Construction and analysis of a Vibrio cholerae delta-aminolevulinic acid auxotroph which confers protective immunity in a rabbit model.

Vibrio cholerae CVD101 is a very effective live vaccine. Although this strain does not produce active cholera toxin because of a mutation in the gene for the cholera toxin A subunit, it still shows residual pathogenicity. To attenuate CVD101 further, we set out to isolate derivatives of CVD101 which were limited in their ability to proliferate in vivo. Two delta-aminolevulinic acid auxotrophs of CVD101, designated V286 and V287, were isolated by transposon mutagenesis and penicillin enrichment. Southern blotting revealed that the mutants differed with respect to the location of the transposon insertion. Under aerobic conditions, in the absence of delta-aminolevulinic acid, both mutants showed diminished growth compared with CVD101. The growth of V286 was most severely affected. Microaerophilic growth of both mutants was less affected. Competition experiments with a rabbit model showed that strain V286 was found in numbers 10(3)- to 10(4)-fold lower than its parental strain. This observation indicates that strain V286 is impaired in its ability to colonize the rabbit intestine. It also supports an important role for aerobic growth in the colonization of the intestine by V. cholerae. Vaccination of rabbits with a single dose of strain V286 resulted in full protection against challenge with a virulent strain. Strain V286 was not shed from rabbits in a cultivatable form. Our results suggest that delta-aminolevulinic acid auxotrophy can attenuate V. cholerae by limiting its ability to colonize without affecting its capacity to induce protective immunity. Furthermore, this type of mutation may prevent the spread of V. cholerae vaccine strains in the environment.

Construction and analysis of Bordetella pertussis mutants defective in the production of fimbriae.

Although the role of fimbriae in bacterial disease has been well established, little is known about the function of Bordetella pertussis fimbriae. To study this function, well-defined fimbrial mutants were constructed. B. pertussis harbours three fimbrial genes, fim2, fim3 and fimX, and strains were constructed in which one or more fimbrial genes were inactivated by means of gene replacement. Analysis of these strains by means of immunoblotting suggested the presence of a fourth fimbrial gene, tentatively designated fimY. A fimbrial mutant was analysed in a mouse respiratory infection model, together with a strain harbouring a deletion in the gene for the filamentous haemagglutinin. Both mutants were affected in their ability to persist in the trachea. Persistence in the nasopharynx was only affected by the mutation in the filamentous haemagglutinin gene. Neither the filamentous haemagglutinin nor the fimbrial mutants were affected in their ability to persist in the lung. Our results suggest that the filamentous haemagglutinin plays a more crucial role than fimbriae in the colonization of the upper respiratory tract of the mouse.

[Aeromonas species as cause of diarrhea and infections outside the gastrointestinal tract in The Netherlands]. / Aeromonas-species als verwekker van diarree en infecties buiten het maag-darmkanaal, in Nederland.

In a national survey in the period May 1986-December 1987, Aeromonas was isolated from 277 out of 16,857 (1.6%) samples of watery, bloody or mucous stool, from patients with diarrhoea. There was a clear seasonal pattern (less than 1% in winter, up to 3% in summer). A. caviae was isolated most frequently (49%), followed by A. sobria (35%) and A. hydrophila (15%). Some non-identifiable strains were isolated as well. Aeromonas were isolated in particular from faeces of patients aged over 70 years (predominantly A. sobria) or under 5 years (predominantly A. caviae). In 67% Aeromonas was isolated as the only possible bacterial cause of diarrhoea, but in 33% other enteropathogenic bacteria were found as well (17% Campylobacter, 14% Salmonella, 2% Shigella). In addition, all Aeromonas isolates were collected which were obtained in normal diagnostic activities in the participating laboratories, among others from blood, from pus or wound fluid, from faecal samples which did not meet the above mentioned criteria. A. caviae was the dominant species in 'other' faeces and 'various' body sites but was not isolated from blood. The results of this study do not indicate that routine examinations for Aeromonas in faeces of patients with diarrhoea are necessary.

Prevalence of enterotoxigenic Escherichia coli strains in outbreaks and sporadic cases of diarrhoea in Spain.

