The release of antimicrobial substances into surface waters is of growing concern due to direct toxic effects on all trophic levels and the promotion of antibiotic resistance through sub-inhibitory concentration levels. This study showcases (1) the variation of antibiotics in sanitary sewage depending on different timescales and (2) a method to assess the antibiotic resistance based on an inhibition test. The test is based on the measurement of the oxygen uptake rate (OUR) in wastewater samples with increasing concentrations of the selected antibiotic agents. The following antibiotics were analysed in the present study: clarithromycin (CLA) was selected due to its high toxicity to many microorganisms (low EC50), ciprofloxacin (CIP) which is used to generally fight all bacteria concerning interstitial infections and doxycyclin (DOX) having a broad spectrum efficacy. Results show that CLA inhibited the OUR by approximately 50% at a concentration of about 10 mg L−1, because Gram-negative bacteria such as Escherichia coli are resistant, whereas CIP inhibited about 90% of the OUR at a concentration equal to or greater than 10 mg L−1. In the case of DOX, a moderate inhibition of about 38% at a concentration of 10 mg L−1 was identified, indicating a significant antibiotic resistance. The results are consistent with the corresponding findings from the Clinical and Laboratory Standards Institute. Thus, the presented inhibition test provides a simple but robust alternative method to assess antibiotic resistance in biofilms instead of more complex clinical tests.

INTRODUCTION

The application of antibiotics in modern life is widely established to inhibit the growth of a microorganism and so prevent the expansion of microorganism-induced diseases, especially in developed areas with a high-density population. As a consequence of the dosage principle, a portion of the consumed antibiotics is inevitably not metabolized in human body and therefore secreted via urine and faeces into the sewer systems (Feldmann 2005). This results in a continuous occurrence of antibiotics in municipal sewage at a sub-inhibitory concentration level, two to three orders of magnitude below the toxic concentration (Kummerer et al. 2004). Through this:

The community matrix of microorganisms frequently changes, caused by selective pressure and continuous attachment/detachment processes within the sewer, induced through varying shear forces during the drainage process (randomly changing dry and rain weather loading). Therefore, the evaluation of the community matrix is always a time- and site-specific snap-shot within the urban drainage system. Resistant bacteria, either directly suspended or detached from the sewer biofilm, may leave the system by two ways: (1) via the wastewater treatment plant (WWTP); or (2) a direct discharge to surface waters via potentially occurring combined sewer overflows (CSOs). The management strategy should therefore aim at reducing close contact of sewage bacteria with microorganisms in the adjacent ecosystems to minimize resistant gene transfer. This is particularly relevant as antibiotic resistance still persists after a long period with a lack of the active antibiotic agent (Sjogren 1995; Morell 1997; Schrag et al. 1997), representing a long-term risk rather than a temporary concern. Consequently, this risk has to be considered in a prospective drainage management to ensure a minimum of direct or indirect discharge via WWTPs and CSOs. The following questions (and the corresponding collection of information) appear relevant for the evaluation of the environmental risk:
  • (1) To what extent are resistant genes present in the urban drainage system?

  • (2) To what extent do scouring effects occur, i.e. biofilm detaches as a function of the shear stress and becomes mobile and occurs in suspended form?

  • (3) To what extent are suspended biofilm fragments (particulate organic matter) discharged via CSOs? Quantification of hydraulics as basis for transport of soluble constituents is state-of-the-art in hydraulic modelling, whereas the model-based description of particulate components is rarely accomplished and highly uncertain.

First, this paper focuses on a method which explores:
  • (1) the presence of antibiotic resistance without using the time-consuming DNA extraction by reading the genetic code; and

  • (2) the behaviour of a mixed biocoenosis (biofilms) rather than a single species.

Secondly, the antibiotic concentrations in sewage leading to the resistance among microorganisms are analysed.

MATERIALS AND METHODS

Sampling site and antibiotic agents

A field sampling campaign aimed at monitoring sewage characteristics was conducted in the southeastern part of the city of Dresden, Germany (ca. 500,000 inhabitants). At the monitored location a residential area of 144 ha with an average population density of 42 inhabitants ha−1 is drained via a combined sewer network. Since 76% of the city's sewer network is operated as a combined sewer system, this site can be considered to be characteristic for Dresden. During dry weather conditions, the considered sewer section conveys sewage of 1.2% of the population of Dresden to the adjacent district. During a period of intense rainfall, on the other hand, the nearby located CSO structure may spill combined sewage, a composite of varying parts of sewage and storm water, into the adjacent river Lockwitzbach. The samples were collected using cooled automatic samplers type TP4 P (MAXX Mess- und. Probenahmetechnik GmbH) configured for different sampling routines:

  • (1) ‘Macro-level’ – sampling every 12 minutes a volume of 100 mL into a bottle for the 24-hour-mixed sample.

