Chemosphere 122 (2015) 176–182

Contents lists available at ScienceDirect

Chemosphere journal homepage: www.elsevier.com/locate/chemosphere

Transfer of oxytetracycline from swine manure to three different aquatic plants: Implications for human exposure Maliwan Boonsaner a,⇑, Darryl W. Hawker b a b

Department of Environmental Science, Faculty of Science, Silpakorn University, Nakhon Pathom 73000, Thailand School of Environment, Griffith University, Nathan, Qld 4111, Australia

h i g h l i g h t s  Shows for the first time transfer of OTC from swine manure to aquatic plants.  Isotherms for desorption of OTC from swine manure constructed.  Comparative bioconcentration of OTC by three contrasting aquatic plants measured.  Per capita daily OTC uptake rates from aquatic plant consumption derived.  ADI comparison shows pathway should not be ignored in determining human exposure.

a r t i c l e

i n f o

Article history: Received 28 October 2013 Received in revised form 18 November 2014 Accepted 22 November 2014 Available online 12 December 2014 Handling Editor: J. de Boer Keywords: Swine manure Oxytetracycline Desorption Bioconcentration Aquatic plants Human exposure

a b s t r a c t Little is known regarding the potential for pharmaceuticals including antibiotics to be accumulated in edible aquatic plants and enter the human food chain. This work investigates the transfer of a widely used veterinary antibiotic, oxytetracycline (OTC), from swine manure to aquatic plants by firstly characterizing desorption from swine manure to water and fitting data to both nonlinear and linear isotherms. Bioconcentration of OTC from water was then quantified with aquatic plants of contrasting morphology and growth habit viz. watermeal (Wolffia globosa Hartog and Plas), cabomba (Cabomba caroliniana A. Gray) and water spinach (Ipomoea aquatica Forsk.). Watermeal and water spinach are widely consumed in Southeast Asia. The OTC desorption and bioconcentration data were used to provide the first quantitative estimates of human exposure to OTC from a manure-water-aquatic plant route. Results show that under certain conditions (plants growing for 15 d in undiluted swine manure effluent (2% w/v solids) and an initial OTC swine manure concentration of 43 mg kg1 (dry weight)), this pathway could provide a significant fraction (>48%) of the acceptable daily intake (ADI) for OTC. While effluent dilution, lower OTC manure concentrations and not all plant material consumed being contaminated would be expected to diminish the proportion of the ADI accumulated, uptake from aquatic plants should not be ignored when determining human exposure to antibiotics such as OTC. Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction Oxytetracycline (OTC) is a member of the tetracycline family of antibiotics and widely used in swine farming (Matsui et al., 2008). Its poor absorption on oral administration results in the contamination of swine manure by unmetabolized material (Winckler and Grafe, 2001). Kumar et al. (2005a) report antibiotic concentrations in animal manures to range from trace levels up to 200 mg kg1. Liquid effluent from swine farms often contains suspended manure. While limited treatment of the effluent may take place at some facilities, OTC can still be discharged into ⇑ Corresponding author. Tel.: +66 3 4245330; fax: +66 3 4245331. E-mail address: [email protected] (M. Boonsaner). http://dx.doi.org/10.1016/j.chemosphere.2014.11.045 0045-6535/Ó 2014 Elsevier Ltd. All rights reserved.

receiving water (Xuan et al., 2010). In the environment, factors such as temperature, pH and light intensity can affect the degradation of OTC in water (Burhenne et al., 1997) and swine manure (Ratasuk et al., 2012). Several studies have investigated the potential for tetracyclines to be taken up from the interstitial water of soil by plants. For example, Kumar et al. (2005b) reported that corn (Zea mays L.), green onion (Allium cepa L.), and cabbage (Brassica oleracea L.) could accumulate chlortetracycline while Boonsaner and Hawker (2010, 2012) found that OTC was uptaken by soybean (Glycine max (L.) Merr.) and rice (Oryza sativa L.). Concern has been expressed that residual pharmaceuticals such as antibiotics can be taken up by edible plants and enter the human food supply (Tanoue et al., 2012). Potential adverse effects

