2001 Progress Report: Evaluation of Endocrine-Distrupting Chemical Effects Across Multiple Levels of Biological Organization: Integration of Physiology Behavior and Population Dynamics In FishesEPA Grant Number: R827399
Title: Evaluation of Endocrine-Distrupting Chemical Effects Across Multiple Levels of Biological Organization: Integration of Physiology Behavior and Population Dynamics In Fishes
Investigators: Thomas, Peter , Rose, Kenneth A. , Fuiman, Lee A.
Institution: The University of Texas at Austin
EPA Project Officer: Klieforth, Barbara I
Project Period: October 1, 1997 through September 30, 1999 (Extended to February 29, 2004)
Project Period Covered by this Report: October 1, 2000 through September 30,2001
Project Amount: $862,290
RFA: Endocrine Disruptors (1997) RFA Text | Recipients Lists
Research Category: Economics and Decision Sciences , Health , Safer Chemicals , Endocrine Disruptors
Objective:The overall aim of this research is to estimate the impacts of several representative endocrine disrupting chemicals (EDCs) on Atlantic croaker (Micropogonias undulatus) populations in marine environments using a suite of reproductive and larval response inputs into an integrated population model.
The specific objectives are to: (1) determine the effects of the representative EDCs on biomarkers of gonad production and gonadal growth; (2) investigate the impacts of the EDCs on biomarkers of gamete maturation, fertilization success, and larval survival; (3) assess the parental transfer of the EDCs to gametes and offspring; (4) determine the effects of parental exposure to the EDCs on ecological performance skills of larvae; (5) determine the influence of parental exposure to the EDCs on larva metabolism, growth, and development; and (6) develop a suite of predictive computer models for scaling individual-level effects of EDCs to fish population responses.
Progress Summary:Objectives 1, 2, and 3
In Year 2, croakers were fed with one of two concentrations of methyl mercury in the diet (target 5 µg and 10 µg/100 g/bw/day) resulting in muscle concentrations of 0.5 µg/g tissue and 4.0 µg/g tissue methyl mercury after 6 weeks of exposure. Mercury treatment resulted in elevated plasma testosterone levels in females but did not significantly alter androgen levels in males. Plasma estradiol levels also were significantly elevated in females exposed to the higher methyl mercury treatment. A similar pattern of increased steroidogenesis was observed when ovarian tissues were incubated in vitro. Interestingly, these increases in ovarian steroidogenesis were not accompanied by elevations in basal and LHRH-stimulated gonadotropin secretion. In contrast, the gonadotropin response to LHRH was alternated in the high dose group. The results suggest that methyl mercury exerts direct effects on teleost ovaries to stimulate androgen production. Moreover, the pattern of a dose-related increase in ovarian steroidogenesis after exposure to methyl mercury also was observed with ovarian growth as assessed by the gonadosomatic index (GSI). Both methyl mercury treatments also significantly impaired final gamete maturation. Sperm motility showed a dose-dependent decline after methyl mercury treatments from 86 percent motile sperm in controls to 62 percent and 58 percent in the low and high dose methyl mercury groups, respectively (P < 0.05). Several measures of reproductive success also were impaired after treatment of parents with methyl mercury, including hatching success (controls 88.7 percent, low dose 63.2 percent, high dose 71.8 percent, both methyl mercury treatments significantly different from controls), and 12-hour posthatch survival (controls 48.1 percent, low dose 27.9 percent, and high dose 41.1 percent). The mean accumulation of mercury per egg was 0.04 ng and 0.4 ng in the low and high dose groups, respectively.
These studies show that methyl mercury can impair three critical stages of the reproductive cycle: gonadal growth, gamete maturation, and fertilization/early egg and larval survival. However, dose-related effects of methyl mercury were not observed with all the reproductive biomarkers. The pattern of endocrine and reproductive disruption differed somewhat from that observed in the previous year with the PCB mixture. However, the sensitivity of several critical reproductive stages to interference by methyl mercury and the lack of a dose-response relationship were characteristics shared with the PCB effects observed in the first year experiment.