  • (2) ‘Meso-level’ – sampling every 2 minutes a volume of 100 mL into a bottle for the 1-hour-mixed sample.

  • (3) ‘Micro-level’ – conducting a fast sampling at every 1.5 minutes of about 1.0 L (accomplished manually).

In total a volume of 0.8 L was needed to analyse traditional sewage quality parameters (chemical oxygen demand (COD) of unfiltered sample, COD of filtered sample (filter pore size 0.45μm), total Kjeldahl nitrogen, PO4-P, total suspended solids, total organic matter) and 0.2 L to analyse the antibiotics, respectively. Generally, substance characteristics, effect specifics and corresponding popularity of antibiotic agents vary to a great deal. To cover at least a part of this variety, the following antibiotics were chosen to be analysed: clarithromycin (CLA), ciprofloxacin (CIP) and doxycyclin (DOX). Main characteristics of these agents are listed in Table 1; effect mechanisms and environmental relevance are described as follows.

Table 1

Characteristics of CLA, CIP and DOX

Item Clarithromycin Ciprofloxacin Doxycyclin 
Spectrum efficacy Mainly Gram-positive, partly Gram-negative Mainly Gram-negative, partly Gram-positive Gram-positive as well Gram-negative, without cell wall 
Enterococcus faecalis 0.15 mg L−1 (EC50a) (Hanisch et al. 20044.0 mg L−1 (MIC90b) (Stille et al. 20064.0 mg L−1 (MIC90) (Ross et al. 2004), 2.0 to 8.0 mg L−1 (MIC90) (CLSI 2007
Escherichia coli resistant (Keller 19910.06 mg L−1 (MIC90) (Stille et al. 20060.5 to 2.0 mg L−1 (MIC90) (CLSI 2007
Item Clarithromycin Ciprofloxacin Doxycyclin 
Spectrum efficacy Mainly Gram-positive, partly Gram-negative Mainly Gram-negative, partly Gram-positive Gram-positive as well Gram-negative, without cell wall 
Enterococcus faecalis 0.15 mg L−1 (EC50a) (Hanisch et al. 20044.0 mg L−1 (MIC90b) (Stille et al. 20064.0 mg L−1 (MIC90) (Ross et al. 2004), 2.0 to 8.0 mg L−1 (MIC90) (CLSI 2007
Escherichia coli resistant (Keller 19910.06 mg L−1 (MIC90) (Stille et al. 20060.5 to 2.0 mg L−1 (MIC90) (CLSI 2007

aEC50: 50% effect concentration level.

bMIC90: 90% minimum inhibition concentration.

CLA is a macrolide which inhibits the protein biosynthesis and is commonly used as a pill or oral suspension. Due to its high toxicity, the concentration to inhibit reproduction or respiration of the microorganisms is in the range of mg L−1 (Kummerer et al. 2004). Further environmentally relevant characteristics are a good solubility in water on the one hand but a low degradability on the other hand. In contrast to CLA, CIP is a fluoroquinolone inhibiting the enzyme gyrase. Its maximum solubility in water is reached at pH 4 until pH 5. The minimum is reached at pH 7, but the solubility again increases as the pH rises above 7, which may be particularly relevant in the case of sewage discharging into receiving waters. Further, CIP has a very low degradability and is still active after 42 days under 37 °C (Mawhinney et al. 1992). The antibiotic DOX, which is a tetracycline inhibiting, like CLA, protein biosynthesis, is easily soluble in water and has a low degradability.

Analysing antibiotics

Collected samples were analysed by the Institute of Clinical Pharmacology, Faculty of Medicine Carl Gustav Carus, Technische Universität Dresden by spiking the sample with 0.8 mg Na2EDTA mL−1, and filtering the sewage water with a filter made of glass (<0.9 μm). Subsequently, the pH value was reduced to 3.5 using formic acid and then the solution was mixed using the standard addition method with the antibiotics of interest. Afterwards, these antibiotics were analysed following the method of solid phase extraction with an HLB cartridge (1 cc 30 mg, Water Oasis, USA), and thereafter the samples were passed through a liquid chromatography tandem mass spectrometry system (ABSciex 4000, USA) with the columns Synergi Hydro (Phenomenex, Germany) and Hilic (Machery-Nagel, Germany) (Rossmann et al. 2014).

