Environmental Project No. 722 2002 Miljøprojekt Occurrence and fate of antibiotic resistant bacteria in sewage Luca Guardabassi and Anders Dalsgaard The Royal Veterinary and Agricultural University, Department of Veterinary Microbiology The Danish Environmental Protection Agency will, when opportunity offers, publish reports and contributions relating to environmental research and development projects financed via the Danish EPA. Please note that publication does not signify that the contents of the reports necessarily reflect the views of the Danish EPA. The reports are, however, published because the Danish EPA finds that the studies represent a valuable contribution to the debate on environmental policy in Denmark. Table of Contents PREFACE 5 SUMMARY AND CONCLUSIONS 7 SAMMENDRAG OG KONKLUSIONER 13 1 INTRODUCTION AND PROJECT BACKGROUND 19 1.1 PROJECT STRUCTURE AND OBJECTIVES 19 1.2 WHAT ARE ANTIBIOTICS? 21 1.2.1 Classification 21 1.2.2 Mechanisms of action 22 1.3 WHAT IS ANTIBIOTIC RESISTANCE? 22 1.3.1 Molecular mechanisms 23 1.3.2 Natural and acquired resistance 23 1.3.3 Acquisition by chromosomal mutations 24 1.3.4 Acquisition by horizontal gene transfer 24 1.3.5 Intracellular migration of resistance genes 24 1.3.6 Measurement of resistance in bacterial populations 25 1.4 THE MICROBIAL THREAT 26 1.4.1 The emergence of resistance in human pathogenic bacteria 26 1.4.2 The spread of resistance among environmental bacteria 27 1.5 SPREAD OF ANTIBIOTIC RESISTANCE IN SEWAGE 28 1.5.1 Antibiotic selective pressure 28 1.5.2 Non-antibiotic selective pressure 29 1.5.3 Optimal conditions for horizontal gene transfer 29 2 METHODOLOGY 31 2.1 SAMPLING SITES, TIMES AND METHODS 31 2.1.1 Sampling at sewers 31 2.1.2 Sampling at sewage treatment plants 32 2.2 MEASUREMENT OF ANTIBIOTIC RESISTANCE 32 2.2.1 Use of Acinetobacter as a bacterial indicator 32 2.2.2 Antibiotic susceptibility testing of Acinetobacter isolates 33 2.2.3 Enumeration of resistant coliforms in sewage 33 2.2.4 Enumeration of resistant acinetobacters in sewage 33 2.2.5 Enumeration of total culturable resistant bacteria in blue mussels 34 2.2.6 Enumeration of resistant E. coli in blue mussels 34 2.2.7 Statistical analysis 34 2.3 IDENTIFICATION AND TYPING OF BACTERIA AND RESISTANCE GENES 35 2.3.1 Identification of Acinetobacter at the genus level 35 2.3.2 Identification of Acinetobacter at the species level 35 2.3.3 Plasmid profiles 35 2.3.4 Ribotyping 35 2.3.5 Typing of tetracycline resistance genes 35 2.4 EXPERIMENTS ON TRANSFER OF TETRACYCLINE RESISTANCE 36 2.4.1 Bacterial strains 36 2.4.2 Mating experiments 36 3 2.5 EXPERIMENTS ON SURVIVAL OF MULTIPLE-RESISTANT BACTERIA IN NATURAL WATERS 37 2.5.1 Bacterial strains 37 2.5.2 Laboratory seawater microcosms 37 2.5.3 In situ pond experiment 37 3 EFFECTS OF HOSPITAL AND PHARMACEUTICAL WASTE EFFLUENT ON THE PREVALENCE OF RESISTANT ACINETOBACTER IN THE RECIPIENT SEWERS 39 3.1 EFFECTS OF HOSPITAL WASTE EFFLUENT 39 3.2 EFFECTS OF PHARMACEUTICAL WASTE EFFLUENT 41 3.3 CONCLUSIONS 44 4 EFFECTS OF SEWAGE TREATMENT ON TOTAL NUMBERS AND PERCENTAGES OF RESISTANT BACTERIA 47 4.1 EFFECTS ON TOTAL NUMBERS OF RESISTANT BACTERIA 47 4.2 EFFECTS ON PERCENTAGES OF RESISTANT BACTERIA 50 4.3 CONCLUSIONS 53 5 SPREAD OF RESISTANT BACTERIA AND RESISTANCE GENES BY MUNICIPAL SEWAGE EFFLUENTS 55 5.1 SURVIVAL IN THE ENVIRONMENT OF RESISTANT BACTERIA ORIGINATING FROM SEWAGE 55 5.1.1 Survival in laboratory seawater microcosms 55 5.1.2 Survival in a freshwater pond 57 5.2 TRANSFER OF RESISTANCE GENES FROM SEWAGE TO AQUATIC BACTERIA 59 5.2.1 Laboratory mating experiments 59 5.2.2 Distribution of tetracycline resistance genes 59 5.3 OCCURRENCE OF RESISTANT BACTERIA IN BLUE MUSSELS EXPOSED TO TREATED SEWAGE 60 5.3.1 Antibiotic resistance of total culturable bacteria in blue mussels 60 5.3.2 Antibiotic resistance of E. coli in blue mussels 62 5.4 CONCLUSIONS 63 6 REFERENCE LIST 65 ANNEX 1. SCIENTIFIC PAPERS AND DISSEMINATION OF RESULTS 70 4 Preface This report contains the results of a project conducted at the Department of Veterinary Microbiology of The Royal Veterinary and Agricultural University (RVAU). The project is composed of three parts (I, II and III). After an introductory chapter on antibiotics, antibiotic resistance and project structure (chapter 1), the methodology used is described in chapter 2. The outcomes of each part of the project are then reported and discussed separately in chapters 3, 4 and 5. The following people have taken part in the project: Luca Guardabassi, Assistant Professor, Department of Veterinary Microbiology, RVAU. Anders Dalsgaard, Associate Professor, Department of Veterinary Microbiology, RVAU. The project steering group consisted of: Linda Bagge, the Danish Environmental Protection Agency. Eva Vestergård, the Danish Environmental Protection Agency. Bo Neegard Jacobsen, Spildevandscenter Avedøre I/S, Denmark. Kim Rindel, Lynettefællesskabet I/S, Denmark. The project was supported by the Danish Environmental Protection Agency (Miljøstyrelsen). Additional financial support was provided by the two sewage treatment plants of Avedøre (Avedøre Spillevandscenter I/S) and Lynetten (Lynettefællesskabet I/S). The results of the project have been published in national and international journals and presented at scientific meetings as listed in Annex 1. 