Environ. Sci. Technol. 2009, 43, 6364–6369
Exposure to Persistent Organic Pollutants and First-Year Survival Probability in Gray Seal Pups A I L S A J . H A L L , * ,† G A R E T H O . T H O M A S , ‡ AND BERNIE J. MCCONNELL† Sea Mammal Research Unit, Scottish Oceans Institute, University of St Andrews, St Andrews, Fife, KY16 8LB Scotland and Center for Chemicals Management, Lancaster Environment Center, Lancaster University, Lancaster, LA1 4YQ U.K.
Received February 11, 2009. Revised manuscript received July 9, 2009. Accepted July 14, 2009.
Many studies have demonstrated that persistent organic pollutants are transferred from mother to pup during lactation in phocid seals, but none have been able to determine the significance of these findings for survivorship. The aim of this study was to investigate the relationship between blubber contaminant concentrations and first-year survival in gray seal pups. A mark-recapture framework was used to estimate survival probabilities and animals were “marked” using novel mobile phone tags. Individual and group covariates (sex, condition, and blubber contaminants) were embedded within a live-resighting model. The most significant covariates remained condition at weaning and sex (males in poor condition had the lowest survival probability), as was found previously, but there was also evidence indicating that higher blubber contaminants additionally decreased survivorship. The models’ Akaike’s Information Criteria (AICs) and their associated weights, point toward the tetrapolybrominated diphenyl ether congeners (dominated by BDE-47) as being the most important group of contaminants affecting survival probability, followed by the total dichlorodiphenyltrichloroethanes (DDTs) and pentapolychlorinated biphenyl congeners. These compounds were not the most abundant in the blubber, suggesting further studies into their toxicological effects in this species are necessary. The specific mechanisms driving the reduction in survivorship remain unknown.
Introduction Marine mammals are exposed through the food chain to a variety of organohalogenated (OH) compounds, such as the polychlorinated biphenyls (PCBs), polybrominated diphenyl ethers (PBDEs), and dichlorodiphenyltrichloroethanes (DDTs). This exposure has been related to a wide variety of adverse health effects including impacts on immunity (1), reproduction (2), and endocrine function (3). Gray seal pups (Halichoerus grypus) ingest relatively high levels of these compounds during weaning, through contaminated maternal milk (4). Over the average 18 day lactation period (5), the pup can quadruple its mass, gaining up to 2.5 kg day-1, most of which is laid down as fat (blubber) stores (6). The maternal * Corresponding author phone: +44 1334 463443; fax: +44 1334 462632; e-mail:
[email protected]. † University of St Andrews. ‡ Lancaster University. 6364
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 16, 2009
contaminants transferred are, thus, accrued in the pups’ blubber. This contaminant legacy makes up a substantial proportion of the total contaminant burden that a seal will acquire during its life (7). As the pup develops into a juvenile, it may be particularly vulnerable to the effects of contaminants that may interfere with maturing functions, such as immunity and growth. During the first year of life, pups undergo rapid physiological change especially as diving capabilities develop (8). This study investigates whether there is any relationship between blubber OH contaminant concentrations (PCBs, PBDEs, DDTs, and other organochlorine compounds) and first-year survival in postweaned gray seal pups. Previous studies investigating factors affecting survival to age one year in U.K. gray seal pups found condition at weaning, sex, and serum total immunoglobulin (IgG) to be significant determinants (9, 10). Males in poor condition (with less blubber) with relatively high levels of total IgG had the lowest survival probability. In the previous studies, these covariates were embedded in a mark-recapture survivorship model using a joint live-resighting, dead recovery, model (11). Here, a similar approach was taken, with the additional advantage of a new marking method using telemetry “phone” tags (12) to improve the resighting probabilities, enabling a simpler live-resighting model to be used. The significant group and condition covariates from the previous studies were again included, with the additional blubber contaminant variables. The blubber contaminants, on a total basis, by chemical class and by congener subgroup, were included in separate models to determine which were likely to be of most importance.