Respiration rate

The method to assess the antibiotic resistance of the community matrix of bacteria is based on an inhibition test measuring the changing oxygen uptake rate (OUR) for increasing antibiotic concentrations, as shown in Equation (1). The OUR is here used to measure the extent of the activity of the bacterial community. In preparation, it is necessary to have an incubation time of 24 hours with continuous aeration at 20 °C and a pH between 6 and 8 before performing the OUR test, as microorganisms need time to metabolize the dosed antibiotic (Alexy 2003). The selected range of concentrations which is linked to toxic concentrations is listed in Table 1. In each case, a blind sample with the actual antibiotic agent concentration in sanitary sewage was analysed to compare the strength of the inhibition with the selected concentrations. The remaining wastewater samples were exposed to between 0.01 and 10 mg L−1 for 24 hours, with continuous aeration before performing the OUR test. For unlimited substrate supply the OUR test was spiked with ethanol. The OUR is calculated as: 
formula
1
where OUR is the observed rate of change of the dissolved oxygen concentration (mg O2 L−1 h−1, cO2,1 and cO2,2 are the initial and final dissolved oxygen concentration respectively, and Δt is the time interval (t2t1) in which the corresponding concentration change of dissolved oxygen has been observed.

RESULTS AND DISCUSSION

Concentration range of antibiotics and typical sewage parameters

Macro level

Figure 1 shows the concentrations of CLA, CIP and DOX determined in 24-hour-mixed sewage samples. The antibiotic agents CLA and CIP were detected across all samples with concentration ranges of 129–1,598 ng L−1 (median 287 ng L−1) and 591–2,151 ng L−1 (median 106.6 ng L−1). In contrast to this, DOX was detected just three times, whereas concentrations in remaining samples were below the detection limit.

Figure 1

Box and whisker plot of antibiotic concentration in sewage (left: 24-hour-mixed samples; middle: 1-hour-mixed samples; right: fast sampling every 2 minutes).

Figure 1

Box and whisker plot of antibiotic concentration in sewage (left: 24-hour-mixed samples; middle: 1-hour-mixed samples; right: fast sampling every 2 minutes).

Meso level

The antibiotic agents CLA and CIP were detected in all samples. As shown in Figure 1, the CLA is in the range of 25–1,500 ng L−1 (median 413 ng L−1), and CIP is in the range of 481–2,500 ng L−1 (median 790.0 ng L−1). DOX was not detected during the sampling period at all. The minor variation of the analysed concentrations between 24- and 1-hour-mixed samples confirms the investigations of Coutu et al. (2013), who analysed the inflow of a WWTP with 40 times more population equivalent than in the district analysed in this study.

Micro level

According to Figure 1, the concentration varies enormously during a few minutes, e.g. CLA from 0 to about 4,600 ng L−1, CIP from 600 to about 3,700 ng L−1 and DOX from 400 to 1,500 ng L−1. This variation was caused by the small number of people using antibiotics, which results in peaks at the moment when the excretion of the antibiotic load reached the sampling point. Furthermore, the dispersion influences peak dynamics; e.g. a high distance between the sampling point and the spot of the excretion results in smoothing the peaks.

Oxygen uptake rate

Clarithromycin

The origin background concentration of CLA in the sewage used for OUR test was between 129 and 1,600 ng L−1 (see Figure 1). Surprisingly, the OUR test with 0.01 mg L−1, which is about 10–100 times higher than the background content, did not show any inhibition effect. The median of the inhibition rate of the higher antibiotic concentrations was 22% (0.1 mg L−1), 41% (1.0 mg L−1) and 48% (10.0 mg L−1) – see Figure 2. These results confirm the EC50 results of between 10 and 100 mg L−1 of the investigations by Alexy (2003). In this case, an increased concentration of CLA would not result in higher inhibition rates. Furthermore, the range of the inhibition rate (1st to 99th percentile) reflects the relation of Gram-positive to Gram-negative bacteria within the tested samples, because CLA mainly affects Gram-positive bacteria and therefore the inhibition rate decreases with a lack of these bacteria. Additionally, it is important to emphasize the fact that a resistance against a macrolide like CLA results in a resistance among all macrolides, called cross-resistance, because of a modification of the ribosomal enzyme system (Hircin 2013). Therefore, it is assumed that the results are equivalent for the other macrolides like erythromycin, roxithromycin and azithromycin, but final evidence should be obtained through further investigations.

