- Volume 13, Issue 1, 1980

The immune response to Salmonella typhimurium was investigated in CB. and BALB/C mice, by an arithmetic Widal agglutination assay. High antibody titres were obtained in CBA and low titres in BALB/C mice after secondary immunisation.

A rabbit antiserum raised against S. typhimurium was tested for agglutinating activity after absorption with spleen- and lymph-node cells obtained from unimmunised CBA or BALB/C mice. BALB/C cells consistently removed more anti-S. typhimurium antibodies than did CBA cells, whilst the quantities of Fc receptor-bearing cells were found to be similar in both strains when measured by the erythrocyte-antibody rosette technique.

It is suggested BALBjC mice give a low antibody response because their cell-surface antigens crossreact to a greater degree with S. typhimurium than do CBA cell-surface antigens.

Mouse peritoneal macrophages and polymorphonuclear leucocytes were examined for their ability to kill Mycoplasma pulmonis in vitro. Killing of mycoplasmas was shown to require specific antibody, suggesting that antibody-dependent phagocytosis may be involved in resistance to infection. Convalescent mouse serum appeared to be less effective than rabbit antiserum in promoting the killing of mycoplasmas by phagocytic cells. Moreover, mycoplasmas recovered directly from the respiratory tract of infected mice were more resistant to killing by macrophage cultures than organisms grown in vitro. The possibility that these observations may contribute to an understanding of the persistent nature of M. pulmonis infections in mice is discussed.

The growth of 95 strains of Klebsiella cultured in air with added CO2 is described. Twenty-two strains of K. ozaenae required CO2 for characteristic growth when cultured at 37°C and growth was promoted by co-trimoxazole, tetracycline and sodium barbitone. Four strains of K. pneumoniae, two of K. edwardsii var. atlantae and one of K. edwardsii var. edwardsii gave increased viable counts after incubation in CO2. None of the 40 strains of K. aerogenes tested required CO2 for growth.

Acute cervical adenitis was investigated in 26 children over a 33-month period. Lymph nodes were aspirated and cultured to yield Staphylococcusaureus or group-A streptococci or both from 22 (84.6%) of the aspirates. Mycobacterium kansasii was isolated from one child.

By combined culture and serial serological tests (antistreptolysin-0, antiDNAase B, and anti-NADase), 50% of the cases were found to be associated with streptococcal infection. In a few children, S. aureus was the only organism recovered, and appeared to be the aetiological agent.

A method has been developed, based on techniques used for isolating bacteria from the rumen, that enables human faeces to be fractionated into three major components. The method requires repeated, vigorous agitation of a suspension of faecal solids with detergent, the use of a stomacher, and high-speed centrifugation. By this means the faecal micro flora are separated from faecal dietary-fibre residues. These two components, with water-soluble material in the stool, comprise 98.3 ± 0.9% of faecal solids. The purity of the microbial fraction was demonstrated by Gram and plant stains and by scanning electron microscopy. Microscopic counts of the bacteria in each fraction of the stool showed that the microbial fraction contained 95% of the total bacteria. Chemical analysis of the component sugars indicated 6–7% possible contamination by non-bacterial polysaccharides. The bacterial pellet was 6% nitrogen, and accounted for 60% of the total faecal nitrogen. When faeces from nine healthy subjects on a metabolically controlled British-type diet were studied, bacteria comprised 54.7 ± 1.7% of the total solids, fibre 16.7 ± 0.8% and soluble material 24.0 ± 1.3%. Bacteria therefore represent a much larger proportion of the faecal mass than was previously thought.

Properties of Escherichia coli considered to be important in the pathogenesis of urinary-tract infection were investigated. The following properties were more common in E. coli strains isolated from urinary infections than in periurethral strains from healthy individuals: (i) 0 serogroups 2, 4, 6, 8, 18ab and 75; (ii) high K-antigen titre; (iii) production of haemolysin; (iv) production of fimbriae; (v) fermentation of salicin. The correlation between isolation of a strain from the urinary tract and possession of any single property was not strong; however, strains rich in a combination of these pathogenic properties were rarely isolated from the periurethral area of healthy subjects but were common in urinary infections. Nevertheless, a significant proportion of urinary strains had few pathogenic properties. Strains rich in pathogenic properties were more commonly isolated from upper urinary-tract infections than from lower-tract infections; this indicates that the properties of the invading organism may influence the localisation of infection.