M. Boonsaner, D.W. Hawker / Chemosphere 122 (2015) 176–182

on humans due to accumulation of OTC in excess of the acceptable daily intake (ADI) include hepatotoxicity, disruption of digestive system functioning and development and spread of antibioticresistant bacteria (Dolliver et al., 2007; Boonsaner and Hawker, 2013). While maximum residue levels for antibiotics including OTC have been established for animal-based products there are no such data for plants (Kim et al., 2011). One reason for this is that relatively little is known of the potential and extent of uptake by plants (Cropp et al., 2010). Work by Boxall et al. (2006) showed that a terrestrial uptake route for crops via interstitial water in soil and manure-amended soil for veterinary medicines could account for up to 10% of the ADI for some compounds. The situation with aquatic plants is however very poorly characterized. The overall aim of this current work then was to assess the extent to which transfer of OTC from swine manure to aquatic plants occurs and the implications of this for human exposure. Manure-water transfer was investigated by using spiked swine manure to determine desorption isotherms. Plant bioconcentration from water was quantified using three aquatic plants with contrasting morphologies and growth habits. Aquatic plants or macrophytes may be classified as floating, submerged or emergent (Zhao et al., 2012). Such plants can uptake chemicals such as tetracyclines directly from water via root cells or other immersed tissues. Watermeal (Wolffia globosa Hartog and Plas) is a free-floating rootless aquatic plant with no distinct stems and leaves (Rahman and Hasegawa, 2011). Cabomba (Cabomba caroliniana A. Gray) is a submerged plant native to South America and introduced to many Asian, Pacific and European countries where it has become an invasive weed (Wilson et al., 2007). Water spinach (Ipomoea aquatica Forsk.), is regarded as an emergent aquatic plant (Ko et al., 2011). It is a plant rooting at the nodes with hollow stems, 2–3 m or longer, that can float. Water spinach and watermeal in particular are widely consumed by Asian people (Zarcinas et al., 2004; Marcussen et al., 2008). Importantly in the context of this current work, both are often grown in polluted rivers and canals in Southeast Asia before harvesting (Söderström and Bergqvist, 2003). Exposure to OTC desorbed from swine manure could occur under these conditions. By combining results from swine manure desorption and plant bioconcentration, the implications for exposure of humans to OTC via this route can be assessed by comparing derived daily intake rates to the ADI. A secondary aim with regard to bioconcentration was to investigate the extent of any translocation with water spinach since it is the upper part of the stem with foliage that is actually consumed (Marcussen et al., 2009). Translocation may be defined as the movement of a chemical from one plant part to another. It is not appropriate to refer to translocation with watermeal and cabomba. In the former, there is no differentiation into separated plant parts. In the later, the whole plant is submerged and has the same exposure to a dissolved chemical in a homogeneous aquatic system. 2. Materials and methods 2.1. Reagents OTC (as the hydrochloride (OTC.HCl)) (>98.5% purity) for use as a standard for HPLC analysis was purchased from Dr. Ehrenstorfer GmbH, Augsburg, Germany. Commercial grade OTC.HCl used in experiments was obtained from Nova Medicine, Bangkok, Thailand and checked for impurities prior to use. Details of all other reagents used are found in Boonsaner and Hawker (2013). 2.2. Analytical protocols for manure, water and plant samples The concentrations of OTC in manure and plant samples (sample mass approximately 10 g) were determined by adding 25 mL

177

of McIlvaine buffer-EDTA solution to the sample and following the method described in Boonsaner and Hawker (2010). Briefly, samples were blended with an Ultra-Turrax homogenizer, then placed in an ultrasonic bath for 2 min and finally filtered through a glass fiber filter. The filtrate was passed through a conditioned SPE cartridge and OTC eluted with 2 mL of methanol. Resulting solutions were analyzed by HPLC (Waters 600, Milford, MA, USA) with Photodiode Array Detection. Analytical details are found in Boonsaner and Hawker (2013). Water samples (100 mL) were passed through a conditioned SPE cartridge, eluted with 2 mL of methanol and analyzed by HPLC using the same conditions as for manure and plant sample extracts. Method detection limits for the analysis of OTC in plants and manure were 0.45 and 0.7 mg kg1 dry weight (dw) respectively and 0.5 mg L1 for water. The mean recoveries from water, plant and swine manure samples were 97%, 70% and 65% respectively. All concentrations reported herein account for these recoveries. Analyses of selected metallic elements in swine manure were carried out by Inductively Coupled Plasma-Optical Emission Spectrometry (ICP-OES). An Agilent 710 instrument was used, with a radio frequency generator power setting of 1.0 kW. Plasma gas (Argon) and auxiliary gas flow rates were 15 and 1.5 L min1 respectively. Yttrium (371.029 nm) was used as the internal standard and analytical emission lines were 259.940 nm (Fe), 285.213 nm (Mg), 396.152 nm (Al) and 422.673 nm (Ca). The organic carbon content of manure samples was determined by the Walkley–Black wet combustion method (Tan, 1996). 2.3. Desorption of OTC from swine manure Swine manure free from antibiotic contamination was obtained from a local farm in Nakhon Pathom province, central Thailand. The manure was sun-dried for 2 days and sieved to 2 mm size. Prior to use, the manure was autoclaved twice at 121 °C and 1.03  105 Pa for 15 min to minimize any microbial activity that may degrade the OTC during the timeframe of the experiments (Ratasuk et al., 2012). OTC was spiked into swine manure by adding an OTC solution in acetone to 10 g of manure to make nominal initial concentrations of 50, 100 and 200 mg OTC kg1 of swine manure. After acetone was removed, actual initial concentrations were determined to be 43, 83 and 203 mg OTC kg1 (dw) of swine manure. Then, 500 mL of sterile water was added to glass jars containing contaminated swine manure to make a total solids content of 20 000 mg L1 or 2% (w/v). Fulhage and Pfost (2001) note that swine lagoon effluent can have a solids content up to 2% (w/v). Nine glass jars for each OTC concentration in swine manure were prepared. All glass jars were covered in aluminum foil and placed in a dark room (28 ± 2 °C). On days 0, 1, 3, 5, 7, 9, 11, 13 and 15, one jar of each test concentration was taken and the water filtered. Duplicate 100 mL samples of filtered water from each test concentration were then analyzed for OTC by SPE and HPLC as described above. After Loke et al. (2002), sorbed OTC amounts at these times were calculated by difference between the initial amount in the system and amounts in the water. To establish the extent of OTC degradation under these conditions, a parallel control experiment was conducted comprising an aqueous OTC solution (40 mg L1) in the absence of swine manure. 2.4. Bioconcentration of OTC by aquatic plants Watermeal, cabomba and water spinach were purchased from a local market in Nakhon Pathom province, central Thailand. Before commencing the experiments, they were analyzed by the methods described above and determined to be OTC free.

