Comparison of a static and a dynamic in vitro model to estimate the bioaccessibility of As, Cd, Pb and Hg from food reference materials Fucus sp. (IAEA-140/TM) and Lobster hepatopancreas (TORT-2)

Abstract

Bioaccessibility, the fraction of an element solubilized during gastrointestinal digestion and available for absorption, is a factor that should be considered when evaluating the health risk of contaminants from food. Static and dynamic models that mimic human physiological conditions have been used to evaluate bioaccessibility. This preliminary study compares the bioaccessibility of arsenic (As), cadmium (Cd), lead (Pb) and mercury (Hg) in two food certified reference materials (CRMs) (seaweed: Fucus sp., IAEA-140/TM; Lobster hepatopancreas: TORT-2), using two in vitro gastrointestinal digestion methods: a static method (SM) and a dynamic multicompartment method (TIM-1). There are significant differences (p < 0.05) between the bioaccessible values of As, Cd, Pb and Hg obtained by SM and TIM-1 in the two CRMs. The specific form in which the elements studied are present in the CRM may help to explain the bioaccessibility values obtained.

Introduction

Arsenic (As), cadmium (Cd), lead (Pb) and mercury (Hg) are toxic trace elements, and their concentrations in foods must be controlled by health authorities because of the possible adverse effects on the health that are associated with their dietary intake. In risk assessment, the exposure assessment stage evaluates the extent, duration, frequency and magnitude of exposures to a chemical pollutant (WHO, 2008). The evaluation of the magnitude of exposure to metal(loid)s through food should consider oral bioavailability, i.e. the “fraction of an administered dose that reaches the central (blood) compartment from the gastrointestinal tract” (Wragg and Cave, 2002), also known as absolute bioavailability (ABA). Relative Bioavailability (RBA) is obtained when the bioavailability of a material that is being studied is compared with the bioavailability determined for a reference material (Ruby et al., 1996, Nagar et al., 2009). A conservative tool for evaluating oral bioavailability is oral bioaccessibility, defined as “the fraction that is soluble in the gastrointestinal environment and is available for absorption” (Wragg and Cave, 2002). Bioaccessibility provides an indication of maximum oral bioavailability and is therefore an important tool to be used in risk assessment (Intawongse and Dean, 2006). In recent years, the evaluation of bioaccessibility has acquired special importance. Studies of soils have shown the suitability of evaluating gastric, intestinal or gastrointestinal bioaccessible metal(loid) concentrations by in vitro assays to predict the in vivo bioavailability of As, Cd and Pb (As: Rodríguez et al., 1999, Juhasz et al., 2007, Nagar et al., 2009; Cd: Schroder et al., 2003; and Pb: Ruby et al., 1996).

For food to be absorbed through the intestinal epithelium, first it has to be digested by a mechanical and chemical process that includes salivary, gastric and intestinal phases. Digestion starts in the mouth, where mechanical trituration is produced by mastication and mixing and moistening of the food with saliva, which contains ptyalin. The food then passes to the oesophagus and stomach, where it mixes with gastric juice which contains pepsin and hydrochloric acid. Digestion continues in the small intestine, where the absorption processes take place. In the first part of the duodenum the food is mixed with bile and pancreatic juice, which, together with a wide variety of enzymes secreted by the activity of enterocytes, constitute the intestinal juice. Nutrients and contaminants in the soluble fraction are then ready for absorption into the systemic circulation.

In vivo studies employing various animal models have been conducted to evaluate the bioavailability of As, Cd and Pb. Most of this work has been done with soils: As (Ruby et al., 1996, Ellickson et al., 2001, Rodríguez et al., 1999, Rodríguez et al., 2003, Juhasz et al., 2007, Nagar et al., 2009); Cd (Schroder et al., 2003); and Pb (Freeman et al., 1996, Ruby et al., 1996, Casteel et al., 1997, Ellickson et al., 2001). The number of in vivo studies for food (Juhasz et al., 2006, Juhasz et al., 2008) is very small. However, “to supplement or supersede the use of animals in determining the bioavailability of potentially harmful elements for human health risk assessment, or to estimate bioavailability where animal studies are not available, in vitro tests can be designed” (Wragg and Cave, 2002). On a laboratory level, these methods model the human physiological conditions involved in gastrointestinal digestion and can predict oral bioaccessibility of metals. A large number of samples can be analysed, and, once the method is fully validated, accurate and reproducible data can be achieved without any ethical constraints (Yoo and Chen, 2006).

