Strains of Escherichia coli associated with urogenital disease in dogs and cats

Selected strains of Escherichia coli associated with urogenital disease in dogs and cats were evaluated for 3 virulence factors associated with human uropathogenic strains. Urogenital strains of E coli from dogs and cats had high prevalence of alpha hemolysin and were clustered in 5 to 10 somatic serogroups, attributes also shared by human uropathogenic strains. However, the canine and feline urogenital strains failed to have increased prevalence of mannose-resistant hemagglutination, as has been reported for human uropathogenic strains.     

Serotypes and virulence factors of Escherichia coli strains isolated from dogs and cats

E. coli strains isolated from urine of dogs and cats with urinary tract infections (UTI) and from feces of healthy one's were serotyped, and the serotypes were correlated with uropathogenic virulence factors. The most prevalent O-serotypes, O4 and O6, were isolated from dogs and cats with UTI. In contrast, O11 and O102 strains were the most frequently found from feces of healthy dogs and cats. Most of type O4 and O6 strains possessed such virulence factors as pil, pap, sfa, hly, and cnf1, while most type O11 and O102 strains pil only or pil and aer. All strains of type O75 possessed afaI and aer. K1 antigen was negative in all strains obtained from UTI.     

Uropathogenic virulence factors in isolates of Escherichia coli from clinical cases of canine pyometra and feces of healthy bitches

Escherichia coli is commonly isolated in canine pyometra, but little is known of the virulence factors that may be involved in the precipitation of this disease. The aim of this study was to compare the prevalence of uropathogenic virulence factor (UVF) genes in E. coli isolates from canine pyometra and from feces of healthy bitches to evaluate their role in the pathogenesis of pyometra. E. coli from 23 cases of canine pyometra and from the feces of 24 healthy bitches were analyzed, by polymerase chain reaction, for UVF genes associated with canine and human urinary tract infections (UTIs). The prevalences of UVFs in E. coli from canine pyometra were similar to that in canine and human uropathogenic E. coli. The prevalence of pap was greater (P=0.036) for E. coli from pyometra (52%) than for fecal isolates (21%), and the papGIII allele was present in all pap-containing isolates. The prevalences of genes for alpha-haemolysin and cytotoxic necrotising factor 1 were not significantly higher (P=0.075) in E. coli from pyometra than from feces. The proportion of pyometra strains with >or=3 UVFs was higher (P=0.039) than that of fecal strains, suggesting that possession of >or=3 UVF genes enhances the pathogenicity of the strain. Our findings demonstrate that E. coli associated with canine pyometra are similar to uropathogenic strains, and that operons that encode P fimbriae, alpha-haemolysin and cytotoxic necrotising factor 1 probably enhance the virulence and pathogenicity of the strain in the canine genital tract.     

Haemolytic Escherichia coli isolated from dogs with diarrhea have characteristics of both uropathogenic and necrotoxigenic strains

Twenty-four haemolytic Escherichia coli strains were isolated from dogs with diarrhea. The strains were serotyped and analysed by polymerase chain reaction (PCR) for genes encoding virulence factors associated with E. coli that cause diarrhea in animals. Adhesion antigen production was deduced from haemagglutination experiments. Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) of heat extracts was also used as an indication for the production of adhesive structures. The majority of the strains was shown to produce this type of virulence factor. Adhesion and invasion tests of the strains and Caco-2 cells showed that all strains adhered and that two were invasive. The two invasive strains were positive in the intimin PCR and one of them also contained genes encoding CS31A. The PCR for heat stable toxin (ST) was positive in only four strains, as was the presence of F17 fimbrial genes. Surprisingly, 19 strains had intact P fimbrial operons, coding for an adhesin involved in urinary tract infection (UTI). The cytotoxic necrotising factor 1 (CNF1) gene, also mainly found in UTI was likewise detected in these 19 strains. Cytolethal distending toxin (Cdt) genes were found in five strains. The high number of strains positive for CNF1 and P fimbriae prompted us to test the strains in a multiplex PCR used to test E. coli isolated from UTI in various species for 30 virulence associated genes. The data showed that the majority of the diarrhea isolates have virulence factor profiles highly similar to UTI E. coli isolates from dogs. This raises the question whether these isolates are real intestinal pathogens or "innocent bystanders". However, since CNF1 producing necrotoxic E. coli (NTEC) strains isolated from humans, pigs and calves with diarrhea appear to be highly related to our strains, it might be that in dogs this type of isolate is capable of causing not only UTI, but also diarrhea. If this is the case and this type of isolate is "bifunctional", domestic animals likely constitute a reservoir of NTEC strains which can be also pathogenic for humans.     

