Using an interlaboratory study to revise methods for conducting 10-d to 42-d water or sediment toxicity tests with Hyalella azteca

INTRODUCTION

Standard methods for conducting toxicity or bioaccumulation tests with freshwater sediments were initially developed in the 1980s and 1990s for amphipods (Hyalella azteca, Diporeia spp.), midges (Chironomus dilutus, Chironomus riparius), oligochaetes (Tubifex tubifex, Lumbriculus variegatus), mayflies (Hexagenia spp.), and cladocerans (Daphnia magna, Ceriodaphnia dubia) 1-4. Toxicity endpoints for H. azteca in these initial methods included 10-d to 28-d survival and growth with a daily diet ration of 1.8 mg (dry wt)/d of yeast–cerophyll–trout chow (YCT) provided to each 300-mL chamber containing 10 test organisms and 100 mL of sediment 2, 5, 6. Standard methods for conducting sediment reproduction tests with H. azteca were initially developed in 2000 5, 6, recommending the same daily diet ration of 1.8 mg YCT/d throughout the 42-d exposures. Test acceptability criteria for the 42-d test based on results of interlaboratory testing were established as ≥80% survival after the initial 28-d sediment exposure 7, with suggested 28-d growth of ≥0.15 dry weight (milligrams per individual) and reproduction of ≥2.0 young/female 5, 6. Subsequently, some laboratories have observed inconsistent control survival, growth, or reproduction of H. azteca in 42-d water or sediment exposures.

A workshop was held in March 2010 in Chicago, Illinois, USA, to discuss the state of the science regarding the use of H. azteca in whole-sediment or water-only toxicity tests and was attended by approximately 40 participants from private, academic, and government laboratories in the United States and Canada. Discussion sessions were divided among topics of culture, water quality, water-only toxicity testing methods, and whole-sediment toxicity testing methods. Survey information assembled before the workshop on H. azteca culturing and testing procedures was compiled to compare practices across laboratories. Culture practices varied widely, ranging from fairly intensive methods involving frequent medium changes and close monitoring of performance to “mesocosm-style” approaches in which mass cultures are maintained for long periods under laboratory conditions that encourage secondary biological growth. Evaluation of water composition data showed that most of the laboratories cultured and conducted toxicity tests in water of moderate to high hardness (∼100 mg/L hardness as CaCO₃ or higher) and with a chloride concentration frequently above 15 mg Cl/L. Both empirical and experimental evidence indicated that chloride was important for supporting good long-term performance of H. azteca and may be a reason why some reconstituted waters have been found unsuitable for H. azteca 5, 6. Among reconstituted waters in use, the formula proposed by Borgmann 8 produced the best results. The most obvious difference between Borgmann 8 reconstituted water and other formulas was the inclusion of bromide. This is a curious finding because there is little evidence in the literature for a critical physiological role of bromide. An informal H. azteca advisory group was 1 of the outcomes from the workshop, a group that periodically exchanges information on methods for culturing or testing with H. azteca.

Over the past 5 yr, studies have been conducted to refine methods for conducting 42-d reproduction tests with H. azteca. Environment Canada published updated guidance on methods for conducting 14-d water-only or sediment toxicity tests with H. azteca 9 and has plans for developing guidance for conducting reproduction studies with H. azteca. Modifications to the H. azteca method have been developed to 1) define ionic composition requirements for exposure water (i.e., >15 mg Cl/L 10 and >0.02 mg Br/L 11), 2) improve growth and reproduction with alternate diets and by increasing rations over time, and 3) define methods for conducting water-only toxicity tests using either daily addition of water (e.g., by using a flow-through system) or complete water replacement (e.g., 3 replacements/week on Monday, Wednesday, and Friday 12). Results of studies using improved waters and diets have demonstrated improved growth (e.g., >0.40 mg/individual at d 28) and improved reproduction (e.g., >8.0 young/female by d 42) of H. azteca using revised test conditions to conduct water-only or sediment tests 10-13. In advance of revising the US Environmental Protection Agency (USEPA) 5 and ASTM International 6 guidance for conducting sediment or water toxicity tests with H. azteca, we wanted to determine whether additional laboratories following these revised methods could also demonstrate improved performance of H. azteca in 10-d, 28-d, or 42-d water or sediment exposures.