Materials and Methods Study Population and Survival Covariate Measures. A random sample of 60 pups born at the Isle of May on the east coast of Scotland in 2002 were tagged with flipper tags (Dalton Animal Identification Systems, Oxford, U.K.) and weighed twice while they were still suckling. Marked animals were recaptured as postweaned pups and reweighed. Pups were sedated with a 10-20 mg dose of Zoletil 100 (Virbac, France). Nose-tail length and axillary girth measurements were also taken (Table 1). Blubber samples were collected from the lateral pelvic region using a 6 mm biopsy punch (Acuderm, Florida) and stored at -20 °C until processing. Mobile phone tags (12) were attached to the back of the neck using epoxy glue, and animals were individually color marked for estimation of tag loss during resighting trips. All procedures were carried out under U.K. Home Office, Animal (Scientific Procedures) 1986 Act licenses. Mass/length, as a condition index that reflects the pup’s stored blubber reserves (9) was calculated for each animal. Various condition indices as surrogates for total body fat are available, but to remain consistent with previous studies on factors affecting survival, mass/length (which was highly correlated with axillary girth, R2 ) 0.81, p < 0.0001) was again used (see discussion in ref 10 for rationale). Because animals could not be sampled exactly on the day they weaned, marked pups were weighed repeatedly before and after weaning and linear regression models fitted to the growth and mass loss data for each animal. Estimated mass at weaning was then calculated from the intersection between the two regression lines (9). The standard length of the pups was assumed to be the same at weaning as when animals were measured, as repeat measures of marked animals indicated no significant growth after weaning (9). Blubber biopsy samples were collected when animals were first recaptured after weaning, and this varied by individual. 10.1021/es9004398 CCC: $40.75
2009 American Chemical Society
Published on Web 07/22/2009
TABLE 1. Geometric Mean, Geometric 95% Confidence Intervals, and Range for Estimated Mass at Weaning, Length, Girth, and Condition Index in Post-Weaned Gray Sealsa estimated mass at weaning (kg) standard length (cm)
mean 95% CI range (min-max) a
males
females
41.9 39.4-44.5 30.6-56.4
38.5 36.6-40.5 28.7-48.8
males
females
axillary girth (cm) males
females
103 101 96.7 93.7 103-106 99-102 95.8-99.2 91.6-94.2 94-112 91-110 86-111 82-104
condition index (mass/length, kg/cm) males
females
0.404 0.386-0.423 0.326-0.504
0.381 0.366-0.397 0.315-0.465
Males n ) 28, females n ) 27.
Contaminant concentrations in gray seal pups increase with days after weaning as fat is used for energy and contaminants are concentrated in less blubber (13, 14). Using the linear relationship between the decrease in mass during the fast and an increase in contaminant concentrations previously established for postweaned gray seal pups from the Isle of May (13), concentrations were adjusted to an estimated contaminant concentration at weaning. Contaminant Concentrations in Blubber Biopsy Samples. The methods used to determine the blubber contaminant concentrations on a congener specific basis are described in refs 7 and 15. Samples were mixed well with anhydrous sodium sulfate, extracted with dichloromethane (DCM) using an accelerated solvent extraction system (Soxhlet). An aliquot was taken for gravimetric lipid determination, and the remaining sample was transferred to hexane. All samples were spiked with seven 13C-labeled PCBs and 13C-labeled BDE 209 before extraction. Samples were then cleaned by chromatography using silica gel treated with concentrated sulphuric acid, eluted with hexane. All samples received a secondary cleanup using gel permeation chromatography before being concentrated to a small volume with internal standards added. Samples were analyzed for 45 PCBs (PCBs 18, 22, 28, 31, 41/64, 44, 49, 52, 54, 60/56, 70, 74, 87, 90/101, 95, 99, 104, 105, 110, 114, 118, 123, 138, 141, 149, 151, 153, 155, 156, 157, 158, 167, 170, 174, 180, 183, 187, 188, 189, 194, 199, and 203) using a GC-MS system (Finnigan TRACE) in SIM mode using an EIC source with 2 ion masses monitored for each chemical of interest. The concentration of 12 organochlorine pesticides was also determined. Nine of these (namely R-chlordane, γ-chlordane, HCB, o,p′-DDD, p,p′-DDD, o,p′-DDE, p,p′-DDE, o,p′-DDT, and p,p′-DDT) were also analyzed using the Finnigan TRACE GC-MS in EI mode. Samples were injected in splitless mode, and separation was achieved on a 50m CP-Sil 8 chromatography column (Chrompak, Varian Ltd., Surrey, U.K.) with a 2 m retention gap. The mobile phase was helium, with a flow rate of 1 mL/min. Twenty one PBDEs (BDEs 17, 28, 32, 35, 37, 47, 49, 71, 75, 77, 85, 99, 100, 119, 138, 153, 154, 166, 181, 183, and 190) and the 3 remaining pesticides (R-HCH, β-HCH and γ-HCH) were analyzed using a GC-quadrupole MS system (Fisons MD800 or Finnigan TRACE) in NCI mode, using ammonia as the reagent gas. Samples were injected in splitless mode, and separation was achieved on a 30m DB5MS chromatography column with a 2 m retention gap. The mobile phase was helium, with a flow rate of 1 mL/min. A set of seven calibration standards, containing all of the internal standards, recovery standards, and analytes, was run on the instrument before and after a batch of up to 24 samples. Within the batch, after every 6 samples, a specially prepared “QC” standard was run. Samples were quantified using the Thermo “Xcaliber”’ instrument software, and calibration and quantification was achieved using an internal standards method. To be accepted as the analyte of interest, a peak in a sample had to be eluted from the GC column at the correct retention time and show the correct ratio (compared to the standards) between the two ions measured.
