Background
Thermophilic campylobacters are the most common bacterial cause of diarrhoea in humans worldwide [
1]. Enteric diseases caused by the thermophilic species
C. jejuni,
C. coli,
C. lari, and
C. upsaliensis range from asymptomatic infections to severe inflammatory bloody diarrhoea [
2].
C. jejuni is often associated with the Guillain–Barré syndrome [
3]. Virulence mechanisms in campylobacteriosis are currently poorly understood.
Poultry and poultry products remain the most common source of foodborne human campylobacteriosis [
4,
5]. The natural habitat of thermophilic
Campylobacter is the intestinal tract of healthy birds and raw meat that can be contaminated during the slaughtering process. Consumption of undercooked chicken meat or contaminated ready-to-eat food is the most common source of infection.
Campylobacter are also found in pigs and cattle. Swine carcasses are often contaminated with faeces at the slaughter and processing facilities during the evisceration process, which ultimately leads to contaminated food products [
6‐
8]. Compared to poultry, the relevance of swine in foodborne campylobacteriosis is not well studied. However, a high incidence of
Campylobacter on pork products at the retail level was found [
9,
10].
South East Asia including Vietnam was often considered a hotspot for emerging infectious diseases [
11]. Vietnam is currently a developing country and knowledge about
Campylobacter and campylobacteriosis is limited. Only few data exist about the prevalence of
Campylobacter in children [
12,
13] and adults [
14]. The prevalence rates of
Campylobacter in cases of diarrhea were between 2 and 4 % for children and <1 % for adults. In a study concerning the incidence of diarrhea in rural Vietnamese children [
15]
Campylobacter was the most frequently identified pathogen comprising 31 % of all isolates.
Fifteen to 32 % of meat samples in different regions of Vietnam contained thermophilic
Campylobacter [
16‐
21]. Duck meat and pork were also contaminated with
Campylobacter in 23.9 and 53.7 % of tested samples, respectively [
21]. Bao et al. isolated thermophilic campylobacters from 35.1 % of chicken carcasses in large and small abattoirs of Ho Chi Minh City and 67.9 % of the isolates belonged to the species
C. jejuni [
22]. However, Schwan investigated meat samples from markets in the Can Tho Province but found no
Campylobacter spp. [
19].
Several molecular biological methods for characterization and discrimination of
Campylobacter isolates have been developed [
23]. PCR and
flaA typing were used as well as multi-locus sequence typing (MLST) and microarray assays for determination of relatedness among isolates [
24,
25].
The molecular genetics of
Campylobacter has been extensively studied but the pathogenesis of
Campylobacter infections is not fully understood. A number of putative virulence and toxin genes that may contribute to pathogenicity in human
Campylobacter infection have partly been identified and sequenced [
26‐
29].
Flagella-mediated motility, adherence to intestinal epithelial cells, invasion and survival in the host cells as well as the ability to produce toxins are important virulence factors [
27]. The involvement of the
flaA gene in
Campylobacter colonization has been shown [
30]. Several
Campylobacter cytotoxins have been identified [
31] and the cytolethal distending toxin (CDT) has been characterized in detail [
32,
33]. CDT is composed of three subunits and it has been suggested that CDT, amongst other functions, may play a role in adhesion and invasion [
34]. Active CDT is lethal for host enterocytes [
35,
36].
It was shown that 19–53 % of
Campylobacter spp. strains contain plasmids of various sizes [
37]. The plasmid-encoded
virB11 gene is a marker potentially associated with the virulence of C
ampylobacter species [
38].
A study with Vietnamese isolates dealt with the identification of possible virulence markers like a novel protein translocation system, the type-6 secretion system [
39].
The antimicrobial resistance of
Campylobacter isolates was investigated also in several studies [
19,
20,
40]. High resistance rates in
C. jejuni were determined against ciprofloxacin, nalidixic acid and tetracycline with 64, 46 and 68 %, respectively. Resistance against antibiotics in
C. coli isolates was higher than in
C.
jejuni. All
C. jejuni isolates were resistant to ciprofloxacin and nalidixic acid while, 83 % showed resistance to tetracycline [
19]. These isolates were recovered from faeces. The broth microdilution method is an easy and reliable method for interpreting minimum inhibitory concentration (MIC) values for
C. jejuni and
C. coli which is also recommended by EUCAST [
41‐
43]. The emergence of antimicrobial resistance in
Campylobacter, particularly to fluoroquinolones, has showed the need for continued monitoring of
Campylobacter resistance.
In this study, Vietnamese Campylobacter isolates were characterized to assess their genetic relatedness, potential virulence factors and antibiotic resistance profiles. The isolates were recovered from chicken and pig meat from two slaughterhouses in Hanoi. The investigation was done using different molecular biological tests, MLST, microarray analysis, and the antimicrobial susceptibility was assessed.
