Arch Environ Contam Toxicol DOI 10.1007/s00244-015-0151-x

The Effects of Chronological Age and Size on Toxicity of Zinc to Juvenile Brown Trout Daniel J. Diedrich1,2 • Ruth M. Sofield1 • James F. Ranville2 • Dale J. Hoff3 V. Dan Wall4 • Stephen F. Brinkman5



Received: 20 November 2014 / Accepted: 13 March 2015 Ó Springer Science+Business Media New York 2015

Abstract A series of toxicity tests were conducted to investigate the role of chronological age on zinc tolerance in juvenile brown trout (Salmo trutta). Four different incubation temperatures were used to control the maturation of the juveniles before zinc exposures. These 96-h exposures used flow-through conditions and four chronological ages of fish with weights ranging from 0.148 to 1.432 g. Time-to-death (TTD) data were collected throughout the exposure along with the final mortality. The results indicate that chronological age does not play a predictable role in zinc tolerance for juvenile brown trout. However, a relationship between zinc tolerance and fish size was observed in all chronological age populations, which prompted us to conduct additional exploratory data analysis to quantify

Electronic supplementary material The online version of this article (doi:10.1007/s00244-015-0151-x) contains supplementary material, which is available to authorized users. & Ruth M. Sofield [email protected] 1

Huxley College of the Environment, Western Washington University, 516 High Street, Bellingham, WA 98225, USA

2

Department of Chemistry and Geochemistry, Colorado School of Mines, 1500 Illinois St., Golden, CO 80401-1843, USA

3

National Health and Environmental Effects Research Laboratory, Mid-Continent Ecology Division, Office of Research and Development, United States Environmental Protection Agency, 6201 Congdon Boulevard, Duluth, MN 55804, USA

4

United States Environmental Protection Agency, U.S. EPA Region 8 EPR-S, 1595 Wynkoop St., Denver, CO 80202-1129, USA

5

Fort Collins Service Center, Colorado Parks and Wildlife, 317 W. Prospect, Fort Collins, CO 80526, USA

how much of an effect size had during this stage of development. The smallest fish (0.148–0.423 g) were shown to be less sensitive than the largest fish (0.639–1.432 g) with LC50 values of 868 and 354 lg Zn/L, respectively. The Kaplan–Meier product estimation method was used to determine survival functions from the TTD data and supports the LC50 results with a greater median TTD for smaller fish than larger juvenile fish. These results indicate that fish size or a related characteristic may be a significant determinant of susceptibility and should be considered in acute zinc toxicity tests with specific attention paid to the expected exposure scenario in the field.

Sensitivity of various developmental life stages of fish to metal toxicity has been well studied. Typically, earlier developmental stages are more sensitive than more mature stages (Holcombe et al. 1979; Hedtke et al. 1982), although later developmental (post-alevin) stages have been shown to be more sensitive in certain cases (Buhl and Hamilton 1991; Chapman 1978; Stubblefield et al. 1999). Generally, effects based on either size (mass or length) or age (e.g., days posthatch) are measured in toxicity tests using only one developmental stage but not both simultaneously. Because they tend to be correlated, it is difficult to distinguish whether size or age is a more important factor; however, both younger and smaller fish have been found to be more sensitive to metal exposures. For example, in a carp toxicity study on the juvenile life-stage, the LC50 value increased as fish size increased (Alam and Maughan 1992). In brook trout, adult fish had greater survival and greater zinc accumulation as size increased (Holcombe et al. 1979). Hoang et al. (2004) showed that fathead minnows\1 day old exposed to nickel under varying water conditions had lower LC50 values than 28-day-old fish.

