Enteroaggregative Shiga toxin-producing Escherichia coli of serotype O104:H4 in Belgium and Luxembourg

In 2011, a large outbreak of infections caused by Shiga toxin-producing Escherichia coli (STEC) O104:H4 occurred in Germany. This exceptionally virulent strain combined virulence factors of enteroaggregative E. coli (EAggEC) and STEC. After the outbreak only a few sporadic cases of infection with this rare serotype were reported, most of which were related to travel to the Middle East or North Africa. Here we describe two cases of enteroaggregative STEC (Agg-STEC) O104:H4 infection that occurred in Belgium in 2012 and 2013 respectively. In both cases travel in a Mediterranean country preceded the infection. The first strain was isolated from the stool of a 42-year-old woman presenting bloody diarrhoea, who had travelled to Tunisia the week before. The second case involves a 14-year-old girl who, upon her return from Turkey to Belgium, suffered from an episode of bloody diarrhoea and haemolytic uraemic syndrome. Extended typing of the isolates with pulsed field gel electrophoresis revealed that the strains were closely related, though not exactly the same as the 2011 outbreak strain. This report supports the previously made hypothesis that Agg-STEC has a human reservoir and might be imported by travellers coming from an area where the pathogen is endemic. Furthermore, it emphasizes the concern that these bacteria may cause future outbreaks as evenly virulent O104:H4 isolates seem to be widespread.


Introduction
O157:H7 is the most common serotype in Shiga toxin-or Vero(cyto)toxin-producing Escherichia coli (STEC/VTEC). A key property of these strains is the production of "Attaching and Effacing" (AE) lesions. STEC of different serotypes can cause severe disease in humans, including haemorrhagic colitis and haemolytic uraemic syndrome (HUS). For this reason, a widely used acronym for these E. coli strains is 'EHEC' (enterohaemorrhagic E. coli). As there is no agreement on the definition of enterohaemorrhagic strains, the acronym 'AE-STEC' will be used throughout this paper to refer to Shiga toxin-and AE-lesion-producing E. coli strains. AE-STEC O157:H7 have the highest incidence and are the most frequently involved in outbreaks (seropathotype A). AE-STEC of serotypes O26:H11, O103:H2, O111:H-, O121: H19 and O145:H-are also involved in severe human disease and outbreaks, though less frequently (seropathotype B) [1]. In 2010, 40.7% of all reported STEC serogroups in Europe were O157. O26 was the most commonly reported non-O157 serogroup (11%) [2].
From May to July 2011, a large outbreak of STEC infections with a serotype that was reported only very rarely before, O104:H4, occurred in Germany. This outbreak differed from previous STEC outbreaks by its sheer size, the large percentage of patients with HUS and the high proportion of adults, mainly women. This exceptionally virulent strain combined virulence factors of enteroaggregative E. coli (EAggEC) and STEC making it a mixed enteroaggregative and Shiga toxin-producing E. coli (designated here as 'Agg-STEC') [1,3]. At the same time, a smaller outbreak of Agg-STEC O104:H4 took place in France. Comparison of the back tracing information from the French and German outbreaks led to the conclusion that fenugreek seeds imported from Egypt were the most likely source of infection although the outbreak strain could not be isolated from the seeds [4].
After the 2011 outbreak, only a few sporadic cases of Agg-STEC of serotype O104:H4 were reported. In this paper, we describe two imported cases of Agg-STEC O104:H4 infection that occurred in Belgium after the 2011 outbreak and compare their phenotypic and genetic profiles with the German outbreak strain recovered from a Luxembourg resident who had travelled to the Hamburg area.

