We examined whether brain glucocorticoid receptor (GR) modulation by polychlorinated biphenyls (PCBs) was involved in the abnormal cortisol response to stress seen in anadromous Arctic charr (Salvelinus alpinus). Fish treated with Aroclor 1254 (0, 1, 10, and 100 mg/kg body mass) were maintained for 5 mo without feeding in the winter to mimic their seasonal fasting cycle, whereas a fed group with 0 and 100 mg/kg Aroclor was maintained for comparison. Fasting elevated plasma cortisol levels and brain GR content but depressed heat shock protein 90 (hsp90) and interrenal cortisol production capacity. Exposure of fasted fish to Aroclor 1254 resulted in a dose-dependent increase in brain total PCB content. This accumulation in fish with high PCB dose was threefold higher in fasted fish compared with fed fish. PCBs depressed plasma cortisol levels but did not affect in vitro interrenal cortisol production capacity in fasted charr. At high PCB dose, the brain GR content was significantly lower in the fasted fish and this corresponded with a lower brain hsp70 and hsp90 content. The elevation of plasma cortisol levels and upregulation of brain GR content may be an important adaptation to extended fasting in anadromous Arctic charr, and this response was disrupted by PCBs. Taken together, the hypothalamus-pituitary-interrenal axis is a target for PCB impact during winter emaciation in anadromous Arctic charr.
- Salvelinus alpinus
- cortisol production
anadromous arctic charr (Salvelinus alpinus) have a very interesting life strategy as they over winter in fresh water without feeding for 9 to 10 mo a year. The annual seawater migration in the summer for ∼7 wk of feeding coincides with dramatic increase in body mass and lipid accumulation (15, 25). The charr depend on this accumulated fat reserve for energy during winter emaciation and preparation for seawater migration the following year (13, 15). However, this lifestyle also predisposes the animal to accumulation of polychlorinated biphenyls (PCBs) in lipid depots. This is particularly of concern in the Arctic because high levels of PCBs have been reported in the Arctic ecosystem (10). Indeed, as Arctic animals fast, the sequestered lipophilic pollutants are mobilized from lipid depots to extra-adipose tissues, including brain and liver (7, 14). Several studies have reported adverse effects by PCBs on health status of Arctic animals (10), including abnormal functioning of the hypothalamus-pituitary-interrenal (HPI) axis in Arctic charr (16).
The activation of the HPI axis is a well-characterized response to stressors and culminates in the elevation of plasma cortisol levels in fish (4). The steroidogenic cells (i.e., interrenal cells) located predominantly in the anterior part of the kidney (head kidney) produce cortisol in response to stimulation by ACTH, the major corticosteroid secretogogue released from the fish pituitary gland (6, 22). The stress-induced elevation of plasma cortisol levels is tightly regulated, and the levels usually fall to resting values over a 24-h period, although the type and intensity of the stressor may alter this profile. This tight control is in part maintained by a negative feedback loop, including GR signaling in the brain that inhibits the release of the trophic hormone, corticotropin releasing factor, in response to elevated cortisol levels (6, 22).
This cortisol response is thought to play a key role in the metabolic and ionic adjustments necessary for coping with stress (22). Indeed, exogenous cortisol administration and/or endogenous elevation of cortisol by stressor manipulations affected intermediary metabolism and ion regulation (20, 22). The metabolic response includes enhanced liver capacity for amino acid catabolism and gluconeogenesis, leading to hyperglycemia, which provides a key source of fuel for coping with increased energy demand, including ion regulation (4, 22). Consequently, disruption of the HPI axis may seriously hamper the adaptive response that allows animals to cope with stress. Recently we showed that PCBs impact the plasma cortisol response to a handling disturbance in Arctic charr (16). Specifically, we showed that high levels of PCBs delayed the cortisol response to stress compared with the control fish. Also, these PCB-exposed fish lacked the ability to regulate their plasma cortisol levels, which kept increasing over time, unlike the control group, where the levels dropped significantly by 24 h posthandling disturbance (16). On the basis of these results, we hypothesized that PCB impacts the negative feedback regulation of cortisol in fish. As GR signaling in the brain is a key component of the negative feedback loop, we tested the above hypothesis by examining whether PCB affects brain GR content in anadromous Arctic charr. We also measured the corticosteroidogenic capacity of the interrenal tissue to shed light on the possible impact of PCBs on this important axis in Arctic charr.