The present interlaboratory study was conducted in 2014 with the goals of determining 1) whether the proposed alternate diets and water requirements would result in improved control survival, growth, and reproduction of H. azteca in 10-d to 42-d water or sediment exposures across a range of laboratories; 2) whether use of the alternate diets and water requirements would support increases in test acceptability criteria for control performance of H. azteca (survival, weight, reproduction); and 3) whether there are other diets or waters that are better than the alternate diets or water requirements being proposed.

MATERIALS AND METHODS

Twenty laboratories in the United States, 2 laboratories in Canada, and 2 laboratories in Germany (24 total), covering a range of experiences from not ever having conducted 42-d H. azteca reproduction tests to routinely conducting 42-d H. azteca reproduction tests, volunteered to participate in the present interlaboratory study. Each laboratory responded to a questionnaire before conducting exposures. Laboratories with no experience or with experience conducting only 4-d or 10-d exposures with H. azteca were classified as low-experience laboratories (n = 8). Laboratories routinely conducting 4-d to 28-d exposures but only occasionally conducting 42-d reproduction exposures with H. azteca were classified as medium-experience laboratories (n = 9). Laboratories routinely conducting 42-d H. azteca reproduction exposures were classified as high-experience laboratories (n = 7).

Participating laboratories were asked to conduct exposures in a control water with >15 mg Cl/L and >0.02 mg Br/L in 300-mL beakers containing 5 mL of sand and 200 mL of water using the water-only method described in the third column in Supplemental Data, Table S1, comparing 2 proposed diets: 1) a ramped diet (where the amount of food being fed is increased over time) of diatoms (Thalassiosira weissflogii) + a ramped diet of Tetramin (amount of both materials provided daily increased each week of the exposure) and 2) a diet of YCT + a ramped diet of Tetramin (constant amount of YCT daily and an increased amount of Tetramin each week of the exposure) 12, 13. Other treatments that laboratories were encouraged to evaluate included a treatment of control sediment (collected from West Bearskin Lake located in northeastern MN, USA) tested with the 2 alternate diets, treatments of other diets or waters, or other treatments of interest to the participating laboratory (e.g., different strains of H. azteca, smaller exposure chambers, daily water additions vs complete water replacement of water 3 times/wk).

All participating laboratories were provided with the following materials: data sheets (for recording biological data and water quality data), control sand (Granusil 4030; Unimin) provided dried and sieved to <0.5 mm before use and to be wetted by the participating laboratory with their control test water (base water) for approximately 1 d before use, West Bearskin Lake control sediment (to laboratories testing sediment as an additional treatment), Nalgene bottles (for sampling major ions, for sampling adult males for morphological determination of genetic strain, and for sampling organisms for gene expression analyses), and diatom stock material (if requested by a laboratory). The shelf life of the diatom stock material obtained from Reed Mariculture is 2 mo when held at 4 °C. Therefore, laboratories were encouraged to obtain a fresh batch of diatom stock material in advance of starting exposures. The procedure for preparing the diatom diet and Tetramin diet can be found in Supplemental Data, Table S2.

Participating laboratories were encouraged to review section 14 in USEPA 5 for details on how to conduct a 42-d sediment reproductive test with H. azteca. Specifically, laboratories were encouraged to review section 14.3.7 describing the procedure for determining the number of males and females in each replicate beaker at the end of the exposures and table 14.2 describing the daily activity schedule for conducting a 42-d toxicity test with H. azteca.

