Embed Size (px)
Citation preview
A
ciggwdasaBBs©
K
1
tpia
0d
International Journal of Antimicrobial Agents 29 (2007) 271–280
Tetracycline susceptibility of the ingested Lactobacillus acidophilusLaCH-5 and Bifidobacterium animalis subsp. lactis Bb-12 strains
during antibiotic/probiotic intervention
Maria Saarela a,∗, Johanna Maukonen a, Atte von Wright b, Terttu Vilpponen-Salmela c,Andrea J. Patterson d, Karen P. Scott d, Heikki Hamynen e, Jaana Matto a
a VTT, Biotechnology, P.O. Box 1000, FI-02044 VTT, Finlandb University of Kuopio, Savilahdentie 9, 70211 Kuopio, Finland
c Harjula Hospital, Niuvantie 4, P.O. Box 38, 70101 Kuopio, Finlandd Rowett Research Institute, Bucksburn, Aberdeen AB21 9SB, UK
e Oma Laakari Ltd., Vuorikatu 20, 70100 Kuopio, Finland
Received 12 September 2006; accepted 26 September 2006
bstract
We investigated the effects of oral therapy with doxycycline, a tetracycline group antibiotic, on the gastrointestinal (GI) survival and tetracy-line susceptibility of probiotic strains Lactobacillus acidophilus LaCH-5 and Bifidobacterium animalis subsp. lactis Bb-12. In addition, thenfluence of doxycycline therapy on the diversity of the predominant faecal microbiota was evaluated by polymerase chain reaction—denaturingradient gel electrophoresis (PCR-DGGE). Faecal samples from the antibiotic group (receiving antibiotics and probiotics) and the controlroup (receiving probiotics only) were analysed for anaerobically and aerobically growing bacteria, bifidobacteria and lactic acid bacteria asell as for the dominant microbiota. Although doxycycline consumption did not have a large impact on GI survival of the probiotics, it had aetrimental effect on the bifidobacteria and on the diversity of the dominant faecal microbiota. A higher proportion of tetracycline-resistantnaerobically growing bacteria and bifidobacteria was detected in the antibiotic group than in the control group. Several antibiotic groupubjects had faecal B. animalis subsp. lactis Bb-12-like isolates with reduced tetracycline susceptibility. This was unlikely to be due to thecquisition of novel tetracycline resistance determinants, since only tet(W), which is also present in the ingested B. animalis subsp. lactisb-12, was found in the resistant isolates. Thus, concomitant ingestion of probiotic L. acidophilus LaCH-5 and B. animalis subsp. lactis
b-12 with the antibiotic did not generate a safety risk regarding the possible GI transfer of tetracycline resistance genes to the ingestedtrains.2006 Elsevier B.V. and the International Society of Chemotherapy. All rights reserved.
eywords: Tetracycline; tet(W); Doxycycline; Susceptibility; Lactobacillus acidophilus; Bifidobacterium animalis
ocpt
. Introduction
Since the disruptive effects of antimicrobial therapies onhe normal microbiota of the human gut are well recognised,
robiotics (live microorganisms), which when administeredn adequate amounts confer a health benefit to the host [1],re commonly used as a complementary therapy with antibi-∗ Corresponding author. Tel.: +358 20 722 4466; fax: +358 20 722 2103.E-mail address: [email protected] (M. Saarela).
[r
sip
924-8579/$ – see front matter © 2006 Elsevier B.V. and the International Societyoi:10.1016/j.ijantimicag.2006.09.020
tics to alleviate the possible gastrointestinal (GI) symptomsaused by the drugs [2,3]. Several studies have shown thatrobiotic Lactobacillus or Bifidobacterium strains can reducehe side effects of Helicobacter pylori eradication therapy4–8]. Probiotic lactobacilli have also proved effective ineducing antibiotic-associated diarrhoea in children [9,10].
The effect of probiotic Lactobacillus or Bifidobacteriumtrains on the composition of the intestinal microbiota dur-ng or following antibiotic therapy has been studied in severalapers [11–16]. The typical result was that probiotics could
of Chemotherapy. All rights reserved.
2 l of An
asbiortaqti1lHit
tGsBtae
2
2
Hcab(tsaoifs
2s
tdatvpcp
a(tBnsflwBn[apvfanii
t[Br(ttwBp
2
iosc1vsntwCsipioc
72 M. Saarela et al. / International Journa
ttenuate ecological disturbances caused by an antibiotic inome bacterial groups studied, but not in all. Although Lacto-acillus and Bifidobacterium (sub)populations were studiedn several of these papers, survival of the ingested probi-tic strains in the GI tract during antibiotic therapy has beenarely studied. In the study by Sullivan et al. [13], Lac-obacillus paracasei F19 was detected in low numbers inpproximately one-third of subjects receiving penicillin oruinolone therapy. Jernberg et al. [15] used terminal restric-ion fragment length polymorphism (T-RFLP) to track thengested probiotic strains (Lactobacillus acidophilus NCBF748, L. paracasei F19 and Bifidobacterium animalis subsp.actis Bb-12) during clindamycin therapy in four subjects.owever, since T-RFLP is a DNA-based detection method,
t cannot differentiate between living and dead cells and isherefore not applicable for studies on probiotic survival.
The aim of this study was to investigate the effects of oralherapy with doxycycline, a tetracycline group antibiotic, onI survival and tetracycline susceptibility of the probiotic
trains L. acidophilus LaCH-5 and B. animalis subsp. lactisb-12. In addition, the influence of tetracycline therapy on
he diversity of the predominant faecal microbiota was evalu-ted by polymerase chain reaction—denaturing gradient gellectrophoresis (PCR-DGGE).