In the second year, several additional analyses of the PCB experiment were completed, including estimations of fecundity that showed a PCB-related decrease. These data were used for initial modeling experiments.
Objectives 4 and 5
Aroclor 1254. In Year 2, we completed the video and data analysis of the Aroclor 1254 study. The growth of low-dose treated larvae (hereafter referred to as the exposed larvae) was significantly lower than the control larvae (P = 0.011). The instantaneous growth coefficients for the control and exposed larvae were 0.012 and 0.008 day-1, respectively. A 3-day delay in development was evident in the exposed larvae following yolk absorption (day 5). On day 9, the exposed larvae were significantly smaller (P = 0.006) than the control larvae and were similar in size to the control larvae at 5 days posthatching (P = 0.95). Similarly on day 13, the exposed larvae were significantly smaller (P = 0.0004) than the control larvae and were similar in size to the control on day 9 (P = 0.67). In Atlantic croaker larvae, development is size-related rather than age-related. Thus, development is impaired in croaker larvae following exposure to Aroclor 1254.
There were no effects of dose or age on the routine swimming activity (percent time active, rate of travel, active swimming speed) of the control and exposed larvae. However, there were significant differences in the response of control and exposed Atlantic croaker larvae to a vibratory stimulus. The percentage of larvae responding to the vibratory stimulus was significantly different (P = 0.002): responsiveness in the control and exposed larvae was low on days 5 and 9 (ca. 15 percent); however, responsiveness was significantly higher (P = 0.02) in the control larvae on day 13 (35 percent) compared to the exposed larvae (15 percent). There were no differences in latency between the control and exposed larvae (P = 0.78), although the exposed larvae tended to respond on average 20 ms later than the control larvae. There was a significant effect of PCB exposure on both mean burst speed (P = 0.02) and maximum burst speed (P = 0.005) during the response to a vibratory stimulus. In the control larvae, mean and maximum burst speeds increased with age, in contrast, the exposed larvae showed very little increase with age. The mean and maximum burst speeds of control larvae were significantly faster than those of the exposed larvae on day 13 (mean burst, P = 0.007; maximum burst, P = 0.005). The mean and maximum burst speeds of the exposed larvae on days 9 and 13 were similar to those for the control larvae on days 5 (mean burst, P = 0.09; maximum burst, P = 0.18) and 9 (mean burst, P = 0.18; maximum burst, P = 0.0.12), respectively. It is likely that reduced responsiveness of the exposed larvae to the vibratory stimulus on day 13 is a result of the 3-day impairment of development.
Methylmercury. The larval "survival skills" investigated were activity, swimming speed, and responsiveness to vibratory and visual stimuli on days 3 (yolk absorption), 6 (oil absorption), 10, and 17 posthatching. These measures are indicators of the ability of a larva to forage for food successfully (routine swimming activity) and to avoid predators (vibratory and visual stimuli). To date, preliminary data analysis of the vibratory stimulus results has been completed. The percentage of larvae responding to the vibratory stimulus in the zero-, low-, and high-dose treatment spawns were low (ca. 10-30 percent). The percentage of larvae responding to the vibratory stimulus was not correlated with the initial methylmercury content in the egg on days 3, 6, 10, or 17 posthatching (all P > 0.05). However, on days 10 and 17 posthatching, responsiveness tended to decrease with increasing initial methylmercury egg loading (P < 0.10). Latency was not correlated with the initial methylmercury loading in the egg on days 3, 6, or 10 (all P > 0.05). However, on day 17 posthatching, latency was positively correlated with increasing methylmercury concentration (P < 0.05). Mean burst speed during the escape response was negatively correlated with the initial methylmercury concentration in the egg on days 3 (P < 0.01) and 17 (P < 0.05); however, no such relationship was found on days 6 and 10 (P > 0.05).
The growth rates (size at age) of the zero, low-, and high-dose treated larvae were similar (P = 0.27). The instantaneous growth coefficients for the zero, low-, and high-dose treated larvae were 0.014, 0.012, and 0.013 day-1, respectively.