Escherichia coli strains isolated 1985-1988 in Spain from patients with diarrhoea were examined; 1170 strains were isolated from 582 sporadic cases of diarrhoea in children, and seven strains were associated with seven outbreaks of diarrhoea. Strains positive for STa enterotoxin production in the infant mouse test were also assayed for production of LT enterotoxin on Vero cells and by a coagglutination test. Thirty-one strains were STa positive: 28 were isolated from 16 (2.7%) sporadic cases of diarrhoea and three were responsible for outbreaks. The majority of STa+LT- strains from both outbreaks and sporadic cases were serotype O153:H45 and expressed the CFA/I colonization factor antigen. Enterotoxigenic STa+LT- strains of serotype O27:H7 and STa+LT+ CFA/II+ strains of serotype O6:K15:H16 were also isolated frequently from sporadic cases.

Infection and immunity to Vibrio cholerae, Salmonella typhimurium and Escherichia coli in a rabbit model.

Cholera disease can be induced in the rabbit by duodenal inoculation (DI) of Vibrio cholerae organisms after ligation of the cecum (C) (DIC model). When ligation of the cecum is omitted, no disease symptoms develop. In contrast, the animals are primed which becomes apparent as vibriocidal protection upon challenge with V. cholerae in the DIC model. This protection coincides with high anti-O antigen IgA levels in the bile. The O antigen was shown to be the protective antigen and it must be presented by live organisms. A non-enterotoxigenic mutant of V. cholerae induced protective immunity in the rabbit but was reported to cause mild diarrhea in human volunteers. Looking for alternatives, we applied cholera toxin, known as a mucosal adjuvant, together with killed V. cholerae cells to rabbits. Unfortunately, the minimum adjuvant dose was equal to the minimum toxic dose. A Salmonella typhimurium strain expressing also the V. cholerae O antigen induced systemic rather than local immunity which was not protective. Several Escherichia coli strains were able to elicit a local immune response, but the animal to animal differences were considerable. Therefore, V. cholerae itself was thought to be the most appropriate carrier organism. Some non-enterotoxigenic and auxotrophic mutants of V. cholerae were able to prime and did not show any undesired side-effects in the DIC model. Therefore, further attenuation of non-toxigenic V. cholerae strains by means of stable deletions in nutritional genes seems to be the most promising way to obtain acceptable vaccine candidates.

Value of the Widal test in the diagnosis of typhoid fever in an endemic area and suggestions for a modification. A preliminary study.

The Widal test for detection of typhoid fever was introduced in two rural hospitals in Nigeria. Because of undetectably low O agglutination titers, 74 sera were sent overseas for further determination by means of passive haemagglutination with O 9, 12-antigen, Vi-antigen and slide agglutination with H-d antigen suspensions. This full test showed discrepancies with the normal Widal test i.e. a false negative Widal H agglutination in 6 out of 17 patients in West Nigeria and a false positive Widal H test in 28 out of 49 patients in Central Nigeria. The full test should be applicable in a small laboratory, but its value has yet to be established.

Priming and boosting of the rabbit intestinal immune system with live and killed, smooth and rough Vibrio cholerae cells.

Cholera disease can be induced in the rabbit by ligation of the cecum (C) followed by duodenal inoculation (DI) of virulent Vibrio cholerae organisms (DIC model). When the cecum is not ligated, DI does not induce disease. In contrast, the animals are primed which becomes apparent upon challenge with live V. cholerae in the DIC model. Such animals are vibriocidally protected. This protection is characterized by absence of disease symptoms, rapid disappearance of V. cholerae from the feces and presence of high levels of anti-lipopolysaccharide Immunoglobulin A in the bile. The present study shows that primed rabbits can also be boosted by duodenal administration of killed, smooth V. cholerae cells. On the other hand, killed cells cannot prime. The minimal lethal dose of a rough derivative of a smooth strain C5, designated R5 and lacking the O antigen part of the LPS, was 100,000 times higher than that of its parent strain C5, in the DIC model. Rabbits which had been duodenally immunized with strain R5 and were subsequently challenged with the smooth strain C5, all developed diarrhea and two out of eight died. This result supports an earlier observation that the specific O antigen part of the V. cholerae LPS is an essential prerequisite for the induction of protective immunity in the rabbit.

Immunoglobulins in bile and serum of the rabbit associated with protection after Vibrio cholerae infection and vaccination.

