned that Cj-P1-DCA-Anaero might induce less colitis in Il10−/− mice compared to Cj-P1. To examine this possibility, we ran another mouse infection experiment. SPF Il10−/− mice were orally gavaged daily with clindamycin for 7 days and then infected with Cj-P0, Cj-P1, and Cj-P1-DCA-Anaero. After an 8-day infection, in consistent with the previous observation, Cj-P1 induced more severe intestinal inflammation in Il10−/− mice compared to Cj-P0, showed as numerous crypt abscesses, extensive immune cell infiltration, and massive goblet depletion (Fig. 5A). Remarkably, Cj-P1-DCA-Anaero induced less intestinal inflammation and histopathological score compared to Cj-P1 (Fig. 5A,B). We also infected Il10−/− mice with Cj-P0, Cj-P1, Cj-P1-DCA, and Cj-P1-DCA-Aero. In line with the in vitro immune cell migration assay, comparable intestinal inflammation was induced in Il10−/− mice infected with Cj-P1 and Cj-P1-DCA (Supple Fig. 2). Surprisingly, Cj-P1-DCA-Aero also induced comparable colitis to Cj-P1, although Cj-P1-DCA-Aero induced fewer white blood cell migration in vitro.

Figure 5
figure 5

Cj-P1-DCA-Anaero induced less colitis compared to Cj-P1. Cohorts of 5–9 SPF Il10−/− mice were orally gavaged daily with clindamycin for 7 days. The mice were then infected with a single dose of 109 CFU /mouse C. jejuni of Cj-P0 (5 mice), Cj-P1 (9 mice), or Cj-P1-DCA-Anaero (5 mice) and were euthanized 8 days post-infection. (A) H&E staining showing representative intestinal histology of C. jejuni-induced colitis in Il10−/− mice. (B) Quantification of histological intestinal damage score. Scale bar is 200 μm. *P < 0.05. Data represent means ± SEM. Results are representative of 3 independent experiments.

Full size image

DCA-Anaero attenuated chicken transmitted Cj-P1 invasion

To investigate why Cj-P1-DCA-Anaero induced less intestinal inflammation, we then evaluated C. jejuni colonization and invasion in colon. In consistent with previous results, Cj-P1 colonized in colon was in the trend but not significant compared to Cj-P0 and Cj-P1-DCA-Anaero (Fig. 6A). Notably, Cj-P1 invades more than 100 folds in colon compared to Cj-P0, an effect attenuated by 99% in Cj-P1-DCA-Anaero (Fig. 6B). Further visualization of C. jejuni spatial invasion by FISH showed that Cj-P1 DNA was detected deeply in the inflamed crypts and the lamina propria, while Cj-P1-DCA-Anaero was mostly absent in the mouse colon (Fig. 6C). These results suggest that DCA-Anaero attenuates C. jejuni transmission-increased virulence on invasion.

Figure 6
figure 6

Cj-P1-DCA-Anaero colonized and invaded less compared to Cj-P1. Cohorts of mice were infected as in Fig. 5. Samples from colon content and tissue were collected and aseptically cultured. (A) C. jejuni count in colon content. (B) C. jejuni count in colon tissue. (C) Clustered C. jejuni (red dots) in colonic section of infected mice was detected using FISH. *, P < 0.05. Data represent means ± SEM. Scale bar is 10 μm. Results are representative of 3 independent experiments.

Full size image

DCA-Anaero attenuated chicken transmitted Cj-P1 induction of inflammatory response

Because of the reduced bacterial invasion and colitis in Cj-P1-DCA-Anaero compared to Cj-P1, we reasoned that the former strains would induce fewer inflammatory responses. To examine this possibility, colon tissue was collected, and RNA was extracted. Gene expression of inflammatory cytokines and chemokines were measured using real-time PCR. Cj-P0 induced inflammatory genes of Il17a, Il1β, and Cxcl1 at 10, 4, and 2 folds, respectively, compared to uninfected mice (Fig. 7A). Remarkably, Cj-P1 induced the gene expression by 19, 10, and 11 folds, respectively, compared to Cj-P0, which was attenuated by 86, 74, and 86%, respectively, by Cj-P1-DCA-Anaero. Since the three pro-inflammatory cytokines mediated immune cell activity, we then visualized inflammatory neutrophil distribution in colon by IHC of MPO. Notably, Cj-P1 induced strong infiltration of MPO positive neutrophil and crypt abscesses, whereas Cj-P1-DCA-Anaero barely elicited neutrophil migration and crypt abscesses (Fig. 7B). These results suggest that DCA-Anaero attenuates C. jejuni transmission-increased virulence on the induction of inflammatory responses.

Figure 7
figure 7

Cj-P1-DCA-Anaero induced less inflammatory response compared to Cj-P1. Cohorts of mice were infected as in Fig. 5. (A) Il17a, Il1β, and Cxcl1 mRNA accumulation was quantified using real-time PCR. (B) IHC representation of MPO-positive cells (brown dots) in the colon tissue of C. jejuni-infected mice. Lower panels are magnified images of the area shown in the upper panels (scale bar, 200 μm). *P < 0.05. Data represent means ± SEM. Scale bar is 200 μm. Results are representative of 3 independent experiments.

