Automated Microdialysis-Based System for in Situ Microsampling and ...

Article pubs.acs.org/est

Automated Microdialysis-Based System for in Situ Microsampling and Investigation of Lead Bioavailability in Terrestrial Environments under Physiologically Based Extraction Conditions María Rosende,† Luis M. Magalhaẽ s,‡ Marcela A. Segundo,‡ and Manuel Miró*,† †

FI-TRACE group, Department of Chemistry, University of the Balearic Islands, Carretera de Valldemossa, km 7.5, E-07122 Palma de Mallorca, Illes Balears, Spain ‡ REQUIMTE, Department of Chemical Sciences, Faculty of Pharmacy, University of Porto, R. de Jorge Viterbo Ferreira, 228, 4050-313 Porto, Portugal S Supporting Information *

ABSTRACT: In situ automatic microdialysis sampling under batch-flow conditions is herein proposed for the first time for expedient assessment of the kinetics of lead bioaccessibility/ bioavailability in contaminated and agricultural soils exploiting the harmonized physiologically based extraction test (UBM). Capitalized upon a concentric microdialysis probe immersed in synthetic gut fluids, the miniaturized flow system is harnessed for continuous monitoring of lead transfer across the permselective microdialysis membrane to mimic the diffusive transport of metal species through the epithelium of the stomach and of the small intestine. Besides, the addition of the UBM gastrointestinal fluid surrogates at a specified time frame is fully mechanized. Distinct microdialysis probe configurations and membranes types were investigated in detail to ensure passive sampling under steady-state dialytic conditions for lead. Using a 3-cmlong polysulfone membrane with averaged molecular weight cutoff of 30 kDa in a concentric probe and a perfusate flow rate of 2.0 μL min−1, microdialysis relative recoveries in the gastric phase were close to 100%, thereby omitting the need for probe calibration. The automatic leaching method was validated in terms of bias in the analysis of four soils with different physicochemical properties and containing a wide range of lead content (16 ± 3 to 1216 ± 42 mg kg−1) using mass balance assessment as a quality control tool. No significant differences between the mass balance and the total lead concentration in the suite of analyzed soils were encountered (α = 0.05). Our finding that the extraction of soil-borne lead for merely one hour in the GI phase suffices for assessment of the bioavailable fraction as a result of the fast immobilization of lead species at near-neutral conditions would assist in providing risk assessment data from the UBM test on a short notice.

■

INTRODUCTION In the assessment of soils potentially contaminated with metals and metalloids, soil ingestion in kindergarden playgrounds, especially by children, is considered to be the most important exposure pathway.1,2 The use of the total soil concentration of individual target metal species for evaluation of soil exposure is nowadays an obsolete practice as it has been demonstrated that only a fraction of a contaminant, namely, the bioavailable fraction, is absorbed following oral ingestion.3,4 The bioavailable fraction in human health-risk assessment refers to the fraction of substance that reaches the systemic circulation from the gastrointestinal tract. This encompasses bioaccessibility, which is defined as the proportion of the overall contaminant that is desorbed from soil in the human gastrointestinal (GI) tract and is potentially available for absorption across a physiological membrane.5 In the estimation of metal bioavailability in soil samples, as in vivo approaches are time-consuming, difficult to perform, and ethically controversial, several in vitro methods in line with the EU-REACH guidelines6 have been developed lately to mimic digestion conditions found in the gastrointestinal tract under © 2013 American Chemical Society

worst case scenarios so as to measure the maximum soluble concentration of target compounds, the so-called bioaccessible fraction.7,8 The BioAccessibility Research Group of Europe (BARGE), with the aim of producing a standardized procedure that harmonizes the various bioaccessibility techniques and their uses in human health risk assessment for contaminated soils, developed recently the Unified Barge Method (UBM),9 which were validated against an in vivo model (young swine) for Cd, Pb, Sb, and As.10 In this test, the human digestive procedure is simulated using digestive fluid surrogates: saliva, gastric juice, duodenal juice, and bile, the composition of which is based on human physiology. The UBM involves two extraction steps, the gastric digestion for 1 h, in which the saliva and gastric juice are added to the solid substrate, followed by the GI digestion for 4 h, which entails the addition of the duodenal and bile juices to the gastric phase. While both the Received: Revised: Accepted: Published: 11668

April 29, 2013 July 23, 2013 September 9, 2013 September 9, 2013 dx.doi.org/10.1021/es401872j | Environ. Sci. Technol. 2013, 47, 11668−11675

Environmental Science & Technology

Article

Figure 1. Diagrammatic description of the automated flow-batch set up incorporating a concentric microdialyzer for monitoring bioavailable lead fractions in contaminated and agricultural soils. PP, peristaltic pump; SV, solenoid valve; G, gastric; GI, gastrointestinal.

