Previous Article | Next Article ![]()
Applied and Environmental Microbiology, February 2007, p. 855-860, Vol. 73, No. 3
0099-2240/07/$08.00+0 doi:10.1128/AEM.02001-06
Copyright © 2007, American Society for Microbiology. All Rights Reserved.
Laboratoire de Microbiologie, EA 4065, UFR des Sciences Pharmaceutiques et Biologiques, Université René Descartes-Paris 5, 75270 Paris Cedex 06, France,1 Centre Hospitalier de Versailles, Le Chesnay, France2
Received 23 August 2006/ Accepted 15 November 2006
|
|
|---|
|
|
|---|
The evolution of bacterial strains toward multiresistance depends on the ecosystem in which they live. The digestive tract, colonized by a complex microbial flora up to 1012/g (wet weight) of gut contents, is a privileged site of horizontal transfer of plasmids carrying antibiotic resistance genes and contributes to the maintaining and the dissemination of resistance (23). Likewise, extended-spectrum ß-lactamases (ESBLs), enzymes capable of efficiently hydrolyzing extended-spectrum cephalosporins and associated with several treatment failures, disseminate through plasmid transmission among Enterobacteriaceae, a main genera of the aero-anaerobic flora of the digestive tract (28). Under normal physiological conditions, resistant bacteria are repressed by the dominant digestive flora (33). However, the ingestion of antibiotics, even at low dosages, may disturb this equilibrium, modify the barrier effect of the dominant flora, and allow the proliferation of resistant strains. In addition, antibiotic therapy can exert selective pressure on plasmid-carrying strains, and certain studies have reported that antibiotics can even induce plasmid transfer under certain circumstances (4, 16, 19).
The concept of modulating the gut microflora by probiotics has emerged these last two decades. Probiotics are defined as live microorganisms that exert beneficial effects on the host physiology when administered at an adequate dosage (6). Probiotic bacteria are mainly part of two genera, i.e., Lactobacillus and Bifidobacterium, and act either by modifying the intestinal microbiota, increasing the number of beneficial bacteria, or direct contact on mucosal cells (5). A number of health effects have been associated with the use of probiotics. For instance, there is a strong evidence of the interest of probiotic administration in the reduction of the duration and the severity of diarrhea associated with rotavirus or in the prevention of antibiotic-associated diarrhea (30, 31). New studies support other beneficial effects for a broad set of disorders, including atopic disease, irritable bowel disease, and inflammatory bowel disease, but these studies need to be confirmed by additional clinical trials (5, 24, 26). However, the potential use of probiotics to limit the emergence of bacterial resistance and the dissemination of resistance genes has been briefly investigated. One team observed the beneficial effect of a Lactobacillus supplementation during treatment with penicillin and quinolone antibiotics (29), and another team revealed the inhibitory effect of the yogurt symbiosis on horizontal antibiotic resistance gene transfers (17, 18).
In this context, we sought to determine the role of bifidobacteria in the transfer of antibiotic resistance genes in vitro and in vivo and to evaluate the impact of different bifidobacterial species on the establishment of transconjugants in the digestive tracts of gnotobiotic mice.
|
|
|---|
Media.
Culture of bifidobacteria was performed in TGYH broth containing tryptone peptone (30 g/liter), glucose (5 g/liter), yeast extract (20 g/liter), and hemin (5 mg/liter) for 48 h at 37°C in an anaerobic chamber (N2:CO2:H2; 80:10:10). Enterobacterial strains were grown in the same media for 18 h at 37°C in aerobiosis. After the mating step, enterobacterial populations were enumerated on eosin methylene blue medium (BioMerieux, Marcy l'Etoile, France) containing ceftazidime (4 mg/liter) for the donors, nalidixic acid (50 mg/liter) for the recipients, and nalidixic acid and ceftazidime for the transconjugants. Bifidobacteria were enumerated on Wilkins-Chalgren modified for bifidobacteria (WCB) medium (2) containing polymyxin B (10 mg/liter) and on WCB containing polymyxin B and ceftazidime (4 mg/liter) for those that could have acquired a bla gene. Eosin methylene blue plates were incubated aerobically for 24 h at 37°C, and WCB plates were incubated for 48 h at 37°C in an anaerobic chamber.
