Differences in virulence gene expression between human blood and stool Campylobacter coli clade 1 ST828CC isolates

Campylobacter colonise the gastrointestinal tract of warm-blooded animals, such as poultry and wild birds, ruminants, pigs, cats and dogs but rarely cause any symptoms in animal hosts. However, when transmitted to humans, Campylobacter cause gastroenteritis worldwide [1]. It has been estimated that about 1% of Europeans develop campylobacteriosis every year [2], and Campylobacter have been the most commonly reported bacterial enteropathogen since 2005 in the European Union, with over 246,000 reported cases in 2016 [3].

Campylobacter infection usually develops one to 5 days after ingestion of the bacteria, with typical symptoms including fever, headache, vomiting and abdominal pain in addition to watery or even bloody diarrhoea [4]. Known post-infection complications are reactive arthritis [5], irritable bowel syndrome [6], and the severe autoimmune demyelinating neuropathy Guillain-Barré syndrome [7], which substantially increase the burden of the disease, which in the USA has been estimated to be 22 500 disability-adjusted life-years [8].

Campylobacter coli is less common than C. jejuni as a human pathogen and seems to account for about 10% of the total number of Campylobacter infections [9]. C. coli infection cannot be distinguished from that of C. jejuni on the basis of the clinical picture [9]. C. coli isolates cluster into three clades [10‐12], where those of clade 1 are mainly clinical and animal isolates, including the clonal complexes of ST828CC and ST1150CC [13], whereas clade 2 and clade 3 lack clonal complex structure and mainly contain isolates of environmental origin [10, 12].

Campylobacter bacteraemia is rare, estimated to occur in less than 1% of the total number of Campylobacter infections [14, 15], and then mainly in the elderly and in patients with immune deficiencies or other underlying conditions [15, 16], although young and otherwise healthy individuals have also been shown to be affected [17, 18]. C. jejuni is the predominant pathogen among campylobacteriosis cases and most commonly isolated from blood while C. coli accounts for around 10% of bacteraemia cases [14, 17, 19].

Campylobacteriosis is considered a multifactorial process in which bacterial properties such as motility, adhesion and invasion seem to be important for virulence [20], and mutational studies in C. jejuni have identified several virulence genes implicated in these processes. Adhesion is thought to be initiated by CadF and FlpA, which are bacterial adhesins known to bind fibronectin on the host cell surface [21‐24]. Secreted Campylobacter invasion antigens, such as CiaB, have been shown to thereafter be needed for efficient invasion into the cells [25, 26]. Also, the invasion-associated marker, IamA, is considered to play a role in invasion, as it has been reported to be present in the majority of invasive, but only in the minority of non-invasive, Campylobacter isolates [27]. The cytolethal distending toxin, CDT, is well-known among Campylobacter and also produced by many other bacterial pathogens, such as Salmonella and E. coli (reviewed in [28]). CDT consists of three subunits; the enzymatically active B subunit and the regulatory A and C subunits. The DNase activity of CdtB induces double-strand breaks in the host chromosomes, eventually leading to cell cycle arrest at the G2/M interphase. CDT might also contribute to inflammation as it has been shown to induce IL-8 secretion from INT-407 cells through its binding to host cell membranes [29].

Why some Campylobacter isolates are more invasive than others and enter the bloodstream is not well understood. In this study, we compared genotypic and phenotypic characteristics of C. coli clade 1 blood and stool isolates and found that, although they are very similar genetically, they showed differences in gene expression levels of several virulence factors.

The aim of this study was to compare various virulence traits in C. coli blood and stool isolates. First, we used the PubMLST/Campylobacter database to get an understanding of the contribution of C. coli to human infections. The percentage of 8.7% C. coli among all human isolates is in line with the reported percentage of 10% C. coli in human Campylobacter infections. However, when looking at Campylobacter isolates from all sources, the contribution of C. coli was higher (16.5%), suggesting that C. coli might be more prevalent among animal and environmental isolates. C. coli environmental isolates mainly belong to clades 2 and 3 and our earlier studies have identified several specific traits that would contribute to their survival in the environment [12, 30].

