Partition-Controlled Delivery of Toxicants: A Novel In Vivo Approach

In conventional static or semi-static embryo toxicity assays with fish, the nominal concentrations of hydrophobic chemicals are often used to establis...
12 downloads 0 Views 71KB Size
Download PDF
Environ. Sci. Technol. 2003, 37, 2262-2266

Partition-Controlled Delivery of Toxicants: A Novel In Vivo Approach for Embryo Toxicity Testing YIANNIS KIPARISSIS,* PARVEEN AKHTAR, PETER V. HODSON, AND R. STEPHEN BROWN* School of Environmental Studies, Queen’s University, Kingston, Ontario, Canada, K7L 3N6

In conventional static or semi-static embryo toxicity assays with fish, the nominal concentrations of hydrophobic chemicals are often used to establish the toxic thresholds, which often far exceed the solubility limits of test compounds. Saturators and continuous-flow diluters have been used to provide stable concentrations below solubility but are complex, use large amounts of test substance, and produce large volumes of waste. We present a partition-controlled delivery (PCD) method that maintains the concentrations of chemicals in test solutions at or below solubility limits for extended exposure times. Concentrations are maintained by equilibrium partitioning of test chemicals from a series of poly(dimethylsiloxane) films loaded with a range of concentrations of each chemical. The efficacy of the PCD assay was tested by comparisons with static (no renewal) and semi-static (24-h renewal) embryo-larval toxicity tests. The test species was Japanese medaka (Oryzias latipes) exposed to retene (7-isopropyl1-methylphenanthrene), a compound causing blue sac disease (BSD) in fish embryos. In the PCD assay, the median effective concentration (EC50) for BSD was 10 µg/L, below retene’s solubility of 17 µg/L. In contrast, the nominal EC50 values for the semi-static 24-h and static assays were about 10 (150 µg/L) and 150 times (2500 µg/L) greater than solubility, respectively. The PCD method is a more sensitive and realistic method for assessing toxicity of nonpolar compounds than (semi)-static assays.

Introduction The toxicity of complex environmental mixtures and individual chemicals is often assayed by tests with embryolarval stages of fish. Embryo-larval tests are ecologically relevant because fish eggs may be exposed continuously to toxicants from the sediments or water column at contaminated sites, and effects may result in changes in recruitment, a population level impact. For instance, after the Exxon Valdez oil spill, larvae of pink salmon and Pacific herring from oiled spawning shoals developed blue sac disease (BSD), a syndrome contributing to high mortality rates (1-3) and reduced recruitment. In addition, embryo toxicity assays offer a surrogate to life-cycle toxicity testing because early life * Address correspondence to either author. (Y.K.) phone: (613)533-6000; fax: (613)533-7716; e-mail: [email protected]. (R,S,B.) phone: (613)533-2655; fax: (613)533-6669; e-mail: browns@ chem.queensu.ca. 2262

