Campylobacter jejuni is responsible for approximatively 90% of campylobacteriosis cases in humans (Kaakoush et al., 2015). Campylobacteriosis is a severe but self-limiting gastro-enteritis with an infectious dose as low as 500 bacteria (Robinson, 1981; Black et al., 1988) that lasts an average of 6 days. In some cases, it can lead to severe conditions such as the Guillain-Barré syndrome, Miller Fisher syndrome, reactive arthritis, irritable bowel syndrome, and celiac disease (Facciolà et al., 2017). In Canada, there was an annual incidence rate of 29.14 per 100,000 population of campylobacteriosis in 2018 (Public Health Agency of Canada, 2021), while in Europe, there were 220,682 human cases attributed to this pathogen in 2019, that was reported as the first cause of notified zoonosis (European Food Safety Authority and European Centre for Disease Prevention and Control, 2021). The principal sources of human contamination by C. jejuni are poultry, dairy products like unpasteurized milk, and water (Domingues et al., 2012). Poultry products are recognized as the primary source of exposure for humans, mainly due to the high C. jejuni loads that can be found in the intestines of poultry carriers (up to 10⁹ CFU/g of intestinal content; Greig and Ravel, 2009). Humans can be infected by the consumption of contaminated or cross-contaminated undercooked chicken products (Nadeau et al., 2002). The majority of campylobacteriosis cases in humans are sporadic, except for virulent Campylobacter species that would appear to be associated with diffuse outbreaks (Llarena et al., 2016). Some outbreaks are reported in the literature after the consumption of undercooked chicken meat (Yu et al., 2010; Llarena and Kivistö, 2020), chicken liver pâté (Little et al., 2010; Moffatt et al., 2016; Lanier et al., 2018), and unpasteurized milk (Centers for Disease Control and Prevention (CDC), 2002; Burakoff et al., 2018).

According to the season and country, up to 95% of broiler chicken flocks can be contaminated with C. jejuni by 2 weeks of age (Stern et al., 2001; Newell and Fearnley, 2003; van Gerwe et al., 2009). Absence of intestinal colonization by C. jejuni before 2 weeks of age could be partially due to the presence of maternal antibodies in younger birds (Sahin et al., 2003) and/or to intestinal microbiota properties (Laisney et al., 2004; Han et al., 2016). In birds, the infection occurs via the fecal-oral route, leading to a rapid spread of C. jejuni in the flock (Sahin et al., 2015). Once ingested, C. jejuni migrates to the chicken intestine and especially to the caeca; this migration is most probably promoted by colonization factors related to chemotaxis and motility (Hermans et al., 2011). After colonizing the gut, the bacterium can disseminate to internal organs such as the liver through the bloodstream (Cox et al., 2005, 2007; Richardson et al., 2011). According to the literature, an increase of the paracellular permeability, associated with alterations of the expression of tight junctions proteins could explain this phenomenon (Lamb-Rosteski et al., 2008; von Buchholz et al., 2022). At the slaughterhouse, sources of contamination are numerous: contaminated feathers, gut leakage caused by the defeathering process, evisceration, and finally cross-contamination via direct contact with various contaminated surfaces within the production environment (Newell et al., 2001). Moreover, C. jejuni extraintestinal spread may present an additional challenge as C. jejuni was detected inside of retail chicken livers, with a prevalence ranging from 10 to 100% and with bacterial loads varying from 10 to 10⁵ CFU/g of liver (Barot et al., 1983; Whyte et al., 2006; Noormohamed and Fakhr, 2012; Harrison et al., 2013; Berrang et al., 2018).

Usually, C. jejuni colonization is asymptomatic in chickens, which explains why this microorganism is often considered by some authors to be a commensal or commensal-like bacterium of the chicken gut. A few studies have observed in experimental settings that particular strains of C. jejuni have caused mild clinical signs in broiler chickens, mainly transient diarrhea, arthritis, and reduction in body weight gain (Ruiz-Palacios et al., 1981; Dhillon et al., 2006; Gharib Naseri et al., 2012; Awad et al., 2015; Alpigiani et al., 2017). Interestingly, one study isolated C. jejuni from the liver of broiler chickens affected with avian vibrionic hepatitis, but a causative link remained unclear (Jennings et al., 2011).

