Selenium (Se) is a naturally occurring, essential trace element. Se-containing proteins, known as selenoproteins, play an important role in thyroid hormone production and the prevention of oxidative stress (Kohrle et al. 2005; Janz et al. 2010). However, high levels of Se can lead to toxicity (Hopkins et al. 1997, 1999a; Finger et al. 2017b; Haskins et al. 2017b, c). Mechanistically, Se toxicity is thought to be mediated through its substitution for sulfur in amino acids (and consequent disruption of protein structure) and/or its induction of oxidative stress (Janz et al. 2010; Spallholz and Hoffman 2002). Selenium toxicity can manifest in reproductive impairment (Roe et al. 2004), mortality (Finger et al. 2017b), alterations in metabolic rate (Hopkins et al. 1999b; Rowe et al. 2001), and/or changes in circulating levels of stress hormones (Hopkins et al. 1997; Patterson et al. 2017).

There is an increasing risk of Se exposure worldwide due to anthropogenic activities, such as smelting, the combustion of fossil fuels for energy production, and mining (Lemly 2004; Lemly and Skorupa 2012). Coal combustion wastes (CCWs) are a byproduct of the incomplete combustion of coal and often enriched with Se (Rowe et al. 2002). Almost 20% of CCWs are deposited into aquatic settling basins (Lemly and Skorupa 2012). These basins serve as habitats for a number of organisms (Hopkins et al. 1997, 1999a, b; Rowe et al. 2002; Roe et al. 2004; Haskins et al. 2017a), heightening their risk of Se exposure.

While many studies have investigated how Se affects animals in vivo, most of these have been performed on animals that are of lower trophic status or shorter-lived (Hopkins et al. 1999a, b, 2004; Rowe et al. 2001; Haskins et al. 2017b). Because the American alligator (Alligator mississippiensis) is a long-lived, apex predator found throughout aquatic habitats in the southeastern United States, it has been suggested to serve as an indicator of environmental quality (Finger and Gogal 2013). Alligators are known to inhabit Se-contaminated areas, including ash settling basins (see Roe et al. 2004; personal observation). Alligators are capable of accumulating high levels of Se when fed prey contaminated with Se (Tuberville et al. 2016; Finger et al. 2017b). Nonetheless, few studies have investigated the effects of Se exposure and accumulation on alligator physiology (Roe et al. 2004; Finger et al. 2016).

This study was designed to determine how the consumption of prey spiked with selenomethionine (SeMet), the most bioavailable form of Se, affects oxidative status and the stress response of alligators to provide better insight into how Se accumulation affects their physiology. Baseline (i.e., unstressed levels, initially after capture) corticosterone (CORT), the main crocodilian stress hormone (Finger et al. 2015a), levels were measured in plasma of alligators exposed to one of three SeMet dietary treatments. In addition, parameters of oxidative status were measured in whole blood and tail muscle, including levels of (1) 4-hydroxynonenal (4-HNE), a lipid peroxidation byproduct; (2) levels of the antioxidant enzymes superoxide dismutase-1 (SOD1), SOD2, and glutathione peroxidase-1 (GPX1); and (3) levels of nuclear factor (erythroid-derived 2)-like 2 (Nrf2), an antioxidant-associated transcription factor that regulates the expression of multiple antioxidants (Ma 2013). Because previous studies in other species have shown that Se exposure can cause oxidative stress and affect the stress response (Hopkins et al. 1997; Patterson et al. 2017), we hypothesized that SeMet exposed alligators would exhibit higher baseline CORT levels and elevated parameters of oxidative status.

