Corresponding author: A.R. Manges, Associate Professor, University of British Columbia, School of Population and Public Health, 137-2206 East Mall, Vancouver, BC V6T 1Z3 Canada
Extraintestinal pathogenic Escherichia coli (ExPEC) is the most common cause of community-acquired and hospital-acquired extraintestinal infections. The hypothesis that human ExPEC may have a food animal reservoir has been a topic of investigation by multiple groups around the world. Experimental studies showing the shared pathogenic potential of human ExPEC and avian pathogenic E. coli suggest that these extraintestinal E. coli may be derived from the same bacterial lineages or share common evolutionary roots. The consistent observation of specific human ExPEC lineages in poultry or poultry products, and rarely in other meat commodities, supports the hypothesis that there may be a poultry reservoir for human ExPEC. The time lag between human ExPEC acquisition (in the intestine) and infection is the fundamental challenge facing studies attempting to attribute ExPEC transmission to poultry or other environmental sources. Even whole genome sequencing efforts to address attribution will struggle with defining meaningful genetic relationships outside of a discrete food-borne outbreak setting. However, if even a fraction of all human ExPEC infections, especially antimicrobial-resistant ExPEC infections, is attributable to the introduction of multidrug-resistant ExPEC lineages through contaminated food product(s), the relevance to public health, food animal production and food safety will be significant.
Extraintestinal pathogenic E. coli (ExPEC) is the most common cause of community-acquired and hospital-acquired extraintestinal infections, including urinary tract (UTI), kidney, bloodstream and other infections. The incidence of extraintestinal infections is thought to exceed 7 million medical visits, 1 million emergency room visits and 100 000 hospitalizations every year in the USA [
In contrast to enteric E. coli pathotypes, such as enterohaemorrhagic, enterotoxigenic, enteroaggregative or enteropathogenic E. coli (which are associated with diarrhoeal illnesses in humans, have been linked to a wide variety of contaminated foods, and have been implicated in outbreaks of human infections), ExPEC does not cause disease in the gut of colonized individuals, but rather persists in the intestine until an opportunity to cause infection presents itself (e.g. sexual intercourse in community-acquired infections or catheter use in hospital-acquired infections). It is this time lag between the acquisition and asymptomatic colonization of the intestine with an ExPEC organism and the development of an infection that presents the biggest difficulty in attributing ExPEC to specific environmental sources or reservoirs. The precise length of this lag is unclear, but it may exceed many months, making the detection of transmission events from food or environmental sources challenging. Hence, there is still some uncertainty over whether ExPEC have a food animal reservoir and are disseminated via food. Environmental E. coli that resemble the ExPEC causing human extraintestinal infections have been recovered from waterways, sewage, domestic and wild animals, soil and other environmental samples; suggesting multiple non-human reservoirs for human ExPEC [
Commonality among fluoroquinolone-resistant sequence type ST131 extraintestinal Escherichia coli isolates from humans and companion animals in Australia.
Identification, antimicrobial resistance profiles, and virulence of members from the family Enterobacteriaceae from the feces of Yellow-Headed Blackbirds (Xanthocephalus xanthocephalus) in North Dakota.
Occurrence of virulence and antimicrobial resistance genes in Escherichia coli isolates from different aquatic ecosystems within the St. Clair River and Detroit River areas.
] has also been demonstrated in several studies; indicating that humans definitely act as reservoirs for ExPEC. The fact that ExPEC may be disseminated along multiple transmission routes is not in dispute; however, the magnitude of the contribution of these various routes is not known. Given that the food-borne route is arguably the biggest contributor to the transmission of enteric E. coli pathotypes, defining the existence of food-borne transmission routes for ExPEC is an obvious research need [
The hypothesis that food, in particular poultry products, may act as a reservoir for human ExPEC is derived from multiple lines of evidence: genetic relationships between avian pathogenic E. coli (APEC) (the E. coli responsible for extraintestinal infections in birds) and human ExPEC; experimental studies showing the pathogenic potential of APEC in mammalian animal models and the pathogenic potential of human-derived ExPEC in avian animal models; molecular epidemiological data showing close genetic relationships between E. coli isolates recovered from human extraintestinal infections, poultry and retail poultry meat (and occasionally pork); and the observation that specific strains of E. coli, over short time periods and in specific communities, may cause a disproportionate number of infections (i.e. cryptic outbreaks).