178

M. Boonsaner, D.W. Hawker / Chemosphere 122 (2015) 176–182

Previous work with watermeal had established that at OTC concentrations of 50 mg L1 or more toxic effects occurred (Boonsaner and Hawker, 2013). Therefore, concentrations of 30 mg L1 and less were chosen for bioconcentration experiments in this investigation. Experiments were conducted with three initial OTC aqueous concentrations, nominally 10, 20 and 30 mg L1. The actual initial concentrations in these experiments were 8.27, 18.45 and 28.8 mg L1. Deionized water used in these studies was autoclaved for 30 min in order to minimize subsequent microbial degradation of added OTC. For each concentration, 500 mL of test solution (pH 4.4–4.6) was added to each of nine clear glass jars (800 mL capacity) followed by 20 g (whole weight or wet weight (ww)) of watermeal. This amount of watermeal covered the surface of the water in a jar. The external sidewall of the jar was covered by aluminum foil to reduce degradation from illumination. For cabomba, approximately 20 g (ww) of plant sample per glass jar was submerged in 500 mL of test solution and the external sidewall of each jar covered with aluminum foil. In the experiments with water spinach, each jar contained of approximately 60 g (ww) of plant material (root mass was 8.5–12.5 g (ww) for each plant sample) and 500 mL of test solution. Roots were submerged and jars were covered with aluminum foil leaving only the shoot exposed to artificial light (5000 lux). The daily illumination regime was 10 h light: 14 h darkness. The nine jars (for each OTC concentration) were placed in a room (28 ± 2 °C) and on days 0, 1, 3, 5, 7, 9, 11, 13 and 15 of exposure, one jar of each concentration was taken for OTC analysis. Plant samples were removed from water at this time and in the case of watermeal, it was filtered from water and separated for analysis. For water spinach, plant samples were separated into root and shoot tissue and root tissue analyzed. On days 3, 9 and 15, shoot samples were also analyzed to characterize the extent of any translocation. The 15 d bioconcentration period was chosen to minimize any OTC degradation in water and also enable comparison with previous work (Boonsaner and Hawker, 2012, 2013). The filtered water was thoroughly stirred before collecting 100 mL samples for analysis. Duplicate samples were analyzed for OTC in both plant tissue and water concentrations at all time periods. The plant concentration factor (CF) was obtained from the ratio of the test compound concentration in whole plant (CPLANT) to the concentration found in the water. For water spinach, root concentration factors (RCF) and shoot concentration factors (SCF) are defined as the ratios of the test compound concentrations in root and shoot tissues respectively, to the concentrations found in the water. CF values for whole water spinach plants at given exposure times were derived from values of RCF and SCF at these times and the masses of root and shoot tissues. A control experiment was performed by placing each plant type in clean, sterilized water at pH = 4.5 and locating jars in the same room as the exposed jars. Following the 15-day experimental period, plant samples and water samples from the control jar were taken for OTC analysis. These showed no contamination and plants appeared normal. 2.5. Model description and data analysis for OTC desorption from swine manure

Here, ksorp and kdesorp are first order sorption and desorption rate constants (time1) respectively. Assuming a mass balance exists in the system at any time such that the initial amount of OTC in spiked swine manure (Mo) is equal to MSM + MW, these equations integrate to:



    ksorp Mo ksorp M o þ Mo  eðksorp þkdesorp Þt ksorp þ kdesorp ðksorp þ kdesorp Þ

ð3Þ

   kdesorp ½Mo  1  eðksorp þkdesorp Þt ksorp þ kdesorp

ð4Þ

MSM ¼  MW ¼

It follows that the percentage (of the initial mass) that is desorbed at any time may then be expressed as:

  M SM MW ¼ 100 % Desorbed ¼ 100 1  Mo Mo     kdesorp 1  eðksorp þkdesorp Þt ¼ 100 ðksorp þ kdesorp Þ

It is noted that expressed in this manner, the percentage desorbed is independent of the starting amount in the system (Mo). From Eqs. (3) and (4) above, an infinite time period is required to achieve equilibrium masses of OTC in swine manure (MSM(1)) and water (MW(1)). If however, effective equilibrium is defined to be 0.99MW(1), a finite time period for OTC equilibrium in the swine manure-water system can be derived (Hawker and Connell, 1985).

