SGM Journals
CLINICAL MICROBIOLOGY AND VIROLOGY

Inhibition of Clostridium difficile strains by intestinal Lactobacillus species

Paul Naaber, Imbi Smidt, Jelena Štšepetova, Tatjana Brilene, Heidi Annuk, Marika Mikelsaar

  • Department of Microbiology, University of Tartu, Ravila 19, 40411, Tartu, Estonia
  • Correspondence Paul Naaber paul.naaberkliinikum.ee

    • Journal of Medical Microbiology 2004; 53(6):551–554 · https://doi.org/10.1099/jmm.0.45595-0

    View at publisher PubMed

    Abstract

    Indigenous intestinal microflora (including lactobacilli) has an important role in protection against Clostridium difficile infection. To assess in vitro interaction between lactobacilli and C. difficile, antagonistic activity of 50 intestinal Lactobacillus spp. strains against 23 pathogenic C. difficile strains was determined. Phenotypic properties of C. difficile strains [production of short-chain fatty acids (SCFAs) and toxin A, and antimicrobial susceptibility] and lactobacilli (production of SCFAs and H2O2) were investigated. Five lactobacilli (Lactobacillus paracasei and Lactobacillus plantarum species) were antagonistic to all C. difficile strains, 18 were antagonistic to some C. difficile strains and 27 showed no antagonistic activity. This antagonistic activity was strain-specific and seemed to correlate with H2O2 and lactic acid production. C. difficile strains that were more sensitive to lactobacilli (n = 9) usually produced higher toxin levels and more SCFAs, and were more resistant to antibiotics, than strains that were resistant to lactobacilli (n = 14). Compatibility of C. difficile strain properties (resistance to lactobacilli or antibiotics) with intestinal microecological conditions (e.g. presence of antagonistic lactobacilli, concentration of antibiotics) may determine expression of disease.

    • Abbreviation: SCFA, short-chain fatty acid.

    Introduction

    Clostridium difficile infection is a frequent complication of antibiotic therapy in hospitalized patients. Impairment of the composition of intestinal microflora is an important prerequisite for colonization by C. difficile and development of infection. It is known that up to 50 % of hospitalized patients become asymptomatic carriers of toxigenic strains and, in symptomatic patients, clinical manifestations vary greatly (Bartlett, 1992; Johnson & Gerding, 1998). The reasons for this may lie in variations in virulence factors of C. difficile strains on one hand and different degrees of disruption of indigenous intestinal microflora and colonization resistance on the other.

    Our previous findings indicate the importance of lactobacilli in maintenance of colonization resistance against C. difficile (Naaber, 1997; Naaber et al., 1997). For prevention and therapy of antibiotic-associated diarrhoea and C. difficile infection, several strains of lactobacilli have been used, but have had varied efficacies (Kotz et al., 1992; Biller et al., 1995; Roffe, 1996; McFarland, 1998; Ouwehand, 1998; Vandenplas, 1999; Pochapin, 2000; D'Souza et al., 2002). Inhibitory potency of different Lactobacillus spp. against C. difficile and mechanisms involved in these interactions have not yet been elucidated.

    The aim of our study was to evaluate the inhibitory activity of various intestinal lactobacilli against different pathogenic C. difficile strains in vitro.

    Methods

    Twenty-three C. difficile strains that belonged to different genotypes and were isolated from faeces of 21 hospitalized patients with sporadic cases of antibiotic-associated diarrhoea were included. Strains were typed with arbitrarily primed PCR as described by Tang et al. (1995). From two patients, two different strains were isolated simultaneously. Strains of intestinal lactobacilli (n = 50) were isolated from faeces of young, healthy children during our previous studies (Mikelsaar et al., 2002).

    In vitro toxin A production by C. difficile strains was detected with the Oxoid toxin A test, according to the manufacturer's instructions. Strains were considered to be low toxin producers if the test was positive with undiluted or 1 : 10 diluted sample and high producers if the test was positive in a 1 : 100 or higher dilution.

    MICs of chloramphenicol, clindamycin, rifampicin, tetracycline, ampicillin, erythromycin, metronidazole and vancomycin were detected with E tests, according to the manufacturer's instructions (AB Biodisk), and interpreted according to the current break-points of the National Committee for Clinical Laboratory Standards.

