Effects of Bordetella avium infection on the pulmonary clearance of Escherichia coli in turkeys

Thirty-six 1-day-old turkeys were inoculated intranasally with Bordetella avium (BA) strain 838. Noninoculated hatchmates (n = 36) were housed separately. At 2 and 4 weeks of age, 15 inoculated (BA+) and 15 noninoculated (BA-) turkeys were exposed to an aerosol of virulent Escherichia coli. The remaining six BA+ turkeys and six BA- turkeys were used as controls (ie, not exposed to E coli). Turkeys were necropsied on postaerosolization days 0 (immediately after aerosolization), 1, 3, 5, and 7. Lung and tracheal specimens were collected from each turkey for bacterial quantitation and histologic examination. A 1-ml blood sample was collected for detection of bacteremia. Numbers of E coli in lung specimens from 2- and 4-week-old turkeys were not significantly different between BA+ and BA- groups (pooled data over time); however, numbers of E coli isolated from tracheal specimens were significantly greater in BA+ turkeys than those in BA- turkeys. Although the incidence of pulmonary abcesses and E coli bacteremia was greater in 2-week-old turkeys than in 4-week-old turkeys, the incidence was not different between BA+ and BA- turkeys. At both ages, air sacculitis developed more often and was more severe in BA+ turkeys than in BA- turkeys. Hyperplastic bronchus-associated lymphoid tissue was found more often in BA+ turkeys than in BA- turkeys and appeared to be the first site of heterophil infiltration after E coli aerosolization.     

鸡致病性大肠杆菌菌毛分型的初步研究

以鸡致病性大肠杆菌F1菌毛特异性单抗1F4－1、2C3、3H3和F11菌毛特异性单抗FA1、FB11作为检测试剂，将111株已知O抗原的鸡致病性大肠杆菌经MD液体培养基连续传代培养后，通过直接玻板凝集法对各菌毛进行初步分型。结果发现F1、F11菌毛与O78、O1及O2三种优势O抗原型菌株之间存在较为明显的相关性，即致病性大肠杆菌主要集中在上。F1、F11菌毛在这三种O抗原型上的总检出率分别为95．6％、75．4％及73．3％。另外，在所检测的111株鸡致病性大肠杆菌中，只表达F1菌毛的大肠杆菌占菌杆总数的33．3％。只表达F11菌毛的大肠杆菌占菌株总数的8．1％，两者都表达的占菌株总数的36％，F1、F11菌毛的总检出率为78．3％。     

Ultrastructural studies of the lung of turkeys (Meleagris gallopavo) inoculated intratracheally with Escherichia coli.

To test the hypothesis that walls of air capillaries are a site for Escherichia coli to pass the air-blood barrier, fimbriated and nonfimbriated strains of E. coli were inoculated intratracheally into 18-day-old turkeys. Venous blood was cultured, and turkeys were necropsied from 0.5 to 8 hours post-inoculation. Lungs were processed for histopathology and electron microscopy. E. coli 078 was identified ultrastructurally using rabbit anti-lipopolysaccharide antibody and protein A-colloidal gold. All birds developed bacteremia; there was no significant difference between groups given fimbriated or nonfimbriated bacteria. Bacteria adhered to the plasma membrane of air capillary epithelial cells and were seen within vacuoles of portions of these cells that lined the fornices of air capillaries. Bacteria were also seen in the basement membrane at the basal surface of air capillary epithelial cells and, rarely, in vacuoles of subjacent endothelial cells. Infected granular and non-granular cells that lined air atria were necrotic 4 hours post-inoculation. Bacteria were within the overlying trilaminar substance and between reticular fibers of the interstitial stroma and pleura at 30 minutes post-infection and thereafter. Thus, the pulmonary air capillaries are a site for entrance of E. coli into the pulmonary blood capillaries, but fimbriae play little or no role in passage across the air-blood barrier.     

Influence of Bordetella avium infection on association of Escherichia coli with turkey trachea

Four-week-old Bordetella avium-infected and B avium-free turkeys were inoculated intratracheally with a suspension of fimbriated or nonfimbriated Escherichia coli. Numbers of E coli associated with tracheal sections were determined at postinoculation hour (PIH) 1 or 6. Significantly (P less than 0.05) greater numbers of E coli were isolated from the tracheas of B avium-infected turkeys compared with numbers in B avium-free turkeys. In B avium-free turkeys, tracheal associated E coli were 90% less at PIH 6 compared with that at PIH 1. However, in B avium-infected turkeys, numbers of E coli were not affected by postinoculation time. Seemingly, B avium-infected turkeys had reduced capacity to clear E coli from the trachea.     

