ARTICLE pubs.acs.org/est
Colon Extended Physiologically Based Extraction Test (CE-PBET) Increases Bioaccessibility of Soil-Bound PAH E. L. Tilston,†,§ G. R. Gibson,‡ and C. D. Collins*,† † ‡
Soil Research Centre, University of Reading, Reading, RG6 6DW, U.K. Department of Food and Nutritional Sciences, University of Reading, Reading, RG6 6AH, U.K.
bS Supporting Information ABSTRACT: Assessment of the risk to human health posed by contaminated land may be seriously overestimated if reliant on total pollutant concentration. In vitro extraction tests, such as the physiologically based extraction test (PBET), imitate the physicochemical conditions of the human gastro-intestinal tract and offer a more practicable alternative for routine testing purposes. However, even though passage through the colon accounts for approximately 80% of the transit time through the human digestive tract and the typical contents of the colon in vivo are a carbohydrate-rich aqueous medium with the potential to promote desorption of organic pollutants, PBET comprises stomach and small intestine compartments only. Through addition of an eight-hour colon compartment to PBET and use of a carbohydrate-rich fed-state medium we demonstrated that colon-extended PBET (CE-PBET) increased assessments of soil-bound PAH bioaccessibility by up to 50% in laboratory soils and a factor of 4 in field soils. We attribute this increased bioaccessibility to a combination of the additional extraction time and the presence of carbohydrates in the colon compartment, both of which favor PAH desorption from soil. We propose that future assessments of the bioaccessibility of organic pollutants in soils using physiologically based extraction tests should have a colon compartment as in CE-PBET.
’ INTRODUCTION Total pollutant concentration is frequently used in the assessment of risk posed by contaminated land to human health. While this has the advantage of being precautionary, it may significantly overestimate the amount of pollutant absorbed by biota, including humans. Such an overestimation of risk can result in significant additional cleanup costs that may reduce the sustainability of brownfield remediation, consequently a number of analytical approaches have been developed to measure likely pollutant exposure. In vivo methods using rats1 or pigs2 enable realistic measurement of the bioavailable contaminant fraction, i.e., the fraction at any given time available for uptake across an organism’s cellular membrane.3 However, the ethical considerations of using mammals and the practical disadvantages of high cost and low sample through-put, make these methods unsuitable for routine testing purposes. Invertebrates are more practical test organisms than mammals,4 but their differing physiologies and ecologies compromise extrapolations of pollutant bioavailability to humans. One of the principal sources of human exposure to contaminated soils is direct ingestion as a result of hand-to-mouth activity, and several in vitro physiologically based extraction tests r 2011 American Chemical Society
have been proposed.58 In vitro alternatives are a cheaper precursor of bioavailability measures, but their shorter duration means they are only able to give information on the bioaccessible fraction, i.e., that which is released from the soil matrix into the gut fluid.3 To date, these systems have been predominantly used to assess the bioaccessibility of polluting trace elements.5,9,10 Comparable work with organic pollutants is more limited1,1113 despite the fact that organic pollutants are equally widespread and a number frequently appear in target lists, e.g., the Stockholm convention (http://chm.pops.int/default.aspx). Polycyclic aromatic hydrocarbons (PAH) are ubiquitous environmental organic pollutants arising from the petroleum industry, old gasworks, and numerous other combustion sources. Soils are the major environmental sink for PAHs and it has been estimated that over 90% of the U.K. PAH burden resides in soil 14 A gastro-intestinal extraction system used to measure exposure to matrix-bound pollutants typically comprises one or more Received: February 11, 2011 Accepted: May 3, 2011 Revised: April 28, 2011 Published: May 13, 2011 5301
dx.doi.org/10.1021/es2004705 | Environ. Sci. Technol. 2011, 45, 5301–5308
Environmental Science & Technology
ARTICLE
Table 1. PhysicoChemical Definition of Each Compartment in the Colon-Extended PBET (CE-PBET) Extraction Test extraction test compartment stomach incubation time
small intestine
colon
1
4
8
2.5
7.0
6.5
(hours) pH of medium biological factors present and status pepsin
present, active present, inactive
absent
pancreatin
absent
absent
present (0.5 g L1)
bile salts
absent
present (1.78 g L1)
present (400 mg L1)
“compartments” maintained at body temperature (37 °C) essentially defined by pH, exposure time, and inclusion of biogenic compounds. The residence time for the mouth and esophagus is less than 2 min and is not considered to have a significant impact on bioaccessibility.15 Therefore gastro-intestinal extraction tests, such as the physiologically based extraction test (PBET)5 begin with 1 h exposure to the stomach, followed by 4 h exposure to the small intestine. Even though some absorption of nutrients, specifically the fermentation products of commensal microbial activity, are absorbed in the colon16 and Cavret et al.17 have demonstrated that PAHs can cross the cell membrane of the colon epithelium cell line Caco-2, there are only a limited number of models which include the colon, e.g., SHIME.18 In addition to extending extraction time, the colon medium itself might be expected to increase PAH desorption from soil because it contains peptides, carbohydrates, and bile salts all of which are known to be a sink for organic pollutants.11,19,20 The aim of this study was to develop a robust system for the assessment of persistent organic pollutant bioaccessibility by adding a colon compartment to PBET. There were three aspects to the development of the colon-extended PBET (CE-PBET) system; first, unfed and fed-state media for use in CE-PBET were assessed; second, the optimum exposure time for the colon compartment was determined and the system was characterized and tested in batch and sequential modes and some of the medium components promoting PAH desorption were identified. Finally, CE-PBET was used to determine the bioaccessibility of field soils.