. Materials and methods
.1. Probiotic preparation
A commercially available Trevis® capsule (produced byansen; marketed by Ipex Medical AB, Danderyd, Sweden)
ontaining probiotic strains L. acidophilus LaCH-5 and B.nimalis subsp. lactis Bb-12 and starter strains Lactobacillusulgaricus LbY-27 and Streptococcus thermophilus StY-31total number of bacteria, log 9–10 per capsule; manufac-urer’s information) was used in the intervention trial. Forpecies-specific enumeration, B. animalis subsp. lactis Bb-12nd L. acidophilus LaCH-5 were cultured from the capsulesn De Man, Rogosa, Sharpe (MRS) agar (Oxoid, Bas-ngstoke, UK), supplemented with cysteine–HCl (0.5 g/L)or bifidobacteria, and were identified by the RiboPrinter®
ystem (DuPont QualiconTM, Wilmington, DE) [17].
.2. Selective culturing and identification of probiotictrains from faecal samples
A pre-feeding trial with two volunteers was performedo screen for optimal sample processing and culturing con-itions and to ensure survival of the probiotic strains L.cidophilus LaCH-5 and B. animalis subsp. lactis Bb-12 inhe GI tract. The probiotic product was given to two healthy
olunteers for 3–4 days at the dose recommended on theroduct label (three capsules per day). Faecal samples wereollected before and after ingestion of the product. The sam-les were cultured immediately after collection, after storagebtwA
timicrobial Agents 29 (2007) 271–280
t −70 ◦C (1 week) or in Cary Blair transport medium [18]2 days) using several culture media and incubation condi-ions (data not shown). Transportation of the samples in Carylair medium (2 days storage) did not markedly reduce theumbers of studied bacterial groups. The following selectiveample processing and incubation conditions were chosenor the intervention study: for L. acidophilus, culture onactic acid bacteria (LAB)-selective Rogosa agar (Oxoid)ith microaerophilic incubation at 37 ◦C for 3 days; and for. animalis, acid pre-treatment of samples (human indige-ous bifidobacteria are more acid sensitive than B. animalis)19], culture on bifidobacteria-selective Beerens agar [20]nd anaerobic incubation at 37 ◦C for 4 days. The ingestedrobiotic strains were detected in faecal samples of botholunteers during the feeding period (log 6.3–6.5 colony-orming units (CFU)/g for B. animalis subsp. lactis Bb-12nd log 4.5–6.5 CFU/g for L. acidophilus LaCH-5), but wereot detected in the baseline samples. The high numbers ofndigenous bifidobacteria (ca. log 9 CFU/g) in faecal samplesndicated the necessity for using selective culturing methods.
Validation of the DNA fingerprinting method for iden-ification of the isolates was based on our previous studies17,21] and on the analysis of 10 Lactobacillus and 11ifidobacterium strains (representing four and five species,
espectively) by randomly amplified polymorphic DNARAPD) with six different primers (data not shown). Ofhe primers tested, OPA-03 (5′-AGTCAGCCAC-3’) provedo be the most discriminatory for L. acidophilus LaCH-5,hereas OPA-02 (5’-TGCCGAGCTG-3′) was best suited for. animalis subsp. lactis Bb-12. RAPD fingerprinting waserformed according to Matto et al. [17].
.3. Intervention study and sample handling
The antibiotic group consisted of 10 patients suffer-ng from respiratory tract infections who were recruitedn their first visit to their general practitioner (GP). Theseubjects (mean age 42 years) subsequently consumed doxy-ycline (Doxymycin®; Orion Pharma, Espoo, Finland; dose50 mg/day for 10 days) and probiotic Trevis® capsules. Tenolunteers (mean age 41 years) consuming probiotic cap-ules only formed the control group. Three capsules (totalumber of L. acidophilus log 9.6 and B. animalis subsp. lac-is log 9.0 in a daily dose) were consumed per day for 2eeks. The investigation was approved by the Human Ethicsommittees of University Hospital of Kuopio and VTT. All
ubjects gave their written informed consent for participationn the study. Faecal samples were collected at three samplingoints: t0, Day 0 for controls (before starting the probioticntervention) and Days 0–2 (most often Day 1) for subjectsn antibiotic therapy (the first faecal sample after the GP’sonsultation); and t1 and t2, 1 week and 2 week samples (for
oth antibiotic and control groups), respectively. An addi-ional sample (t3, 1 month after discontinuation of therapy)as obtained from four individuals in the antibiotic group.ntibiotic group subjects were asked to collect the first fae-
l of An
ctTpfwti
m1(bFMtp(HsasTeagtM(Polsfa
sfi(clssonsedb
2a
e
(((bswcwcpfies(adtaMas
2s
Bt2sww
2a
dagssab2sgbdo(
M. Saarela et al. / International Journa
al sample before antibiotic/probiotic consumption, but dueo ethical reasons no delay in starting therapy was requested.hus, mainly results of sampling point t1 are discussed in theresent study. However, data handling was additionally per-ormed by separating the antibiotic group subjects with andithout detectable B. animalis subsp. lactis at sampling point
0, and no difference between these subgroups was observedn any of the other studied bacterial groups (data not shown).
The samples were transferred to Cary–Blair transportedium, transported to the laboratory and analysed within
–2 days. The samples were cultured on Rogosa agarLAB), Beerens agar (bifidobacteria), supplemented Brucellalood agar (anaerobes) (Tammer Tutkan Maljat, Tampere,inland) and sheep blood agar (aerobes) (Tammer Tutkanaljat) plates with and without tetracycline supplementa-
ion (8 �g/mL). Rogosa, Beerens and Brucella blood agarlates were incubated in anaerobic jars (H2/CO2/N2 10:5:85)Anoxomat WS8000; Mart® Microbiology, Lichtenvoorde,olland) for 3, 4 and 7 days, respectively, at 37 ◦C, and
heep blood agar plates were incubated aerobically for 4 dayst 37 ◦C. Lactobacillus acidophilus LaCH-5 and B. animalisubsp. lactis Bb-12 were detected as described in Section 2.2.o confirm the specificity of the selective culture media, sev-ral tentative LAB (43 from the control group and 29 from thentibiotic group) and Bifidobacterium (32 from the controlroup and 27 from the antibiotic group) isolates were iden-ified to genus level by applying API50CHL (bioMerieux,
arcy l’Etoile, France) (LAB), partial 16S rDNA sequencingLAB and bifidobacteria) or bifidobacterial genus-specificCR. All tentative LAB isolates represented Lactobacillusr Pediococcus spp., and all tentative bifidobacteria iso-ated from the control group represented Bifidobacteriumpp. However, 44% of the tentative bifidobacterial isolatesrom the antibiotic group were not actually bifidobacteria,nd these were thus excluded from further analysis.