The method detection limit was defined as the higher of either (i) three times the standard deviation of the blank values or (ii) the instrument detection limit. Recoveries averaged 67-81% for each of the 13C labeled standards. First-Year Survival Probability. In this study, we extended the models of first-year survival in gray seal pups (9, 10) to include blubber contaminants that are putative endocrine and immune function disrupters. We used a mark-recapture framework with a Cormack Jolly Seber (CJS) live resighting model (16) where animals were “resighted” using mobile phone tags (12). Marked animals are released into the population then subsequently “encountered”, usually physically or visually. In our study, the phone tags sent an SMS text message every 6 h. The successful receipt of a message ashore, therefore, constituted a resighting event or “encounter” within the coastal zone of phone coverage (gray seals return to shore to haul out). If marked animals are released into the population on occasion 1, then each succeeding encounter (successful receipt of a text message) is one encounter occasion, i.e., release f S1 f encounter 1 f S2 f encounter 2 f etc. Animals survive from initial release to the first reencounter with probability S1 and from the first encounter occasion to the second with probability S2, etc. The recapture probability at encounter occasion 1 is p2, and p3 is the recapture probability at encounter occasion 2, etc. The survival rate between the last two encounter occasions is not estimable because only the product of survival and recapture probability for this occasion is identifiable. Generally, the survival rates of the CJS model are denoted as Φ1, Φ2, etc., because the quantity estimated is the probability of remaining available for recapture. Thus, animals that emigrate from the study area appear to have died. However, our phone tags had “roaming” agreements with European mobile phone service providers, thus, allowing the study area to extend to all the European coastal zones where gray seal pups can disperse. For pups that moved away from U.K. coverage, we were still able to receive messages (messages were received from Germany and Norway), and Φ is equivalent to survival without emigration confounding. Text messages received were then concatenated into monthly encounter histories from which first-year survivorship could be estimated. Tags lasted 6 months due to battery life, so survival estimates were calculated for the period December to June and annual estimates were adjusted accordingly. Information from double tagging and resighting surveys allowed us to estimate tag loss rates. Of the 60 deployed, data from 55 were used in the final analysis due to the failure of a small number of tags. Model fitting was carried out using the Program MARK (17). Individual and group covariates (i.e., variables which may explain the variation in the animals’ encounter histories, namely, sex (group covariate), condition, PCBs, PBDEs, DDTs, other OC pesticides, and total OHs (individual covariates, specific to each animal)) were embedded in the survival models in an approach similar to VOL. 43, NO. 16, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6365
TABLE 2. Geometric Means and Geometric 95% Confidence Intervals for Concentrations of Total PCBs, Total PBDEs, Total DDTs, Total HCHs, Total Chlordanes and HCB, and Total OH Compounds in Post-Weaned Gray Sealsa
geometric mean 95% CI a
total PCBs (ng/g of lipid)
total PBDEs (ng/g of lipid)
total DDTs (ng/g of lipid)
total HCHs (ng/g of lipid)
total chlordanes and HCB (ng/g of lipid)
total OH compounds (ng/g of lipid)
1096 933-1288
141 129-158
229 195-269
132 120-144
14 12-17
1660 1454-1894
n ) 55.
logistic regression (17). A set of candidate models were fitted to the data and model selection carried out using Akiake’s Information Criterion (AIC) and its small sample size form, QAICc, to determine the best model from the set, given the data. The aim is to choose models that adequately fit the data with as few parameters as possible, making a trade off between bias from having too few parameters and poor precision due to too many. The best approximating model has the lowest QAICc value. As QAICc is measured on a relative scale, the ∆QAICc values are reported as the difference in QAICc between each model in the set and that for the model with the lowest QAICc. Akaike weights (see ref 18 and SI.1 in the Supporting Information for the definition) are also reported where a given weight is considered as the “weight of evidence” (probability) in favor of model i being the best model for the given situation.
