in situ esbl conjugation from avian to human e. coli ... · 1 in situ esbl conjugation from avian...
TRANSCRIPT
1
In situ ESBL conjugation from avian to human E. coli during cefotaxime 1
administration 2
Annemieke Smet,1,2*
Geertrui Rasschaert,2 An Martel,
1 Davy Persoons,
2,3 Jeroen Dewulf,
3 Patrick 3
Butaye,1,4
Boudewijn Catry,5 Freddy Haesebrouck,
1 Lieve Herman,
2 and Marc Heyndrickx
2 4
5
Department of Pathology, Bacteriology and Avian Diseases, Faculty of Veterinary Medicine, 6
Ghent University, Salisburylaan 133, 9820 Merelbeke, Belgium1;
7
Institute for Agricultural and Fisheries Research, Unit Technology and Food, Brusselsesteenweg 8
370, B-9090 Melle, Belgium2; 9
Department of Reproduction, Obstetrics and Herd Health, Veterinary Epidemiology Unit, 10
Faculty of Veterinary Medicine, Ghent University, Salisburylaan 133, B-9820 Merelbeke, 11
Belgium3; 12
Department of Bacteriology and Immunology, CODA-CERVA-VAR, Groeselenberg 99, 1180 13
Brussels, Belgium4 14
Public Health & Surveillance, Scientific Institute of Public Health, J. Wytsmanstraat 14 1050 15
Brussels, Belgium5
16
17
* Corresponding author. Mailing address: Department of Pathology, Bacteriology and Avian 18
Diseases, Faculty of Veterinary Medicine, Ghent University, Salisburylaan 133, 9820 Merelbeke, 19
Belgium. Phone: 003292647435. Fax: 003292647494. 20
E-mail: [email protected] 21
Keywords: Keywords: conjugation, extended-spectrum β-lactams, transfer frequency, continuous 22
flow model 23
2
Abstract 24
Aims: The behaviour of an E. coli isolate of broiler origin harbouring a blaTEM-52-carrying 25
plasmid (lactose negative mutant of B1-54, IncII group) was studied in an in situ continuous flow 26
culture system, simulating the human caecum and the ascending colon during cefotaxime 27
administration. 28
Methods and Results: Fresh faeces from a healthy volunteer, negative for cephalosporin-29
resistant E. coli, were selected to prepare inocula. The microbiota was monitored by plating on 30
diverse selective media, and a shift in the populations of bacteria was examined by 16S rDNA 31
PCR denaturing gradient gel electrophoresis. E. coli transconjugants were verified by plasmid 32
and pulsed field gel electrophoresis profiles (PFGE). The avian extended-spectrum β-lactamase 33
(ESBL) positive E. coli was able to proliferate without selective pressure of cefotaxime and E. 34
coli transconjugants of human origin were detected 24 h after inoculation of the donor strain. 35
Upon administration of cefotaxime to the fresh medium, an increase of the population size of E. 36
coli B1-54 and the transconjugants was observed. PFGE and plasmid analysis revealed a limited 37
number of human E. coli clones receptive for the blaTEM-52-carrying plasmid. 38
Conclusions: These observations provide evidence of the maintenance of an E. coli strain of 39
poultry origin and the horizontal gene transfer in the human commensal bowel microbiota even 40
without antimicrobial treatment. 41
Significance and Impact of Study: The fact that an E. coli strain of poultry origin might 42
establish itself and transfer its bla gene to commensal human E. coli raises public health 43
concerns. 44
45
46
3
Introduction 47
Escherichia coli, the best studied member of the bacterial family of Enterobacteriaceae, is a 48
Gram-negative rod-shaped bacterium that mostly occurs as a normal commensal in the intestinal 49
tract of animals and humans. Some strains are, however, important intestinal and extraintestinal 50
pathogens. Human and animal pathogenic E. coli are able to cause a spectrum of illnesses ranging 51
from self-limiting gastrointestinal infections to life threatening bacteraemia. 52
In human and veterinary medicine, β-lactams are extensively used. The predominant cause of 53
resistance to β-lactam antibiotics in Gram-negative bacteria is the production of β-lactamases. 54
These enzymes can be located on the chromosome or on plasmids (Sheldon 2005). 55
Extended-spectrum β-lactamases (ESBLs) (e.g. TEM-52), capable of efficiently hydrolyzing 56
extended-spectrum cephalosporins and associated with several treatment failures, disseminate 57
through plasmid transmission among the family of the Enterobacteriaceae. These bacteria are 58
part of the total microbiota of the digestive tract (Shah et al. 2004; Sheldon 2005; Smet et al. 59
2009). 60
The presence of cephalosporin-resistant E. coli in the intestinal tract of food-producing animals is 61
extensively described in many reports (Cloeckaert et al. 2007; Hasman et al. 2005; Kojima et al. 62
2005; Liu et al. 2007; Riaño et al. 2006). There is recently a high concern among clinical 63
microbiologists about ESBL-producing E. coli on poultry meat as a result of monitoring reports 64
in the EU, e.g. in the Netherlands and Belgium, showing high prevalence levels (Kojima et al. 65
2005; Liu et al. 2007; Riaño et al. 2006; Smet et al. 2009). The diversity of ESBLs (e.g TEM-52, 66
CTX-M-1, CTX-M-2) among these bacteria is high and they may act as a reservoir of ESBL 67
genes for pathogens causing disease in humans. This raises a potential public health concern 68