membrane16−19 with a specified pore size is employed to simulate the diffusive transport of metal species through the GI epithelium. It involves a simulated GI digestion followed by evaluation of the dialyzable concentration of target species that pass across the membrane of a dialysis bag. This in vitro method has been shown to provide bioavailability measurements of iron in foodstuffs that correlated well with human in vivo assessments.20 However, the manipulation and use of dialysis bags is not straightforward because of potential membrane leaking and unwanted release of the bag from the holder clip for beakers. Despite DGT and dialysis-based procedures being appealing tools for in vitro bioavailability explorations, both methods have not been harnessed to obtain knowledge on the GI absorption kinetics, because a single measurement of leached analytes is commonly undertaken after stopping the simulated GI digestion. Taking into account the fact that merely free metal ions and labile metal complexes could diffuse across physiological membranes,3,12,21 there is a quest for novel methods to assess the kinetics of metal complexation/ dissociation and metal precipitation under simulated GI conditions. In this paper, an automatic continuous-flow system furnished with a concentric microdialysis probe having a molecular weight cutoff (MWCO) resembling the human epithelium is presented for ascertainment of the kinetics of oral bioaccessibility and bioavailability of trace metals in contaminated and agricultural soils using the UBM test with minimal analyst intervention.

gastric and the GI phases have been shown to correlate well with animal bioavailability, BARGE recommended the use of the stomach phase as the most appropriate fluid surrogate for worst-case risk assessment because it is more conservative than the GI phase.10 Physiologically based extraction tests do not, however, take into account the transport of bioaccessible ions across biological membranes and may result in overestimation of the contaminant bioavailability. In fact, bioaccessibility is a necessary precursor of bioavailability but is not, on its own, sufficient for bioavailability to occur.11,12 In vitro bioaccessible tests could therefore serve as conservative tools to estimate oral bioavailability, yet the level of conservancy is for some models deemed too high and questionable.13 To tackle this shortcoming, Pelfrêne et al.14,15 reported a diffusive gradient in thin film (DGT) technique as a model to simulate the human intestinal epithelium and evaluate the absorption of ingested trace elements from highly contaminated soils. DGT devices were deployed in the GI solutions obtained after carrying out the UBM test. Major disadvantages of this approach involve the impossibility to reuse the DGT device, and because of the negligible complexation efficiency of the Chelex 100 resin at low pH, DGT methods cannot be used for monitoring labile metals in the gastric milieu at pH < 1.5. Another in vitro bioavailability method usually applied to estimate trace element and mineral bioavailability in solid substrates (e.g., food commodities) is based on the measurement of the dialyzable fraction, in which a semipermeable 11669

dx.doi.org/10.1021/es401872j | Environ. Sci. Technol. 2013, 47, 11668−11675

Environmental Science & Technology

Article

Spain). The flow manifold was constructed from PTFE tubing of 0.25 mm ID. A multichannel peristaltic pump (Type minipuls 3, Gilson, France) furnished with Tygon tubing of 2.29 mm OD and 1.30 mm ID was used to provide the duodenal and the bile juices to the extraction medium once the gastric digestion was concluded. In order to inject concurrently and precisely the volumes required of both synthetic GI juices in the UBM test, namely, 67 and 22.5 mL of duodenal and bile solutions, respectively, a solenoid valve was incorporated so as to recirculate the surplus of bile juice to the reservoir. A heating and magnetic stirring device coupled to a digital thermoregulator supplied by VELP Scientifica (Usmate Velate Monza e della Brianza, Italia) was employed for the control and the stabilization of the overall temperature of the UBM extraction test at 37 ± 2 °C. All the programmable flow sequences were executed by a personal computer running the lab-made software written in Visual Basic 6.0 (Microsoft, Redmond, WA, USA). The software fosters through an RS232 interface the activation/ deactivation of the microsyringe pump, the control of flow rates, the selection of position in the autosampler and solenoid valve, as well as the activation of the peristaltic pump via the 6pin barrier strip connector allied to a relay. The content of bioavailable lead in the microdialysates was determined by resorting to an electrothermal atomic absorption spectrometer (ETAAS, PinAAcle 900Z, Perkin-Elmer, Waltham, MA, USA) equipped with Zeeman background correction, and a transversely heated graphite atomizer housing pyrolytically coated graphite tubes. The lead hollow cathode lamp was set to an analytical wavelength of 238.31 nm with a slit width of 0.7 nm and an operating current of 10 mA. The ETAAS temperature program for lead assays using 50 μg of NH4H2PO4 and 3 μg of Mg(NO3)2 per assay as combined chemical modifier is shown in Table S4. The manufacturer’s recommendations were slightly modified with the addition of a calcination step using air as a carrier gas so as to eliminate the potential dissolved organic carbon in the samples. Analytical Procedure. Before starting the analytical procedure, 1.5 g of soil was accurately weighed in a 150 mL extraction vessel, where 22.5 mL of simulated saliva was added. The resulting suspension was manually shaken whereupon 33.75 mL of simulated gastric fluid was added and the pH of the resulting mixture adjusted to 1.2 ± 0.1 with HCl. The amount of soil and volumes of the gut fluids specified by BARGE were 2.5-fold increased to ensure that the membrane of the microdialysis probe is entirely immersed in the extraction milieu. The extraction vessel containing the UBM gastric phase and the microdialyzer was placed onto a magnetic stirrer with programmable temperature control to ensure 37 ± 2 °C in the course of the 5-h in vitro digestion. The operational sequence of the automatic analytical protocol including cleaning, filling, and stabilization of the microdialysis probe with the gastric perfusate; collection of the gastric microdialysates; automated addition of duodenal and bile juices followed by cleaning, filling, and stabilization of the microdialysis probe with the GI perfusate and collection of GI microdialysates is explained in detail in the SI. The analytical workflow is schematically illustrated in Figure S1 under SI.

