- Volume 11, Issue 4, 1978

Earlier quantitative investigations of antagonism between skin bacteria were based on the use of liquid cultures, but a more realistic model has now been devised, based on the use of the surfaces of solid media. Pure or mixed inocula were spread evenly over suitable agar media in Petri dishes marked out with a standard grid. Growth curves were constructed from viable counts of the surface bacteria after they had been removed from excised squares of the agar media and dispersed. The method was highly reproducible, and competitive interactions were revealed more clearly than in studies with liquid media. An antibiotic-producing strain of Staphylococcus epidermidis (S6+) readily suppressed strains of Micrococcus, Corynebacterium and Streptococcus species. However, a Staphylococcus aureus strain which was less sensitive to the antibiotic effect of S6+ interacted in a complex manner, depending on the absolute and relative size of the S6+ inoculum.

Antibody responses were studied in human patients from whom leptospiral serovars—mainly pomona or hardjo—had been isolated and identified. The antibody to the polysaccharide F4 antigen belonged exclusively to the IgM class, even as late as 10 months after infection. Human sera cross-reacted widely with F4 antigen from heterologous serovars. The antibodies involved in leptospiral agglutination were mainly IgM, but some patients also produced IgG agglutinins. The titres of IgM agglutinins were higher than those of IgG agglutinins and persisted for many months, regardless of the presence or absence of IgG agglutinins. Both types of immunoglobulin from patients with serovar pomona infection protected hamsters against lethal infections with homologous leptospires. The hamster-protective capacity of human sera correlated well with agglutinin titres. Sera from patients infected with serovars other than pomona protected hamsters against challenge with pomona only if they contained agglutinins to that organism.

The rabbit antibody response to leptospiral F4 antigen extracted from serovar pomona depended on the method of immunisation. Intravenous injection of whole leptospires stimulated F4 antibodies that were confined to the IgM class, but leptospires injected intramuscularly with adjuvant stimulated F4 antibodies in both the IgM and the IgG classes. Both methods of immunisation stimulated agglutinins in both the IgM and IgG classes. F4 antigen in soluble form was immunogenic when injected intradermally with adjuvant, but not when given alone. The F4 antibodies were distinct from the agglutinins in that they did not protect hamsters from acute infection with homologous leptospires, nor did they kill leptospires in vitro although they reacted with leptospires. The hamster-protective capacity of rabbit sera depended on the level of agglutinin.

Serum specimens from 14 infants with congenital rubella were examined for specific IgM antibody by six different methods. IgM-containing fractions were separated either by sucrose density-gradient centrifugation or by gel filtration through Sephadex G-200, and were then tested by the indirect immunofluor-escence technique and by the haemagglutination-inhibition (HI) test (long- and short-incubation methods).

Immunofluorescence staining of density-gradient fractions detected specific IgM in all 14 infants. The HI test (long method), applied to density-gradient fractions, was almost as sensitive, detecting antibody in 13 infants; the short method was less sensitive. The gel-filtration technique proved to be generally less satisfactory than sucrose density-gradient centrifugation.

Evidence was obtained for the occurrence of as yet unclassified non-specific inhibitors in the serum of some infants. These inhibitors were deposited with the IgM on sucrose-density gradients and they could have been mistaken for rubella-specific IgM antibody, particularly in the HI test (long method).

Ninety-nine strains of streptococci were isolated from 97 cases of pyogenic infections, most of which involved the teeth. Physiological and serological tests were performed on these streptococci and on 37 strains of streptococci from culture collections. The results were used for a numerical classification. Seventy-nine of the strains isolated from patients formed a cluster with Streptococcus milleri and group-F reference strains, and were therefore considered as streptococci resembling S. milleri. By the use of an antiserum prepared against strain z3, protein antigens were demonstrated in acid extracts of 65% of the strains of S. milleri. These antigens were found in only five strains not included in the S. milleri cluster.

The microbial flora of the jejunal lumen of 28 infants with acute gastroenteritis was compared with that of a group of 10 normal infants. The jejunum of control subjects harboured an “oral” type of flora and in a few instances enterobacteria in small numbers. The concentrations of all but one of the groups of organisms were higher in the patients than in controls, and the differences were of statistical significance for enterobacteria and lactobacilli. In eight subjects, the same pathogen was identified in the jejunum and the stool. In six subjects with rotavirus infection, there were almost no Gram-negative aerobic rods in the jejunum. The possible role of other Gram-negative aerobic rods in producing gastroenteritis is discussed. It is suggested that studies of jejunal flora are of considerable importance in assigning an aetiological role to bacteria in the causation of acute gastroenteritis.