Figure 2

Relative OUR with different concentration ranges of CLA, CIP or DOX.

Figure 2

Relative OUR with different concentration ranges of CLA, CIP or DOX.

Ciprofloxacin

In the case of CIP, the background concentration in raw sewage varies between 600 and 2,200 ng L−1 (see Figure 1). The OUR test clarifies the sub-inhibitory concentration range in sewage up to 1.0 mg L−1, which is about 10–100 times higher than the background contents (see Figure 2). The median inhibition rate is about 91% in the case of 10.0 mg L−1. Further investigations with increasing concentration of CIP did not result in higher inhibition rates. In conclusion, the investigations represent the successful impact of CIP of both sides of the spectrum. In detail, Gram-positive and Gram-negative bacteria are mainly inhibited, resulting in low ranges between the 1st and 99th percentile (see Figure 2). These results reflect the high efficiency of CIP within an aerobic and not acidic milieu (Burgis 2005), whereas pH was between 5.6 and 5.8 during the respiration test.

Doxycyclin

In the case of DOX, the background concentration in the sewage was several times lower than the detection limit and just once about 500 ng L−1 (see Figure 1). The inhibition test with concentrations of DOX up to 10.0 mg L−1, which is over 1,000 times higher than the background contents, leads to a median OUR inhibition rate of about 38% (see Figure 2). This result reflects the positive effect of maintaining a resistance despite absence of the antibiotic agent. These results clearly confirm the cut-off point between 2 and 8 mg L−1 (CLSI 2007), which is obviously reached with the represented concentration of DOX between 1.0 and 10.0 mg L−1.

SUMMARY AND CONCLUSIONS

Observations in the present study show that excreted antibiotics occur in inhomogeneous ‘cloud-like’ concentration patterns in the sewer system owing to the fact that just a minimal number of connected inhabitants excrete these drugs. These phenomena are observable despite homogenizing effects like dispersion, advection, diffusion, degradation and even sorption. Estimating the in-sewer load of the drug which is recently prescribed in this district would require a more detailed look at the characteristics of the antibiotic agents and the dynamics of the drainage process during dry and wet weather periods. This is accomplished in a follow-up study and will be discussed in a future contribution. The analysis shows that the OUR test allows a robust and straightforward quantification of inhibitory effects in the microbial community due to antibiotic agents in raw sanitary sewage. The results reveal that the inhibitory effect level depends on:

  • (1) the antibiotic agent itself; and

  • (2) the composition of the bacterial community (ratio of Gram-negative to Gram-positive bacteria).

It is documented that the ‘no effect concentration’ of the community within the sewage is in the lower range of mg L−1, and in the case of CLA the inhibition of bacteria could be assumed for all other macrolides (cross-resistance). Considering the majority of active heterotrophic bacteria which are not inhibited by the background concentration at the sewer system, previous studies show that microorganisms develop resistance. Of course the presented OUR test is just a phenotypic test which cannot discriminate between a pure tolerance due to protecting extracellular polymeric substance or other mechanisms and a real-genetic resistance. Still, the inhibition rate as one effect caused by antibiotics is similar to previous investigations of biological analyses, e.g. the standard deviation of 20% in the case of EC50 value (Nusch 1995). For research into the discrimination between tolerance and resistance, the authors will collaborate with molecular biologists to better understand the governing processes.

ACKNOWLEDGEMENTS

The authors would like to gratefully thank Reinhard Oertel and Sara Schubert from Institute of Clinical Pharmacology, Faculty of Medicine Carl Gustav Carus, Technische Universität Dresden for laboratory analysis. The authors gratefully acknowledge the German Federal Ministry of Education and Research (BMBF) for funding the project of ANTI-Resist: 02WRS1272A. Jin Zhang also gratefully acknowledges the state-sponsored scholarship program (No. 2010605021) provided by the China Scholarship Council (CSC).