Participating laboratories were provided as much flexibility as possible regarding their level of participation and timing of their study. At a minimum, each participating laboratory was asked to evaluate the water-only testing method (with 5 mL of sand substrate) with the 2 alternate diets described in the third column of Supplemental Data, Table S1 (item 13). Each diet and substrate treatment involved testing with a total of 16 replicate beakers of 300 mL: 4 replicates dedicated to survival and growth measurements at day 10, 4 replicates dedicated to survival and growth measurements at day 28, and the remaining 8 replicates dedicated to survival, growth, and reproduction measurements through day 42. Survival in exposures conducted with a sand substrate was documented weekly, and reproduction was documented weekly starting on approximately day 21 (there was the potential for reproduction before day 28 as a result of the improved diet(s) 12, 13). The number of treatments tested by each of the participating laboratories depended on the resources and interests of each participating laboratory.

Laboratories were given the option to do daily additions of water during the exposures or complete water renewals 3 times/wk (Supplemental Data, Table S1). The advantage of daily addition of water (2 volume additions/d) is that it is an approach consistent with sediment testing methods and concentrations of test materials can be maintained more consistently than with complete replacement of water 3 times/wk (i.e., Monday, Wednesday, Friday). The disadvantage of daily additions of water is that either specialized equipment is needed to deliver water (e.g., a diluter or other delivery system) or daily manipulation of the beakers is required (e.g., manually pouring off and adding water to each beaker; see appendix A in USEPA 5).

A spreadsheet with templates was provided to participating laboratories to summarize biological data (by replicate) and routine water quality data (i.e., temperature, hardness, alkalinity, conductivity, pH, ammonia, dissolved oxygen). Errors or questionable data found in the spreadsheet data file were sent back to the participating laboratory for verification. All verified data are compiled in Table 1 and Table 2 and in Supplemental Data, Tables S3 to S7.

RESULTS

DISCUSSION

The improved performance observed in the control sediment exposures compared with the water-only exposures might have been the result of the natural sediment substrate providing a nutritional or water quality supplement to H. azteca. Alternatively, this improved performance in control sediment may have been the result of the high-experience laboratories testing sediment (i.e., no low-experience laboratory tested West Bearskin Lake sediment, and only 1 low-experience laboratory tested an alternate control sediment using the alternate diets). Subsequent 42-d testing with H. azteca with either of these 2 alternate diets has resulted in US Geological Survey-Columbia and USEPA-Duluth meeting the proposed test acceptability criteria in sand and in various natural control sediments, with growth and reproduction of H. azteca typically up to approximately 20% greater in natural sediment compared with sand, suggesting that H. azteca can receive a nutritional or water quality supplement from natural sediment (C. Ivey, unpublished data). Periodic testing of a sand control will be recommended in the planned revisions to the USEPA 5 and ASTM International 6 sediment testing methods to determine whether the water or diet are adequate to support performance of H. azteca independent of nutritional or water quality supplements from sediments.

High-experience laboratories were able to more frequently meet test acceptability criteria compared with medium-experience and low-experience laboratories. Other than historically performing more long-term H. azteca exposures, what factors might result in a high-experience laboratory being better able to meet proposed test acceptability criteria? Factors that may contribute to improved test performance include experience in preparing and using the alternate diets, in weighing small masses of H. azteca, in transferring H. azteca during the exposures, and in recovering offspring of H. azteca during the reproductive phase of the exposures.

Most of the high-experience laboratories had previous experience preparing and using the alternate diets, which may have improved performance. Additional guidance will be provided in the updated USEPA 5 and ASTM International 6 methods regarding how to prepare and deliver the alternate diets. The diets are added as solids suspended in water, and care is required to deliver a consistent amount of food to each exposure chamber each day. The food containers must be agitated to ensure a homogeneous suspension each time food solution is withdrawn. An adjustable-volume repeating pipette can be used to add food to each exposure chamber; however, the food suspension in the reservoir of the pipette should be remixed (e.g., inverted or gently shaken) after each delivery of food to each replicate exposure chamber to prevent settling and unequal allocation of food. Failure to provide a consistent amount of food to each exposure chamber will compromise the performance of test organisms in a study. Moreover, to avoid flushing food out of the exposure chambers, feeding should take place soon after completion of a water renewal, allowing the food time to settle to the sediment surface before the next water renewal.