GRAM-NEGATIVE anaerobic bacilli are a major component of the normal flora of human faeces. In quantitative studies, Bacteroides spp. have been isolated in counts of 1011 organisms/g of faeces (Hill et al., 1971; Finegold, Attebury and Sutter, 1974; Finegold, 1977). Drasar (1974) isolated Bacteroides in greater numbers from subjects from developed countries whereas gram-positive anaerobes were predominant in the faeces of subjects from developing countries.

Investigators who have identified faecal Bacteroides have found that the B. fragilis group are the dominant species. A previous taxonomic scheme (Hol-deman and Moore, 1974a) assigned all the bile- and penicillin-resistant faecal strains to the single species B. fragilis, with five subspecies; when detailed identification was undertaken the commonest faecal subspecies were ss. vulgatus and ss. thetaiotaomicron (Werner, 1974; Moore and Holdeman, 1975; Finegold et al., 1975). However, DNA-homology studies have shown that each subspecies is unique and should be assigned species status (Cato and Johnson, 1976) and several new species have been described to accommodate strains with different properties within the B. fragilis group (Holdeman and Moore, 1974b; Werner, Rintelen and Kunstek-Santos, 1975). Many other Bacteroides and Fusobacterium spp. have been found inconsistently in small numbers in faeces. This may reflect the problems of isolation, identification and classification of these organisms. B. asaccharolyticus (B. melaninogenicus ss. asaccharolyticus; Finegold and Barnes, 1977) has been isolated regularly from faeces (Werner, Pulverer and Reichertz, 1971) but in much smaller numbers than the B. fragilis group.

The present study was undertaken to assess the contribution of the different species and subspecies of Bacteroides to the normal faecal flora as part of the evaluation of a scheme for the identification of gram-negative anaerobic bacilli by conventional bacteriological tests (Duerden et al., 1980).

The occurrence of gram-negative anaerobic bacilli in the normal vaginal flora was studied in 20 normal healthy women attending a family-planning clinic. A swab was taken from the cervix and posterior fornix and Bacteroides spp. were isolated on a selective medium from 13 (65%) subjects. A heavy growth of Bacteroides was obtained from 11 specimens but only a few colonies were isolated from two specimens. Where possible, 10 representative colonies from each subject were studied and 113 isolates were identified by conventional bacteriological tests. Most isolates (78%) belonged to the B. melaninogeni-cusjoralis group. The commonest species identified in this group were the B. biviusjdisiens complex (42%), B. melaninogenicus ss. intermedius (22%) and ss. melaninogenicus (16%). Asaccharolytic strains were isolated in smaller numbers from 54% of subjects, but only five strains of the B.fragilis group were isolated from two subjects; fusobacteria were not detected.

Gram-negative anaerobic bacilli were isolated on an enriched selective medium from specimens of subgingival plaque from 20 normal, healthy adults without periodontal disease. A heavy growth of Bacteroides and Fusobacterium spp. was obtained from all specimens and 10 representative colonies from each subject were selected for identification by conventional bacteriological methods. Results were obtained in these tests with 188 isolates; some strains of the B. melaninogenicusloralis group would not grow in pure culture on solid media and showed satellitism in mixed culture, and some strains of fusobacteria did not survive repeated subculture. Most isolates (68%) belonged to the B. melaninogenicus/oralis group. The commonest species identified in this group were B. oralis (41%), B. melaninogenicus ss. melaninogenicus (26%) and ss. intermedius (17%). B. oralis was isolated from all subjects and at least one strain of B. melaninogenicus was isolated from 18 out of 20 subjects. Only eight strains of the B. fragilis group were detected and they probably represented transient carriage. Fusobacterium spp. and L. buccalis were isolated from 15 out of 20 subjects and accounted for 21% of the isolates studied, but most could not be identified specifically by the set of tests used. Asaccharolytic species of Bacteroides were isolated in small numbers from seven subjects.

Intestinal absorption of arbutin (p-hydroxyphenyl-β-glucoside), a non-metabolised analoque of d-glucose, and net flux of water out of the small intestinal lumen were found to be decreased in vitro in rats fed with Candida albicans. In rats on a protein-deficient diet and fed C. albicans there was net secretion of water. Sugar uptake and net water transport were also impaired in infected animals in vivo. These abnormalities were present although there was no histological evidence of invasion of the small-intestinal mucosa by C. albicans. The findings suggest that C. albicans has significant enteric pathogenicity and that this is particularly important in malnutrition.

