Wild fish are frequently exposed to multiple stressors, but the influence of previous or ongoing stress on an animal's subsequent response is poorly understood. Using wild‐caught bluegill sunfish (Lepomis macrochirus) as a model, we used exogenous hormone implants to experimentally raise circulating cortisol in a group of fish for ∼10 days. We also maintained sham‐treated and control groups of fish. We subjected all animals to a secondary stressor in the form of either a heat challenge or fasting challenge. We compared survival, body condition, and plasma‐borne indicators of physiological status among cortisol‐treated, sham‐treated, and control groups following the secondary stressor. In order to compare short‐ and long‐term effects of cortisol treatment, we initiated the secondary stressor either 4 or 30 days following initial cortisol treatment. Cortisol‐treated fish succumbed to the fasting challenge sooner than sham‐treated and control fish at both 4 and 30 days. Interestingly, cortisol‐treated fish lost equilibrium sooner than sham‐treated and control fish during the heat challenge when conducted at 30 days, but not at 4 days. These results demonstrate that multiple simultaneous stressors have cumulative effects on bluegill sunfish. Furthermore, these results demonstrate that supraphysiological cortisol doses alter the long‐term responses of bluegill sunfish to additional challenges, even after apparent recovery. Such cumulative and long‐term effects may be an important factor in mediating the response of wild animals to natural and anthropogenic stressors, and should be considered in ecological studies.
All fish were sampled under an Ontario Ministry of Natural Resources Scientific Collection Permit and handled in accordance with the guidelines of the Canadian Council on Animal Care (Carleton University, B09–11). Experimentation took place at Queen's University Biological Field Station in eastern Ontario, Canada (44°31′N, 76°20′W). Adult wild bluegill sunfish (total length > 150 mm) were used as a model species, because they are abundant and can easily be captured and held in captivity. In May and June, 2009, fish were captured from Lake Opinicon by angling and were then held in 1,000‐L outdoor tanks supplied with flow‐through lake water (∼18 ˚C). Fish were randomly distributed into three groups: cortisol treatment, sham treatment, and control. Cortisol‐treated fish received an intraperitoneal injection of 50 mg kg−1 cortisol (hydrocortisone 21‐hemisuccinate; Sigma H4881, Sigma‐Aldrich, Oakville, Ontario, Canada) emulsified in pure cocoa butter (0.005 mL g−1). Sham‐treated fish received only the cocoa butter vehicle, while control fish were not injected. All fish were identified by individual and treatment type using anchor tags. We conducted two distinct experiments, each using these initial cortisol treatments. Both experiments consisted of applying a secondary stressor at either 4 or 30 days following the initial treatment. All fish were fed pellets (2% adjusted body weight; Martin Mills, Floating Feed, 3 mm) twice‐daily during the 4‐ or 30‐day holding periods, and were monitored closely at regular intervals for loss of condition (e.g., emaciation) leading to mortality and/or abnormal behavior (e.g., loss of equilibrium).
Time Course of Cortisol Effects
To provide context for the elevation of cortisol used in the experiments, we conducted a time‐course experiment in which cortisol, glucose, hematocrit, and condition values were measured over a 30‐day period. In the spring of 2010, groups of 60 fish (n = 20 per treatment) were angled, randomly subjected to one of our three treatments, and placed in a 1,000‐L outdoor tank. Fish were fed twice‐daily and monitored closely for 24 hr, 4 days, 10 days, or 30 days. At the end of the monitoring period, a subset of fish (n = 10 per treatment) was removed from the tank and individually placed into individual opaque sensory deprivation chambers (∼2 L) supplied with a constant flow of lake water for 24 hr. Fish were then removed from the chambers individually and blood sampled by caudal puncture, using lithium‐heparinized 1‐mL syringes and 25 gauge, 38‐mm needles. Blood samples were collected within 2 min of removing fish from the chambers or samples were excluded from analysis. Previous research by our group has revealed that acute cortisol responses for bluegill are still below 170 ng mL−1 at 10 min and peak at 40 min post‐stressor (Cook, 2011), which makes a 2‐min sample appropriate as a nonstressed value. In addition, a separate group of fish (n = 20) was sampled immediately after angling from the lake for baseline field values, an approach that has previously used to obtain baseline values for wild fish (O'Connor et al., 2009). Whole blood was analyzed for glucose concentration, measured on 10 μL of whole blood with a hand‐held glucose meter (ACCU‐CHEK glucose meter; Roche Diagnostics, Basel, Switzerland), a device previously validated for use on fish (Cooke et al., 2008). Hematocrit values were determined using microhematocrit tubes centrifuged for 5 min (CritSpin‐Micro‐Hematocrit Centrifuge, Norwood, MA, USA). The remaining blood was centrifuged (Fisher Scientific Micro‐Fuge, Toronto, Ontario, Canada) at 2,000 × g for 5 min. Plasma samples were frozen immediately in liquid nitrogen and then transferred to a –80°C freezer and stored until cortisol analysis (see below). Fish were euthanized by cerebral percussion and dissected, collecting information relevant to condition and health of the individual.