Environmental Research 150 (2016) 482–488
Contents lists available at ScienceDirect
Environmental Research journal homepage: www.elsevier.com/locate/envres
Oral relative bioavailability of Dichlorodiphenyltrichloroethane (DDT) in contaminated soil and its prediction using in vitro strategies for exposure refinement Albert L. Juhasz a,n, Paul Herde b, Euan Smith a a b
Future Industries Institute, University of South Australia, Mawson Lakes, SA 5095, Australia South Australian Health and Medical Research Institute, Gilles Plains, SA 5086, Australia
art ic l e i nf o
a b s t r a c t
Article history: Received 12 April 2016 Received in revised form 7 June 2016 Accepted 25 June 2016 Available online 14 July 2016
In this study, the bioavailability of DDTr (sum of DDT, DDD and DDE isomers) in pesticide-contaminated soil was assessed using an in vivo mouse model. DDTr relative bioavailability (RBA) ranged from 18.77 0.9 (As35) to 60.87 7.8% (As36) indicating that a significant portion of soil-bound DDTr was not available for absorption following ingestion. When DDTr bioaccessibility was assessed using the organic Physiologically Based Extraction Test (org-PBET), the inclusion of a sorption sink (silicone cord) enhanced DDTr desorption by up to 20-fold (1.6–3.8% versus 18.9–56.3%) compared to DDTr partitioning into gastrointestinal fluid alone. Enhanced desorption occurred as a result of the silicone cord acting as a reservoir for solubilized DDTr to partition into, thereby creating a flux for further desorption until equilibrium was achieved. When the relationship between in vivo and in vitro data was assessed, a strong correlation was observed between the mouse bioassay and the org-PBET þsilicone cord (slope ¼ 0.94, y-intercept¼ 3.5, r2 ¼0.72) suggesting that the in vitro approach may provide a robust surrogate measure for the prediction of DDTr RBA in contaminated soil. & 2016 Elsevier Inc. All rights reserved.
Keywords: Bioaccessibility Correlation DDT Relative bioavailability Sorption Sink
1. Introduction Dichlorodiphenyltrichloroethane (DDT) is a persistent, hydrophobic organic compound historically utilised for the control of malaria and typhus and as an agricultural insecticide. Although highly effective at eradicating malaria, DDT was banned in the US in the early 1970s as a consequence of its potential adverse effects on human (e.g. endocrine disruption) and environmental health (e.g. eggshell thinning, biomagnification) (Beard, 2006; Guo et al., 2009; Mitra et al., 2011; Wu et al., 2012; Gerber et al., 2016). In 2004, the Stockholm Convention on Persistent Organic Pollutants was ratified by over 170 countries which restricted DDT usage to vector control. Although agricultural use of DDT may continue in some countries, DDT may also be present at elevated concentrations in urban, agricultural or industrial soil as a consequence of its historical use and DDT's recalcitrant nature. In addition to p,p′ and o,p′ DDT isomers, contaminated soil may also contain other related organochlorine compounds including p,p′ and o,p′ isomers of dichlorodiphenyldichloroethane (DDD) and dichlorodiphenyldichloroethylene (DDE). These compounds may be present as a n
Corresponding author. E-mail address:
[email protected] (A.L. Juhasz).
http://dx.doi.org/10.1016/j.envres.2016.06.039 0013-9351/& 2016 Elsevier Inc. All rights reserved.
consequence of impurities during the manufacturing process or as a result of (a)biotic degradation processes (Foght et al., 2001; Bosch et al., 2015; Qu et al., 2015; Wang et al., 2015). However, due to their structural and chemical relatedness, DDTr is often used to refer to the sum of all DDT related compound. With an increase in urbanisation of former agricultural land, people may be exposed to DDTr from historically contaminated soil via incidental ingestion. Children are particularly at risk from this exposure pathway due to the prevalence of hand-to-mouth contact. However, as a consequence of hydrophobicity (log organic carbon partitioning coefficients [Koc] of 4.7–5.4), DDTr may strongly sorb to soil organic matter (SOM) (Hilber et al., 2008; Bielská and Hofman, 2015; Li et al., 2016) which will influence its release from the soil matrix under gastrointestinal conditions. Subsequently, exposure may be significantly reduced as a result of DDTr bioavailability limitations. In a recent study utilising an in vivo mouse model, Smith et al. (2012) determined that DDTr relative bioavailability (RBA) in soil from former sheep and cattle dip sites ranged from 3 to 25% (n ¼7). Similarly, Li et al. (2016) determined that DDTr RBA in soil from agricultural (sheep and cattle dip sites) and pesticide manufacturing locations was significantly less than 100% (18–65%; n¼ 6) as a consequence of the limited release of DDTr from SOM, resulting from black carbon's
A.L. Juhasz et al. / Environmental Research 150 (2016) 482–488