From each sample tentative LAB (294 isolates (5–39 perubject) from controls and 448 isolates (28–72 per subject)rom subjects on antibiotic therapy) and bifidobacteria (415solates (32–57 per subject) from controls and 463 isolates23–70 per subject) from subjects on antibiotic therapy) wereollected from selective plates for further analysis. The iso-ates were characterised by RAPD (see above) to reveal theimilarity with the consumed L. acidophilus and B. animalisubsp. lactis strains as well as to reveal the heterogeneityf indigenous LAB and Bifidobacterium populations. Whenecessary, a more discriminatory analysis of B. animalisubsp. lactis isolates was performed with pulsed-field gellectrophoresis (PFGE) and ribotyping [17]. In addition, theominant bacterial population in faecal samples was assessedy PCR-DGGE according to Matto et al. [22].
.4. Antibiotic susceptibility testing of L. acidophilus
nd B. animalis subsp. lactisSusceptibility of the strains/isolates to tetracycline wasvaluated using disk diffusion with 30 �g (Oxoid) and 80 �g
tu(t
timicrobial Agents 29 (2007) 271–280 273
Rosco, Taastrup, Denmark) tetracycline disks and by EtestBiodisk, Solna, Sweden) on LAB susceptibility test mediumLSM) (supplemented with 0.3 g/L cysteine (LSM + cys) forifidobacteria) [23]. Disk diffusion was used for preliminarycreening of susceptibility of the isolates, whilst the Etest,hich has previously been shown to be applicable for sus-
eptibility testing of bifidobacteria [24], was used for isolatesarranting determination of the actual minimum inhibitory
oncentration (MIC). The inoculum was prepared by sus-ending colonies from cultures grown on LSM(+cys) agaror 2 days to LSM(+cys) broth to a cell density correspond-ng to 1 McFarland standard. The suspension was spreadvenly on the pre-reduced agar plates using a sterile cottonwab. The disks or Etest strips were placed on the air-dried15–30 min) agar surface. The plates were incubated undernaerobic conditions at 37 ◦C for 48 h. The inhibition zoneiameters (disks) or the MICs (the lowest antibiotic concen-ration at which growth was inhibited) were read at the pointt which 80% inhibition of the growth occurred. TetracyclineICs of the control strains were 0.5 �g/mL for Trevis® L.
cidophilus LaCH-5 and 4–8 �g/mL for Trevis® B. animalisubsp. lactis Bb-12 (Bb-12 results are of four repeats).
.5. Persistence of tetracycline resistance in B. animalisubsp. lactis Bb-12-like isolates
To study the persistence of tetracycline resistance in. animalis subsp. lactis Bb-12-like faecal isolates, four
etracycline-resistant Bb-12-like isolates (initial MICs >256,56, 32 and 32 �g/mL) obtained from antibiotic groupubjects were continually subcultured in LSM + cys brothithout tetracycline for 4 weeks (five cycles during eacheek) and the MICs were determined by Etest as above.
.6. Identification of tetracycline resistance genes in B.nimalis subsp. lactis-like isolates
Three faecal B. animalis subsp. lactis-like isolates withifferent MIC values (32, 32 and >256 �g/mL) were char-cterised for the presence of tetracycline resistance (TcR)enes. The isolates were obtained from two antibiotic groupubjects either from sample t1 or t3. Trevis® B. animalisubsp. lactis Bb-12 with a MIC of 4–8 �g/mL was used ascontrol. Bacterial isolates were grown overnight in LSM
roth and 800 �L of the overnight culture was mixed with00 �L of glycerol. The glycerol stocks were frozen andhipped on dry ice to Rowett Research Institute for TcR
ene analysis. Bacterial stock cultures were concentratedy centrifugation, re-suspended in 1/5 volume of sterileH2O and boiled for 10 min to lyse cells, before spottingnto a Biodyne® B nylon membrane, 0.45 �m pore sizePierce, Rockford, IL). Approximately 0.4 �L of each cul-
ure was applied to the membrane by successive spottingsing the MicroGridII TAS robotic spotter from BioRoboticsBiorobotics Ltd., Haslingfield, UK). DNA was cross-linkedo air-dried membranes in a GS Gene Linker UV chamber
2 l of An
(MpLaSpttKrtGaptctoPa
2
ttPs
3
3it
astg22t
3da
bsbob
gbIfpalAngisctL
3B
2gRsot
3da
(mtte(Tcbc
sctwbItlo
74 M. Saarela et al. / International Journa
Bio-Rad, Hercules, CA) and stored at room temperature.embranes were hybridised using standard conditions to
robes consisting of radioactively labelled (Random Primedabelling Kit; Roche, Penzberg, Germany) PCR productsmplified from selected TcR genes ([25], Patterson andcott, unpublished data). These comprised nine ribosomerotection-type TcR genes (including tet(W), tet(M), tet(O),et(Q) and tet(32)), 12 efflux type genes and two inactivationype genes. Exposures with Bio-imaging plates (Fuji Film,anagawa, Japan) were captured using the FLA3000 fluo-
escence scanner (Fuji Film) and results were analysed withhe AIDA image analyser program (Raytest, Straubenhardt,ermany). The bacterial culture stocks (3 �L) were also
mplified directly using specific PCR primers to confirm theresence of TcR genes. The degenerate ribosome protectionype gene primers (tet1 and tet2; anneal 55 ◦C) [26], primersorresponding to the full-length tet(W) gene (tetWFF andetWFR; anneal 52.5 ◦C) [27] and an internal 1 kb fragmentf the gene (tetWfor-array and tetWrev-array; anneal 51 ◦C;atterson et al., unpublished data) were used in standard PCRmplifications.
