aco
se
f Ap
tle u
ed f
extraction test (PBET) or more generally, a simulated in vitro
gastro-intestinal extraction procedure.
resentative of the human digestive tract include stomach and
small intestinal pH and chemistry, soil-to-solution ratio, stom-
ach mixing, and stomach emptying rates. Several in vitro
methods have been developed (Miller et al., 1981; Crews
et al., 1983; Ruby et al., 1993; Hack and Selenka, 1996).
* Corresponding author. Tel.: þ44 191 227 3047; fax:þ44 191 227 3519.
E-mail address: john.dean@unn.ac.uk (J.R. Dean).
Available online at www.sciencedirect.com
Environmental Pollution 151. Introduction
The environmental risk to humans from metals by consum-
ing soil and contaminated vegetables can be assessed by mea-
suring their bioaccessibility. In this context, bioaccessibility
has been defined as the fraction of a compound that is released
from its matrix in the gastrointestinal tract, and thus becomes
available for intestinal absorption, i.e. enters the blood stream
(Oomen et al., 2002). Several in vitro approaches have been
developed in attempts to mimic the effects of the human diges-
tion process. They are commonly described in the scientific lit-
erature under the specific name of the physiologically-based
There are a number of in vivo approaches available to eval-
uate the bioaccessibility of metals for humans, e.g. from soil,
dust or food. However, each method has its limitations. Infor-
mation obtained from in vivo studies can be difficult to inter-
pret due to physiological discrepancies between humans and
the experimental animals adopted. Such problems led to the
development of in vitro systems, based on gastrointestinal
(GI) extraction including the so called physiologically-based
extraction test (PBET). The technique measures the fraction
of a metal which is solubilised from a sample under simulated
gastrointestinal conditions and which therefore is available for
absorption (Kelley et al., 2002). The simulated parameters rep-Keywords: Physiologically-based extraction test (PBET); Metals; Oral bioaccessibility; ICP-MS; PlantsEvaluation of a physiologically-based extraction test to assess the risk to humans of consuming contaminated vegetables.
Abstract
The oral bioaccessibility of metals in vegetable plants grown on contaminated soil was assessed. This was done using the physiologically-
based extraction test (PBET) to simulate the human digestion of plant material. A range of vegetable plants, i.e. carrot, lettuce, radish and spin-
ach, were grown on metal contaminated soil. After reaching maturity the plants were harvested and analysed for their total metal content (i.e. Cr,
Cd, Cu, Fe, Mn, Mo, Ni, Pb and Zn) by inductively coupled plasma-mass spectrometry (ICP-MS). The plant samples were then subsequently
extracted using an in vitro gastrointestinal approach or PBET to assess the likelihood of oral bioaccessibility if the material was consumed by
humans.
 2007 Elsevier Ltd. All rights reserved.Use of the physiologically-b
the oral bioaccessibility
plants grown in
Marisa Intawong
Biomolecular and Biomedical Research Centre, School o
Ellison Building, Newcas
Received 5 March 2007; received in revis0269-7491/$ - see front matter  2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.envpol.2007.05.022sed extraction test to assess
of metals in vegetable
ntaminated soil
, John R. Dean*
plied Sciences, University of Northumbria at Newcastle,
pon Tyne NE1 8ST, UK
orm 7 May 2007; accepted 11 May 2007
2 (2008) 60e72
www.elsevier.com/locate/envpol

can be simple, rapid and low in cost and may provide insights
1).
In vitro extraction procedures make use of simulated gastric
were used for this study. A control soil is one to which no spik-
nmeand intestinal juices which are applied to samples to try to pre-
dict the availability of metals for human absorption. During
the past two decades, several approaches have been investi-
gated to assess bioaccessibility of metals in soil and food
samples. The system was developed originally to assess the
bioavailability of iron from food for nutrition studies (Miller
et al., 1981). The methods are both rapid and inexpensive, re-
quiring only a day and only a small fraction of the cost of an in
vivo study (Kelley et al., 2002). Most methods are static gas-
trointestinal models which simulate transit through the human
digestive tract by sequential exposure of the samples to simu-
late mouth, gastric, and small intestinal conditions (Oomen
et al., 2002). Only a few groups of researchers have carried
out dynamic gastrointestinal models which mimic the gradual
transit of extracted mixtures through the simulated physiolog-
ical conditions in the digestive tract. Static models are more
convenient than dynamic models in terms of ease of use.