MATERIALS AND METHODS
Porcine ACTH-(1–39), L-15 media, protease inhibitor cocktail, and 2-phenoxyethanol were obtained from Sigma (St. Louis, MO). Bicinchoninic acid (BCA) reagent was from Pierce Chemical. Multiwell (24-well plate) tissue culture plates were obtained from Falcon (Becton Dickinson Labware). All the electrophoresis reagents and molecular weight markers were from BioRad. Antibody to trout GR was developed in our laboratory (26), whereas hsp90-monoclonal rat anti-human hsp90 antibody was obtained from StressGen (Canada) and trout total hsp70 antibody was from Dr. Peter Candido (Biochemistry Department, University of British Columbia). The secondary antibodies were alkaline phosphatase-conjugated either goat anti-rat IgG for hsp90 (StressGen) or goat anti-rabbit IgG for GR and hsp70 (BioRad). Nitroblue tetrazolium (NBT) and 5-bromo-4-chloro-3-indolyl phosphate p-toluidine salt (BCIP) were obtained from Fisher Scientific (Ontario).
Experimental fish were fourth generation offsprings from anadromous Arctic charr collected at Svalbard (79°N) in 1990. The fish were hatched in 1998 at Karvik Research Station, Tromsø, Norway. The fish had been held under natural water temperatures (0.5°C) and winter light conditions (10:14-h light/dark cycle) before the experiment and fed ad libitum. At the beginning of the experiment, fish were distributed into centrally drained 500-liter capacity circular tanks supplied with a continuous flow of fresh water. The water temperature was gradually increased to 7°C over 2 wk and light conditions to a 12:12-h light/dark photoperiod, and these conditions were maintained until the end of the experiment.
At the start of the experiment in February 2000, four groups of fish were fasted and two groups were fed to satiety with pelleted commercial dry feed (Skretting, Stavanger, Norway) 5 days a week. The experimental protocol has been described in detail before (16). Briefly, the four fasted groups were force-fed Aroclor 1254 (Promochem, Ulricehamn, Sweden) at 0 (control), 1 (low PCB), 10 (medium PCB), and 100 (high PCB) mg/kg body mass, respectively, and the two fed groups were treated with 0 (control) and 100 (high PCB) mg/kg body mass of Aroclor 1254. The rationale for using Aroclor 1254 to represent PCB contamination was because this technical mixture has a congener composition very similar to that found in wild fish in the Arctic (23). Plasma cortisol concentration and the cortisol response to stress for this study were published before (16). The whole brain samples were quickly collected and frozen on dry ice and stored frozen at −70°C for GR, hsp70, and hsp90 protein analyses.
This study (February-July 2001) mimicked the first experiment, except we only had one fed control group but no high-dose Aroclor 1254 (100 mg/kg body mass)-treated fed group. In both experiments, the fed group was only used for comparison because feeding was artificial for these animals, because, in the wild, they normally fast during the time these experiments were conducted. At the end of the experimental period, fish were killed with a high dose of benzocaine (300 ppm), and blood samples were collected with Li-heparinzed vacutainers. The plasma was separated by centrifugation (6,000 g, 10 min) and stored at −70°C for later analyses of plasma cortisol concentrations. After blood collection, head kidney was dissected and kept in petri dishes with L-15 media for in vitro ACTH challenge test.