One systematic review has been conducted that addresses the question of potential food-borne transmission of ExPEC, but is focused on extended spectrum β-lactamase (ESBL) -producing ExPEC [
Do human extraintestinal Escherichia coli infections resistant to expanded-spectrum cephalosporins originate from food-producing animals? A systematic review.
]. In this narrative review, the current evidence for the existence of a poultry reservoir for human ExPEC or food-borne transmission of ExPEC from poultry meat to humans is presented. The review does not focus on antimicrobial-resistant ExPEC per se; however, the majority of recent studies have focused on antimicrobial-resistant ExPEC, specifically ESBL-producing E. coli. Therefore studies of ESBL-positive ExPEC lineages tend to be over-represented in the literature. This review primarily focuses on the genetic evidence, specifically virulence genotyping, multilocus sequence typing (MLST) designation, pulsed-field gel electrophoresis (PFGE), or other molecular typing methods. The epidemiological challenges related to investigating the food–ExPEC hypothesis and limits to inferring attribution to poultry, are also described.
Extraintestinal pathogenic E. coli
Extraintestinal pathogenic E. coli are typically defined either by the presence of common virulence factors, including adhesins (e.g. P and S fimbriae), iron-acquisition systems, capsules, and toxins (e.g. haemolysin) [
], which capitalizes on known sequence variation within a set of housekeeping genes to assign a sequence type (ST), has been the classification method of choice for ExPEC recently. Classification of MLST reflects evolutionary relatedness and E. coli population structure. Phylogenetic grouping is another common classification system, where E. coli are defined by the A, B1, B2 and D phylogroups [
]. ExPEC generally fall into the B2 and D phylogroup categories, whereas groups A and B1 are more often associated with commensal E. coli. Finally, as a link to historical studies of E. coli, serotyping information is included as part of ExPEC classification, if known. A common nomenclature has emerged that describes ExPEC strains using a combination of Serotype—Phylogroup—Sequence Type, such as E. coli O25:H4-B2-ST131; this review will adhere to this ExPEC strain naming convention, or ST designations, whenever possible. Virulence gene profile-based genotyping has been performed in many studies, and is another common method of classifying E. coli into similar groups.
There are highly successful lineages or groups, many multidrug-resistant, which are responsible for the majority of human extraintestinal infections [
]. For example, E. coli O25:H4-ST131, a globally disseminated strain that has been shown to be responsible for up to 60% of all E. coli infections; and accounts for up to 78% of infections caused by fluoroquinolone-resistant and/or ESBL-producing ExPEC [
Do human extraintestinal Escherichia coli infections resistant to expanded-spectrum cephalosporins originate from food-producing animals? A systematic review.
]. Common human ExPEC STs include: 10, 12, 38, 69, 73, 95, 117, 127, 131, 394, 405 and 1193 (Fig. 1; unpublished data), although the distribution of STs responsible for infections varies by geography [
National survey of Escherichia coli causing extraintestinal infections reveals the spread of drug-resistant clonal groups O25b:H4-B2-ST131, O15:H1-D-ST393 and CGA-D-ST69 with high virulence gene content in Spain.
Fig. 1Distribution of Escherichia coli sequence types (STs) responsible for the majority of unique community-acquired urinary tract infections (n = 114) in 2012–13 in Vancouver, British Columbia (unpublished data). STs associated with poultry and retail poultry meat include E. coli ST10, ST69, ST95, ST117 and ST131.
Unlike many of the infections caused by enteric E. coli pathotypes, extraintestinal infections caused by E. coli are not usually recognized as being the result of a common-source epidemic or outbreak; however, potential community outbreaks of E. coli causing UTIs, in addition to other, more severe extraintestinal infections, have been identified in London, UK (O15:K52:H1-D-ST393) (1986), Copenhagen, Denmark (O78:H10-A-ST10) (1988), Berkeley, USA (O11:K52:H18-D-ST69) (1999), Calgary, Canada (2000) and elsewhere [
]. However, none of these outbreak investigations have identified the source for the ExPEC implicated in the outbreak.
