Environ. Sci. Technol. 2007, 41, 1766-1770
Selenium Toxicity To Invertebrates: Will Proposed Thresholds for Toxicity To Fish and Birds Also Protect Their Prey? ADRIAN M. H. DEBRUYN* AND PETER M. CHAPMAN Golder Associates Ltd., 195 Pemberton Avenue, North Vancouver, British Columbia, Canada V7P 2R4
Efforts to manage the environmental risks of selenium (Se) in freshwater ecosystems have focused primarily on fish and birds, with invertebrates most often considered only as dietary sources of Se to higher trophic levels. Relatively little attention has been given to the risk of Se toxicity to invertebrates. Based on a review of 156 aqueous, dietary, or internal Se concentrations associated with toxic effects in 29 macroinvertebrate species, we found that water concentrations associated with acute lethality varied >1000-fold among taxa, whereas toxic dietary concentrations varied ∼100-fold and toxic internal concentrations varied about 30-fold. Sublethal effects occurred at ∼10fold lower concentrations than lethality. Sublethal effects occurred at 1-30 µg Se/g dry weight in invertebrate tissue, a range that encompasses proposed dietary thresholds for toxicity to fish and water birds, suggesting that Se may cause toxic effects in some invertebrate species at concentrations considered to be “safe” for the organisms consuming them.
Introduction Selenium (Se) is an essential metalloid that occurs naturally in association with coal seams, phosphate deposits, and other mineral formations. Selenium becomes of ecotoxicological concern when natural processes or anthropogenic activities such as mining or irrigation mobilize the element, producing elevated concentrations in water, sediment, and biota. The biogeochemistry of Se in aquatic systems is complex, with four stable oxidation states and multiple organic and inorganic forms. Selenium bioaccumulates (1, 2), biotransforms (3, 4), biomagnifies in some species (5, 6), and has been implicated in observed impacts to invertebrates, fish, water birds, and mammals (e.g., 7-10). Efforts to manage the environmental risks of Se in freshwater ecosystems have focused primarily on fish and birds, with invertebrates most often considered only as dietary sources of Se to higher trophic levels (e.g., 7, 11-13). Relatively little attention has been given to the risk of Se toxicity to invertebrates, possibly because of the assertion by Lemly (14) that “[f]ood-chain organisms can ... build up tissue concentrations of selenium that are toxic to predators while remaining unaffected themselves.” This conclusion was based on several studies that found no effect on macroinvertebrate populations in which individual animals had tissue Se concentrations from 20 to 370 µg/g dry weight, whereas the * Corresponding author phone: 604 904-6044; fax: 604 662-8548; e-mail:
[email protected]. 1766
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 5, 2007
estimated threshold dietary concentration for fish was only 3 µg/g dry weight (14). However, Swift (15) noted that at least some of the studies cited by Lemly (14) had infrequent sampling and low statistical power to detect effects: for example, the presence of a species in multiple years was in some cases taken as evidence that there was no effect on the populations. In a field experiment, Swift (15) reported dramatic reductions in abundances of some invertebrate taxa at internal Se concentrations around 60 µg/g dry weight. In addition, several laboratory experiments (16-18) and at least one field study (10) suggest that toxic effects occur in some invertebrate species at internal concentrations lower than this. In this paper, we revisit the evidence for Se toxicity to freshwater invertebrates. We compiled existing toxicity data (156 aqueous, dietary, or internal Se concentrations associated with toxic effects to 29 macroinvertebrate species) to estimate the distributions of waterborne, dietary, and internal Se concentrations associated with toxic effects. Our objective was to answer the following question: “Are invertebrate tissue Se concentrations considered “safe” for fish and water birds also “safe” for invertebrates?”
Methods Toxicity Data Compilation. We restricted our analysis to studies that reported effects with clear relevance to population-level processes, including mortality, growth, reproduction, and population abundance, but excluding behavioral and biochemical changes. We also included only studies that isolated the effect of Se vs other potential stressors. Laboratory studies of toxicant mixtures and field studies of complex stressors such as fly ash were excluded. Aqueous Se concentrations associated with toxic effects were available for 18 species of freshwater benthic invertebrates exposed to waterborne (usually inorganic) Se. Dietary or internal concentrations associated with toxic effects were available for only two freshwater benthic taxa (a midge larva and an isopod), so we supplemented these with data for other freshwater (i.e., zooplankton) and terrestrial invertebrates. By pooling the data in this way, we assume that the major ecotoxicological differences between freshwater zoobenthos and the other ecotypes (zooplankton and terrestrial invertebrates) are in characteristics that influence exposure to and accumulation of Se (e.g., diet and habitat), rather than sensitivity to bioaccumulated Se. These ecotypic differences may have a large influence on the waterborne concentrations that elicit toxic effects, but would not be expected to influence dietary or internal toxic concentrations (19). Dietary toxicity data were available for 1 zoobenthic species, 2 zooplankton species, and 8 terrestrial insects. Internal toxicity data were available for 2 zoobenthic species, 2 zooplankton species, and 3 terrestrial insects. Effect of Sulfate. There is evidence in the literature that, for some zooplankton and fish, the toxicity of dissolved selenate (Se VI) in acute tests declines as the concentration of sulfate in ambient water increases. However, evidence for an effect of sulfate on selenate toxicity in zoobenthos is equivocal. Brix et al. (20) reported a relationship between sulfate and selenate toxicity to Gammarus pseudolimnaeus (Amphipoda), but this relationship may have been driven by a single low LC50 value at low ambient sulfate. A similar analysis for Hyalella azteca (Amphipoda) revealed no relationship between selenate toxicity and ambient sulfate (20). Hansen et al. (21) found that bioconcentration of selenate by Chironomus (Diptera) larvae declined with increasing 10.1021/es062253j CCC: $37.00
2007 American Chemical Society Published on Web 02/02/2007
FIGURE 1. Toxicity of waterborne Se to freshwater benthic invertebrates. Data are genus mean concentrations associated with 1-4 d lethality (solid line) of selenate (b), selenite (O) and selenomethionine (×), >4 d lethality (dashed line) of selenate (2) and selenite (4), and sublethal toxicity of a selenate/selenite mixture (9) and selenite (0). Test genera are Hyalella (Hy), Chironomus (Ch), Bulinus (Bu), Gammarus (Ga), Hydra (Hd), Culex (Cu), Tubifex (Tu), Paratanytarsus (Pa), Physa (Ph), Aplexa (Ap), Nephelopsis (Ne), and Cyclocypris (Cy). Symbol codes indicate test details in Supporting Information.