PLATES III-IV

THE ABILITY of bacteria to adhere to epithelial surfaces has been correlated with their pathogenicity. The bacteria studied include enteropathogenic Escherichia coli in the gut of man (McNeish et al., 1975) and of animals (Sellwood et al., 1975; Nagy, Moon and Isaacson, 1977), streptococci in the human oral cavity (Ellen and Gibbons, 1974), gonococci in the urogenital tract (Ward and Watt, 1975), Vibrio cholerae in the small intestine (Freter and Jones, 1976; Jones, Abrahams and Freter, 1976) and mycoplasmas in the respiratory tract (Powell et al., 1976).

The adhesion between bacteria and epithelial surfaces may be specific and selective, involving bacterial cell-surface components and specific surface components of the host mucosal cells. For example, pathogenic strains of E. coli adhere to the ileum rather than the duodenum of pigs and calves, and within the ileum adhere to the tips rather than the bases of the villi (Arbuckle, 1970; Smith, 1977). However, the surface components of bacteria and of host cells responsible for this adhesion have not been fully defined. The bacterial surface components may be fimbriae. These are non-flagellar filamentous appendages, the term "fimbriae" being introduced by Duguid and co-workers (1955) who described the haemagglutinating ability of fimbriate bacteria. Other factors in addition to fimbriae may also be involved or the mechanism of adhesion may depend upon factors that do not include fimbriae.

Studies on adhesion of E. coli to epithelial surfaces have included observations on adhesion to and penetration of monolayers of cultured HeLa and Hep 2 cells (Du Pont et al., 1971), guinea-pig eyes (Serény, 1957; Du Pont et al., 1971; Guerrant et al., 1975), adhesion studies with epithelial cells from the urinary tract (Svanborg Edén et al., 1976) and with epithelial cells from the buccal cavity (Ellen and Gibbons, 1974).

The aim of our investigation was to compare the ability of strains of E. coli to adhere to and to invade a variety of cell surfaces, and to compare the adhesive properties of strains of E. coli from urinary-tract infections and from normal faeces, to determine whether there is a correlation between the source of a strain and its adhesive and invasive properties.

Sulphadiazine and trimethoprim in a wide range of concentrations were added to urine from patients with untreated urinary-tract infections. At therapeutic concentrations, the antibacterial activity of trimethoprim was not increased by the addition of sulphadiazine. Exposure of Escherichia coli to trimethoprim in urine was not associated with an increase in resistance to that agent. It was also not possible to select, in vitro, stable resistance to trimethoprim in sensitive cultures of E. coli.

At therapeutic levels in blood, trimethoprim and sulphadiazine singly produced mainly a bactericidal action on pathogens responsible for urinary-tract infections. Sulphadiazine occasionally enhanced the effect of trimethoprim at subtherapeutic levels. These findings support the need for further evaluation of trimethoprim alone, rather than its use as a combination with a sulphonamide.

PLATE V

GROUP-B streptococci were recognised first as animal pathogens (Stableforth, 1932; Lancefield, 1934). However, Colebrook and Purdie (1937) related group-B streptococci to human disease, and there have since been reports linking group-B streptococci with human infections including septic abortion, puerperal sepsis, bacteraemia, neonatal sepsis, mastitis, meningitis, arthritis, osteomyelitis and inflammation of the urogenital tract (Duma et al., 1969, Eickhoff et al., 1964; Jelinkova, Neubauer and Duben, 1970; Bayer et al., 1976); the most fully documented of these are neonatal sepsis and meningitis. The first published account of neonatal meningitis caused by this organism in Great Britain was given by Jones and Howells (1968). Since then reports from all over the world have implicated group-B streptococci as neonatal pathogens (Jelinkova, 1977). It has been suggested that the newborn acquire the organism during passage through the birth canal (Hood, Janney and Dameron, 1961; Eickhoff et al., 1964; Baker and Barrett, 1973; Franciosi, Knostman and Zimmerman, 1973; Finch, French and Phillips, 1976). Group-B streptococcal infections in neonates appear to be of two types: one, an "early-onset" disease in the first week of life which may be septicaemic or meningitic, the other a "late-onset" disease which is almost always meningitic. There is little doubt that early-onset infections are the result of maternal carriage of the organism, but the source of the infecting streptococcus in late-onset disease is still unknown (Paredes et al., 1977).