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 10, 2003

stages of fish are very sensitive to pollutants and good predictors of effects in complete life-cycle tests (4). Embryo toxicity assays can be completed within 2 weeks and use small amounts of test solutions. By comparison to life-cycle tests, a greater number of chemicals can be tested, costs can be reduced, and less handling of hazardous chemicals is involved (5). There are several exposure protocols for embryo toxicity tests, usually static (nonrenewal of test solutions) or semistatic (e.g., daily renewal) (6-9). These protocols do not require specialized instrumentation, and there is no substantial disturbance of fish embryos during observations (10). However, for testing hydrophobic or easily degraded chemicals, excess concentrations (up to severalfold above solubility) must be added to test solutions to cause adverse responses in test organisms. Between additions, concentrations decline with absorption or degradation in test chambers, which makes the estimation of toxic thresholds problematic (e.g., see Basu et al.; 11). Consequently thresholds are often calculated from nominal rather than actual concentrations (6-9, 12), and toxicity is wrongly estimated. To overcome this uncertainty, chemicals dissolved in carrier solvents are sometimes deposited topically on fertilized eggs (i.e., topical application exposures) for a brief period (13) or are injected (i.e., injection exposures) (14, 15). In these protocols, embryos are exposed to known doses of chemicals. However, topical application exposures imply a relationship between egg chorion permeability and octanol-water partition coefficient (Kow) for the test chemicals (13), and injection methods are invasive and require sophisticated instruments and experienced personnel (14, 16). More importantly, toxicity estimates based on a single “dose received” provide little information about the consequences of continuous exposures from water. A more complex scheme for providing stable concentrations uses a combination of a saturator and continuous-flow dilution (17). The saturator contains test compound in pure solid form or adsorbed onto a porous support such as glass wool in a column. Water flows continuously through the column, and a saturated concentration of the compound emerges. The saturator output is mixed continuously at various flow ratios with a dilution water to generate various concentrations up to saturation. A stable concentration of test compound is provided in the toxicity test apparatus after passage of several solution volumes, and continuous replacement results in a controlled concentration over the time of the test; however, these systems are quite complex, especially the constant-flow and mixing components. Large quantities of test compound are required to provide saturation with continuous flow in a many-day exposure, and large volumes of contaminated solution are generated that must be properly disposed. As a result, this approach is not widespread, and a simpler method for providing stable solutions is required for routine use. An alternative method for controlling aqueous concentrations of sparingly soluble compounds based on partition from a hydrophobic phase has been suggested (18). Recently, poly(dimethylsiloxane) (PDMS) films were used to deliver polycyclic aromatic hydrocarbons (PAHs) to aqueous solutions at or below their solubility limits for toxicity tests with microorganisms (19). The partition-controlled delivery (PCD) system generated a rapid equilibrium between films and aqueous solutions for a number of compounds with Kow values between 2.8 and 6.1, and it was versatile and reliable (19). The purpose of the present study was to adapt the PCD technique for fish embryo toxicity tests and to determine its 10.1021/es026154r CCC: $25.00

 2003 American Chemical Society Published on Web 04/04/2003

FIGURE 1. Molecular structure of retene (7-isopropyl-1-methylphenanthrene): CAS Registry No. 483-658. reliability and usefulness by comparing it with static nonrenewal and semi-static 24-h renewal assays. A continuousflow system was not used for reasons of complexity and cost, as described above. Retene, an alkyl-substituted phenanthrene, was used as the model compound because it is very toxic to embryo-larval stages of different fish species (9) and occurs in aquatic environments near some pulp mills in relatively high concentrations (20). Embryos from Japanese medaka (Oryzias latipes) were used as the test organism because they are visible through a transparent chorion, their embryonic development is well-characterized, and teratogenic responses are comparable to those observed in embryos from other fish species (10).

Experimental Section Design of Embryo Toxicity Testing. The efficacy of PCD as an exposure method was assessed by comparison to two widely used exposure protocols of static bioassays (one addition of chemical with no solution renewal) and semistatic bioassays (renewal of chemical and test solution every 24 h). In semi-static tests with juvenile rainbow trout, waterborne retene concentrations declined exponentially by more than 95% within 24 h of creating a new solution (11). Hence, static exposure systems for 17-day medaka tests should be characterized by one spike of high concentration in the first 24 h, followed by 16 days of low and declining concentrations. For semi-static tests, exposure should be characterized by 17 daily high concentration spikes with exponential declines to low concentrations. In contrast, PCD exposures should be characterized by steady-state concentrations of retene throughout the 17-day exposure. To verify this model, we used the PCD exposure regime in a 17-day medaka embryo toxicity test with periodic analyses of waterborne retene (Figure 1) and then calculated the EC50 for BSD based on these measured concentrations. We did not measure retene concentrations in the parallel static and semi-static embryo toxicity tests because of previous experience with juvenile trout (11), which demonstrated the difficulty in generating a stable exposure estimate despite a high frequency of sampling. There were no clear advantages to using nominal concentrations, average exposure characteristics (arithmetic or geometric means), or areas under the curve (concentration × time) in the previous test. Hence, we used nominal concentrations as indicators of contaminant level in static and semi-static tests, as is the standard procedure for most reported tests. Chemicals and Reagents. Retene (98% purity) was obtained from ICN Biomedicals Inc (Costa Mesa, CA). The embryo rearing solution (1 mL of 10% NaCl, 1 mL of 0.3% KCl, 1 mL of 0.4% CaCl2‚2H2O, 1 mL of 1.63% MgSO4‚7H2O, and 95 mL of glass-distilled H2O) was purchased from Carolina Biological Supply Company, Burlington, NC. Poly(dimethylsiloxane) (PDMS) aquarium sealant (Dow Corning Corporation Midland, MI) was purchased from a local pet store. Hexane, ethanol, and acetone were obtained as HPLC grade solvents from Fisher Scientific (Toronto, ON, Canada). Japanese Medaka Embryo Toxicity Assays. Three different exposure protocols were used to determine the embryo toxicity of retene. In each approach, fertilized Japanese medaka eggs from several females were collected, separated