TABLE 3. QAICc, ∆QAICc, QAICc Weights, Model Likelihoods, Number of Parameters, and Model Deviances (First Stage of the Model Selection Process)a model 1. Φ(s), p(t) 2. Φ(t), p(t) 3. Φ(s), p(s) 4. Φ(s), p(s × t) 5. Φ(s × t), p(s × t) a
delta QAICc model no. QAICc QAICc weight likelihood parameters deviance 309.84
0.000 0.804
1.000
8
73.343
313.97
4.136 0.102
0.126
12
67.935
314.19
4.353 0.091
0.113
4
86.605
321.05 11.209 0.003
0.004
14
69.976
339.73 29.880 0.000
0.000
24
60.465
t ) time, Φ ) survival, p ) recapture, s ) sex.
Results Individual and Group Covariates of Survival. The study pups all were overtly healthy, were successfully weaned by their mothers, and survived the postweaning fasting period, determined both from surveying the colony until all the animals had departed to go to sea for the first time and from the subsequent text messages received. One individual was a possible outlier, weighing 56.4 kg at weaning. No animals were less than 25 kg at weaning, and weaning masses were highly comparable to the previous cohorts studied (4, 9, 10). Males were significantly heavier than females (two-sample t test, unequal variances, p ) 0.03, Table 1). The geometric mean concentrations and 95% geometric confidence limits for the contaminant congeners measured in the blubber are given in Tables S1a-S1c in the Supporting Information. The proportions of PCBs, DDTs, and PBDEs congener groups are also given in Table S2 in the Supporting Information. The concentrations of the PCB congeners on a lipid weight basis were all lognormally distributed, and parametric statistical analyses were carried out on the log10 transformed concentrations. The geometric mean concentrations and 95% geometric confidence limits for the congener sums within each chemical group are shown in Table 2. Profiles were dominated by the total PCBs whose concentrations made up 66% of the total OH contaminants measured in the blubber. First-Year Survival Probability. In the first stage of modeling, a goodness of fit test of the model to the data without covariates was carried out. Males and females were considered separately, using sex as a group covariate, and survival and recapture probabilities were allowed to vary with time. All combinations of sex and time varying survival and recapture were included (Table 3). The goodness of fit tests found there were no significant model assumption violations. To account for any lack of fit effect on model selection, a parametric bootstrap goodness of fit procedure (100 simulations) incorporated into program MARK was used (17). This estimated an overdispersion factor, which was then used to adjust the AICc statistics to quasi-AICc values (QAICc) (19). The overdispersion factor was estimated using the ratio of the observed deviance statistic to the average value obtained from the bootstrap replicates. This indicated only mild overdispersion (cˆ ) 1.157). 6366
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 16, 2009
The best model from the first stage was found to be survival varying by sex (males and females having different survival probabilities) and recapture varying with time (Table 3). Lower recaptures in February to April (Figure 1) probably reflect an exploratory behavioral phase where the pups disperse to sea before establishing regular foraging patterns (20). The ∆QAICc is essentially a “strength of evidence” comparison; the larger the ∆QAICc, the less plausible is the fitted model as being the best approximating model in the candidate set. Some “rules of thumb” useful in assessing the relative merits of the various models were outlined by ref 21: models having ∆QAICc e 2 have substantial support (evidence), those where 4 e ∆QAICc e 7 have less support, and models having ∆QAICc g 10 have essentially no support. The second stage of modeling also incorporated sexdependent survival and time-dependent recapture. The results of the model selection process for the second stage are shown in Table 4. The best model from the set of
FIGURE 1. Recapture probabilities (bar ) standard error) by time (1 ) January) from model 1 Φ(sex × condition), p(t).