’ MATERIALS AND METHODS Gastro-Intestinal Extraction. The gastro-intestinal extraction test comprised three compartments, specifically the stomach, small intestine, and colon (Table 1). Incubations were performed in 130-mL-capacity glass bottles, placed in a shaking water bath at 37 °C. Soil (1 g d.w.) was added to 100 mL of medium prewarmed for 30 min. PAHs are semivolatile so the incubation vessels were not continuously sparged with N2 as described by Ruby et al.,5 but sealed with Teflon-lined screw-caps for the duration of the incubation. The integrity of bottles and their seals was confirmed in separate experiments where aliquots of media were spiked with a mixture of seven PAHs in acetone (Table S1, Supporting Information (SI)). There was no significant
difference in the amount of individual PAHs extracted immediately after spike addition or after up to 8 h of incubation. Two formats of the extraction test were operated: (a) batch where test substrates were exposed to each compartment in isolation and (b) sequential where test substrates were exposed to the three compartments in succession. In the sequential format, exposure to stomach and small intestine compartments was achieved by exposing the test substrate to the stomach compartment for 1 h before converting the stomach medium to small intestine medium. The bottles were then resealed and incubation continued for a further 4 h. The transition between small intestine and colon compartments was effected by physical transfer: the test substrate was recovered by centrifugation (3000g, 10 min), added to a bottle of prewarmed colon medium, and incubated for a further 8 h. Media. Stomach medium was based on the unfed-state medium used in PBET,5 with the small intestine created by addition of bile salts and pancreatin and adjustment of the pH. Fed-state colon medium was that used by Macfarlane et al.21 which has been validated against the intestinal contents of human sudden death victims22 (Table S2, SI). For unfed-state colon medium the dietary components were omitted and were likewise included to create fed-state stomach and small intestine media. While not included in the original PBET media, 4.0 g L1 mucin and 800 mg L1 cysteine hydrochloride were included in media for both stomach and small intestine compartments. Mucin is present throughout the gastro-intestinal tract23, and cysteine hydrochloride is a reducing agent used to promote establishment of anaerobic conditions.24 Analytical grade inorganic salts (Fisher Scientific, Loughborough, U.K.) were used throughout and all other reagents were obtained from Sigma-Aldrich (Gillingham, U.K.). PAH Extraction. The contents of each incubation bottle were separated into solid (soil and particulate medium components) and liquid phases by centrifugation. PAHs in solution were extracted by a single liquidliquid extraction optimized in preliminary work; briefly, 2 mL of supernatant were removed to 10-mL capacity round-bottomed glass culture tubes with Teflon-lined screw-caps, 4 mL of 5-to-4 acetone-hexane extractant (GLC, pesticide residue grade, Fisher Scientific) containing 500 μg mL1 biphenyl (Alfa Aesar, Heysham, U.K.) as an internal standard was added, and the contents of the tube were vortexed at 1600 rpm for 1 min. The tubes were then left to stand for 15 min during which time mucin precipitated and uncharacterized, unstable emulsions separated into aqueous and hexane phases. The hexane phase was then transferred to GC vials and stored at 20 °C until analysis. The pellet (soil and insoluble medium components) was extracted with 4 mL of 5:4 acetonehexane extractant in glass vials placed on a roller shaker for 20 min, 2 mL of distilled water was then added to promote partitioning of PAHs into the hexane component, and the vial was returned to the roller shaker for a further 5 min. Recovery of individual PAHs from spiked solutions used in the different CEPBET compartments ranged between 50.1 and 96.0, mean 76.2 (Table S1). Gas Chromatography. PAHs in the hexane extracts of the pellet were quantified by an Agilent 6890N gas chromatograph fitted with a 30 m 320 μm 0.25 μm 5% phenyl methyl siloxane capillary column (Agilent, Santa Clara, CA). A CombiPAL (CTC Analytics, Zwingen, Switzerland) autosampler introduced 2 μL into the inlet (300 °C), which was operated in splitless mode. The temperature program was 50 °C, ramped at 15.3 °C 5302