Analysis of the AIDA-I gene sequence and prevalence in Escherichia coli isolates from pigs with post-weaning diarrhoea and oedema disease

Three hundred and twenty-four strains of Escherichia coli isolated from weaned pigs with diarrhoea or oedema disease in Eastern China were screened by multiplex PCR for the presence of the gene encoding adhesin involved in diffuse adhesion I (AIDA-I). Two AIDA-I positive strains were subjected to analysis of the nucleotide sequence of the complete orfA and orfB of the AIDA gene. The AIDA-I positive E. coli isolates were also assessed for five fimbriae (F4, F5, F6, F18 and F41) by monoclonal antibodies and for toxin genes (STa, STb, LT, EAST1, Stx2e) by PCR. Twenty-one (6.5%) of the isolates possessed AIDA-I genes. Of these isolates, two carried AIDA-I genes as the only demonstrated virulence factors, and the remaining isolates carried other virulence factor genes. Comparing the AIDA-I sequence from porcine and human sources, a high homology of orfA both in porcine E. coli and human E. coli was observed. However, each orfB of the two porcine E. coli isolates was 3864 nucleotides long compared with 3861 for the E. coli 2787 orfB, and showed 96.5% homology to E. coli 2787. The data indicated (1) that AIDA-I may be an occasional virulence factor in post-weaning diarrhoea and oedema disease in pigs, (2) that it has the potential to transfer between porcine and human E. coli, and (3) that there is a genetic diversity in orfB between human and porcine E. coli.     

Virulence genes and P fimbriae PapA subunit diversity in canine and feline uropathogenic Escherichia coli.

In this study, a total of 118 Escherichia coli strains isolated from dogs (93) and cats (25) with urinary tract infection (UTI) were tested in a multiplex polymerase chain reaction for the presence of adhesin-encoding genes (pap, sfa, and afa), hemolysin encoding genes (hly), cytotoxic necrotizing factor 1 (cnf1) and aerobactin (aer) genes. Virulence gene frequencies detected in those isolates which had been randomly collected (68 canine strains) were: 43% pap, 57% sfa, 1% afa, 44% hly, 41% cnf1 and 34% aer. These frequencies were much higher in the remaining 50 hemolytic strains of either cat or dog origin. Virulence factor associations in the 80 hemolytic strains studied revealed that 50/80 simultaneously had two adhesin genes (pap and sfa) and two cytotoxin genes (hly and cnf1), and 15/80 in addition had the aer gene. The major structural subunit and antigenic determinant of P fimbriae of uropathogenic E. coli is PapA. Polymorphism in this subunit was studied by an F antigen-specific papA allele polymerase chain reaction in 51 canine and 22 feline pap positive E. coli strains. The most prevalent canine papA alleles were F10 (39%), F15 (37%) and F12 (35%). In feline strains F15 (50%) was more frequent, other allele frequencies were F12 (45%), F14 and F10 (27%) and F16 (23%). Only nine canine and two feline strains were negative for one of the 11 serologically defined F types of P fimbriae. Three copies of the pap operon were found in 16/51 canine and 9/22 feline UTI E. coli pap positive strains. In this study, we show that a particular combination of virulence genes appears with high frequency in dog and cat urinary tract E. coli strains (pap, sfa, hly, and cnf1). In spite of the more frequent presence of F10, F12 and F15 papA alleles in this virulence gene combination, the occurrence of different papA alleles in strains where up to three copies of the pap operon are present accounts for the observed P fimbriae diversity.     