In Vitro Cortisol Production
Head kidney tissue from treatment groups was challenged in vitro with porcine ACTH-(1–39), and cortisol production rate was measured using well-established static incubation protocol described previously (2, 33). Briefly, head kidney was excised immediately, rinsed in ice-cold L-15 medium, and finely minced into ∼1-mm3 cubes. The incubation consisted of distributing tissues from each fish equally into two wells (each part incubated either with or without ACTH; 6 fish × 2 wells = 12 wells) in a 24-well Falcon tissue culture plate. Tissue pieces were allowed to incubate with gentle shaking for 2 h at 7°C (equilibration period), after which the supernatant was replaced with fresh L-15 media. The tissue was incubated for 1 h, and the media were collected and stored for determination of cortisol concentration (basal cortisol production rate). The media were then replaced with either fresh media alone (control) or fresh media containing ACTH (0.5 IU/ml) and incubated for an additional 2 h, after which the supernatant was frozen for later determination of cortisol concentration (stimulated cortisol production rate). Preliminary time course and dose response were conducted to establish the concentration of ACTH required for eliciting a maximal response in Arctic charr head kidney tissues. Wet weight of the tissue in each well was recorded and cortisol production rate was expressed as nanograms per hour milligram wet weight.
Brain PCB Analysis
The whole brain was removed and frozen for later PCB analyses (see below). The brains from the 10 fish sampled from each group were divided into three pooled samples representing 3, 3, and 4 fish. The pooled samples were homogenized and extracted with acetone followed by hexane-acetone (3:1). After clean-up, PCB was determined by gas chromatography using an HRGC 5890 Series II (Fision Instruments, Milano, Italy) equipped with a spitless injector, electron capture detector (ECD), and a high-performance HP-5 capillary column (50 m × 0.20 μm × 0.11 μm; Hewlett-Packard, Folsom, CA). Nitrogen was used as carrier gas (42 ml/min). The injector and detector temperatures were 280°C and 320°C. The temperature and gradient were 60°C (1 min), 15°C/min; 160°C (0 min), 1.5°C/min; 270°C (20 min). Calculations were done on the basis of external (PCB 53) and internal standards to which the chromatograph had been calibrated. The analyzed congeners (IUPAC numbers), including percent recovery were: PCB 28–50%, PCB 31–51%, PCB 52–67%, PCB 99–54%, PCB 101–50%, PCB 105–69%, PCB 118–53%, PCB 138–63%, PCB 153–57%, PCB 156–80%, and PCB 180–72%. The PCB data were adjusted in accordance with the recovery values.
Plasma samples and L-15 media from in vitro ACTH-stimulated cortisol production samples were ether extracted before determining the plasma cortisol levels (16). Plasma cortisol concentrations were estimated using commercially available ImmuChem 125I-RIA kit (ICN Biochemicals) according to established protocols (32).
SDS-PAGE and Western Blotting
The brain GR, hsp90, and hsp70 protein expression were determined using SDS-PAGE followed by immunodetection as explained below. The tissue protein concentration was determined using the bicinchoninic acid (BCA) method using bovine serum albumin (BSA) as the standard. The procedure for SDS-PAGE and Western blotting were according to established protocols (8). Briefly, samples (40 μg protein/sample) were separated on 8% polyacrylamide gels using discontinuous buffer system (19). The gels were transferred onto nitrocellulose membrane (20 V for 30 min) with a semidry transfer unit (BioRad) using transfer buffer [25 mM Tris pH 8.3, 192 mM glycine, and 20% (vol/vol) methanol] and were blocked with 5% skimmed milk in TBS-t [20 mM Tris pH 7.5, 300 mM NaCl, and 0.1% (vol/vol) Tween 20 with 0.02% sodium azide] for 60 min. Primary and secondary antibodies were diluted in the blocking solution to appropriate concentrations as indicated. For GR and hsp70, polyclonal trout-specific antibodies (8, 28) were used at 1:1,000 and 1:3,000 dilutions respectively, whereas for hsp90, a monoclonal rat anti-human hsp90 antibody was used at 1:3,000 dilution. The secondary antibody was alkaline phosphatase-conjugated either goat anti-rabbit (GR and hsp70: 1: 3,000 dilution) or anti-rat (hsp90: 1:3,000 dilution) IgG. The membranes were incubated in primary antibody for 60 min at room temperature, washed with TBS-t (2 × 5 min), incubated with secondary antibody for 60 min and finally washed with TBS-t (1 × 15 min). Visualization of bands was carried out with NBT (0.033% wt/vol) and BCIP (0.017% wt/vol). The molecular mass was visualized using prestained low-range molecular weight markers (112 kDa phosphorylase B, 81 kDa bovine serum albumin, 49.9 kDa ovalbumin, 36.2 kDa carbonic anhydrase, 29.9 kDa soybean trypsin inhibitor, 21.3 kDa lysozyme). Quantification of bands was done with Chemi Imager using the AlphaEase software (Alpha Innotech).