.7. Statistical analysis
Statistical significance of the differences between the con-rol and antibiotic groups were analysed with the Students-test (Microsoft Excel 2002). A P-value <0.05 (presented as< 0.05, P < 0.01 or P < 0.001) was considered statistically
ignificant.
. Results
.1. Diversity and temporal stability of the predominantntestinal bacterial population during consumption ofhe probiotic product and the antibiotic
Each individual (both in control and antibiotic groups) hadunique and diverse dominant bacterial population in faecal
amples (Fig. 1). However, the intraindividual diversity ofhe dominant bacterial population was larger in the controlroup (mean number of bands in DGGE profile 35, range5–42) than in the antibiotic group (mean number of bands0, range 14–25): the Simpson index of diversity between thewo groups was 51%.
.2. Culturable bacterial numbers in faecal samplesuring consumption of the probiotic product and thentibiotic
The numbers of anaerobically and aerobically growingacteria, bifidobacteria and LAB that were culturable on the
pecific media used are shown in Fig. 2a. Whilst the types oflood agars used will not permit growth of all viable anaer-bic and aerobic faecal bacteria, valid comparisons can stille made between the samples. In the antibiotic and controlaopd
timicrobial Agents 29 (2007) 271–280
roup subjects, the total numbers of anaerobically and aero-ically growing bacteria were similar throughout the study.n both groups, anaerobes outnumbered aerobes 1000-fold inaecal samples at all sampling points (Fig. 2a) (e.g. at sam-ling point t1, anaerobes control group log 10.6 ± 0.1 CFU/g,ntibiotic group log 10.6 ± 0.4 CFU/g; aerobes control groupog 7.0 ± 1.1 CFU/g, antibiotic group log 6.7 ± 1.0 CFU/g).
statistically significant difference was seen in the totalumbers of bifidobacteria between the antibiotic and controlroup: the mean number of bifidobacteria was log 9.5 CFU/gn the faecal samples of control subjects during probiotic con-umption (sample t1), whereas in the antibiotic group theorresponding figure was log 7.7 CFU/g (P < 0.05). No sta-istically significant differences were seen in the numbers ofAB between the groups or sampling points (Fig. 2a).
.3. Heterogeneity of faecal Lactobacillus andifidobacterium populations
From each control subject, one to four RAPD types (mean.7) of LAB were detected in sample t1, whereas in antibioticroup subjects the corresponding figures were one to sevenAPD types (mean 4.1). Control subjects harboured three to
ix RAPD types (mean 4.5) of bifidobacteria, whereas antibi-tic group subjects had one to four bifidobacterial RAPDypes (mean 2.1) in sample t1.
.4. Tetracycline-resistant bacteria in faecal samplesuring consumption of the probiotic product andntibiotic
The breakpoint for intermediate resistance to tetracyclineTcR/I) is 8 �g/mL [28]. The TcR/I population comprised aean of 17% of the total anaerobically growing bacteria in
he t1 sample of control subjects. The corresponding figure forhe antibiotic group subjects was 70% (Fig. 2b). The differ-nce between antibiotic and control groups was significantP < 0.01) at all sampling points for anaerobically growingcR/I bacteria. In the control group, the TcR/I populationomprised a mean of 33% of the total aerobically growingacteria in sample t1, whereas in the antibiotic group theorresponding figure was 57% (Fig. 2b).
The proportion of TcR/I population on the bifidobacteria-elective medium (Beerens) was low (mean 1.4–1.6%) in theontrol group, whereas in the antibiotic group the propor-ion of TcR/I population on bifidobacteria-selective mediumas 52% (sample t1) during doxycycline therapy (P < 0.01etween antibiotic and control group samples) (Fig. 2b).n most of the samples, no growth was observed on theetracycline-containing LAB-selective medium (detectionimit log 3). When observed, the proportion of TcR/I LABf the total LAB population remained low (mean 8–14%)
nd their occurrence was sporadic (often in different subjectsn different sampling occasions). In the antibiotic group, theroportion of TcR/I LAB was higher than in the control groupuring all sampling occasions (mean 23–30%) and remained
M. Saarela et al. / International Journal of Antimicrobial Agents 29 (2007) 271–280 275
Fig. 1. Polymerase chain reaction—denaturing gradient gel electrophoresis (PCR-DGGE) profiles of the predominant bacterial populations in (a) two controlgroup subjects (subjects consuming probiotics) (Nos 16 and 18) and (b) three antibiotic group subjects (subjects consuming probiotics and doxycycline) (Nos5, 6 and 7). Faecal samples were obtained at three sampling points (sample t0, 0–1 days; sample t , 1 week; and sample t , 2 weeks). In subject 7, an additionals probio
fs
3s
afsBopbwcitTo4cafiiBdgtsboot
Raaasbnad
3
ctsfiitwbbMcmf
ww(
ample (t3) was obtained 1 month after discontinuation of the antibiotic and
airly constant. However, this difference was not statisticallyignificant.
.5. Specific identification of the ingested probiotictrains from faecal samples
Isolates with an identical RAPD type to the ingested B.nimalis subsp. lactis Bb-12 strain were recovered fromaecal samples of all control subjects during probiotic con-umption (sample t1; Fig. 2c). B. animalis subsp. lactisb-12 (mean count log 6.8 CFU/g) did not comprise partf the predominant bifidobacterial population (<1%) duringrobiotic consumption (sample t1) in the controls (total num-er of bifidobacteria log 9.5 CFU/g). However, this strainas frequently the only Bifidobacterium on tetracycline-
ontaining bifidobacteria-selective medium, suggesting thatt comprised the predominant part of the TcR/I bifidobac-erial population in the control subjects (data not shown).he B. animalis subsp. lactis Bb-12 strain in the probi-tic product is intermediately resistant to tetracycline (MIC–8 �g/mL), and growth of this strain on tetracycline-ontaining (8 �g/mL) agar was thus expected. All ninentibiotic group subjects (one subject dropped out after therst sampling occasion) had B. animalis subsp. lactis Bb-12
n the 1-week sample (t1) (Fig. 2c). B. animalis subsp. lactisb-12 represented 20% of the total bifidobacterial populationuring probiotic consumption (sample t1) in the antibioticroup (B. animalis subsp. lactis Bb-12 log 6.0 CFU/g vs.otal number of bifidobacteria log 7.7 CFU/g). B. animalisubsp. lactis Bb-12 predominated on tetracycline-containing
ifidobacteria-selective medium in sample t1 in only threef nine antibiotic group subjects, indicating the presence ofther tetracycline-resistant bifidobacteria besides Bb-12 inhese subjects (data not shown). L. acidophilus LaCH-5-likelt>c
1 2
tic therapy. M, marker.
