Environ. Sci. Technol. 2006, 40, 2469-2477
Biomonitoring Human Exposure to Organohalogenated Substances by Measuring Urinary Chlorophenols Using a High-Throughput Screening (HTS) Immunochemical Method MIKAELA NICHKOVA† AND M.-PILAR MARCO* Applied Molecular Receptors Group (AMRg), Department of BiologicalOrganic Chemistry, IIQAB-CSIC, Jorge Girona, 18-26, 08034-Barcelona, Spain
The widespread contamination of the environment by persistent organochlorinated substances is well-known. Highthroughput immunochemical methods may improve routine assessment of the exposure of the population to these chemicals by analyzing urinary biomarkers. Trichlorophenols (TCP) have often been considered as biomarkers of many organochlorinated compounds. With the aim to assess exposure of the population to these substances a high-throughput immunosorbent solid-phase extraction (HTS-IS-SPE) procedure coupled to ELISA for simultaneous analyses of 2,4,6-TCP immunoreactivity equivalents (2,4,6-TCP-IR equiv) in multiple hydrolyzed urine samples has been developed. Around 100 urine samples can be processed simultaneously with an inter- and intraassay precision lower than 23% CV and a limit of detection of 0.3 µg L-1. The analyses by HTS-IS-SPE-ELISA and HTSIS-SPE-GC/MS of urine samples (N ) 117) collected from three different population groups point to a broad exposure of the Catalonian population to organohalogenated substances including the recently emerging organobrominated pollutants. Environment and edible products seem to be the most likely sources of exposure, since excretion of 2,4,6-TCP-IR equiv has been found to be independent from the occupational sector. An excellent correlation was observed between the 2,4,6-TCP-IR equiv determined by HTSIS-SPE-ELISA and the concentrations measured by HTSIS-SPE-GC/MS (R2 ) 0.912). The results show that immunochemical screening methods, based on the quantification of urinary biomarkers, can be excellent tools for exposure assessment. The HTS-IS-SPE-ELISA presented here has proved to be efficient, precise, accurate, rapid, and specific, which opens up the possibility for a broad variety of applications where routine testing of large number of samples is required.
Introduction Biological monitoring is gaining increasing popularity in occupational toxicology and environmental health as an approach for early detection of individual chemical exposure, * Corresponding author phone: 93 4006171; fax: 93 2045904; e-mail:
[email protected]. † Current address: Department of Entomology, University of California-Davis, Davis, CA 95616. 10.1021/es0518629 CCC: $33.50 Published on Web 03/08/2006
2006 American Chemical Society
risk assessment, and prevention of further adverse effects (1). Chlorophenols are well-known environmental pollutants due to their widespread use as plant protection agents, preservatives, and disinfectants, and as intermediates in the chemical synthesis of a variety of pesticides, dyes, and other products. They have been classified by the International Agency for Research on Cancer (IARC) as possibly carcinogenic agents to humans (2). The toxicological effects of trichlorophenols are mainly related to their mutagenic and immunosuppressive properties, dechlorination mechanisms in tissues, and induction of chromosome breakage and aneuploidy, etc. (3). Chlorophenols are among the chemicals with biological exposure indices (BEIs) established by the American Conference of Governmental Industrial Hygienists (ACGIH) (4). The German Commission for Investigation of Health Hazards of Chemical Compounds in Work Area (GCIHHCCWA) has already proposed monitoring trichlorophenol levels in urine as biomarkers of exposure to lindane and hexachlorobenzene (5). The urinary excretion of trichlorophenols has often been used as an indication of the exposure to them, or to chlorobenzenes, hexachlorocyclohexanes, and chlorophenoxy acids that are metabolized and excreted as chlorophenols (6, 7). While 2,4,6-trichlorophenol (2,4,6-TCP) urinary excretion has been reported for pesticide sprayers (8, 9) and workers of waste incinerators (10) and sawmill plants, reaching to levels of 30 µg L-1, the urinary median concentration levels depicted for the general population are usually lower (0.6-2 µg L-1 2,4,6-TCP) (11-14). As chemical precursors of dioxins, chlorophenols can be contemplated as PCDD (polychlorinated dibenzodioxins) indicators (15, 16). In addition, it has been demonstrated that chlorophenols are transformed in vitro to PCDD or PCDF (polychlorinated dibenzofuranes) by