All statistical analyses were performed with SPSS version 10.0 (SPSS, Chicago, IL), and data were expressed as means + SE. The data were transformed (logarithmic), wherever necessary, for homogeneity of variance, but nontransformed values are shown in the figures. One-way ANOVA was used to assess the changes in plasma cortisol levels, ACTH-stimulated cortisol production, GR protein expression, and brain total PCB content in fasted group exposed to different doses of Aroclor 1254. Student's t-test was employed to assess the differences between sham and high dose of Aroclor in the fed group as well as the fed and fasted control groups. A post hoc (least significant difference) test was used for pair-wise comparison wherever significant differences were observed. A probability level of P ≤ 0.05 was considered significant.
Accumulation of PCB content in the brain of fasted fish showed a significant dose-response relationship with Aroclor 1254 treatment. The PCB content in the brain was ∼10-fold greater in the medium dose group compared with the low-dose group and sixfold greater in the high-dose group compared with the medium-dose group, respectively (Fig. 1A). Among fed fish, the high-dose PCB group had significantly higher PCB content compared with the fed control. However, the brain PCB accumulation with high Aroclor treatment in the fasted fish was at least threefold higher than that seen in the fed fish treated the same way (Fig. 1, A and B).
Fasting significantly increased brain GR content in Arctic charr compared with the fed fish (Fig. 2, A and B). However, high PCB dose, but not medium and low PCB doses, significantly decreased (75%) brain GR content in the fasted fish compared with other treatments (Fig. 2C). No differences in GR protein expression were observed between PCB-treated and control fed groups (Fig. 2D).
The brain hsp90 expression was significantly decreased with fasting in Arctic charr compared with fed fish (Fig. 3, A and B). Treatment of fasted fish with low and medium dose of Aroclor 1254 had no significant effect on brain hsp90, but high dose of PCB significantly lowered brain hsp90 expression (Fig. 3C). This decrease in hsp90 content with high PCB was not seen in the fed fish (Fig. 3D). Fasting did not significantly effect brain hsp70 expression in Arctic charr (Fig. 4, A and B). PCB exposure resulted in a significantly lower hsp70 expression in the low and high PCB group, but not the medium group, among fasted fish (Fig. 4C). High PCB had no significant impact on hsp70 expression in the brain of fed fish (Fig. 4D).
The objective of this experiment was to investigate whether the altered plasma cortisol levels seen in the PCB-treated fish [plasma cortisol levels for the fish used in experiment 1 were presented in Jorgensen et al. (16)] were due to altered interrenal secretory capacity in Arctic charr. As occurred in the previous experiment (16), plasma cortisol levels were elevated significantly in the fasted, control group compared with the corresponding fed group (Fig. 5). Among fasted fish, all three PCB doses significantly decreased plasma cortisol levels compared with the controls, whereas medium and high-dose PCB treatments also resulted in significantly lower cortisol levels compared with the low-dose PCB group (Fig. 5).
Fasting per se significantly lowered in vitro ACTH-induced cortisol production by interrenal cells compared with fed, control group (Fig. 6A). However, PCB treatment did not significantly affect ACTH-induced cortisol production in the fasted Arctic charr (Fig. 6B).
We have shown for the first time that PCB disrupts brain GR content in fasted Arctic charr. The decrease in brain GR content with high PCB dose corresponded with an abnormal cortisol profile in response to stress in these fish (16), leading us to propose that HPI axis is a target for PCB impact in anadromous Arctic charr. As this response was evident only in the fasted fish, it seems likely that the Arctic life strategy of this species, especially natural fasting for extended periods, predisposes these animals to neurotoxicity by PCBs. This is further underscored by the high PCB levels in the Arctic environment due to atmospheric transport and other routes of deposition, resulting in “hot spots” based on PCB body burdens in animals, including Arctic charr (10, 11).