Previous studies have identified indistinguishable PCR and PFGE patterns, suggesting that unrelated women were colonized and then infected by the same strain of ExPEC. This could occur through person-to-person contact or environmental exposures, but the more likely hypothesis was that there was a common contaminated food source in the community, which led to a large number of colonized people, from which a fraction emerged with an infection [
]. Subsequent investigations began to link poultry, and occasionally pork, to human ExPEC, with an emphasis on important multidrug-resistant ExPEC lineages such as E. coli O25:H4-B2-ST131.
Epidemiological challenges
The time lag between ExPEC colonization and infection is the fundamental challenge facing studies attempting to attribute ExPEC transmission to food, human or environmental sources. Since extraintestinal infections caused by E. coli are not generally reportable to public health institutions, increases in incidence over historic baseline rates cannot be assessed. The time lag makes traceback or case–control investigations, common to food-borne disease epidemiology, challenging. Conducting prospective cohort studies to capture and document direct ExPEC transmission events is logistically infeasible. Employing an experimental approach, whereby volunteers are orally infected with a pathogenic strain of ExPEC and then followed for intestinal colonization or infection is ethically questionable. Therefore the most useful epidemiological approaches for determining source attribution for ExPEC are unavailable. These limitations need to be kept in mind when reviewing the evidence for ExPEC reservoirs and transmission. Genomic epidemiology studies, which contrast the whole genome of E. coli recovered from food or environmental sources and E. coli recovered from human extraintestinal infections, are the best hope for determining attribution and several are currently underway. However, debate around the degree of genetic similarity based on whole genome sequences that will convincingly demonstrate attribution to a particular source promises to be lively (Fig. 2; unpublished data). Typically whole genome sequence results for E. coli strains linked to food-borne outbreaks exhibit in the order of ten single nucleotide polymorphisms or fewer [
]; whereas these ecologically sampled ExPEC isolates may vary by 100 or more SNPs. This conundrum persists from the PFGE era, when the Tenover criteria [
] (indistinguishable isolates were defined as isolates exhibiting zero PFGE band differences and closely related isolates exhibited one to three PFGE band differences) were used to infer relationships between ExPEC sampled from human and food sources, and not in the context of an outbreak investigation for which the Tenover criteria were originally defined. Finally, colonization resistance afforded by the intestinal microbiota [
] could make it difficult for incoming ExPEC from food or environmental sources to establish themselves in the gut. This might lead to transient intestinal colonization with ExPEC, rather than to long-term colonization. This argues that more recent exposures may be more relevant for epidemiological and attribution studies, although this needs to be tested.
Fig. 2Distribution of single nucleotide polymorphism (SNP) distances between pairs of isolates. Of the 6441 distinct pairwise SNP distances derived from 114 whole genome sequenced isolates analysed by Cortex
(from Fig. 1; unpublished data), 447 pairs have an SNP distance <2000 and 5994 pairs have an SNP distance ≥2000. This plot depicts the distribution of SNP distances <2000. The frequency distribution of SNP distances is shown in the histogram and non-linear least-squares fit of the SNP distances to the γ-distribution is shown in red. This figure illustrates that beyond the 10–20 SNP distances identified in a typical food-borne outbreak, there is a peak exhibiting >400 SNP distances. This may indicate some genetic relatedness between certain isolates, but drawing epidemiological inferences regarding relationships and attribution based on these data will remain challenging.
Avian pathogenic E. coli from poultry and human ExPEC
Avian pathogenic E. coli cause systemic extraintestinal infections such as aerosacculitis, polyserositis and septicaemia in chickens, turkeys and other avian hosts. APEC are generally found in the intestinal microbiota of healthy birds and infections result from environmental exposures and increased host susceptibility [
]. Genome content, virulence gene profiles, phylogeny, biofilm formation and in vivo transcriptional activation were demonstrated to be shared by APEC and human ExPEC strains (ExPEC serotypes O18:K1:H7, O78 and O2:K1:H7) [
Pathogenomic comparison of human extraintestinal and avian pathogenic Escherichia coli – search for factors involved in host specificity or zoonotic potential.
]. Other experimental studies of pathogenesis in vivo and in vitro have shown that APEC can cause disease in mammalian hosts and ExPEC recovered from human infections can likewise cause disease in avian animal models [
Acquisition of avian pathogenic Escherichia coli plasmids by a commensal E. coli isolate enhances its abilities to kill chicken embryos, grow in human urine, and colonize the murine kidney.