high affinity for DDTr. Although limited studies have been undertaken to assess DDTr RBA, data to date suggests that DDTr exposure may be overestimated if total soil-bound DDTr concentrations are utilised for risk assessment purposes. While in vivo assays are an appropriate approach for the determination of DDTr RBA for the refinement of human health exposure, their routine use is impractical due to cost and ethical considerations. However, an alternative approach for refining exposure is the use of in vitro gastrointestinal extraction methodologies (i.e. bioaccessibility assays) which are simple, rapid and inexpensive compared to in vivo bioassays. In order to utilise the bioaccessibility approach for predicting contaminant RBA, a strong relationship between in vivo and in vitro data is required (as in the case for arsenic and lead) (Smith et al., 2011; Juhasz et al., 2014; 2015; Diamond et al., 2016). However, the in vivo-in vitro relationship for DDTr is limited due to the dearth of studies investigating DDTr RBA and limitations (i.e. solubility constraints) associated with traditional approaches for the assessment of hydrophobic organic contaminant (HOC) bioaccessibility. Recently, a number of researchers have demonstrated the potential of a sorption sink approach for measuring HOC bioaccessibility (James et al., 2011; Gouliarmou et al., 2013; Li et al., 2015; Zhang et al., 2015; Li et al., 2016; Juhasz et al., 2016). Gastrointestinal solutions are supplemented with a sink (e.g. poly[dimethylsiloxane], Tenax TA, C18 solid phase extraction membranes) which simulates the passive diffusion of HOCs across the small intestines. Absorption of HOCs form the in vitro solution into the sink creates a concentration gradient which facilitates the mobilisation of desorbable HOCs from the soil matrix. This approach has been shown to significantly enhance HOC bioaccessibility from soil and soot matrices compared to its assessment by partitioning into gastrointestinal fluid (Li et al., 2015; Zhang et al., 2015; Li et al., 2016; Juhasz et al., 2016). The objective of this study was to assess DDTr RBA in a variety of DDTr-contaminated soils using an in vivo mouse model and to determine the suitability of a sorption sink bioaccessibility approach for predicting DDTr RBA. As the desorbable fraction is considered to be the maximum fraction available for absorption across the intestinal epithelium (Oomen et al., 2003; Gouliarmou et al., 2013; Zhang et al., 2015), it was hypothesised that the assessment of DDTr bioaccessibility using the sorption sink approach may provide a conservative estimate of DDTr RBA.
2. Materials and methods 2.1. DDTr-contaminated soil DDTr-contaminated soil was collected from former dip sites in north eastern New South Wales. At these sites, DDT was applied to control cattle and sheep ticks from 1957 to 1967. Although DDT usage ceased decades ago, elevated concentrations of the organochlorine pesticide still persists in surrounding soil due to the recalcitrant nature of DDT (McDougall, 1997; Menchai et al., 2008). Soil was air dried then sieved with the o250 mm soil particle size fraction retained for characterisation, bioaccessibility and RBA analysis. Soil physicochemical properties were determined in duplicate for each soil. Soil organic carbon content was determined by oxidation / combustion (Nelson and Sommers, 1996) while DDTr concentration was determined by GC-ECD following extracs tion using a Dionex ASE 200 (see DDT Extraction and Quantification). Table 1 outlines selected physico-chemical properties for soil used in this study.
483
Table 1 Properties of DDTr-contaminated soils used in this study. Characterisation was undertaken on the o 250 mm soil particle size fraction. Soil
As28 As30 As32 As35 As36 As37 As38 As46 a
Organochlorine (mg kg 1) DDT
DDD
DDE
DDTr
990 5920 475 7080 707 17,710 532 18,837
104 663 50 639 80 1328 69 1787
221 516 53 275 98 356 64 787
13137 36 71007 750 5787 73 7990 7206 885 7110 19,390 7 440 6657 43 21,380 7 1395
pHa
TOC (%)a
As (mg kg 1)a
6.0 5.9 6.0 5.0 5.7 4.6 5.3 4.7
5.5 3.1 5.1 5.8 1.2 6.8 5.2 5.3
1458 482 1346 1835 713 3605 315 1465
Data represents the mean of duplicate analysis. Values varied by less than 5%.
2.2. Assessment of DDTr relative bioavailability DDTr RBA was assessed using female Balb/c mice weighing 22–24 g as detailed by Smith et al. (2012). Previously, Smith et al. (2012) determined a linear relationship between DDT administered and DDTr accumulation in liver, kidney and adipose tissue. Although higher concentrations of DDTr accumulated in adipose tissue, utilising this endpoint for DDTr RBA offers challenges with respect to sample collection and extraction/clean up prior to analysis. As a consequence, liver and kidneys were utilised for the assessment of DDTr RBA in this study. Animal care was compliant with the Standard Operating Procedures of the South Australian Health and Medical Research Institute and the Guidelines for the Care and Use of Laboratory Animals (NRC, 1996). Mice were acclimatised to animal housed conditions (12/ 12 light/dark cycles with water supplied ad libitum) prior to the assessment of DDTr RBA. DDTr-contaminated soil was administered to mice following incorporation of soil in standard mouse chow (Glen Forest Specialty Feeds, Australia) to achieve dosing concentrations of 10–1250 mg DDT g BW day1. Mice (n ¼4 per soil) were exposed to DDTr containing mouse chow for 7 days after which animals were humanly killed by cervical dislocation and the kidneys and liver collected for DDTr determination. Kidneys and liver were frozen on collection, following by freeze drying prior to accelerated solvent extraction using an ASE200 (Dionex). DDTr RBA was calculated by comparing the kidney/liver concentration of DDTr in soil and reference spiked sand dosed mice following dose normalisation as detailed in Eq. (1). ⎛ Kidney/LiverDDTr⎡⎣ soil⎤⎦ D⎡⎣ spiked sand⎤⎦ ⎞ ⎟ × 100 DDTrRBA( %) = ⎜ × ⎜ Kidney/LiverDDTr⎡ D⎡⎣ soil⎤⎦ ⎟⎠ ⎣ spiked sand⎤⎦ ⎝