a biochemically catalyzed oxidation. This metabolic pathway could lead to a higher inner exposure to PCDD/F after chlorophenol exposure (16). Determination of the actual chlorophenol urinary excretion of the population would provide a reliable estimation of the individual exposure. With this purpose, rapid, inexpensive, and reliable techniques are essential for routine assessment. However, sample preparation is often the bottleneck step for speeding up the overall analysis (17). In the past few years there has been an increasing trend in developing parallel sample processing approaches. Although the 96-well format has been utilized for many years for immunoassays and in vitro binding assays, solid-phase extraction (SPE) devices in a 96-well plate format were introduced later in 1996 and, since then, have enjoyed widespread application and rapid acceptance in biotechnology and pharmaceutical laboratories where high-throughput screening (HTS) is sought. The 96-well SPE formats have been designed to fit automated plate handling systems and to reduce the void volume, as well as the sorbent bed masses, to achieve minimum desorption volumes. Sample preparation in a 96-well format is mainly used in conjunction with liquid chromatography-tandem mass spectrometry (LCMS-MS) (18) or coupled to optical detection (19). The commercialization of 96-channel robotic liquid handling workstations and the wide selection of 96-well SPE sorbents (C2, C8, C18, ion-exchange, etc.) allows rapid development and automation of SPE methods, eliminating traditional timeconsuming and labor-intensive sample preparation development procedures for complex samples (i.e., plasma (20), urine (21), and food (22), etc.). New trends in SPE research are also oriented toward the development of more selective procedures. Immunosorbents VOL. 40, NO. 7, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2469
(IS) based on the specific molecular antigen-antibody recognition (23, 24) allow selective preconcentration of analytes from large sample volumes minimizing the use of organic solvents (25-28). Traditionally IS-SPE is performed sequentially in both online or offline configurations, followed by chromatography (29-31), and to a lesser extent, by immunochemical analysis (32, 33). In this context, we recently have reported an immunochemical method for urinary determination of 2,4,6-TCP based on IS-SPE followed by ELISA quantification (33). The IS-SPE has proven to be an excellent sample preparation method eliminating completely matrix effects including intrinsic individual urine differences. The IS-SPE-ELISA procedure allowed measurement of free (unconjugated) and total chlorophenol content down to 0.66 and 0.83 µg L-1 in nonhydrolyzed and hydrolyzed urine samples, respectively. However, the efficiency was limited by the tedious, manually performed serial immunoaffinity urine extractions. To meet the ever-increasing demands for higher sample throughput in biological monitoring, it is crucial to streamline all aspects of the process, from sample treatment to analyte measurement. There are only a few reports on 96- or 384-parallel IS-SPE (34, 35). In this context the work presented here demonstrates the potential of a HTSIS-SPE-ELISA procedure as a biomonitoring tool to assess human exposure to organohalogenated substances through the immunochemical analysis of chlorophenols urinary excretion.
Experimental Section Apparatus. The pH and the conductivity of all buffers and solutions were measured with a pH meter pH 540 GLP and a conductimeter LF 340, respectively (WTW, Weilheim, Germany). Polystyrene microtiter plates were purchased from Nunc (Maxisorp, Roskilde, DK). Washing steps were performed on a SLY96 PW microplate washer (SLT Labinstruments GmbH, Salzburg, Austria). Absorbances were read on a SpectramaxPlus (Molecular Devices, Sunnyvale, CA). A MD800 gas chromatograph equipped with a mass spectrometer detector (Fisons Instruments, VG, Manchester, UK) was used. Reagents. Chloro- and bromophenols were purchased from Aldrich Chemical Co. (Milwaukee, WI). The preparation of the antisera (As43 and As45) against 2,4,6-TCP by immunization with hapten 5-KLH conjugate (3-(3-hydroxy2,4,6-trichlorophenyl)-propanoic acid coupled to keyhole limpet hemocyanin) has already been reported (36). Isolation of the IgG fraction from the antiserum As45 was accomplished by 35% (NH4)2SO4 precipitation. The IgG fraction was restored with 10 mM PBS and dialyzed against 0.5 mM PBS (3 × 5 L) and Milli Q water (1 × 5 L). The aqueous solution was finally freeze-dried and stored at -30 °C until use. The synthesis of the coating antigen (4-BSA) (3-(2-hydroxy-3,6-dichlorophenyl)propanoic acid) used in ELISA was previously described (37). Buffers. PBS is 10 mM phosphate buffer, 0.8% saline solution, and, unless otherwise indicated, the pH is 7.5. The expression X % EtOH-PBS is PBS with the corresponding percentage of EtOH. PBST is PBS with 0.05% Tween 20. 