123

Arch Environ Contam Toxicol

Both groups of fish were in the same developmental stage, but organism size was not explicitly reported (Hoang et al. 2004; Holcombe et al. 1979). Similarly, Williams and Holdway (2000) found that 9- to 10-day-old Australian rainbow fish exposed to pulses of zinc or cadmium had greater LC50 values than did fish \24 h or 3–4 days old; size was not considered. In general, fewer studies have found greater tolerance in smaller individuals than in larger ones. This research was conducted as part of a research project on the upper Arkansas River of Colorado, where aqueousphase zinc has been identified as a toxicant of concern to juvenile trout (Brinkman et al. 2006; Davies and Woodling 1980). The need for an assessment of the contribution of size and age on zinc sensitivity is evident when the fluctuating incubation temperatures in natural habitats are considered. Greater regional climate variability may induce fluctuation of temperatures in the field (Beniston 2005; Brown et al. 1992) and lead to more frequent drought and greater variability in spring runoff intensity, timing, and duration (Barnett et al. 2008; Clow 2010). Additionally, in colder temperatures, growth rate decreases (Crespo and Balasch 1980; Elliott and Hurley 1998; Jensen 1990). The result is that populations incubated in colder water will be older when they reach the same size as populations incubated in warmer water. Depending on how in-stream temperatures change compared with the time of early snowmelt, greater variability in metal concentrations and timing of pulses in the streams could be expected. These pulses could occur during the time interval of the most sensitive life stage and the most sensitive fish size resulting in different exposure scenarios than has historically been the case. In this experiment, we identified three influences that are important to consider for modeling field toxicity: (1) the lack of clear understanding as to which factor—size or age—is a more important modifier of metal sensitivity; (2) predicted changes in temperature and toxicity in systems, which will affect both the age and size at which fish are exposed to aqueous metal pulses; and (3) the reliance on standardized laboratory toxicity-testing methods that account either for size by requiring a specific mean weight as in ASTM E729-96 (2007) or for age with a specific age requirement. These influences make it difficult to predict how changing environmental temperatures will affect toxicity to fish and, therefore, to know what modifications to standardized toxicity tests should be made to better match environmental conditions. Given the unknown effects of environmental variability, this study was used to determine the influence of chronological age and fish size on their tolerance to zinc within a developmental stage to better inform toxicity models that consider temperature and the early life stages.

123

Materials and Methods In this work, the term ‘‘chronological age’’ is used to emphasize that a distinction between the effects of age and size was made by holding size approximately constant while varying age with different rearing temperatures. Organisms and Conditions Brown trout (Salmo trutta) in the eyed-egg stage were obtained from the Colorado Parks and Wildlife research hatchery in Bellevue, Colorado, USA. The original source of the eggs was the Colorado Delaney Buttes Reservoir spawning operation. To manipulate developmental stage and chronological age independently, eggs were subjected to different temperature treatments before zinc exposures. Eggs were distributed to four 100-gallon tanks and gradually acclimated to four different temperatures (12, 9, 6, and 3 °C). The rearing time, spanning from the placement of the fish into the four treatment groups through the initiation of the zinc exposures, will be referred to as the ‘‘rearing-temperature treatment.’’ The four rearing-temperature treatments ended when the mean weight of ten randomly captured juveniles was approximately 0.5 g. This was achieved at 50, 83, 106, and 141 days after swim-up going from high to low rearing temperatures, respectively; for simplicity, the chronological ages and rearing times will refer to the number of days after swim-up that fish were exposed to the rearing-temperature treatment. At 40 days in the 3 °C rearing-temperature treatment, the temperature was increased 1 °C/day up to 6 °C because fish weight was not increasing at 3 °C. This modification in the rearing-temperature treatment remained in adherence to the study design because juveniles were of similar sizes at the time of exposure and still had staggered chronological ages. Once the fry had reached the target weight, temperatures were increased by 1 °C/day to a final temperature of 12 °C and maintained there for 7 days before the zinc exposures. The time from first zinc exposure through the end of each toxicity test will be referred to as the (zinc-exposure period,) which lasted 96 h. Zinc Exposures The four zinc exposures were conducted at 12 °C with a flow-through serial dilutor system according to American Society for Testing and Materials (ATSM) standard method (ASTM E1241 2005). Juveniles of the appropriate chronological age were gathered randomly and assigned, one at a time, to 1 of 24 exposure tanks (2.7 L), until each contained 15 fish. The exposure tanks were placed