MW ¼ MW1 ð1  eðksorp þkdesorp Þt Þ ¼ 0:99M Wð1Þ )t¼

dM SM ¼ ksorp M W  kdesorp MSM dt

ð1Þ

dM W ¼ kdesorp M SM  ksorp MW dt

ð2Þ

ln 0:01 4:605 ¼ ðksorp þ kdesorp Þ ðksorp þ kdesorp Þ

ð6Þ

The equilibrium masses of OTC in the swine manure (MSM(1)) and water (MW(1)) themselves may also be derived from Eqs. (3) and (4):

MSMð1Þ ¼

ksorp Mo ; ðksorp þ kdesorp Þ

M Wð1Þ ¼

kdesorp Mo ðksorp þ kdesorp Þ

ð7Þ

These can be converted into equilibrium concentrations using the mass of swine manure (0.010 kg) and the volume of water (0.5 L) employed in desorption experiments. Linear, Freundlich and Langmuir sorption isotherms were constructed using this equilibrium concentration data. For linear isotherms or Nernst partitioning:

KD ¼

C SMð1Þ C Wð1Þ

ð8Þ

where KD (L kg1) is the (linear) sorption coefficient. The expression for the empirical Freundlich isotherm is:

C SM ¼ K F C nWð1Þ

ð9Þ

where KF is the Freundlich sorption coefficient ((mg kg1) (mg L1)n) and n the nonlinearity parameter while the Langmuir isotherm is described by:

C SMð1Þ ¼

Desorption of OTC was considered using a one-compartment swine manure model. On this basis, the changes in mass of OTC in the swine manure (MSM) and water (MW) with time are given by:

ð5Þ

QK L C W1 ð1 þ K L C Wð1Þ Þ

ð10Þ

Here Q is the maximum sorption capacity (mg kg1) of the sorbent and KL the Langmuir sorption coefficient (L mg1) (Loke et al., 2002). Data were plotted and statistical results (regression, t-test and ANOVA) obtained using GraphPad Prism Version 5.0c (GraphPad Software, San Diego, USA, 2009). Octanol/water distribution ratio (DOW) values for OTC were calculated using Advanced Chemistry Development Software V11.02 (ACD/Labs, Toronto, Ontario, Canada) and collected from the SciFinder ScholarÒ 2014 database.

179

M. Boonsaner, D.W. Hawker / Chemosphere 122 (2015) 176–182

3. Results and discussion 3.1. Desorption of OTC from swine manure From Eq. (5), plots of the percentage of the original amount of OTC spiked into swine manure that is desorbed with time are shown in Fig. 1. Goodness of fit data contained therein (the coefficient of determination (R2) and the standard error of the estimate (Sy.x)) show that a single homogenous compartment for swine manure and simple first order sorption and desorption rate constants are an adequate description of the system. Further evidence is that the mean (± standard error) of the derived equilibrium ‘‘% Desorbed’’ values from Fig. 1 is 50.4 ± 4.98% (n = 3). This is in agreement with the prediction from Eq. (5) that the proportion of OTC desorbed is independent of the starting amount in the system, which varies by a factor of almost five. A basic tenet of the derivation is the operation of a mass balance in the swine manure-water system. These desorption results therefore confirm control experiment data that showed minimal OTC degradation (t1/2 = 39 d) within the timespan of the experiments. Times to equilibrium under the conditions employed (9.76 ± 1.36 d) can also be derived from this approach. The standard error (1.36 d or 13.9% of the mean) is relatively low considering the variation of OTC mass in the system. This is consistent with Eq. (6) that shows equilibrium time to be a function of rate constants only, and not OTC amounts. Equilibrium OTC concentrations derived as described above (Eq. (6) et seq.) were employed to construct Linear, Freundlich and Langmuir isotherms (Fig. S1, Supplementary Material). Isotherm constants and goodness of fit data are shown in Table 1. In this work, manure-water equilibrium for OTC was determined from desorption of spiked manure for the first time. This is appropriate since unmetabolized OTC often enters the aquatic environment initially associated with swine manure (Pan et al.,

2011). Moreover, the occurrence of dissolved OTC in surface waters suggests that sorption to sorbents is not irreversible (Xu and Li, 2010). Equilibrium can also be reached from sorption of aqueous OTC by uncontaminated swine manure (Loke et al., 2002). Equilibrium manure and aqueous concentrations for OTC and other antibiotics derived in this manner have previously been fit to various types of sorption/desorption isotherms. For example, sorption of tylosin, a macrocyclic lactone antibiotic, as well as OTC have been fit to Freundlich isotherms (Kolz et al., 2005; Loke et al., 2002). Langmuir isotherms have been fitted to data from OTC sorption to marine sediments (Xu and Li, 2010). Although isotherms are typically nonlinear, KD values from linear isotherms have also been reported. Loke et al. (2002) found a KD value of 77.6 L kg1 for sorption of OTC by uncontaminated swine manure, compared with 38.3 L kg1 derived in this work from desorption of spiked manure. Such discrepancies have been attributed to sorption/desorption hysteresis resulting from irreversible sorption (Xu and Li, 2010). It is not possible to conclude hysteresis exists from the present comparison however because of the variability in composition of swine manures employed. The primary mechanisms for sorption of OTC to manure are acting as a chelating agent and binding to divalent metal cations as well as partitioning into organic matter (Jones et al., 2005). The organic carbon content in the swine manure employed by Loke et al. (2002) was 42% but elemental composition was unspecified. In this current work, organic carbon content was 33%, while ICP-OES analysis found 0.24% w/w Al, 2.58% w/w Ca, 0.20%w/w Fe, and 1.42% Mg. The multiple mechanisms of OTC sorption with environmental sorbents including manure mean that the extent of sorption can’t be reliably predicted from compound lipophilicity as expressed by the octanol/water distribution ratio (DOW) and the organic carbon content of the sorbent as for many other organic sorbates (Jones et al., 2005). KD values are larger than might be expected on the basis of a fractional DOW value (102.19 at pH = 4.5)). This means experimental determinations such as those carried out in this work are necessary. 3.2. Bioconcentration of OTC by plants