FIGURE 2. Experimental setup for the partition-controlled delivery (PCD) embryo toxicity assay (ERS, embryo rearing solution; PDMS, poly(dimethylsiloxane)). from each other, and thoroughly mixed in a Petri dish. Embryos were placed into experimental vials or beakers and examined under a dissecting microscope to ensure that no damage occurred during handling. During the 17-day toxicity test, the embryos were examined at the same time each day to determine the stage and pattern of development and the occurrence of lesions, hatching success, and mortality. The ambient temperature throughout the experiments was 25 ( 2 °C. Preparation of PDMS Films. A PDMS solution was prepared for film deposition by dissolving 6 mg/mL uncured PDMS in hexane. A solution volume of either 600 or 1000 µL was deposited in each 20-mL vial. The vials were in a holder at 45° so that only one side was coated with PDMS. The hexane was evaporated, and the residual film was cured at room temperature for about 1 h. A series of cured films were postloaded with retene by the addition of 600 µL of hexane solution containing sufficient retene to provide final film concentrations of 1, 3, 9, 19, 37, and 190 mg/g. After evaporation of hexane, the vials containing films were filled with 15 mL of embryo-rearing solution. The vials were covered with aluminum foil, placed on an orbital shaker at 700-900 rpm, and left at ambient temperature for 45 min to establish equilibrium of retene concentrations between the PDMS film and the solution. Each vial was rotated one-half-turn in the 45° holder, and five medaka eggs were placed in each vial (n ) 50 for each treatment) on the side opposite to the PDMS film. Hence, eggs were not in contact with the film, and effects could be attributed only to water dissolved retene (Figure 2). At various times throughout the experiment, 2.5 mL of reteneequilibrated solution was removed and mixed with 2.5 mL of ethanol to measure aqueous concentrations. The retene aqueous/ethanol solutions were sonicated prior to HPLC analysis. A Varian HPLC system with ProStar 240 quarternary gradient solvent delivery module, ProStar 360 fluorescence detector, and ProStar 430 robotic auto sampler were used for sample analysis, and Star Chromatography Workstation software was used for data acquisition and analysis (Varian, Mississauga, ON, Canada). A C18 Zorbax column (25 × 0.46 cm; Chromatographic Specialties, Brockville, ON, Canada) was used for the separation of retene with a 1.25-cm C18 guard column placed before the analytical column. The injection volume was 50 µL, and a gradient program was used to separate analytes. Detection of retene was carried out at 254/370 nm excitation and emission wavelengths, respectively. Static Nonrenewal Embryo Toxicity Assay. The embryo toxicity of retene was estimated with a slightly modified version of the protocol described by Wisk and Cooper (6) and Harris and co-workers (7). Retene was dissolved in acetone to make stock solutions, and 10-µL aliquots were delivered into 7-mL scintillation vials prior to testing. In VOL. 37, NO. 10, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2263

FIGURE 3. Partition-controlled delivery of retene provides aqueous solution concentrations (CS) in proportion to a range of retene concentrations in the PDMS film (Cf). Error bars indicate 95% confidence intervals (n ) 3). The curved line is a fit of an asymptote with log-transformed Cf data. The inset panel shows expansion of the 0-10 mg/g region with linear regression of day 1 data. control vials, only 10 µL of acetone was added. The acetone was allowed to evaporate in the fume hood, and 2 mL of embryo-rearing solution was added. Nominal concentrations of retene were 0 (control), 100, 1000, 2500, 5000, and 10 000 µg/L. In each of 10 exposure vials per treatment, two eggs were added. Semi-Static 24-h Renewal Embryo Toxicity Assay. Briefly, 50-µL aliquots from different acetone stock solutions of retene were delivered into 400-mL beakers having 200 mL of dechlorinated Kingston tap water (hardness ) 135 mg/L CaCO3; alkalinity ) 90 mg/L CaCO3; pH 7.8; conductivity ) 280 µS/cm; oxygen concentrations >80% saturation). The nominal concentrations of retene in the experimental beakers were 0, 10, 32, 100, 320, 560, and 1000 µg/L. Fertilized medaka eggs (n ) 30) were placed into each beaker, and the test solutions were changed every day until the end of the experiment.