TABLE 4. QAICc, ∆QAICc, QAICc Weights, Model Likelihoods, Number of Parameters, and Model Deviances (Second Stage of the Model Selection Process)a model
QAICc
delta QAICc
QAICc weight
model likelihood
no. parameters
deviance
1. Φ(s × cond), p(t) 2. Φ(s × cond + tetra-PBDEs), p(t) 3. Φ(s × cond + DDTs), p(t) 4. Φ(s × cond + penta-PCBs), p(t) 5. Φ(s × cond + hexa-PCBs), p(t) 6. Φ(s × cond + hexa-PBDEs), p(t) 7. Φ(s × cond + other pesticides), p(t) 8. Φ(s × cond + hepta-PCBs), p(t) 9. Φ (s × cond + HCHs), p(t) 10. Φ(s × cond + mono-ortho-PCBs), p(t) 11. Φ(s × cond + tri-PCBs), p(t) 12. Φ(s × cond + penta-PBDEs), p(t) 13. Φ(s × cond + tetra-PCBs), p(t) 14. Φ(s × cond + total PBDEs), p(t) 15. Φ(s × cond + hepta-PBDEs), p(t) 16. Φ(s × cond + tri-PBDEs), p(t) 17. Φ(s × cond + total PCBs), p(t)} 18. Φ(s × cond + total Ohs), p(t) 19. Φ(s × cond × tetra-PBDEs), p(t) 20. Φ(s), p(t) 21. Φ(t), p(t)
295.67 296.97 297.26 297.37 297.59 297.63 297.67 297.69 297.78 297.85 298.06 298.06 298.06 298.06 298.06 298.07 298.07 298.08 299.18 309.84 313.97
0.000 1.293 1.583 1.694 1.915 1.956 2.001 2.016 2.105 2.175 2.382 2.386 2.388 2.388 2.391 2.392 2.400 2.404 3.508 14.164 18.300
0.1393 0.0723 0.0631 0.0598 0.0535 0.0524 0.0512 0.0509 0.0486 0.0450 0.0423 0.0423 0.0422 0.0422 0.0421 0.0421 0.0419 0.0419 0.0241 0.0001 0.00001
1.0000 0.5238 0.4531 0.4287 0.3838 0.3760 0.3676 0.3650 0.3491 0.3371 0.3039 0.3034 0.3030 0.3030 0.3026 0.3024 0.3012 0.3006 0.1731 0.0009 0.0001
10 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 8 12
273.74 272.63 272.92 273.03 273.25 273.29 273.33 273.35 273.44 273.51 273.72 273.72 273.72 273.73 273.73 273.73 273.73 273.74 272.40 292.60 287.19
a
t ) time, Φ ) survival, p ) recapture, cond ) condition, s ) sex.
21 included condition-dependent survival by sex and timedependent recapture. Models 2 to 19 included the various contaminant covariates. The top two models (with no covariates) from the first stage now ranked 20th and 21st. Although the best model did not include any of the contaminant covariates and was better supported than the others in the candidate set, there was some evidence to support inferences from the second, third, and fourth models. Their ∆QAICc values were all 0.4). The selected best model had a weight of only 0.139, and the summed weight for all the models containing contaminant covariates was 0.836. On the basis of this evidence (as discussed in ref 22), contaminants as factors affecting survivorship in gray seal pups are worthy of further consideration. Due to the very high number of potential models that could have been fitted to the data if all congeners were included separately (over 1000), congeners were grouped for inclusion in the models. The second best model, given the data, included the tetra-PBDEs (Table 4, BDEs 47, 49, 66, 71 and 75). These were dominated by BDE-47, which made up 55% of all PBDE congeners and 89% of the tetra-PBDEs. The relationship between the monthly survival probabilities and log10(tetra-PBDEs) with 95% confidence intervals for males and females (holding condition at a constant mean value) is shown in Figure 2. The effect of increasing blubber concentrations of tetra-PBDEs was more marked in males than females (monthly Φ reduced by 0.217 in males and 0.157 in females, over the range of concentrations measured in the study). The next two models in the candidate set also had ∆QAICc values 0.05). From the best model that included condition-dependent survival, the coefficients for sex and condition effects were as follows: males 1.61 ( SE 0.55; male condition 2.42 ( SE 1.08; females 2.04 ( SE 0.48; female condition 0.69 ( SE 0.42; thus, the odds of survival were estimated to be 1.5 times higher for females than males, regardless of condition. This gives an estimated annual survival probability for males in average condition of 47.5% and females of 63.9% (after accounting for the effect of tag loss, see ref 9). VOL. 43, NO. 16, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6367
Discussion We used a mark-recapture survival study to investigate the impact of blubber contaminants in postweaned gray seal pups on their first-year survival probability. While the most important predictors of survival remained condition at weaning and sex, there was evidence to support the hypothesis that higher blubber contaminants also decrease first-year survival probability. This study is, to our knowledge, the first to investigate the ultimate effect of contaminant exposure on early survivorship in seals. The models were a good fit to the data, the standard errors around the resulting survival estimates were not large, and the first-year survival estimates were in line with our earlier findings (9, 10). The study pups appeared to be in good health when they were last seen on the breeding colony, with no obvious signs of infection. One animal dropped out of the study, as it was taken into a rehabilitation center in Germany after a few months at sea due to malnourishment, and the tag was removed. The study animals were comparable in size (with an even number in each sex class) to weaned pups studied at this colony in the past (4, 9) and, therefore, are probably a representative sample of the population of weaned pups. The finding that postweaning survival in gray seal pups is influenced by their condition at weaning and that this relationship varies between the sexes confirms our previous studies (9). Male pups in poor condition again had the lowest first-year survival probability. This confirmation suggests that marking and resighting individual animals using phone tags is robust. However, the assumptions underlying