Porcine intestinal epithelial cell lines as a new in vitro model for studying adherence and pathogenesis of enterotoxigenic Escherichia coli

Enterotoxigenic Escherichia coli (ETEC) infections result in large economic losses in the swine industry worldwide. The organism causes diarrhea by adhering to and colonizing enterocytes in the small intestines. While much progress has been made in understanding the pathogenesis of ETEC, no homologous intestinal epithelial cultures suitable for studying porcine ETEC pathogenesis have been described prior to this report. In the current study, we investigated the adherence of various porcine ETEC strains to two porcine (IPEC-1 and IPEC-J2) and one human (INT-407) small intestinal epithelial cell lines. Each cell line was assessed for its ability to support the adherence of E. coli expressing fimbrial adhesins K88ab, K88ac, K88ad, K99, F41, 987P, and F18. Wild-type ETEC expressing K88ab, K88ac, and K88ad efficiently bound to both IPEC-1 and IPEC-J2 cells. An ETEC strain expressing both K99 and F41 bound heavily to both porcine cell lines but an E. coli strain expressing only K99 bound very poorly to these cells. E. coli expressing F18 adhesin strongly bound to IPEC-1 cells but did not adhere to IPEC-J2 cells. The E. coli strains G58-1 and 711 which express no fimbrial adhesins and those that express 987P fimbriae failed to bind to either porcine cell line. Only strains B41 and K12:K99 bound in abundance to INT-407 cells. The binding of porcine ETEC to IPEC-J2, IPEC-1 and INT-407 with varying affinities, together with lack of binding of 987P ETEC and non-fimbriated E. coli strains, suggests strain-specific E. coli binding to these cell lines. These findings suggest the potential usefulness of porcine intestinal cell lines for studying ETEC pathogenesis.     

Molecular investigation of Escherichia coli strains associated with apparently persistent urinary tract infection in dogs

Persistent Escherichia coli urinary tract infection (UTI) in dogs is a frustrating clinical problem. Affected dogs often appear to fail to respond to therapy or to reacquire infection shortly after therapy is completed. Urovirulence factors (UVFs) of the infecting E. coli, antibiotic resistance, and tissue colonization may be contributory but have not been evaluated in dogs with persistent E. coli UTI. In this study, the strain types of E. coli in dogs with persistent UTI were evaluated with pulsed-field gel electrophoresis (PFGE) to determine whether persistence was due to acquisition of new isolates or failure to eradicate existing isolates. UVFs in these isolates, assessed by polymerase chain reaction, and antibiograms were correlated with treatment outcome in these dogs. Results documented a mixed pattern: 9 dogs remained chronically infected with 1 or 2 strains, each with distinct reproducible UVFs, but 1 dog was infected with numerous unrelated E. coli strains over time. Two dogs had a mixed pattern, consisting of 1 or more episodes of persistent E. coli infection attributable to a single strain in addition to episodes caused by unrelated strains. Many isolates had no detectable UVFs, highlighting the likely importance of impaired colonization resistance in the affected dogs. Antibiotic resistance was common, often in response to previous treatments, especially with trimethoprim-sulfamethoxazole. Antibiotic resistance patterns differed significantly within PFGE strain types, suggesting lateral acquisition of resistance plasmids or integrons. These results can be used to help guide testing for and management of persistent E. coli UTI in dogs.     

Identification of a new Escherichia coli She haemolysin homolog in avian E. coli

Haemolysin is one type of virulence factor that assists in the pathogenesis of Escherichia coli. Currently, hemolytic activity in E. coli has been attributed to haemolysin genes found in either uropathogenic or enterohemorrhagic E. coli. Both haemolysins are classified as RTX toxins because they both have repeats in toxin domains and share similar operon organization, sequence homology, and mechanisms of action. Haemolytic avian E. coli isolates, however, lack either E. coli haemolysin gene. To investigate the avian E. coli haemolysin, a genomic library was made from an avian pathogenic E. coli. A haemolytic clone that was isolated was shown to contain homology with sheA, an E. coli K- 12 gene which causes haemolysis when present in high copy number. The cloned haemolysin gene, hlyE, lacked the conserved amino acid sequence and accessory genes common to all RTX toxins. DNA hybridizations and polymerase chain reaction amplifications showed that the nucleotide sequences homologous to hlyE were not present in a collection of three O157: H7 E. coli, five haemolytic canine uropathogenic E. coli, one haemolytic O26 E. coli, and three haemolytic avian pathogenic E. coli. Thus we have identified a new E. coli haemolysin distinct from the RTX haemolysins and have shown that some avian pathogenic E. coli possess a haemolysin with no apparent homology to hlyE or RTX haemolysins.     