Interaction of turkey complement with Escherichia coli isolated from turkeys

The role of turkey complement in a serum bactericidal reaction was determined using serum-sensitive and serum-resistant Escherichia coli isolated from turkeys. Inactivation of complement by heating serum (56 C for 40 minutes) or by treating serum with 10 mM EDTA eliminated bactericidal activity. Serum-sensitive E coli organisms were killed by turkey serum treated with 10 mM ethylene glycol-bis-beta-(aminoethyl ether)-N,N,N',N'-tetraacetic acid and 5 mM MgCl2. Exposure of normal turkey serum to serum-sensitive or serum-resistant E coli resulted in equivalent reductions in hemolytic activity of serum. Treatment of serum-resistant E coli with antibody rendered the bacteria sensitive to bactericidal effects of normal turkey serum. Serum-sensitive E coli organisms were readily killed by an alternative complement pathway, serum-sensitive and serum-resistant E coli activated the complement system equally well, and antibody was required for complement-mediated killing of certain serum-resistant E coli organisms from turkeys.     

Active oral immunization of suckling piglets to prevent colonization after weaning by enterotoxigenic Escherichia coli with fimbriae F18

Immunoprophylaxis of porcine oedema disease and post-weaning diarrhoea caused by strains of Escherichia coli expressing fimbriae F18 is an unsolved problem. The study was designed to examine whether vaccination with a live F18ac vaccine of unweaned pigs born to sows with F18ac antibody in the colostrum requires preformed fimbriae in the vaccine, and whether protection against the heterologous fimbrial variant F18ab is induced as well. Genetically susceptible pigs were vaccinated orally on three consecutive days, beginning 10 days before weaning with 10(11) CFU of an F18ac culture. Challenge with a dose of 10(7) CFU of E. coli F18 on three consecutive days was initiated 9 or 11 days after weaning. Eighteen pigs given the fimbriated F18ac vaccine and challenged with a strain of the homologous fimbrial variant were protected against colonization; mean faecal viable counts of the challenge strain were >3 log10 lower than those from the 17 non-vaccinated control pigs. The vaccinated pigs developed a significant rise of F18ac IgA serum antibodies. The 23 pigs which had received the non-fimbriated vaccine showed no significant protection and exhibited much lower serum F18ac IgA ELISA reactivities. Eighteen pigs vaccinated with the fimbriated F18ac and challenged with an F18ab strain had faecal viable counts nearly as high as those from 16 non-vaccinated control pigs. It is concluded that only oral vaccines having preformed fimbriae induce protection limited to the homologous fimbrial variant.     

Development of a sandwich ELISA and comparison with PCR for the detection of F11 and F165 fimbriated Escherichia coli isolates from septicaemic disease in farm animals

The P fimbriae F11 and F165 that have been demonstrated on Escherichia coli septicaemic strains in poultry and calves, respectively, possess a nearly identical major subunit that demonstrates a serological cross-reaction. A polyclonal antibody-based sandwich ELISA (sELISA) that was specific for both F11 and F165 fimbriated strains was compared with a PCR method to detect F11/F165 fimbriated strains, in a collection of E. coli strains isolated from diseased animals. Of 298 isolates tested, 36 were positive by PCR of which only 14 were sELISA positive. There were no sELISA positive but PCR negative results. The 36 PCR positive isolates comprised 11 avian strains of which 10 were sELISA positive, 20 bovine strains of which 4 were sELISA positive and 3 ovine strains, 1 porcine strain and 1 equine strain all of which were sELISA negative. The F11/F165 incidence of 10.7% in 103 poultry and 18.3% in 109 bovine isolates demonstrates a moderate level of these factors in E. coli septicaemic cases in Northern Ireland.     

Polyclonal sandwich ELISAs to detect F18 fimbriated Escherichia coli in pigs in Northern Ireland.