the statistical approach remain the same as for the previous studies (9, 10). Animals in a covariate group are assumed to have the same survival and resighting probabilities, and the survival of an animal does not depend on whether it is at risk of resighting. Information on the movements of postweaned gray seal pups born at the same study site, using satellite relay data loggers (23), indicates that pups do not closely associate with each other and using phone tags to “resight” animals ensured resighting was not affected by observation error. While the model containing sex and condition covariates was the best model in the set, those models incorporating the various OH contaminant covariates were also considered useful for making inferences with substantial evidence for support. The rational behind considering all models in the set that had ∆QAICc < 2 is based on discussions in the literature (21, 22, 24, 25). Models are approximations of reality, and model selection methods can be utilized to determine if the information in the data is simple (with a few dominant effects only) or more complex (22). By employing a parsimonious method in model fitting (26), the optimum number of parameters can then be justified, given the size of the sample. Our results suggest that the best model was twice as likely to be the “true” model given the data as the others in the set, but it did not have overwhelming support. There was thus a 1:3 chance that the second model was in fact the best approximating model. Notably, those contaminants that were the best predictors of survival were not the most abundant in the blubber. The tetra-PBDEs were the most important compounds (together with condition and sex), and their congener profiles were dominated by BDE-47. This is rather surprising given that the PCBs are thought to be generally more toxic than the PBDEs (27). In addition, the tetra-PBDEs were not significant predictors of preweaning growth rate or condition or mass at weaning. The mechanism of action linking exposure to reduced survival is not known but various toxic effects for tetra-PBDEs and BDE-47 have been reported. For example, thyroid hormone levels were decreased in lambs born to exposed sheep (28), and thyroxine levels were significantly reduced in BDE-47 exposed rats and mice (29). In phocid 6368
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 16, 2009
seals, correlations between circulating thyroid hormone levels (both positive and negative) and blubber PBDEs have been reported (13, 30). Exposure of neonatal rats to low doses of BDE-47 increased their locomotor activity (hyperactivity (31)), and significant immunotoxic effects have been found in children (32) and mice (33). A technical PBDE mixture, which included BDE-47, triggered aryl hydrocarbon receptor mediated gene expression in rat hepatoma cells and in zebrafish embryos (34). Neonatal exposure of mice to BDE-47 indicated permanent aberrations in spontaneous behavior (35) that appeared to worsen with age. Neurotoxic effects of various PBDE congeners on the developing brain in vitro have also been found, including effects of BDE-47 (36). More toxicity studies have been carried out using BDE-47 alone, and although the tetra-PBDE profile was dominated by BDE-47, extrapolating findings from studies using individual congeners is often not recommended due to the distinct toxicities associated with the different congeners and how they may interact in a mixture. DDTs and penta-PCBs were also potentially important survival predictors that share similar endocrine and immune function effects with the PBDEs (37, 38). They are all also neurotoxic to varying degrees (39). Neurological effects could affect dispersal and successful foraging, which is of key importance to gray seals that have no parental guidance and limited fat reserves when they leave the breeding beach. Pups, therefore, need to develop successful foraging strategies relatively quickly. However, given the variety of effects reported in laboratory animals (28, 29, 31-36), the correlations with circulating thyroid hormones in free-living seals (13, 30), and the earlier finding that IgG levels are important first-year survival covariates in gray seal pups (10), exposure may result in numerous end points. In addition, the composition of the mixture to which the pups are exposed (including extra contaminants, such as mercury not measured here) could, in fact, be more important. The results of this study do not prove causality, and the extent to which our observations support a causal linkage between contaminants and survivorship remains unclear. Further, research using controlled in vivo and in vitro approaches are needed in order to provide mechanistic evidence about the nature of PBDE and related OH toxicity in gray seals.
Acknowledgments This study was supported by the U.K. Department for Environment, Food, and Rural Affairs as part of the EDAQ programme, the Natural Environment Research Council, and Seimens Mobile. All procedures were carried out under Home Office Licence No. 60/3303, and authors thank Scottish Natural Heritage for permission to work at the Isle of May NNR. The authors have no competing financial interests.
Supporting Information Available The geometric mean concentrations and 95% geometric confidence limits for all the contaminant congeners measured in the postweaned gray seal pups are given in Tables S1a-S1c. The proportions of PCBs, DDTs, and PBDEs congener groups are also given in Table S2. The calculation of AIC weights is also described. This material is available free of charge via the Internet at http://pubs.acs.org.