    For determination of production of short-chain fatty acids (SCFAs), C. difficile strains were cultivated in peptone/yeast/glucose broth and lactobacilli in MRS broth for 48 h. In vitro production of SCFAs was measured by GC as a concentration of these compounds, as described by Holdeman & Moore (1972). Production of H2O2 by lactobacilli was detected on tetramethyl benzidine medium, as described by Eschenbach et al. (1989).

    Antagonistic activity of lactobacilli was detected on agar plates as inhibition of C. difficile growth. Wilkins–Chalgren blood agar was inoculated with a 10 μl aliquot of a 48 h culture of lactobacilli in MRS broth by streaking the plates. After 48 h incubation in a 10 % CO2 environment, 1 μl C. difficile saline suspension (0.5 on the McFarland turbidity scale) was streaked perpendicularly with a lactobacillus growth streak. Inhibition assays with each Lactobacillus sp. and C. difficile combination were repeated at least four times on two separate plates. Plates were incubated for 24 h in an anaerobic environment and inhibition zones (mm) of C. difficile growth from the Lactobacillus sp. streak were measured.

    Data were analysed with Jandel SigmaStat 2.0 software by using Fisher, Spearman correlation and Mann−Whitney tests.

    Results and Discussion

    Based on antagonistic activity against C. difficile strains, we divided the lactobacilli into three groups: (i) those able to inhibit growth of all C. difficile strains investigated (LB+); (ii) those able to inhibit growth of some (nine of 23) C. difficile strains (LB+/−); and (iii) those without any antagonistic activity (LB−). Species composition of these groups is shown in Table 1.

    Table 1. Antagonistic activity of different species of lactobacilli against C. difficile

    Five strains of lactobacilli were able to inhibit growth of all C. difficile strains tested. These belonged to Lactobacillus paracasei and Lactobacillus plantarum [facultatively heterofermentative (FHEL) group]. At the same time, all obligately heterofermentative (OHEL) group species (except one) did not show antagonistic activity against any C. difficile strain. However, the antagonistic activity of lactobacilli was not strictly correlated with a particular species of lactobacilli.

    Based on the interaction with LB+/− group lactobacilli, C. difficile strains could be divided into two groups: nine strains were sensitive to all these lactobacilli and 14 C. difficile strains were resistant. We found that 11 C. difficile strains were high toxin producers, 11 were low toxin producers and one strain did not produce toxin A in vitro. High toxin producers were usually sensitive to LB+/− group lactobacilli, in contrast to low toxin producers (P < 0.001). All C. difficile strains tested were highly sensitive to ampicillin, vancomycin and metronidazole. With other antibiotics, MIC values varied greatly. Strains with higher MICs to chloramphenicol, rifampicin, tetracycline and erythromycin were usually sensitive to the antagonistic activity of LB+/− lactobacilli and were also highly toxigenic (P < 0.001). Thus, based on these phenotypic properties, C. difficile strains could be divided into two typical groups: phenotype A (antibiotic-resistant, more susceptible to antagonistic activity of lactobacilli and highly toxigenic) and phenotype B (sensitive to antibiotics, more resistant to inhibition by lactobacilli and low toxin A producers).

    By comparing quantitative production of SCFAs, we found that C. difficile strains of phenotype A produced significantly higher amounts of isobutyric (mean, 1.84 versus 1.16 mg l−1; P < 0.001), butyric (mean, 6.56 versus 5.15 mg l−1; P < 0.05), isovaleric (mean, 2.14 versus 1.68 mg l−1; P < 0.001) and valeric (mean, 1.85 versus 1.31 mg l−1; P < 0.01) acids, compared to phenotype B strains. There was no significant difference in production of acetic or isocapronic acids. All lactobacilli tested produced acetic and lactic acids. We found that LB+ group strains usually produced more lactic acid (median, 20.9 mg l−1) than LB+/− (median, 18.5 mg l−1; P < 0.005) and LB− (median, 10 mg l−1; P < 0.005) strains. No differences were found in production of acetic acid.

    H2O2 was produced by 42 % of lactobacilli in vitro. We found that LB+ and LB+/− strains were more commonly H2O2 producers (14 of 23) than LB− strains (seven of 27; P = 0.02).

    In our study, we found that both in vitro antagonistic activity of lactobacilli against C. difficile and susceptibility of C. difficile strains to this antagonism was strain-dependent. Previous studies that used only a single indicator C. difficile strain to screen several lactobacilli or a single Lactobacillus sp. strain against different C. difficile strains did not describe this important phenomenon (Bogovic-Matijašić et al., 1998; Forestier et al., 2001; Strus et al., 2001).