Avian-pathogenic Escherichia coli strains are similar to neonatal meningitis E. coli strains and are able to cause meningitis in the rat model of human disease.
Identification of potentially diarrheagenic atypical enteropathogenic Escherichia coli strains present in Canadian food animals at slaughter and in retail meats.
]. This experimental evidence, as well as the similarities between APEC and ExPEC in virulence gene transcription suggests the zoonotic potential of APEC [
Comparison of virulence factors and expression of specific genes between uropathogenic Escherichia coli and avian pathogenic E. coli in a murine urinary tract infection model and a chicken challenge model.
Comparison of extraintestinal pathogenic Escherichia coli strains from human and avian sources reveals a mixed subset representing potential zoonotic pathogens.
]. In a retrospective study of E. coli O45:K1:H7-B2-ST95 and E. coli O45:HNM/H19-D-ST371/ST2676 isolates recovered from patients with colibacillosis and from poultry from Spain and France, multiple clusters of PFGE patterns similar at >85% were detected [
Overlapped sequence types (STs) and serogroups of avian pathogenic (APEC) and human extra-intestinal pathogenic (ExPEC) Escherichia coli isolated in Brazil.
]. Comparative genomic analysis of an APEC strain O2:K1:H5-ST Complex 95 (IMT5155) and APEC O1:K1:H7 showed that they were genetically similar to human ExPEC E. coli O18:K1 strains (specifically reference strains IHE3034 and UTI89). In vivo experiments again indicated that these APEC are able to cause avian colisepticaemia and septicaemia and meningitis in a neonatal rat model [
Comparative genomic analysis shows that avian pathogenic Escherichia coli isolate IMT5155 (O2:K1:H5; ST Complex 95, ST140) shares close relationship with ST95 APEC O1:K1 and human ExPEC O18:K1 strains.
Several groups have independently identified E. coli O25:H4-ST131 and other pandemic ExPEC lineages (ST69, ST394, ST95, ST10 and ST117) in both human extraintestinal infections and in food animal or retail meat sources [
]. In Sweden, E. coli ST69, ST10, ST117 and ST10 comprised 50% of the ESBL-producing E. coli recovered from domestic chicken meat and several genetically indistinguishable (by PFGE) clusters of E. coli ST38 and ST69 were observed [
]. A large proportion of chicken meat and chickens imported into Sweden were contaminated with ESBL-producing or AmpC-producing E. coli and these resistant E. coli, primarily ST10, ST38, ST131 and ST69, were found to spread from imported parent broilers to broiler meat [
Spread of extended spectrum cephalosporinase-producing Escherichia coli clones and plasmids from parent animals to broilers and to broiler meat in a production without use of cephalosporins.
]. This indicates that the occurrence of these drug-resistant ExPEC lineages on chicken meat is due to faecal contamination at slaughter. Similarly, potential ExPEC were recovered from 22% of retail poultry meat products in Finland [
Denmark and The Netherlands have been leaders in the investigation of the relationship between poultry meat and human extraintestinal infections caused by E. coli. In the studies of isolates from human infections, domestic and imported poultry and pigs, Danish investigators identified closely related ExPEC strains based on virulence genotyping [
] and found two broiler chicken isolates and one UTI isolate that differed by only three bands by PFGE and identified one pork isolate and one E. coli isolate recovered from human stool that differed by only two bands by PFGE [
]. This investigation was strengthened by the testing of these related human- and meat-associated isolates in a mouse model of ascending UTI, demonstrating that all isolates could cause extraintestinal infections in a mammalian model. The same group identified E. coli phylogroup D-ST69 isolates in broiler and chicken meat and demonstrated their virulence and zoonotic potential in the same mouse UTI model [
Detection of clonal group A Escherichia coli isolates from broiler chickens, broiler chicken meat, community-dwelling humans, and urinary tract infection (UTI) patients and their virulence in a mouse UTI model.