2X PBST is PBST double concentrated. Borate buffer is 0.2 M boric acid-sodium borate, pH 8.7. Coating buffer is 50 mM carbonate-bicarbonate buffer, pH 9.6. Citrate buffer is a 40 mM solution of sodium citrate, pH 5.5. The substrate solution contains 0.01% TMB (tetramethylbenzidine) and 0.004% H2O2 in citrate buffer. The coupling buffer used in the immunosorbent preparation was 0.2 M NaHCO3, 0.5 M NaCl, pH 8.3. Capping buffer A was 0.5 M ethanolamine buffer (0.5 M NaCl, pH 8.3). Capping buffer B was 0.1 M acetate (0.5 NaCl, pH 4). Materials. N-Hydrohysuccinimide (NHS)-activated Sepharose 4 Fast Flow was purchased from Pharmacia Biotech (Uppsala, Sweden). The gel is based on highly cross-linked 2470
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 7, 2006
4% agarose matrix with a ligand density of 16-23 µmol NHSgroups mL-1 drained medium. Polystyrene microtiter plates used for the ELISA analysis were purchased from Nunc maxisorb (Roskilde, Denmark). VersaPlate 96-Well SPE System (Varian, Palo Alto, CA) consists of a 96-well baseplate, removable 96 empty columns, and a vacuum manifold set. The vacuum manifold was connected to a water pump. The vacuum controller (mechanical gauges) was purchased from Hucoa-Erlo¨ss (Barcelona, Spain). VersaPlate accessories, such as disposable waste reservoir, cartridge removal tool, 20 µm pore frits, 96 glass vials (0.75 mL) in a collection rack, 96-well microplate Teflon-coated silicone rubber seal, sealing tape pads, and sealing caps, were purchased from Varian. High Throughput Immunosorbent SPE (HTS-IS-SPE). Immunosorbent (IS) Preparation. Antibody immobilization was performed by covalently coupling the obtained IgGs (Ab45) to NHS-activated Sepharose 4 Fast Flow as follows: A suspension (1 mL) of the Sepharose was washed with cold 1 mM HCl (12 mL). The drained gel (0.5 mL) was transferred to a tube and mixed with the solution of the purified IgG (20 mg mL-1, 0.5 mL) in coupling buffer. The suspension was left shaking (100 rpm) for 3h at room temperature and finally washed with coupling buffer (3 mL). The nonreacted NHSgroups were blocked for 1 h at room temperature with capping buffer A. After three repetitive cycles of washing alternating the capping buffers A and B (3 × 1 mL), the gel was washed with 10 mM PBS buffer (3 mL). Finally, it was reconstituted to the initial suspension volume (1 mL) and stored at 4 °C in 0.05 M Na2HPO4 (pH 7) containing 0.1% NaN3 until needed. Antibody coupling efficiency to the Sepharose gel was estimated by measuring the concentration of the total IgG initially loaded and the IgG and after coupling by UV at λ ) 280 nm. Previously, the collected solution after coupling of IgG to the Sepharose gel was purified from the coeluting NHS groups using a HiTrap desalting column prepacked with SephadexG-25 Superfine. Preparation of the 96-IS-SPE Rack. A set of 96 VersaPlate empty columns was assembled in the 96-format SPE vacuum manifold (VersaPlate System, Varian). The Ab45 derivatized Sepharose suspension (0.4 mL) was loaded on each column, packed by suction under vacuum and washed with PBS (10 mL) to obtain immunosorbent columns with 0.2 mL of bed volume. The theoretical binding capacity of the columns was estimated according to the amount of IgG coupled, the bivalent binding capabilities of the IgG molecules, and on the assumptions that the IgG fraction contains about 10% of specific IgGs and that around 50% of the antibody molecules may not be accessible due to steric hindrance or to a random antibody orientation. The experimental IS binding capacity was determined by loading PBS samples (6 mL) with different 2,4,6-TCP content (3.6, 12, 100, 500, and 750 ng). General HTS-IS-SPE Procedure. HTS-IS-SPE was performed with the VersaPlate 96-well SPE system, which consists of vacuum manifold set equipped with a vacuum controller and a water pump. The solutions (sample application, washes, and elution) were passed through the immunosorbent under low vacuum maintaining the flowrate in the range of 1-2 mL min-1. All liquid loadings were performed using an 8-channel electronic pipet (Eppendorf). Waste from sample loading, washing, and regeneration steps was collected in a disposable reservoir. Eluted fractions were collected in a rack of 96 glass vials (0.75 mL). When not in use, the VersaPlate 96-well IS-SPE assembly was stored sealed with caps at 4 °C in PBS containing 0.1% NaN3. Each HTSIS-SPE cycle consisted of sample loading, washing, eluting, and regenerating. The columns were brought to room temperature, washed with PBS (5 mL), and preconditioned with 70% EtOH (1.2 mL) and PBS (1.2 mL). After loading the sample (spiked PBS or urine, 6 mL) the columns were washed with PBS or 20% EtOH-PBS (1.2 mL). The bound analyte