Small round viruses (SRV) were isolated from the faeces of diarrhoeic calves from three farms. All three SRV preparations caused diarrhoea experimentally in gnotobiotic calves. Each preparation contained viral particles of two morphological types, “astrovirus-like” and “calicivirus-like”, and from one preparation the two particle types were separated from each other. The calicivirus-like agent (“Newbury agent”) was 33 nm in diameter, and caused diarrhoea in gnotobiotic calves, with villous atrophy and D-xylose malabsorption. This virus did not infect cell cultures. The astrovirus-like agent did not cause diarrhoea in two gnotobiotic calves; however, it infected cell cultures (primary calf kidney) and the infected cells immunofluoresced with convalescent gnotobiotic-calf antiserum. The astrovirus-like agents in the three preparations were antigenically related. Experiments in calves showed that there was a degree of cross-protection between the three SRV preparations, as judged by the presence or absence of diarrhoea, but that at least three unrelated pathogens were present.

Purified M proteins were recovered from acid extracts of Streptococcus pyogenes, M-types 1, 3, 6, 12 and 17, by elution from columns of hydroxyapatite of the proteins precipitated with ammonium sulphate. M protein free from non-type-specific antigens was recovered only from M-type 12. Although similar fractions were not recovered from M-types 1, 3, 6 and 17, purified preparations containing a single cross-reactive antigen were obtained. In addition to the M proteins associated with cross-reactive antigens, type-specific antigens that did not stimulate opsonic antibodies were isolated from the acid extracts of M-types 3 and 17. Further study of the purified M proteins revealed molecular weights that ranged from 32 000 to 63 000 daltons, total amino acid compositions that were similar, and N-terminal amino acids that were variable.

Culture filtrates of micro-organisms isolated from the upper intestinal secretions of malnourished children and grown in pure culture were shown to impair the intestinal absorption of water and electrolytes in live rats. Decreased net movement out of the intestinal lumen, or actual secretion of water, sodium or potassium into the intestinal lumen, was found with culture filtrates of single isolates of Staphylococcus epidermidis, Escherichia coli 055, Escherichia coli B7A, Shigella sonnei, Klebsiella pneumoniae, Candida albicans and Candida tropicalis. These organisms have been found to contaminate upper intestinal secretions in malnourished children and it is suggested that the effects observed in these experiments might be relevant to the production of the diarrhoea that is a dominant clinical feature of childhood malnutrition.

By plasmid manipulation and by other techniques, forms of non-pathogenic and of enteropathogenic Escherichia coli strains were obtained that possessed different combinations of the genes that control the production of enterotoxin (Ent) and of 88, 99, 987P, polysaccharide K and O antigens and the utilisation of raffinose and lactose. Mixtures of forms derived from the same strain, each resistant to a different antibiotic, together with much larger numbers of non-pathogenic E. coli organisms and lactobacilli were then given orally to colostrum-deprived calves and lambs and to colostrum-deprived piglets that were genetically resistant (R) or genetically susceptible (S) to infection with 88+ strains of E. coli. The concentrations that each form subsequently attained in different parts of the alimentary tract of these animals was then determined by performing bacterial counts on culture media containing appropriate antibiotics. By this procedure it was possible to compare the part played by the different bacterial components in colonising the alimentary tract.

A non-pathogenic 09:K36:H19 strain, but not E. coli K12, was rendered enteropathogenic for 88S piglets by implanting an 88 and an Ent plasmid in it and for calves and lambs by implanting a 99 and an Ent plasmid in it. The 88 + form of this strain and of wild enteropathogenic strains proliferated throughout the small intestine of 88S piglets, but the concentrations they attained in the small intestine of 88R piglets, calves and lambs were, if anything, less than those attained by their 88− counterparts; the small intestines of 88S and 88R piglets appeared equally susceptible to other colonising factors. E. coli 0149:K91,88 strains that had been isolated from sick calves produced neither small-intestinal colonisation nor diarrhoea in calves. The 99+ form of the 09:K36:H19 strain and of all the wild enteropathogenic strains examined proliferated in the small intestine of calves, but only some of them proliferated in the small intestine of piglets. The enteropathogenicity for piglets of a wild 99+ enteropathogenic strain of calf origin was confirmed by producing small-intestinal colonisation and diarrhoea with it in two colostrum-fed conventionally reared piglets. Forms of some wild enteropathogenic strains from which the 99 plasmid had been removed proliferated in the small intestine of calves and piglets to an extent sufficient to cause diarrhoea, but they did not proliferate to the extent shown by the corresponding 99+ form in mixed infections.