Case Descriptions
The first case is a 42-year-old woman who presented on the first day of September 2012 to the emergency department of the Universitair Ziekenhuis Brussel (Belgium) with complaints of heavy abdominal cramps, nausea and watery diarrhoea. She mentioned the presence of blood in the faeces. She had returned 3 days before from a 1-week journey in Tunisia where diarrhoea started in the morning, just before leaving. She was not aware of other cases. A blood analysis was normal, without signs of HUS (creatinine 0.62 mg/dL, haemoglobin 13.4 g/dL, platelets 393 000/mm³). At physical examination only discrete signs of dehydration were present. She was discharged with dietetic advice and a prescription of EnterolTM (Saccharomyces boulardii) and domperidone. She did not consult anymore, as she recovered uneventfully.
The second case involves a 14-year-old girl who was referred from the regional hospital AZ Vesalius to the Universitair Ziekenhuis Leuven (Belgium) in the third week of July 2013 because of abdominal pain, bloody stools, fever and general malaise for 4 days. Her blood examination revealed signs of HUS (anaemia with low haemoglobin 11.3 g/dL, increased levels of lactate dehydrogenase and bilirubin total 10.21 mg/dL and thrombocytopenia with platelet count 27 000/mm³), with presence of schistocytes in peripheral blood smear and decreased kidney function (oliguria, creatinine 1.07 mg/L, urea 68 mg/dL). She was treated supportively with spontaneous recovery of her kidney function. No dialysis was needed. The patient had travelled to Turkey during the first 2 weeks of July where she developed watery diarrhoea that spontaneously resolved after 5 days. Upon her return to Belgium she suffered from a new episode of diarrhoea, this time with the presence of blood in stools, and she developed HUS 6 days later.
The strains isolated from these patients were compared with an isolate related to the German outbreak. A 51-year-old woman residing in Luxembourg travelled for several days during the second week of May 2011 in the Hamburg region in Germany. Hamburg was the German state with the highest incidence (10 cases per 100 000 population) of reported HUS during the outbreak [3]. A few days after her return, she developed diarrhoea, which gradually worsened until she noticed blood in her faeces and consulted a physician who prescribed a stool culture. The patient recovered without developing HUS.

Methods and Results
In both Belgian cases a STEC O104:H4, referred to as EH2211 and EH2303, respectively, was recovered out of the stools by using methods described elsewhere [5]. Slide agglutination was performed with the K9 antiserum (Statens Serum Institut, Copenhagen, Denmark), indicative of the O104 serogroup [6]. The full serotype was confirmed by a positive PCR result for O104 (wzxO104) and H4 (fliCH4) [7]. Both strains grew on sorbitol-MacConkey agar with cefixime and tellurite (SMAC-CT), fermented sorbitol and were b-glucuronidase positive. By contrast with EH2211, EH2303 did not ferment lactose. A PCR was positive for aaiC and aggR genes [8], but negative for the STEC virulence genes intimin (eaeA) and enterohaemolysin (hlyA) [9]. Both strains possessed the Shiga toxin 2a (stx2a) subtype, as described by Scheutz et al. [10]. Disc diffusion susceptibility testing according to EUCAST recommendations (or CLSI when no breakpoints were available from the former) was performed using antibiotics selected by the Programme for Food-and Waterborne Diseases and Zoonoses (FWD) and revealed that strain EH2211 had an intermediate susceptibility to nalidixic acid and was resistant to ampicillin and trimethoprim. EH2303 on the other hand was resistant to ampicillin, amoxicillin/clavulanic acid, cefazolin, cefuroxime, cefotaxime, ceftriaxone, ceftazidime, cefepime, aztreonam, sulfonamides, trimethoprim and nalidixic acid. No extended-spectrum b-lactamase (ESBL) was detected for EH2211 according to CLSI screening Pulsed-field gel electrophoresis using XbaI was performed using the PulseNet standard operating procedure [11] to compare these two Agg-STEC O104:H4 strains with C110638, an Agg-STEC O104:H4 from Luxembourg linked to the 2011 German outbreak, and EH1651, an unrelated STEC O104:H2 strain included in the study of Miko et al. [12] (Table 1).