Fig. 1. A plot of the percentage of the original amount of OTC spiked into swine that is desorbed with time from Eq. (5) for the three different initial amounts of OTC, together with goodness of fit statistics. Data are presented as the mean and standard error of the mean.

The uptake into plants may be an important exposure pathway of OTC to humans and other biota. The results from this study showed that all plant species investigated accumulated OTC from ambient water. The concentration accumulated depends on species, exposure time and ambient OTC levels. CF values are only a function of species and exposure time however. Fig. 2 depicts mean CF data at each exposure time, together with the standard error of these means. They are based on aggregating duplicate CF values obtained from each of the three aqueous concentration treatments. (Initial aqueous concentrations were 8.27, 18.45 and 28.8 mg L1). The averages of the standard errors of the mean for each exposure time, expressed as a percentage, are 9.8% for watermeal, 14.4% for cabomba and 11.2% for water spinach roots. For comparative purposes, CF data are also based on those plant parts actually in contact with water. Hence whole plant CF values are shown for watermeal and cabomba and root (RCF) data for water spinach. Furthermore, CF values are also expressed on a

Table 1 Isotherm constants and goodness of fit data for desorption of OTC from swine manure. Standard errors of all constants are shown in parentheses. Linear isotherm

Freundlich isotherm

Langmuir isotherm

KD (L kg1)

r2

KF (mg kg1) (mg L1)n

n

R2

Q (mg kg1)

KL (L mg1)

R2

38.3 (4.35)

0.90

49.8 (0.798)

0.614 (0.020)

0.99

149 (6.09)

0.543 (0.042)

0.99

180

M. Boonsaner, D.W. Hawker / Chemosphere 122 (2015) 176–182

Fig. 2. Concentration factors (L g1) for plant parts in contact with water with exposure time. For watermeal and cabomba, this is the whole plant. For water spinach it is the root only. Data are expressed as the mean and standard error of the mean (n = 6) of duplicate plant concentrations (dw) from the three different aqueous concentration treatments.

dry weight (dw) basis for these plant tissues. The magnitude of the CF values increase with time, reaching values of 2–3 L g1 after 15 d exposure. There is no discrimination between plant CF values for short exposure times (1 and 3 d), but after this watermeal  water spinach > cabomba (ANOVA, Tukey’s post hoc test, P < 0.05). Comparable CF data were obtained from separate investigations of bioconcentration of OTC by watermeal (Boonsaner and Hawker, 2013) and rice roots (Boonsaner and Hawker, 2012) after a similar 15 d exposure. 3.3. Translocation of OTC by water spinach Translocation of OTC within the 15 d timeframe of the experiments was observed for water spinach. Plant tissue was subdivided into root and shoot tissues and concentration factor values for root (RCF), shoot (SCF) and whole plant (CF) with time are shown in Fig. 3. Again, for a given tissue type and exposure time, data from different aqueous concentration treatments are aggregated. RCF, SCF and CF values increase with time. SCF values are less than RCF data at equivalent exposure times (t-test; P < 0.05) with the difference increasing with time. Appreciable translocation of OTC in plants is relatively unusual but is significant with regard to this work because it is the upper plant parts that are widely used as a vegetable in South East Asia. Water spinach has also been shown to accumulate and translocate metals including cadmium (Marcussen et al., 2008). The reasons for the translocation of OTC are unclear however. Translocation has not been observed with rice and soybean over 15 d for example (Boonsaner and Hawker, 2010, 2012). Lack of chemical translocation with these plants has been attributed to relatively hydrophilic compounds such as OTC being unable to traverse hydrophobic membranes in the root structure (Tanoue et al., 2012). There is little evidence however that the structure of water spinach roots is markedly different to that of rice and soybean in this regard. 3.4. Implications for human exposure Desorption of antibiotics from swine manure subsequent uptake and translocation by edible water spinach represents a poorly characterized for humans. To quantify this process for OTC,

into water and plants such as exposure route from the mass

Fig. 3. A plot of concentration factors (L g1) with time for the whole water spinach plant, root tissues and shoot tissues. Data are expressed as the mean and standard error of the mean (n = 6) from duplicate plant tissue concentrations (dw) from the three different aqueous concentration treatments.

balance for the swine manure/water system described earlier where Mo is the initial OTC mass in swine manure, the equilibrium OTC concentration in water is given by:

Mo ¼ C SMð1Þ MassSM þ C Wð1Þ V W ¼ C Wð1Þ K D MassSM þ C Wð1Þ V W Mo Co ) C Wð1Þ ¼ ¼ ðK D MassSM þ V W Þ ½K D þ ðV W =MassSM Þ