Methods

Animal Husbandry

The details of alligator husbandry employed herein have been previously described by Finger et al. (2017b). Briefly, 24 sexually immature alligators (average ± standard error, head length: 13.12 ± 0.16 cm; total length: 104.35 ± 1.73 cm), originally obtained from Rockefeller Wildlife Refuge in Grand Chenier, LA, were randomly allocated to one of three pens (n = 8) inside a climate-controlled facility (22.7 °C) with a translucent fiberglass panel roof on the Savannah River site (SRS) near Jackson, SC (Finger et al. 2015b; Hamilton et al. 2016a). Unheated (ambient temperature) water was continuously filtered throughout each pen and alligators were maintained on natural circadian rhythms through light filtration. Alligators were allowed to habituate to pens for > 5 months and fed Mazuri Crocodilian Diet pellets three times weekly until satiation before commencing treatment with SeMet. During the SeMet dosing experiment, alligators were fed pellets twice per week.

Selenomethionine Dietary Treatment

In-depth treatment details employed for dosing alligators were previously described by Finger et al. (2017b). Briefly, alligators were fed thawed, dead, small, “fuzzy” mice spiked with 1000 or 2000 ppm seleno-d,l-methionine (µg SeMet/g mouse dry weight; Sigma S3875 St. Louis, MO) or DI water (control treatment) via oral gavage once per week for 7 weeks (October–December 2014). Each pen comprised a different treatment group (Pen 1, 1000 ppm SeMet; Pen 2, 2000 ppm; Pen 3, control treatment). Alligators that were fed SeMet accumulated significantly more Se (all values in dry mass) in their kidneys (1000 ppm SeMet: 101.60 ± 8.63 ppm Se; 2000 ppm SeMet: 96.38 ± 5.81 ppm Se) and livers (1000 ppm SeMet: 35.20 ± 6.35 ppm Se; 2000 ppm SeMet: 49.97 ± 4.00 ppm Se) than control (kidney Se: 6.51 ± 0.22 ppm Se; liver Se: 2.22 ± 0.14 ppm Se) alligators (Finger et al. 2017b).

Blood Sampling and CORT Quantification

Alligators were captured and blood was sampled from the occipital sinus (with a 25-gauge, 2.54-cm nonheparinized needle and 3-mL syringe) within 3 min of capture 1 week after the last SeMet treatment at 1000 h Eastern Standard Time (December 1, 2014) to determine baseline (i.e., initial levels after capture) CORT levels (Hamilton et al. 2016a). Before entering the building housing the alligators, a timer was started to monitor the cumulative time required to capture an individual and sample it (CumTime, range: 92–708 s; Finger et al. 2015a) and the cumulative time required to bleed an individual after its capture (HandlingTime, range: 36–132 s). Within 1 h of collection, blood samples were centrifuged 3 min at 1640×g and plasma was stored at − 60 °C (Hamilton et al. 2016b).

Plasma CORT was extracted in a mixture of ethyl-acetate:hexane as described previously (Hamilton 2016; Lance and Elsey 1999; Lance et al. 2004). After extraction, all samples were analyzed with an enzyme immunoassay (EIA; ADI-900-097 Enzo, Farmingdale, NY) according to the manufacturer’s protocol. American alligator plasma samples have been previously optimized and validated for use with the specific EIA used in this study (Hamilton 2016). Intra-assay variation was 7.15%. The limit of detection (LOD; 2.40 pg/mL) for the kit was determined as described previously (Wada et al. 2007) by the following equation: \({\text{LOD}} = {\text{mean}}\, {\text{blank}}\, {\text{optical}}\, {\text{density}} + \left( {2 \times {\text{standard}}\, {\text{deviation}}\, {\text{of}}\, {\text{blank}}} \right)\). Any results that fell below the LOD for the kit were assigned this concentration.

Euthanasia and Western Blotting

Three-to-four days after this blood sampling (4–5 December 2014), all alligators were euthanized as described previously (Finger et al. 2017b). At euthanasia, whole trunk blood and tail muscle samples were obtained and stored at − 60 °C until eventual analysis of protein levels by Western blotting at the School of Kinesiology at Auburn University. Because unexpected mortalities prevented obtaining fresh tissue samples from all alligators (Finger et al. 2017b), we only investigated the oxidative status of individuals that were euthanized within 5 min of capture. Thus, Western blots were only used to analyze tissue samples from 12 alligators (i.e., 4 alligators/treatment).