Virulence and fitness gene patterns of Shiga toxin-encoding Escherichia coli isolated from pigs with edema disease or diarrhea in Germany

Fecal Escherichia coli isolates (n = 3,218) from piglets with edema disease or diarrhea were screened for the genes of Stx2 and Stx2 variants. A total of 283 E. coli isolates (8.8%) proved exclusively positive for Stx2e and most of these (85.1%) harbored genes for F18 fimbria. No recognized adhesins were detectable in 14.5% of the isolates. Genes for heat-stable or heat-labile E. coli enterotoxins were found in F18+ as well as F18 isolates (51.9% and 33.3%, respectively). Five isolates also harbored fyuA and irp2 genes which are indicative of a high pathogenicity island in E. coli. All Stx2e+ isolates lacked genes for intimin, EHEC hemolysin, STEC autoagglutinating adhesin, subtilase cytotoxin, serine protease Espl. The majority of Stx2e+ isolates belonged to phylogenetic groups A (59.3%) and D (38.9%) and only few isolates were classified as B1 and B2 (1.8%). The results suggest that Stx2e-producing E. coli strains are highly prevalent in diseased pigs in Germany. Despite their significant diversity, most strains possess all typical features (Stx2e, F18) of porcine edema disease E. coli. However, a considerable portion of porcine strains resembles published human Stx2e+ strains in that they lack any recognized pig-associated adhesin. Thus, a zoonotic potential cannot be excluded for these strains.     

Serotypes and intimin types of intestinal and faecal strains of eae+ Escherichia coli from weaned pigs

Enteropathogenic Escherichia coli (EPEC) that are known to cause severe diarrhoea in children and young rabbits are well characterized, but there are few reports on the serotypes and intimin (eae) types of EPEC in weaned pigs. Based on detection of the eae gene by PCR and by DNA-hybridisation with LEE specific gene probes, 20 intestinal and 17 faecal eae(+) strains from diarrhoeal (164) and non-diarrhoeal (57) weaned pigs from 13 Hungarian farms, representing 12.8% of diarrhoeal and 14.0% of non-diarrhoeal pigs, were identified. The dominant serotype was O123:H11 (40%) among intestinal, and O108:H9 (23%) among faecal strains. The majority (85%) of the intestinal strains possessed eae-beta and 10% carried eae-gamma gene. In contrast, significantly (p<0.025) fewer faecal strains (53%) harboured the eae-beta gene, and 23% were eae-gamma positive. In vitro adhesion tests of intestinal and faecal eae(+) strains indicated adhesion of 20/37 of the strains to PK15 (porcine kidney) cells while only 3/37 strains adhered to HeLa cells. The ultrastructure of intimate bacterial attachment of representative porcine eae(+) strains to PK15 cells showed no pedestal formation, in contrast to the human EPEC (O127:H5, eae-alpha) strain. In conclusion, the data do not demonstrate a significant role for the eae(+)E. coli in porcine post-weaning diarrhoea, but provide new information on a dominant porcine serotype (O123:H11, eae-beta), and on differences of serotypes and intimin types of porcine eae(+) strains according to their site of isolation. Furthermore there was an indication that the PK15 cell line could be used as a model to study in vitro adherence of eae(+)E. coli of some human and porcine origin.     

