Proposal for a New Inclusive Designation for Extraintestinal Pathogenic Isolates of Escherichia coli: ExPEC

Although Escherichia coli is probably the best-known bacterial species and one of the most common isolates in clinical microbiology laboratories, misconceptions abound with regard to the various types of E. coli and the infections that they cause. E. coli strains of biological significance to humans can be broadly classified (on the basis of both genetic and clinical criteria) into 3 major groups: commensal strains, intestinal pathogenic (enteric or diarrheagenic) strains, and extraintestinal pathogenic strains.

Commensal strains of E. coli constitute the bulk of facultative fecal flora in most healthy humans, other mammals, and birds. These strains appear to be adapted for peaceful coexistence with the host and do not appear to cause disease within the intestinal tract. Furthermore, in humans they usually do not cause disease outside the intestinal tract, except when precipitating factors are present, such as an indwelling foreign body or impairment of host defenses. Most human commensal strains originate from E. coli phylogenetic group A, as defined by multilocus enzyme electrophoresis, as do the familiar laboratory strains of E. coli (e.g., K-12 and its derivatives) [1, 2]. Commensal strains of E. coli typically lack the specialized virulence traits that are present in intestinal and extraintestinal pathogenic strains.

In contrast with commensal strains, intestinal pathogenic strains of E. coli are rarely encountered in the fecal flora of healthy hosts and, instead, appear to be essentially obligate pathogens, causing gastroenteritis or colitis when ingested in sufficient quantities by a naive host. Six distinct pathogenic categories (pathotypes) of intestinal pathogenic strains of E. coli currently are recognized: enterotoxigenic (ETEC), Shigatoxin producing/enterohemorrhagic (STEC/EHEC), enteropathogenic (EPEC), enteroinvasive (EIEC), enteroaggregative (EAEC), and diffusely adherent (DAEC); Shigella (although generally not thought of as a type of E. coli) can be included as a seventh pathotype. Although there is some overlap between certain diarrheagenic pathotypes, with respect to virulence traits, each pathotype possesses a unique combination of virulence traits that results in a distinctive pathogenic mechanism. Phylogenetic diversity exists within each of the pathotypes, the members of which are derived from phylogenetic groups A, B1, or D or from other miscellaneous lineages. Thus, the unifying theme within each of the pathotypes is not a common phylogenetic origin but a specific combination of virulence traits, which can be acquired, through horizontal transfer (e.g., plasmid or lysogenic phage), by members of distantly related evolutionary lineages. Despite their ability to cause enteric disease, these strains are mostly incapable of causing disease outside the intestinal tract [3].

Extraintestinal infections (EIs) due to E. coli are common in all age groups and can involve almost any organ or anatomical site. Typical EIs include urinary tract infection (UTI), meningitis (mostly in neonates and after neurosurgery), diverse intraabdominal infection, pneumonia (particularly in hospitalized and institutionalized patients), intravascular-device infection, osteomyelitis, and soft-tissue infection, which usually occurs when the tissue is compromised. Bacteremia can accompany infection at any of these sites [4]. Although E. coli is primarily considered to be a community-acquired pathogen, it also is the most frequently isolated gram-negative bacillus found in longterm—care facilities and hospitals [5]. Severe illness and death can occur in otherwise healthy hosts, but adverse outcomes are considerably more likely with comorbid disease and impaired host defenses.

Current understanding of the pathogenesis of EIs due to E. coli is derived almost entirely from analysis of isolates from UTIs, bacteremia, and neonatal meningitis [6, 7]. As a group, these strains are epidemiologically and phylogenetically distinct from both the commensal and the intestinal pathogenic strains; thus, they have been classified as extraintestinal pathogenic strains of E. coli. They appear to be incapable of causing enteric disease, but they can stably colonize the host intestinal tract and may constitute the predominant strain in ∼20% of healthy human hosts [8]. In contrast with intestinal pathogenic strains of E. coli, host acquisition of an extraintestinal pathogenic E. coli strain is insufficient for infection to occur. Instead, entry of the organism into an extraintestinal site (e.g., the urinary tract) is required. In the past, the ability of such strains to asymptomatically colonize the human intestinal tract caused considerable uncertainty about whether the strains actually were pathogens or merely opportunistic commensal strains.