ð11Þ

The equilibrium concentration in water (mg L1) is a function of the initial concentration in swine manure (CO mg kg1 (dw)), swine manure/water sorption coefficient (KD L kg1) and VW/MassSM (L kg1), the reciprocal of the solids content in the system. If bioconcentration by plants does not deplete the equilibrium aqueous OTC concentration greatly, antibiotic levels in plant tissue (CPLANT mg kg1 (dw)) can be related to initial concentration in swine manure (CO) by:

C PLANT ðdwÞ ¼ 103  CF



 Co K D þ ðV W =MassSM Þ

ð12Þ

This expression brings together the sorption/desorption and plant bioconcentration aspects of the current study. For human consumption of watermeal and cabomba, the relevant plant concentrations are whole plant but for water spinach they are shoot concentrations. The factor 103 is necessary for unit conversion. In their work on uptake of veterinary medicines from contaminated OTC soils via interstitial water into plants, Boxall et al. (2006) noted that based on data from the WHO Global Environment Monitoring System – Food Contamination Monitoring and Assessment Program (GEMS/Food), an adult consumes 0.512 kg of aboveground plant material daily. Using measured OTC concentrations in lettuce as a surrogate for above-ground crops and an acceptable daily intake (ADI) of 30 lg kg1 d1, they found uptake from this route would contribute 0.06% of the ADI for OTC. Considering a range of veterinary medicines, a maximum contribution approaching 10% was found for the antibiotic trimethoprim. It was acknowledged that the results of this derivation were exaggerated due to the assumption of all vegetable material being contaminated. Nonetheless, it provided amongst the first quantitative estimates of the contribution of plant-based antibiotics to human exposure. OTC concentrations in watermeal, cabomba and water spinach were derived using Eq. (12). A range of initial OTC swine manure concentrations (43, 83 and 203 mg kg1 (dw)) were employed in

M. Boonsaner, D.W. Hawker / Chemosphere 122 (2015) 176–182

this work so as to construct sorption/desorption isotherms. While manure concentrations in the upper part of this range have been found in some studies (e.g. Pan et al., 2011), levels are often in the lower region (Martínez-Carballo et al., 2007; Dolliver et al., 2007). Therefore the minimum initial OTC swine manure concentration of 43 mg kg1 (dw), a solids content of 20 000 mg L1 or 0.02 kg L1, a KD value of 38.3 L kg1 and 15 d concentration factor values for plants and plant tissues were used to calculate plant concentrations. Values of CPLANT so obtained were 1330, 970 and 760 mg kg1 (dw) for watermeal, cabomba and water spinach shoots respectively. Using tissue water contents of 91.5% (Naskar et al., 1986), 93% (Hasan and Chakrabarti, 2009) and 88% (Chu et al., 1980), wet weight OTC concentrations are 110, 68 and 91 mg kg1 respectively. Water spinach is extensively cultivated in peri-urban regions of South-East Asia and an important staple food in these areas (Marcussen et al., 2008). Daily human uptake rates of OTC from aquatic plant consumption and a comparison with the ADI may be obtained with the same general approach as used by Boxall et al. (2006), but with a specific human consumption rate for water spinach. Marcussen et al. (2008) quote a per capita water spinach consumption rate of 77.3 g d1 in peri-urban Hanoi, Vietnam. Comparable data for Cambodia were 30 and 21 g d1 for vegetable and non-vegetable farmers respectively (Marcussen et al., 2009). Thus, multiplying a conservative consumption rate of 21 g d1 by the plant concentrations derived above affords uptake rates for a 70 kg adult of 34.0, 20.3 and 27.3 lg kg1 d1 with watermeal, cabomba and water spinach shoots respectively. This treatment assumes daily consumption rates of watermeal and cabomba are the same as for water spinach since there is little specific information on these. As an example, detailed calculation of uptake rates for OTC from water spinach shoots is found in Text S1 in Supplementary Material. As calculated, it can be seen per capita daily uptake rate of OTC from water spinach is 91% of the ADI. Results for watermeal and cabomba are 113% and 67% respectively. These results are based on a linear manure-water sorption coefficient which simplifies the relationship between CW(1) and C0 (Eq. (11)). Using nonlinear isotherms, nonlinear relationships between CW(1) and C0 are obtained that cannot be solved analytically for CW(1). If solved iteratively however, based on a Freundlich isotherm (Eq. (9) and Table 1), proportions of the ADI for OTC from water spinach, watermeal and cabomba consumption are 63%, 80% and 48% respectively. Using a Langmuir isotherm (Eq. (10) and Table 1), the proportions are 68%, 85% and 51% respectively. Overall, regardless of the isotherm employed, a significant proportion of the ADI could be acquired from consumption of aquatic plants under certain circumstances. These data provide the first quantitative estimates of human exposure to an important antibiotic, OTC, from a manure-wateraquatic plant route. However there are some important caveats. The first is that swine manure concentrations could be lower than considered here. Secondly, the aqueous medium in which the plants grow in these calculations represents undiluted effluent (2% w/v solids) from a swine farm. To the extent that dilution occurs, which is difficult to predict, the fraction of the ADI uptaken by humans is reduced. For example, if the concentration of OTC in the water to which plants were exposed were reduced by a factor of 100 as a result of dilution, consumption of water spinach would result in levels of OTC uptaken approaching 1% of the ADI. Thirdly, calculations are based on all plant material consumed being contaminated. Finally, 15 d plant concentration factors are employed. With longer timespans, the known lability of OTC toward light and hydrolysis (Pouliquen et al., 2007; Xuan et al., 2010) for example would likely mean decreased aqueous concentrations and hence decreased plant concentrations.