5 6 Summary and conclusions In this project, we investigated different aspects concerning the occurrence and fate of antibiotic resistant bacteria in sewage: • the effects of waste effluent from a hospital and a pharmaceutical plant on the prevalence of resistant Acinetobacter in the recipient sewers (Part I) • the effects of tertiary sewage treatment on total numbers and percentages of resistant bacteria (Part II) • the survival in natural aquatic habitats of multiple-resistant bacteria originating from municipal sewage effluents, and more in general the impact of such effluents on the spread of antibiotic resistance (Part III). Part I In Part I of the project, Acinetobacter was used as a bacterial indicator for monitoring antibiotic resistance in the sewers receiving waste effluent from two potential sources of resistant bacteria and/or antibiotic residues: a hospital and a pharmaceutical plant manufacturing antibiotics. The choice of Acinetobacter was prompted by the normal occurrence of these organisms in water and their remarkable ability to develop antibiotic resistance. The levels of susceptibility to six antibiotics were determined by the disc- diffusion method in 385 Acinetobacter isolates from sewage collected upstream and downstream from the discharge points of the hospital (n=180) and the pharmaceutical plant (n=205). Isolates from the sewers at the pharmaceutical plant were further analysed by plasmid DNA profiling and phenotypic tests to detect any changes in the distribution of Acinetobacter species/strains associated with the discharge of waste effluent from this source. Statistical analysis of the data from antibiotic susceptibility testing showed that the discharge of waste effluent from the pharmaceutical plant was associated with a statistically significant increase in the prevalence of both single and multiple antibiotic resistance profiles amongst Acinetobacter isolates from the recipient sewers (logistic regression P < 0.01). This increase of antibiotic resistance was observed throughout the period of study and was evident up to 250 m downstream from the discharge point. Strains isolated downstream from the pharmaceutical plant discharge point also demonstrated different plasmid profiles and phenotypic traits compared with those isolated upstream. In contrast, the hospital waste effluent increased only the prevalence of resistance to one antibiotic, namely oxytetracycline, amongst Acinetobacter isolates from the recipient sewers. Furthermore, the increase in tetracycline resistance observed immediately downstream from the hospital discharge point was significantly reduced 500 m further downstream from the discharge point. 7 Based on this evidence, it was concluded that the discharge of waste effluent from the pharmaceutical plant caused an increase in the prevalence of both single and multiple antibiotic resistance amongst Acinetobacter spp. in the recipient sewers and had a higher impact on antibiotic resistance compared with the discharge of hospital waste effluent. Furthermore, waste effluent from the pharmaceutical plant determined a change in the distribution of Acinetobacter species/strains in the sewers, probably due to introduction of either antibiotic residues or resistant strains. Part II Part II of the project dealt with the effects of tertiary sewage treatment on the prevalence of resistant bacteria in two large-scale municipal treatment plants during a period of six months. Total and relative numbers of resistant bacteria in raw sewage, treated sewage and anaerobically digested sludge were determined by bacteriological counts on agar media with and without inclusion of ampicillin, tetracycline, gentamicin or all three antibiotics. Two different agar media, one selective for coliforms (MacConkey agar) and one selective for Acinetobacter (Baumann agar), were used in order to study the effect of treatment on different bacterial populations. In addition, the levels