Arch Environ Contam Toxicol

randomly at 24 positions in a temperature-controlled water bath. Each exposure tank received 40 ml/min of exposure solution from the diluter. Because of the early removal of live fish for gill tissue analysis (results did not pass qualityassurance requirements and so are not presented here), an uneven number of juveniles were exposed to each zinc treatment and in each chronological age; the total number of juveniles exposed in the 50-, 83-, 106-, and 141-day rearing time populations was 242, 254, 270, and 273, respectively. Analytical reagent grade ZnSO4 was obtained from Mallinckrodt (Paris, Kentucky, USA). Dechlorinated municipal tap water from Fort Collins with a mean hardness ranging from 43.7 to 54.9 mg/L CaCO3 (Diedrich 2007) was used for the control and was spiked with a series of five serially diluted concentrations of the zinc sulfate in four replicates. Water hardness was determined according to standard methods (American Public Health Association 1985). Acidity was measured using an Orion Research pH meter. The range of sample pH was 7.2–7.6 with a mean of 7.4. Total zinc concentrations were serially diluted and sampled daily during exposures. Concentrations were determined with an atomic absorption spectrometer and measured in lg Zn/L at 12 °C (157, 286, 538, 1044, and 1996), at 9 °C (266, 471, 858, 1556, and 2953), at 6 °C (238, 447, 834, 1623, and 3005), and at 3 °C (234, 462, 832, 1656, and 3305); mean concentrations reported. For the 50-, 83-, 106-, and 141-day rearing time populations, controls were measured at 2, 2, 1, and 3 lg Zn/L, respectively. TTD data were collected starting from 24 h and continued through 96 h. Individual fish were removed from the tanks within 1 h of mortality for the interval from 24 to 96 h. Mortality was more specifically defined as ecological death and was characterized as inverted fish or fish remaining motionless after prodding. On removal, dead fish from the 83- and 104-day rearing times groups were weighed to the nearest 0.001 g. Fish weight was also collected for the 50- and 141-day rearing times fish except that wet weights of individual dead fish were measured after gills were removed. The average weight of excised gills was added to those individual weights to obtain total body weights that were comparable with the 82- and 106-day rearing time fish. Surviving fish from all age groups were sacrificed and weighed at the end of the exposure. Data Analysis Size Categorization This experiment was not originally designed to analyze for zinc tolerance based on fish size, but during the

confirmatory data analysis of chronological age effects it was apparent that fish size over the range of tested zinc concentrations was related to survival and time to death. Exploratory data analysis was conducted to determine if this apparent size-related pattern could be quantified. The exploratory data analysis required a categorization of fish in each chronological age; several categorizations were used with all showing similar patterns of size-related tolerance [e.g., Diedrich (2007) and supplementary data Fig. S1 and Tables S1 and S2). The categorization presented here used four size groups (smallest, small, large, and largest), based on quartiles for each respective population. The fish were not fed over the duration of the exposures (to avoid introduction of additional metalbinding biotic ligands), and they may have lost weight while they were in the exposure tanks. Because wet weight was measured when individuals were removed from the exposure tanks at death or experiment conclusion, it was possible that fish could be classified into the next smaller size category due to this weight loss. To avoid including these potentially misclassified fish in the exploratory data analysis, we applied a standard linear rate of weight loss to be 0.91 %/day (Brown 1951; Fredenburgh 1969; Lawrence 1941). The weight lost for each individual was estimated based on the number of days it was in the exposure. This weight-loss calculation was considered to be conservative because fish metabolism is generally decreased in low food environments (Brown 1951). Fish that may have been categorized into a different size category based on this calculation were removed from the quartiles and thus excluded from sizebased LC50 calculations and survival analysis. Statistical Analysis All 96-h LC50 values were calculated using a three-parameter log-logistic function (chronological age tests) or a two-parameter log-logistic function (size categorization) for binomial data in the (drc) package for R statistical software version 3.1.1 [R Foundation for Statistical Computing (2008), Vienna, Austria] (Ritz and Streibig 2006). Likelihood ratio tests were conducted with the (compParm) function in R to statistically compare whether the parameter estimates of the models were different based on chronological age or size classification (Ritz and Streibig 2006; Wheeler et al. 2006). An a of 0.05 was used as the level of significance. The GraphPad Prism 5 (GraphPad Software,) statistics software package was used to estimate survival curves with the Kaplan–Meier product estimation method. The relationship between curves in different size categories were determined with the Mantel–Haenszel method at a = 0.05. Data were left- and right-censored at 24 and 96 h, respectively.