The wide genetic diversity of C. jejuni strains in poultry production has been revealed using molecular biology techniques (Wassenaar and Newell, 2000). Most of the studies have shown that multiple C. jejuni genotypes could be simultaneously isolated from the cecal content or feces of broiler chickens from the same flock (Thomas et al., 1997; Hiett et al., 2002; Messens et al., 2009), suggesting that birds can be colonized with different C. jejuni strains on the same farm. Among these multiple genotypes, the dominance of one or sometimes two C. jejuni strains in the cecal content or feces of broiler chickens within the same flock is noted most of the time, from the time of contamination until slaughter (Newell et al., 2001; Nadeau et al., 2002; Ring et al., 2005; Ellerbroek et al., 2010), and this has been reported to be due to the different abilities of these strains to colonize broiler chickens (Ringoir and Korolik, 2003; Chaloner et al., 2014; Pielsticker et al., 2016). Thus, when a multi-strain colonization occurs, the involved C. jejuni strains compete and some may become predominant, displacing the others, as it has been confirmed experimentally (Korolik et al., 1998; Konkel et al., 2007; Coward et al., 2008; Calderón-Gómez et al., 2009). In a previous study aimed at better understanding these different abilities to colonize the chicken gut, C. jejuni strains isolated from the caeca of broiler chickens were characterized in vitro and classified according to their autoagglutination, chemotaxis, adhesion, and invasion properties known to contribute to an effective gut colonization. It was also found that these strains showing different phenotypes possess different abilities to compete for gut colonization in birds (Thibodeau et al., 2015a).

As for extraintestinal spread, one or more genotypes of C. jejuni have been found inside retail chicken livers but only one genotype seemed to be dominant (Berrang et al., 2018, 2019). Experimentally, the strains’ ability to colonize the chicken intestine and to spread to organs, including the liver, was determined using a model where birds were inoculated with a dominant C. jejuni isolate (Jennings et al., 2011; Firlieyanti et al., 2016; Pielsticker et al., 2016). Despite successful intestinal colonization, hepatic spread was highly variable. Using animal experimentations, other authors also demonstrated that C. jejuni strains from human and avian origin would have different abilities to spread to the liver in broiler chickens (Chaloner et al., 2014; Pielsticker et al., 2016). Finally, in a recent study conducted at the slaughterhouse, the authors isolated a dominant strain of C. jejuni from the cecal content of broilers that was different from the one isolated from the inside of the liver of the same bird (Berrang et al., 2019). Thus, the extra-intestinal C. jejuni spread to the liver appears to also be strain specific and would involve other unknown factors and mechanisms that appear to be different from those governing intestinal colonization.

These new insights into the relationship between some C. jejuni strains inside the chicken intestinal environment and liver is therefore raising concerns that the presence of the pathogen in birds might impact animal and public health. However, how intestinal colonization and extraintestinal spread are linked or related, especially in the context of a multi-C. jejuni strain colonization, remains unclear.

The aim of this study was therefore to further analyze the intestinal co-colonization of two thoroughly characterized C. jejuni strains with an opposite potential to compete for gut colonization in broiler chickens. Using different ratios of these strains, we investigated the impact of this multi-strain inoculation on each strain, on cecal and ileal colonization of birds, on the extra-intestinal spread of each strain to the liver and on the expression of tight junction proteins from ileal tissue. We also explored the underlying mechanisms of the intestinal competition between the same two C. jejuni strains by in vitro co-culture assays.

Most in vivo experiments in broilers available in the literature have been conducted using a single C. jejuni strain. However, considering the wide variety of C. jejuni strains observed at the broiler farm level, the current study evaluates the impact of the presence of two distinct C. jejuni strains – alone and together – on the intestinal colonization and hepatic spread of the bacterium in broiler chickens to approximate the reality on farms more closely.