of susceptibility to 14 antibiotics were determined by the disc-diffusion method in 442 Acinetobacter isolates identified by colony hybridisation with a genus- specific DNA probe. Depending on the different antibiotics and media used for bacteriological counts, the total numbers of resistant bacteria ranged between 10 to 1000-fold lower in treated sewage than in raw sewage. For both bacterial populations under study, the prevalence of resistant bacteria in treated sewage and digested sludge were not significantly higher than in raw sewage. On the contrary, the prevalence of ampicillin-resistant acinetobacters (i.e. presumptive Acinetobacter not identified at the genus level) was significantly reduced by sewage treatment at one plant (linear regression P <0.05). Similarly, sludge treatment determined a reduction in the prevalence of ampicillin-resistant acinetobacters, as well as ampicillin- and gentamicin- resistant putative coliforms (linear regression P <0.05). The results obtained by bacteriological counts were confirmed by antibiotic susceptibility testing of Acinetobacter isolates. Based on logistic regression analysis, the frequencies of antibiotic resistance in isolates from treated sewage and digested sludge were not significantly higher in comparison with those in isolates from raw sewage. Comparison of the levels of resistance to 14 antimicrobial agents between isolates obtained from raw and treated sewage allowed detection of a statistically significant increase in the prevalence of antibiotic resistance to only one antibiotic (nalidixic acid) and restricted to one plant. It can therefore be concluded that tertiary sewage treatment did not determine a selection for resistant bacteria. In accordance, sewage treatment appears to reduce numbers of bacteria irrespective of their susceptibility to antibiotics. Although the results of the study clearly indicated that the overall prevalence of antimicrobial resistant bacteria was not increased by sewage treatment, the final effluent of one plant was found to contain low numbers (10 to 102 CFU/ml) of bacteria resistant to ampicillin, gentamicin and tetracycline, collectively. This multiple resistance phenotype is not likely to occur naturally in aquatic bacteria, as suggested by the absence of bacterial growth following inoculation of freshwater and seawater samples on agar plates containing these 8 three antibiotics. Consequently, it was decided to investigate the ability of such multiple-resistant bacteria to survive in natural aquatic environments (Part III). Part III In the final part of the project, three multiple-resistant strains isolated from treated sewage were investigated for their ability to survive in natural waters and retain antibiotic resistance. In parallel, survival experiments in laboratory seawater microcosms and membrane-filter chambers immersed in a freshwater pond were also carried out. The three strains were representative of three different bacterial species: Acinetobacter johnsonii, Escherichia coli and Citrobacter freundii. The multiple resistance patterns of these strains were used as selective markers for their detection among the indigenous bacteria. The experiments were performed using low bacterial inoculums (103 to 104 CFU ml-1) appropriate to reproduce the actual conditions occurring when treated sewage is discharged into natural aquatic recipients. Two of the three multiple-resistant strains (Escherichia coli and Citrobacter freundii) survived for at least one month in the seawater microcosms and the freshwater pond, whereas, the Acinetobacter johnsonii strain survived for shorter times in both settings. The results demonstrated that some multiple- resistant strains occurring in municipal sewage effluents were able to survive for relatively long periods following their release into natural aquatic habitats. The strains survived longer in autoclaved water better than in untreated water, suggesting that the presence of the indigenous microflora affected survival, presumably due to antagonism or predation. However, the strains maintained their multiple resistance properties following one month of incubation under natural conditions, indicating that stress and nutrient depletion did not affect the stability of their resistance phenotypes. An aspect of the study in Part III focused on the possibility that antibiotic resistance genes occurring in sewage could be transferred to bacteria living in natural aquatic environments. This was studied by laboratory mating experiments. Ten unrelated tetracycline-resistant Acinetobacter strains isolated from sewage (n=10) were mated with a