F18 fimbriated Escherichia coli are a newly described cause of postweaning diarrhea in pigs. Polyclonal rabbit antisera were raised to the antigenic variants, F18ab and F18ac, of these fimbriae and were used to develop monospecific sandwich enzyme-linked immunosorbent assays (ELISAs). The ELISAs were standardized with type cultures characterized by polymerase chain reaction techniques (PCR) and then used to conduct a study of the prevalence of F18 fimbriated E. coli in pigs in Northern Ireland. A total of 176 isolates were tested by ELISA and PCR. Eight isolates were positive for F18 by ELISA, of which 2 were shown to be false positives by PCR and one was PCR positive but ELISA negative. Of the 6 confirmed ELISA positives, all produced VT2 toxin and 3 produced ST toxin. Four positives were from serogroups O138 and O139, previously associated with porcine diarrhea.     

Attachment of different Escherichia coli strains to cultured rumen epithelial cells

The attachment to fully characterized primary rumen epithelial cell cultures of Escherichia coli strains isolated from different animal species and expressing F1–F4 or F17 fimbriae was examined. As the cell cultures contained stratified (keratinized) and non-stratified (non-keratinized) cells which grew either confluently or non-confluently, the strength of attachment of the different bacterial strains was assessed in relation to the differentiation state of the cells. Thus, strains having F1 fimbriae attached to all types of cultured cells, while strains with F2 and F3 fimbriae did not bind at all. E. coli strains having F4 or F17 fimbrae attached only to non-keratinized cells, particularly to confluent areas. As membrane glycosylation is known to change with differentiation (keratinization), our results suggest that the attachment of fimbriated E. coli strains which were capable of binding to rumen cells was more likely to be dependent on differentiation than the host specificity of the bacteria.     

Fimbriae in Escherichia coli isolated from the small intestine of piglets

Ninety E. coli strains, isolated from piglets which had died from neonatal diarrhea, were tested for the presence of K88, K99, 987P and type 1 fimbriae. Two or more types of fimbriae were demonstrated in 14 of the strains, a single fimbria! type in 44 strains while in 32 strains no fimbriae were detected. Of the 14 E. coli strains with more than 1 type of fimbriae, 12;, 10, 8 and 4 strains showed K88, K99, 987P and type 1, respectively.Twelve E. coli strains were isolated from piglets which had died in the neonatal period without showing signs of neonatal diarrhea at necropsy. One strain showed 987P and 3 strains showed type 1 fimbriae, while the remaining 8 strains were unfimbriated.Sixteen fimbriated E. coli strains were subcultured in order to examine the stability of fimbrial expression in the strains. The K88 and the type 1 fimbriae were regularly expressed, while the K99 and 987P were inconsistently demonstrated.     

Monoclonal antibodies reveal a weak interaction between the F18 fimbrial adhesin FedF and the major subunit FedA

F18+ Escherichia coli can cause post-weaning diarrhoea and oedema disease in pigs. These diseases are responsible for substantial economic losses, but a vaccine is not available. A good knowledge of the characteristic of the fimbriae is useful for the development of a vaccine composed of the fimbrial virulence factor. F18 fimbriae are composed of the major subunit FedA and the minor subunits FedE and the adhesin FedF. In the present study monoclonal antibodies (mAbs) against FedA and FedF were produced. In addition to their diagnostic value, these mAbs revealed a weaker interaction between FedA and FedF compared to the subunit-subunit interactions in other fimbriae, like type 1 and P pili. Further experiments are needed to investigate if this weak interaction could be one of the reasons for the slow colonisation of the small intestinal mucosa by F18+ E. coli.     

In vitro adhesion of an avian pathogenic Escherichia coli O78 strain to surfaces of the chicken intestinal tract and to ileal mucus

The role of fimbria in adherence of an avian pathogenic Escherichia coli (APEC) O78 strain 789 to chicken intestine was studied. Bacterial adhesion to tissue sections representing the regions within the chicken intestinal tract was determined by using immunohistochemical methods. E. coli 789 grown to express the type 1 fimbria adhered efficiently to the crop epithelium, to the lamina propria of intestinal villi, and to the apical surfaces of both the mature as well as the crypt-located enterocytes in intestinal villi, whereas no adhesion to mucus-producing goblet cells was detected. The adhesion was inhibited by mannoside and the role of type 1 fimbriae in the observed adhesion was confirmed with a recombinant strain expressing type 1 fimbriae genes cloned from E. coli and Salmonella enterica. E. coli 789 strain grown to favor AC/I fimbriae expression as well as the recombinant E. coli strain expressing the fac genes adhered to goblet cells but only poorly to the other epithelial sites. E. coli strain 789 as well as S. enterica serovar Typhimurium IR715 and S. enterica serovar Enteriditis TN2 strains were able to multiply in ileal mucus medium. The type 1 fimbria expressing bacteria adhered to the ileal mucus, whereas the AC/I fimbriated strains showed poor adherence to the mucus. The adhesion of E. coli 789 onto the crop epithelium and the follicle associated epithelium of the chicken ileum was efficiently inhibited by an adhesive strain ST1 of Lactobacillus crispatus isolated from chicken, whereas poor inhibition of E. coli adherence was observed with the weakly adhesive L. crispatus strain 134mi. The type 1 fimbriae may be important in colonization of the chicken intestine by APEC and Salmonella.     