123

Arch Environ Contam Toxicol

Results Statistical Results

Table 1 Concentration-response LC50 (lg/L) value and slope model estimates for chronological age and size categories with lower and upper confidence levels reported LC50

A Bartlett’s test of weight within the individual chronological age groups indicated that the variances were similar with a test statistic of 1.34 and significance of p = 0.72. The 106-day rearing time fish was the only group that was normally distributed based on the D’Agostino and Pearson test. The mean weights of each population were (0.51 ± 0.13), (0.62 ± 0.22), (0.54 ± 0.13) and (0.46 ± 0.13) g for the 50-, 83-, 106-, and 141-day rearing time groups, respectively (Fig. S2). A Kruskal–Wallis test showed that the mean weights were significantly different (p \ 0.0001) between chronological age groups. Dunn’s multiple comparison tests indicated significant weight differences between all chronological age groups except the 50- and 106-day rearing time populations. Zinc Toxicity and Chronological Age

LCL

UCL

Slope

LCL

UCL

Time group (days) 141

681

529

833

2.20

1.37

3.04

106 83

342 649

286 289

398 810

2.63 2.01

1.47 1.26

3.79 2.77

50

572

402

741

2.26

1.19

3.32

Largest

354

313

395

3.40

2.10

4.70

Large

303

253

354

2.10

1.34

2.86

Small

536

463

610

2.40

1.73

3.06

Smallest

868

737

998

2.03

1.47

2.59

Size group

LCL lower confidence level, UCL upper confidence level

Table 2 Likelihood ratio pairwise comparison of LC50 and slope model estimates Slope

LC50

Mean control mortality was B7.5 % in each of the four chronological age tests. The LC50 values were 572, 649, 342, and 681 lg/L for the 50-, 83-, 106-, and 141-day rearing time populations, respectively (Fig. 1). Confidence intervals (CIs) and slope estimates are listed in Table 1. There were no significant differences in the slope estimates for any of these models (Table 2). The LC50 value was significantly different for the 106-day rearing time population compared with the 50-, 83-, and

Comparison groups

Estimate

p

Estimate

p

Chronological age (days) 141/106

1.992

0.0006

0.838

0.5113

141/83

1.049

0.7789

1.094

0.7496

141/50

1.911

0.3918

0.975

0.9335

106/83

0.527

0.0000

1.306

0.4244

106/50

0.598

0.0002

1.164

0.6667

83/50

1.135

0.5411

0.891

0.6886

Size from combined 83-, 106-, and 141-day rearing time fish Smallest/small

1.618

0.0011

0.849

0.3477

Smallest/large

2.860

0.0000

0.966

0.8724

Smallest/largest Small/large

2.451 1.768

0.0000 0.0006

0.598 1.138

0.0080 0.5802

Small/largest

1.515

0.0010

0.704

0.0785

Large/largest

0.857

0.0997

0.619

0.0247

Bold p values indicate a significant difference at a = 0.05

141-day-old populations. No other chronological age LC50 estimates were significantly different from each other (Table 2). Zinc Toxicity and Juvenile Size

Fig. 1 Concentration-response curves from 96-h zinc exposure to four S. trutta populations. Each population was reared at a different temperature resulting in chronological ages ranging from 50 to 141 days post-swim-up at the start of the exposure