Literature Cited (1) Ross, P. S.; De Swart, R. L.; Timmerman, H. H.; Reijnders, P. J. H.; Vos, J. G.; Van Loveren, H.; Osterhaus, A. D. M. E. Suppression of natural killer cell activity in harbour seals (Phoca vitulina) fed Baltic Sea herring. Aquat. Toxicol. 1996, 34, 71–84. (2) Reijnders, P. J. H. Reproductive failure in common seals feeding on fish from polluted coastal waters. Nature (London) 1986, 324, 456–457.
(3) Tanabe, S. Contamination and toxic effects of persistent endocrine disrupters in marine mammals and birds. Mar. Pollut. Bul. 2002, 45 (1-12), 69–77. (4) Pomeroy, P. P.; Green, N.; Hall, A. J.; Walton, M.; Jones, K.; Harwood, J. Congener-specific exposure of grey seal (Halichoerus grypus) pups to chlorinated biphenyls during lactation. Can. J. Fish Aquat. Sci. 1996, 53, 1526–1534. (5) Pomeroy, P. P.; Fedak, M. A.; Rothery, P.; Anderson, S. Consequences of maternal size for reproductive expenditure and pupping success of grey seals at North Rona, Scotland. J. Anim. Ecol. 1999, 68, 235–253. (6) Fedak, M. A.; Anderson, S. S. The energetics of lactation: accurate measurements from a large wild mammal, the grey seal (Halichoerus grypus). J. Zool. (London) 1982, 198, 473–479. (7) Thomas, G. O.; Moss, S. E. W.; Asplund, L.; Hall, A. J. Absorption of decabromodiphenyl ether and other organohalogen chemicals by grey seals (Halichoerus grypus). Environ. Pollut. 2005, 133, 581–586. (8) Burns, J. M.; Lestyk, K. C.; Folkow, L. P.; Hammill, M. O.; Blix, A. S. Size and distribution of oxygen stores in harp and hooded seals from birth to maturity. J. Comp. Physiol., B 2007, 177, 687–700. (9) Hall, A.; McConnell, B.; Barker, R. Factors affecting first-year survival in grey seals and their implications for life history strategy. J. Anim. Ecol. 2001, 70, 138–149. (10) Hall, A. J.; McConnell, B. J.; Barker, R. J. The effect of total immunoglobulin levels, mass and condition on the first-year survival of grey seal pups. Funct. Ecol. 2002, 16, 462–474. (11) Barker, R. J. Joint modelling of live-recapture, tag-resight and tag-recovery data. Biometrics 1997, 53, 666–677. (12) McConnell, B. J.; Beaton, R.; Bryant, E.; Hunter, C.; Lovell, P.; Hall, A. J. Phoning home - a new GSM mobile phone telemetry system to collect mark-recapture data. Mar. Mammal Sci. 2004, 20, 274–283. (13) Hall, A.; Kalantzi, O.; Thomas, G. Polybrominated diphenyl ethers (PBDEs) in grey seal pups during their first year of life - are they thyroid hormone endocrine disrupters? Environ. Pollut. 2003, 126, 29–37. (14) Kalantzi, O.; Hall, A. J.; Thomas, G. O.; Jones, K. C. Polybrominated diphenyl ethers (PBDEs) and selected organochlorines in grey seals (Halichoerus grypus) in the North Sea. Chemosphere 2005, 58, 345–354. (15) Thomas, G. O.; Sweetman, A. J.; Parker, C. A.; Kreibich, H.; Jones, K. C. Development and validation of methods for the trace determination of PCBs in biological matrices. Chemosphere 1998, 36, 2447–2459. (16) Pollock, K. H.; Alpizar-Jara, R. Classical Open-population Capture-Recapture Models. In Handbook of Capture-Recapture Analysis; Amstrup, S. C., McDonald, T. L., Manly, B. F. , Eds.; Princeton University Press: Princeton, 2005; pp 36-58. (17) White, G. C.; Burnham, K. P. Program MARK: survival estimation from populations of marked animals. Bird Study 1999, 46, 120–139. (18) Buckland, S.; Burnham, K.; Augustin, N. Model selection: an integral part of inference. Biometrics 1997, 53, 603–618. (19) Lebreton, J.-D.; Burnham, K. P.; Clobert, J.; Anderson, D. R. Modeling survival and testing biological hypotheses using marked animals: A unified approach with cast studies. Ecol. Monogr. 1992, 62, 67–118. (20) McConnell, B. J.; Fedak, M. A.; Lovell, P.; Hammond, P. S. Movements and foraging areas of grey seals in the North Sea. J. Appl. Ecol. 1999, 36, 573–590. (21) Burnham, K.; Anderson, D. Model Selection and Inference. A practical information-theoretic approach; Springer-Verlag: New York, 1998; p 353. (22) Burnham, K. P.; Anderson, D. R. Multimodel Inference. Understanding AIC and BIC in model selection. Socio. Meth. Res. 2004, 33, 261–304.