(1)
where: Kidney/Liver DDTr [soil] ¼ DDTr (mg) accumulated in the kidneys and liver following oral administration of DDTr-contaminated soil. Kidney/Liver DDTr [spiked sand] ¼DDTr (mg) accumulated in the kidneys and liver following oral administration of DDT spiked sand. D [spiked sand] ¼Dose of DDT administered in DDT spiked sand (mg kg 1). D [soil] ¼Dose of DDTr administered in DDTr-contaminated soil (mg kg 1). 2.3. Assessment of DDTr bioaccessibility DDTr bioaccessibility was assessed using the organic Physiologically Based Extraction Test (org-PBET) (Ruby et al., 2002) with and without the inclusion of a sorption sink; silicone cord (poly
484
A.L. Juhasz et al. / Environmental Research 150 (2016) 482–488
[dimethylsiloxane]). Previous studies determined that the bioaccessibility of hydrophobic contaminants may be underestimated without the inclusion of a sorption sink as a result of the limited capacity of the in vitro fluid (solubility constraints) to accommodate desorbable contaminants (James et al., 2011; Gouliarmou et al., 2013; Li et al., 2015; Zhang et al., 2015; Juhasz et al., 2016; Li et al., 2016). Silicone cord (CBC, Adelaide; diameter of 3 mm, 8 g m 1) was selected as the sorption sink as it has previously been shown to be compatible for the assessment of hydrophobic contaminant bioaccessibility (Gouliarmou et al., 2013; Juhasz et al., 2016). Silicone cord was prepared by soaking overnight in ethyl acetate followed by three overnight soaks in methanol, three overnight soaks in acetone and five overnight soaks in MilliQ water (Gouliarmou et al., 2013; Juhasz et al., 2016). This washing protocol was followed in order to remove organic constituents from the silicone matrix that may potentially interfere during DDTr analysis. The gastric phase of the org-PBET comprised glycine (15.0 g l 1), sodium chloride (8.8 g l 1), pepsin (1.0 g l 1), bovine serum albumin (5.0 g l 1) and mucin (2.5 g l 1) adjusted to pH to 1.5 70.05 using concentrated HCl. Intestinal phase conditions were achieved by adding porcine pancreatin (0.6 g l 1) and bovine bile (4.0 g l 1) to gastric solution and adjusting the pH to 7.270.2 using sodium hydroxide (50% w/v). Prior to the assessment of DDTr bioaccessibility in contaminated soil, the rate, efficacy and capacity of the silicone cord to recover DDT from org-PBET solution was assessed using p,p′ DDT spiked sand (100–10,000 mg kg 1). p,p′ DDT-spiked sand (0.8 g) was added to Pyrex schott bottles (100 ml) containing 8 g (1 m) of silicone cord and 80 ml of org-PBET gastric phase solution and incubated at 37 °C /40 rpm on a Ratek suspension mixer. After gastric phase extraction (1 h), replicate flasks were sacrificed and the silicone cord removed for DDTr quantification. For additional flasks, gastric phase conditions were modified to the intestinal phase and incubation continued with replicate flasks removed at selected time points (2, 4, 8, 16 and 24 h) for DDT quantification in silicone cord. When DDTr bioaccessibility was assessed in contaminated soils (n ¼7), 0.8 g of soil, 80 ml of org-PBET solution and a silicone mass of approximately 100 times that of the soil organic matter content were combined. Following a 1 h gastric phase, in vitro conditions were modified to the intestinal phase and incubation continued for a further 22 h. At the end of the incubation period, the silicone cord was removed, rinsed with MilliQ water and gently wiped with lint free tissue prior to DDTr extraction. Following DDTr quantification in silicone cord, DDTr bioaccessibility was calculated according to Eq. (2).
surrogate during GC-ECD analysis. Initially, the oven temperature was held at 100 °C for 0.5 min followed by a linear increase to 200 °C at 25 °C min 1, holding for 5.6 min. The temperature was then increased to 240 °C at 6 °C min 1 and finally ramped to 285 °C at 20 °C min 1, holding at this temperature for 5 min. Injector and detector temperatures were maintained at 240 and 300 °C, respectively. Recovery of surrogates during DDTr quantification averaged 90.1 715.3% for soil and silicone cord extracts and 95.7 712.8% for tissue extracts. Sample blanks, duplicate analysis and spiked sample recoveries were included in the analysis to ensure quality assurance/quality control (QA/QC). Recovery of PCNB ranged between 83–102% of the original spike solution. The average deviation between duplicate samples (n ¼6) was 5.1% while the average recovery from spiked samples (n¼ 6) was 96.2% (90.2–106.0%).
