Comparison of Human and Cattle E. coli O26 Isolates by Pulsed Field Gel Electrophoresis (PFGE)

In recent years, Escherichia coli (E. coli) O157, a verotoxin-producing E. coli (VTEC), has emerged as a pathogen of increasing importance, responsible for numerous outbreaks of gastro-intestinal infection worldwide. It is now apparent,however, that E. coli isolates of serogroups other than O157 (termed non-O157 VTEC) also make a significant contribution to human diarrhoeal disease and in many European countries, non-O157 VTEC are isolated more frequently than O157 VTEC.
<P>

To date, in excess of 200 non-O157 VTEC serotypes have been identified worldwide with over 100 types associated with human illness, including Haemolytic Uraemic Syndrome (HUS). Although non-O157 VTEC are not thought to not contribute significantly to the development of HUS in Scotland at present (surveillance of childhood HUS in Scotland between1997 and 2001 identified that 89% of reported cases had VTEC O157 infection), their contribution to diarrhoeal disease remains to be established.
<P>
In 2003, the Scottish E. coli O157 Reference Laboratory (SERL) received or isolated 31 non-O157 E. coli from 30 individuals, of which 16 possessed virulence factors (vtx1, vtx2, eae or hly). These included four apparently unrelated strains of E. coli O26 isolated over a six-week period. Although it was not possible to trace the source of any of the non-O157 VTEC isolated in Scotland, some of the serogroups isolated from Scottish human cases have recently been isolated from Scottish cattle.
<P>
Preliminary results of a Food Standard Agency funded project (Project SO1014)determining the prevalence of specific non-O157 VTEC in Scottish beef cattle farms has shown that VT positive E. coli O26 isolates are common in Scottish beef cattle.
<P>
Over the past few years, E. coli O26 strains have become the most frequently isolated non-O157 VTEC associated with human disease, having been reported in Europe,Japan, Asia, USA, Australia, New Zealand, Argentina and Brazil. The organism has been isolated from cattle and humans and is considered as pathogenic for both. E. coli O26 has recently been associated with outbreaks in the Republic of Ireland and Japan and is known to survive well in bovine faeces and withstand substantial periods of stress. Typing of E. coli O26, by various methods has revealed great genetic heterogeneity within the group.
<P>
It is well documented that Scottish E. coli O157 isolates responsible for clinical disease are also circulating in the cattle population, sharing common virulence profiles, the same phage types and indistinguishable pulsed-field gel electrophoresis profiles. However, although cattle are thought to be a reservoir of infection of E. coli O26 isolates, there has been no direct evidence of bovine VTEC O26 causing human infection to date.
<P>
A total of 185 E. coli O26 isolates from human (33) and cattle (152) sources were analysed in this study to determine whether common PFGE profiles of E. coli O26 are circulating in both the cattle and clinical populations of Scotland. All bovine isolates were obtained from finishing cattle aged 12 to 30 months of age collected during the Food Standards Agency funded project (SO1014). The clinical collection consisted of all human O26 strains, isolated from Scottish cases between January 2002 and October 2003 and a selection of English and European E. coli O26 isolates.
<P>
All isolates were examined for virulence determinants (vtx1, vtx2, eae & hly) and were analysed by pulsed field gel electrophoresis using Xba I. Digital images of PFGE profiles were analysed visually and with BioNumerics (v 3.0) software. BioNumerics data was further analysed using SAS (v 8.2) and differences between human and cattle isolates were evaluated by assigning isolates to statistically significant groups.
<P>
Results demonstrated that the most common virulence profile of E. coli O26 was the possession of vtx1 in conjunction with the eae gene and the hly gene (vtx1+, eae+, hly+). This was also the most common virulence profile in Scottish bovine isolates analysed in this study. This concurs with recently published data on the virulence factors of non-O157 VTEC in a cohort of Scottish beef calves and their dams and investigators have previously reported the preponderance of this profile in human isolates of E. coli O26. None of the Scottish E. coli O26 isolates investigated possessed the vtx2 gene. Although VTEC, including isolates of E. coli O26, which produce VT1, have been shown to cause HUS, there is a more established association between the production of VT2 and severe clinical disease. However, vtx genes in non-O157 VTEC are known to be unstable and stressful conditions, such as the exposure of VTEC to a number of antibiotics and growth-promoting compounds in animals may influence the horizontal transfer of vtx genes between bacteria.
<P>
Surveillance of E. coli O26 in Germany and the Czech Republic has established that they can acquire vtx2 genes over a short timeframe. It is possible that the acquisition of the vtx2 gene may then confer on E. coli O26 a greater pathogenic potential for humans.
<P>
Following PFGE analysis, the human isolates were assigned to 30 unique PFGE profile groups and the animal isolates were allocated to 88 unique PFGE profile groups, 83 of which were unique to individual farms. The human strains did not share 100 % similarity with any of the animal strains. Statistical analysis of PFGE profiles obtained in this study established that clinical isolates were widely distributed among cattle isolates. The assignment of isolates to four groups confirmed that human and cattle isolates were intermingled, with each group including E. coli O26 of human and cattle origin.
<P>
E. coli O26 occurs commonly on Scottish farms and isolates also share common virulence profiles with the clinical isolates. Although an exact interpretation of relatedness of human and animal isolates is not possible, comparison of E. coli O26 PFGE profiles has shown that, within the limits of technology used, human and cattle isolates cannot be regarded as being distinct populations, and the possibility therefore remains that Scottish cattle are a credible source of human infection of E. coli O26. Despite this, the epidemiological association between cattle and human infection is lacking and may not be readily established. The link may be elucidated by information gained from enhanced screening for all non-O157 VTEC, establishment of their reservoirs of infection and epidemiological studies to follow up clinical cases.
<P>
It is clear that further work is required to monitor the incidence of non-O157 VTEC, in humans, animals and the environment, to determine their contribution to human diarrhoeal disease and to identify their reservoirs and routes of transmission to humans.