was eluted with 70% EtOH-PBS (0.6 mL). Subsequently, the columns were regenerated with PBS (1.2 mL). Eluted fractions were analyzed by ELISA and/or by GC-MS as described below. ELISA. Unless otherwise indicated, the fractions eluted after HTS-IS-SPE cleanup were diluted 10 times with PBS and measured by ELISA following the protocol described previously (37) using 2,4,6-TCP standards prepared in 7% EtOH-PBS. The standard curve was fitted with a 4-parameter logistic equation using the software SoftMaxPro v2.6 (Molecular Devices) and GraphPad Prism (GraphPad Software Inc., San Diego, CA). Unless otherwise indicated, data presented correspond to the average of at least two well replicates. The main analytical parameters of the immunoassay are as follows: IC50 ) 1.13 µg L-1, LOD ) 0.2 µg L-1, dynamic range from 0.3 to 3.12 µg L-1, slope ) 1.2. The most important cross-reactants are the following chlorinated and brominated phenols: 2,4,5-TCP (12%), 2,3,4,6-TtCP (21%), 2,4-DBP (119%), and 2,4,6-TBP (710%) (37). GC-MS. The fractions (0.5 mL) eluted after the HTS-ISSPE procedure fractions were diluted 10 times with PBS, acidified to pH 3 with 1 N HCl, and extracted with toluene (0.5 mL) by agitating the mixture on an automatic shaker at 300 rpm for 50 min. The organic layer was separated by centrifugation and the toluene extracts (0.1 mL) were derivatized by BSTFA (bis-trimethylsilyl)trifluoroacetamide, 2 µL) for 2 h at room temperature. 2,3,5,6-Tetrachloronitrobenzene was used as internal standard (ISt). The toluene extraction efficiency of the halogenated phenols ranged between 70 and 107% with CVs less than 14% (33). Injections (1 µL) were splitless (48 s) with solvent delay (5 min). A HP5MS (cross-linked 5% Ph Me siloxane, 30 m × 0.25 mm .d. × 0.25 µm (film thickness) column was used. The ion source temperature was set at 200 °C, and He was the carrier gas employed at 1 mL/min. The ionization energy was 70 eV. GC conditions were as follows: temperature program, 100 to 300 °C (7 °C/min); injector temperature, 250 °C; and the ions 253/255/268 (2,4,6-TCP and 2,4,5-TCP), 307/309/324 (2,4DBP); 387/389/402 (2,4,6-TBP); 287/289/304 (2,3,4,6-TtCP); 203/259/261 (ISt) were monitored in the SIM mode. The mass spectrometer was operated in the time varied selected ions monitoring mode (i.e., with different group ions monitored for various time specific absolute intervals: 2,4,6-TCP and 2,4,5-TCP, 253, 255, 268 m/z, 8-10 min; 2,4-DBP, 307, 309, 324 m/z, 10-11 min; 2,3,5,6-tetrachloronitrobenzene, 203, 259, 261 m/z, 11-12 min; 2,3,4,6-TtCP, 287, 289, 304 m/z, 12-13 min; 2,4,6-TBP, 387, 389, 402 m/z, 13-16 min). For the SCAN mode, the mass range explored was 45-550. All data is reported as m/z (relative intensity). The urinary LOD of the HTS-IS-GC-MS method was 0.1 µg L-1 for 2,4,6-TCP and 2,4,5-TCP and 0.2 µg L-1 for 2,4-DBP, 2,3,4,6-TtCP, and 2,4,6-TBP. Samples. A pooled urine sample from different individuals (tested negative for specific drugs of abuse, 0.1% NaN3) was purchased from Bio-Rad Laboratories (Irvine, CA). According to ELISA and GC-MS analyses performed previously (33), the 2,4,6-TCP-IR equiv was 1.12 µg L-1, corresponding to 2,4,6-TCP (0.2 µg L-1), 2,3,4,6-tTCP (0.12 µg L-1), and 2,4,6TBP (0.12 µg L-1). Individual urine samples from workers of the industrial sector (petrochemical company, Tarragona, N ) 43, collected in 2001) and administrative sector (Barcelona, N ) 41, collected in 2002), as well as from population living in the vicinity of a municipal incinerator (Mataro´, N ) 33, collected in 1999) were kindly provided by the Instituto de Higiene y Seguridad en el Trabajo (Ministerio del Trabajo y Asuntos Sociales) and by the Centre de Seguretat i Condicions de Salut en el Treball (Generalitat de Catalunya). All urine samples had been aliquoted and stored at -30 °C until use. The concentration of creatinine was determined by the Jaffe´ method (38).