Mycoplasma fermentans inoculated directly into the uterine tubes of female grivet monkeys produced a self-limiting acute salpingitis and parametritis. The inflammation was accompanied by a significant rise in the titre of specific indirect haemagglutinating antibodies. Inoculation of M. fermentans into the uterine cavity through the cervical canal without dilatation of the cervix produced practically no signs of inflammation and no antibody response. However, when the intrauterine inoculation of mycoplasmas was followed by curettage of the endometrium, in animals whose uterine tubes had been closed by ligatures, pronounced upper genital-tract inflammation developed, together with a significant antibody response.

PURIFIED δ-haemolysin from several canine strains of Staphylococcus aureus give reactions of only partial identity in immunodiffusion test with purified δ-haemolysin from S. aureus strain CN4108 (Newman) and its corresponding rabbit antibodies (Turner, 1978). This raises the question of whether these antibodies to the δ-haemolysin of strain CN4108 would neutralise the haemolytic activity of δ-haemolysin from the canine strains. Because the δ-haemolysin from both sources was neutralised by a non-immunoglobulin component of normal rabbit serum (Jackson and Little, 1958; Gladstone and Yoshida, 1967; Turner, 1978), it was essential to use only the immunoglobulin fraction of immune sera in this investigation.

THE HORNY layer of epidermis, which can be colonised by dermatophytes, is covered by surface lipids. The lipolytic activity of dermatophytes may therefore be of particular importance in the early stages of infection, before the fungus has established permanent contact with keratin; but little is known about the possible presence of lipases in these pathogenic fungi. The early observations of Mallinckrodt-Haupt (1927), Tate (1929), Nékám (1935) and, more recently, Pinetti and Lostia (1961) and Böhme (1969) have demonstrated that dermatophytes can hydrolyse lipids of animal and of vegetable origin. The occurrence of lipase has been demonstrated histochemically (Knight, 1957) and in dermatophyte mycelial filtrates (Böhme, 1969). This paper describes the screening of the more common dermatophytes to test for the presence of lipases.

PLATES VIII-XI

IN previous studies Bordetella pertussis was shown to be resistant to the mucociliary clearance mechanism of the rabbit trachea (Matsuyama, 1977). Scanning electron microscopy readily demonstrated B. pertussis organisms associated with clusters of the cilia of the epithelial cells. A similar association has been described in chick organ cultures (Iida and Ajiki, 1974, 1975) and hamster organ cultures (Collier, Peterson and Baseman, 1977; Muse, Collier and Baseman, 1977). B. pertussis is not a natural pathogen of the rabbit, chick or hamster; on the other hand, Bordetella bronchiseptica is a natural respiratory-tract pathogen of the rabbit and the dog (Wilson and Miles, 1975). Bemis, Greisen and Appel (1977) demonstrated experimentally the strong affinity of B. bronchiseptica for the canine respiratory tract; they also mentioned similarities between the respiratory infections produced by B. pertussis and B. bronchiseptica.

In the present study, scanning electron microscopy has been used to observe infections of the rabbit tracheal mucosa with B. bronchiseptica, and with Staphylococcus aureus---an organism that has no special predilection for the respiratory tract.

Dane-particle cores labelled with [3H]TTP were subjected to ultracentrifugation in a metrizamide density gradient. Two populations of core particles with different densities were obtained, and radioactivity was found only in the cores that sedimented at the lower density (1.19–1.23 g/cm3). All the cores in this group, when spread in a monolayer, were found to expel a closed circular double-stranded DNA molecule. In contrast, the core particles that sedimented at the higher density (1.23–1.27 g/cm3) were not associated with radioactivity, nor was any DNA strand extruded from them. These results show that metrizamide density gradients allow the separation of complete hepatitis B virions for the study of viral DNA.

An enzyme-linked immunosorbent assay (ELISA) for Brucella abortus antibody was compared with direct agglutination, complement fixation and Coombs tests in studies of 112 sera. Of these, 15 gave positive results by ELISA tests which included the 13 sera that gave positive results in the other tests. The assay aided recognition of four groups: (i) sera with very high levels of specific IgG and significant levels of IgM and IgA from patients with acute brucellosis; (ii) sera with high levels of IgG only from patients with chronic brucellosis; (iii) sera with low levels of IgG only, representing residual antibody; (iv) sera with little or no brucella-specific antibody.

ELISA was antibody-class specific, and the results were more readily interpreted than conventional serological data. Moreover, ELISA was more sensitive, more rapid and simpler than the battery of agglutination, Coombs and complement-fixation tests commonly used.