- Volume 51, Issue 3, 2002

The role of interferon (IFN)-γ in host inflammatory responses, including inflammatory cytokine production, in experimental pneumonia with Legionella pneumophila was examined in IFN-γ knockout (IFN-γ−/−) mice. IFN-γ−/− mice and wild-type BALB/cA mice were inoculated intranasally with L. pneumophila strain KC. The survival rate of IFN-γ−/− mice was significantly lower than that of control mice. Viable bacterial counts in lungs and blood showed a rapid and continuous increase in IFN-γ−/− mice, in contrast to a gradual decrease in the lungs and an intermittent bacteraemia in control mice. Histopathological analysis of L. pneumophila-infected lung tissues demonstrated mild pneumonia in control mice, whereas severe pneumonia was shown in IFN-γ−/− mice. During the late stages of infection, the number of total bronchoalveolar lavage (BAL) cells was significantly higher in IFN-γ−/− than in control mice. The concentrations of tumour necrosis factor-α and interleukin-1β in sera of IFN-γ−/− mice were significantly lower in control mice during the early stages of infection, suggesting suppressed production of inflammatory cytokines in IFN-γ−/− mice. In contrast, during the late stages of infection, the levels of these cytokines were significantly higher in sera of IFN-γ−/− mice than in control mice, suggesting severe and systemic infection in IFN-γ−/− mice. The findings suggest that retardation of host immune responses, including inflammatory cytokine production caused by deficiency of IFN-γ, might allow the bacteria to grow and cause fulminant pneumonia.

A multiparametric flow cytometry antimicrobial susceptibility test was developed and its performance was evaluated on clinical urine isolates and samples in comparison with standard methods. Alterations in cytoplasmic membrane integrity were monitored by propidium iodide, and the anionic probe bis-(1,3-dibutylbarbituric acid) trimethine oxonol (DiBAC4(3)) was used to measure changes in membrane potential. Microbial size and cellular content were analysed by light scattering. Twelve antibiotics were tested on 6 ATCC control strains, 22 urine isolates and 19 clinical urine samples, variously containing Escherichia coli, Pseudomonas aeruginosa, Klebsiella pneumoniae, Proteus mirabilis, Enterococcus faecalis, Staphylococcus aureus, S. saprophyticus and S. epidermidis. Agreement between the flow cytometry results, broth microdilution and disk diffusion tests was 93.9% (n=3285mutests). Of the 20 discrepancies observed, 18 were for species other than E. coli. Perfect correlation was obtained with five antibiotics, whereas norfloxacin, nitrofurantoin and tetracycline were responsible for 13(65%) of the 20 discrepancies.

A fatal case of cholera caused by Vibrio cholerae O1 El Tor serotype Ogawa occurred in Aichi Prefecture, Japan in 1995. The patient was identified locally, but the route of the infection was unknown. The causative isolate and 38 other domestic and imported V. cholerae O1 isolates, obtained between 1984 and 1997, were analysed by prophage typing, antimicrobial susceptibility testing and pulsed-field gel electrophoresis (PFGE). This was done to determine whether the isolate from this case differed from others associated with either mild cholera infections or asymptomatic carriage, and to elucidate the route of infection. Cholera toxin (CT) from 37 toxigenic isolates was assayed semi-quantitatively. The 39 isolates were divided into 12 temporary types in accordance with the results of the three typing techniques. The isolate from the fatal infection and nine other isolates were classified as temporary type IV. No difference in CT production was found between the isolate from the fatal case and the other 36 toxigenic isolates. Taken together, it is unlikely that a V. cholerae O1 isolate of distinguishable type was responsible for the fatal illness. Temporary type IV isolates were frequently present in both domestic and imported cases from 1994 to 1997 in Aichi, but they did not emerge before 1993. These results suggest that a new clone was introduced after 1993 from overseas and then disseminated into Aichi, and this may have been an important step in triggering the fatal case of cholera.

An experimental Helicobacter pylori infection in miniature pigs was developed and investigated. Eighteen miniature pigs were inoculated with an H. pylori strain that has high virulence in mice at c. 5×1010cfu. H. pylori infection in miniature pigs was achieved by the administration of agar 1% in brucella broth with fetal bovine serum 10% just before inoculation. The bacterial colonisation and distribution were analysed by mapping of viable cell counts in the stomach in pigs of three different ages. The mapping assay was achieved on post-infection day 3 for the 5-day-old and 2-week-old pigs, and between days 41 and 43 for 3-month-old pigs. The highest cell counts were observed in 5-day-old pigs, which averaged 4.9×1065mucfu/g of mucosa (n=4). The bacteria were colonised mainly in the cardiac and fundus gland region in the 5-day-old and 2-week-old pigs, whereas the colonisation sites did not depend on the region in the 3-month-old pigs. Biopsy assay of the antral mucosa of a 3-month-old pig after H. pylori infection showed that this infection persisted for >22 months. Serum antibody against H. pylori was detected in the infected pigs but not in the uninfected animal. Immunostaining demonstrated the presence of bacteria on the epithelial surface of the infected pigs. A microscopic finding common to all the infected pigs, focal gastritis with infiltration of lymphocytes detected on the lesser curvature of the stomach, resembled the microscopic appearance in H. pylori-infected human patients. These results suggest that miniature pigs might be a suitable model for studying H. pylori infection.