(23) Bennett, K. A. Causes of individual variation in fasting fuel allocation and consequences for diving behaviour in grey seal pups (Halichoerus grypus). PhD, University of St Andrews, St Andrews, Scotland, 2004. (24) Anderson, D. R.; Burnham, K. P. Avoiding pitfalls when using information-theoretic methods. J. Wildl. Manage. 2002, 66 (3), 912–918. (25) Spiegelhalter, D. J.; Best, N. G.; Carlin, B. R.; van der Linde, A. Bayesian measures of model complexity and fit. J. Roy. Stat. Soc. B 2002, 64, 583–616. (26) Akaike, H. Likelihood of a model and information criteria. J. Economet. 1981, 16, 3–14. (27) Chen, G.; Bunce, H. J. Polybrominated diphenyl ethers as Ah receptor agonists and antagonists. Toxicol. Sci. 2003, 76, 310– 320. (28) Abdelouahab, N.; Suvorov, A.; Pasquier, J. C.; Langlois, M. F.; Praud, J. P.; Takser, L. Thyroid Disruption by Low-Dose BDE47 in Prenatally Exposed Lambs. Neonatology 2009, 96, 120– 124. (29) Hallgren, S.; Sinjari, T.; Hakansson, H.; Darnerud, P. O. Effects of polybrominated diphenyl ethers (PBDEs) and polychlorinated biphenyls (PCBs) on thyroid hormone and vitamin A levels in rats and mice. Arch. Toxicol. 2001, 75 (4), 200–208. (30) Tabuchi, M.; Veldhoen, N.; Dangerfield, N.; Jeffries, S.; Helbing, C. C.; Ross, P. S. PCB-related alteration of thyroid hormones and thyroid hormone receptor gene expression in free-ranging harbor seals (Phoca vitulina). Environ. Health Perspect. 2006, 114, 1024–31. (31) Suvorov, A.; Girard, S.; Lachapelle, S.; Abdelouahab, N.; Sebire, G.; Takser, L. Perinatal exposure to low-dose BDE-47, an emergent environmental contaminant, causes hyperactivity in rat offspring. Neonatology 2009, 95, 203–9. (32) Ashwood, P.; Schauer, J.; Pessah, I. N.; de Water, J. V. Preliminary evidence of the in vitro effects of BDE-47 on innate immune responses in children with autism spectrum disorders. J. Neuroimmunol. 2009, 208 (1-2), 130–5. (33) Darnerud, P. O.; Thuvander, A. Studies on immunological effects of polybrominated diphenyl ethers (PBDE) and polychlorinated biphenyls (PCB) exposure in rats and mice. Organohalogen Compd. 1998, 35, 415–418. (34) Wahl, M.; Lahni, B.; Guenther, R.; Kuch, B.; Yang, L.; Straehle, U.; Strack, S.; Weiss, C. A technical mixture of 2,2′,4,4′-tetrabromo diphenyl ether (BDE47) and brominated furans triggers aryl hydrocarbon receptor (AhR) mediated gene expression and toxicity. Chemosphere 2008, 73, 209–15. (35) Eriksson, P.; Jakobsson, E.; Fredriksson, A. Brominated flame retardants: a novel class of developmental neurotoxicants in our environment. Environ. Health. Perspect. 2001, 109, 903– 8. (36) Darras, V. M. Endocrine disrupting polyhalogenated organic pollutants interfere with thyroid hormone signalling in the developing brain. Cerebellum 2008, 7 (1), 26–37. (37) Sonnenschein, C.; Soto, A. M. An updated review of environmental estrogen and androgen mimics and antagonists. J. Steroid Biochem. Mol. Biol. 1998, 65 (1-6), 143–150. (38) Wade, M. G.; Foster, W. G.; Younglai, E. V.; McMahon, A.; Leingartner, K.; Yagminas, A.; Blakey, D.; Fournier, M.; Desaulniers, D.; Hughes, C. L. Effects of subchronic exposure to a complex mixture of persistent contaminants in male rats: Systemic, immune, and reproductive effects. Toxicol. Sci. 2002, 67, 131–143. (39) Eriksson, P.; Ahlbom, J.; Fredriksson, A. Exposure to DDT during a defined period in neonatal life induces permanent changes in brain muscarinic receptors and behaviour in adult mice. Brain Res. 1992, 582, 277–81.
ES9004398
VOL. 43, NO. 16, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6369