TABLE 1. Recoveries Obtained for the HTS-IS-SPEa 2,4,6-TCP ng
µg L-1
Nb
recovery ( SD, %
750 500 100 12 3.6
125 83.3 16.6 2 0.6
5 10 3 16 16
69.7 ( 5.67 101.05 ( 4.74 84.81 ( 13.25 97.75 ( 16.84 95.77 ( 26.04
a PBS solutions (6 mL) of 2,4,6-TCP were loaded on the 96-SPE rack, washed with PBS (1.2 mL), and eluted with 70% EtOH (0.6 mL). The fractions eluted were diluted 10 times with PBS and measured by ELISA. b N is the number of the columns used. Columns were placed in different positions of the rack to assess reproducibility.
Quality Controls (QCs). Performance of the HTS-IS-SPE (sorbent capacity, precision, accuracy) was assessed by passing standard solutions (6 mL). Buffer QCs were standard solutions of 2,4,6-TCP at different concentration levels (83, 16.6, 8, 2, and 0.7 µg L-1) prepared in PBS buffer. Urine QCs (8, 2, and 0.7 µg L-1 2,4,6-TCP) were prepared in hydrolyzed pooled urine sample (Bio-Rad Laboratories, Irvine, CA). All stock and QC solutions were stored at 4 °C. Sample Preparation. Urine samples were hydrolyzed under alkaline conditions as previously described (33). KOH (15 M, 2 mL) was added to the urine samples (10 mL), and they were heated on a sand bath for 30 min at 100 °C. Then, the samples were neutralized to pH 7.5 by adding concentrated H2SO4 and centrifuged to eliminate the precipitate formed.
Results and Discussion Evaluation of the HTS-IS-SPE Procedure. Immunosorbents (IS) were prepared by covalently linking the IgG fraction of As45 to the Sepharose gel with a coupling efficiency of 97%. A set of 96 minicolumns (200 µL drained gel bed) placed on a rack were packed with this immunosorbent. A theoretical maximum binding capacity of 500 ng (2.22 nmol) 2,4,6-TCP was estimated for each minicolumn attending the amount of antibody immobilized. Further experiments performed loading PBS solutions (6 mL) containing different 2,4,6-TCP concentrations proved that the experimental binding capacity, in buffer, was also around this value (see Table 1). It is also important to notice the reproducible behavior observed among the different minicolumns, independently from their position in the 96-section rack, despite the manual packing process. Thus, a %CV of 4.69 (N ) 10) was recorded when 500 ng of 2,4,6-TCP was loaded. At the lower concentration level (0.6 µg L-1) tested the CV was higher (27.19%, N ) 16). Elution and washing conditions were investigated with the aim to know the retention capabilities of the IS. A certain EtOH content in the PBS of the washing step could help wash out potential nonspecific urine interferences coeluting with the analyte while improving the selectivity of the cleanup method (33). To prove this fact, a step elution experiment increasing the EtOH concentration from 10% to 70% was performed loading 2,4,6-TCP PBS samples (6 mL) at three concentration levels (83.3, 33.3, and 8.3 µg L-1). The results obtained demonstrated that the ethanol content of the washing step had to be kept below 20% to ensure recovery of the analyte in the elution step. Under these conditions, particular halophenols were retained by the IS on a significant percentage (2,4,6-TCP, 80%; 2,3,4,6-TtCP, 80%; 2,4,6-TBP, 70%; and PCP, 25%), while the recoveries for the rest of the halophenols tested remained below 10%. PBS samples with 2,4,6-TCP concentrations of up to 33 µg L-1 (200 ng in 6 mL of PBS) could be loaded, reaching acceptable recoveries (>80%) applying this procedure. These levels of 2,4,6-TCP urinary excretion are only expected in population groups with a high probability of exposure. VOL. 40, NO. 7, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2471
The analytical HTS-IS-SPE-ELISA protocol established ((i) Sample loading, 6 mL of hydrolyzed urine; (ii) Washing, 1.2 mL of 20%EtOH-PBS; (iii) Elution, 0.6 mL of 70% EtOHPBS and dilution to 6 mL with PBS for ELISA analysis) results in no preconcentration, nor dilution, of the analyte concentration in the raw urine sample. Therefore, the LOD of the protocol is determined by that of the ELISA (IC50 ) 1.13 µg L-1, LOD ) 0.2 µg L-1) and the immunoextraction recovery values.