Higher variability in weights of H. azteca was evident among inexperienced laboratories compared with the medium-experience or high-experience laboratories. Laboratories not accustomed to weighing small masses may not possess the experience, quality balances, or proper weighing vessels needed to accurately measure weights (e.g., the total mass of 10 H. azteca in a replicate would be 0.5 mg at the 10-d test acceptability criteria of 0.05 mg/individual). To reduce variability in dry weight determination, weigh pans should be as small as is practical (e.g., pans weighing less than ∼100 mg when empty should be used). Weighing and reweighing dried blank weigh pans (with no organisms added) provides an indication of potential variability in weight because of factors other than the mass of test organisms being weighed (e.g., variability associated with the balance, handling, or drying of pans).

The reproduction test acceptability criteria were nearly always the most difficult for inexperienced laboratories to meet. Finding young H. azteca that may be only days old in sand particles that are nearly the same size can be difficult for inexperienced individuals. Isolating young or repeatedly transferring adult H. azteca can result in losing organisms (e.g., H. azteca can stick to the glass surface of a pipette). Laboratories are encouraged to practice isolating and transferring H. azteca (e.g., add a known number of organisms to a chamber and after 1 h determine the ability of individuals to consistently isolate small organisms 14).

A goal of the present interlaboratory study was to determine whether survival, growth, or reproduction of H. azteca was better using either of the 2 alternate diets. The relationship between 28-d weight and 42-d reproduction across the 2 diets is illustrated in Figure 1A for exposures conducted in sand and in Figure 1B for exposures conducted in sediment. Reproduction increased with increasing weights, with laboratories that met test acceptability criteria for 28-d weight typically meeting test acceptability criteria for reproduction in water exposures (Figure 1A) and in sediment exposures (Figure 1B). Growth and reproduction were typically greater in sand for high-experience and medium-experience laboratories compared with low-experience laboratories, with no consistent pattern between the 2 diets (Figure 1A).

The influence of the 2 alternate diets on survival, growth, or reproduction of H. azteca is illustrated in Figure 2. Survival at day 42 was slightly higher in the ramped diatom + ramped Tetramin diet compared with the YCT + ramped Tetramin diet (Figure 2A). There was no discernable difference in 10-d weight data, but 28-d weight tended to be greater in the YCT + ramped Tetramin diet (Figure 2B). No consistent pattern was observed between the 2 diets in day 42 weight (Figure 2C) or reproduction (Figure 2D). Hence, both diets resulted in consistent performance of H. azteca in 10-d to 42-d exposures. However, dissolved oxygen tended to be lower in exposures conducted with the YCT + ramped Tetramin diet compared with the ramped diatom + ramped Tetramin diet. Dissolved oxygen lower than 2.0 mg/L can lead to mortality in H. azteca. An added advantage with the ramped diatom + ramped Tetramin diet is the ease of preparation and consistency of stock concentration over the YCT + ramped Tetramin diet. The diatom diet can be prepared in a matter of hours, whereas it takes approximately 1 wk to prepare the YCT diet.

Two systems of water renewal were evaluated in the present interlaboratory study: renewal of 2 volume additions per day and Monday–Wednesday–Friday complete water renewal. The percentage of laboratories meeting test acceptability criteria performing daily renewals did not substantially differ from the percentage of all laboratories meeting test acceptability criteria (Table 1).

Five participating laboratories used a base water with <15 mg Cl/L (laboratories E, M, P, T, and X), and all but 1 laboratory used a base water with >0.02 mg Br/L (laboratory Y; Supplemental Data, Table S7). The 5 laboratories that used a base water <15 mg Cl/L generally exceeded test acceptability criteria. However, the laboratory that used a base water <0.02 mg Br/L did not meet the test acceptability criteria for reproduction in 4 of the 6 diet or substrate exposures. The percentage of laboratories meeting test acceptability criteria did not improve with increasing concentrations of chloride above 15 mg Cl/L or in increasing bromide above 0.02 mg Br/L in the base waters. The revisions to the USEPA and ASTM International methods will recommend using a base water >15 mg Cl/L and with >0.02 mg Br/L 10, 11.