Levels of 4-HNE (ab46545; Abcam, Cambridge, MA), Nrf2 (GTX103322; GeneTex, Irvine, CA), SOD1 (GTX100554; GeneTex), SOD2 (GTX116093; GeneTex), and GPX1 (GTX116040; GeneTex) in muscle and whole blood (Hill et al. 2013) were determined by Western blot as described previously (Hyatt et al. 2016; Finger et al. 2017b). Each membrane was stained with Ponceau as both a loading and transfer control. Proteins were visualized with a chemiluminescent system (GE Healthcare Life Sciences, Pittsburgh, PA), and captured images were analyzed with a ChemiDocIt Imaging System (UVP, LLC, Upland, CA).

Statistical Analysis

All statistical analyses were performed in JMP Pro 14. Linear regression was used in all analyses. Dietary treatment was used as a factor to investigate the effects of SeMet administration on plasma CORT and oxidative status. CumTime and/or HandlingTime were included as covariates (when significant) to account for disturbance/capture-associated effects on oxidative status or CORT. Post hoc multiple comparisons were made using t tests.

Unstandardized effect sizes, including regression coefficients (RC) ± standard errors (SE) and simple effect sizes (i.e., group mean differences) ± SE are presented below to indicate effect size (Finger et al. 2017b). When means are presented, these are indicated explicitly. Western blot results are in arbitrary units (AU). Significance was set at α = 0.05.

Results

Plasma CORT

Plasma CORT was significantly affected by dietary treatment (F2,17 = 9.12; p = 0.0020; Fig. 1). Both 1000 ppm (p = 0.0356; simple effect size: 5.39 ± 2.36 ng/mL) and 2000 ppm (p = 0.0005; simple effect size: 10.99 ± 2.58 ng/mL) SeMet alligators had significantly higher plasma CORT than control alligators. CORT levels also were significantly higher in alligators fed 2000 ppm SeMet than those fed 1000 ppm SeMet alligators (p = 0.0392; simple effect size: 5.59 ± 2.50 ng/mL).

Fig. 1
figure 1

The effect of Dietary Treatment on baseline plasma corticosterone levels. Bars are raw means (1 ± SE) of respective treatments. Plasm CORT levels are reported in ng/mL

Full size image

Both HandlingTime (p = 0.0082, RC: 0.11 ± 0.04 ng/mL) and CumTime (p = 0.0012, RC: 0.03 ± 0.01 ng/mL) significantly affected plasma CORT when they were added individually to the Dietary Treatment model. However, when both were included together (along with Dietary Treatment) in the same model, they both had only suggestive effects on plasma CORT (HandlingTime: p = 0.0663, RC: 0.08 ± 0.04; CumTime: p = 0.103, RC: 0.02 ± 0.01). Because HandlingTime had a greater effect on CORT levels (in combined and individual models), it was included in the final Dietary Treatment model.

Simple regression analysis (independent of Dietary Treatment) revealed that neither CumTime (p = 0.85) nor HandlingTime (p = 0.1587) affected plasma CORT. Therefore, we examined their respective effects on each treatment independently. Plasma CORT levels of control alligators and alligators fed 1000 ppm SeMet were not affected by CumTime (Control, p = 0.469; 1000 ppm SeMet, p = 0.2512) or HandlingTime (Control, p = 0.1304; 1000 ppm SeMet, p = 0.624). In alligators fed 2000 ppm SeMet, both HandlingTime (p = 0.0238, RC: 0.24 ± 0.06 ng/mL) and CumTime (p = 0.0013, RC: 0.11 ± 0.01 ng/mL) significantly affected plasma CORT.

Oxidative Status

Dietary Treatment had no effect on any muscle parameter investigated (SOD1, p = 0.6436; SOD2, p = 0.2755; 4-HNE, p = 0.6638; Nrf2, p = 0.3745; GPX1, p = 0.3501; Table 1). Neither HandlingTime nor CumTime had any effects on muscle parameters (all p > 0.05).