Plasma Cortisol Regulation in Fasted Arctic Charr
The extended fasting of fish in the present study mimicked the natural fasting experienced by anadromous Arctic charr during their overwintering in fresh water (15). Although feeding is unnatural during this period in anadromous Arctic charr, our fed group did provide critical insight into the adaptive strategies, especially with respect to cortisol regulation, associated with winter emaciation in this species. Cortisol, the primary glucocorticoid in bony fishes, is an important metabolic hormone playing a key role in the energy substrate repartitioning process (22). This metabolic role of cortisol would be of particular importance with long-term fasting in anadromous Arctic charr, especially because these animals usually fast for 9–10 mo a year before their seawater migration (15). The higher plasma cortisol levels seen in the fasted fish compared with the fed fish in our companion study (16) as well as this study (Fig. 5) clearly attest to a role for this hormone in the longer-term metabolic adjustments. This notion is further supported by the elevated plasma glucose levels in the fasted fish compared with the fed fish (16), pointing to increased cortisol stimulation, especially because one of the well-established roles for this steroid is enhanced gluconeogenesis (22, 32).
The changes in endocrine and metabolic parameters in these naturally fasting animals, important for maintaining stability in response to unfavorable conditions, support the concept of allostasis (21). The maintenance of elevated cortisol levels, albeit at a mild stress level seen in salmonids (4), for extended periods with fasting may involve a resetting of the cortisol feedback loop, including HPI axis adjustments. The depression in ACTH-induced cortisol production capacity by fasted Arctic charr clearly argues against increased hormone production as a reason for the elevated plasma cortisol levels. Our results are in accordance with other studies that showed similar attenuation of in vitro ACTH-induced cortisol production even with short-term fasting in rainbow trout (Oncorhynchus mykiss) (2) and brook charr (Salvelinus fontinalis) (30). This was also true in vivo as evidenced by the attenuated cortisol response to a handling disturbance in the fasted Arctic charr compared with the fed charr (16). Together, these results point to a fasting-induced downregulation of cortisol production capacity. Although the mechanism(s) is not clear, our recent finding that 6 days of fasting in trout depressed ACTH-, but not cAMP-induced cortisol production (2), argues for ACTH receptor downregulation as a possible mechanism.
The lack of any impact on the interrenal response to fasting led us to propose that adjustments to the brain-pituitary axis are a likely scenario for elevated cortisol levels in Arctic charr. The corresponding upregulation of brain GR content with fasting (Fig. 2, A and B) is interesting, especially given the fact that plasma cortisol levels were also elevated in this group. This is mainly because previous studies showed a downregulation of GR content with elevated plasma cortisol levels in teleosts (22, 28, 32). As far as we are aware, this is the first study to examine the effect of fasting on brain GR content in fish. Our results implicate a role for plasma cortisol levels in upregulating their own receptors in the brain of fasted charr, which may be an important adaptation to maintain HPI axis stability during winter emaciation in this species. Although the precise mechanism for the receptor upregulation is not known, the lower hsp90 content in the fasted brain (Fig. 3, A and B), an important molecular chaperone involved in this steroid receptor signaling and stability (28), implies a possible link with GR regulation. Also, the lack of any change in total hsp70 content (Fig. 4, A and B), an indicator of cellular stress, alludes to a fasting-specific hsp90 response that was independent of the overall cellular stress status.