Results and Discussion BSD Syndrome. The developmental defects in Japanese medaka embryos as a consequence of retene exposure were similar for all three exposure systems. Signs assessed included impaired circulation, pericardial and peritoneal edemas, and craniofacial deformities, described elsewhere as dioxin-like (21-23) or as BSD in larval rainbow trout (Oncorhynchus mykiss) and zebra fish (Danio rerio) (9). To simplify data analysis and presentation, the presence of at least one of the hallmark signs was taken as an indication of BSD. The patterns in developmental timing and intensity of BSD signs were similar to those noted in medaka exposed to several PAHs (6, 7, 24, 25) and to natural or synthetic flavonoids such as flavone and β-naphthoflavone (26). Characterization of PCD of Retene. The use of PDMS films to deliver retene to aqueous solutions was characterized prior to experiments with medaka eggs. The criteria for successful delivery were to reach equilibrium concentrations rapidly and to maintain those concentrations for the 17-day duration of the test. As observed previously (19), equilibrium retene concentrations were established within 45 min with orbital shaking. Equilibrated concentrations varied linearly with film concentrations from 0 to about 15-20 µg/L, presumably the solubility limit (Figure 3). The aqueous solubility of retene is not reported in the literature but has been calculated to be 17 µg/L at 25 °C ((software KOWWIN v 1.65, W. Meylan, 1993-1995, Syracuse Research Corp., Syracuse, NY). The film: solution partition constant (Kfs) was determined from the linear regression slope for the first five data points (see inset in Figure 3, Kfs calculated after conversion to common units), 2264

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 10, 2003

FIGURE 4. Partition-controlled delivery of retene in vials containing medaka eggs. Measured solution concentrations (CS) before (day 7) and after (day 17) hatching of the eggs are compared with the expected CS for the same films in solutions without eggs (as in Figure 3). The Y-X line is drawn to indicate agreement between measured and expected values. giving a log Kfs value of 5.8. This is higher than our previous value for phenanthrene (4.16) and similar to the value for benzo[a]pyrene (5.47), consistent with the higher molecular weight of retene as compared with phenanthrene. For comparisons, the estimated log Kow value for retene has been reported as 6.4 (11). Figure 3 also shows solution concentrations after 17 days. Variations over this time period are likely the result of aging effects of the films and possibly recrystallization of retene during film loading. The higher concentration films were visibly cloudy, consistent with the presence of small crystals. The most pronounced differences between day 1 and day 17 values were for the 19 and 37 mg/g films, where the solution concentration dropped by 20% and 15%, respectively, over this time. Further characterization of film aging and formation methods is needed to fully understand this change. However, none of the differences was great enough to affect further experiments using the films. These results demonstrate that the films maintained a controlled concentration with precision generally within 15% for the required period of time and that varying the film concentration could adjust the solution concentration to any value up to aqueous solubility limits. Since the 37 mg/g film contained sufficient retene to provide a saturated concentration, this was the highest film concentration used for subsequent toxicity tests. It should be noted that log Kfs values are correlated with log Kow values (19) and that log Kfs can be estimated from log Kow when unknown. An estimate of log Kfs can be used to determine the appropriate film concentration range for a new compound, greatly simplifying the development of PCD methods. PCD of Retene into Vials with Medaka Embryos. The addition of medaka eggs to vials with PDMS films had a significant effect on the apparent solution concentrations of retene. In Figure 4, the concentrations before eggs hatched (day 7) and after hatch (day 17) were compared with the expected solution concentrations (i.e., measured for the same film in a solution with no eggs). Before hatch, these levels varied randomly above and below the expected values, indicating no significant effect of the eggs. After hatch, however, the concentrations were systematically higher than expected, indicating a direct effect on the overall partition. During hatch, choriolytic hatching enzymes are secreted which digest the egg envelope (27) and cause its fragmentation. The fragments of the egg chorion as well as the filamentous villi that are attached to the chorion created a suspension of fine organic (lipidic/proteinaceous) material in the aqueous phase, likely due to bacterial degradation.