Table 1 Mean (± 1 SE) treatment differences in levels of oxidative parameters in tail muscle and whole blood
Full size table

SOD2 levels were the only blood parameter significantly affected by Dietary Treatment (F2,9 = 7.15, p = 0.0138; Table 1). SOD2 levels were higher in alligators fed 1000 (p = 0.0499; simple effect size: 0.54 ± 0.24 AU) and 2000 ppm SeMet (p = 0.0045; simple effect size: 0.88 ± 0.24 AU) than control alligators. No difference in blood SOD2 levels was observed between those fed 1000 or 2000 ppm SeMet (p = 0.1701). Blood Nrf2 (p = 0.8786), GPX1 (p = 0.7055), 4-HNE (p = 0.1036), and SOD1 (p = 0.2658) levels were unaffected by Dietary Treatment (Table 1). No effects of CumTime and HandlingTime were observed on blood parameters.

Discussion

This is the first study to investigate how dietary ingestion of Se affects stress hormone levels (i.e., CORT) and oxidative status in the American alligator, a long-lived, top-trophic carnivore. Alligators were fed SeMet-contaminated prey over a 7-week period. At the end of this period, blood and tissue samples were obtained to measure plasma CORT and oxidative status. The current study demonstrates that although SeMet treatment resulted in increased CORT levels as expected, SeMet treatment inconsistently affected oxidative status.

Plasma CORT

As important mediators of the stress response, plasma glucocorticoids rapidly increase above their baseline level in response to a stressor (Sapolsky et al. 2000). After an acute or transient stressor, negative feedback mechanisms are usually activated that eventually return plasma glucocorticoids back to baseline levels. However, a chronic or persistent stressor, such as toxicant exposure, can result in prolonged elevation of glucocorticoids (Gunderson et al. 2003; Patterson et al. 2017). As such, baseline levels of plasma glucocorticoids often are measured to assess the effects of a chronic stressor on an animal (Guillette et al. 1997; Lance and Elsey 1999; Finger et al. 2015a).

Alligators treated with 1000 and 2000 ppm SeMet had significantly higher baseline (i.e., unstressed, initial levels) plasma CORT levels than control alligators (Fig. 1), supporting our original hypothesis. Mean baseline plasma CORT levels of control alligators (0.62 ng/mL; range: 0.28–1.34 ng/mL) were low, equivalent to levels previously reported in both juvenile and adult alligators (Elsey et al. 1990a, b; Guillette et al. 1997). In contrast, alligators treated with SeMet had levels of baseline plasma CORT that were as high or higher than previously reported stress-induced levels (Lance and Elsey 1999; Gunderson et al. 2003). The mean plasma CORT level (9.66 ± 4.55 ng/mL) of alligators fed 2000 ppm SeMet was comparable to levels found in juvenile alligators housed at high stocking densities for 3.5 months (Elsey et al. 1990a), subadult alligators held in a bag for 2 h to simulate an acute stressor (Lance et al. 2004), and juvenile alligators subjected to restraint for 4 h (Lance and Elsey 1999). Alligators fed 1000 ppm SeMet (3.00 ± 1.27 ng/mL) had mean plasma CORT levels that were higher than those found in captive adult alligators housed at a high stocking density (Elsey et al. 1990b).

Similar to our results, Se exposure is associated with higher plasma glucocorticoid levels in some amphibians and fish (Hopkins et al. 1997, 1999a; Miller et al. 2007, 2009; Thomas and Janz 2011; Wiseman et al. 2011; Patterson et al. 2017). For example, white sturgeon (Acipenser transmontanus) exposed to SeMet for 72 days had higher baseline cortisol levels (the main fish glucocorticoid) (Patterson et al. 2017). Likewise, in southern toads inhabiting a Se-contaminated ash basin, higher Se levels were associated with higher baseline CORT levels (Hopkins et al. 1997). Because higher baseline plasma CORT levels are indicative of chronic stress, our results demonstrate that chronic Se exposure is a stressor to alligators.