(Smet et al. 2008; Vollaard and Clasener 1994). The human digestive tract, colonized by a 69
4
complex microbiota, has been proposed as a suitable environment for horizontal transfer of 70
plasmids carrying antimicrobial resistance genes and contributes to the maintaining and the 71
dissemination of resistance (Netherwood et al. 1997). In the case of ESBL genes, it is not exactly 72
known how easy humans become carriers as the result of the consumption of poultry meat. The 73
determination of the transfer frequency between E. coli populations of human and poultry origin 74
is a first step in this study. 75
In vitro transfer of β-lactamase encoding (bla) genes is relatively easily to demonstrate between 76
individual strains and has been extensively studied (Smet et al. 2009). In vivo transfer of a bla 77
gene in the intestine of mice has been shown to occur (Moubareck et al. 2003; Schjorring et al. 78
2008) as also indications of transfer of bla genes between commensal members of 79
Enterobacteriaceae of human origin in clinical cases due to the treatment with a β-lactam 80
antimicrobial agent (Bidet et al. 2005; Karami et al. 2007). 81
More information is however needed on the transfer of specific antibiotic resistance genes from 82
animal or food derived strains to human strains in the gut. A recent controlled experiment where 83
healthy humans ingested a sulphonamide-susceptible recipient strain of human origin and a 84
sulphonamide-resistant donor strain of animal origin, showed the transfer of this sul gene to E. 85
coli of human origin in the gut. It has to be noted that the volunteer where this transfer took place, 86
was already colonized with commensal sulphonamide-resistant E. coli before the start of the 87
experiment. The reason that these volunteers also ingested an E. coli strain of animal origin was 88
to shed light on the risk associated with antimicrobial drug resistant E. coli in food. However, 89
evidence of transfer of resistance genes from an animal-derived E. coli to a recipient of human 90
origin was not provided in the human intestinal tract (Trobos et al. 2009). Another study reported 91
the transfer of vancomycin resistance from an E. faecium isolate of animal origin to an E. faecium 92
isolate from human origin in the intestines of humans (Lester et al. 2006). 93
5
These human controlled experiments requiring volunteers are possibly not without risk and 94
usually do not investigate the behaviour of the antibiotic resistant strains (donors and 95
transconjugants) with and without antibiotic treatment. Ethical and health constraints however 96
hamper experimental studies in humans. Therefore, the objective of this study was to investigate 97
whether an E. coli isolate of broiler origin could transfer its blaTEM-52-carrying plasmid to E. coli 98
of human origin in an in situ continuous flow culture system, simulating the human intestinal 99
tract in different situations (disturbed or stabilized gut microbiota) and this to highlight the risk 100
for public health associated with ESBL-producing E. coli in food. The in situ continuous flow 101
culture system underwent different culture scenarios that simultaneously allowed assessing the 102
impact of a 4 day course of a third-generation cephalosporine (cefotaxime). This study also 103
allowed us to measure to what extent the cephalosporin-resistant E. coli strains, both the donor 104
from broiler origin and the human transconjugants, could colonize in this fermentation system in 105
different situations (antibiotic treatment or not). Such data about transfer frequency and 106
colonization rate are necessary to perform risk assessment of human intake of food containing 107
antimicrobial drug resistant bacteria. 108
109
Materials and Methods 110
Strain preparation 111
E. coli strain B1-54, isolated from a faecal sample of broilers during an earlier survey (Smet et al. 112
2009) was used as donor, after selection of a lactose negative mutant to allow accurate 113
enumeration of transconjugants. The B1-54 isolate carried a blaTEM-52-carrying plasmid that 114
belonged to the incompatibility group IncI1 and had a size of > 100 kb. B1-54 also showed 115
phenotypic resistance to nalidixic acid and tetracycline but these resistance genes are not located 116
6
on the blaTEM-52-carrying plasmid (Smet et al. 2009). The methodology was slightly modified 117
from the one described by Varela and colleagues (1997). Briefly, mutants were isolated on 118
minimal succinate (0.2%) plates containing 3mM β-thio-o-nitrophenyl-β-D-galactopyranoside 119
(TONPG) (Sigma, Bornem, Belgium) plus 0.5mM isopropyl-β-D-thio-galactopyranoside (IPTG) 120
(Sigma, Bornem, Belgium) and incubated at 37°C. Mutants which grew on these plates were 121
streaked on 1% (w/v) melibiose MacConkey agar plates (plus 0.5mM IPTG). Red colonies were 122
picked and inoculated on 1% (w/v) lactose MacConkey agar plates. One white colony was 123
selected for this study, and this lactose negative mutant of E. coli B1-54 was grown overnight on 124
a MacConkey agar plate supplemented with 2.5 mg/l cefotaxime. 125
The presence of red colonies (lactose positive) on the MacConkey agar plates (Oxoid LTD, 126
