Background
Campylobacter jejuni is a leading cause of bacterial enteritis in developed countries and the most commonly reported zoonosis in the European Union [
1].
C. jejuni colonizes the gastrointestinal tract of many animals including poultry and wild birds, cattle, pigs, cat and dog. Eating undercooked poultry has been shown to be a risk factor for campylobacteriosis also in Finland [
2], however, poultry is colonized with
Campylobacter to a significantly lower extent than in many other countries [
3]. Epidemiological studies using serotyping and genotyping methods have revealed a high diversity among
C. jejuni from different sources in Finland and the risk factors for human
Campylobacter infection may vary according to the geographical area and even with age [
4,
5].
Although the genomes of several
C. jejuni strains have been sequenced [
6‐
8], very little is still known about the pathogenicity mechanisms and the virulence factors of this common enteropathogen [
9]. The acute
Campylobacter infection is often self-limited but in severe cases antimicrobial and hospital treatment may be needed. The reasons why certain patients develop a more serious acute infection or late sequelae of the disease are not understood.
Several studies have searched for the presence of putative virulence factors among
Campylobacter isolates of human and animal origin but only few studies have been able to show an association between certain bacterial factors and the outcome of human
Campylobacter infection. Genes studied have usually included those suggested to have a role in adhesion, colonization, invasion and toxin production. The plasmid-associated gene
virB11 [
10], as well as the genes
ciaB (
Campylobacter invasive antigen B) [
11,
12], and
cj0486 encoding a putative sugar transporter [
13] have been suggested to be involved in invasiveness. In addition,
pldA encoding outer membrane phospholipase A [
14], and
ceuE encoding enterochelin uptake binding protein [
15] have been studied. The genes
cgtB [
16] and
wlaN [
17] are involved in the biosynthesis of lipooligosaccharide (LOS), which may show ganglioside-mimicking structures important for the triggering of Guillain-Barré syndrome, an acute peripheral polyneuropathy, after
C. jejuni infection [
18]. Cytolethal distending toxin (CDT, encoded by the gene cluster
cdtABC) has been present in most of the tested isolates [
19] but its role in the outcome of the disease still remains uncertain. Likewise, a recent report showed γ-glutamyl transpeptidase (GGT, encoded by the gene
ggt) to have a role in the persistent colonization of the avian gut [
20], but its importance in the course of human campylobacteriosis is not known.
We have recently demonstrated that
C. jejuni isolates of domestic origin and those highly susceptible for ciprofloxacin were associated with a more severe form of enteritis characterized by bloody stools [
21]. Our aim in the present study was to reveal other bacterial factors that may affect the outcome of the
C. jejuni infection in a well-characterized clinical material including both infections acquired abroad or in Finland. For that purpose, we analyzed 166 clinical isolates of
C. jejuni for the production of GGT and the presence of the genes
virB11,
ciaB,
cgtB,
wlaN,
cj0486,
ceuE,
pldA and
cdtABC.
Results
Of the 166 cases of
C. jejuni infection in the study 126 were acquired abroad and 40 were acquired in Finland. Resistance to erythromycin (4 isolates) and resistance to doxycycline (59 isolates) were only detected in isolates of foreign origin. All except one of the isolates of domestic origin were susceptible for ciprofloxacin whereas 83/126 (66%) of isolates of foreign origin were resistant to ciprofloxacin, as published earlier [
21]. Prevalence of the putative virulence markers among the isolates according to MICs to ciprofloxacin and doxycycline, respectively, is presented in Table
2. GGT-production was associated with susceptibility to ciprofloxacin and doxycycline in the univariate analysis. On the other hand, both ciprofloxacin- and doxycycline-resistant isolates were more likely than the susceptible ones to harbor the genes
cj0486 and
ceuE (Table
2).