Efficacy of avian pneumovirus vaccines against an avian pneumovirus/Escherichia coli O2:K1 dual infection in turkeys

The clinical, pathological and microbiological outcome of a challenge with avian pneumovirus (APV) and Escherichia coli O2:K1 was evaluated in turkeys vaccinated with an attenuated APV vaccine and with or without maternally derived antibodies. Two groups of two-week-old poults, one with and one without maternally derived antibodies against APV, were vaccinated oculonasally with attenuated APV subtype A or B. A third group remained unvaccinated. Eleven weeks later, the turkeys were inoculated intranasally with either virulent APV subtype A, or E. coli O2:K1, or with both agents three days apart. After the dual infection, birds vaccinated with attenuated subtype A or B, and with or without maternally derived antibodies, had lower mean clinical scores than the unvaccinated birds. In the vaccinated birds, virus replication was significantly reduced and no bacteria were isolated, except from the birds vaccinated with attenuated subtype B. In the unvaccinated turkeys, large numbers of E. coli O2:K1 were isolated from the turbinates of the dually infected birds between one-and-a-half and seven days after they were inoculated.     

Avian pathogenic Escherichia coli (APEC).

Avian pathogenic Escherichia coli (APEC) cause aerosacculitis, polyserositis, septicemia and other mainly extraintestinal diseases in chickens, turkeys and other avian species. APEC are found in the intestinal microflora of healthy birds and most of the diseases associated with them are secondary to environmental and host predisposing factors. APEC isolates commonly belong to certain serogroups, O1, O2 and O78, and to a restricted number of clones. Several experimental models have been developed, permitting a more reliable evaluation of the pathogenicity of E. coli for chickens and turkeys. Hence, virulence factors identified on APEC are adhesins such as the F1 and P fimbriae, and curli, the aerobactin iron sequestering system, K1 capsule, temperature-sensitive hemagglutinin (Tsh), resistance to the bactericidal effects of serum and cytotoxic effects. Experimental infection studies have shown that the air-exchange regions of the lung and the airsacs are important sites of entry of E. coli into the bloodstream of birds during the initial stages of infection and that resistance to phagocytosis may be an important mechanism in the development of the disease. They have also demonstrated that F1 fimbriae are expressed in the respiratory tract, whereas P fimbriae are expressed in the internal organs of infected chickens. The role of these fimbrial adhesins in the development of disease is not yet, however, fully understood. The more recent use of genetic approaches for the identification of new virulence factors will greatly enhance our knowledge of APEC pathogenic mechanisms. Diagnosis of APEC infections is based on the clinical picture, lesions and isolation of E. coli. This may be strengthened by serotyping and identification of virulence factors using immunological or molecular methods such as DNA probes and PCR. Approaches for the prevention and control of APEC infections include the control of environmental contamination and environmental parameters such as humidity and ventilation. Antibiotherapy is widely used, although APEC are frequently resistant to a wide range of antibiotics. Vaccines containing killed or attenuated virulent bacteria protect against infection with the homologous strain but are less efficient against heterologous strains. Hence, vaccination for colibacillosis is not widely practised because of the large variety of serogroups involved in field outbreaks.     