APD types were detected in sample t1 in all control andntibiotic group subjects (log 5.6 CFU/g in the control groupnd log 5.3 CFU/g in the antibiotic group) (Fig. 2c). Bb-12-nd LaCH-5-like RAPD types were also detected in someubjects (both in the antibiotic group and in the control group)efore consumption of probiotic capsules (Fig. 2c and dataot shown). This finding is likely due to the common avail-bility and consumption of Bb-12- and LaCH-5-containingairy products in Finland [21].
.6. Antibiotic susceptibility of ingested probiotic strains
B. animalis subsp. lactis Bb-12-like isolates showinglearly smaller (>10 mm) inhibition zone diameters in theetracycline disk diffusion test compared with the controltrain (isolated from the probiotic capsule) were detected inve of nine antibiotic group subjects (55 of the 152 assayed
solates), whilst all assayed isolates (n = 124) from the 10 con-rol group subjects showed similar inhibition zone diametersith the control strain (Table 1). A subset of isolates fromoth groups was further assayed for tetracycline suscepti-ility by the Etest method. Isolates with higher tetracyclineICs (up to >256 �g/mL) compared with isolates from the
ontrol group (MICs 4–16 �g/mL) and the control B. ani-alis subsp. lactis Bb-12 strain (4–8 �g/mL) were detected
rom five antibiotic group subjects (Table 1).In the four antibiotic group subjects harbouring isolates
ith increased tetracycline resistance, tetracycline MICsere only two log2 values higher than in the control strain
32 �g/mL vs. 4–8 �g/mL), whilst in one subject even six
og2 value higher tetracycline MICs compared with the con-rol strain were detected at three sampling points (MICs up to256 �g/mL in samples t0 to t2) (Table 2). However, after dis-ontinuation of antibiotic and probiotic therapy (sample t3),
276 M. Saarela et al. / International Journal of Antimicrobial Agents 29 (2007) 271–280
Fig. 2. (a) Culturable numbers of anaerobically and aerobically growing bacteria, bifidobacteria and lactic acid bacteria (LAB) in faecal samples obtainedfrom the control group (subjects consuming probiotics) and antibiotic group (subjects consuming probiotics and doxycycline) on three sampling occasions. (b)Proportion of tetracycline resistant/intermediately resistant (TcR/I) anaerobically and aerobically growing bacteria, bifidobacteria and LAB in faecal sampleso of the i( obtained units.
tfTlst0b
3s
btained from control group and antibiotic group subjects. (c) DetectionBb-12) and Lactobacillus acidophilus LaCH-5 (LaCH-5) in faecal samplesifferences between the control and antibiotic groups. CFU, colony-forming
he B. animalis subsp. lactis Bb-12-like isolates recoveredrom this subject had tetracycline susceptibility ≤32 �g/mL.he similarity of the tetracycline-resistant B. animalis subsp.
actis Bb-12-like faecal isolates with the ingested B. animalis
ubsp. lactis Bb-12 strain was confirmed by RAPD with twoo five additional primers (OPA-03, OPA-05, OPA-13, OPL-1 and OPL-04), ribotyping with EcoRI (data not shown) andy PFGE with XbaI and SpeI (Fig. 3).otft
ngested probiotic strains of Bifidobacterium animalis subsp. lactis Bb-12d from control group and antibiotic group subjects. *Statistically significant
.7. Persistence of tetracycline resistance in B. animalisubsp. lactis Bb-12-like isolates
After 3–4 weeks subculturing (15–20 cycles) with-
ut antibiotic challenge, a 1–2 log2 value decrease inhe tetracycline MIC was observed in three of theour studied tetracycline-resistant B. animalis subsp. lac-is Bb-12-like isolates (MICs before >256, 256, 32
M. Saarela et al. / International Journal of Antimicrobial Agents 29 (2007) 271–280 277
Table 1Tetracycline susceptibility of faecal isolates with an identical randomly amplified polymorphic DNA (RAPD) type to the ingested Lactobacillus acidophilusLaCH-5 or Bifidobacterium animalis subsp. lactis Bb-12 strain, as determined by disk diffusion or the Etest method
Group L. acidophilus LaCH-5-like B. animalis subsp. lactis Bb-12-like
Disk diffusion Disk diffusion Etest
TcR (n) Total (n) TcR (n) Total (n) TcR (n) Total (n) MIC distribution (�g/mL)
4 8 16 32 64 128 ≥256
ControlIsolatesa 0 36 0 124 0 21 6 11 4 0 0 0 0Subjectsb 0 10 0 10 0 10 4 7 3 0 0 0 0
AntibioticIsolatesa 0 39 55 152 36 51 0 2 13 12 5 9 10Subjectsb 0 9 5 9 5 7 0 1 5 5 1 1 1
MIC, minimum inhibitory concentration.a TcR = number of isolates showing reduced tetracycline susceptibility compared with the control strain (L. acidophilus or B. animalis subsp. lactis isolated
from the probiotic capsules). Reduced susceptibility was defined as follows: disk diffusion (30 �g tetracycline), >10 mm smaller inhibition zone diameter thanthe control; Etest, MIC > 16 �g/mL.
b TcR = number of subjects harbouring isolates with reduced tetracycline susceptibility.