FIGURE 1. Effect of sample volume and loading level on the recovery of the HTS-IS-SPE procedure. Alkaline hydrolyzed pooled urine sample was spiked with different amounts of 2,4,6-TCP (200, 75, and 20 ng corresponding to 40%, 15%, and 4% loading levels) and loaded at different volumes to the columns, washed with 20% EtOH, and eluted with 70% EtOH. The 2,4,6-TCP concentrations in the purified urine were measured by ELISA. Recovery values were calculated with respect to the spiked amount subtracting the background level of 2,4,6-TCP-IR in the urine sample. The standard deviation corresponds to the average of three HTS-IS columns processed simultaneously. To evaluate performance of the HTS-IS-SPE procedure to determine total (free and conjugated) chlorophenol urinary excretion, a commercial urine sample pooled from different individuals was used. This sample had been characterized by GC-MS for its content of halophenols (chloro- and bromophenols, see Experimental Section for description) (33) revealing that it contained a concentration equivalent to 1.12 µg L-1 of 2,4,6-TCP (2,4,6-TCP-IR equiv). Considering the actual recoveries of the present protocol the expected response in 2,4,6-TCP-IR equiv would be 0.81 µg L-1. Thus, prior to HTS-IS-SPE, the urine was hydrolyzed and spiked with 2,4,6-TCP at three concentration levels (33.3, 12.5, and 3.3 µg L-1). Each spiked sample was then loaded at three different volumes (6, 12, and 18 mL) into the HTS-IS 96-rack, washed with 20% EtOH-PBS (1.2 mL), and the fractions eluted with 70% EtOH/PBS (0.6 mL) were measured by ELISA. The binding capacity of the HTS-IS minicolumns was similar to the experimental capacity determined in buffer. As it can be observed in Figure 1, the breakthrough volume was found to be around 12 mL (>80% recovery) for urine samples with 2,4,6-TCP concentration values below 33.3 µg L-1. At this concentration level the recovery was 70% but almost quantitative when 6 mL of hydrolyzed urine were loaded.
Inter- and intra-day precision and accuracy of the HTSIS-SPE-ELISA procedure were evaluated by simultaneously anlyzing 96 samples composed of nonspiked pooled urine samples (N ) 15), buffer quality controls (bQC, N ) 9), and urine quality controls (uQCs, N ) 72). The analysis was repeated on three consecutive days varying the sample distribution through the 96-well SPE system. Both, bQC and uQCs were prepared at three 2,4,6-TCP concentration levels, within the range of the most frequently found values in the general and occupational exposed population groups: low (0.7 µg L-1), medium (2 µg L-1), and high (8 µg L-1). The data summarized in Table 2 show the good accuracy (recoveries >80%) and reliability of the HTS-IS-SPE-ELISA method. Both inter- and intra-assay precision (% CV) were lower than 20% except when the measurements took place at the limit of detection, where they were slightly higher (23%). The 2,4,6TCP-IR equiv measured in the nonspiked urine sample was 0.81 ( 0.13 µg L-1 (N ) 12, 3 different days and 4 replicates each day), in accordance with the expected results (see above). Immunosorbent regeneration and stability were assessed by regularly passing buffer QCs through the minicolumns to test the binding capacity and the performance of the HTS-IS within several applications of urine samples. The buffer QCs used were PBS solutions of 2,4,6-TCP at three concentration levels corresponding to 100%, (HQC, 83.3 µg L-1), 20% (MQC, 16.6 µg L-1), and 2% (LQC, 1.6 µg L-1) of the maximum capacity determined initially. Figure 2 shows the average long-term stability of three HTS-IS columns processed simultaneously. As it can be observed, quantitative recovery (in the range of 80-100%) was maintained for up to 35 analyses for all levels showing that repetitive loading of urine samples does not adversely affect the IS capacity and performance. Moreover, the IS minicolumns are efficiently regenerated after each cycle and therefore can be used for several series of chlorophenol extractions from urine samples. Validation of the HTS-IS-SPE-ELISA procedure by GCMS was performed analyzing urine samples (N ) 117) from different individuals. The urine samples were hydrolyzed and processed following the HTS-IS-SPE protocol described above. The eluted fractions were diluted 10 times with PBS and split into two parts: one part (0.5 mL) was measured directly by ELISA and the other part (5 mL) was acidified, extracted with toluene, derivatized, and analyzed by GCMS. All samples could be processed several times due to the high efficiency of the HTS-IS-SPE and ELISA procedures.