Although uninoculated birds were raised in the same room as inoculated birds, they were not colonized by C. jejuni throughout the animal experiments. Moreover, birds inoculated with a single C. jejuni strain were shown to have been colonized only by that strain. All measures taken in the animal facility to prevent contamination between groups were very effective, thereby to compare the intestinal colonization level, the hepatic spread of C. jejuni and the transcript level of tight junctions for birds inoculated by single or both strain(s) without room effect.

In both rooms, few birds were colonized in caeca at 1 day post infection (dpi), and even fewer birds were found to be colonized in ileum after analysis, due especially to the fact that the dose and inoculated strain(s) would have had an influence on early intestinal colonization. It has already been described that, depending on the strain, the cecal colonization of broilers could start at 1 dpi or later, and that all birds were colonized between 3 dpi and 14 dpi (Young et al., 1999; Knudsen et al., 2006). In this study, at 7 and 21 dpi, broiler chickens inoculated with 10³ CFU of D2008b alone were never shown to be colonized throughout the experiment, unlike birds inoculated with 10³ CFU of G2008b alone, with 10⁷ CFU of D2008b, or with both strains. Although both strains were originally isolated from the caeca of broiler chickens, it seems that a 10³ CFU inoculum of D2008b was not sufficient to induce colonization and the minimum dose of D2008b required for an effective intestinal colonization appears to be higher for this strain compared to G2008b. It has already been shown that a minimum dose of 10² to 10⁴ CFU per bird, according to the C. jejuni strain used, is required for effective colonization in 14-day-old chickens (Ringoir and Korolik, 2003; Dhillon et al., 2006; Knudsen et al., 2006). These variations between strains are likely due to differences in phenotypic properties such as autoagglutination, chemotaxis, adhesion, and invasion, as previously reported (Thibodeau et al., 2015a). Although similar cecal colonization levels were observed between birds inoculated with 10⁷ CFU of the D2008b and those inoculated with 10³ CFU of G2008b at 21 dpi, D2008b appears to be better able to colonize the ileum. We suspect that the inoculum difference could impact colonization levels in the ileum or that the colonization dynamics of these two C. jejuni strains at the caeca and ileum levels could be different. Interestingly, our observations support similar observations reported by in broiler chickens inoculated with a same inoculum of two C. jejuni reference strains separately: M1 or 13126 (Chaloner et al., 2014). These authors showed that M1 preferentially colonized the caeca, whereas 13126 colonized the distal ileum at a higher level. The authors suggested that these strains would have different metabolic capacity and that variations throughout the gastrointestinal tract, including pH, oxygen tension, levels of bile, mucins, or even differences in the intestinal microbiota could explain these findings.

For birds housed in room 1 inoculated with both strains, D2008b seemed to benefit from the presence of G2008b for its establishment in the caeca and in the ileum as both strains were found at similar average levels at 21 dpi; D2008b alone did not colonize the birds. As the presence of D2008b did not affect the colonization levels of G2008b in the caeca and in the ileum, we suggest that a commensalism could occur between both C. jejuni strains at the intestinal level. Commonly, the term commensalism is used to define an interaction between two organisms that is beneficial for one without negatively or positively impacting the other. On the contrary, no commensalism was observed for the in vitro co-culture assays. Consequently, we think that direct interactions between strains would not underlie this proposed relationship observed in vivo. It was also demonstrated that other bacterial species can enhance Campylobacter colonization in mice (Wang et al., 2018) and in chickens (Yan et al., 2021), such as Clostridium. Therefore, investigation on the microbiota of the study animals is mandatory to see if different bacterial populations are associated to the different inoculated strains. When the inoculum of D2008b was increased by 4 log₁₀ for birds housed in room 2, similar cecal colonization levels of D2008b were observed at 7 dpi, whether or not birds were inoculated with G2008b. However, we noted an average decrease of about 3 log₁₀ in the colonization levels of D2008b in the presence of G2008b in caeca at 21 dpi. We therefore suggest that competition occurs between C. jejuni strains for cecal colonization between 7 dpi and 21 dpi, as reported in other experiments (Konkel et al., 2007; Coward et al., 2008; Calderón-Gómez et al., 2009). Our findings show that the in vitro classification method developed by our group successfully predicted G2008b as a dominant strain during this cecal co-colonization (Thibodeau et al., 2015a), despite a higher inoculum of D2008b. In addition, in room 2, we found variations in strain ratios between birds inoculated with both strains at 21 dpi, as already observed by other authors (Konkel et al., 2007; Coward et al., 2008). For these authors, stochastic effects including bird-to-bird transmission, population crashes, and phenotypic variations would contribute to these variations. Despite lower levels found in the ileum, we saw that D2008b colonization was decreased in the presence of G2008b at 21 dpi. We therefore demonstrated that C. jejuni strains compete for establishment in the ileum. Authors have also reported this competition in their dual infection model with two C. jejuni reference strains in chickens (Chaloner et al., 2014). For birds from room 1, there might be a dominance effect observable at the last time point in coherence with the differences in the amount of both strains for some birds.