Previous studies have shown that cholera, as well as protective immunity against infection with Vibrio cholerae, can be induced in the rabbit. This protection is long-lasting (up to 30 months) and is characterized on challenge by rapid, symptom-free disappearance of V. cholerae from the intestine; we therefore believe this to be vibriocidal protection. In this study, we analysed the humoral and secretory immune response against various subcellular V. cholerae components in vibriocidally protected, non-vibriocidally protected, and unprotected animals. Only vibriocidal protection was found to be associated with high levels of biliary IgA directed against lipopolysaccharide O antigen. We did not find such a correlation between either type of protection and response in serum. Therefore, anti-lipopolysaccharide antibodies are essential in protection against experimental infection with V. cholerae.

Protective immunity against Vibrio cholerae infection in the rabbit.

The DIC model (Duodenal Inoculation with ligation of the Cecum in rabbits) was employed to study experimentally induced cholera and the related protective immunity. Duodenal inoculation (DI) without ligation of the cecum with live V. cholerae organisms did not cause any disease symptom but induced protection against subsequent challenges with homologous and heterologous organisms for up to 24 months. After 30 months this protective immunity began to decrease. A similar protective immunity could be induced by administration of the A- B+ derivative CVD101 of V. cholerae strain 395. This type of experiment can only be done successfully with conventional, healthy rabbits held under low stress conditions. A so-called specific pathogen-free rabbit breed was found to be entirely unsuitable. Duodenal inoculation with heat- or merthiolate-inactivated V. cholerae for a prolonged period of time by means of an intestinal osmotic minipump did not induce protection. Injection of heat-inactivated V. cholerae material into the Peyer's patches sometimes led to protection, suggesting that a thermostable antigen, possibly lipopolysaccharide, is one of the major protective antigens. Duodenal administration of a combination of inactivated V. cholerae serotypes Ogawa and Inaba cells and 1 mg B subunit of the V. cholerae enterotoxin by up to three inoculations protected only 3 out of 12 rabbits against challenge. The results obtained on the rabbit model are discussed in relation to the efficacy of this vaccine in human volunteers and in a recent field test.

In vivo adherence and colonization of Vibrio cholerae strains that differ in hemagglutinating activity and motility.

A scanning electron microscopic study was carried out to compare the in vivo pathogenicity of two strains of Vibrio cholerae in an adult rabbit ligated-gut test model. V. cholerae C5 (serotype Ogawa, biotype El Tor), a motile strain possessing hemagglutinating activity in vitro, and C21 (serotype Ogawa, classical biotype), a nonmotile strain possessing no hemagglutinating activity, were tested. Tissue samples from small intestinal loops were examined 3, 6, 9, and 12 h postinoculation. Contradictory to most published data, neither hemagglutinating activity nor motility appeared to be essential prerequisites for the pathogenesis of cholera in the experimental animal model used: nonmotile hemagglutinin-negative strain C21 adhered to and colonized the small intestine at least to the same extent as did motile hemagglutinin-positive strain C5. Maximum colonization was seen at 9 h postinoculation for both strains. C5 and C21 vibrios caused comparable damage to the villi of the small intestine. The villous epithelium showed only mild changes during the first 9 h postinoculation. However, after 12 h the epithelium was seriously damaged concomitant with a decrease in the number of vibrios. Many villi showed partial or total denudation, owing to repelled epithelium, leaving a bare basal lamina with only some to moderate numbers of vibrios attached. Since similar changes were induced by pure cholera enterotoxin, these changes were likely the result of excessive fluid accumulation. From this study it is concluded that, at least in the animal model used, factors other than hemagglutinating activity and motility may also play a role in the association of V. cholerae with the small intestinal surface.

Serotyping of Aeromonas species using passive haemagglutination.

Antisera were raised with heat-killed vaccines prepared from 25 Aeromonas hydrophila strains, 4 A. sobria strains and one A. caviae strain. Twenty-seven of these 30 antisera gave high titers when tested in the microtiter tray agglutination (MTA) test with their homologous antigen heated at 100 degrees C for 30 min. Three O antisera gave low titers in the MTA test but reacted to high titers in the haemagglutination (HA) test with conserved sheep red blood cells coated with alkali-treated heat extract. All 27 antisera showing high titers in the MTA test, showed even higher titers in the HA test. Therefore, the HA technique was employed to type 306 strains isolated from surface and drinking water, from food samples and from faeces of human patients with diarrhea. Of 155 A. hydrophila strains 21 (14%) could not be typed. For A. sobria and A. caviae, the percentage of untypable strains was 46% and 68% respectively. Many A. sobria and A. caviae strains reacted to titer in A. hydrophila O antisera. A limited number of strains reacted to titer in more than one specific O antiserum. Immunoelectrophoretic studies indicated that the O antigen is often represented by more than one precipitation line and that true K antigens occur in Aeromonas.