Full size image

Discussion

Although C. jejuni is a prevalent foodborne pathogen mainly transmitted from chickens, few approaches available to control the bacterial chicken colonization. Moreover, the microbiota and cellular events responsible for host resistance or susceptibility to C. jejuni infection remain largely elusive27,28. Foodborne C. jejuni is mainly monitored through enumerating the bacteria in food such as chickens. The limitation of the practice is the lack of understanding of the virulence change after chicken colonization/transmission. In previous studies, we infected chickens with C. jejuni AR101 (labeled as Cj-P0) and the chickens didn’t show any clinical signs and grew comparably to uninfected birds21. The result was consistent with previous researches that C. jejuni asymptomatically colonizes chickens as a commensal-like pathogen29. Interestingly, C. jejuni passaging through chickens increases colonization potential in chickens30 and mice31, however, it hasn’t been well studied if the asymptomatic transmission in chickens influenced C. jejuni induction of intestinal inflammation in susceptible hosts such as human or Il10−/− mice. In this study, we investigated how C. jejuni transmitted from birds raised in different husbandry influenced its virulence in subsequent infection using Il10−/− mice. The results reveal new insights regarding C. jejuni chicken transmission, pathogen virulence, and husbandry conditions.

One of the notable findings in this study was that chicken-transmitted C. jejuni (Cj-P1) increased virulence to induce stronger campylobacteriosis in mice, although chickens colonized with C. jejuni were healthy, shown with comparable body weight gain between uninfected and infected birds21. Campylobacter jejuni colonizes 95% flock of 20,000 chickens within 7 days after the initial one bird infected with the bacterium32. Transmitted C. jejuni increases phase-variable controlled flagellar33, which may contribute to its fast horizontal transmission rate in the flock. Furthermore, C. jejuni passaging through chicken reservoir promotes phase variation in specific contingency genes, and the populations with the variations colonize mice31. In consistent with these previous reports, in this study, the chicken-transmitted C. jejuni colonized mice with more number than the original pathogen Cj-P0. A new observation in this study is that beyond colonization ability increase, the chicken-transmitted C. jejuni also increased virulence to induce more severe colitis. At the host cellular level, the transmitted C. jejuni induced more immune cell infiltration in colon and increased immune cell migration in in vitro assay. Immune cell migration is one of the important steps in inducing campylobacteriosis25. Innate immune cells can directly sense bacterial cellular molecules with various receptors such as Toll-like receptors (TLR) and Nod-like receptors (NLR)34. It would be helpful to investigate in the future whether the chicken-transmitted C. jejuni had phase variations in cell surface virulence genes, such as lipooligosaccharide. In addition, it remains to be determined whether the chicken-transmitted C. jejuni increased virulence genes on mobility, growth, and toxin production. Together, these findings suggest that an equal count of C. jejuni may induce quite a different campylobacteriosis, and monitoring chicken-transmitted C. jejuni virulence is important for preventing foodborne campylobacteriosis.

The remarkable finding in this study is that DCA-modulated anaerobes in chickens reduced the transmitted C. jejuni virulence on inducing campylobacteriosis in mice (Cj-P1-DCA-Anaero vs. Cj-P1). DCA-modulated anaerobes reduce C. jejuni chicken colonization21. Conventionalized anaerobic microbiota reduces C. jejuni-induced intestinal inflammation in gnotobiotic mice20. It is well documented that microbiota prevents intestinal pathogen colonization through competitive exclusion such as virulence expression35, nutrition exclusion36, altered pH37, and bactericidal products38. However, it remains largely elusive whether microbiota influences co-inhabited pathogens on their-transmitted infection capacity. Here, we showed that DCA-modulated anaerobes in chickens reduced the transmitted C jejuni colonization and invasion in the large intestine of mice. Mechanistically, DCA-modulated anaerobes reduced transmitted C jejuni induction of immune cell migration in vitro and into the intestine as well as decreased the bacterial induction of proinflammatory mediator expression. It will be important to conduct future research on which C. jejuni virulence factors were down-regulated by DCA-Anaero and were responsible for the reduction of campylobacteriosis. These findings suggest that the increased chicken-transmitted C. jejuni virulence is feasible to be controlled with select microbiota (e.g. DCA-modulated anaerobes).

Interestingly, DCA-modulated C. jejuni (Cj-P1-DCA) failed to reduce intestinal inflammation, while DCA in the diet reduces C. jejuni chicken colonization21. Campylobacter jejuni multidrug efflux transporter gene, CmeABC, is significantly up-regulated during the initial exposure to bile acids such as DCA39. In consistent with the elevated production of CmeABC, bile salts in culture media promote C. jejuni resistance to multiple antimicrobials39. DCA induces C. jejuni virulence gene expression and invasion into epithelial cell in vitro22. DCA failed to reduce C. jejuni in vitro growth21. Consistently, Cj-P1-DCA showed similar motility and induction of immune cell migration compared to Cj-P1, suggesting comparable virulence on activating immune cells.

Taken together, our data revealed that DCA-modulated anaerobic microbiota not only reduced C. jejuni colonization in chickens, but also reduced chicken-transmitted campylobacteriosis. The reduction of virulence reflected on immune cell migration and infiltration into the intestine, C. jejuni invasion, pro-inflammatory response, and collectively intestinal inflammation. These findings highlighted the importance to monitor chicken-transmitted C. jejuni virulence in addition to colonization counts. Although it is a new concept, using select microbiota in poultry production is a key step to successfully prevent foodborne campylobacteriosis.