To explore the underlying mechanisms, both strains were co-cultivated in vitro in MH broth. After 48 h of culture, the lower gene copies number observed for D2008b in the presence of G2008b suggests that an interaction occurs between strains. We hypothesize that G2008b could secrete antimicrobial factors in the presence of D2008b to stop its growth, which would explain the absorbance increase at T = 24 h when strains were grown together. This increase in the optic density could also be due to an increased overall bacterial death of D2008b. Another hypothesis would be based on a different susceptibility of both strains to oxidative stress resulting from bacterial growth (Nennig et al., 2022). Moreover, both strains might not be able to use the same nutrients as the culture media is becoming depleted, therefore giving an advantage to one strain over the other. Similar in vitro experiments have been performed with two C. jejuni strains competing for cecal colonization in chickens (Konkel et al., 2007). However, the authors observed no significant difference in the growth rate of strains when cultured alone or together. Future work should focus on investigating the nature of interactions that occur between C. jejuni strains. Although differences observed between both strains in our in vitro co-culture assays are lower than that of our in vivo trial, this suggests that other mechanisms could be involved in strain dominance. However, all of these findings remain to be validated with other competitive strains by performing in vitro assays.

Additionally, we observed that a 4 log₁₀ increase in the inoculum of D2008b led to its greater dissemination to the liver, likewise reported in the literature (Awad et al., 2020). Moreover, we also noted that G2008b appears to facilitate and accelerate the hepatic dissemination of mainly D2008b, and this was clearly observed when a low dose of D2008b – insufficient on its own to cause intestinal colonization – was used. Recently, similar observations were reported in female turkeys co-inoculated with C. jejuni and C. coli (Rzeznitzeck et al., 2022). For the first time, we therefore demonstrate that an inoculation with two C. jejuni strains would increase liver contamination. Our findings could explain why the majority of studies conducted with a single C. jejuni strain show a much lower rate of liver contamination than is found inside retail chicken livers (Knudsen et al., 2006; Pielsticker et al., 2016). Based on these new data, the hepatic spread variability of isolates inoculated separately (Knudsen et al., 2006; Jennings et al., 2011; Pielsticker et al., 2016) would illustrate the need for additional C. jejuni strains to improve the hepatic dissemination of a specific strain. In the current study, the dominance of D2008b inside the liver confirms that C. jejuni strains have different abilities to invade the internal organs of broiler chickens (Young et al., 1999; Chaloner et al., 2014; Pielsticker et al., 2016). Moreover, in some birds inoculated with both strains, a dominance of D2008b in the liver was associated with a dominance of G2008b in intestinal content. Similar observations have already been reported in a recent study conducted by at the slaughterhouse, showing the isolation of distinct C. jejuni strains from the caeca and liver of the same chickens (Berrang et al., 2019). It therefore seems that intestinal colonization and extraintestinal spread involve different mechanisms. In addition, the transcript level decrease of JAM2, CLDN5 and CLDN10 at 7 dpi observed for some chicken groups inoculated with both strains in both rooms and having livers contaminated by C. jejuni suggest that the presence of C. jejuni could alter the intestinal epithelial barrier function and enhance the intestinal paracellular permeability, likewise reported in the literature (Lamb-Rosteski et al., 2008; Awad et al., 2020; von Buchholz et al., 2022), leading to extraintestinal spread to internal organs through the bloodstream because of a leaky gut, likely for specific strains only. In view of our results, we suggest that this hypothetical mechanism, leading to the hepatic spread of D2008b, would be enhanced in the presence of G2008b. However, the reason why G2008b did not seem to benefit from this leaky gut by entering the liver remains to be determined. Moreover at 21 dpi, the host response to the hepatic dissemination, like a tissue repair, could explain the increased transcript level of ZO1, JAM2, OCLN and CLDN10 for the groups of birds having livers contaminated by C. jejuni. This hypothesis could thus explain the transient extraintestinal dissemination of C. jejuni to the liver. Future research should thoroughly study genotypic and phenotypic characteristics of C. jejuni strains able to invade the liver and study the underlying mechanisms of this hepatic dissemination during a multi-strain colonization in broiler chickens.