TABLE 2. Precision and Accuracy of the 2,4,6-TCP Analysis in Urine by the HTS-IS-SPE-ELISA Method urine QCs
buffer QCs
spiked concentration, µg L-1 mean µg L-1 mean recovery, %a interday precision (% CV)b intraday precision (% CV)c
spiked concentration, µg L-1
Nd
0.7
2
8
Nd
0.7
2
8
72 72 24 72
0.61 87.5 22.9 11.3
1.55 77.4 20.6 6.20
7.06 88.3 17.4 10.4
9 9 3 9
0.65 92.7 17.1 5.40
1.74 87.2 19.2 9.35
7.62 95.3 13.7 12.5
a The recovery for urine QCs is calculated as [(pre-extracted spiked urine)/(post-extracted spiked urine)] × 100; The recovery buffer QCs is calculated as (measured/nominal value) × 100. b The % CV is the average of the % CVs for each concentration for each day (within one HTS-IS--SPE procedure). c The % CV corresponds to the recovery obtained for each concentration for 3 days (between three 96-SPEs). d N is the number of the samples used for each concentration.
2472
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 7, 2006
FIGURE 2. Stability of the HTS-IS after several cycles of use loading urine samples. Buffer QCs (500, 100, and 10 ng 2,4,6-TCP) were regularly loaded into the columns and the eluted fractions were analyzed by ELISA. Data correspond to the average of three IS columns processed simultaneously in the HTS-IS-SPE procedure. The only limitation was the sequential injections into the GC-MS chromatograph and the data processing. The correlation between the values determined by ELISA (ELISA 2,4,6-TCP-IR equiv) and the concentration of 2,4,6-TCP determined by GC-MS (GC-MS 2,4,6-TCP) is presented in Figure 3A. Almost all samples tested positive for 2,4,6-TCPIR equiv with a significant overestimation by the ELISA method. This result was not surprising, because of the presence and ELISA cross-reactivity of other structurally related halophenols co-extracted from the urine samples (37). Thus, usually chlorophenols are used and produced as technical mixtures. Figure 3B shows the excellent correlation (y ) 1.14x - 0.21 with a regression coefficient R2 ) 0.912) between both techniques when the concentration of the individual halophenols recorded by GC-MS (2,4,5-TCP, 2,4DBP, 2,3,4,6-TtCP, and 2,4,6-TBP) were converted into 2,4,6TCP IR equiv according to their crossreactivity (% CR) values, demonstrating that the HTS-IS-SPE-ELISA procedure can be an excellent biomonitoring tool. Table 3 summarizes the main features of the HTS-ISSPE-ELISA method. The significant increase in sample throughput, excellent detectability, and the reduced urine sample volume required for the assays are some of the main advantages. More than 100 samples can be processed per day with a overall LOD of 0.3 µg L-1 and a MQC of 30 µg L-1 2,4,6-TCP-IR equiv. Concentrations at, or exceeding, the MQC indicate that further analyses reducing the sample volume will be required to ensure accurate analysis. About 6 mL of urine is sufficient to provide enough ELISA replicates and to perform GC-MS confirmation with the samples testing positive in the ELISA. The evaluation studies have shown the absence of false negative or false positive results. Human Exposure Assessment Model Study using HTSIS-SPE-ELISA. As part of the evaluation, three population groups belonging to two different working sectors (petrochemical industry, N ) 43, and administration, N ) 41) and to a vicinity close to an incinerator plant (N ) 33) were biomonitored using the established HTS-IS-SPE. Since our aim was the validation of the screening method more than performing a rigorous exposure assessment, no additional information on the health status and particular habits of the individuals was provided to us at the time this study was performed. The urines were treated following the procedure described above. Each urine sample was extracted in duplicate following the HTS-IS-SPE procedure and analyzed by ELISA and by GC-MS. Regarding frequency of detection (positive sample vs total number of samples), practically all urine samples had
FIGURE 3. Validation of the HTS-IS-SPE-ELISA procedure by GCMS. Urine samples from different individuals (N ) 117) were hydrolyzed by KOH and their 2,4,6-TCP concentration was determined by HTS-IS-SPE-ELISA. Cross-reactants in purified urines were identified and quantified by GC-MS. (A) Correlation between 2,4,6TCP-IR equiv measured by ELISA and 2,4,6-TCP concentration determined by GC-MS. (B) Correlation between ELSIA 2,4,6-TCP-IR equiv and GC-MS TCP-IR equiv defined as ∑(%CR × cross-reactant concentration determined by GC-MS).