FIGURE 5. Frequency of BSD for various retene concentrations delivered by three separate methods (see text). Concentrations are “nominal” (total µg added divided by volume) except for the PCD case, where they were measured. Curved lines were generated by a sigmoidal fit of the log-transformed CS data. These particulate and dissolved materials would be expected to sorb hydrophobic retene. An overall shift in partition toward the aqueous phase would result, even if the truly water-dissolved concentration remained constant. The most severely affected medaka embryos had also died during that period so that decomposition would also contribute to a higher apparent retene concentration in the aqueous phase. The variability in concentration measurements was also much greater for samples containing eggs (data not shown), consistent with a heterogeneous suspension. Our analytical method would measure total retene and not discriminate between dissolved and bound forms. One measurement made just after all 5 embryos hatched on the 10th day produced an apparent aqueous concentration over three times the expected result. This means that apparent differences in the film-water partitioning with time may not represent any real differences in exposure of embryos. The role of dissolved organic matter in controlling exposure requires further investigation. While waterborne retene concentrations varied after hatching of embryos in the PCD vials, the retene concentrations were still highly regulated. The data for day 7 in Figure 4 shows that the concentration was held close to expected values for the embryo growth phase before hatch, with the three highest concentration films providing aqueous concentrations close to saturation. Variation in retene concentrations after hatch probably had little effect on measured toxicity because BSD was evident before hatch (days 5-9). Hence, the measured aqueous retene concentrations at the end of the exposure were not the cause of these effects. Comparisons among Embryo Toxicity Assays. Figure 5 demonstrates that the prevalence of BSD in all three embryo toxicity assays was concentration-dependent. However, the PCD method appears to be the most sensitive because the median effective concentration (EC50) for BSD was just under 10 µg/L (Probit Analysis, SoftTox, Los Angeles, CA) based on measured retene concentrations. In contrast, the EC50 for the semi-static 24-h renewal test (150 µg/L) and the static nonrenewal test (2500 µg/L), estimated from nominal concentrations, were much higher. The EC50 for the PCD method is below solubility and more realistic because this exposure regime could be attained in a natural situation. The nominal concentration EC50 values for the semi-static and static tests were about 10 and 150 times solubility, respectively. In all three embryo toxicity assays, exposure to retene started at the early embryonic stages, usually at the 2- or 4-cell stage of embryonic development (1-2 h post-fertilization (28)). Certain developmental stages of fish embryos show different susceptibility to environmental toxicants, and details including previous examples of retene toxicity may be found elsewhere (28-31). In previous semi-static assays with

juvenile trout, the amount of retene in solution declines exponentially to less than 5% of the initially added concentration after a 24-h exposure (11). Thus, in this static nonrenewal experiment with medaka embryos, retene concentrations are probably below the threshold of toxicity between days 2 and 3, which implies that BSD would have to be initiated either prior to day 2 or at later stages (beyond day 3) when the egg tissue was sufficiently metabolized to release further retene. On the other hand, in the 24-h renewal exposure, medaka embryos were exposed to retene at all developmental stages, and because it was added daily, a lower nominal retene concentration was needed to elicit the same syndrome. However, once again, the actual concentrations of retene that caused BSD were not well-defined as the retene is rapidly lost during 24-h exposure periods (11). The PCD technique provides a superior embryo toxicity testing method for risk assessment because uncertainties related to actual concentrations of hydrophobic compounds during different developmental stages are removed. In routine use, only a limited number of concentration measurements are needed to confirm performance of the film over time. In contrast to flow-through diluter experiments, the PCD method is actually simpler to implement than the static and semi-static assay procedures.

Acknowledgments Funding for this research was provided by grants from the Canadian Network of Toxicology Centres (CNTC). The authors thank Nikolaos Kyparissis for drawing Figure 2 and Tice Post for making the angled vial support assemblies.