181

Comparisons with exposure from a contaminated soil-interstitial water-plant uptake route (Boxall et al., 2006) show higher concentrations in aquatic plants in the present work. Differences between plant species employed notwithstanding, manureamended soil would be expected to exhibit lower concentrations than manure itself as a result of dilution, translating into lower plant concentrations. 4. Conclusion OTC can be transferred from swine manure entrained in water to aquatic plants, some of which may be used for human consumption. Separate experiments characterized desorption from swine manure into water and bioconcentration from water into aquatic plants. Equilibrium desorption data can be fit to various isotherms. The three different aquatic plants whose behavior was investigated all bioconcentrated OTC from water with plant concentration factors increasing with time and reaching 2–3 L g1 after 15 d exposure. Translocation was observed with water spinach in contrast to previous work with rice and soybean that showed no appreciable translocation over similar timespans and under similar conditions. In assessing implications for human exposure, for an initial swine manure concentration of 43 mg kg1 (dw) and a relatively conservative consumption of water spinach (21 g d1), this pathway could result in uptake of approximately 65–90% of ADI for this plant grown in undiluted swine manure effluent. Results for the other two aquatic plants investigated were also significant under these conditions (>48% ADI). Factors such as lower manure concentrations, effluent dilution and not all plant material consumed being contaminated would diminish the proportion of ADI accumulated. Results show however that this uptake route should not be ignored when determining human exposure to antibiotics such as OTC. Acknowledgements This study was supported by the Research Fund of the Faculty of Science, Silpakorn University, Thailand. The authors would like to thank Miss C. Teprak for laboratory assistance and S. Jaipan for plant and manure preparations. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.chemosphere. 2014.11.045. References Boonsaner, M., Hawker, D.W., 2010. Accumulation of oxytetracycline and norfloxacin from saline soil by soybeans. Sci. Total Environ. 408, 1731–1737. Boonsaner, M., Hawker, D.W., 2012. Investigation of the mechanism of uptake and accumulation of zwitterionic tetracyclines by rice (Oryza sativa L.). Ecotoxicol. Environ. Safe. 78, 142–147. Boonsaner, M., Hawker, D.W., 2013. Evaluation of food chain transfer of the antibiotic oxytetracycline and human risk assessment. Chemosphere 93, 1009– 1014. Boxall, A.B.A., Johnson, P., Smith, E.J., Sinclair, C.J., Stutt, E., Levy, L.S., 2006. Uptake of veterinary medicines from soils into plants. J. Agric. Food Chem. 54, 2288–2297. Burhenne, J., Ludwig, M., Nikoloudis, P., Spiteller, M., 1997. Primary photoproducts and half-lives. Environ. Sci. Pollut. Res. Int. 4 (1), 10–15. Chu, Y.-H., Chang, C.-L., Hsu, H.-F., 1980. Flavonoid content of several vegetables and their antioxidant activity. J. Sci. Food Agric. 80 (5), 561–566. Cropp, R.A., Hawker, D.W., Boonsaner, M., 2010. Predicting the accumulation of organic contaminants by plants. Bull. Environ. Contam. Toxicol. 85, 525–529. Dolliver, H., Kumar, K., Gupta, S., 2007. Sulfamethazine uptake by plants from manure-amended soil. J. Environ. Qual. 36, 1224–1230. Fulhage, C.D., Pfost, D., 2001. Swine Manure Management Systems in Missouri. University of Missouri MU Extension, Columbia, Mo. (accessed 24.10.14).