tetracycline-sensitive Acinetobacter strain isolated from an unpolluted stream. Only two out of ten donor strains were able to transfer tetracycline resistance under laboratory conditions (in vitro). In one instance, transfer of tetracycline resistance was associated with relocation of multiple small plasmids from the donor to the recipient strain, whereas, in another, transfer was apparently not mediated by plasmid conjugation. These results confirmed that sewage is a possible vehicle for the dissemination of antibiotic resistance genes in the indigenous microflora of aquatic habitats. However, the limited number of strains used in the mating experiments did not allow conclusions to be drawn on the occurrence of this phenomenon in nature. The impact of municipal sewage effluents on the spread of antibiotic resistance was further investigated by comparing the occurrence of resistant bacteria in blue mussels exposed to sewage effluents and in blue mussels originating from unpolluted sites. Blue mussels were selected as a biological niche due to their ability to harbour high numbers of bacteria through daily filtration of large volumes of water. Higher percentages of ampicillin-resistant bacteria were found in mussels exposed to treated sewage (12.9 to 95.5%) in comparison with mussels not exposed to treated sewage (1.5 to 5.4%), whereas, the percentages of gentamicin- and tetracycline-resistant bacteria 9 were low (<3%) independent of the origin of the mussels. Small traces (#0.1%) of multiple-resistant bacteria were found only in mussels exposed to treated sewage, suggesting that municipal sewage effluents are potentially a source for the spread of these bacteria in the aquatic environment. The isolation of resistant bacteria was generally higher in mussels collected from the immediate proximity of the outlets of sewage effluents compared with mussels collected at 100 m from the outlets, indicating a correlation between prevalence of resistant bacteria and distance from the outlet. Conclusions The results of this project indicate that the occurrence of single and multiple- resistant bacteria in sewage can be increased by the discharge of waste effluent from pharmaceutical plants producing or manufacturing antibiotics. To our knowledge, this was the first study demonstrating an impact of antibiotic manufacturing on the occurrence of resistant bacteria in sewage. The observed increase in the prevalence of resistant Acinetobacter could have been due to the presence in the effluent of resistant bacteria selected inside the plant by the presence of high antibiotic concentrations. Alternatively, antibiotic residues may have caused a selection of resistant bacteria in the recipient sewers, or a combination of both. These results indicate that waste effluents from pharmaceutical plants manufacturing antibiotics are an important source for the occurrence and selection of resistant bacteria in sewage. As this study was conducted on a single pharmaceutical plant, further investigation is needed to assess the role of antibiotic manufacturing on selection and/or introduction of resistant bacteria in sewage. The occurrence and fate of resistant bacteria in sewage should be given carefully consideration due to the ubiquitous nature of bacteria. Sewage is a convenient and suitable vehicle for the dissemination of resistant bacteria, in that it connects antibiotic selective environments, such as hospitals, chemical industries, farms and slaughterhouses to natural environments. Risk assessment was not included as an objective in this study. However, risks for human health may result from the dissemination in the environment of resistant bacteria occurring in sewage and the possible contamination of bathing and drinking water with these organisms. Our investigation indicates that sewage treatment causes a reduction in the total numbers of resistant bacteria without increasing their percentage in treated sewage compared with raw sewage. Therefore, it appears that treatment of sewage has a positive effect in limiting the dissemination of resistant bacteria. However, the investigation also demonstrates that: • Multiple-resistant bacteria occurring in raw sewage can survive treatment and reach natural aquatic environments by municipal sewage effluents. • Multiple-resistant bacteria occurring in municipal sewage effluents can survive for relatively long periods and maintain their resistance properties following introduction into natural aquatic habitats. • Bacteria resistant to three or four different classes of antibiotics were found in shellfish exposed to municipal sewage effluents but appear to 10