dx.doi.org/10.1021/es2004705 |Environ. Sci. Technol. 2011, 45, 5301–5308
Environmental Science & Technology
ARTICLE
Table 2. Recovery of PAHs from Artificially Spiked OECD Soil Incubated in Unfed Stomach and Small Intestine Media for PBET and CE-PBET Systemsa residual soil, T0 system unfed PBET
unfed CE-PBET
PAH
1
(μg g )
stomach, Tþ1 soil (%)
small intestine, Tþ4 soil (%)
supernatant (%) 30.3
naphthalene
209 b
63.0
37.2 a
acenaphthene
33 d
78.9
68.3 bc
28.5
fluorene phenanthrene
40 d 77 c
77.3 80.0
74.8 bc 86.5 a
nd nd
anthracene
37 d
84.1
88.2 d
nd
fluoranthene
86 c
86.9
91.5 d
nd
pyrene
52 d
80.7
90.4 d
nd
naphthalene
243 a
74.1
39.8 a
30.6
acenaphthene
36 d
92.3
71.3 bc
29.0
fluorene
43 d
93.0
77.0 c
nd
phenanthrene anthracene
79 c 38 d
94.5 95.1
64.4 b 90.8 d
nd nd nd
fluoranthene
85 c
91.6
76.0 bc
pyrene
50 d
88.3
92.1 d
nd
coefficient of variation range (%)
4 7
2 12
2 10
6 12
statistical significance of interaction term, P
0.009
0.757
0.003
0.948
a
Each value is the mean of three replicates (six for residual soil, T0). Means within a column identified by a common letter are not significantly different from each other according to Tukey’s honestly significant difference test (P < 0.050). There were no significant differences among PAHs in the stomach treatment. Percentages were analyzed as arc-sine transformed values, but are presented as untransformed values. nd = not detected.
min1 to 280 °C and held for 15 min. The carrier gas was helium at 2 mL min1. Individual PAHs were detected with a FID at 300 °C with identification and quantification according to the retention times and peak areas of pure compounds dissolved in hexane. PAHs extracted from aliquots of supernatant were quantified by an Agilent 7890A network GC system fitted with a 5% phenyl methyl siloxane capillary column (30 m 250 μm 0.5 μm) and coupled to an Agilent 5975C mass spectrometer operated in single ion monitoring (SIM) mode, used as described by Gomez-Eyles et al.25 Limits of detection for GC-FID analysis were 22 μg g1 soil for naphthalene and 5 μg g1 soil for all other PAHs and those for GC-MS analysis were 400 μg L1 medium for naphthalene and 100 μg L1 medium for all other PAHs. Artificially Contaminated Soil. OECD standard soil comprising 10% milled moss peat, 70% 50200 μm sand, 20% kaolin clay, and adjusted to pH 6 with calcium carbonate (CaCO3) was used.4 The soil was artificially contaminated with a mixture of seven PAHs by placing 1 kg dry soil to a depth of approximately 2 cm in an aluminum foil tray and applying 100 mL of PAHs (97 or 98% purity, Sigma-Aldrich) in acetone; uncontaminated soil was treated similarly with just acetone. Soils were mixed well and the acetone was allowed to evaporate for 1 h in a fume cupboard, the moisture content was then adjusted to 50% water holding capacity before storage at 20 °C in glass screw-capped jars. The initial concentrations of PAHs (μg g1 soil) were as follows: naphthalene, 460; acenaphthene, 94; fluorene, 91; phenanthrene, 95; anthracene, 93; fluoranthene, 96 and pyrene, 95. These concentrations would be typical for the uppermost 15 cm of heavily contaminated soil at old gasworks sites in the U.K..14 Periodically nine replicate aliquots of soil as stored were extracted, the variability among replicates ranged between 4 and 7% for individual PAHs and no significant decline in PAH concentrations was observed.