Static in vitro models use sequential exposure to simulate digestion in different compartments (mouth, stomach and intestine). The selection of compartments varies according to the static models applied, with those that emulate the stomach and small intestine being the most frequent. Human physiological conditions such as pH, gastric and intestinal enzymes, temperature and residence times are emulated to some extent during application of the model. Dynamic in vitro models accurately reproduce the gradual transit of ingested compounds through the gastrointestinal tract and simulate in vivo human digestion. The TIM dynamic model (TNO Gastrointestinal Model) is a dynamic multicompartment computer-controlled system which has four computer-controlled chambers simulating the conditions of the stomach, duodenum, jejunum and ileum for TIM-1 (Minekus et al., 1995) and also the large intestine for TIM-2 (Minekus et al., 1999). The conditions simulated in TIM-1 are reliable, reproducible and consistent with in vivo data, and validation experiments have demonstrated the system's predictive value for evaluating the availability for human absorption of minerals, vitamins and food mutagens (Blanquet et al., 2004). Another in vitro simulator is SHIME (Simulator of the Human Intestinal Microbial Ecosystem), a model which mimics physicochemical, enzymatic and microbiological conditions in the stomach, small intestine and various colon regions (Yoo and Chen, 2006, Laird et al., 2009a, Laird et al., 2009b, Cave et al., 2010).

Static in vitro models have been used to study the bioaccessibility of a great variety of nutrient compounds (tocopherols, fatty acids, carotenoids and vitamins) and chemical compounds (metal(loid)s, polycyclic aromatic hydrocarbons, pesticides and phenols). With regard to toxic trace elements in foods, various static in vitro models have been used to determine As bioaccessibility from seaweed (Laparra et al., 2003, Laparra et al., 2004, Almela et al., 2005, Koch et al., 2007a), rice (Laparra et al., 2005), Chinese pills (Koch et al., 2007b) and seafood (Koch et al., 2007a, Laparra et al., 2007). For Cd and Pb, bioaccessibility has been studied in vegetable matrices (Mounicou et al., 2002, Versantvoort et al., 2004; Chan et al., 2007, Intawongse and Dean, 2008) and in seafood products (Amiard et al., 2008, Metian et al., 2009). Lastly, for Hg bioaccessibility, research has concentrated on fish (Cabañero et al., 2004, Shim et al., 2009, Laird et al., 2009a, Torres-Escribano et al., 2010).

The dynamic in vitro model TIM-1 has been applied to research on the behaviour of drug dosage forms (Blanquet et al., 2004) and the development of pharmaceutical formulations (Blanquet et al., 2005). In nutrients it has been employed to study the bioavailability of folate from fortified milk products (Verwei et al., 2006) and the parameters influencing the digestive stability of carotenoids (Blanquet-Diot et al., 2009). The TIM model has also been used in recent years to evaluate the bioaccessibility of As, Cd and Pb in soils (Oomen et al., 2002, Van de Wiele et al., 2007). To our knowledge, dynamic in vitro assays using the TIM-1 model have not been used to estimate the bioaccessibility of As, Cd, Pb and Hg in food certified reference materials (CRMs) or food products.

Currently, in soil samples the application of in vitro digestion models in risk assessment associated with oral exposure to metal(loid)s is frequent. However, for food contaminants there are two aspects that should be investigated before extending their use to improve exposure assessment. Firstly, there are no studies comparing the results generated by the various in vitro methods, in contrast to the large amount of information available for soil samples (Oomen et al., 2002, Van de Wiele et al., 2007, Hagens et al., 2009; Cave et al., 2010). Secondly, it has been shown that speciation is important in the understanding of metal bioaccessibility in soil samples. The arsenic and lead mineral forms that constitute the mineralogy of the soil explain the trends in the bioaccessibility of these elements, as shown by Ruby et al. (1996) and more recently by Meunier et al. (2010). For example, higher arsenic bioaccessibility in soils (up to 49%) is associated with the presence of calcium–iron arsenate, and soil samples containing arsenic predominantly as arsenopyrite or scorodite have a very low bioaccessibility < 1% (Meunier et al., 2010). In foods, too, speciation should be considered as a factor that may condition the bioaccessibility results obtained with each in vitro method.

The aim of this preliminary work is to compare a static method (SM) and the TIM-1 model to evaluate the bioaccessibility of toxic trace elements in foods. Two food CRMs were used to achieve this aim. The homogeneity and availability of these samples make them suitable candidates for this study.