Discussion
As in other countries, thermophilic
Campylobacter are common bacterial agents in Vietnam which cause gastro-enteric illness in humans, especially in children [
13,
15]. Meat and meat products serve as the main sources of human infections. Huong et al. showed that approximately 30 % of raw chicken samples from school and hospital canteens and retail markets in Hanoi city were contaminated with
Campylobacter [
17]. Other studies from different regions of Vietnam came to comparable results [
20,
21]. Here, chicken and pork meat samples from a market in Hanoi were investigated for the presence of thermophilic campylobacters. 13.3 % of the samples were
Campylobacter positive.
C. jejuni and
C. coli were detected as contaminants of chicken and pork meat. In comparison, a report from Germany showed that 52.3 % of chicken carcasses in slaughterhouses and nearly 40 % of raw meat in retail were positive for these microorganisms [
53]. It is possible that the lower percentage of
Campylobacter findings in the meat samples from the Hanoi market is a result of slaughter procedures. In Vietnam the meat is not prepared in large slaughterhouses for retail as in Germany. Chicken and also pigs are slaughtered in low numbers and often directly on-site and the risk of contamination for example by generation of aerosols in slaughterhouses is smaller.
In this study, 20
Campylobacter isolates were obtained from meat samples in Vietnam. Unfortunately, after transfer to Germany only 9 isolates could be re-cultivated which were subject to further investigation. Eight of them were identified as
C. jejuni and one as
C. coli by mPCR. Several methods were used to type the isolates. As a rapid and simple method to illustrate heterogeneity within the
C. jejuni isolates,
flaA-RFLP typing was used. Four different strain types were detected by
DdeI digestion of the amplified
flaA gene. This enzyme was used because it showed the highest discriminatory power in former investigations [
24] in comparison with
AluI or
Sau3AI. The digestion pattern of
C. coli isolate 09CS0051 was completely different. Limitations of this typing method resulted from the use of only a very small part of the
Campylobacter genome and difficulties in standardization of the analytical process. This complicates an inter-laboratory comparison of results between different laboratories.
Microarray analysis worked as a PCR-based comparative genomic fingerprinting (CGF) assay [
54] and confirmed the heterogeneity of the
C. jejuni isolates. An advantage of this method is the use of the whole genome data instead of only one or a few genes. The basis of this method is the detection of the presence or absence of several gene loci that are spread over the whole genome. SplitsTree analysis of the hybridization results showed high genetic diversity as no isolate is identical with another one. The
C. coli isolate 09CS0051 was clearly distinct from the
C. jejuni isolates. Additionally, sequence types and clonal complexes of the isolates determined by MLST are given in Fig.
2.
Both methods showed differences concerning the relatedness of different C. jejuni isolates among each other. Isolates 09CS0049 and 09CS0066 represented an identical sequence type and belonged to the same clonal complex but in microarray investigation they showed only poor relatedness. 09CS0043 and 09CS0047 were part of the clonal complex ST-353 but differed in the sequence type. Genetic relatedness based on microarray data was marginal. In contrast, two isolates (09CS0040 and 09CS0046) were found with identical sequence type and pattern in flaA typing after DdeI digestion and even microarray analysis showed a high degree of similarity.
The major advantage of MLST is the comparability of results independent from the laboratory and the local working conditions (technicians, machines etc.). The relatively high costs of this complex technique are outweighed by the hard facts that are obtained in the form of DNA sequences of seven house-keeping genes. In this study, six sequence types in the group of
C. jejuni isolates were detected. These sequence types were compared with the database on the
Campylobacter MLST Home Page (
http://pubmlst.org/campylobacter/). ST 2837 and ST 4395 belonged to clonal complex ST-353 whereby 09CS0047 (ST 4395) was identical with an isolate which was recovered from a stool sample of a hospital inpatient with gastroenteritis in Vietnam in 2010. 09CS0049 and 09CS0066 belonged to CC ST-443. Sequence type 5799 was previously isolated from human stool samples in Japan. Three isolates could not be assigned to any known clonal complex. The sequence type of isolate 09CS0068 had previously been discovered once in a human stool sample in Thailand, two others were not described yet. Isolate 09CS0067 represented sequence type 4099. This type belongs to the ST-460 complex and was previously identified in a human sample in Canada.
C. coli isolate 09CS0051 belonged to sequence type 860 and ST-828. Identical isolates were found several times during the last two decades in Europe and the USA. Records from Asia are lacking until now. In summary, the investigated Vietnamese isolates in their majority seemed to represent strains typical for the Asian region. A route of infection of
Campylobacter from meat to humans can be assumed.
The
Campylobacter isolates were characterized regarding virulence factors associated with adhesion and invasion of host cells. All isolates harboured flagellin genes
flaA,
flaB,
flhA and
fliM. Similar observations have been reported previously [
27,
55]. Molecular genetic approaches with defined mutants showed that
flaA is essential for colonization [
30]. The complex flagellum of
Campylobacter species is encoded by two tandem-oriented flagellin genes (
flaA and
flaB). While the function of the
flaA gene seems to be fully elucidated, there are many speculations about the function of the
flaB gene, which may play a role in antigenic variation or influence the motility in various environmental conditions [
56].
fliY, a gene of flagellar motor switch proteins, could not be detected.