Experiments. Experiment 1 was a comparison of unfed PBET with unfed CE-PBET medium, effectively testing the effects of including mucin and cysteine hydrochloride in the medium. Only the stomach and small intestine compartments were considered because of the absence of a colon compartment in PBET. Experiment 2 determined the optimum incubation time for the unfed CE-PBET colon compartment via a time course experiment with sampling after 4, 8, 16, and 32 h. Experiment 3 contrasted the presence or absence of medium components such as the dietary components (a comparison of unfed-state CE-PBET with fedstate CE-PBET) and concentration of bile salts in the fed-state CE-PBET. Batch test format was used for these experiments. Experiment 4 used the sequential test format to compare exposure to individual CE-PBET compartments with all three compartments in series. Experiment 5 used a number of samples from an industrially impacted soil. These were transported in cool boxes and sieved to 2 mm in the laboratory. They were then subject to the CE-PBET test in the fed-state, sequential format. Quality Control. In all experiments there were three replicates of each treatment, unless otherwise stated. Aliquots of uncontaminated soil were also incubated and acted as controls, and bottles without soil were included as blanks, but in subsequent analysis by gas chromatography no peaks with retention times corresponding with those of the seven target PAHs were ever detected in hexane extracts of control and blank bottles. Data Handling and Statistics. Prior to statistical analysis, T0-referenced concentrations of PAHs were converted to percentage values and arc-sine transformed. This reduced the influence of both the underlying differences in the initial concentrations of the individual PAHs and unequal variances associated with a wide range of percentage values.26 Analyses of variance were performed using MINITAB (release 15, MINITAB, State College, PA) and means were separated according to Tukey’s honestly significant difference test (P < 0.050). 26 5303
dx.doi.org/10.1021/es2004705 |Environ. Sci. Technol. 2011, 45, 5301–5308
Environmental Science & Technology
ARTICLE
Figure 1. Recovery of PAHs from artificially spiked OECD soil (a) and in solution (b) in unfed CE-PBET colon medium over a 32-h incubation period. Values are means of three replicates. ( = naphthalene, 2 = acenaphthene, 4 = fluorene, 9 = anthracene, 0 = phenanthrene, b = fluoranthene, O = pyrene. HSD bar indicates Tukey’s honestly significant difference for comparing means.
Table 3. Recovery of PAHs from CE-PBET Run in Batch Format with Artificially Spiked OECD Soil Incubated in Unfed- and FedState CE-PBET Mediaa residual soil, T0 fed state unfed CE-PBET
fed CE-PBET
PAH
(μg g1)
stomach, Tþ1
small intestine, Tþ4
colon, Tþ8
soil (%)
soil (%)
supernatant (%)
soil (%)
supernatant (%)
naphthalene
235 d
62.1 a
33.0 b
25.7
28.5 b
22.0 def
acenaphthene
35 a
81.8 cd
63.2 c
25.8
28.3 b
22.8 def
fluorene
41 ab
81.0 cd
67.0 cd
nd
29.4 b
7.5 a
phenanthrene anthracene
77 c 40 ab
87.7 de 86.1 c
59.8 ef 82.3 f
nd nd
44.7 d 58.0 e
9.5 ab 15.0 bcd
fluoranthene
86 c
89.4 e
74.3 g
nd
74.5 g
6.5 a
pyrene
53 b
90.4 e
94.2 h
nd
77.0 g
12.4 abc
naphthalene
202 d
65.7 a
50.1 a
32.5
24.9 a
19.6 cdef
acenaphthene
32 a
68.7 ab
59.3 cd
31.3
31.3 b
24.5 efg
36 ab
69.0 ab
64.2 d
26.5
39.0 c
24.6 fg
phenanthrene
70 c
69.7 a
71.2 c
15.8
61.5 e
15.4 bcdef
anthracene fluoranthene
31 a 73 c
75.1 bc 67.9 ab
74.7 g 82.3 f
7.0 10.2
60.3 e 86.3 h
34.4 g 12.4 abc
40 ab
fluorene
65.6 a
89.3 i
19.3
67.3 f
20.4 cdef
coefficient of variation range (%)
pyrene
919
29
210
111
111
796
statistical significance of interaction term, P