sulfate concentration, but this decline was modest (less than a factor of 2) over a very large range of sulfate concentrations (Se/sulfur ratios from 1:0 to 1:480). Given the lack of a clear relationship between Se toxicity and sulfate for zoobenthos, and given that sulfate measurements were not available for many of the compiled toxicity data, we elected not to perform any adjustment of waterborne Se toxicity data to account for differences in sulfate concentration among acute toxicity tests. No similar adjustment would be possible or appropriate for chronic or field exposures (20, 22) or for dietary and internal toxicity data. Analysis of Toxicity Data. We summarized the compiled toxicity data as cumulative frequency distributions, in a manner analogous to that used to construct species sensitivity distributions. Where multiple waterborne or dietary Se toxicity data were available for a single species and/or genus, we calculated a geometric mean value. Averaging values within a taxon in this way gives equal weight to all taxa for which data are available, and avoids having the form of the distribution dominated by more intensively studied taxa. Relatively few data were available for toxic internal Se concentrations, so all individual values are shown. Where data were sufficient, separate distributions were constructed for the various chemical forms of Se, for acute vs chronic exposures, and for lethal vs sublethal effects. The objective of this analysis was not to calculate an effects threshold or concentration that would be “safe” for a specified percentage of exposed taxa, but rather to determine the range of exposures over which effects occur, and to examine patterns of toxicity among forms of Se, among receptor taxa, and as a function of exposure duration, exposure route, and the type of toxic effects.
Results Sensitivity to Waterborne Selenium. Waterborne Se toxicity data revealed a broad range of sensitivity among taxa, with acutely (1-4-d test duration) toxic water concentrations spanning more than 3 orders of magnitude (Figure 1). Hyalella (Amphipoda) was the most sensitive taxon (genus mean LC50 ∼ 0.4 mg Se/L as selenite) and Nephelopsis (Hirudinea) was the most tolerant (LC50 ∼ 400 mg Se/L as
FIGURE 2. Toxicity of dietary Se to freshwater benthic, zooplanktonic, and terrestrial invertebrates. Data are genus mean concentrations associated with lethality (solid line) of selenate (b), selenite (O) and organoselenium (×), and sublethal toxicity (dashed line) of selenate (2), selenite (4) and organoselenium (+). Test genera are as in Figure 1 and Myzus (My), Corcyra (Co), Spodoptera (Sp), Podisus (Po), Megaselia (Me), Tribolium (Tr), Drosophila (Dr), Daphnia (Da), Brachionus (Br), and Pieris (Pi). Symbol codes indicate test details in Supporting Information.
selenate). Acute lethality distributions for selenate (Se VI) and selenite (Se IV) were nearly identical, and the two values available for organoSe (as selenomethionine) fell within the range of toxic concentrations for the inorganic forms (details of all test data are provided in the Supporting Information). Lethal and sublethal toxic effects were observed at somewhat lower water concentrations in laboratory tests of longer duration, and at much lower water concentrations in experimental streams. In laboratory tests, Hyalella exhibited median lethality at 0.1 (selenite, 10-14-d tests) to 0.2 (selenate, 7-10-d tests) mg Se/L, and >50% reduction in reproduction at 0.1 mg Se/L (selenate, 1-d test). In a 30-d laboratory test, Chironomus exhibited impaired development at 0.5 mg Se/L (selenate/selenite mixture). In a >2-year study of selenite-spiked experimental streams, Caecidotea (Isopoda) and Tubifex (Tubificidae) exhibited ∼50% reductions in abundance at 0.01 mg Se/L (not shown in Figure 1). Sensitivity to Dietary Selenium. Dietary toxicity data showed a narrower range of sensitivity among taxa than was observed for waterborne toxicity data (Figure 2). The range for diets spiked with or cultured in selenate was about 2 orders of magnitude, with lethality observed at dietary concentrations in the range of 10 to 1000 µg Se/g dry weight, and sublethal effects in the range of 1 to 80 µg Se/g dry weight. Myzus (Aphididae) was the most sensitive taxon, exhibiting a 50% decrease in population size at a dietary concentration of 1.5 µg Se/g dry weight. The genus-mean value for Chironomus was somewhat higher, although one study reported a 40% decrease in growth at a dietary concentration of