Efficacy of enrofloxacin, florfenicol and amoxicillin against Ornithobacterium rhinotracheale and Escherichia coli O2:K1 dual infection in turkeys following APV priming

Experimental groups of 15 susceptible 3-week-old turkeys were inoculated oculonasally with avian metapneumovirus (APV) subtype A and susceptible Escherichia coli O2:K1 and Ornithobacterium rhinotracheale (ORT) bacteria, with a 3 days interval between viral and bacterial inoculation and approximately 8h between the two bacterial inoculations. The aims of the present study were to assess the efficacy of drinking-water administration of enrofloxacin for 3 and 5 days, amoxicillin for 5 days and florfenicol for 5 days for the treatment of the resulting respiratory disease, based on clinical and bacteriological examinations. Antimicrobial treatment started 1 day after dual bacterial inoculation. After infection, the birds were examined and scored for clinical signs daily, weighed at different times, and their tracheae swabbed daily. Five birds were euthanised and examined for macroscopic lesions at necropsy at 5 days post-bacterial inoculation (dpbi) and the remainder at 15dpbi. Samples of the turbinates, trachea, lungs, sinuses, air sacs, heart, pericardium and liver were collected for bacteriological examination. Recovery from respiratory disease caused by an APV/E. coli/ORT triple infection in 3-week-old turkey poults was overall most successful after enrofloxacin treatment, irrespective of treatment duration, followed by florfenicol treatment. Compared with the untreated group, clinical signs as well as ORT and E. coli multiplication in the respiratory tract were significantly reduced by both enrofloxacin treatments and the florfenicol treatment, with the enrofloxacin treatments showing significantly better reductions than the florfenicol treatment. Five-day treatment with amoxicillin, compared with the untreated group, did not cause a significant reduction in any of the aforementioned parameters.     

Enteric-coated pellets of F4 fimbriae for oral vaccination of suckling piglets against enterotoxigenic Escherichia coli infections

To prevent enterotoxigenic Escherichia coli (ETEC) induced postweaning diarrhoea, the piglet needs an active mucosal immunity at the moment of weaning. In the present study, the feasibility of oral vaccination of suckling piglets against F4+ETEC infection with F4 fimbriae was studied. Furthermore, oral vaccination with enteric-coated pellets of F4 fimbriae was compared to vaccination with F4 fimbriae in solution. Therefore, piglets were orally administered 1mg F4 fimbriae in pellets or in solution during three successive days at the age of 7 and 21 days, whereas control piglets were not vaccinated. Five days postweaning (33 days of age), all animals were orally challenged with F4+ETEC. Despite the induction of an immune response upon oral administration of both F4 fimbriae in pellets as in solution, the colonisation of the small intestine by F4+ETEC upon oral challenge could not be prevented. However, a marginal but significant reduction in F4+ E. coli faecal excretion was found in the piglets vaccinated with F4 fimbriae in pellets, indicating that the use of an enteric-coat which protects the F4 fimbriae against inactivation by milk factors and degradation by enzymes and bile improves vaccination.     

Pasteurella anatipestifer infections in California turkey flocks: circumstantial evidence of a mosquito vector

Outbreaks of Pasteurella anatipestifer infections in California turkey flocks were investigated and found to have a seasonal distribution, with a peak incidence in fall, coinciding with peak Culex mosquito populations. An experiment was conducted to test the hypothesis that mosquitoes may serve as vectors for P. anatipestifer infections in turkeys. Four 7-week-old turkey poults were exposed for 7 days to mosquitoes captured from turkey barns during a field outbreak of P. anatipestifer serotype 1 infection. One turkey developed serum antibodies to serotype 1, detectable by enzyme-linked immunosorbant assay, and was resistant to an intravenous inoculation of P. anatipestifer serotype 1 at 4 weeks postexposure. Giemsa-stained blood smears from this bird and from three 7-week-old turkeys inoculated intravenously with P. anatipestifer revealed the presence of rod-shaped bacteria in or on the surface of host erythrocytes. No such rod-shaped bodies were found on erythrocytes of an uninoculated control turkey.     

Quantitative resistance level (MIC) of bacterial pathogens (Escherichia coli, Pasteurella multocida, Pseudomonas aeruginosa, Salmonella sp., Staphylococcus aureus) isolated from chickens and turkeys: national resistance monitoring by the BVL 2004/2005