⎛ DDTr[μg ]in silicone cord ⎞ DDTr bioaccessibility( %) = ⎜ ⎟ × 100 DDTr[μg ]in soil ⎝ ⎠
The RBA of DDTr in contaminated soil was assessed using an in vivo mouse assay. Previously, Smith et al. (2012) utilised DDTr accumulation in mouse adipose tissue, kidneys or liver as the endpoint for bioavailability assessment. Although the dose-response for DDTr accumulation in each biological endpoint was linear, Smith et al. (2012) observed some variability (although not significantly different; p4 0.05) in calculated DDTr RBA values when the three endpoints were compared. In order to increase the robustness of DDTr RBA determinations, DDTr accumulation in the present study was determined in combined kidney and liver samples (i.e. to increase sample size). Adipose tissue was omitted as a bioavailability endpoint due to the requirement for a more complex post-extraction clean up procedure prior to GC-ECD analysis. As detailed in Appendix A, dose-response studies determined a linear relationship between DDT administered and DDTr accumulation in combined kidney and liver samples (Fig. A1). The dose-response relationship was utilised in order to calculate DDTr RBA (see Eq. (1)) following administration of DDTr-contaminated soil to mice and quantification of DDTr accumulation in kidney and
(2)
2.4. DDTr extraction and quantification DDTr was extracted from contaminated soil and mouse kidneys and liver using an accelerated solvent extractor (ASE200; Accelerated Solvent Extraction System, Dionex Pty Ltd, Lane Cove, NSW, Australia) according to Smith et al. (2012). Briefly, freeze dried material was ground, spiked with surrogate (100 ml pentachloronitrobenzene [PCNB]; 100 mg ml 1) prior to extraction using hexane: acetone (1:1 v/v) and U. S. EPA method 3545 (Dionex Application note 320) (USEPA, 2007a). DDTr was extracted from silicone cord using acetone according to Gouliarmou et al. (2013). DDTr was quantified using an Agilent Technologies 7890 A gas chromatograph equipped with an electron capture detector and a Zebron ZB-1701 P column (30 m 0.25 mm 0.25 mm; Phenomenex Australia, Lane Cove, NSW, Australia) using USEPA method 8081B (USEPA, 2007b). Dibromo-DDE was added as a
3. Results and discussion 3.1. Soil characterisation Table 1 shows the concentration of DDTr in contaminated soil (n¼ 8) which were utilised for the assessment of DDTr RBA and bioaccessibility. The concentration of DDTr ranged from 5787 73 (As32) to 21,380 71395 mg kg 1 (As46) with p,p′ DDT being the predominant organochlorine compound (75–91% of the total DDTr concentration). p,p′ DDD and p,p′ DDE were also present in all soils, comprising 6.8–10.4% and 1.8–16.8% of the total DDTr concentration respectively. The varying proportions of p,p′ DDD and p, p′ DDE in contaminated soil may be reflective of impurities associated with dipping solutions and/or (a)biotic transformation processes. p,p′ DDT may be degraded via cometabolism under both aerobic (Aislabie et al., 1997) and anaerobic conditions (Wedemeyer, 1967; Plimmer et al., 1968), although the resulting transformation products (p,p′ DDE and p,p′ DDD respectively) may be more toxic or recalcitrant than DDT itself. In addition, Van Zwieten et al. (2003) suggested that the extent of p,p′ DDT transformation may be influenced by co-contaminants, particularly arsenic, which inhibit aerobic and anaerobic degradation of p,p′ DDT. Although a general trend of increasing DDT: DDD and DDT: DDE ratios were observed with increasing arsenic concentration (Table 1), ratios may also have been influenced by anthropogenic impacts such as dipping solution impurities. 3.2. DDTr relative bioavailability in contaminated soil
A.L. Juhasz et al. / Environmental Research 150 (2016) 482–488
485
via reductive dechlorination to p,p′ DDD (Esaac and Matsumura, 1980; Kitamura et al., 2002). p,p′ DDD may undergo further metabolism via dehydrochlorination resulting in the formation of p,p′ DDE. The decreasing proportion of DDT in soil, kidney and liver samples was not as pronounced when soils with 410,000 mg DDTr kg 1 were administered to mice presumably due to the combination of elevated absorbed DDT concentrations, rate of dechlorination/dehydrochlorinaiton and the transit time in the corresponding compartments. For example, the ratio of DDT:DDD: DDE in As46 was 88:8:4 whereas the ratio in kidney and liver samples was 57:29:14 and 60:21:19 respectively. 3.3. Sorption sink approach for assessing DDTr bioaccessibility
Fig. 1. DDTr relative bioavailability in contaminated soil. Values represent the mean and standard deviation of four replicates.
liver samples. DDTr RBA in contaminated soil ranged from 18.7 70.9 (As35) to 60.87 7.8% (As36) (Fig. 1). Although a limited number of soils have been assessed for DDTr RBA using an in vivo mouse model, the values determined were (on average) higher than those reported by Smith et al. (2012) (3–25%; n ¼7) and similar to those of Li et al. (2016) (18–65%; n¼6). Smith et al. (2012) determined DDTr RBA in soil from former cattle dip sites in northeastern New South Wales, Australia while Li et al. (2016), determined DDTr RBA in mice following exposure to soil from former pesticide factories (Chinese) and dip sites (Australia). Although similar approaches were used for the assessment of DDTr RBA (mouse model; contaminated soil incorporated into mouse chow; accumulation of DDTr as the endpoint for RBA determination), Li et al. (2016) utilised a shorter exposure period (4 days) compared to Smith et al. (2012) and the current study. While the differences in exposure periods may not influence DDTr RBA determinations, it may have an impact on the transformation of DDT (due to enzyme induction) and the proportion of DDD and DDE accumulated in target organs. In vivo transformation of p,p′ DDT was evident as indicated by increasing proportions of p,p′ DDD and p,p′ DDE in kidney and liver samples compared to administered soil (Fig. 2). For example, the ratio of DDT:DDD:DDE in As28 was 75:8:17 whereas the ratio in kidney and liver samples was 59:19:22 and 26:49:25 respectively. In mammals, p,p′ DDT is primarily transformed in the liver
Fig. 2. Proportion of p,p′ DDT (■), p,p′ DDD ( ) and p,p′ DDE (☐) in four representative soil samples prior to administration to mice and following the 10 day exposure period in kidney and liver samples.