]. A study from the Netherlands, focused on ESBL-producing E. coli, identified four sets of E. coli isolates of human and poultry or retail chicken meat origin with indistinguishable ESBL genes (blaCTX-M-1 and blaTEM-52), plasmids and MLST genotypes (ST10, ST58, ST117 and ST10) [
]. A second study from the Netherlands also identified ESBL-positive E. coli recovered from human infections, human stool and retail chicken meat, which all shared common STs, including ST10, ST117, ST168, ST156 and ST23 [
]. Giufrè et al., in Italy also recovered genetically similar E. coli (ST410, ST23, ST559, ST617 and ST10) from UTIs, bloodstream infections and poultry meat isolates [
Several studies in North America have investigated the role of retail meat in human ExPEC infections. In a US study of retail meat isolates (specific commodities were not analysed separately), human clinical infections and stool, ExPEC was identified in 14% of isolates tested, virulence characteristics appeared similar, but PCR-based and PFGE patterns were not related [
Molecular analysis of antimicrobial-susceptible and -resistant Escherichia coli from retail meats and human stool and clinical specimens in a rural community setting.
Broiler chickens, broiler chicken meat, pigs and pork as sources of ExPEC related virulence genes and resistance in Escherichia coli isolates from community-dwelling humans and UTI patients.
], suggesting that beef cattle are not a reservoir for human ExPEC. Indirect epidemiological associations have also been observed between development of multidrug-resistant UTI and report of frequent chicken or pork meat consumption [
Food-borne disease and antimicrobial resistance surveillance systems such as the Canadian Integrated Programme for Antimicrobial Resistance Surveillance, the US National Antimicrobial Resistance Monitoring System, Danish Integrated Antimicrobial Resistance Monitoring and Research Programme, and other national surveillance systems often include the collection of ‘generic’ E. coli as part of their surveillance programmes. These ‘generic’ E. coli are thought to represent E. coli present in the food animals themselves; however, more detailed information linking E. coli population dynamics in the animals and those E. coli recovered through retail meat sampling is needed. These surveillance programmes characterize E. coli in part to measure global resistance in non-pathogenic bacteria and also to test for the presence of possible faecal contamination or to detect sanitation issues in meat processing. Several of the studies highlighted in this review leveraged these national surveillance systems to study the ‘generic’ E. coli from retail meat, finding that a fraction of these ‘generic’ E. coli may actually represent human ExPEC organisms, exhibiting MLST and virulence factor profiles that overlap with ExPEC from human infections.
Conclusions
Results from multiple investigations from around the world implicate poultry, and possibly pork, but not other meat commodities, as a potential reservoir for human ExPEC. The direct transmission of poultry-associated ExPEC to humans has not been observed, and is not likely to be observed due to the lag between ExPEC acquisition and infection and the epidemiological challenges alluded to earlier. Instead, whole genome sequencing and comparative genomics analysis of E. coli recovered from food animals, retail meat and human infections may offer the discrimination necessary to address whether and how much the poultry reservoir contributes to the burden of human ExPEC infections. This approach will still pose challenges, given the degree of genetic variation exhibited by ExPEC from within and across multiple sources (Fig. 2; unpublished data). The totality of the evidence presented in this review is compelling. First, multiple independent, experimental studies of extraintestinal disease pathogenesis demonstrate that human ExPEC, when introduced into avian disease models, behaves like APEC; and APEC, when introduced into a mammalian model of human extraintestinal infection, behaves like human ExPEC. Second, the consistent observation of certain E. coli STs associated with human ExPEC infections (ST10, ST23, ST69, ST95, ST117 and ST131) in poultry by multiple investigators around the world strongly supports the hypothesis that specific lineages of human ExPEC are likely to have a reservoir in poultry and poultry meat. These lineages may be the primary contributors to food-borne ExPEC transmission, possibly because they exhibit less host specificity, a notion put forward several decades ago by Bettelheim et al. [
]. These lineages may then contribute to cryptic outbreaks, where increases in the proportion of extraintestinal infections caused by one or a few highly related strains may emerge over a short time in one region [
There is no disagreement that human ExPEC reservoirs may also exist in water, sewage, humans and other environmental sources; although, many would argue that one of the bigger contributors to ExPEC transmission is likely to be food [
]. It is difficult to estimate how much ExPEC exchange can be attributed to person-to-person contact. However, this route probably plays a prominent role in the dissemination of more host-specific lineages of ExPEC. For example, O6:K15:H31-B2-ST92 and O6:H1-B2-ST73, which tend to be recovered exclusively from human sources, and rarely from poultry products or other food sources, suggesting that more human-adapted ExPEC lineages exist. These lineages may then contribute to sporadic and endemic human extraintestinal infections [
If attributing a portion of human ExPEC infections to food sources is difficult, then attempting to attribute ExPEC transmission and human infections to other reservoirs will be equally if not more challenging. Investigators have suggested using methods such as Bayesian network modelling or model-guided fieldwork to estimate the contribution of these other non-food sources [
]; however, these methods still rely on data for model building that are weak or non-existent, and they do not address the fundamental problem associated with attribution when there is a lag between exposure and disease. This is in contrast to diseases such as campylobacteriosis [
The focus on poultry as a major potential source for ExPEC is the result of significant public health concern. If the increase in antimicrobial-resistant extraintestinal infections caused by E. coli is attributable to the introduction of new multidrug-resistant ExPEC lineages through contaminated food product(s), then the relevance to public health, food animal production and food safety would be significant.