In the study 2004/2005, the current quantitative resistance level of Escherichia coli, Pasteurella multocida, Pseudomonas aeruginosa, Salmonella sp. and Staphylococcus aureus from chickens and turkeys was determined for the first time within the framework of the National Resistance Monitoring of the Federal Office of Consumer Protection and Food Safety (BVL).The objective was to implement a valid database on the basis of which the development and spread of resistance can be evaluated and monitored. During the investigation period from January 2004 to February 2005,927 strains were collected and 857 (92%) bacteria strains which corresponded to the specifications of the study protocol were tested with the broth microdilution method to determine the in vitro susceptibility (minimum inhibitory concentration) to 22 to 28 antimicrobial agents or antibiotic combinations. The results document a prevalence of resistance that exceeds that of bacterial pathogens of other animal species, especially in the case of tetracycline. Apart for S. aureus, clinical resistance to fluoroquinolones can still be considered low in poultry pathogens (E. coli approx. 2%). By applying the MICG of 4 mg/L for enrofloxacin, a susceptibility of approximately 78 % was calculated for S. aureus. A comparison of the prevalence of resistance between chickens and turkeys, showed that a slightly higher prevalence of resistance can be expected in turkeys. Differences between the susceptibility data of chicks and adult animals could only be found in turkeys. In the case of E. coli, the prevalence of resistance of strains isolated from adult turkeys was up to 10% higher than those isolated from chicks for the corresponding antimicrobial agents. It must be pointed out that the number of E. coli strains from adult turkeys was much higher (n = 194) than the number from turkey chicks (n = 21). The results indicate clearly that in a resistance monitoring system it is necessary to categorise poultry by animal species (chicken, turkey) as well as by production stage and type (broiler, laying hen), so that the epidemiology of resistance can be correctly represented and evaluated. This information is the basis for the development of long-term management options for minimizing antibiotic resistance. Furthermore, the knowledge of prevailing resistance levels in Germany is a valuable tool for veterinary practitioners when determining an empirical therapy.The data collected by the BVL make an important contribution to the optimisation of the safety of food from animals, and thus to improving consumer protection.     

Inactivated Salmonella expressing the receptor-binding domain of bacterial adhesins elicit antibodies inhibiting hemagglutination

We examined the potential of inactivated Salmonella strains to induce protective antibodies against two adhesins of pathogenic Escherichia coli. The receptor-binding domains of the F17a-G adhesin of F17a fimbriae and of the FimH adhesin of type 1 fimbriae were fused to the translocator domain of the autotransporter AIDA-I. An IgG response against F17a-G or FimH was induced after immunization of mice with acetone-inactivated Salmonella displaying the corresponding fimbrial receptor-binding domain. These sera inhibit in vitro agglutination of erythrocytes by E. coli carrying these fimbriae. Our results demonstrate that induced and subsequently acetone-inactivated Salmonella are useful delivery vehicles for the stimulation of an IgG antibody response against heterologous antigens.     

Role of fimbriae F18 for actively acquired immunity against porcine enterotoxigenic Escherichia coli

Enterotoxigenic (ETEC) and enterotoxaemic (ETEEC) Escherichia (E.) coli that express F18 (F107) fimbriae colonize the small intestine and cause diarrhoea and/or oedema disease in weaned pigs. So far, two antigenic variants of F18 can be distinguished with a common antigenic factor designated ‘a’ and two specific factors called ‘b’ and ‘c’. In this study the existence of crosswise anti-colonization immunity between E. coli strains that express F18ab or F18ac fimbrial variants, respectively, was demonstrated. Weaned pigs of susceptible genotype with respect to susceptibility to adhesion of E. coli with fimbriae F18 were inoculated with E. coli strains 3064^STM (O157:K-:H-:F18ab; resistant to streptomycin) and 8199^RIF (O141ab:K-:H4:F18ac; resistant to rifampicin). The faecal shedding was compared subsequent to immunization and homologous or heterologous challenge. An enzyme-linked immunosorbent assay (ELISA) was applied to measure IgA, IgM and IgG antibodies against the F18ab and F18ac antigens in saliva, faeces, serum and intestinal wash samples. About 8 log CFU/g of the inoculated strains were found in faeces of all pigs following immunization as well as in non-immunized controls after challenge. Bacterial counts of the inoculated strains after challenge were between 2 and 5 log lower, without any difference between homologous and heterologous challenge. Intestinal colonization with fimbriated E. coli resulted in production of significantly increased levels of anti-fimbrial antibodies, especially IgA, in serum and intestinal wash samples. There were higher levels of homologous than of heterologous anti-fimbrial antibodies. Production of antibodies against F18a or against another common fimbrial antigen is probably responsible for crosswise anti-colonization immunity between E. coli strains with F18ab and F18ac fimbrial variants. Serum F18-specific IgA may be a useful indicator of a mucosal immune response directed against F18 fimbriae.