Table 2
Contingency table results for antimicrobial susceptibility and putative virulence markers among 166 C. jejuni isolates
Ciprofloxacin MIC ≤ 1 (82/166) * | 20/82 (p = 0.001) § | 1/82 | 81/82 | 16/82 | 17/82 | 31/82 | 62/82 | 52/82 | 69/82 |
Ciprofloxacin MIC ≥ 4 (84/166) † | 5/84 | 3/84 | 83/84 | 15/84 | 22/84 | 50/84 (p = 0.0051) ¶ | 80/84 (p = 0.0003) ¶ | 49/84 | 62/84 |
Doxycycline MIC ≤ 2 (102/166) * | 22/102 (p = 0.0032) § | 1/102 | 101/102 | 22/102 | 20/102 | 40/102 | 82/102 | 59/102 | 82/102 |
Doxycycline MIC ≥ 4 (64/166) ‡ | 3/64 | 3/64 | 63/64 | 9/64 | 19/64 | 41/64 (p = 0.0018) # | 60/64 (p < 0.0001) # | 42/64 | 49/64 |
No. positive isolates (%) | 25/166 (15%) | 4/166 (2.4%) | 164/166 (99%) | 31/166 (19%) | 39/166 (23%) | 81/166 (49%) | 142/166 (86%) | 101/166 (61%) | 131/166 (79%) |
Contingency tables were also used to assess whether the presence of putative virulence factors correlated with clinical data. In the univariate analysis (Table
3), GGT production and the presence of the gene
cgtB were associated with bloody stools. In addition, isolates lacking the
ceuE gene were associated with hospitalization, as 8/24 (33%) of the patients infected with
ceuE-negative isolates were hospitalized, compared to 21/139 (15%) of those with
ceuE-positive isolates (p = 0.031). GGT production was strongly associated with infections acquired in Finland as compared to infections from abroad. On the other hand, the genes
cj0486 and
ceuE were markedly more common among isolates of foreign origin. Of all the domestic isolates, 34/40 (85%) showed at least one of the following characteristics; GGT-production, lack of
cj0486 or absence of
ceuE whereas the number of foreign isolates was 60 (48%), respectively (p < 0.0001). A total of 69/126 (55%) of the imported isolates, but only 7/40 (18%) of the domestic isolates were positive for both
cj0486 and
ceuE (p < 0.0001).The significant findings in the univariate model were controlled with a multivariate analysis to assess whether some of the variables were independently associated with each other. GGT-production was independently associated with domestic infections, and the genes
cj0486 and
ceuE were independently associated with imported infections, respectively (Table
4). Although bloody diarrhea was significantly associated with both GGT-production and the presence of
cgtB in the univariate analysis, this factor could not be further analyzed with a multivariate analysis as the proportion of missing data in the questionnaires regarding this specific finding was too high (28%).
Table 3
Characteristics of 166 patients and putative virulence factors present in the respective C. jejuni isolates
Female sex (99/166) | 17/99 | 3/99 | 98/99 | 19/99 | 23/99 | 52/99 | 85/99 | 62/99 | 78/99 |
Underlying disease (38/162) | 9/38 | 3/38 | 37/38 | 10/38 | 9/38 | 18/38 | 32/38 | 23/38 | 31/38 |
Domestic infection (40/166) | 16/40 (p < 0.0001) * | 1/40 | 40 | 9/40 | 8/40 | 7/40 | 26/40 | 25/40 | 34/40 |
Infection from abroad 126/166) | 9/126 | 3/126 | 124/126 | 22/126 | 31/126 | 74/126 (p < 0.0001) † | 116/126 (p < 0.0001) † | 76/126 | 97/126 |
Vomiting (41/156) | 8/41 | 1/41 | 40/41 | 8/41 | 9/41 | 17/41 | 34/41 | 25/41 | 32/41 |
Fever (136/156) | 20/136 | 4/136 | 134/136 | 27/136 | 30/136 | 66/136 | 115/136 | 82/136 | 104/136 |
Bloody stools (21/119) | 6/21 (p = 0.025) | 1/21 | 21 | 8/21 (p = 0.034) | 3/21 | 9/21 | 17/21 | 16/21 | 18/21 |
Long-lasting (≥ 10 d) diarrhea (42/161) | 6/42 | 1/42 | 42 | 12/42 (p = 0.051) | 7/42 | 18/42 | 35/42 | 24/42 | 31/42 |
Hospitalization (29/163) | 3/29 | 1/29 | 29 | 7/29 | 8/29 | 15/29 | 21/29 (p = 0.031) ‡ | 19/29 | 23/29 |
No. positive isolates (%) | 25/166 (15%) | 4/166 (2.4%) | 164/166 (99%) | 31/166 (19%) | 39/166 (23%) | 81/166 (49%) | 142/166 (86%) | 101/166 (61%) | 131/166 (79%) |
Table 4
Multivariate analysis showing independent association between origin of infection and certain C. jejuni markers
GGT* | 8.67 | 3.43-21.91 | < 0.0001 |
cj 0486† | 6.71 | 2.75-16.39 | < 0.0001 |
ceu E† | 6.71 | 2.67-16.95 | < 0.0001 |
PFGE analysis with KpnI digested samples revealed that among the 40 domestic isolates, a total of 32 PFGE types were identified indicating a high diversity. Furthermore, all the 35 isolates of foreign origin analyzed had different PFGE profiles, which did not overlap with those of domestic isolates. Thus, the isolates were highly diverse and did not seem to have a common source.