Basingstoke, Hampshire, England) (with 2.5 mg/ ml cefotaxime) refers to transconjugants (of 127
human origin), while white colonies (lactose negative) refers to the lactose mutant of E. coli B1-128
54. 129
130
Fermentation system 131
Growth conditions for the caecum and ascending colon were mimicked in an in situ system, 132
which operated as a continuous system. The growth medium for the human caecum and the 133
ascending colon was prepared as described by MacFarlane et al. (2005). One 1.3 l glass 134
fermentation vessel, operated under the control of a BioFlo 110 unit (New Brunswick Scientific, 135
Edison, New Jersey, USA), was used in this study to simulate the microbial growth conditions of 136
the caecum and ascending colon. Via a peristaltic pump, fresh sterilized medium was added at a 137
constant rate of about 1.8 ml/min and spent culture liquid was wasted at the same rate to maintain 138
a constant working volume of 500 ml. Culture pH was maintained at 6.2, the temperature at 37°C 139
and moderate agitation at 200 rpm. Cultures were sparged by flushing the headspace of the vessel 140
7
with a nitrogen carbon dioxide gas mixture (80 and 20%, respectively) at a flow rate of 20 ml/min 141
to create anaerobic conditions. Fresh faeces from a healthy volunteer (adult), negative for 142
cephalosporin-resistant E. coli, were used to prepare inocula. The stools were homogenized with 143
a kitchen blender, divided in aliquots of 30 ml, supplemented with 15% glycerol and frozen at -144
80°C. After inoculation with 5 ml pooled stool content (corresponding to 1g feces), the fermentor 145
was operated in batch mode for 24 h. Continuous culture was started by switching on the 146
peristaltic pumps (Messens et al. 2009). The lactose mutant of E. coli B1-54 was grown 147
overnight on a MacConkey agar plate (supplemented with 2.5 mg/l cefotaxime). Two loopfull of 148
colonies were suspended in 10 ml PBS up to a final concentration of approximately 2 x 108 149
CFU/ml and added to the fermentation system at the indicated times (Fig. 1-3). Immediately after 150
inoculation of the lactose mutant of E. coli B1-54 (the donor strain), a sample was taken in order 151
to calculate the start number of the donor strain in the fermentation system. 152
153
Study design 154
Three experiments were carried out and in all three cases cefoxatime was supplied during 4 155
consecutive days. In a first negative control experiment (I) no ESBL+ donor strain (lactose 156
mutant of E. coli B1-54) was added in order to evaluate the stability of the model. 157
A second experiment (II) consisted of adding the ESBL+ donor strain simultaneously with the 158
intenstinal content at the iniation of the batch mode, 24 h prior to the peristaltic simulation. 159
During a final third experiment (III) the ESBL+donor strain was added 24 hours after the 160
peristaltic simulation had started. 161
162
163
164
8
Bacterial population dynamics 165
The cultivatable microbiota was monitored and counted by dilution plating on diverse selective 166
media (Oxoid LTD, Basingstoke, Hampshire, England): Reinforced Clostrial Medium (RCM) for 167
total anaerobes (37°C, 24h, anaerobic conditions); de Man, Rogosa and Sharpe medium (MRS) 168
for lactic acid bacteria (37°C, 24h, anaerobic conditions); Tryptone Soy Agar (TSA) for total 169
aerobes (37°C, 24h, aerobic conditions); and MacConkey Agar (with and without 2.5 mg/l 170
cefotaxime) for lactose negative donor strains of E. coli (of poultry origin, white colonies), 171
population of potential recipient E. coli (of human origin and cefotaxime susceptible, pink 172
colonies) and E. coli transconjugants (of human origin and cefotaxime resistant, pink colonies) 173
(37°C, 24h, aerobic conditions). 174
A shift in bacterial populations, due to a continuous administration of cefotaxime (2.5mg/l) 175
during 4 consecutive days, was examined by 16S (V3 region) rDNA PCR denaturing gradient gel 176
electrophoresis (PCR-DGGE), as described previously (Boon et al. 2000). 177
178
Transconjugant identification and horizontal gene transfer confirmation 179
A representative number of pink colonies were phenotypically identified as E. coli (Smet et al. 180
2008) after overnight aerobic incubation at 37°C on MacConkey agar with 2.5 mg/l cefotaxime. 181
The selected E. coli transconjugants were genotyped using XbaI-pulsed field gel electrophoresis 182
(PFGE) (Bertrand et al. 2006). In order to demonstrate whether these transconjugants obtained 183
the blaTEM-52-carrying plasmid, plasmid DNA was obtained as described by Takahashi and 184
Nagano (1984), the incompatibility (Inc) group was defined by the PCR-based replicon typing 185
(PBRT) method (Carattoli et al. 2005) and PCR and sequencing was performed for the 186
identification of the blaTEM-52 gene (Smet et al. 2009). These transconjugants were tested for 187
susceptibility to: ampicillin, amoxicillin-clavulanic acid, cefoxitin, ceftazidim, cefepime, 188
9
nalidixic acid, tetracycline, sulfonamides, trimethoprim, streptomycin, gentamicin and 189
chloramphenicol as earlier described (Smet et al. 2008).