    The important role of in vitro antagonistic activity of lactobacilli against potential pathogens and also the ability to sustain colonization resistance in microecosystems in vivo has been attributed by several investigators to the production of H2O2 and SCFAs (Eschenbach et al., 1989; Brashears et al., 1998; Ouwehand, 1998; Brashears & Durre, 1999; Malakar et al., 1999). Others have found that, in some cases, where lactobacilli show high antagonistic activity against pathogens, these mechanisms are not involved or have only a supportive effect (Su et al., 1987; Coconnier et al., 1997). Bogovic-Matijašić et al. (1998) suggested the importance of bacteriocins of lactobacilli in inhibition of C. difficile. Our finding that lactobacilli that were more antagonistic to C. difficile produced H2O2 and high levels of lactic acid more frequently indicates the probable role of these compounds in lactobacilli–C. difficile interactions. However, the role of other mechanisms of antagonism, such as production of bacteriocins, should be investigated and the relevance of in vitro findings should be evaluated by in vivo studies.

    In C. difficile strains, we found an inverse relationship between toxicity, resistance to antibiotics and susceptibility to antimicrobial agents that were produced by lactobacilli. Correlation (positive or negative) between virulence and antibiotic resistance has been described in several other species. On one hand, these properties could be selected by the same environmental factors or mediated by related mechanisms; on the other hand, expression of resistance may require extra energy and thus reduce virulence and metabolic activity (Martínez & Baquero, 2002). In the case of antibiotic-associated diarrhoea, concentration of antibiotics and the antagonistic effect of indigenous microflora are the main selective factors for C. difficile strains in the intestine. Based on our findings, we suggest that if antibiotic concentration is high and the indigenous microflora is deeply suppressed, antibiotic-resistant (but lactobacilli-sensitive) and more toxigenic strains could have an advantage. In cases of milder alteration of indigenous microflora and lower concentrations of antibiotics, strains of C. difficile that are lactobacilli-resistant and more antibiotic-sensitive, but less toxic, have the advantage to survive and cause infections. If this is true, administration of some antibiotics (e.g. clindamycin) induces more serious C. difficile infection, not only due to deeper alteration of indigenous microflora, but also by selection of more toxigenic C. difficile strains.

    We have found a substantial difference in antagonistic activity of intestinal lactobacilli against C. difficile strains with different virulence. Evaluation of the in vivo effect of highly antagonistic strains of lactobacilli and description of the mechanisms involved enable a more complete understanding of the pathogenesis of C. difficile diarrhoea and may introduce new possibilities in an ecological approach to its prophylaxis and therapy.