Two genetically distinct clades of H. azteca have been routinely used in toxicity testing with water or sediment: the US lab clade and the Burlington clade 15. The Burlington clade is morphologically differentiated from the US clade by characteristics of the second gnathopod of males and setation of the inner plate of the first maxilla (D. Soucek, personal observation). The US lab clade was originally obtained in the early 1980s from Corvallis, Oregon, by A. Nebeker (of the USEPA), and the Burlington clade was originally obtained in the 1990s from Burlington, Ontario, Canada, by U. Borgmann (of Environment Canada). The Burlington clade has been observed to exhibit similar growth but lower reproduction compared with the US clade (K.M. Major, Master's thesis, University of Illinois at Urbana-Champaign, Champaign, IL, USA). All but 1 laboratory used the US clade in the present study. Variable growth and reproduction of H. azteca were observed for the laboratory that tested the Burlington clade; however, additional study is needed to determine whether a base water of >15 mg Cl/L and >0.02 mg Br/L with the 2 alternate diets would support good performance of the Burlington clade in water or sediment testing. The 10-d growth studies and acute toxicity tests with the Burlington clade indicated that low chloride (e.g., <15 mg/L) did not have the negative effect on growth or contaminant sensitivity observed for the US clade 10.

Reproduction of H. azteca is typically more variable than growth 16, but variability in reproduction is lower with H. azteca fed improved diets 12. In the interlaboratory testing of the 2 alternate diets for laboratories that met test acceptability criteria, the coefficient of variation (CV) was typically <20% for growth and typically >20% for reproduction of H. azteca (Supplemental Data, Tables S3–S6). This difference in variation affects the statistical power of the comparisons and the number of replicates required for a test. For example, detection of a 20% difference between treatment means at a statistical power of 0.8 would require approximately 4 replicates at a CV of 10% and 14 replicates at a CV of 20% (figure 16.5 in USEPA 5). Fewer replicates would be required if detection of larger differences among treatment means were of interest. An 8-replicate design for conducting reproduction studies with H. azteca is recommended by the USEPA 5 and ASTM International 6 as a compromise between logistical constraints and statistical considerations.

The results of the present interlaboratory study were compared with 2 previous interlaboratory sediment studies 7, 17. Burton et al. 17 conducted an H. azteca interlaboratory study with 10 participating laboratories evaluating the 10-d whole-sediment toxicity test method. Each laboratory had to have its own culture of H. azteca, was experienced in conducting toxicity tests with H. azteca, and used the traditional diet of 1.8 mg/beaker YCT daily. In the present study, 100% of the 26 laboratories met the 10-d survival test acceptability criteria of 80%, with mean CVs across laboratories <3.7% with the 2 alternate diets. In the Burton et al. 17 study, 90% of the laboratories met the 10-d test acceptability criteria survival of 80%, with a mean CV across laboratories of 5.8%. Hence, consistent 10-d interlaboratory performance was observed in both the Burton et al. 17 study and the present study. No 10-d growth endpoints were evaluated in the Burton et al. 17 interlaboratory study.

Norberg-King et al. 7 conducted an H. azteca interlaboratory study evaluating short-term and long-term sediment toxicity testing methods with up to 18 participating laboratories. The testing was split up over 2 phases: 1) phase 1 with laboratories testing 2 control sediments to become familiar with the new long-term methods and 2) phase 2 including 10-d and long-term testing with control and contaminated sediments. Participating laboratories were selected based on experience in sediment testing and whether they had their own cultures; all laboratories used the traditional diet of 1.8 mg/beaker YCT daily. In the definitive phase 2 testing, 82% of laboratories met the 10-d survival test acceptability criteria of 80% in West Bearskin Lake sediment, with a mean survival across laboratories of 92% and a mean CV across laboratories of 5.0%. In the present study, the 10 laboratories conducting West Bearskin Lake sediment exposures met the 10-d survival test acceptability criteria of 80%, with mean survival >97% and mean CVs across laboratories <3.0% with the alternate diets. No 10-d growth data were evaluated by Norberg-King et al. 7.