Previously, we observed that SeMet treatment decreased alligator body condition (Finger et al. 2017b). Specifically, alligators fed 1000 and 2000 ppm SeMet decreased in body condition over the 7-week administration period, whereas the body condition of control alligators did not change. Combined, our results presented here suggest that SeMet may exert a negative effect on alligator body condition through increasing CORT levels.

Oxidative Status

Selenium is thought to induce oxidative stress through increasing reactive oxygen species (ROS) and decreasing levels of the antioxidant glutathione (Spallholz and Hoffman 2002). However, because Se displays nutritional hormesis, its effects on oxidative stress are dose dependent. In fact, Se deficiency can lead to oxidative stress, whereas Se supplementation may prevent oxidative stress through increasing levels or the activity of antioxidant enzymes (Elsey and Lance 1983; Cheng et al. 1999; Monteiro et al. 2009).

Alligators treated with SeMet had higher blood SOD2 levels than control alligators (Table 1). Higher blood SOD2 levels suggest that SeMet treatment altered alligator oxidative status and provides support for our original hypothesis. Se exposure has been shown to increase superoxide anion, a potent ROS, generation (Spallholz and Hoffman 2002). SODs detoxify superoxide by reducing it to hydrogen peroxide (H2O2; Fukai and Ushio-Fukai 2011). Interestingly, SOD1 and muscle SOD2 levels were unaffected by SeMet treatment. These null effects on SOD1 and muscle SOD2 may be a consequence of differences in tissue expression or due to the intracellular location of SODs; SOD1 is mainly found in the cytoplasm, whereas SOD2 is mainly localized in the mitochondrial matrix (Fukai and Ushio-Fukai 2011).

In contrast to our original hypothesis, levels of 4-HNE (a lipid peroxidation byproduct) were similar among control and SeMet-treated alligators (Table 1) even though alligators accumulated Se to some of the highest levels ever reported in a reptile (Finger et al. 2017b). This suggests that SeMet treatment does not increase oxidative damage in alligators. However, augmented superoxide generation (and/or glutathione depletion) induced by SeMet may have led to increased expression of certain antioxidant enzymes, such as blood SOD2, as a compensatory mechanism. In turn, this may have prevented or mitigated oxidative damage caused by SeMet exposure, masking any effects of SeMet treatment on oxidative damage (Jing et al. 2015). Regardless of the ultimate cause, similar to our results, there was no effect of chronic SeMet treatment on lipid peroxidation levels in white sturgeon (Acipenser transmontanus) and rainbow trout (Oncorhynchus mykiss) larvae (Vidal et al. 2005; Zee et al. 2016). A number of studies in birds also have failed to find a relationship between Se accumulation and lipid peroxidation (Hoffman and Heinz 1998; Ware et al. 2012; Brady et al. 2013).

Similar to 4-HNE, SeMet treatment had no effect on Nrf2 and GPX1 levels. GPX1 is a selenoprotein that reduces H2O2 and certain hydroperoxides using glutathione (Brigelius-Flohe and Maiorino 2013). Nrf2 is involved in regulating the expression of a number of antioxidant enzymes, including GPX1, in response to increased ROS and oxidative damage (Ma 2013). Because lipid peroxidation levels were not affected by SeMet treatment (i.e., damage), it is unsurprising that no changes in Nrf2 and GPX1 levels were observed. Likewise, there was no change in GPX1 levels of white sturgeon chronically treated with SeMet (Zee et al. 2016).

There are a few possible reasons why we observed no consistent effect of SeMet treatment on oxidative status. First, the oxidative effects of Se appear to be dependent on the Se species being investigated. In this study, alligators were administered SeMet, which in vitro has been shown to have no effect on superoxide generation or glutathione depletion (Stewart et al. 1999; Spallholz et al. 2001; however, see Misra et al. 2012). In vivo studies in fish have found similar findings (Vidal et al. 2005; Zee et al. 2016). Second, unexpected mortalities prevented obtaining fresh tissue samples from all alligators. Only individuals that were euthanized quickly after capture were investigated. Mortalities may also have contributed to an unintentional selection of the most Se-tolerant alligators for sampling (see Finger et al. 2017a for a discussion of this in relation to methylmercury). Lastly, because kidney and liver tissues were used to determine Se levels, we were unable to measure oxidative status of these organs. Selenium tends to preferentially bioaccumulate in the liver and kidney (Burger et al. 2000; Hopkins et al. 2002). Therefore, it is possible that using only blood or muscle to investigate oxidative status may have prevented an accurate reflection of the effects of Se on oxidative status in alligators even though chronic SeMet treatment can lead to significant Se accumulation in both blood and muscle and lead to histological alterations (Haskins et al. 2017b, c). Future studies should investigate whether nondestructive tissues (i.e., blood) can be used to adequately inform oxidative status in relation to Se exposure.