During the second year, we have completed the development of preliminary, but fully functional, versions of three linked models. The three models are: statistical, individual-based, and matrix projection. The statistical model relates the swimming speed and behavioral responses of the fish larvae exposed to EDCs to the probability of escaping a real fish predator. The individual-based model tracks 10,000 individual larvae of a cohort through their daily growth and mortality. The individual-based model predicts daily growth rate from encounters with zooplankton, and predicts daily mortality from encounters and capture by individual jellyfish (sea nettle and ctenophore) predators and fish predators. The 10,000 initial larvae are configured with swimming speeds and probabilities of escaping predator attacks characteristic of the control larvae or the contaminant-exposed larvae. The individual-based model then predicts ocean larva stage growth and mortality for control and contaminant-exposed larvae. These growth and mortality rates are used to determine how the two elements of the matrix model corresponding to ocean larvae should change between control and contaminant-exposed larvae. The third model is a matrix projection population model that simulates 100 years of fish population abundances and age-structure. Our matrix model is quite detailed to permit realistic simulation of stage-specific effects of contaminants. The matrix model allows for multiple time steps (e.g., eggs are daily; adults are annual), multiple spatial regions, one or two spawning cohorts within each year, density-dependence, and stochastic variation. One-hundred year simulations are performed under baseline and assumed contaminant exposure.
Preliminary versions of the statistical, individual-based, and matrix projection models have been developed and applied to the Atlantic Bight croaker population and the PCB experiments performed in the first year of the project. We used linear regression, multivariate methods, and regression trees to successfully relate behavioral responses to probability of escaping predator attacks. Our attempt to use neural networks were unsuccessful, as network models were highly sensitive to initial estimates assigned to parameter values. The individual-based larval cohort model was configured to roughly represent the conditions experienced by croaker larvae in the Atlantic Bight. According to the individual-based model, the low-dose PCB treatment resulted in slowed growth (longer stage duration) and higher mortality rate than control larvae. This translated into an 8 percent lower diagonal element of the matrix model and a 22 percent lower subdiagonal element of the matrix model, which together determine the survival of ocean larval stage. We also used the fecundity experiments from year 1 to impose a 13 percent reduction in the first-time fecundity of croaker, and the experimental results that measured survival at 24 hours after fertilization to increase the mortality rate of eggs in the matrix model.
The matrix model was configured to simulate croaker population dynamics for the Mid- and South Atlantic Bights. Six life stages that comprise the first year of life were represented: egg, yolk-sac larva, ocean larva, estuarine larva, early juvenile, and late juvenile. We also separated the first year of life into two regions (Chesapeake Bay in Virginia and estuaries of North Carolina), and two spawning cohorts (Fall and Spring). As much as possible, field data from Virginia and North Carolina long-term monitoring of young of the year croaker were used to determine growth and mortality rates of each of the six life stages. Finally, we included density-dependent survival on the late juvenile stage and interannual variation in mortality rates of eggs and yolk-sac larvae.
The culmination of the three linked models is shown in Figure 1, 100-year simulations of croaker population abundances under baseline and with all croakers assumed to experience the low-dose PCB effects. Only larvae from age-1 spawners are affected by PCBs in the simulations because we assume that almost all of the PCBs in the mothers are released in the eggs of their first spawning. It is important to note that the predicted effects of the PCB shown in Figure 1 are not meant to be what actually happened to croakers. Rather, the results demonstrate that our approach of coupled laboratory and linked models can be used to predict population effects of EDC exposure. We simulated PCB effects on multiple life stages, including sublethal (behavioral) effects in these simulations.
Future Activities:No major changes to the original project plan are envisioned for Year 3. The experiments with nonylphenol are ongoing. Several manuscripts on the results of the studies in Years 1 and 2 are in preparation.
Journal Articles on this Report : 1 Displayed | Download in RIS Format
|Other project views:||All 9 publications||1 publications in selected types||All 1 journal articles|
||Murphy CA, Rose KA, Alvarez MC, Fuiman LA. Modeling larval fish behavior:scaling the sublethal effects of methylmercury to population-relevant endpoints. Aquatic Toxicology 2008;86(4):470-484.||