In the Norberg-King 7 study, the 28-d survival test acceptability criterion of 80% and the 28-d weight test acceptability criterion of 0.15 mg/individual were met by 88% of laboratories, with mean CVs across laboratories of 6.8% for survival and 28% for weight. The present study also had a 28-d survival test acceptability criterion of 80% but a nearly 2.5-fold higher 28-d weight test acceptability criterion of 0.35 mg/individual. Regardless of the higher 28-d weight test acceptability criteria, laboratories in the present study testing West Bearskin Lake sediment met 28-d survival and weight test acceptability criteria 100% of the time, with mean survival CVs across laboratories of <1.6% and mean weight CVs across laboratories of <20%.

In the Norberg-King et al. 7 study, the test acceptability criterion for reproduction was set at 2.0 young/female; 71% of laboratories met the test acceptability criterion in control sediment, with a mean CV across laboratories of 49% (n = 5). In the present study, the reproduction test acceptability criterion was set 3-fold higher at 6.0 young/female and laboratories were able to meet the reproduction test acceptability criteria in control sediment 100% of the time, with a mean CV across laboratories of 38% (n = 6), using the YCT + ramped Tetramin diet and 83% of the time, with a mean CV across laboratories of 51% (n = 6), using the ramped diatom + ramped Tetramin diet. No test acceptability criteria for 42-d survival or growth were provided in the Norberg-King et al. 7 study, but mean 42-d survival was 89%, with a mean CV across laboratories of 7.4% (n = 7). Mean survival at 42-d in the present interlaboratory study was >92%, with mean CVs across laboratories of <8.3% (n = 6).

The laboratories participating in the Norberg-King et al. 7 study used the traditional YCT diet of 1.8 mg/beaker YCT daily and may not have met the updated water quality requirements of >15 mg Cl/L and >0.02 mg Br/L. The laboratories participating in the Norberg-King et al. 7 study frequently met the 28-d survival test acceptability criterion in control sediment of 80% control survival (88% of laboratories); but only 13% of the laboratories would have met the updated 28-d test acceptability criterion for weight of 0.35 mg/individual (100% of the laboratories in the present study met the weight test acceptability criteria), and none of the laboratories would have met the updated 42-d test acceptability criterion for reproduction of 6.0 young/female (>83% of the laboratories in the present study met the reproduction test acceptability criterion).

In conclusion, improved growth and reproduction of H. azteca were observed in tests conducted with the 2 alternate diets of ramped diatoms (Thalassiosira weissflogii) + ramped Tetramin or YCT + ramped Tetramin, especially when compared with results from tests conducted using the traditional diet of 1.8 mg/d of YCT. Laboratories in the present interlaboratory study were able to meet test acceptability criteria and, in most cases, had lower CVs compared with previous interlaboratory studies using the traditional YCT diet. Laboratory success in conducting 42-d H. azteca exposures benefited from adherence to several key requirements of the detailed testing, culturing, and handling method. Although there is no substitute for experience in conducting 42-d H. azteca exposures, laboratories can likely increase their success rate by training technicians in testing, culturing, and handling procedures along with following more detailed guidance on conducting long-term exposures described in ongoing revisions to the USEPA 5 and ASTM International 6 methods. The 2 alternate diets, along with added water quality requirements of >15 mg Cl/L and >0.02 mg Br/L, supported increases in minimum control performance of H. azteca (e.g., weight and reproduction). In the revised methods, both diets will be acceptable for use in H. azteca 10-d to 42-d water or sediment toxicity testing. Results from the present interlaboratory study are being used in the revisions of the USEPA 5 and ASTM International 6 methods for conducting 10-d to 42-d water or sediment toxicity exposures with H. azteca.

Supplemental Data

The Supplemental Data are available on the Wiley Online Library at DOI: 10.1002/etc.3417.

Disclaimer

Any use of trade, product, or firm names is for descriptive purposes only and does not imply endorsement by the US government. Funding for the present study was provided in part by the Great Lakes Restoration Initiative. The USEPA has not formally reviewed the present study; the views expressed herein may not reflect the views of the USEPA.

Data availability

Data, associated metadata, and calculation tools are available from the corresponding author ([email protected]).