Serotyping of Serratia marcescens using passive haemagglutination.

Serratia marcescens O antisera prepared with boiled bacterial suspensions gave relatively low titers when tested with boiled broth cultures in tube or tray agglutination tests, but extremely high titers when tested in the haemagglutination test with conserved sheep erythrocytes coated with alkali-treated heat extract. However, an antiserum prepared with boiled O group 6 bacteria, did not react at all. A proper O6 antiserum could only be prepared if the O6 cells were not heated at 100 degrees C. The antigenic relationships between standard strains of S. marcescens were investigated by means of haemagglutination tests. The antigenic relationships already described in the literature were confirmed. Additional reciprocal relationships were observed, some of which could be explained by 8 antigenic factors which are designated I-VIII. O antisera could be made monospecific by absorption. It was confirmed that standard strain O6H3 has two antigenic determinants which should be written as O6ab, that standard strain O6H8 has O6bc and that O14 has O6b. Standard strain O7 was found to possess O6a in addition to the specific determinant O7, whereas the antigen O6c turned out to be identical with O8. The standard strain O14 has, in addition to O6b only the factor hitherto described as Co12, 13, 14 and no specific O14 antigen. We propose to redesignate Co12, 13, 14 as O14 and to write the formula of O12 and O13 strains possessing this antigen as O12, 14 and O13, 14 respectively. O antigen O23 should then be written as O14, 23. For similar reasons we propose to redesignate Co2, 3 as O25. The O antigen O20 was found to be related to the antigens O16ab, O16ac and O16acd in such a way that it should be written O16ae. For similar reasons O22 should be written as O10ac. When serotyping Dutch isolates of Serratia marcescens, hitherto undescribed antigenic combinations such as O2, 16ab; 03, 6a; O3, 21 and O4, 6b were encountered. For H antigen typing we found the slide agglutination with cultures grown on semisolid medium the most appropriate.

Monoclonal antibodies against the K99 antigen of Escherichia coli for diagnostic purposes.

Hybridomas secreting monoclonal antibodies against the K99 antigen of Escherichia coli were produced by the fusion of spleen cells from immunized BALB/c mice with P3/X63-Ag8.653 myeloma cells. The seven hybridomas which produced the highest antibody titers in vitro, as detected by enzyme-linked immunosorbent assay (ELISA) and Perma slide agglutination test (PSAT), were chosen for antibody production in vivo. No cross reaction was observed with K88ab, F41 and P987 antigens in the ELISA. The titer of each ascitic fluid was established by the ELISA and the slide agglutination (SAT) tests. The two ascitic fluids with the highest titer in the SAT were incorporated into the set of antisera used for serotyping at our laboratory. The results were satisfactory both in terms of stability and specificity.

[Resistance to antibiotics in Salmonella]. / Resistentie tegen antibiotica bij Salmonella.

Approximately 20,000 strains of Salmonella were screened annually for resistance to tetracycline, chloramphenicol, kanamycin and ampicillin since 1959, and also to trimethoprim since 1978. Tetracycline-resistant strains increased in human subjects and pigs from 1961. After the ban on incorporation of tetracycline in animal feeds for nutritive purposes in 1974, the proportion of tetracycline-resistant strains in pigs and human subjects decreased. In veal calves, the number of strains of S. typhimurium and S. dublin resistant to multiple drugs increased from 1972. Strains resistant to multiple antibiotics in man were mainly isolated from adoptive children from Indonesia. No further spread of these strains was observed. So far, strains similar to those in calves resistant to multiple drugs were only incidentally isolated from human patients.

Effect of iron on serotypes and haemagglutination patterns of Escherichia coli in bottle-fed infants.