The
ciaB gene, coding for a
Campylobacter invasion antigen, was present in most of the
C. jejuni isolates. It was absent in 09CS0049 and
C. coli 09CS0051. Another gene which is important in the invasion process of
Campylobacter to host cells is
iamA. It was detected in all
C. jejuni isolates. Carvalho et al. described the detection of the
iamA gene in 85 % of invasive
C. jejuni but in non-invasive isolates it is rare [
57]. Also the
cadF gene was detected in all
C. jejuni isolates. It encodes for an outer-membrane protein which mediates the binding of the bacteria to fibronectin [
58]. Based on the results it can be concluded that these Vietnamese isolates represented invasive
C. jejuni strains.
Cytolethal distending toxin causes direct DNA damage leading to induction of DNA damage checkpoint pathways [
35]. The
cdt gene cluster consists of 3 genes
cdtA,
cdtB and
cdtC. The
cdt genes were shown to be conserved among different
Campylobacter strains [
59]. Bang et al. observed that the presence of these genes in isolates from different sources exceeds 90 % [
27]. In all Vietnamese
C. jejuni strains isolated from chicken and pork meat the complete
cdt gene cluster was observed. Rozynek et al. obtained results for
C. jejuni strains isolated from children with diarrhea and found that
cdtA,
cdtB and
cdtC were present in 98.4, 97.0 and 98.0 % of all isolates, respectively [
60]. However,
cdtC was not detected in
C. coli isolate 09CS0051 from chicken meat which was in agreement with a previous study [
60]. On the other hand, a similar frequency of
cdt genes and the
cdt gene cluster was observed in dog and chicken isolates [
55]. In this study all investigated isolates harboured the
cdtB gene. It is indeed generally accepted that the
cdtB genes are widespread amongst poultry and cattle as well as in human isolates in Denmark, Japan, Poland, and Belgium [
27,
60‐
62]. However, low percentages of occurrence of
cdtB have been reported in humans (28 %) and chickens (20 %) in India, which could be due to genetic reasons or variations in the isolates from different geographic areas [
63].
Only a minority of
C. jejuni isolates gave positive PCR results for
virB11 encoding a putative component of a type IV secretion system. It is located in the pVir plasmid and could be involved in virulence [
38]. The 25.0 % prevalence of the
virB11 gene in
C. jejuni isolates in this study is higher than 10.3 % in human isolates reported by Bacon et al., but much lower than in pig isolates (35.7 %). Until now, the role of the protein encoded by the
virB11 gene in the invasion and colonization process of eukaryotic cells by
Campylobacter species could not be elucidated.
Macrolides, quinolones and tetracycline are among the common antimicrobials recommended for testing, because they can be of therapeutic relevance in severe cases of infection. High levels of resistance of Campylobacter to tetracycline and ciprofloxacin were frequently reported but resistance to erythromycin and gentamicin remained low.
The antimicrobial susceptibility profiles among the Vietnamese isolates were analyzed based on the guidelines of CLSI (2008) [
41]. In this study standardization of the protocol for the commercially available broth microdilution test as a method for the determination of the minimum inhibitory concentration (MIC) of antibiotics was done [
52]. All isolates were sensitive to gentamicin and most of isolates (88.8 %) were sensitive to chloramphenicol, erythromycin and streptomycin. Similar results were reported in several previous studies [
4,
42,
64‐
68]. In contrast to our findings, a previous study reported high resistance to streptomycin with 60.0 % [
64]. The resistance rate to ciprofloxacin was 66.7 % which is in agreement with a previous study showing high resistance [
51,
64], but in contrast to another study with only 9.5 % [
69]. Resistance to nalidixic acid was 88.9 % which is similar to several aforementioned reports [
51,
64,
68,
70]. However, other studies found either low resistance [
71,
72] or none at all [
65]. Resistances to tetracycline was higher (77.8 %) than previously reported (32.0 %) [
72], but it was lower than in isolates recovered from conventionally grown turkeys [
68].
The gene loci responsible for antibiotic resistance were detected in all resistant isolates to ciprofloxacin and erythromycin and 66.7 % of resistant isolates to tetracycline. Ciprofloxacin resistance among
C. jejuni and
C. coli isolates was conferred by threonine-to-isoleucine mutation of amino acid 86 of the
gyrA protein (Thr-86-Ile), a finding that is in agreement with other previous studies [
73‐
76]. Tetracycline resistance was attributed to the presence of the
tet(O) gene [
51]. All resistant isolates in this study were carrying
tet(O); none of the susceptible isolates gave a positive result using specific PCR.