REFERENCES

REFERENCES
Alexy
R.
2003
Antibiotika in der aquatischen Umwelt: Eintrag, Elimination und Wirkung auf Bakterien. (Antibiotics in the Aquatic Environment: Input, Elimination and Effects on Bacteria). Dissertation, Albert-Ludwigs-Universität Freiburg. http://www.freidok.uni-freiburg.de/volltexte/660/ (accessed 19 December 2014)
.
Burgis
E.
2005
Intensivkurs Allgemeine und Spezielle Pharmakologie. (Crash Course General and Specific Pharmacology)
.
Elsevier, Munich, Germany
.
CLSI
2007
Performance Standards for Antimicrobial Susceptibility Testing. 17th Information Supplement. Clinical and Laboratory Standards Institute, Wayne, Pennsylvania, USA
.
Coutu
S.
Wyrsch
V.
Wynn
H. K.
Rossi
L.
Barry
D. A.
2013
Temporal dynamics of antibiotics in wastewater treatment plant influent
.
Science of the Total Environment
458–460
,
20
26
.
Feldmann
D.
2005
Modellberechnungen zum Verhalten und Verbleib von Arzneimittelrückständen im Krankenhausabwasser und Beurteilungsmöglichkeiten ihres ökotoxikologischen Gefährdungspotentials. (Model-Based Calculations on the Behavior of Pharmaceutical Residues in Hospital Wastewater and the Evaluation of their Ecotoxicological Risk Potential,) Dissertation
,
Technische Universität Berlin
,
Berlin, Germany
.
Feuerpfeil
I.
Lopez-Pila
J.
Schmidt
R.
Schneider
E.
Szewzyk
R.
1999
Antibiotikaresistente Bakterien und Antibiotika in der Umwelt (Antibiotic-resistent bacteria and antibiotics in the environment)
.
Bundesgesundhbl. – Gesundheitsforsch. – Gesundheitsschutz
42
,
37
50
.
Hanisch
B.
Abbas
B.
Kratz
W.
Schueuermann
G.
2004
Human drugs in aquatic ecosystems: approach for the environmental risk assessment of drug residues
.
Umweltwissenschaften und Schadstoff-Forschung
16
(
4
),
223
238
.
Hircin
E.
2013
Macrolide antibiotic. http://flexikon.doccheck.com/de/Makrolid (accessed 13 September 2013)
.
Hoffmann
L. R.
D'Argenio
D. A.
MacCoss
M. J.
Zhang
Z.
Jones
R. A.
Miller
S. I.
2005
Aminoglycoside antibiotics induce bacterial biofilm formation
.
Nature
436
,
1171
1175
.
Keller
M.
1991
Clarithromycin. Pharma-kritik
13
(
19
),
73
76
.
Kummerer
K.
Alexy
R.
Huttig
J.
Scholl
A.
2004
Standardized tests fail to assess the effects of antibiotics on environmental bacteria
.
Water Research
38
(
8
),
2111
2116
.
Mawhinney
W. M.
Adair
C. G.
Gorman
S. P.
McClurg
B.
1992
Stability of ciprofloxacin in peritoneal-dialysis solution
.
American Journal of Hospital Pharmacy
49
(
12
),
2956
2959
.
Miller
C.
Thomsen
L. E.
Gaggero
C.
Mosseri
R.
Ingmer
H.
Cohen
S. N.
2004
SOS Response induction by beta-lactams and bacterial defense against antibiotic lethality
.
Science
305
,
1629
1631
.
Morell
V.
1997
Antibiotic resistance: road of no return
.
Science
278
,
575
576
.
Nusch
E.
1995
Möglichkeiten und Grenzen der Aussagefähigkeit konventioneller Biotestverfahren (Possibilities and limitations of the significance of conventional bioassays). Umwelttechnologieforum UTECH95, Berlin, Germany
.
Ross
J. E.
Jones
R. N.
& Quality Control Working Group
2004
MIC quality control guidelines for doxycycline when testing Gram-positive control strains by the reference methods
.
Diagnostic Microbiology and Infectious Disease
50
(
4
),
295
297
.
Rossmann
J.
Schubert
S.
Gurke
R.
Oertel
R.
Kirch
W.
2014
Simultaneous determination of most prescribed antibiotics in multiple wastewater by SPE-LC-MS/MS
.
Journal of Chromatography B
969
,
162
170
.
Schrag
S. J.
Perrot
V.
Levin
B. R.
1997
Adaption to the fitness costs of antibiotic resistance in Escherichia coli
.
Proceedings of Royal Society of London. Series B
264
,
1287
1291
.
Stille
W.
Brod
H-R.
Groll
A. H.
Just-Nübling
G.
2006
Antibiotika-Therapie: Klinik und Praxis der antiinfektiösen Behandlung (Antibiotic Therapy: Clinic and Practice of Anti-Infective Treatment)
.
Schattauer Verlag, Stuttgart, Germany
.