TABLE 3. Features of the Urinary Analysis of 2,4,6-TCP by IS-SPE-ELISA and HTS-IS-SPE-ELISA IS-SPEELISAe(33)
HTS-IS-SPEELISAf
10 mL 3-fold dilution 96 samples/8 days 96 samples/9 days 0.99 1.4 20 5.1% (5) 18.4% (30)
6 mL no dilution 96 samples/1 h 96 samples/1 day 0.3 0.55 30 17.4% (0.7) 22.9% (8)
parameters sample volume cleanup speed of IS cleanup speed of total analysis LODa, µg L-1 LOQb, µg L-1 MQCc, µg L-1 % CVd (2,4,6-TCP, µg L-1)
number of false positives 0 number of false negatives 0
0 0
a The limit of detection is evaluated according to the LOD of the ELISA (LOD7%EtOH/PBS ) 0.2 µg L-1, 90% of the signal at zero analyte concentration) and the corresponding IS-SPE and hydrolysis recoveries. b The limit of quantification is evaluated according to the LOQ of the ELISA (LOQ7%EtOH/PBS ) 0.37 µg L-1, 80% of the signal at zero analyte concentration) and the corresponding IS-SPE and hydrolysis recoveries. c MQC ) maximum quantifiable concentration with recovery >70%. d The coefficient of variation (% CV) corresponds to the interday analysis. e Immunosorbents consisted of 1 mL bed NHS-activated high-trap columns derivatized with Ab45. f Immunosorbents consisted of 200 µL bed minicoluns of NHS-activated Sepharose derivatized with Ab45.
detectable concentrations of 2,4,6-TCP IR equiv. Detailed data of the results obtained given in µg g-1 creatinine are VOL. 40, NO. 7, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2473
TABLE 4. Urinary Levels of 2,4,6-TCP-IR Equivalents (Mean Concentration, Range, Concentrations at Selected Population Percentiles) for the Three Groupsa population group
N
f,%b
mean ( SD
min
5%
25%
median
75%
95%
max
administration petrochemical vicinity incinerator
41 43 33
92 92 100
1.33 ( 0.78 1.90 ( 1.68 3.86 ( 4.36
0.42 0.18 0.59
0.61 0.65 0.78
0.85 0.92 1.54
1.13 1.33 2.19
1.44 2.02 3.93
2.67 5.68 14.47
4.69 8.06 22.69
a The concentrations were determined by HTS-IS-SPE-ELISA and expressed in µg g-1 creatinine. The detection limit is 0.17 µg g-1 creatinine based on the LOD of the ELISA (0.2 µg L-1) and the mean urinary creatinine concentration of all the subjects (1.24 g L-1 urine). b Frequency of immunoreactive samples expressed in percentage
TABLE 5. Urinary Levels of All the Analytes (Mean Concentration, Range, Concentrations at Selected Population Percentiles) for the Three Population Groupsa analyte/group 2,4,6-TCP administration petrochemical vicinity incinerator 2,4,5-TCP administration petrochemical vicinity incinerator 2,3,4,6-TCP administration petrochemical vicinity incinerator 2,4-DBP administration petrochemical vicinity incinerator 2,4,6-TBP administration petrochemical vicinity incinerator
N
f,%b
mean
37 41 29
89 100 86
37 41 29
min
5%
0.18 0.74 0.43