Acknowledgements
This manuscript was supported by funding from the Canadian Institutes of Health Research MOP-114879 to ARM. I would like to thank Eugenia Wong for her help in completing the literature review.
Transparency declaration
The authors declare that they have no conflicts of interest.
References
Foxman B.
Epidemiology of urinary tract infections: incidence, morbidity, and economic costs.
Commonality among fluoroquinolone-resistant sequence type ST131 extraintestinal Escherichia coli isolates from humans and companion animals in Australia.
Identification, antimicrobial resistance profiles, and virulence of members from the family Enterobacteriaceae from the feces of Yellow-Headed Blackbirds (Xanthocephalus xanthocephalus) in North Dakota.
Occurrence of virulence and antimicrobial resistance genes in Escherichia coli isolates from different aquatic ecosystems within the St. Clair River and Detroit River areas.
Do human extraintestinal Escherichia coli infections resistant to expanded-spectrum cephalosporins originate from food-producing animals? A systematic review.
National survey of Escherichia coli causing extraintestinal infections reveals the spread of drug-resistant clonal groups O25b:H4-B2-ST131, O15:H1-D-ST393 and CGA-D-ST69 with high virulence gene content in Spain.
Pathogenomic comparison of human extraintestinal and avian pathogenic Escherichia coli – search for factors involved in host specificity or zoonotic potential.
Acquisition of avian pathogenic Escherichia coli plasmids by a commensal E. coli isolate enhances its abilities to kill chicken embryos, grow in human urine, and colonize the murine kidney.
Avian-pathogenic Escherichia coli strains are similar to neonatal meningitis E. coli strains and are able to cause meningitis in the rat model of human disease.
Identification of potentially diarrheagenic atypical enteropathogenic Escherichia coli strains present in Canadian food animals at slaughter and in retail meats.
Comparison of virulence factors and expression of specific genes between uropathogenic Escherichia coli and avian pathogenic E. coli in a murine urinary tract infection model and a chicken challenge model.
Comparison of extraintestinal pathogenic Escherichia coli strains from human and avian sources reveals a mixed subset representing potential zoonotic pathogens.
Overlapped sequence types (STs) and serogroups of avian pathogenic (APEC) and human extra-intestinal pathogenic (ExPEC) Escherichia coli isolated in Brazil.
Comparative genomic analysis shows that avian pathogenic Escherichia coli isolate IMT5155 (O2:K1:H5; ST Complex 95, ST140) shares close relationship with ST95 APEC O1:K1 and human ExPEC O18:K1 strains.
Spread of extended spectrum cephalosporinase-producing Escherichia coli clones and plasmids from parent animals to broilers and to broiler meat in a production without use of cephalosporins.
Detection of clonal group A Escherichia coli isolates from broiler chickens, broiler chicken meat, community-dwelling humans, and urinary tract infection (UTI) patients and their virulence in a mouse UTI model.
Molecular analysis of antimicrobial-susceptible and -resistant Escherichia coli from retail meats and human stool and clinical specimens in a rural community setting.
Broiler chickens, broiler chicken meat, pigs and pork as sources of ExPEC related virulence genes and resistance in Escherichia coli isolates from community-dwelling humans and UTI patients.
00992-1&id=gr1.jpg){kind=link}
00992-1&id=gr2.jpg){kind=link}