Knowledge of oral bioaccessibility of a contaminant is use-
ful and valid for estimating potential human health risks. As
bioaccessibility data is essentially related to the amount of
contaminant in the animal/human bloodstream then data
must be produced from the dosing of animals with contami-
nated samples and the subsequent measurement of the contam-
inant in the blood or organs of the animals, i.e. the use of in
vivo animal models (Cave et al., 2002). However, since in
vivo studies are both expensive and laborious, and the possi-
bility of measuring certain parameters during the experiments
is often limited (Cabanero et al., 2004), the data are normally
determined in an in vitro environment and represents the
amount of contaminant dissolved in the gastrointestinal tract.
This paper uses the physiologically-based extraction test
(PBET) to determine the oral bioaccessibility of metals in
contaminated plant samples. The procedure applied was first
described by Ruby et al. (1996) and subsequently modified
Cave et al., 2002. No attempt is made to assess the uptake
by plants from metal contaminated soil as this has been previ-
ously reported (Intawongse and Dean, 2006).
2. Experimental
2.1. Chemicals and apparatus
All chemicals used were of analytical grade. Concentrated
hydrochloric acid, acetic acid, sodium bicarbonate, pepsin-A
powder 1 Anson unit per gram (lactose as diluent) and pancre-
atin were provided by BDH Chemicals Ltd. (Poole, UK),
while bile salts, sodium malate, sodium citrate and lacticnot achievable in whole animal studies (Miller et al., 198All of the PBET models involve simulated gastric extraction
with pepsin and with a mixture of pancreatin, amylase and bile
salt in the intestinal stage. Researchers have shown that the in
vitro study results can be correlated to bioavailability deter-
mined by in vivo studies (Ruby et al., 1996). The approaches
M. Intawongse, J.R. Dean / Enviroacid were provided by Sigma (Missouri, USA). The soils
were adulterated using ammonium molybdate and lead(II)ing of heavy metal has been made. Before spiking, compost
soil was passed through a 2 mm sieve and air dried for 48 h.
For the low concentration level (approximately 5 times un-
adulterated concentration), a metal solution at the approximate
levels (50, 5, 150, 3000, 400, 100, 40, 25 and 100 mg/kg for
Cr, Cd, Cu, Fe, Mn, Mo, Ni, Pb and Zn, respectively) was
prepared to spike into the soil. The soil was weighed, approx-
imately 300 g for each mixing, and placed in a stainless steel
tray. Then, the soil was thoroughly mixed with the metal solu-
tion to ensure homogeneity and air dried for a day to allow
excessive water to evaporate. The spiked soil was left for
2 weeks before planting to ensure chemicalesoil contact.
The medium and high concentration level soils (10 and 15
times unadulterated concentration) were prepared in the
same manner as described above.
2.2.2. Growing of vegetable plants
The seeds of spinach, lettuce, radish and carrot which were
obtained directly from local markets were germinated in plas-
tic trays. After 2 weeks the seedlings were transplanted into
individual plastic pots containing 100 g of metal contaminated
soil; low, medium and high metal concentration soils were
used, as indicated above. Plants were also planted in unadul-
terated soil as control samples. The plants were watered dailynitrate from BDH Chemicals Ltd., cadmium(II) nitrate, cop-
per(II) nitrate, iron(III) nitrate, manganese(II) nitrate and zin-
c(II) nitrate from Fisher Scientific UK Ltd. (Loughborough,
Leicestershire), chromium(III) nitrate from Merck (Darmstadt,
Germany), and nickel(II) nitrate from Acros Organics (New
Jersey, USA).
Amulti-element standard for Cr, Mn, Fe, Ni, Cu, Zn, Mo, Cd
and Pb and an internal standard solution containing Sc, In and
Tb were purchased from SPEXCertiPrep (Middlesex, UK).
18.2 MU-cm ultra pure water used was produced by a Direct-
Q Millipore System (Molsheim, France). Soil (Levington
multipurpose compost) was obtained from a local garden centre.
The certified reference materials (CRMs) used were tea
leaves (INCT-TL-1) obtained from the Institute of Nuclear
Chemistry and Technology (Warszawa, Poland) and spinach
leaves (SRM 1570a) purchased from the National Institute
of Standards and Technology (Gaithersburg, MD, USA).
All ICP-MS measurements were carried out with an ICP
mass spectrometer XSeries II (Thermo Electron Corporation,
Cheshire, UK). A controlled temperature shaking water bath
(Grant Instruments Ltd., OLS 200, Cambridge, UK) was
employed for the PBET experiment. A heating block (2006
Digestor, Foss Tecator, Hoganas, Sweden) was used for acid
digestion of plant samples.
2.2. Methodology
2.2.1. Preparation of heavy metal contaminated soils
Contaminated soils spiked at three concentration levels
(low, medium and high) and control soils (unadulterated)
61ntal Pollution 152 (2008) 60e72with distilled water and grown under artificial light, using a
sodium lighting system at 150 Wm2 for periods of 16 h