A variety of in vitro methodologies have been utilised for the assessment of organic contaminant bioaccessibility. However, a constraint of some assays is the limited capacity of the in vitro solution to accommodate desorbable contaminants (Zhang et al., 2015; Juhasz et al., 2016). As DDT and related compounds exhibit low aqueous solubility, DDTr bioaccessibility may be underestimated if bioaccessibility evaluations are based solely on dissolved DDTr in gastrointestinal fluid (Smith et al., 2012; Li et al., 2016). Previously, the suitability of utilising a sorption sink approach (silicone cord) in org-PBET solution was demonstrated for measuring desorbable/bioaccessible PAHs in contaminated soil (Juhasz et al., 2016). This approach was re-evaluated to determine the rate, efficacy and capacity of the silicone cord to recover p,p′ DDT from org-PBET solution using p,p′ DDT spiked sand (100– 10,000 mg kg 1). The sand matrix was utilised to supply p,p′ DDT at elevated concentrations in a readily desorbable form as silica has limited capacity for retention due to the weak electrostatic interactions between p,p′ DDT and the matrix (Smith et al., 2012). Fig. 3 shows the sorption efficacy of silicone cord for p,p′ DDT when the organochlorine was supplied at 100, 1000 and 10,000 mg kg 1 in silica sand. Sorption efficacy was calculated by determining the mass of p,p′ DDT sorbed to the silicone cord at the corresponding time point divided by the initial mass supplied in the sand matrix. Following the initial 1 h gastric phase, the mass of p,p′ DDT that partitioned onto the silicone cord was negligible, however, when extraction conditions were modified to reflect intestinal phase conditions, the rate of partitioning was enhanced considerably. The enhancement was presumably due to the influence of bile and pancreatin which as a consequence of their
Fig. 3. Silicone cord sorption kinetics for DDT supplied at 100 (●), 1000 (☐) and 10,000 mg kg 1 (■) during gastrointestinal extraction of DDT-spiked sand using the organic Physiologically Based Extraction Test. After 1 h of gastric phase extraction, the in vitro solution was modified to reflect intestinal phase conditions and incubation continued for up to 24 h. Values represent the mean and standard deviation of triplicate analyses.
486
A.L. Juhasz et al. / Environmental Research 150 (2016) 482–488
surfactant like properties (Tang et al., 2006; Gorelick and Jamieson, 1994) increased the capacity of the gastrointestinal fluid to solubilise p,p′ DDT. For sand spiked with 100 and 1000 mg p,p′ DDT kg 1, 495% of the initial p,p′ DDT mass partitioned onto the silicone cord within 4 h (Fig. 3). However, for p,p′ DDT supplied at the upper concentration (10,000 mg kg 1), a 21 h intestinal phase was required for 95% of the initial mass to partition onto the sorption sink (linear sorption from 1 to 17 h of 550 mg h 1). For other organochlorine compounds (e.g. o,p′ DDT, p,p′ DDD, o,p′ DDD, p,p′ DDE, o,p′ DDE) sorption kinetics into in vitro sinks have been shown to be comparable to p,p′ DDT. Using Tenax (5 g l 1) as the sorption sink and the org-PBET spiked with organochlorine compounds (10 mg l 1), Li et al. (2016) determined that the time required for 99% of p,p′ DDT, o,p′ DDT, p,p′-DDD, o,p′ DDD, p,p′-DDE and o,p′ DDE to partition into the Tenax was 8.4–14.7 min. The consistency in sorption kinetics is not surprising given the similarity in organic carbon partitioning coefficients (ATSDR, 2002). Although the sorption sink approach has been adopted for the assessment of PAH bioaccessibility (Gouliarmou et al., 2013; Li et al., 2015; Zhang et al., 2015; 2016; Juhasz et al., 2016), limited research has been undertaken on other organic compounds such as p,p′ DDT. In a recent study, Li et al. (2016), utilised the PBET intestinal phase supplemented with Tenax as the sorption sink to assess p,p′ DDT sorption kinetics. When p,p′ DDT, p,p′ DDD and p,p′ DDE were supplied in the PBET intestinal phase at a concentration of 10 mg l 1, the time required to reach 99% of equilibrium was o15 min when 0.1 g of Tenax was utilised. A considerably longer timeframe was required for equilibrium to be reached in this study compared to Li et al. (2016) as a consequence of the significantly higher amount of p,p′ DDT supplied (80–8000 mg compared to 0.2 mg). Irrespective of the timeframe required to reach equilibrium, initial experiments demonstrated the capacity of the silicone cord to recover elevated concentrations of desorbable p,p′ DDT which may be encountered in highly contaminated soil at former dip sites. As a consequence, in order to measure the desorbable fraction (i.e. measure o 95% of equilibrium), an extended intestinal phase extraction (22 h) was employed as a conservative estimate of DDTr bioaccessibility due to the elevated DDTr concentration in As37 and As46. 3.4. DDTr bioaccessibility in contaminated soil DDTr bioaccessibility in contaminated soil was assessed using the org-PBET. Experiments were conducted with and without the incorporation of silicone cord in order to determine the influence of the sorption sink on DDTr desorption / bioaccessibility. As detailed in Gouliarmou et al. (2013), a silicone mass of approximately 100 times the soil organic matter content was included in order to provide sufficient sorptive capacity for desorbed DDTr relative to the soil. As illustrated in Fig. 4, DDTr bioaccessibility was enhanced significantly (up to 20-fold) when the sorption sink was included in the in vitro assay. In the absence of the sorption sink, DDTr bioaccessibility was low (1.6–3.8%) as a consequence of the limited capacity of the org-PBET gastrointestinal fluid to accommodate desorbable DDTr. However, enhanced DDTr desorption was facilitated through the incorporation of the silicone cord (18.9– 56.3%) which provided a reservoir for solubilized DDTr to partition into, thereby creating a flux for further desorption until equilibrium was achieved. Limited information is available regarding the bioaccessibility of DDTr in contaminated soil. In a previous study, Li et al. (2016) determined that the bioaccessibility of DDTr in soil from former pesticide facilities and dip sites (5.4–9254 mg DDTr kg 1; n¼ 6) was low when assessed using the org-PBET (1.2–15%). However, when Tenax was include in the assay as a sorption sink, DDTr bioaccessibility (27–56%) was enhanced 3.4–22 fold. Although lower DDTr bioaccessibility was observed in soil with
Fig. 4. DDTr bioaccessibility in contaminated soil determined using the organic Physiologically Based Extraction Test (org-PBET) (■) and the org-PBET incorporating silicone cord (☐). Values represent the mean and standard deviation of triplicate analyses.