The 987P antigen was a potent factor in the colonisation of the small intestine in piglets but not in calves. In piglets and in calves, the polysaccharide K and O antigens enhanced colonisation. The 88+ forms usually proliferated to a similar degree throughout the small intestine, but forms possessing the other colonising factors either proliferated only in the posterior small intestine or to a much greater degree in the posterior than in the anterior small intestine. Forms possessing both the 987P and 88 antigens proliferated throughout the small intestine of 88S piglets; they proliferated only in the posterior small intestine of 88R piglets and to a lesser degree than the corresponding 987P+88− forms. Neither the Ent plasmid, the Raf (raifinose) plasmid nor the ability to utilise lactose influenced small intestinal colonisation. Wild 88−99−987P− E. coli strains that were enteropathogenic for weaned pigs did not proliferate in the small intestines of piglets.

Pregnant ewes were vaccinated with partially purified, cell-free K99 antigen isolated from an enteropathogenic strain of Escherichia coli, strain B41 (O101:K99:NM), to induce passive immunity via the colostrum in their off-spring against an oral challenge with heterologous “calf-lamb” entero-pathogenic strains of E. coli B44. After sucking their dams, lambs were dosed orally with 7 × 1010-2.2 × 1011 organisms within 4-21 h of birth. One group of 10 lambs was dosed with cultures of the mucoid (09:K30(A),K99:NM) form of strain B44 and another group of 10 lambs with the non-mucoid (09:K99:NM) form; two groups of four control lambs from unvaccinated dams were similarly challenged. All four control lambs challenged with mucoid B44, but none of 10 lambs from vaccinated dams, developed severe, watery diarrhoea. In the group of lambs challenged with the non-mucoid form of strain B44, loose faeces were detected in only two of the four control lambs and in none of the lambs from vaccinated dams. This suggests that the polysaccharide K antigen may contribute to the virulence of “calf-lamb” enteropathogenic strains that possess the K99 antigen. However, lambs passively immunised with colostrum from dams vaccinated with K99 antigen alone were protected against the production of enteric colibacillosis by oral challenge with EPEC strain B44.

The variation in metabolism of biochemical test substrates by klebsiella isolates has been demonstrated. It is suggested that this variation is likely to result in false-negative reactions in biochemical tests incubated for short periods. The observations made may explain the reported difficulties in obtaining reproducible results in biotyping Klebsiella strains.

Preliminary work suggests that differences in substrate metabolism will provide a means of increasing the sensitivity of methods for the biochemical typing of Klebsiella spp.

Known numbers of standard strains of Escherichia coli and Bacteroides fragilis were inoculated together into surgical incisions in guinea-pigs. Histo-logical and quantitative bacteriological proof is presented that pathogenic synergy occurred in vivo with copious pus formation when combinations of subinfective doses of each organism were inoculated. A further series of experiments showed that the amounts and ratios of E. coli and B. fragilis in the septic inocula were critical, and that the resultant pus contained equal numbers of both the aerobic and the anaerobic organisms. This model may be suitable for the experimental validation of short-term regimens of prophylactic antibiotics in surgery; the present results seem to have important therapeutic implications.

A radioimmunoassay developed to measure antibody against Myco-bacterium leprae antigen 7 in man was applied to the nine-banded armadillo (Dasypus novemcinctus). Normal armadillo sera had low but significant antibody activity in the test. Fourteen of 17 armadillos with systemic myco-bacterial infection after inoculation with M. leprae showed increased antibody activity in the assay, and in some instances the activity was higher than in a pool of sera from patients with lepromatous leprosy. Crossed immunoelectro-phoresis with armadillo serum in the intermediate gel revealed antibodies against five distinct antigenic components of M. leprae. Development of systemic mycobacterial infection after inoculation with M. leprae is thus associated with a distinct humoral immune response. The use of radioimmunoassay for selection of animals for inoculation and for following the development of the infection is discussed.

Marmosets, from which endogenous ureaplasmas had been eradicated by treatment with minocycline, were tested for susceptibility to infection by urea-plasmas from the genital and respiratory tracts of other animal species. They could be infected with ureaplasmas of human and simian origin, but were resistant to bovine and canine ureaplasmas. The results indicated that human, marmoset and squirrel-monkey ureaplasmas may form a biological subgroup, distinct from bovine and canine ureaplasmas, and that host range should not be ignored as a parameter for classification.