Antibiotic sensitivity profiles do not reliably distinguish relapsing or persisting infections from reinfections in cats with chronic renal failure and multiple diagnoses of Escherichia coli urinary tract infection

Older cats with chronic renal failure (CRF) commonly develop urinary tract infections (UTI). Uropathogenic Escherichia coli (UPEC) is identified as the causal agent of UTI in most affected cats. Infections are often complicated, and UPEC infections may persist or recur in these cats. Antibiotic sensitivity profiles have been used to distinguish relapsing or persisting UTI from reinfection by different clones of the same species. However, the accuracy with which antibiograms discriminate different urinary E coli clones in cats is uncertain. We studied 17 cystocentesis-derived UPEC isolates collected from 5 cats with stable CRF and multiple diagnoses of UTI. UTIs were classified as relapses versus persistent infections or reinfections using antibiograms determined by Kirby-Bauer discs and Etests. Subsequently, clonality of UPEC isolates was determined by pulsed-field gel electrophoresis (PFGE). A comparison of PFGE results with antibiograms indicated that antibiotic resistance patterns varied considerably within several individual E coli clones. Both antibiotic susceptibility tests differentiated between relapsing or persistent infections and reinfections with only 58% overall efficiency. Thus, antibiotic sensitivity profiles cannot be relied upon to distinguish between persisting or relapsing infections as compared to reinfections in cats with CRF and multiple diagnoses of E coli UTI.     

Rapid detection of virulence-associated genes in avian pathogenic Escherichia coli by multiplex polymerase chain reaction

Based on recently published prevalence data of virulence-associated factors in avian pathogenic Escherichia coli (APEC) and their roles in the pathogenesis of colibacillosis, we developed a multiplex polymerase chain reaction (PCR) as a molecular tool supplementing current diagnostic schemes that mainly rely on serological examination of strains isolated from diseased birds. Multiple isolates of E. coli from clinical cases of colibacillosis known to possess different combinations of eight genes were used as sources of template DNA to develop the multiplex PCR protocol, targeting genes for P-fimbriae (papC), aerobactin (iucD), iron-repressible protein (irp2), temperature-sensitive hemagglutinin (tsh), vacuolating autotransporter toxin (vat), enteroaggregative toxin (astA), increased serum survival protein (iss), and colicin V plasmid operon genes (cva/cvi). In order to verify the usefulness of this diagnostic tool, E. coli strains isolated from fecal samples of clinically healthy chickens were also included in this study, as were uropathogenic (UPEC), necrotoxigenic, and diarrhegenic E. coli strains. The application of the multiplex PCR protocol to 14 E. coli strains isolated from septicemic poultry showed that these strains harbored four to eight of the genes mentioned above. In contrast, those isolates that have been shown to be nonpathogenic for 5-wk-old chickens possessed either none or, at most, three of these genes. We found only one enterohemorrhagic (EHEC), one enteropathogenic (EPEC), and two enterotoxic (ETEC) E. coli strains positive for irp2, and another two ETEC strains positive for astA. As expected, UPEC isolates yielded different combinations of the genes iss, papC, iucD, irp2, and a sequence similar to vat. However, neither the colicin V operon genes cva/cvi nor tsh were amplified in UPEC isolates. The multiplex PCR results were compared with those obtained by DNA-DNA-hybridization analyses to validate the specificity of oligonucleotide primers, and the protocol was concluded to be a useful, sensitive, and rapid assay system to detect avian pathogenic E. coli and differentiate them from nonpathogenic strains and those belonging to other pathotypes.     

Antibacterial activity of antisera against homologous and heterologous Escherichia coli of porcine origin.

Fourteen enteropathogenic and five nonenterotoxigenic Escherichia coli strains isolated from pigs were used for producing antisera in rabbits and pigs. These antisera were used in an vitro test system for antibacterial activity against homologous and heterologous porcine E. coli strains. Antibacterial titres were determined against the homologous strains and the percent reduction in CFU/ml caused by a 1/200 dilution of the sera against heterologous strains was determined. The results indicated that following immunization the antibacterial activity of serum against homologous and heterologous strains was significantly increased. This activity did not appear to be influenced by O and K antigen relationships among the organisms or by enterotoxigenicity of the vaccine strains. When antiserum produced against a combination of three enteropathogenic E. coli was tested against 20 strains a wider spectrum of heterologous antibacterial activity was obtained than with antiserum produced against any individual strain. The results indicate the existence in E. coli strains of porcine origin of common antigenic determinants not related to the serological formula and that a selected combination of strains can be expected to induce antibacterial acitivity against a wide variety of serological types of porcine enteropathogenic E. coli.     