Lead was chosen as model analyte because it has been previously validated by BARGE.10 The microdialysis approach used in this work is a dynamic and in situ sampling technique where the driving force behind the separation is solely the concentration gradient across a semipermeable hydrophilic barrier. The membrane separates the compounds according to their molecular dimension and only low molecular weight species can cross the permselective membrane.22−25 The most salient attributes of the coupling of microdialysis sampling with the UBM test are as follows:26,27 (I) the extraction medium is to be monitored without removal of gut fluids while maintaining the equilibria of the in vitro digestion because the small recipient volume for dialysates ensures minute uptake of species from the gastric and GI phases; (II) the possibility of straightforward automation of the sampling step itself; (III) the exclusion from the perfusate of proteins and humic acid and humin species from the soil matrix and metal chelates of large molecular weight; (IV) the monitoring of passive absorption as a function of time, giving rise to a more realistic insight into the temporal bioavailability of soil-borne metal species in the course of oral ingestion; and (V) as reported by van de Wiele et al.,13 dialysis-based separation methods are deemed most suitable to measure bioaccessible fractions of contaminants in the human gut and closely approach in vivo oral bioavailability. To the best of our knowledge, this is the first time that microdialysis sampling has been allied to the UBM test for investigation of the transport of bioaccessible metal fractions across permselective membranes.

■

EXPERIMENTAL SECTION Reagents, Samples, and Dialysis Membranes. Detailed information on the sampling procedure and physicochemical characterization of soils along with the UBM reagents and the variety of microdialysis hollow fibers assayed are given in Supporting Information (SI) and Tables S1, S2, and S3. Flow System and Instrumentation. The automated flow system for assessment of oral bioavailability of lead in soils is schematically illustrated in Figure 1. It comprises a flowthrough concentric microdialysis probe for in situ sampling of bioavailable lead, and an auxiliary flow system to add the synthetic gut fluids at a given time frame to mimic GI extraction conditions. The custom-built concentric microdialysis probe made of 5cm-long stainless steel capillary of 0.50 mm OD and 0.4 mm ID (G Kinnvall AB, Sweden) contains a tunable 10-cm nonpolar fused silica inner cannula of 0.25 mm ID and 0.38 mm OD. The probe was fitted with a 3-cm-long polysulfone capillary membrane of 0.5 mm ID with averaged MWCO of 30 kDa (see magnified inset in Figure 1). Cyanoacrylate instant glue was used to fix the membrane onto the stainless steel capillary tubing. The outlet of the probe is connected to a 10-cm-long nonpolar fused silica tubing using a short piece of Tygon tube. The perfusion liquid was pumped through the microdialysis probe by resorting to a unidirectional microsyringe infusion pump (CMA 402 syringe pump, CMA/Microdialysis, Stockholm, Sweden) equipped with two 1-mL syringes, one of them containing the gastric perfusion liquid and the other one the GI juice. The connection of both syringes with the concentric microdialyzer was effected via a Y-junction (IDEX Corp., Lake Forest, IL). The microdialysates were automatically collected in a 45-position rack autosampler (Crison Instruments, Barcelona,

■

RESULTS AND DISCUSSION Selection of the Membrane and the Probe Design. Preliminary batchwise investigations using aqueous standard 11670

dx.doi.org/10.1021/es401872j | Environ. Sci. Technol. 2013, 47, 11668−11675

Environmental Science & Technology

Article

water and two saline solutions containing identical electrolyte concentrations to those of the UBM gastric and GI phases (without enzymes), respectively, were tested. When using water as perfusate carrier, the RR of lead in the gastric medium was ≫100%. This is probably a consequence of the ultrafiltration process resulting from unbalanced osmotic pressure at both sides of the dialysis membrane.26 The perfusate water tends to pass across the membrane to equate the ionic strength between the outer and inner probe medium, thereby reducing the diluent volume in the lumen with the consequent analyte preconcentration. This was evidenced with the decrease of the gastric dialysate volume in ca. 55% compared with the nominal volume whenever distilled water was perfused, instead of recoveries >90% for the saline gastric perfusate. Gastric and GI perfusates were thus selected for the remainder of the studies to equilibrate the ionic strength between the outer medium and probe perfusate. To evaluate the percentage of lead which could dialyze across the membrane, the content of lead in the extraction vessel in both the gastric and GI digestion phases was fractionated into two pools: the dissolved fraction (