    References

    1. Bartlett, J. G. (1992). Antibiotic-associated diarrhea. Clin Infect Dis 15, 573–581.
    2. Biller, J. A., Katz, A. J., Flores, A. F., Buie, T. M. & Gorbach, S. L. (1995). Treatment of recurrent Clostridium difficile colitis with Lactobacillus GG. J Pediatr Gastroenterol Nutr 21, 224–226.
    3. Bogovic-Matijašić, B., Rogelj, I., Nes, I. F. & Holo, H. (1998). Isolation and characterization of two bacteriocins of Lactobacillus acidophilus LF221. Appl Microbiol Biotechnol 49, 606–612.
    4. Brashears, M. M. & Durre, W. A. (1999). Antagonistic action of Lactobacillus lactis toward Salmonella spp.and Escherichia coli O157 : H7 during growth and refrigerated storage. J Food Prot 62, 1336–1340.
    5. Brashears, M. M., Reilly, S. S. & Gilliland, S. E. (1998). Antagonistic action of cells of Lactobacillus lactis toward Escherichia coli O157 : H7 on refrigerated chicken meat. J Food Prot 61, 166–170.
    6. Coconnier, M.-H., Liévin, V., Bernet-Camard, M.-F., Hudault, S. & Servin, A. L. (1997). Antibacterial effect of the adhering human Lactobacillus acidophilus strain LB. Antimicrob Agents Chemother 41, 1046–1052.
    7. D'Souza, A. L., Rajkumar, C., Cooke, J. & Bulpitt, C. J. (2002). Probiotics in prevention of antibiotic associated diarrhoea: meta-analysis. BMJ 324, 1361–1366.
    8. Eschenbach, D. A., Davick, P. R., Williams, B. L., Klebanoff, S. J., Young-Smith, K., Critchlow, C. M. & Holmes, K. K. (1989). Prevalence of hydrogen peroxide-producing Lactobacillus species in normal women and women with bacterial vaginosis. J Clin Microbiol 27, 251–256.
    9. Forestier, C., De Champs, C., Vatoux, C. & Joly, B. (2001). Probiotic activities of Lactobacillus casei rhamnosus: in vitro adherence to intestinal cells and antimicrobial properties. Res Microbiol 152, 167–173.
    10. Holdeman, L. V. & Moore, W. E. C. (editors) (1972). Anaerobe Laboratory Manual. Blacksburg, VA: Virginia Polytechnic Institute and State University.
    11. Johnson, S. & Gerding, D. N. (1998). Clostridium difficile-associated diarrhea. Clin Infect Dis 26, 1027–1036.
    12. Kotz, C. M., Peterson, L. R., Moody, J. A., Savaiano, D. A. & Levitt, M. D. (1992). Effect of yogurt on clindamycin-induced Clostridium difficile colitis in hamsters. Dig Dis Sci 37, 129–132.
    13. Malakar, P. K., Martens, D. E., Zwietering, M. H., Béal, C. & van't Riet, K. (1999). Modelling the interactions between Lactobacillus curvatus and Enterobacter cloacae.II. Mixed cultures and shelf life predictions. Int J Food Microbiol 51, 67–79.
    14. Martínez, J. L. & Baquero, F. (2002). Interactions among strategies associated with bacterial infection: pathogenicity, epidemicity, and antibiotic resistance. Clin Microbiol Rev 15, 647–679.
    15. McFarland, L. V. (1998). Epidemiology, risk factors and treatments for antibiotic-associated diarrhea. Dig Dis 16, 292–307.
    16. Mikelsaar, M., Annuk, H., Shchepetova, J., Mändar, R., Sepp, E. & Björkstén, B. (2002). Intestinal lactobacilli of Estonian and Swedish children. Microb Ecol Health Dis 14, 75–80.
    17. Naaber, P. (1997). Clostridium difficile infection and intestinal microbial ecology. PhD thesis, University of Tartu, Tartu, Estonia.
    18. Naaber, P., Klaus, K., Sepp, E., Björksten, B. & Mikelsaar, M. (1997). Colonization of infants and hospitalized patients with Clostridium difficile and lactobacilli. Clin Infect Dis 25, S189–S190.
    19. Ouwehand, A. C. (1998). Antimicrobial components from lactic acid bacteria. In Lactic Acid Bacteria: Microbiology and Functional Aspects, 2nd edn, pp. 139–160. Edited by S. Salminen & A. von Wright. New York: Marcel Dekker.
    20. Pochapin, M. (2000). The effect of probiotics on Clostridium difficile diarrhea. Am J Gastroenterol 95, S11–S13.
    21. Roffe, C. (1996). Biotherapy for antibiotic-associated and other diarrhoeas. J Infect 32, 1–10.
    22. Strus, M., Pakosz, K., Gosciniak, H., Przondo-Mordarska, A., Rozynek, E., Pituch, H., Meisel-Mikolajczyk, F. & Heczko, P. B. (2001). Antagonistic activity of Lactobacillus bacteria strains against anaerobic gastrointestinal tract pathogens (Helicobacter pylori, Campylobacter coli, Campylobacter jejuni, Clostridium difficile). Med Dosw Mikrobiol 53, 133–142 (in Polish).
    23. Su, W. J., Waechter, M. J., Bourlioux, P., Dolegeal, M., Fourniat, J. & Mahuzier, G. (1987). Role of volatile fatty acids in colonization resistance to Clostridium difficile in gnotobiotic mice. Infect Immun 55, 1686–1691.
    24. Tang, Y. J., Houston, S. T., Gumerlock, P. H., Mulligan, M. E., Gerding, D. N., Johnson, S., Fekety, F. R. & Silva, J., Jr (1995). Comparison of arbitrarily primed PCR with restriction endonuclease and immunoblot analyses for typing Clostridium difficile isolates. J Clin Microbiol 33, 3169–3173.
    25. Vandenplas, Y. (1999). Bacteria and yeasts in the treatment of acute and chronic infectious diarrhea.Part I. Bacteria. Clin Microbiol Infect 5, 299–307.