Conservation of deduced amino acid sequence of FimH among Escherichia coli of bovine, porcine and avian disease origin

The FimH subunit of type 1 pili mediates adhesion of Escherichia coli to epithelium in different animal hosts. In this study, we sequenced and analyzed the fimH genes of 24 E. coli strains from bovine and porcine clinical cases. The obtained sequences were compared among each other and also with 24 known fimH sequences from avian E. coli strains. This comparison revealed a substantial homology (>99%) among strains from the different animal species origins. Moreover, specific mutations were found, some of which were present more frequently in avian strains or in bovine and porcine strains.     

Differentiation of avian pathogenic Escherichia coli (APEC) strains by random amplified polymorphic DNA (RAPD) analysis

Here we describe the application of a random amplified polymorphic DNA (RAPD) analysis for molecular genetic typing avian pathogenic Escherichia coli (APEC) strains. The RAPD technique was shown to be highly reproducible. Stable banding patterns with a high discriminatory capacity were obtained using two different primers. Overall, 55 E. coli strains were analyzed with a RAPD technique. The RAPD analysis showed that the E. coli strains isolated from poultry in Thailand and Sweden could be grouped into 50 of RAPD types by using these two different primer sets. Most of these different E. coli RAPD types were not geographically restricted. There was, as expected, a tendency of higher genetic relationship among E. coli strains isolated from the same farm. It is suggested that the RAPD technique may provide a rapid, low cost, simple and powerful tool to study the clonal epidemiology of avian E. coli infections.     

Adherent and invasive Escherichia coli are associated with persistent bovine mastitis

Bovine mastitis caused by Escherichia coli has traditionally been viewed as a transient infection. However, E. coli can also cause clonal persistent intramammary infection (IMI) in dairy cows. In this study, we explored the possibility that E. coli strains associated with persistent IMI are better able to adhere to, invade, survive and replicate in cultured mammary epithelial cells (MAC-T) than transient strains, and examined their serotype, overall genotype, phylogenetic group, and the presence of known virulence genes. Both transient and persistent E. coli strains adhered to MAC-T cells, but persistent strains invaded MAC-T cells 2.6-63.5 times more than transient strains. Blocking the adhesin/invasin FimH with mannose diminished but did not eliminate adhesion and invasion of any strain. Cytoskeletal and protein kinase inhibitors cytochalasin D, colchicine, genistein and wortmannin dramatically reduced invasion of MAC-T cells by both strains. All of the persistent strains, but only one transient strain, were able to survive and replicate intracellularly in MAC-T cells over 48 h. Transient and persistent strains displayed heterogeneous serotypes and overall genotypes, but similar phylogeny (group A), and lacked virulence genes of invasive E. coli. We have found that E. coli strains associated with persistent IMI are better able to invade and replicate within cultured mammary epithelial cells than transient strains. The invasion process involves the host cytoskeleton and signaling cascades and is not FimH dependent. Our findings suggest that the invasion of mammary epithelial cells and intracellular survival play an important role in the pathogenesis of persistent E. coli mastitis.     

Comparison of ruminant and human attaching and effacing Escherichia coli (AEEC) strains

The presence of 12 genes associated with virulence in human attaching and effacing Escherichia coli (AEEC) was studied within a collection of 20 enterohemorrhagic E. coli (EHEC) and 206 atypical enteropathogenic E. coli (EPEC) isolated from ruminants. In addition, virulence genes and the clonal relationship of 49 atypical EPEC O26 strains isolated from humans and ruminants were compared to clarify whether ruminants serve as a reservoir of atypical EPEC for humans. A great diversity in the content of virulence gene was found. Thus, the espH, espG and map genes were detected in more than 85% of ruminant AEEC strains; the tccP2, espI, efa1/lifA, ehxA and paa genes were present in 50-70% of strains; and other genes such as tccP, espP, katP and toxB were detected in <25% of strains. EHEC strains contained more virulence genes than atypical EPEC strains. Our results suggest for the first time that the efa1/lifA gene is associated with diarrhea in newborn ruminants and that the AEEC strains with the H11 flagellar antigen are potentially more virulent than the non-H11 AEEC strains. Importantly, we identified a new intimin variant gene, eaeρ, in three ruminant atypical EPEC strains. The comparison of ruminant and human EPEC O26 strains showed that some ruminant strains possess virulence gene profiles and pulse-field gel electrophoresis pulsotypes similar to those of human strains. In conclusion, our data suggest that atypical EPEC is a heterogeneous group with different pathogenic potential and that ruminants could serve as a reservoir of atypical EPEC for humans.     