Concern exists over recent unexplained deaths among intravenous drug users. This report describes a patient with crepitant cellulitis who was admitted complaining of severe pain in the right forearm. Ultrasonography demonstrated gas in the tissues and he was referred for early surgical debridement of the arm. He was treated with intravenous benzyl penicillin, gentamicin and metronidazole and made a full recovery. Aspirate samples grew Bacillus cereus, morphologically similar to the isolate obtained from a sample of the patient's own heroin. Antibiogram and API 50CHB profiles were also similar. Further typing included ‘H’ flagellar serotyping, which found both blood and heroin strains to be non-typable, and amplified fragment polymorphism analysis, which showed that the strains were indistinguishable. Genotyping of two selected genes from B. cereus confirmed almost certain identity between the two strains. This case illustrates the potential virulence of B. cereus when inoculated into tissues, and to our knowledge, is the first report to demonstrate a conclusive microbiological link between contaminated heroin and serious sepsis in a drug user due to B. cereus.

Ninety-eight aerobic, gram-negative bacterial isolates from subgingival samples from family-owned dogs with naturally occurring periodontitis were characterised phenotypically by conventional biochemical testing, by cellular fatty acid profiling and by the use of commercial identification systems. The majority (48, 81%) of the fermentative isolates but only 18% of the non-fermenters were identified by conventional biochemical testing alone. With additional cellular fatty acid profiling, another 7 (12%) fermentative and 23 (59%) non-fermentative isolates were identified to genus or group level. Cellular fatty acid analysis was essential for the identification of most non-fermenters, many of which are difficult to identify due to a paucity of positive reactions in routine biochemical tests. Commercial identification systems were less useful and did not contribute to further identification of these problematic isolates. This study underlines the difficulties encountered in the identification of canine oral bacteria – a group of potential bite wound pathogens – and presents schemes for microbiology laboratories to characterise such isolates.

A suspension of human faeces (FS) and its anaerobic culture (FC), bacterial metabolic products and organic acids were examined for inhibitory effects on growth and verotoxin 2 (VT2) production of Escherichia coli O157:H7 in vitro. FS and FC showed a marked inhibitory activity to growth and production of VT2 by E. coli O157:H7 under anaerobic conditions. They may have both bacteriostatic and bactericidal effects on E. coli O157. The growth of E. coli O157 was markedly suppressed by acetic, propionic and butyric acids compared with hydrochloric acid and lactic acid at concentrations between 25 mm and 40 mm, being proportional to the pH values. At pH 5.5, 40 mm of short-chain fatty acids (SCFAs) almost completely inhibited the growth of E. coli O157. SCFAs markedly inhibited the growth of E. coli O157 at pH 6.0 rather than pH 7.0. Propionic acid is likely to be more suppressive to E. coli than acetic and butyric acids. The production of VT2 was approximately proportional to the growth of E. coli O157. However, incubation for 24 h in vitro showed that the growth and VT2 production of E. coli O157 decreased but were not completely inhibited at pH 6.5 and 7.0 in a mixture of acetic, propionic and butyric acids at a physiological concentration (110 mm, 60:25:25, respectively, in molar ratio). It is probable that the colonic microflora could contribute to a reduction of E. coli O157:H7 infections via the activation of intestinal fermentation by dietary manipulation or something similar to give pH 6.0 or <6.0 and that factors such as age, chemical therapy and body condition, which have effects on the balance of the intestinal microflora, would be associated with the incidence rates of E. coli O157 infections.

A cosmid DNA library had been constructed previously from 40-kb fragments of genomic DNA from a virulent invasive strain of Salmonella enterica serotype Typhimurium (TML) in an avirulent hypo-invasive Typhimurium strain (LT7). Selection of invasive clones from the library was attempted by iterative passage through a rabbit ileal organ culture. After the fourth passage, a clone, designated LT7(pHC20uu.2), was isolated. Exposure to both gut tissue and Caco-2 cells enhanced the growth, invasiveness for gut and Caco-2 cells, and flagellin expression of LT7(pHC20uu.2) although its invasiveness was less than that of strain TML. Expression of appendages (surface structures c. 60–70 nm diameter) was shown to play a role in but not to confer invasiveness, and was demonstrated in the absence of direct contact with eukaryotic cells. Exposure to gut tissue also affected the expression of several outer-membrane proteins (OMPs) in all four Salmonella strains – TML, LT7, LT7(pHC79), LT7(pHC20uu.2) – used in this work. As the genes involved in flagella, invasin and porin expression are distributed around the salmonella chromosome, it is possible that pHC20uu.2 encodes a pleiotropic regulator of genes involved in gastro-enteritic virulence and adaptation to the in-vivo gut environment. pHC20uu.2 mapped at c. centisome 25 on the salmonella chromosome close to, but distinct from, SPI-5.

Candida ID is a new chromogenic medium for the identification of yeasts from clinical specimens. C. albicans produces blue pigmentation, whereas pink pigmentation is produced by C. tropicalis, C. lusitaniae, C. guilliermondii and C. kefyr; other Candida species appear white. In this study, 240 clinical samples (throat swabs and stool samples) from haematology patients were inoculated on to Candida ID and Sabouraud-chloramphenicol agar in parallel, yielding a total of 105 yeasts; the media had overall detection rates of 85.7% and 86.7% respectively. The sensitivity of Candida ID for identification of C. albicans by blue pigmentation was 52.9% at 24 h and 94.1% at 48 h. Specificity of the blue pigmentation was 100% at 48 h. Two strains of C. tropicalis were identified, one produced pink pigmentation at 72 h, the other strain did not produce any pigmentation after 5 days. Candida ID was superior in detecting mixtures of yeasts compared with Sabouraud-chloramphenicol agar. Candida ID is a suitable primary isolation medium for yeasts from clinical specimens, providing rapid direct identification of C. albicans and enhanced detection of mixtures.