190
At a given sample point, the average transfer frequency was estimated by dividing the number of 191
transconjugants per ml by the number of recipients (human E. coli) per ml. 192
193
Results 194
Experiment I: Caecal bacterial dynamics without ESBL+ donor strain 195
In a blank experiment (without administration of the lactose mutant of E. coli B1-54), simulating 196
the human caecum and the ascending colon, E. coli of human origin (and cefotaxime susceptible) 197
were able to stabilize at a population size of 6±0.2 log10 cfu/ml for 12 days in continuous culture 198
(Fig. 1). No spontaneous lactose negative E. coli mutants were detected during the experiment. 199
From day 12 onwards, the normal incubation medium was replaced by a stock containing 2.5 200
mg/l cefotaxime and maintained until day 15. Antimicrobial treatment gave an immediate 201
decrease of the population size (from 6 log10 to 3 log10 cfu/ml) for E. coli of human origin. At day 202
16, the medium with 2.5 mg/l cefotaxime was replaced again by the normal incubation medium. 203
After day 16, the population size of the cefotaxime susceptible E. coli of human origin increased 204
again and restored itself after discontinuation of the antimicrobial administration (Fig. 1). 205
206
Experiment II: Caecal bacterial dynamics with ESBL+ donor strain at the initiation of 207
experiment 208
The aim of the second experiment was to investigate whether the lactose mutant of E. coli B1-54 209
of broiler origin was able to maintain itself and transfer its blaTEM-52-carrying plasmid to E. coli 210
10
strains of human origin as well as the behaviour of the human transconjugants, in a fermentation 211
condition simulating a disturbance of the human microbial community. 212
The lactose mutant of E. coli B1-54 was added as donor strain together with the human stool 213
sample in the fermentation system, which was operated in batch mode for 24 h. Continuous 214
culture was then started by switching on the peristaltic pumps. Fluctuations in population size of 215
several bacterial groups were seen during the first five days (Fig. 2a). The lactose mutant of E. 216
coli B1-54 and the E. coli of human origin (and cefotaxime susceptible) could establish 217
themselves after 5 days at a population size of 7 ± 0.1 log10 cfu/ml and 7 ± 0.2 log10 cfu/ml, 218
respectively (Fig. 2a). Other average population sizes were 8 ± 1 log10, 9 ± 2 log10 and 9 ± 2 log10 219
cfu/ml for total aerobic bacteria, lactic acid bacteria and total anaerobes, respectively (Fig. 2b). 220
Twenty-four hours after administration of the lactose mutant of E. coli B1-54, transconjugants of 221
human origin, resistant to cefotaxime, were detected at a population size of 5 log10 cfu/ml (Fig. 222
2a). The average transfer frequency of ESBL-resistance was 2.5 x 10-3
at day 2 of the 223
fermentation experiment. From day 11 onwards, the normal incubation medium was replaced by 224
a stock containing 2.5 mg/l cefotaxime. The antimicrobial treatment was maintained until day 14. 225
During antimicrobial treatment (day 11-14), a decrease of the population size of the normal 226
cefotaxime susceptible human microbial community (E. coli (from 6 to 3 ± 0.04 log10) and total 227
aerobic bacteria (from 6 to 3 ± 0.1 log10)) was seen. On the contrary, an increase of the 228
population size of the lactose mutant of E. coli B1-54 (from 6 to 9 ± 0.3 log10) and of the human 229
transconjugants (from 5 to 8 ± 0.2 log10), both resistant to cefotaxime, was noted (Fig. 2a and 2b). 230
After stopping the antimicrobial treatment, the population size of the human E. coli community, 231
susceptible to cefotaxime, restored itself to the same level as before cefotaxime administration, 232
while the population size of the cefotaxime resistant lactose mutant of E. coli B1-54 and the 233
11
human transconjugants was maintained at the same high level as during antimicrobial treatment 234
(Fig. 2a). 235
The microbiota was disturbed as indicated by a different denaturing gradient gel electrophoresis 236
(DGGE) pattern compared to the fermentation condition without antibiotic (data not shown). At 237
the start of antimicrobial treatment, one band disappeared, but reappeared during antimicrobial 238
treatment. Also a few bands appeared more strongly during antimicrobial treatment. After 239
stopping the antimicrobial treatment, the DGGE pattern of the microbiota returned to the initial 240
condition, but the few more prominent bands which appeared during antimicrobial treatment, 241
remained in the pattern. Pure cultures of the lactose mutant of E. coli B1-54, of a human 242
transconjugant and of a human cefotaxime susceptible E. coli showed similar DGGE patterns 243
(data not shown). 244
245
Experiment III: Caecal bacterial dynamics with ESBL+ donor strain 24 hours after the 246
initiation of experiment 247
The aim of this third experiment was to verify whether the lactose mutant of E. coli B1-54 of 248
broiler origin was able to maintain itself and to transfer its blaTEM-52-carrying plasmid to E. coli 249
strains of human origin as well as the behaviour of the human transconjugants in a fermentation 250
condition in which the human microbial community was already dominantly present. 251
The human stool sample was added in the fermentation system and operated in batch mode for 24 252
h so that the human microbial community was predominantly present in the fermentation system. 253
The lactose mutant of E. coli B1-54 was supplemented after the switch to continuous culture 254
(after 1 day). The lactose mutant of E. coli B1-54 and the cefotaxime susceptible E. coli of 255
human origin could establish themselves at a population size of 3 ± 1 log10 and 7 ± 1 log10 cfu/ml, 256
respectively (Fig 3a). Other average population sizes were 7 ± 0.1 log10 cfu/ml for total aerobic 257
12
bacteria, lactic acid bacteria and total anaerobes (Fig. 3b). As seen in experiment II, 24h after 258
administration of the lactose mutant of E. coli B1-54, transconjugants of human origin were 259
detected. The population size (1 log10 cfu/ml) of the transconjugants was remarkably lower in 260
comparison with the second experiment. The average transfer frequency was 2.5 x 10-5
at day 3.