Enrofloxacin resistance in Escherichia coli isolated from dogs with urinary tract infections

OBJECTIVE: To assess the strain heterogeneity of enrofloxacin-resistant Escherichia coli associated with urinary tract infections in dogs at a veterinary medical teaching hospital (VMTH). In addition, strains from other veterinary hospitals in California were compared with the VMTH strains to assess the geographic distribution of specific enrofloxacin-resistant E. coli isolates. DESIGN: Bacteriologic study. SAMPLE POPULATION: 56 isolates of E. coli from urine samples (43 isolates from dogs at the VMTH, 13 isolates from dogs from other veterinary clinics in California). PROCEDURES: Pulsed field gel electrophoresis was performed on 56 isolates of E. coli from urine samples from 56 dogs. All 56 isolates were tested for susceptibility to amoxicillin, chloramphenicol, enrofloxacin, tetracycline, trimethoprim-sulphamethoxazole, cephalexin, and ampicillin. Enrofloxacin usage data from 1994 to 1998 were obtained from the VMTH pharmacy. RESULTS: Several strains of enrofloxacin-resistant E. coli were collected from urine samples from the VMTH, and strains identical to those from the VMTH were collected from other veterinary clinics in California. For the isolates that did share similar DNA banding patterns, variable antibiotic resistance profiles were observed. CONCLUSIONS AND CLINICAL RELEVANCE: The increased occurrence of enrofloxacin-resistant E. coli from urine samples from dogs at the VMTH was not likely attributable to a single enrofloxacin-resistant clone but may be attributed to a collective increase in enrofloxacin resistance among uropathogenic E. coli in dogs in general.     

Colonisation of conventional weaned pigs by enteropathogenic Escherichia coli (EPEC) and its hazard potential for human health

Enteropathogenic Escherichia coli (EPEC) bacteria frequently cause severe enteric diseases primarily in children and in young rabbits. Their pathogenicity for pigs has been tested by oral infection of colostrum-deprived newborn, and of severely immunosuppressed weaned pigs, but colonisation of conventional weaned pigs by porcine EPEC has not been experimentally studied. EPEC show similarities to enterohaemorrhagic E. coli (EHEC) additionally carrying shiga toxin genes integrated into the chromosome by lambdoid phages. We have demonstrated earlier that the porcine EPEC prototype strain P86-1390 (O45) could be transduced in vivo (in ligated loops of weaned pigs), by Stx2 phage derived from a human EHEC. Thus, the ability of this porcine EPEC strain to colonise conventional weaned pigs under farming conditions became a question of relevance to human health. To clarify this question, four intragastric infection experiments were performed on a total of 95 conventional weaned pigs. The EPEC P86-1390 and other well-characterised porcine EPEC strains were applied to 54 pigs, leaving 41 weaned pigs as negative controls. In three experiments moderate predispositions were applied: coinfections with enterotoxigenic E. coli (ETEC) or with low-virulence TGE coronavirus, application of fumonisin B1 with a normal therapeutic dose of dexamethasone, and the increase of soybean protein concentration in the feed. A total of 41 weaned pigs served as negative controls inoculated with a commensal porcine E. coli. Housing conditions simulated the farm environment. As an overall result, ileal segments of 18.5% of infected pigs were shown to be colonised by EPEC, while no EPEC were detected in the ilea of controls. Among predisposing factors occurring on farms, feed protein content increased by 20% (26.3% crude protein, provided by 48% soybean meal) seemed to enhance EPEC colonisation and resulted in the mobilisation of spontaneous latent EPEC/ETEC infection. The results indicate that under normal farm conditions porcine EPEC may colonise conventional weaned pigs by inducing ileal attaching effacing (AE) lesions with reasonable frequency, without clinical signs. The results also suggest that conventional weaned pigs may represent undetected reservoirs of porcine EPEC, potentially giving rise to the emergence of new types of EHEC due to natural transduction by Stx phages.