Table 2Detection of Bifidobacterium animalis subsp. lactis Bb-12-like isolates with reduced tetracycline susceptibility (TcR) in faecal samples of antibiotic groupsubjects on different sampling occasions (only those sampling points when TcR isolates were detected are shown)
Group/subject Tetracycline susceptibility
Disk diffusion Etest
TcR (n)a Total (n) TcR (n) a Total (n) MIC range of TcR
isolates (�g/mL)
Antibiotic (subject 7)Sample t0 10 10 5 5 64–256Sample t1 10 18 8 9 16 to >256Sample t2 12 19 10 10 64 to >256Sample t3 6 10 6 10 16–32
Antibiotic (subject 3)Sample t1 2 14 2 5 16–32
Antibiotic (subject 6)Sample t2 4 9 3 3 32
Antibiotic (subject 9)Sample t1 7 10 1 4 16–32
Antibiotic (subject 12)
Montrol s
ar
3l
catalaa
aad[
4
t
Sample t1 4 9
IC, minimum inhibitory concentration.a Number of isolates showing reduced susceptibility compared with the c
nd 32 �g/mL; MICs after 64, 128, 8 and 32 �g/mL,espectively).
.8. Identification of TcR genes in B. animalis subsp.actis Bb-12-like isolates
All three TcR B. animalis subsp. lactis Bb-12-like fae-al isolates (regardless of the MIC for tetracycline) as wells Trevis® B. animalis subsp. lactis Bb-12 hybridised onlyo the tet(W) probe. None of the other probes hybridised
bove background. The presence of tet(W) in all four iso-ates was confirmed by PCR amplification and sequencing ofn internal 1 kb product. The full-length product could not bemplified using the primer combination tetWFF and tetWFR,fBtt
3 3 32
train (see Table 1).
nd selective primer combinations revealed that the sequencet the 3’ end of the tet(W) gene in these bifidobacterial isolatesiffered from the original Butyrivibrio fibrisolvens isolate27].
. Discussion
In this study, we have demonstrated that oral tetracyclineherapy results in increased tetracycline resistance among
aecal anaerobic bacteria, including the ingested probiotic. animalis subsp. lactis population. To our knowledge,his is the first paper investigating the effect of antibioticherapy on the antibiotic susceptibility of simultaneously
278 M. Saarela et al. / International Journal of An
Fig. 3. Pulsed-field gel electrophoresis (PFGE) with XbaI and SpeI pro-files of Bifidobacterium animalis subsp. lactis: lane 1, control isolated fromprobiotic capsules; lanes 2–5, isolates from antibiotic group subject 7 (sam-ples t0 to t3; tetracycline minimum inhibitory concentrations (MICs) 256,>256, 256 and 32 �g/mL, respectively); lane 6, isolate from antibiotic groupsar
iG
aibdhtoGaicsU1(oBi6ett((etc
itb
dbccgtogtatf
wmddopirgcow(
tsmons(daasv
cDscg(wcwoso
ubject 3 (sample t1; tetracycline MIC 32 �g/mL); and lane 7, isolate fromntibiotic group subject 12 (sample t1; tetracycline MIC 32 �g/mL). M, lowange PFGE marker (New England Biolabs, Ipswich, MA).
ngested probiotic strains and on their survival through theI tract.Although there are several studies where the effect of
ntibiotic therapy combined with probiotic ingestion on thentestinal microbiota or the side effects of the therapy haveeen investigated [4–16,29], the fate of the probiotic strainuring GI transit in an antibiotic-containing environmentas attracted little interest. Our results show that even ifhe ingested probiotic strain is susceptible to the antibi-tic in question (in our case L. acidophilus), it can surviveI transit quite well; during doxycycline consumption, L.cidophilus LaCH-5-like isolates were detected at approx-mately equal levels (log 5.3–5.6 CFU/g) in antibiotic andontrol group subjects. In the case of the ingested B. animalisubsp. lactis strain, the situation was somewhat different.nlike L. acidophilus LaCH-5, B. animalis subsp. lactis Bb-2 is intermediate with regard to tetracycline susceptibilityMIC 4–8 �g/mL vs. 0.5 �g/mL). However, this fact hadnly a small effect on the GI survival of the strain, sinceb-12 was also detected in approximately equal numbers
n both groups (log 6.8 CFU/g in the control group and log.0 CFU/g in the antibiotic group). In the study of Sullivant al. [13], three of the ten subjects on penicillin therapy andhree of the eight subjects on quinolone therapy harbouredhe ingested probiotic strain, L. paracasei F19, at low levelslog 3.3–5.2 CFU/g) after 10 days of simultaneous ingestionprobiotic daily dose log 10.3 CFU). As the study of Sullivant al. [13] did not contain a group consuming probiotics only,he effect of antibiotics on the L. paracasei F19 GI survivalannot be evaluated.
Although doxycycline consumption did not have a largempact on the survival of the two probiotic strains throughhe GI tract or on the total number of anaerobically growingacteria cultured, it did have a detrimental effect on the bifi-
wipd
timicrobial Agents 29 (2007) 271–280
obacterial population in the present study. The numbers ofifidobacteria at sampling point t1 (after 1 week of antibioticonsumption) were markedly lower in the antibiotic groupompared with the control group. Furthermore, antibioticroup subjects had a less diverse Bifidobacterium populationhan the controls (as evidenced by detecting a lower numberf bifidobacterial RAPD types per subject in the antibioticroup). Human intestinal bifidobacteria are generally suscep-ible to tetracycline and show median MICs ≤2 �g/mL [30],lthough in recent studies tetracycline-resistant Bifidobac-erium strains have been reported with prevalences varyingrom 14% [30] to 30% [31].