Strains of Escherichia coli isolated from faecal specimens of ten infants receiving breast milk, six receiving a cow-milk preparation with iron supplement (5 mg/l) and six the preparation without iron supplement (less than 0.5 mg/l), were serotyped and examined for their haemagglutinating activity. The Escherichia coli flora of breast-fed and bottle-fed infants consisted of one resident strain, accompanied by one or more transient strains. Changes in the serotype of the Escherichia coli flora and in the frequency of occurrence of strains associated with urinary tract infections were more often seen in bottle-fed than in breast-fed infants. In breast-fed and bottle-fed infants without iron supplement most strains of Escherichia coli were non-haemagglutinating, while most strains in infants bottle-fed with iron supplement showed mannose-resistant haemagglutination. It is concluded that human milk favours the establishment of a stable non-pathogenic Escherichia coli flora and that a low iron content in standard cow-milk preparation favours colonization with non-adherent strains of Escherichia coli.

Vibrio cholerae infection and acquired immunity in an adult rabbit model.

We modified the rabbit model for enteric infection by Vibrio cholerae developed by Spira et al. and designated the RITARD (for removable intestinal tie-adult rabbit diarrhea) model (20). Our modification DISC comprises a permanent ligation of the cecum (C) to prevent resorption of the fluid secreted by the small intestine, a temporary ligation of the small intestine (S) to enable the bacteria to colonize, and duodenal inoculation (DI) of the challenge material. The main difference between RITARD and DISC is that in the latter model the challenge material is injected into the duodenum approximately 10 cm distal to the stomach instead of into the jejunum. Four out of 5 V. cholerae strains tested, including 2 serotypes and 2 biotypes, were able to elicit a massive and usually fatal cholera-like diarrhea. The virulence depended strongly on the culturing conditions. One strain, C5, caused fatal diarrhea in a dose of about 1000 organisms, even if the temporary ligation was omitted (DIC model). Other modifications were the DIS and the DI model in which the permanent ligature of the cecum or both ligatures were omitted. Duodenal inoculation of organisms in a dose of 100 X the minimum infective dose (MID) in the DIS or DI model did not cause any disease symptom. However, such inoculations were found to cause protection against subsequent challenges with 100 X MID of homologous and heterologous organisms up to 52 weeks after duodenal inoculation. Subcutaneous injection with classical, whole cell cholera vaccine gave only partial protection of short duration. This model might contribute to the understanding of the pathogenesis of cholera as well as to the improvement of efficacy testing of cholera vaccines.

Mechanized procedures for the serology of Salmonella.

A mechanized microtechnique originally designed for the serotyping of E. coli was adapted to the specificity control of diagnostic Salmonella agglutinating antisera as well as for the serotyping of Salmonella. Salmonella strains sent for serotyping were inoculated into a tube with broth which was used as H antigen and, after heating also as O antigen. The agglutination reactions were carried out in clear plastic trays with U formed wells. The antigens were strained with gentian violet in order to obtain a better contrast. With only 6 monofactor O sera, 97% of 100,000 cultures received for serotyping in 1979-1982 could be O antigen typed. Eight H antisera were sufficient for the complete serotyping of 27% of these strains whereas 64% was partly typed. The technique has a high degree of versatility.

Virulence of urinary and faecal Escherichia coli in relation to serotype, haemolysis and haemagglutination.

The virulence of faecal and urinary Escherichia coli strains was studied in relation to serotype, haemolysin production and haemagglutination pattern. By means of an experimental mouse model E. coli strains can be divided into avirulent (I), mouse nephropathogenic (II), and generally virulent (III) strains. Virulent group II and group III strains were more often haemolytic and haemagglutinating than avirulent group I strains. Presence of K antigen could not be associated with virulence. Discriminant analysis for qualitative variables revealed that no combination of the investigated properties contributed more to a strain's virulence level than did one single property. It is concluded that other virulence factors, apart from haemolysin production in group II strains and haemagglutinins in group III strains, must be involved in the determination of a strain's virulence level. All O2, O6 and O18 ac strains tested were virulent, and by far the most O75 strains were avirulent, whereas other O groups were more variable with regard to virulence. Pyelonephritis strains were more often mannose-resistance haemagglutinating than faecal and other urinary isolates, indicating that mannose-resistant adhesins may be important in the pathogenesis of pyelonephritis.