In vitro broth mating.
Conjugation experiments using K. pneumoniae (SHV-5) and E. coli or serovar Typhimurium (CTX-M-15) and E. coli were performed according to the broth mating procedure as follows. After culture in TGYH broth, 2 ml of the donor, 2 ml of the recipient, and 2 ml of a bifidobacterial culture were mixed in a sterile tube. A control tube was prepared by replacing the bifidobacterial culture with 2 ml of TGYH broth. After 5 h of anaerobic incubation at 37°C, the cells were diluted in regenerated sterile water supplemented with peptone in 10-fold series, and 100 µl of each dilution was plated on selective medium. Transfer frequencies were calculated by dividing the number of transconjugants by the number of donor colonies. Each experiment was repeated three times.
In order to investigate the mechanism of action of bifidobacteria on gene transfer, experiments were performed by replacing the bifidobacterial culture by either a heat-killed culture of bifidobacterial strain (culture in TGYH broth treated 20 min at 100°C), the supernatant of a culture in TGYH broth, or the bacterial pellet of a culture in TGYH broth, washed and suspended in TGYH broth.
In vivo mating experiments. (i) Mice.
Protocol for in vivo experiments was approved by the animal care and use committee of the Faculté des Sciences Pharmaceutiques et Biologiques (Paris, France). Germfree consanguineous C3H mice were maintained in separate isolators (J.C.E. Biotechnology, Vichy, France) and fed ad libitum with a commercial diet sterilized by gamma irradiation (4 Mrad) and supplied with autoclaved drinking water. The germfree status of mice was ensured before the beginning of the experiment by testing fecal samples for aerobic and anaerobic growth of bacteria and yeast.
(ii) Experimental design.
In vivo transfer was studied in the absence (control groups) or presence of bifidobacteria. In the control groups, the recipient strain E. coli K-12 was inoculated first (8 log10 CFU) and then, 8 days later, after checking the implantation of the recipient strain, the mice were inoculated with the donor strain (8 log10 CFU) K. pneumoniae (referred to as blaSHV-5 mice) or serovar Typhimurium (referred to as blaCTX-M-15 mice). In the following experiments, germfree mice received, in addition to the donor and the recipient strains, one bifidobacterial strain. Every germfree mouse received first intragastrically 8 log10 CFU of the recipient E. coli K-12. Three days later, after verification of the implantation of the recipient strain, the mice were inoculated with 8 log10 CFU of the bifidobacterial strain, and 5 days later, after we checked the implantation of the recipient and the bifidobacterial strains, the mice were inoculated with 8 log10 CFU of the donor strain. Four experiments were carried out, the first three using the donor strain K. pneumoniae and the bifidobacterial strains B. longum CUETM-89-215, B. bifidum CIP-56.7T, and B. pseudocatenulatum CIP-104168T. The fourth experiment used the donor strain serovar Typhimurium and the bifidobacterial strain B. bifidum CIP-56.7T. Each experiment lasted 2 months and included six mice.
(iii) Analysis of fecal samples.
For each model, the evolution of the various bacterial populations in mice feces was studied since this reflects the number of bacteria in the cecum (10). Fecal samples (one to two pellets) collected five times per week by provoked defecation were diluted in 10-fold series and spread onto the appropriate plates supplemented with antibiotics by using the automated WASP spiral system (AES Laboratoire, Bruz, France). Bacterial counts were expressed per gram of feces and as log10 values.
Antibiotic susceptibility testing and ESBL production detection.