CumTime, HandlingTime, and Plasma CORT

Both CumTime and HandlingTime increased baseline plasma CORT levels. These results suggest that our presence within the facility (i.e., the disturbance and/or the capture and sampling of conspecifics) and the act of capture raised CORT levels, respectively, despite all samples being obtained within 3 min of capture. Nevertheless, when we examined each treatment independently, we found that CumTime and HandlingTime only affected CORT levels of alligators fed 2000 ppm SeMet. This highlights the possibility that chronic Se exposure enhances sensitization to stressors (such as disturbance or capture). If this was the case, however, we would have expected CumTime and HandlingTime to have also increased plasma CORT in alligators fed 1000 ppm SeMet, but this was not observed. Because one alligator in the 2000 ppm SeMet treatment had extremely high plasma CORT levels (26.63 ng/mL), which is equivalent to levels of wild juvenile alligators held in a bag for 2 h after initial capture (Guillette et al. 1997), this individual likely biased the effects of CumTime and HandlingTime on plasma CORT. In fact, when this individual was removed, there was no effect of CumTime or HandlingTime on plasma CORT levels overall or within the 2000 ppm SeMet treatment (data not shown).

Even though control alligators were sampled after all SeMet-treated alligators (507–793 s post-pen entrance), CumTime had no effect on their plasma CORT levels. This suggests that control alligators were not affected by prolonged human presence and/or the auditory (and likely visual) stimuli associated with the capture of other individuals in the SeMet treatments in nearby pens. Similar to these results, Elsey et al. (1990a) found that sampling order did not affect juvenile alligator CORT levels. In a previous study conducted on hatchling farmed saltwater crocodiles (Crocodylus porosus), CumTime significantly increased plasma CORT (Finger et al. 2015a). However, the disparity of results observed in saltwater crocodiles was most likely a consequence of agricultural management practices aimed at minimizing crocodile-human interactions to increase welfare and reduce stress. Our results suggest that crocodilians (or at least American alligators) are capable of acclimating to human presence and/or disturbance in captive settings. Consequently, these results may have significant implications in the realm of crocodilian farming and captive husbandry.

Conclusions

In conjunction with our previous study (Finger et al. 2017b), the current study is one of the first to investigate how Se accumulation affects crocodilians. Moreover, this is the first study to investigate how Se affects stress hormones and oxidative stress in a crocodilian. Our results demonstrate that chronic Se administration causes stress in alligators: alligators fed SeMet had higher levels of baseline (i.e., at initial capture) CORT. This highlights a potential concern for alligators, and other aquatic wildlife, inhabiting Se-contaminated environments.

In contrast to CORT, SeMet treatment had no consistent effect on alligator oxidative status, which is similar to a number of other studies in other taxa (Hoffman and Heinz 1998; Vidal et al. 2005; Miller et al. 2007; Ware et al. 2012; Brady et al. 2013; Zee et al. 2016). The inconsistent effects of SeMet treatment on alligator oxidative status suggests that the mortality observed in our previous study (Finger et al. 2017b) occurred independent of oxidative stress (Zee et al. 2016). However, because ours is the first study to investigate the effects of Se exposure on crocodilian oxidative status, future studies are required to investigate this further. Future studies should investigate the effects of Se on alligator physiology, such as how Se influences immune function, health, and reproduction and provides a more thorough endocrine profile following Se exposure.