261
After 11 days of normal operation, the switch was made to medium stock with 2.5 mg/l 262
cefotaxime. Gradual replacement of the regular medium in the fermentor by medium 263
supplemented with 2.5 mg/l cefotaxime, resulted in a decrease of the E. coli of human origin (and 264
cefotaxime susceptible) (from 7 to 5 ± 0.04 log10) and an increase of the cefotaxime resistant 265
lactose mutant of E. coli B1-54 (from 3 to 5 ± 0.1 log10 cfu/ml) and the transconjugants of human 266
origin (from 2 to 5 ± 0.1 log10 cfu/ml) (Fig. 3a and 3b). At day 14, the feed was switched back to 267
the medium without 2.5 mg/l cefotaxime. After day 14, the population size of the cefotaxime 268
resistant lactose mutant of E. coli B1-54 and the transconjugants was maintained at the same high 269
level as during antimicrobial treatment, and the population size of the cefotaxime susceptible 270
human E. coli community gradually restored (Fig. 3a and 3b). 271
DGGE observations and characterisation showed highly comparable results with experiment II. 272
273
Verification of the E. coli transconjugants 274
A total of 60 E. coli transconjugants were retrieved from the MacConkey agar plates 275
(supplemented with 2.5 mg/l cefotaxime) during fermentation experiments II and III and were 276
verified by plasmid analysis (data not shown) and PFGE profiles. All isolated transconjugants 277
contained a large plasmid of >100 Kb from incompatibility group IncI1 and PCR and sequencing 278
revealed the presence of a blaTEM-52 gene on this plasmid (data not shown). Antimicrobial 279
susceptibility revealed that the E. coli transconjugants, besides showing an ESBL phenotype (i.e. 280
a marked synergistic effect between resistance to extended-spectrum cephalosporins and 281
13
amoxicillin with clavulanic acid), were susceptible to all other tested non-β-lactam antimicrobial 282
agents. PFGE revealed only two different fingerprint patterns among the isolated E. coli 283
transconjugants of human origin (Fig. 4). The most predominant fingerprint pattern was obtained 284
for 55 of the 60 isolates (91.7%) (lane 3 and 5 in Fig. 4), with a distinct fingerprint pattern 285
obtained for the other isolates. PFGE was also performed for a total of 20 random cefotaxime 286
susceptible E. coli strains isolated from the stool sample. Surprisingly, only one PFGE pattern 287
was seen and it was interesting to notice that this fingerprint pattern was identical to the 288
predominant fingerprint pattern found among the transconjugants. 289
290
Discussion 291
In this study, it was experimentally shown that a lactose mutant of ESBL-producing E. coli B1-292
54, isolated from a faecal sample of broilers, was able to proliferate in the fermentation system 293
simulating a part of the human gastrointestinal tract. The results suggest that the population size 294
of the donor strain depends on ongoing microbial competitions during the experiments that might 295
suffer from standardisation. Previous reports suggested that in order for antibiotic resistance 296
transfer to occur, high numbers (> 106 cfu/g) of both donor and recipient (compatible) strains 297
should coincide at the same time in the gut (Lester et al. 2006). In our study, E. coli 298
transconjugants of human origin were detected 24 h after inoculation of the donor strain B1-54 in 299
both experiments, even with the low concentration of the donor strain (3 log cfu/ml) in 300
experiment 3. Nevertheless, a substantial difference in transfer frequency (10-5
versus 10-3
) was 301
calculated in the two fermentation conditions. Thus, attention should be given to assess the 302
reproducibility of the fermentation conditions to allow accurate comparison of different 303
scenarios. Other limitations of the system are the fact that only the latter part of the digestive tract 304
14
is simulated while dynamics in resistance throughout the digestive tract (Catry et al. 2007) can 305
have an influence on transfer frequencies. Also factors like age, immunity, disease conditions 306
cannot be evaluated in the presented model (Graffunder et al. 2007; Levy 2002). The normal 307
human bacterial microbiota, dominantly present and acting as barrier for other bacteria, gives this 308
in situ model nevertheless advantages in mimicking the human gastrointestinal tract in 309
comparison to animal models. Also the possibility to measure the transfer frequency of 310
antimicrobial resistance determinants and to monitor the behaviour of the donor, recipient and 311
transconjugants are important advantages of this in situ model in comparison to animal models 312
(Schjorring et al. 2008). Despite the continuous input and output of the growth medium, 313
mimicking gastrointestinal transit, it has to be noted that conjugation of the blaTEM-52-carrying 314
plasmid still appeared even without selective pressure of an antimicrobial agent. This easy 315