Surprisingly, when only looking at the human isolates in the PubMLST database, there was a substantially larger fraction of C. coli among the blood isolates (25.9%) compared to the stool isolates (8.9%). This can be interpreted in several ways. Either that C. coli might be better than C. jejuni to invade the blood stream and cause a more severe disease and that these numbers truly reflect different characteristics of the two species, or simply that there has been more interest to characterise C. coli blood isolates, as they are a less common cause of bacteraemia. Interestingly, a large fraction (27.3%) of the C. coli blood isolates were not assigned to any clonal complex, suggesting that these particular isolates might be genetically different from the commonly detected C. coli stool isolates.

We were further interested in whether human C. coli blood and stool isolates could show phenotypic and genetic differences that might explain the successful invasion of the blood stream. In our established in vitro infection model using HT-29 colon cancer cells, all C. coli blood and stool isolates were able to adhere to the cells and evoke an IL-8 response. The stool isolates adhered significantly better than the blood isolates at 3 h post infection while induced IL-8 levels were similar for all isolates tested. Neither did we see a difference in adhesion between C. jejuni blood and stool ST21CC isolates in our earlier study [31], suggesting that adherence potential rather varies on the isolate level than depending on isolation source and genetic origin.

Nucleotide and amino acid sequences of genes involved in adhesion and invasion, such as cadF, flpA, ciaB and iamA, were very similar for the C. coli blood and stool isolates. However, the genes did show differences in RNA expression levels between the two groups of isolates. flpA, which is involved in adhesion through its binding to fibronectin, and ciaB and iamA, which are involved in invasion, were expressed to higher levels in stool isolates than in blood isolates. For the cdt toxin genes cdtA, B and C, the differences between the blood and stool isolates were even larger, although the stool isolates varied in their expression levels. C. coli clade 1 stool isolates were earlier shown to have lower expression levels of several virulence genes compared to C. coli clade 2 and 3 water isolates [30]. All these results together highlight the importance of measuring actual expression levels of genes and not only looking at presence of the genes, as many studies show vague correlation between presence of virulence genes demonstrated by PCR and severity of disease. Interestingly, we also found several disruptive mutations in the cdt genes. The most intact gene was cdtB, which had a complete ORF in 84% of all isolates. This is expected since cdtB is the enzymatically active component of the toxin, and enzymes usually are more evolutionary conserved than regulatory genes. These types of mutations were not seen in any of the virulence genes involved in adhesion and invasion studied here or in any of the 80 C. coli clade 2 and 3 environmental isolates investigated in our previous study [30], excluding, in our opinion, the risk of this phenomenon being due to methodological errors during sequencing.

Campylobacter show great capability to adapt to different hosts and environments. The expression level differences of the virulence genes and the differences in cdt gene structure between isolates from different sources suggest that C. coli might be able to downregulate and/or inactivate selected virulence mechanisms once they established in the human gut but even more so after invading the blood stream. C. jejuni stool isolates of ST21CC, as compared to blood isolates, have been shown to possess larger amounts of accessory gene content, such as bacteriophage-derived integrated elements (CJIEs) and plasmids [31], possibly resulting in increased invasive potential.

Induction of cell cycle arrest in the small intestine by the CDT toxin would prevent the regeneration of the intestinal epithelium, which then might facilitate invasion. This particular activity would not be so important for bacterial survival and maintenance when the infection has already established in the gut and/or spread to the blood stream. It was earlier shown that C. jejuni human isolates produced much higher levels of the CDT toxin compared to C. coli human isolates [33], indicating that the CDT toxin might not be as important in C. coli virulence.

In this study, we present interesting differences in virulence gene structure and expression levels between C. coli clade 1 blood and stool isolates. This data suggests that C. coli might downregulate and/or inactivate their virulence mechanisms once they have established the infection in the human gut and invaded the blood stream, and contributes to the understanding of Campylobacter pathogenesis.