Literature Cited (1) Brown, E. D.; Norcross, B. L.; Short, J. W. Can. J. Fish. Aquat. Sci. 1996, 53, 2337. (2) McGurk, M. D.; Brown, E. D. Can. J. Fish. Aquat. Sci. 1996, 53, 2343. (3) Marty, G. D.; Short, J. W.; Dambach, D. M.; Willits, N. H.; Heintz, R. A.; Rice, S. D.; Stegeman, J. J.; Hinton, D. E. Can. J. Zool. 1997, 75, 989. (4) McKim, J. M. J. Fish. Res. Board Can. 1977, 34, 1148. (5) Birge, W. J.; Black, J. A.; Westerman, A. G. Environ. Toxicol. Chem. 1985, 4, 807. (6) Wisk, J. W.; Cooper, K. R. Chemosphere 1990, 20, 361. (7) Harris, G. E.; Kiparissis, Y.; Metcalfe, C. D. Environ. Toxicol. Chem. 1994, 13, 1405. (8) Hamm, J. T.; Hinton, D. E. Aquat. Toxicol. 2000, 48, 403. (9) Billiard, S. M.; Querbach, K.; Hodson, P. V. Environ. Toxicol. Chem. 1999, 18, 2070. (10) Metcalfe, C. D.; Gray, M. A.; Kiparissis, Y. In Impact Assessment of Hazardous Aquatic Contaminants: Concepts and Approaches; Rao, S. S., Ed.; Lewis Publishers: Boca Raton, FL, 1999; pp 2952. (11) Basu, N.; Billiard, S. M.; Fragoso, N.; Omoike, A.; Tabash, S.; Brown, R. S.; Hodson, P. V. Environ. Toxicol. Chem. 2001, 20, 1244. (12) Klamer, H. J. C.; Villerius, L. A.; Roelsman, J.; De Maagd, P. G.-J.; Opperhuizen, A. Environ. Toxicol. Chem. 1997, 16, 857. (13) Helmstetter, M. F.; Alden, R. W., III. Aquat. Toxicol. 1995, 32, 1. (14) Metcalfe, C. D.; Cairns, V. W.; Fitzsimons, J. D. Aquat. Toxicol. 1988, 13, 347. (15) Walker, M. K.; Zabel, E. W.; Akerman, G.; Balk, L.; Wright, P.; Tillitt, D. E. In Techniques in Aquatic Toxicology; Ostrander, G. K., Ed.; Lewis Publishers: New York, 1996; pp 41-72. (16) Papoulias, D. M.; Noltie, D. B.; Tillitt, D. E. Aquat. Toxicol. 1999, 48, 37. (17) Kahl, M. D.; Russom, C. L.; DeFoe, D. L.; Hammermeister, D. E. Chemosphere 1999, 39, 539. (18) Mayer, P.; Wernsing, J.; Tolls, J.; De Maagd, P. G. J.; Sijm, D. T. H. M. Environ. Sci. Technol. 1999, 33, 2284. (19) Brown, R. S.; Akhtar, P.; Akerman, J.; Hampel, L.; Kozin, I. S.; Villerius, L. A.; Klamer, H. J. C. Environ. Sci. Technol. 2001, 35, 4097. (20) Leppanen, H.; Oikari, A. Environ. Toxicol. Chem. 1999, 18, 1498. (21) Guiney, P. D.; Smolowitz, R. M.; Peterson, R. E.; Stegeman, J. J. Toxicol. Appl. Pharmacol. 1997, 143, 256. VOL. 37, NO. 10, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2265

(22) Spitsbergen, J. M.; Walker, M. K.; Olson, J. R.; Peterson, R. E. Aquat. Toxicol. 1991, 19, 41. (23) Villalobos, S. A.; Anderson, M. J.; Denison, M. S.; Hinton, D. E.; Tullis, K.; Kennedy, I. M.; Jones, A. D.; Chang, D. P. Y.; Yang, G. S.; Kelly, P. Environ. Health Perspect. 1996, 104, 734. (24) Kim, Y.; Cooper, K. R. Chemosphere 1999, 3. (25) Metcalfe, C. D.; Metcalfe, T. L.; Cormier, J. A.; Huestis, S. Y.; Niimi, A. J. Environ. Toxicol. Chem. 1997, 16, 1749. (26) Kiparissis, Y. Ph.D. Dissertation, Trent University at Peterborough, ON, Canada, 2001. (27) Inohaya, K.; Yasumasu, S.; Ishimaru, M.; Ohyama, A.; Iuchi, I.; Yamagami, K. Dev. Biol. 1995, 171, 374. (28) Iwamatsu, T. Zool. Sci. 1994, 11, 825.

2266

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 10, 2003

(29) Von Westernhagen, H. In Fish Physiology; Randall, D. J., Ed.; Academic Press: San Diego, CA, 1988; Vol. XIA, pp 253-345. (30) Fragoso, N. M.; Parrott, J. L.; Hahn, M. E.; Hodson, P. V. Environ. Toxicol. Chem. 1998, 17, 2347. (31) Brinkworth, L. C.; Hodson, P. V.; Tabash, S.; Lee, P. J. Toxicol. Environ. Health Part A 2003, 66, 47.

Received for review September 13, 2002. Revised manuscript received February 14, 2003. Accepted February 19, 2003. ES026154R