transfer of the blaTEM-52-carrying plasmid to another recipient was not so surprising since the 316
transfer of this plasmid has also been investigated by in vitro liquid mating experiments between 317
individual strains in a previous report, for which the transfer frequency was 1.27 x 10-3
(Smet et 318
al. 2009). A comparison between these transfer frequencies reveals that the transfer of blaTEM-52 319
encoded resistance between E. coli strains can be influenced by the human gut microbiota under 320
stabilized conditions. Probably the lower transfer frequency of 10-5
is a more realistic estimation 321
of the normal situation. 322
Due to antimicrobial cefotaxime administration, both the lactose mutant of E. coli B1-54 of 323
broiler origin and the human E. coli transconjugants were able to proliferate quickly. In both 324
simulated conditions (disturbed and stabilized gut microbiota), an increase with 2-3 log of both 325
resistant E. coli populations was observed. After antimicrobial administration, the population size 326
of the normal microbial community, susceptible to cephalosporins, increased again to previously 327
obtained levels, while the cephalosporin resistant lactose mutant of E. coli B1-54 and the E. coli 328
15
transconjugants were able to maintain at the increased levels in the fermentation system. 329
Nevertheless, our experiments clearly support other reports (Cloeckaert et al. 2007; Persoons et 330
al. 2010) in which third-generation cephalosporins substantially enhance the proliferation of 331
ESBL-producing bacteria in the digestive tract. From the molecular cultivation independent 332
DGGE approach, it was also evident that after antimicrobial administration some bacterial 333
groups, other than the E. coli population and probably cefotaxime (inherently or acquired) 334
resistant, had increased in population size. These results indicate that antimicrobial agents may 335
not only select for resistant bacteria, but may also increase the potential of these resistant bacteria 336
as donors of resistance genes by increasing their population size. 337
The human E. coli transconjugants were genotyped using PFGE. Since E. coli has a polyclonal 338
population structure, it would be expected to find several PFGE patterns. However, only two 339
PFGE patterns were found, indicating that only two human E. coli clones received the blaTEM-52-340
carrying plasmid. Besides a dominant transconjugant clone which was genetically identical to the 341
dominant clone in the faecal E. coli population of the human donor, also a minor transconjugant 342
clone was selected with a genotype which was not picked up in the random selection of the faecal 343
E. coli isolates. The reason behind this phenomenon is unknown and needs further investigation. 344
To our knowledge, this is the first time that an in situ system simulating the human caecum and 345
the ascending colon was used to study the behaviour of an E. coli isolate of broiler origin as well 346
as of the human E. coli transconjugants after the transfer of an ESBL gene from a poultry-derived 347
E. coli to (a) recipient(s) of human origin. These observations provide evidence of the 348
maintenance of an E. coli strain of poultry origin and the horizontal gene transfer in the human 349
commensal bowel microbiota even without antimicrobial treatment. 350
In conclusion, the ability of an ESBL-producing E. coli strain of animal origin to be maintained 351
under normal or disturbed simulated physiological conditions in the human gut and to transfer the 352
16
ESBL-resistance gene residing on a large plasmid to commensal bacteria of human origin without 353
antimicrobial treatment raises public health concerns. In addition, cephalosporin treatment 354
increases potential health risk by increasing the population size of the cephalosporin resistant 355
microbial population. Further research is necessary to determine the risk of carriage in the human 356
gut of ESBL-producing microbiota from consumption of food products such as poultry with a 357
high prevalence of ESBL-producing E. coli and the potential risk of long-term carriage of a high 358
population level of this resistant microbiota in the human gut after antimicrobial cephalosporin 359
treatment. 360
361
Acknowledgements 362
This work was supported by a grant of Federal Public Service of Health, Food Chain Safety and 363
Environment (Grant number RT 06/3 ABRISK). 364
365
366
367
368
369
370
371
372
373
374
17
References 375
Bertrand, S., Weill, F.X., Cloeckaert, A., Vrints, M. Mairiaux, E., Praud, K., Dierick, K., 376
Wildemauve, C., Godard, C., Butaye, P., Imbrechts, H., Grimont, P.A. and Collard, J.M. 377
(2006) Clonal emergence of extended-spectrum beta-lactamase (CTX-M-2)-producing 378
Salmonella enterica serovar Virchow isolates with reduced susceptibilities to 379