Ten transconjugants from each in vitro experiment and ten others from each fecal sample collected from mice were selected randomly for antibiotic susceptibility determination using the disk diffusion method according to the recommendations of the Comité de l'Antibiogramme de la Société Française de Microbiologie (3). The following commercial disks (Bio-Rad, Marne-la-Coquette, France) were tested: ceftazidime (30 µg), amoxicillin (20 µg)-clavulanic acid (10 µg), cefotaxime (30 µg), kanamycin (30 µg), tetracycline (30 IU), and nalidixic acid (30 µg). ESBL production was detected by the double-disk synergy test according to the method of Jarlier et al. (13). The presence of blaSHV and blaCTX-M genes in donor strains and in transconjugants selected during in vitro and in vivo experiments was confirmed by using PCR as described by Perilli et al. (25) and Mulvey et al. (22), respectively.
Statistical analysis.
Transfer frequencies with or without bifidobacteria were compared for in vitro experiments by using the Student test and for in vivo experiments by using the Mann-Whitney U test for small samples.
|
|
|---|
0.01) (Table 1) . The highest decreases in transconjugant counts were observed in conjugations including B. pseudocatenulatum CIP-104168T and were equivalent to 2.6 and 2.1 logs for blaSHV-5 and blaCTX-M-15 transfers, respectively. In contrast, insignificant changes in the transfer frequencies were noticed with B. bifidum BS119 and B. pseudocatenulatum BS4. In all experiments, the presence of bifidobacteria did not modify the levels of donors and recipients. |
View this table: [in a new window] |
TABLE 1. In vitro transfer of bla genes between enterobacteria in the absence or presence of bifidobacteriaa
|
![]() View larger version (36K): [in a new window] |
FIG. 1. Effects of different bifidobacterial preparations of B. longum CUETM-89-215, B. bifidum CIP-56.7T, and B. pseudocatenulatum CIP-104168T on transconjugant counts in the conjugation medium. (A) Transfer of blaSHV-5; (B) transfer of blaCTX-M-15. Data, expressed as means ± the standard errors, are from three independent experiments. Symbols: , control; 1% [Student test]).
|
Experiments in the absence of bifidobacteria (control groups).
Transconjugants were detected in the first samples collected 1 day after donor inoculation, revealing the ease with which transfers of blaSHV-5 or blaCTX-M-15 gene occurred in vivo. The transfer of both genes was confirmed by phenotype and genotype analysis: both ESBL production and either blaSHV-5 or blaCTX-M-15 genes were detected in all transconjugants resistant to nalidixic acid and ceftazidime. Transconjugants established at a mean level of 5 log10 CFU per g of feces in blaSHV-5 mice and 8 log10 CFU per g of feces in blaCTX-M-15 mice.
Experiments in the presence of bifidobacteria.
On days 10, 11, and 14 the level of transconjugants in blaSHV-5 mice was the same in all three groups treated with bifidobacteria, i.e., B. bifidum CIP-56.7T, B. pseudocatenulatum CIP-104168T, or B. longum CUETM-89-215. Starting at day 14, a significant decrease in transconjugant levels was observed when mice were associated with B. bifidum CIP-56.7T or B. pseudocatenulatum CIP-104168T (P
0.05). Indeed, B. bifidum CIP-56.7T and B. pseudocatenulatum CIP-104168T both led to 3.3-log decreases at day 17 and to 1.4- and 2.8-log decreases, respectively, in the last 3 weeks of the experiments (Fig. 2) . In contrast, B. longum CUETM-89-215 was not associated with a significant decrease in the transconjugant level in gnotobiotic mice. In blaCTX-M-15 mice, the bifidobacterial strain B. bifidum CIP-56.7T led to a 2.1-log decrease in transconjugant levels throughout the experiment (data not shown).
![]() View larger version (8K): [in a new window] |
FIG. 2. Bacterial counts in homogenized feces of gnotobiotic mice. On day 0, 18 germfree mice were inoculated with 108 E. coli K-12 recipients. On day 3, six of these mice were inoculated with 108 B. bifidum CIP-56.7T and six others were inoculated with 108 B. pseudocatenulatum CIP-104168T. The remaining six mice corresponded to our control group. On day 8, all mice were inoculated with 108 K. pneumoniae CIP-104771::blaSHV-5 donors. The figure shows the geometric averages of the recipient and donor counts in the 18 mice and the geometric averages of transconjugant counts in each group of six mice. The error bars indicate the standard deviations. Symbols: , E. coli K-12; , K. pneumoniae CIP-104771::blaSHV-5; , transconjugants in the absence of bifidobacteria; , transconjugants in the presence of B. bifidum CIP-56.7T; , transconjugants in the presence of B. pseudocatenulatum CIP-104168T.