Strain EH2303 clustered closest to the 2011 outbreak strain with 91.9% similarity. Using a cut-off value of 80% similarity the three O104:H4 strains could be divided in the same cluster A, whereas strain EH1651 showed only 69.7% similarity and formed a separate group B (Fig. 1).
Common characteristics of EH2211, EH2303 and the German outbreak strain include: ability to grow on SMAC-CT agar; positive PCR results for aaiC, aggR, stx2a, wzxO104 and fliCH4; negative PCR results for eaeA, hlyA and stx1; agglutination with K9 antiserum; fermentation of sorbitol and production of b-glucuronidase. The outbreak strain and strain EH2303 were both multiresistant and produced an ESBL CTX-M-15 while EH2211 was susceptible to most tested antibiotics [6,22,26]. A remarkable characteristic of strain EH2303 is the fact that it did not ferment lactose. To the best of our knowledge, this is the first lactose-negative Agg-STEC O104:H4 strain ever reported.
Strain EH2303 shows more similarity to the strain from Luxembourg than EH2211 based on their pulsed field gel electrophoresis profiles. Closely related, but not identical, pulsed field gel electrophoresis XbaI patterns of E. coli O104: H4 strains isolated before and after the German 2011 outbreak have already been described [12,21].
In the cases described in this paper, travel to Turkey and Tunisia preceded the infection with Agg-STEC O104:H4, suggesting that the infection was acquired in these countries. Indeed, other sporadic cases of infection with E. coli O104:H4, as well as an outbreak associated with travel to Turkey or North Africa, have been reported in Europe. According to the information reported to European Centre for Disease Prevention and Control, four of the STEC O104 cases in the EU Member States in the period 2004-2010 were travel-related; the countries of origin of infection were Afghanistan, Egypt, Tunisia and Turkey. One of those was of the serotype O104: H4; a Finnish case with infection acquired in Egypt [26]. Infections of Agg-STEC O104 were noted in Italy (2009) and Sweden (2010). As in the present case, they were linked to travel to Tunisia [19,20]. In September 2011, an outbreak of HUS due to Agg-STEC O104:H4 occurred among French tourists returning from Turkey. Around that same period, two patients from Germany and one from Denmark were diagnosed with an Agg-STEC O104:H4 infection, also after travel to Turkey [21]. Out of four sporadic isolates of E. coli O104:H4 from children with HUS in France, one was associated with recent travel to North Africa [22].
Whereas AE-STEC is clearly a zoonosis, mainly related to ruminants, no animal reservoir has been reported for EAggEC nor Agg-STEC. Studies conducted in cattle in Germany and France during the 2011 outbreak were not able to detect the presence of the outbreak strain, indicating that cattle are not the reservoir of Agg-STEC [27,28]. As no other natural reservoir has been established, it has been hypothesized Agg-STEC has a human reservoir and might be imported into the European Union by travellers or migrants coming from an area where the pathogen is endemic [29]. Without an animal reservoir, it is not expected that Agg-STEC strains can be maintained in countries with high standards of hygiene.
Comparative genomics of E. coli O104:H4 strains isolated before, during and after the 2011 outbreak provided us with better insight into the evolution of this serotype. Genomic analysis of a number of post-outbreak O104:H4 strains has shown that they are not derived from the 2011 outbreak but share a recent common ancestor [22]. A nearly identical Shiga toxin-carrying prophage was found in several of the historical O104:H4 strains, which suggests that the prophage was transferred into the O104:H4 population at least a decade  ago. On the other hand, acquisition of the antibiotic resistance plasmid pESBL and gene cassettes for antibiotic resistances seems to be one of the most recent events in the evolutionary history [20,22]. The two cases reported here had differing antibiotic resistance profiles. Strain EH2211 was susceptible to most antibiotics, whereas EH2303 was multiresistant, as was the 2011 outbreak strain, and carried a CTX-M-15 type ESBL.
In conclusion, the cases described here support the previously expressed concern that these bacteria may cause future severe outbreaks as evenly virulent, though not exactly the same, O104:H4 isolates seem to be widespread.