ciprofloxacin among poultry and humans in Belgium and France (2000 to 2003). J Clin 380
Microbiol 44, 2897-2903. 381
Bidet, P., Burghoffer, B., Gautier, V., Brahimi, N., Mariani-Kurkdjian, P., El-Ghonemi, A., 382
Bingen, E. and Arlet, A. (2005) In vivo transfer of plasmid-encoded ACC-1 AmpC from 383
Klebsiella pneumoniae to Escherichia coli in an infant and selection of impermiability to 384
imipenem in K. pneumoniae. Antimicrob Agents Chemother 49, 3562-3565. 385
Boon, N., Goris, J., De Vos, P., Verstraete, W. and Top, E.M. (2000) Bioaugmentation of 386
activated sludge by an indigenious 3-chloroaniline-degradin Comamonas 387
testosteronistrain, 12gfp. Appl Environ Microbiol 66, 2906-2913. 388
Carattoli, A., Bertini, A., Villa, L., Falbo, V., Hopkins, K.L. and Threlfall, E.J. (2005) 389
Identification of plasmids by PCR-based replicon typing. J Microbiol Methods 63, 219-390
228. 391
Catry, B., Dewulf, J., Goffin, T., Decostere, A., Haesebrouck, F. and De Kruif, A. (2007) 392
Antimicrobial resistance patterns of Escherichia coli through the digestive tract of veal 393
calves. Microb Drug Resist 13, 147-150. 394
Cloeckaert, A., Praud, K., Doublet, A., Bertini, A. Carattoli, A., Butaye, P., Imbrechts, H., 395
Bertrand, S., Collard, J.M., Arlet, G. and Weill, F.X. (2007) Dissemination of an 396
extended-spectrum-β-lactamase blaTEM-52 gene-carrying IncI1 plasmid in various 397
18
Salmonella enterica Serovars isolated from poultry and humans in Belgium and France. 398
Antimicrob Agents Chemother 51, 1872-1875. 399
Graffunder, E.M., Preston, K.E., Evans, A.M. and Venezia, R.A. (2005) Risk factors associated 400
with extended-spectrum beta-lactamase producing organisms at a tertiary hospital. J 401
Antimicrob Chemother 56, 139-145. 402
Hasman, H., Mevius, D., Veldman, K., Olesen, I. and Aarestrup F.M. (2005) β-lactamases among 403
extended-spectrum β-lactamase (ESBL)-resistant Salmonella from poultry, poultry 404
products and human patients in the Netherlands. J Antimicrob Chemother 56, 115-121. 405
Karami, N. Martner, A., Enne, V.I., Swerkersson, S., Adlerberth, A. and Wold, A.E. (2007) 406
Transfer of an ampicillin resistance gene between two Escherichia coli strains in the 407
bowel microbiota of an infant treated with antibiotics. J Antimicrob Chemother 60,1142-408
1145. 409
Kojima, A., Ishii, Y. and Ishihara, K. (2005) Extended-spectrum-beta-lactamase-producing 410
Escherichia coli strains isolated from farm animals from 1999 to 2002: report from the 411
Japanese Veterinary Antimicrobial Resistance Monitoring Program. Antimicrob Agents 412
Chemother 49, 3533-3537. 413
Lester, C.H., Frimodt-Moller, N., Sorensen, T.L., Monnet, D.L. and Hammerum, A.M. (2006) In 414
vivo transfer of the vanA resistance gene from an Enterococcus faecium isolate of animal 415
origin to an E. faecium isolate of human origin in the intestine of human volunteers. 416
Antimicrob Agents Chemother 50: 596-599. 417
Levy, S.B. (2002) The 2000 Garrod lecture: factors impacting on the problem of antibiotic 418
resistance. J Antimicrob Chemother 49, 25-30. 419
Liu, J.H., Wei, S.Y., Ma, J.Y., Zeng, Z.L., Lü, D.H., Yang, G.X. and Chen, Z.L. (2007) Detection 420
and Characterization of CTX-M and CMY-2 β-lactamases among Escherichia coli 421
19
isolates from farm animals in Guangdong province of China. Int J Antimicrob Agents 29, 422
576-581. 423
MacFarlane, S., Woodmansey, E.J. and Macfarlane, G.T. (2005) Colonization of mucin by 424
human intestinal bacteria and establishment of biofilm communities in a two stage 425
continuous culture system. Appl Environ Microbiol 71, 7483-7492. 426
Messens, W., Goris, J., Dierick, N., Herman, L. and Heyndrick, M. (2009) Inhibition of 427
Salmonella Typhimurium by medium-chain fatty acids in an in vitro simulation of the 428
porcine cecum. Vet Microbiol doi:10.1016/j.vetmic.2009.08.002. 429
Moubareck, C., Bourgeois, N., Courvalin, P. and Doucet-Populaire, F. (2003) Multiple antibiotic 430
resistance gene transfer from animal to human enterococci in the digestive tract of 431
gnotobiotic mice. Antimicrob Agents Chemother 47: 2993-2996. 432
Netherwood, T., Bowden, R., Harrison, P., O’Donnell, A.G., Parker, D.S. and Gilbert, H.J. 433
(1999) Gene transfer in the gastrointestinal tract. Appl Environ Microbiol 65, 5139-5141. 434
Persoons, D., Dewulf, J., Smet, A., Herman, L., Heyndrickx, M., Martel, A., Catry, B., Butaye, P. 435
and Haesebrouck, F. (2010) Prevalence and persistence of antimicrobial resistance in E. 436
coli from broiler chickens. Microb Drug Res 16, 67-74. 437
Riaño, I., Moreno, M.A., Teshager, T., Saenz, Y., Dominguez, L. and Torres, C. (2006) Detection 438
and characterization of extended- spectrum beta-lactamases in Salmonella enterica strains 439
of healthy food animals in Spain. J Antimicrob Chemother 58, 844-847 440