123

The percent mortality at 96 h for all fish in each chronological age exposure separated into size category based on quartiles (Table 3) was calculated. This comparison was performed for zinc concentrations that resulted in 15–90 % mortality (average of the replicates). In all cases but one (83-day rearing time population in 266 lg/L), the largest fish had greater mortality than the smallest fish. There was

Arch Environ Contam Toxicol Table 3 Size categories determined from weight at time of removal from zinc exposures and based on quartiles of each respective group with potential misclassifications because of starvation removed from the combined categories after being grouped into the quartiles Group

Smallest size range (mean ± 1 SD) (g)

Small size range (mean ± 1 SD) (g)

Large size range (mean ± 1 SD) (g)

Largest size range (mean ± 1 SD) (g)

Chronological age (days) 141

0.148–0.389 (0.311 ± 0.064)

0.392–0.456 (0.430 ± 0.020)

0.457–0.516 (0.476 ± 0.019)

0.526–0.894 (0.645 ± 0.076)

106 83

0.163–0.459 (0.376 ± 0.072) 0.213–0.457 (0.372 ± 0.06)

0.46–0.543 (0.502 ± 0.024) 0.458–0.597 (0.531 ± 0.04)

0.546–0.617 (0.579 ± 0.022) 0.599–0.759 (0.672 ± 0.048)

0.618–0.943 (0.693 ± 0.068) 0.76–1.432 (0.911 ± 0.139)

50

0.256–0.441 (0.374 ± 0.057)

0.445–0.533 (0.497 ± 0.022)

0.535–0.585 (0.563 ± 0.015)

0.586–0.907 (0.691 ± 0.099)

Size from combined 83-, 106-, and 141-day rearing time fish Largest

NA

NA

NA

0.639–1.432 (0.771 ± 0.133)

Large

NA

NA

0.515–0.634 (0.569 ± 0.034)

NA

Small

NA

0.428–0.51 (0.464 ± 0.019)

NA

NA

Smallest

0.148–0.423 (0.333 ± 0.062)

NA

NA

NA

NA not applicable Size categories for chronological age populations (not combined ages) were only used for overall percent mortality comparisons at 96 h. The size categories for the combined 83-, 106-, and 141-day rearing time fish were used for concentration–response models and survival curves

Fig. 2 Percent of S. trutta that died during a 96-h exposure to zinc when the post-swim-up chronological age at start of test was a 50 days, b 83 days, c 106 days, and d 141 days. Fish are categorized by size and exposure concentration. The numbers on the graph are the number of individuals in each category. An asterisk indicates that this category was not included because the n was \5. Only concentrations that resulted in 15 to 90 % average mortality are included here

also a general trend of decreasing percent mortality as size decreased (Fig. 2). Individual chronological age exposure groups (‘‘Results’’ section) did not provide sufficient statistical power for LC50 estimates of size categories. To increase power, the 83-, 106-, and 141-day-exposure fish data were combined and then recategorized into the four size groups (smallest, small, medium, and large) based on the quartiles of this combined group; as with the individual chronological age quartiles, fish that were potentially misclassified

because of weight loss were removed from the quartiles and not included in further analysis. The exposure concentrations were similar for these three chronological ages, so the zinc concentrations were averaged for the exposure– response model [246, 460, 841, 1611, and 3117 lg Zn/L (Table 4)]; the 50-day rearing time population was not included in this combined group because the zinc concentrations for this test were very different than the concentrations used in the three subsequent tests. The weights in the group of three temperatures combined ranged from

123

Arch Environ Contam Toxicol Table 4 Comparisons of survival curves of the smallest, small, large, and largest size categories for data with similar exposure concentrations in the combined 83-, 106-, and 141-day rearing time fish averaged Mean concentration ± SD (lg/L) 246 (17)

v2 6.911

p 0.0748

460 (12)

13.83

0.0031

841 (14.5)

23.21

The effects of chronological age and size on toxicity of zinc to juvenile brown trout.

A series of toxicity tests were conducted to investigate the role of chronological age on zinc tolerance in juvenile brown trout (Salmo trutta). Four ...
489KB Sizes 0 Downloads 11 Views