The effects of doxycycline therapy on the LAB populationere detected as an increased intraindividual diversity (deter-ined by RAPD) but not in actual LAB numbers. Increased
iversity probably reflects altered population dynamics: if theominant LAB strains/species are suppressed by the antibi-tic, the likelihood of detecting less susceptible and lessrevalent strains/species increases. Tetracycline susceptibil-ty in lactobacilli varies between different species. Strainsepresenting the Lactobacillus casei gr. and the L. acidophilusr. species typically show MICs ≤2 �g/mL, although as in thease of bifidobacteria, occasional atypically resistant strainsccur in lactobacilli [32]. Pediococci, which in this studyere mainly detected on tetracycline-supplemented medium
data not shown), are commonly resistant to tetracycline [33].The effects of doxycycline therapy could also be seen in
he dominant microbiota, as determined by DGGE. DGGEhowed that doxycycline therapy simplified the dominanticrobiota. To the best of our knowledge, there are only two
ther studies where the effects of antibiotics on the domi-ant faecal bacteria have been studied by DGGE (or the veryimilar technique, temperature gradient gel electrophoresisTGGE)). Donskey et al. [34] detected differences in theominant faecal bacterial DGGE profiles of four patientsfter some antibiotic therapies in some subjects but not inll, whilst in the study of De La Cochetiere et al. [35] a majorhift in the faecal dominant microbiota was seen in six healthyolunteers owing to amoxicillin therapy.
Doxycycline therapy had a profound effect on the fae-al B. animalis subsp. lactis Bb-12-like bacterial population.uring antibiotic therapy, five of the nine antibiotic group
ubjects had isolates with clearly reduced tetracycline sus-eptibility (MICs up to >256 �g/mL), whereas in the controlroup reduced tetracycline susceptibility was not observedMICs ≤16 �g/mL). The reduced tetracycline susceptibilityas probably not due to the acquisition of novel tetracy-
line resistance determinants during antibiotic therapy since,hen three of the less susceptible isolates were studied,nly tet(W), which is also present in Trevis® B. animalisubsp. lactis Bb-12 strain, was found. tet(W) has previ-usly been reported in B. animalis subsp. lactis [36] as
ell as in several human intestinal Bifidobacterium speciesncluding some Bifidobacterium longum, Bifidobacteriumseudocatenulatum, Bifidobacterium adolescentis and Bifi-obacterium bifidum strains [27,30,36]. When selected
l of An
rsdtwt
coiobdatoLdBa
A
tRRttnwMf
R
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
M. Saarela et al. / International Journa
esistant B. animalis subsp. lactis Bb-12-like isolates wereubcultured without antibiotic selection, tetracycline MICsecreased but not to the original level. However, without fur-her studies it is impossible to say whether this phenomenonas due to changes that had occurred in the tet(W) gene or
o its regulation in these isolates during doxycycline therapy.In conclusion, the present study showed that doxycy-
line therapy markedly reduced the numbers and diversityf faecal bifidobacteria, diminished the diversity of dom-nant faecal microbiota and increased the proportionf tetracycline-resistant culturable anaerobically growingacteria. Concomitant ingestion of probiotic bacteria L. aci-ophilus LaCH-5 and B. animalis subsp. lactis Bb-12 with thentibiotic did not generate a safety risk regarding the possibleransfer of tetracycline resistance genes to the ingested strainsver the time period of this study. Lactobacillus acidophilusaCH-5, which is susceptible to tetracycline, remained souring antibiotic therapy, and in B. animalis subsp. lactisb-12 (which is TcI owing to the presence of tet(W)) nodditional tetracycline resistance genes were detected.
cknowledgments
This study was carried out with financial support fromhe Commission of the European Communities, specificTD program ‘Quality of Life and Management of Livingesources’, QLK2-CT-2002-00843, ‘Antimicrobial resis-
ance transfer from and between gram-positive bacteria ofhe digestive tract and consequences for virulence’. It doesot necessarily reflect the views of the Commission and in noay anticipates the Commission’s future policy in this area.s Marja-Liisa Jalovaara and Ms Niina Torttila are thanked
or skilful technical assistance.
eferences
[1] Food and Agriculture Organization of the United Nations/World HealthOrganization. Guidelines for the evaluation of probiotics in food.Report of a Joint FAO/WHO Working Group on Drafting Guidelinesfor the Evaluation of Probiotics in Food. London, Ontario, Canada:FAO/WHO; 2002.
[2] Saavedra JM. Probiotics plus antibiotics: regulating our bacterial envi-ronment. J Pediatr 1999;135:564–8.
[3] Sullivan A, Nord CE. Probiotics and gastrointestinal diseases. J InternMed 2005;257:78–92.
[4] Armuzzi A, Cremonini F, Bartolozzi F, et al. The effect of oral admin-istration of Lactobacillus GG on antibiotic-associated gastrointestinalside-effects during Helicobacter pylori eradication therapy. AlimentPharmacol Ther 2001;15:163–9.
[5] Armuzzi A, Cremonini F, Ojetti V, et al. Effect of Lactobacillus GGsupplementation on antibiotic-associated gastrointestinal side effectsduring Helicobacter pylori eradication therapy: a pilot study. Digestion
2001;63:1–7.[6] Cremonini F, Di Caro S, Covino M, et al. Effect of different probioticpreparations on anti-Helicobacter pylori therapy-related side effects: aparallel group, triple blind, placebo-controlled study. Am J Gastroen-terol 2002;97:2744–9.
[
[
timicrobial Agents 29 (2007) 271–280 279
[7] Sheu B-S, Wu JJ, Lo C-Y, et al. Impact of supplement withLactobacillus- and Bifidobacterium-containing yogurt on triple ther-apy for Helicobacter pylori eradication. Aliment Pharmacol Ther2002;16:1669–75.
[8] Tursi A, Brandimarte G, Giorgetti GM, Modeo ME. Effect of Lac-tobacillus casei supplementation on the effectiveness and tolerabilityof a new second-line 10-day quadruple therapy after failure of afirst attempt to cure Helicobacter pylori infection. Med Sci Monit2004;10:CR662–6.
[9] Arvola T, Laiho K, Torkkeli S, et al. Prophylactic LactobacillusGG reduces antibiotic-associated diarrhea in children with respiratoryinfections: a randomized study. Pediatrics 1999;104:e64.
10] Vanderhoof JA, Whitney DB, Antonson DL, Hanner TL, Lupo JV,Young RJ. Lactobacillus GG in the prevention of antibiotic-associateddiarrhea in children. J Pediatr 1999;135:564–8.
11] Lidbeck A, Edlund C, Gustafsson J-A, Kager L, Nord CE. Impactof Lactobacillus acidophilus on the normal intestinal microflora afteradministration of two antimicrobial agents. Infection 1988;16:329–36.