Schjorring S., Struve, C. and Krogfelt, K.A. (2008) Transfer of antimicrobial resistance plasmids 441
from Klebsiella pneumoniae to Escherichia coli in the mouse intestine. J Antimicrob 442
Chemother 62, 1086-1093. 443
20
Shah, A.A., Hasan, F., Ahmed, S. and Hameed, A. (2004) Characteristics, epidemiology and 444
clinical importance of emerging strains of gram-negative bacilli producing extended-445
spectrum β-lactamases. Res Microbiol 155, 409-421. 446
Sheldon, A.T. (2005) Antibiotic resistance: a survival strategy. Clin Lab Sci 18, 170-180. 447
Smet A., Martel, A., Persoons, D., Dewulf, J., Heyndrickx, M., Cloeckaert, A., Praud, K., 448
Claeys, G., Catry, B., Herman, L., Haesebrouck, F. and Butaye, P. (2009) Comparative 449
analysis of Extended-spectrum-β-lactamase (ESBL)-carrying plasmids from different 450
members of Enterobacteriaceae isolated from poultry, pigs and humans: evidence for a 451
shared β-lactam resistance gene pool? J Antimicrob Chemother 63, 1286-1288. 452
Smet, A., Martel, A., Persoons, D., Dewulf, J., Heyndrickx, M., Catry, B., Herman, L., 453
Haesebrouck, F. and Butaye, P. (2008) Diversity of extended-spectrum β-lactamases and 454
class C β-lactamases among cloacal Escherichia coli in Belgian broiler farms. Antimicrob 455
Agents Chemother 52, 1238-1243. 456
Takahashi, S. and Nagano, Y. (1984) Rapid procedure for isolation of plasmid DNA and 457
application to epidemiological analysis. J Clin Microbiol 20, 608-613. 458
Trobos, M., Lester, C.H., Olsen, J.E., Frimodt-Moller, N. and Hammerum, A.M. (2009) Natural 459
transfer of sulphonamide and ampicillin resistance between Escherichia coli residing in 460
the human intestine. J Antimicrob Chemother 63, 80-86. 461
Varela, M.F., Brooker, R.J. and Wilsons, T.H. (1997) Lactose carrier mutants of Escherichia coli 462
with changes in sugar recognition (lactose versus melibiose). J Bacteriol 179, 5570-5573. 463
Vollaard, E.J. and Clasener, H.A. (1994) Colonization resistance. Antimicrob Agents Chemother 464
38, 409-414. 465
466
467
21
Figure Legends 468
Figure 1 Influence of antimicrobial treatment on the microbial community of the human caecum 469
and the ascending colon (blank experiment). Start continuous culture at day 1. The start and end 470
of the antimicrobial treatment (from day 12 to day 15) is indicated by the box. 471
472
Figure 2 Administration of E. coli B1-54 before the microbial community of the human ceacum 473
and the ascending colon reached a steady state. Start continuous culture at day 1. The start and 474
end of the antimicrobial treatment (from day 11 to day 14) is indicated by the box. 475
476
Figure 3 Administration of E. coli B1-54 after the microbial community of the human ceacum 477
and the ascending colon reached a steady state. Start continuous culture at day 1. The start and 478
end of the antimicrobial treatment (from day 11 to day 14) is indicated by the box. 479
480
481
Figure 4 PFGE profiles of the donor, recipient and the E. coli transconjugants (experiment 2). M: 482
marker; 1 and 2: E. coli of human origin (susceptible to cefotaxime; recipient); 3: E. coli 483
transconjugant (dominant PFGE fingerprint pattern, experiment 2 before antimicrobial 484
treatment); 4: E. coli transconjugant; 5: E. coli transconjugant (dominant PFGE fingerprint 485
pattern, experiment 3, before antimicrobial treatment); 6: E. coli B1-54 (donor). 486
22
Figure 1 487
1,00E+00
1,00E+01
1,00E+02
1,00E+03
1,00E+04
1,00E+05
1,00E+06
1,00E+07
1,00E+08
1,00E+09
1,00E+10
0 1 2 3 4 5 6 8 12 13 15 16 17 18 19 20 21 24 27
time after inoculation (days)
CF
U/m
l
E. coli (human)
total aerobic bacteria
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
23
Figure 2 505
(a) 506
1,00E+00
1,00E+01
1,00E+02
1,00E+03
1,00E+04
1,00E+05
1,00E+06
1,00E+07
1,00E+08
1,00E+09
1,00E+10
1,00E+11
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
time after inoculation (days)
CF
U/m
l E. coli (human)
B1-54
transconjugants
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
24
(b) 524
1,00E+00
1,00E+01
1,00E+02
1,00E+03
1,00E+04
1,00E+05
1,00E+06
1,00E+07
1,00E+08
1,00E+09
1,00E+10
1,00E+11
1,00E+12
0 1 2 3 4 6 7 8 9 10 11 12 13 14 15 16 17
time after inoculation (days)
CF
U/m
l
lactic acid bacteria
anaerobes
total aerobic bacteria
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
25
Figure 3 542
(a) 543
544
1,00E+00
1,00E+01
1,00E+02
1,00E+03
1,00E+04
1,00E+05
1,00E+06
1,00E+07
1,00E+08
0 1 2 3 4 6 7 8 9 10 11 12 13 14 16 17
time after inoculation (days)
CF
U/m
l E. coli (human)
B1-54
transconjugants
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
26
(b) 560
1,00E+00
1,00E+01
1,00E+02
1,00E+03
1,00E+04
1,00E+05
1,00E+06
1,00E+07
1,00E+08
0 1 2 3 4 6 7 8 9 10 11 12 13 14 16 17
time after inoculation (days)
CF
U/m
l lactic acid bacteria
anaerobes
total aerobic bacteria
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577