higher TOC in the present study, the relationship was not statistically significant (r2 ¼ 0.13; Fig. A2). A similar non-significant relationship between TOC and bioaccessibility has been reported by a number of researchers for HOCs (Li et al., 2015; Juhasz et al., 2016; Li et al., 2016) whereby the poor correlation was influenced by the variability in TOC composition (e.g. amorphous versus black carbon). 3.5. Correlation between DDTr relative bioavailability and bioaccessibility In order to determine the suitability of the org-PBET to predict DDTr RBA, the relationship between in vivo and in vitro assessment methodologies was determined. Previous studies determined that a poor correlation was observed between DDTr RBA, determined using a mouse model, and DDTr bioaccessibility, determined using in vitro assays without the inclusion of a sorption sink (Smith et al., 2012; Li et al., 2016). As detailed in Li et al. (2016), the slope of the in vivo-in vitro relationship was 4 2 (i.e. DDTr RBA values were double bioaccessibility measurements), with a large y-intercept (þ 18.2) and a poor goodness of fit (r2 ¼0.36). The poor relationship resulted from low DDTr bioaccessibility as a consequence of the limited capacity of the in vitro solution to facilitate desorption of DDTr from the soil matrix (i.e. solubility constraints). Although DDTr bioaccessibility in soil utilised in this study was o4% when assessed using the org-PBET (Fig. 4), a strong relationship was observed between DDTr RBA and bioaccessibility (Fig. 5; r2 ¼0.89). However, the slope of the in vivo-in vitro correlation was 415 indicating that values determined by the org-PBET were over a magnitude lower that values determined using the in vivo mouse model. A slope of this magnitude may introduce significant variability in the predictive capabilities of the in vivo-in vitro relationship as small changes in measured DDTr bioaccessibility may result in large differences in predicted DDTr RBA. Although there is limited guidance on acceptability criteria for assessing the robustness of in vitro assays for predicting contaminant RBA, it has been suggested that in addition to a strong goodness of fit (r 40.8) and a small y-intercept, the slope of in vivo-in vitro relationship should be 0.8–1.2 (Wragg et al., 2011). In contrast, when the in vivo-in vitro relationship was assessed using DDTr bioaccessibility determined using the org-PBET with the inclusion of the sorption sink, a strong in vivo-in vitro correlation (r2 ¼0.72) was observed with a slope of near unity (0.94) and a small y-intercept (3.5; Fig. 5). Previously, Li et al. (2016) determined that inclusion of Tenax as a sorption sink into the org-
DDTr accumulation (µg g -1 DW)
A.L. Juhasz et al. / Environmental Research 150 (2016) 482–488
487
400
300
200
100
0 0
500
1000
1500
DDT Dose ( µ g day ) -1
PBET provided a suitable approach for predicting DDTr RBA (slope ¼1.17, y-intercept¼ 14.8, r2 ¼ 0.62). When DDTr in vivo-in vitro correlations determined in this study and that of Li et al. (2016) were compared, there was no significant difference in the slope (p ¼0.5) and y-intercepts (p ¼0.9) of these relationships due in part to the large standard deviation for some DDTr RBA measurements. Notwithstanding, this research suggests that the inclusion of a sorption sink, such as silicon cord or Tenax, into gastrointestinal solution provides a good surrogate measure of the passive diffusion of DDTr across the small intestines.
Fig. A1. Dose-response of DDT administered to mice and DDTr accumulation in liver and kidneys.
80
DDTr Bioaccessibility (%)
Fig. 5. Relationship between DDTr relative bioavailability, determined using an in vivo mouse assays and DDTr bioaccessibility, determined using the org-PBET with (■) and without (☐) the incorporation of silicone cord. The solid line represents the line of best fit with 95% confidence intervals.