12] Sullivan A, Barkholt L, Nord CE. Lactobacillus acidophilus, Bifi-dobacterium lactis and Lactobacillus F19 prevent antibiotic-associatedecological disturbances of Bacteroides fragilis in the intestine. JAntimicrob Chemother 2003;52:308–11.
13] Sullivan A, Johansson A, Svenungsson B, Nord CE. Effect of Lacto-bacillus F19 on the emergence of antibiotic-resistant microorganismsin the intestinal microflora. J Antimicrob Chemother 2004;54:791–7.
14] Madden JAJ, Plummer SF, Tang J, et al. Effect of probiotics on prevent-ing disruption of the intestinal microflora following antibiotic therapy:a double-blind, placebo-controlled pilot study. Int Immunopharmacol2005;5:1091–7.
15] Jernberg C, Sullivan A, Edlund C, Lansson JK. Monitoring ofantibiotic-induced alterations in the human intestinal microflora anddetection of probiotic strains by use of terminal restriction fragmentlength polymorphism. Appl Environ Microbiol 2005;71:501–6.
16] Plummer SF, Garaiova I, Sarvotham T, et al. Effects of probiotics on thecomposition of the intestinal microbiota following antibiotic therapy.Int J Antimicrob Agents 2005;26:69–74.
17] Matto J, Malinen E, Suihko M-L, Alander M, Palva A, Saarela M.Genetic heterogeneity and functional properties of intestinal bifidobac-teria. J Appl Microbiol 2004;97:459–70.
18] Atlas RM. In: Parks LC, editor. Handbook of microbiological media.2nd ed. Boca Raton, FL: CRC Press; 1997. p. 253.
19] Alander M, Matto J, Kneifel W, et al. Effect of galacto-oligosaccharidesupplementation on human fecal microflora and on survival and persis-tence of Bifidobacterium lactis Bb-12 in the gastrointestinal tract. IntDairy J 2001;11:817–25.
20] Beerens H. Detection of bifidobacteria by using propionic acid as aselective agent. Appl Environ Microbiol 1991;57:2418–9.
21] Matto J, Fonden R, Tolvanen T, et al. Intestinal survival and persistenceof probiotic Lactobacillus and Bifidobacterium strains administered intriple-strain yoghurt. Int Dairy J 2006;16:1174–80.
22] Matto J, Maunuksela L, Kajander K, et al. Composition and temporalstability of gastrointestinal microbiota in irritable bowel syndrome—alongitudinal study in IBS and control subjects. FEMS Immunol MedMicrobiol 2005;43:213–22.
23] Klare I, Konstabel C, Muller-Bertling S, et al. Evaluation of new brothmedia for microdilution antibiotic susceptibility testing of lactobacilli,pediococci, lactococci, and bifidobacteria. Appl Environ Microbiol2005;71:8982–6.
24] Matto J, Suihko M-L, Saarela M. Comparison of three test media forthe antimicrobial susceptibility testing of bifidobacteria with the Etestmethod. Int J Antimicrob Agents 2006;28:42–8.
25] Sambrook J, Fritsch EF, Maniatis T. Molecular cloning: a laboratorymanual. 2nd ed. Cold Spring Harbor, NY: Cold Spring Harbor Labora-tory Press; 1989.
26] Barbosa TM, Scott KP, Flint HJ. Evidence for recent intergenerictransfer of a new tetracycline resistance gene, tet(W), isolated from
2 l of An
[
[
[
[
[
[
[
[
[
short-course antibiotic challenge. J Clin Microbiol 2005;43:5588–
80 M. Saarela et al. / International Journa
Butyrivibrio fibrisolvens, and the occurrence of tet(O) in ruminal bac-teria. Environ Microbiol 1999;1:53–64.
27] Scott K, Melville CM, Barbosa TM, Flint HJ. Occurrence of the newtetracycline resistance gene tet(W) in bacteria from the human gut.Antimicrob Agents Chemother 2000;44:775–7.
28] National Committee for Clinical Laboratory Standards. Methods forantimicrobial susceptibility testing of anaerobic bacteria. Approvedstandard. 6th ed. M11-A6. Wayne, PA: NCCLS; 2004.
29] Wullt M, Johansson Hagslatt M-L, Odenholt I. Lactobacillus plantarum299v for the treatment of recurrent Clostridium difficile-associateddiarrhea: a double-blind, placebo-controlled trial. Scand J Infect Dis2003;35:365–7.
30] Moubareck C, Gavini F, Vaugien L, Butel MJ, Doucet-Populaire F.
Antimicrobial susceptibility of bifidobacteria. J Antimicrob Chemother2005;55:38–44.31] Delgado S, Florez AB, Mayo B. Antibiotic susceptibility of Lactobacil-lus and Bifidobacterium species from the human gastrointestinal tract.Curr Microbiol 2005;50:202–7.
[
timicrobial Agents 29 (2007) 271–280
32] Danielsen M, Wind A. Susceptibility of Lactobacillus spp. to antimi-crobial agents. Int J Food Microbiol 2003;82:1–11.
33] Tankovic J, Leclerco R, Duval J. Antimicrobial susceptibility of Pedio-coccus spp. and genetic basis of macrolide resistance in Pediococcusacidilactici HM3020. Antimicrob Agents Chemother 1993;37:789–92.
34] Donskey CJ, Hujer AM, Das SM, Pultz NJ, Bonomo RA, RiceLB. Use of denaturing gradient gel electrophoresis for analysis ofthe stool microbiota of hospitalized patients. J Microbiol Methods2003;54:249–56.
35] De La Cochetiere MF, Durand T, Lepage P, Bourreille A, GalmicheJP, Dore J. Resilience of the dominant human fecal microbiota upon
92.36] Masco L, Van Hoorde K, De Brandt E, Swings J, Huys G. Antimicro-
bial susceptibility of Bifidobacterium strains from humans, animals andprobiotic products. J Antimicrob Chemother 2006;58:85–94.