182

M. Boonsaner, D.W. Hawker / Chemosphere 122 (2015) 176–182

Hasan, M., Chakrabarti, R., 2009. Use of Algae and Aquatic Macrophytes as Feed in Small-scale Aquaculture. A Review. FAO Fisheries and Aquaculture Technical Paper No. 531, FAO, Rome. 123p. Hawker, D.W., Connell, D.W., 1985. Prediction of bioconcentration factors under non-equilibrium conditions. Chemosphere 14, 1835–1843. Jones, A.D., Bruland, G.L., Agrawal, S.G., Vasudevan, D., 2005. Factors influencing the sorption of oxytetracycline to soils. Environ. Toxicol. Chem. 24 (4), 761–770. Kim, K.-R., Owens, G., Kwon, S.-I., So, K.-H., Lee, D.-B., Ok, Y.S., 2011. Occurrence and environmental fate of veterinary antibiotics in the terrestrial environment. Water Air Soil Pollut. 214, 163–174. Ko, C.-H., Lee, T.-M., Chang, F.-C., Liao, S.-P., 2011. The correlations between system treatment efficiencies and aboveground emergent macrophyte nutrient removal for the Hsin-Hai Bridge phase II constructed wetland. Bioresource Technol. 102, 5431–5437. Kolz, A.C., Ong, S.K., Moorman, T.B., 2005. Sorption of tylosin onto swine manure. Chemosphere 60 (2), 284–289. Kumar, K., Gupta, S.C., Chander, Y., Singh, A.K., 2005a. Antibiotic use in agriculture and its impact on the terrestrial environment. Adv. Agron. 87, 1–54. Kumar, K., Gupta, S.C., Baidoo, S.K., Chander, Y., Rosen, C.J., 2005b. Antibiotic uptake by plants from soil fertilized with animal manure. J. Environ. Qual. 34 (6), 2082– 2085. Loke, M.-L., Jette Tjornelund, J., Halling-Sørensen, B., 2002. Determination of the distribution coefficient (log Kd) of oxytetracycline, tylosin A, olaquindox and metronidazole in manure. Chemosphere 48, 351–361. Marcussen, H., Joergensen, K., Holm, P.E., Brocca, D., Simmons, R.W., Dalsgaard, A., 2008. Element contents and food safety of water spinach (Ipomoea aquatica Forssk.) cultivated with wastewater in Hanoi, Vietnam. Environ. Monit. Assess. 139, 77–91. Marcussen, H., Dalsgaard, A., Holm, P.E., 2009. Element concentrations in water spinach (Ipomoea aquatic Forssk.), fish and sediment from a wetland production system that receives wastewater from Phnom Penh, Cambodia. J. Environ. Sci. Health A 44, 67–77. Martínez-Carballo, E., González-Barreiro, C., Scharf, S., Gans, O., 2007. Environmental monitoring study of selected veterinary antibiotics in animal manure and soils in Austria. Environ. Pollut. 148, 570–579. Matsui, Y., Ozu, T., Inoue, T., Matsushita, T., 2008. Occurrence of a veterinary antibiotic in streams in a small catchment area with livestock farms. Desalination 226 (1–3), 215–221.

Naskar, K., Banerjee, A.C., Chakraborty, N.M., Ghosh, A., 1986. Yield of Wolffia arrhiza (L.) Horkel ex Wimmer from cement cisterns with different sewage concentrations, and its efficacy as a carp feed. Aquaculture 51 (3–4), 211–216. Pan, X., Qiang, Z., Ben, W., Chen, M., 2011. Residual veterinary antibiotics in swine manure from concentrated animal feeding operations in Shandong Province, China. Chemosphere 84, 695–700. Pouliquen, H., Delépée, R., Larhantec-Verdier, M., Morvan, M.-L., Le Bris, H., 2007. Comparative hydrolysis and photolysis of four antibacterial agents (oxytetracycline, oxolinic acid, flumequine and florfenicol) in deionised water, freshwater and seawater under abiotic conditions. Aquaculture 262 (1), 23–28. Rahman, M.A., Hasegawa, H., 2011. Aquatic arsenic: phytoremediation using floating macrophytes. Chemosphere 83, 633–646. Ratasuk, N., Boonsaner, M., Hawker, D.W., 2012. Effect of temperature, pH and illumination on abiotic degradation of oxytetracycline in sterilized swine manure. J. Environ. Sci. Health A 47, 1687–1694. Söderström, H.S., Bergqvist, P.-A., 2003. Polycyclic aromatic hydrocarbons in a semiaquatic plant and semipermeable membrane devices exposed to air in Thailand. Environ. Sci. Technol. 37, 47–52. Tan, K.H., 1996. Soil Sampling, Preparation and Analysis. Marcel Dekker Inc., Hong Kong. 1408p. Tanoue, R., Sato, Y., Motoyama, M., Nakagawa, S., Shinohara, R., Nomiyama, K., 2012. Plant uptake of pharmaceutical chemicals detected in recycled organic manure and reclaimed wastewater. J. Agric. Food Chem. 60, 10203–10211. Wilson, C.E., Darbyshire, S.J., Jones, R., 2007. The biology of invasive alien plants in Canada. 7. Cabomba caroliniana A. Gray. Can. J. Plant Sci. 87, 615–638. Winckler, C., Grafe, A., 2001. Use of veterinary drugs in intensive animal production: evidence for persistence of tetracycline in swine slurry. J. Soil Sed. 1, 66–70. Xu, X.-R., Li, X.-Y., 2010. Sorption and desorption of antibiotic tetracycline on marine sediment. Chemosphere 78, 430–436. Xuan, R., Arisi, L., Wang, Q., Yates, S.R., Biswas, K.C., 2010. Hydrolysis and photolysis of oxytetracycline in aqueous solution. J. Environ. Sci. Health B 45, 73–81. Zarcinas, B.A., Ishak, C.H., McLaughlin, M.J., Cozens, G., 2004. Heavy metals in soils and crops in Southeast Asia. 2. Thailand. Environ. Geochem. Health 26, 343– 357. Zhao, D., Jiang, H., Yang, T., Cai, Y., Xu, D., An, S., 2012. Remote sensing of aquatic vegetation distribution in Taihu Lake using an improved classification tree with modified thresholds. J. Environ. Manage. 95, 98–107.