60
40
20
0
4. Conclusions Human exposure to HOCs via incidental soil ingestion is influenced by the propensity of the contaminant to desorb from the soil matrix under gastrointestinal conditions and the propensity for absorption across intestinal epithelia. While DDTr RBA may be determined using an in vivo mouse model with accumulation in kidneys and liver as the biomarker of exposure, outcomes of this research demonstrated that DDTr RBA may be predicted using an in vitro approach incorporating a sorption sink. The in vitro approach provides a methodology that is inexpensive and rapid compared to in vivo strategies. Although further data is required to support the robustness of the in vivo-in vitro relationship, the application of the sorption sink approach has the potential to inform exposure assessment refinement which may impact risk assessment and remediation outcomes.
Acknowledgements In vivo assays were approved and conducted according to application SAM73 of the South Australian Health and Medical Research Institute Animal Ethic Committee. The authors would like to acknowledge the support of the Future Industries Institute, University of South Australia and the South Australian Health and Medical Research Institute for this research. Appendix A Details of the in vivo DDT dose-response (Fig. A1) and the relationship between DDTr bioaccessibility and TOC (Figure A2) are
0
2
4
6
8
TOC (%) Fig. A2. Relationship between DDTr bioaccessibility, determined using the orgPBET with silicone cord inclusion, and soil total organic carbon content.
provided in Appendix A.
References Aislabie, J.M., Richards, N.K., Boul, H.L., 1997. Microbial degradation of DDT and its residues: a review. N.Z. J. Agric. Res. 40, 269–282. ATSDR, 2002. Toxicological profile for DDT, DDE and DDD. US Department Of Health And Human Services, Agency For Toxic Substances And Disease Registry,, Atlanta, Georgia. Beard, J., 2006. DDT and human health. Sci. Total Environ. 355, 78–89. Bielská, L., Hofman, J., 2015. The kinetics of solid-phase microextraction measured for freshly added and aged hydrophobic compounds in two different soils. Int. J. Environ. Anal. Chem. 95, 635–649. Bosch, C., Grimalt, J.O., Fernández, P., 2015. Enantiomeric fraction and isomeric composition to assess sources of DDT residues in soils. Chemosphere 138, 40–46. Diamond, G.L., Bradham, K.D., Brattin, W.J., Burgess, M., Drexler, J.W., Griffin, S., Hawkins, C.A., Juhasz, A.L., Klotzbach, J.M., Nelson, C., Lowney, Y.W., Scheckel, K. G., Thomas, D.J., 2016. Predicting oral relative bioavailability of arsenic in soil from in vitro bioaccessibility. J. Toxicol. Environ. Health Part A 79, 165–173. Esaac, E.G., Matsumura, F., 1980. Metabolism of insecticides by reductive systems. Pharmacol. Ther. 9, 1–26. Foght, J., April, T., Biggar, K., Aislabie, J., 2001. Bioremediation of DDT-contaminated soils: a review. Bioremediat. J. 5, 225–246. Gerber, R., Smit, N.J., Van Vuren, J.H.J., Nakayama, S.M.M., Yohannes, Y.B., Ikenaka, Y., Ishizuka, M., Wepener, V., 2016. Bioaccumulation and human health risk assessment of DDT and other organochlorine pesticides in an apex aquatic predator from a premier conservation area. Sci. Total Environ. 550, 522–533. Gorelick, F.S., Jamieson, J.D., 1994. The pancreatic acinar cell: structure-functional relationship. In: Johnson, L.R. (Ed.), Physiology of the Gastrointestinal Tract vol. 2. Raven Press, New York, N.Y.
488
A.L. Juhasz et al. / Environmental Research 150 (2016) 482–488
Gouliarmou, V., Collins, C.D., Christiansen, E., Mayer, P., 2013. Sorptive physiologically based extraction of contaminated solid matrices: incorporating silicone rod as absorption sink for hydrophobic organic contaminants. Environ. Sci. Technol. 47, 941–948. Guo, Y., Yu, H.-Y., Zeng, E.Y., 2009. Occurrence, source diagnosis, and biological effect assessment of DDT and its metabolites in various environmental compartments of the Pearl River Delta, South China: a review. Environ. Pollut. 157, 1753–1763. Hilber, I., Mäder, P., Schulin, R., Wyss, G.S., 2008. Survey of organochlorine pesticides in horticultural soils and there grown Cucurbitaceae. Chemosphere 73, 954–961. James, K., Peters, R.E., Laird, B.D., Ma, W.K., Wickstrom, M., Stephenson, G.L., Siciliano, S.D., 2011. Human exposure assessment: a case study of 8 PAH contaminated soils using in vitro digestors and the juvenile swine model. Environ. Sci. Technol. 45, 4586–4593. Juhasz, A.L., Herde, P., Herde, K., Boland, J., Smith, E., 2014. Validation of the predictive capabilities of the SBRC-G in vitro assay for estimating arsenic relative bioavailability in contaminated soils. Environ. Sci. Technol. 48, 12962–12969. Juhasz, A.L., Herde, P., Herde, K., Boland, J., Smith, E., 2015. Predicting arsenic relative bioavailability using multiple in vitro assays: validation of in vivo-in vitro correlations. Environ. Sci. Technol. 49, 11167–11175. Juhasz, A.L., Tang, W., Smith, E., 2016. Using in vitro bioaccessibility to refine estimates of human exposure to PAHs via incidental soil ingestion. Environ. Res. 145, 145–153. Kitamura, S., Shimizu, Y., Shiraga, Y., Yoshida, M., Sugihara, K., Ohta, S., 2002. Reductive metabolism o fp.p′-DDT and o,p′-DDT by rat liver cytochrome P450. Drug Metab. Dispos. 30, 113–118. Li, C., Cui, X.Y., Fan, Y.Y., Teng, Y., Nan, Z.R., Ma, L.Q., 2015. Tenax as sorption sink for in vitro bioaccessibility measurement of polycyclic aromatic hydrocarbons in soils. Environ. Pollut. 196, 47–52. Li, C., Sun, H.-J., Juhasz, A.L., Cui, X., Ma, L.Q., 2016. Prediction of DDTr relative bioavailability in historically-contaminated soils using a Tenax-improved physiologically based extraction test (TI-PBET). Environ. Sci. Technol. 50, 1118–1125. McDougall, K.W., 1997. Arsenic and DDT residues at cattle tick dip sites in NSW. Land Contam. Reclam. 5, 323–336. Menchai, P., Van Zwieten, L., Kimber, S., Ahmad, N., Rao, P.S.C., Hose, G., 2008. Bioavailable DDT residues in sediments: laboratory assessment of ageing effects using semi-permeable membrane devices. Environ. Pollut. 153, 110–118. Mitra, A., Chatterjee, C., Mandal, F.B., 2011. Synthetic chemical pesticides and their effects on birds. Res. J. Environ. Toxicol. 5, 81–96. National Research Council (NRC), 1996. Guide for the Care and Use of Laboratory Animals. National Academy Press,, Washington DC. Nelson, D.W., Sommers, L.E., 1996. Total carbon, organic carbon, and organic matter. In: Page, A.L., et al. (Eds.), 2nd ed. Methods of Soil Analysis, Part 2 9. American Society of Agronomy, Inc., Madison, WI, pp. 961–1010.