This question has been largely resolved in favor of the “pathogen” hypothesis [9]. Unlike most commensal E. coli strains, EI isolates typically are derived from phylogenetic group B2 or D [10] and possess genes for various combinations of adhesins (e.g., P and S fimbriae), iron-acquisition systems (e.g., aerobactin), host defense—avoidance mechanisms (e.g., capsule or O-specific antigen), and toxins (e.g., hemolysin), which collectively are now regarded as extraintestinal virulence factors [7, 9]. Animal model experiments have confirmed the enhanced virulence of EI isolates, compared with that of commensal strains, and the contribution of extraintestinal virulence factors to pathogenicity [10–12]. Genes for multiple virulence factors often are present together on large blocks of chromosomal DNA, termed “pathogenicity-associated islands” (PAIs), which are analogous to (but distinct from) the diarrhea-associated PAIs that are present in many intestinal pathogenic E. coli [13].

The concentration of extraintestinal virulence genes within certain evolutionary lineages of E. coli has resulted in the concept of virulent clones. The existence of these clones was initially suggested because of the finding that certain E. coli serotypes (e.g., O4:K12:H5, O6:K2:H1, and O18:K1:H7) were strikingly overrepresented among isolates from patients with pyelonephritis, bacteremia, and/or neonatal meningitis, when compared with fecal isolates from healthy hosts. This led to the concepts of pyelonephritogenic, uropathogenic, sepsis-associated, and meningitis-associated E. coli types.

Unfortunately, these syndrome-specific designations, although valid in a limited sense, are misleadingly narrow. They imply that certain extraintestinal pathogenic E. coli strains possess site-specific virulence traits that uniquely enable the strains to cause infection at a given anatomic site. However, no single clone or virulence factor is limited to (or absolutely required for) infection at any one site or for any particular EI syndrome. This is not entirely surprising, since certain host defense mechanisms (e.g., complement and professional phagocytes), bacterial nutritional requirements (e.g., iron acquisition), and receptors for bacterial attachment (e.g., membrane glycolipids and matrix proteins) are common to many of the extraintestinal sites that E. coli typically infects. Furthermore, although extraintestinal pathogenic strains of E. coli tend to possess representatives from multiple virulent-trait categories (e.g., adhesins, protectins, etc.), the specific virulence gene profiles of isolates from any particular EI can be extremely diverse [14], suggesting that there exist multiple alternative bacterial solutions to the challenge of being a successful pathogen at various host sites.

Consequently, by analogy to the acronyms currently used for various intestinal pathogenic E. coli types, we propose the acronym “ExPEC” (extraintestinal pathogenic E. coli) as a new inclusive designation for certain strains of E. coli that cause EIs. Operationally, a clinical isolate of E. coli would be defined as an ExPEC strain if it possessed currently recognized extraintestinal virulence factors or if it demonstrated enhanced virulence in an appropriate animal model of EI. Isolation of an E. coli strain from a patient with an EI does not, by itself, confer the designation of ExPEC, since commensal strains of E. coli can cause EI when the host is compromised. Furthermore, virulence potential probably exists as a continuum and depends on the number and type of virulence factors, the inoculum size, and the presence of copathogens. Therefore, the categorical separation of commensal strains from extraintestinal pathogenic strains implied by use of the term ExPEC is undoubtedly somewhat artificial. Nonetheless, this designation is a more rational conceptualization for these strains. It avoids the misconceptions generated by use of the more traditional, but inappropriately narrow, acronyms UPEC (uropathogenic E. coli) and BMEC (bacterial meningitis E. coli). These morerestrictive designations would best be reserved for use when true site- or syndrome-specific pathotypes or clones are delineated within the larger ExPEC population.

References