PCB's Impact on Cortisol Response in Fasted Charr
All PCB doses clearly impacted the fasting-induced elevation of cortisol levels in Arctic charr (Fig. 5). This is particularly significant given the fact that the PCB tissue burden reported here is not very different from the levels reported previously for feral Arctic charr (3, 29). As an elevated cortisol response may be crucial for the energy repartitioning process, our results suggest a metabolic impact of PCB, which in turn will affect the homeostatic process essential for Arctic charr to cope with the extended fasting. The attenuation of plasma cortisol levels did not correspond to a lack of interrenal capacity for cortisol production (Fig. 6B), clearly pointing to other control points, including a decrease in ACTH release and/or increased clearance of the steroid from circulation. As brain GR signaling is thought to be crucial for the negative feedback loop, the lack of any significant effect of low and medium PCB dose on brain GR content argues against altered ACTH secretion as the reason for the lower cortisol levels in these two groups. However, one cannot rule out the possibility that PCB affects the adrenocorticotropic cells, independent of brain stimulus, as was shown before for the yellow perch (Perca fluviatilis), resulting in lower cortisol levels to stress (12). In the present study, cortisol response to handling disturbance in the low and medium PCB groups was similar to that of the control-fasted fish (16), arguing against any impact on the ACTH secretion pathway in those two groups. This raises the possibility that the clearance of the hormone may be enhanced by PCBs, resulting in lower circulating cortisol levels. Indeed, previous studies did suggest a possible role for arylhydrocarbon receptor-induced CYP1A activation as a mechanism for increased steroid uptake and catabolism by the liver (31). Even in the present study, liver CYP1A expression showed a dose-dependent increase with PCB treatment in the fasted fish (data not shown) implying an enhanced potential for cortisol clearance.
Whereas the low and medium dose of PCB did not affect the cortisol response to stress, the high dose PCB group did show an abnormal cortisol response to a standardized handling stress (16). This was characterized by a delayed elevation in plasma cortisol levels that kept increasing with time, unlike the other groups where the levels peaked, but dropped significantly by 24 h posthandling disturbance (16). This lack of a negative feedback regulation in the high PCB group also corresponded with a lower brain GR content, which would limit CRF due to decreased brain responsiveness to cortisol stimulation (27). The lower brain GR content in the high PCB group, but not in the low and medium groups, also correlated positively with brain hsp90 and hsp70 content, suggesting an overall reduction in proteins that are critical in the cellular stress response process in fish. As this response was evident only in the fasted fish given high PCB concentration and correlated with the highest PCB burden in the brain relative to all other treatments (including high-dose PCB, fed fish), our results suggest that the brain PCB concentration exceeded a threshold for neurotoxicity only in the fasted fish.
The threefold higher PCB content in the brain of fasted fish compared with the fed fish is not surprising given the fact that fasted Arctic charr mobilize lipid depots for energy, resulting in the redistribution of PCBs to extra-adipose tissues, specifically liver and brain (14, 17). Consequently, the anadromous lifestyle predisposes these animals to PCB-mediated neurotoxicity. This is further reflected by their abnormal HPI axis response to a standardized stress in the high PCB-treated charr (16). As GR is crucial for the negative feedback regulation, the lower brain GR, hsp90, and hsp70 expression may point to a loss of neurons, especially those involved in the HPI axis, in response to high PCB exposure in Arctic charr. In support of this argument, downregulation of brain GR content due to chronic stress has been linked to neurotoxicity, including the loss of neurons containing GR (27). Also, PCB-mediated neurotoxicity is associated with neuronal loss in several animals (9, 26), including fish (18).
In conclusion, maintaining elevated plasma cortisol levels and upregulation of their receptors may be an important adaptation for longer-term repartitioning of energy resources to cope with winter emaciation in anadromous Arctic charr. PCBs lowered plasma cortisol levels in Arctic charr, and this may lead to impairment of the energy repartitioning process that is crucial for coping with extended seasonal fasting in this species. In addition, high dose of PCB appears to result in neuronal loss, as reflected by a decrease in GR, hsp90, and hsp70 expression, leading to the disruption of the negative feedback regulation of cortisol during stress (16). Overall, winter emaciation that is part of the natural life history strategy of anadromous charr, coupled with the high PCB deposition in the Arctic, makes these animals vulnerable to the neurotoxic effects of PCBs. As the hormonal response to fasting in Arctic charr is not solely limited to the HPI axis, but also involves other hormones, including the somatotropic axis (1), the impact of PCBs on this Northernmost living salmonid may be far reaching and requires more study.
This study was funded under Project No. OPP-9908890 from the Office of Polar Programs, National Science Foundation; the Norwegian Research Council, Project No. 151462/720; and the Natural Sciences and Engineering Research Council, Canada, Discovery Grant.
Thanks are extended to Roula Raptis, University of Waterloo, and Judith Wolkers, University of Tromsø, for technical assistance.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2004 the American Physiological Society