Oomen, A.G., Rompelberg, C.J.M., Bruil, M.A., Dobbe, C.J.G., Pereboom, D.P.K.H., Sips, A.J.A.M., 2003. Development of an in vitro digestion model for estimation of bioaccessibility of soil contaminants. Arch. Environ. Contam. Toxicol. 44, 281–287. Plimmer, J.R., Kearney, P.C., von Endt, D.W., 1968. Mechanism of conversion of DDT to DDD by Aerobacter aerogenes. J. Agric. Food Chem. 4, 594–597. Qu, J., Xu, Y., Ai, G.-M., Liu, Y., Liu, Z.-P., 2015. Novel Chryseobacterium sp. PYR2 degrades various organochlorine pesticides (OCPs) and achieves enhancing removal and complete degradation of DDT in highly contaminated soil. J. Environ. Manag. 161, 350–357. Smith, E., Kempson, I.M., Juhasz, A.L., Weber, J., Rofe, A., Gancarz, D., Naidu, R., McLaren, R.G., Grafe, M., 2011. In vivo - in vitro and XANES spectroscopy assessments of lead bioavailability in contaminated peri-urban soils. Environ. Sci. Technol. 45, 6145–6152. Smith, E., Weber, J., Rofe, A., Gancarz, D., Naidu, R., Juhasz, A.L., 2012. Assessment of DDT relative bioavailability and bioaccessibility in historically contaminated soils using an in vivo mouse model and fed and unfed batch in vitro assays. Environ. Sci. Technol. 46, 2928–2934. Tang, X.Y., Tang, L., Zhu, Y.G., Xing, B.S., Duan, J., Zheng, M.H., 2006. Assessment of the bioaccessibility of polycyclic aromatic hydrocarbons in soils from Beijing using an in vitro test. Environ. Pollut. 140, 279–285. U.S. Environmental Protection Agency., 2007a. U.S. EPA Method 600/4-81-055, Interim Methods for the Sampling and Analysis of Priority Pollutants in Sediments and Fish Tissue, Section 3.1.3. Dionex Corporation, Sunnyvale, CA, USA. 〈http://www.dionex.com/en-us/webdocs/4321-AN320-ASE-Chlorinated-Pesti cides-07Apr2011-LPN0637-04.pdf〉 (downloaded 18.07.2011). U.S. Environmental Protection Agency., 2007b. U.S. EPA Method 8081B, Organochloride Pesticides by Gas Chromatography. 〈http://www.epa.gov/osw/ha zard/testmethods/sw846/pdfs/8081b.pdf〉 (downloaded 18.07.2011). Van Zwieten, L., Ayresm, M.R., Morris, S.G., 2003. Influence of arsenic co-contamination on DDT breakdown and microbial activity. Environ. Pollut. 124, 331–339. Wedemeyer, G., 1967. Dechlorination of 1,1,1-trichloro-2,2-bis(p-chlorophenyl) ethane by Aerobacter aerogenes. Appl. Microbiol. 15, 569–574. Wragg, J., Cave, M., Basta, N., Brandon, E., Casteel, S., Denys, S., Gron, C., Oomen, A., Reimer, K., Tack, K., Van de Wiele, T., 2011. An inter-laboratory trial of the unified BARGE bioaccessibility method for arsenic, cadmium and lead in soil. Sci. Total Environ. 409, 4016–4030. Wu, J.-P., Zhang, Y., Luo, X.-J., Chen, S.-J., Mai, B.-X., 2012. DDTs in rice frogs (Rana limnocharis) from an agricultural site, South China: tissue distribution, biomagnification, and potential toxic effects assessment. Environ. Toxicol. Chem. 31, 705–711. Zhang, Y., Pignatello, J.J., Tao, S., Xing, B., 2015. Bioaccessibility of PAHs in fuel soot assessed by an in vitro gastrointestinal model: effect of including an absorption sink. Environ. Sci. Technol. 49, 3905–3912.