Discussion
We recently showed that bloody stools were more common among patients infected with
C. jejuni isolates of domestic origin and those highly susceptible for ciprofloxacin [
21].
C. jejuni isolates of Finnish origin have even earlier been shown to differ significantly from those of foreign origin as being almost exclusively susceptible for ciprofloxacin [
28,
29]. In the present study, we further analyzed possible differences between
C. jejuni isolates acquired in Finland and those from abroad and screened for the presence of certain putative virulence markers. Significant differences were detected as GGT production was independently associated with infection of domestic origin and the isolates of foreign origin significantly more often harbored the genes
cj0486 and
ceuE, findings also verified with a multivariate analysis.
GGT is an enzyme present in both bacteria and eukaryotes. It has a role in glutathione and glutamine metabolism in
C. jejuni [
30]. The presence of
ggt has been shown to vary among
C. jejuni isolates [
20,
31] and we recently demonstrated that
ggt was common among human and chicken
C. jejuni isolates but significantly less common among bovine isolates [
25]. GGT activity in
C. jejuni has been suggested to affect the persistent colonization of the avian gut [
20] and in a mouse model for
C. jejuni it was shown to enhance colonization [
30]. In our study, GGT-production was present in 15% of the isolates and associated with bloody diarrhea in the univariate analysis, although the latter, due to a lack of data available, could not be further analyzed in a multivariate analysis. Interestingly, the domestic
C. jejuni isolates were able to produce GGT significantly more often than the imported isolates, a finding also verified by the multivariate analysis. PFGE typing of all domestic isolates verified the sporadic nature of the domestically acquired infections confirming also that GGT production was not linked with a certain genotype.
The gene
cj0486, encoding a putative sugar transporter and suggested to be related to invasiveness [
13], as well as the gene
ceuE, encoding a transport protein for uptake of the siderophore enterochelin [
32] were detected in 49% and 86% of the isolates in the present study, respectively, in line with some other reports [
19]. Although their presence was not associated with the outcome of the disease, interestingly they were significantly more often found among isolates of foreign than among those of domestic origin. Furthermore, the isolates lacking
ceuE seemed to cause infections requiring hospital treatment, but this finding was not verified by the multivariate analysis. In addition to the typing of all domestic isolates, PFGE typing of travel-associated isolates from July indicated that almost all patients had unique genotypes and none of the studied characteristics was linked with a genotype. Travel-associated isolates originated from a total of 36 different countries.
Only few studies have been promising in trying to show correlation between putative virulence factors or the characteristics of
C. jejuni and the outcome of the disease. Of the different
C. jejuni markers studied in the present report, GGT production and the presence of
cgtB were associated with bloody stools in the univariate analysis, but the other putative virulence factors did not correlate with any specific clinical findings. The ß-1,3 galactosyltransferase gene
cgtB in the LOS gene clusters A and B involved in the biosynthesis of ganglioside-like LOS [
16,
17] also showed a trend of being associated with longer-lasting diarrhea.
C. jejuni LOS gene clusters A, B and C have even earlier been associated not only with a more severe outcome of the disease as characterized by bloody stools and longer duration of diarrhea but also with the development of post-infectious complications [
33]. However, in the present study, the other ß-1,3 galactosyltransferase gene
wlaN, expressed in the LOS gene cluster C [
34] was not associated with any clinical characteristics. The gene
ciaB has been suggested to be involved in invasiveness [
11,
12] and thus, could be needed for the development of the disease. Indeed, all except two of the 166 isolates in our study were
ciaB positive, which is in line with earlier reports [
19,
35]. The presence of another putative virulence factor the gene
pldA, encoding phospholipase A [
14], was detected in the present study to a somewhat lower extent (61%) as compared to other reports (91-100%) [
19,
35], and did not correlate with the clinical outcome of the disease. As CDT and
virB11 were concerned, our results supported those of others showing CDT activity in the great majority [
19] but the presence of
virB11 in only a tiny proportion [
13,
19] of clinical
C. jejuni isolates. Thus, it seems very unlikely that these particular markers would play a role in the diversity of the outcome of the human disease.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
BF participated in the design of the study, collected and analyzed data, performed statistical analyses and prepared the draft manuscript. PE participated in the selection of virulence factors, performed and supervised some PCR experiments. HH performed the PFGE and conducted some PCR experiments. SS provided expertise in statistical analyses. MLH participated in the design of the study and supervised the performance of some experiments. HR participated in the design of the project, coordinated and supervised the study and helped to draft the manuscript. All authors provided ideas and comments on the draft manuscript and approved the final manuscript.