In conclusion, we showed that the in vitro classification developed by our group successfully predicted the competition between D2008b and G2008b, with a dominance of G2008b in the ileum and caeca of birds. For the first time, we observed that commensalism seems to exist between both C. jejuni strains, and that this relationship could contribute to intestinal colonization and enhance the hepatic spread of one strain in particular. These results highlight the importance of limiting the introduction of other C. jejuni strains by implementing effective biosecurity measures that would help avoid the potential amplification of gut colonization and liver dissemination of certain strains already present on the farm.

However, given the variance of Campylobacter colonization in chickens, more animal experimentations should be performed with C. jejuni strains in order to confirm the hypotheses of the current study since the lack of replicates in the present study is a clear limitation. For instance, it would be interesting to vary the strains used for inoculation and to sequence them to be able to increase the number of strains for comparative genomics to narrow down the list of genes that might be involved in the differential spread of the strains to the liver. Additionally, using different dose for inoculation and following the animals every day instead of every week would help to better understand what is going on when multiple strains of C. jejuni colonize the chicken at the same time. One last limitation of our study is that the number of copy numbers of genes found in each strain was not validated, results that could slightly modify the qPCR results. Future studies should focus on the underlying mechanisms of the commensalism and competition between C. jejuni strains during multi-strain colonization in chickens. Such as analysis of bird immune response and intestinal microbiota as an improved understanding of the colonization dynamics between C. jejuni strains, alone and together, should be considered in order to develop adequate on-farm control strategies.

The original contributions presented in the study are included in the article/Supplementary material, further inquiries can be directed to the corresponding authors.

The animal study was reviewed and approved by Comité d’Éthique sur l’Utilisation des Animaux (CÉUA) of the Faculté de Médecine Vétérinaire of the Université de Montréal (certificate number: 19-Rech-2039).

SC, PF, and AT conceived and designed the experiments. WT provided essential scientific and technical advice. SC, M-LG, WT, and AT actively participated in conducting the in vivo experimentations, including bird inoculations, necropsies, and sample processing. SC performed all the experiments in the laboratory. SC and AT analyzed data and wrote the manuscript. M-LG, WT, and PF revised the manuscript. All authors contributed to the article and approved the submitted version.

The current study was financially supported by the Natural Sciences and Engineering Research Council of Canada (NSERC grant no. RDCPJ 520873-17), the Consortium de recherche et innovations en bioprocédés industriels au Québec, the Swine and Poultry Infectious Diseases Research Center (CRIPA), The Groupe de Recherche sur les maladies infectieuses en production animale (GREMIP) and the Fonds du Centenaire of the Faculté de Médecine Vétérinaire of the Université de Montréal.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.