|
|
|
|---|
The emergence of multiresistant bacteria is now an increasing problem that could be handled by applying antibiotic policies, such as the administration of shorter courses of antibiotics and/or the cycling of antibiotics (12). The use of probiotics can also be an alternative to the latter policies as suggested by Levy (16). In fact, probiotics, acting as a competitive flora, could avoid or suppress the gut colonization by resistant bacteria through their barrier effects. Sullivan et al. examined whether Lactobacillus sp. strain F19, in conjunction with antibiotics, prevents the establishment of resistant bacteria in the gastrointestinal tract and reported that the overall difference between the placebo and supplemented groups was limited (29). However, our study revealed that probiotics might interfere by reducing antibiotic resistance genes transfers.
Such an inhibitory effect was never described for bifidobacteria. A previous study demonstrated that the daily consumption of yogurt symbiosis, i.e., Lactobacillus bulgaricus and Streptococcus thermophilus, decreased plasmid transfer between two Escherichia coli strains in the digestive tract of mice colonized with human fecal flora (17). A later study involving the same model evaluated the effect of yogurt, heat-treated fermented milk, milk, and lactose solution intake on plasmid transfer and the establishment of the resulting transconjugants in the digestive tracts of mice. It was proposed, in that study, that the lactose, available at the same concentration in the above supplementations, was the main agent responsible for the inhibiting effects (18).
In our attempt to understand the inhibitory effect observed in our study, we monitored the evolution of bacterial populations at the end of the experiments. In fact, bifidobacterial strains can exert inhibitory effects on the growth of other bacterial species (11). However, in our experiments the level of donor and recipient populations remained stable in vitro and in vivo. The decrease in transfer efficiency was not due therefore to an inhibitory impact of bifidobacterial strains on either the donor or the recipient strains. This inhibitory effect could be related to an impact of bifidobacteria on the conjugation mechanism itself, altering its efficiency. To investigate this hypothesis, in vitro experiments were performed with heat-killed bacteria, culture supernatant, and bacterial pellet. Since no inhibition was observed with the bacterial pellet, the decrease in the transfer was not due to a problem of steric hindrance affecting the contact between donor and recipient. The threefold decrease in transconjugant levels observed when the conjugation medium included either heat-killed bifidobacterial strains or bacterium-free supernatants suggests that the inhibitory action is more likely due to thermostable metabolites produced by bifidobacteria. Bacterial conjugation is a complex phenomenon that involves a mating-pair formation system belonging to the donor strain (27). The components of the mating-pair formation system, comprising a minimal set of 10 proteins, form a membrane-spanning protein complex and a surface-exposed sex pilus, which both serve to establish intimate physical contact with a recipient bacterium. Thus, these bifidobacterial metabolites might act on the conjugation process by affecting the tip of the donor pili or altering the recipient cell surface (8). Inhibition of conjugation was previously described with agents that affect the formation of sex pili or allow plasmid curing of the donor strains, such as sodium dodecyl sulfate, organic solvents, proteases, high temperature, and acridine dyes (32).
To conclude, little is known about the environmental factors capable of limiting the spread of resistance genes, whereas the evolution of bacteria toward resistance to antibiotics, and even multidrug resistance, has become a major cause of concern worldwide (1). We demonstrated here the antagonistic impact of bifidobacteria on the transfer of antibiotic resistance genes. Further investigations are in progress in order to identify the strains associated with the highest inhibitory impact and to elucidate the mechanism of this effect. Nevertheless, the use of probiotics might provide a new means to limit the dissemination of antibiotic genes.
C.M. was supported by a research grant from the Chancellerie des Universités de Paris.
Published ahead of print on 22 November 2006. ![]()
|
|
|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»