Physiological stress levels predict survival probabilities in wild rabbits
Sonia Cabezas a,⁎, Julio Blas b , Tracy A. Marchant b , Sacramento Moreno a
a Department of Applied Biology, Estación Biológica de Doñana, Spanish Council for Scientific Research–CSIC, Avenida María Luisa s/n. E-41013, Sevilla, Spain
b Department of Biology, University of Saskatchewan, 112 Science Pl., Saskatoon, SK, Canada S7N 5E2
Abstract
Among vertebrates, short-term elevations of glucocorticoid hormones (corticosterone or cortisol) facilitate a suite of physiological and behavioral changes aimed at overcoming environmental perturbations or other stressful events. However, chronically elevated glucocorticoids can have deleterious physiological consequences, and it is still unclear as to what constitutes an adaptive physiological response to long-term stress. In this study, we experimentally exposed European wild rabbits Oryctolagus cuniculus to a source of long-term stress (simulated through a 2- to 4- week period of captivity) and tested whether glucocorticoid physiology predicted two major components of rabbit fitness: body condition and survival probability. Following exposure to long-term stress, moderately elevated serum corticosterone and fecal glucocorticoid metabolites levels in the wild rabbits were negatively associated with body condition, but positively associated with subsequent survival upon release. Our results suggest that the cost of maintaining elevated corticosterone levels in terms of decreased body condition is balanced by the increased chance of survival upon release.
.
Keywords: Body condition; Corticosterone; European wild rabbit; Fecal glucocorticoid metabolites; Non-invasive monitoring; Oryctolagus cuniculus; Stress; Survival; Translocation
Introduction
Animals respond physiologically to the perception of an array of external noxious or stressful stimuli, including predation attempts, harsh weather, habitat change and anthro- pogenic disturbances, through a rapid cascade of endocrine secretions within the hypothalamic–pituitary–adrenal (HPA) axis (Axelrod and Reisine, 1984; Sapolsky et al., 2000; Wingfield et al., 1997, 1998). This response is widely conserved across vertebrates and ultimately involves the secretion of glucocorticoids (GC, cortisol or corticosterone) from the adrenals. The presence of elevated GC allows individuals to cope with short-term stressors through a complex set of physiological and behavioral changes aimed at halting non- essential activities while the stressful perturbation persists. Mobilization of body energy stores, increased gluconeogenesis,
⁎ Corresponding author. Fax: +34 954 621 125.
E-mail addresses: ,
(S. Cabezas).
and the promotion of food-searching and dispersal behaviors are some of the adaptive changes activated by GC. Wingfield et al. (1998) proposed that the activation of the HPA axis under stressful situations constitutes a distinct “emergency” life history stage in vertebrates.
The degree of stress response activation often correlates with the overall health of an individual, and the quantification of GC levels is often used to infer the health status of animal populations (Creel et al., 1996; Morton et al., 1995; Romero,
2004; Wasser et al., 2000; Wingfield et al., 1997). In particular, the utilization of a capture and restraint protocol (Wingfield,
1994) has become a popular approach whereby the adrenocor- tical response of wild individuals to a short period of capture and human handling is measured, and inferences are drawn about how the population would cope with other sources of stress. Although this approach is a valuable tool to evaluate the response to stress over the course of a few minutes to hours (i.e., acute stressors), it is not clear whether short-term responsive- ness can predict adaptation to longer term stress (i.e., chronic stressors). This is important because many anthropogenic disturbances (e.g., habitat degradation, exposure to tourism,
contamination episodes) and a number of wildlife management practices (e.g., habitat restoration, reintroductions and reinfor- cements of animal populations) involve perturbations that last for considerable periods of time, and a prolonged activation of the HPA axis may reverse earlier adaptive responses to the stressful event. Some studies have shown that chronic GC elevation may produce potentially detrimental effects, especial- ly reduced survival and fitness through infertility, impaired resistance to disease, inhibition of growth and atrophy of body tissues (Rogovin et al., 2003; Romero and Wikelski, 2001; Sapolsky et al., 2000; Suorsa et al., 2003). Others, however, suggest that chronic but moderate GC elevation provides adaptive advantages to the individual (Cote et al., 2006; Pravosudov, 2003). Therefore, the interpretation of inter- individual differences in the HPA axis responsiveness may yield opposite conclusions depending whether the stressor is acute or chronic and also depending on the actual level of activation of the GC release (Romero, 2004).
Despite differences between short- and long-term respon- siveness to stressors, few studies have experimentally evaluated what constitutes an adequate response to long-term perturba- tions or whether differences in the associated stress response can actually predict the fitness of an individual. In this context, the translocation of animals, used to improve and recover natural populations of wildlife, might provide useful informa- tion about adaptive responses to chronic stress. The establish- ment of quarantine periods prior to release is a highly recommended veterinary practice in translocation programs because the potential transmission of new disease agents into the release area, may have a major impact on native populations and on the overall success of the translocation program (Woodford and Rossiter, 1994). Quarantine periods imply handling and captivity of animals over several days or weeks, can be performed in a standardized manner, and often involve subsequent monitoring of the introduced animals. Consequent- ly, translocation programs constitute a unique opportunity to study how individuals cope with a long-term source of stress and whether assessment of GC secretion can provide predictive information on major components of fitness such as body condition and survival. In this sense, the study of HPA responsiveness during a quarantine period constitutes an experimental capture and restraint protocol, but modified so that the exposure to the source of perturbation lasts for weeks rather than minutes or hours. This information could be then used to evaluate and predict a population-level response to chronic stress exposure as occurs with many of the anthropo- genic disturbances threatening wildlife populations.
We tested these possibilities in European wild rabbits (Oryctolagus cuniculus) during a translocation program within the Doñana Natural Park in Huelva province of south-west Spain. In particular, we studied whether GC secretion and metabolism during the quarantine period could explain two major components of individual rabbit fitness: body condition and survival probability. Native populations of European wild rabbits have suffered a considerable decline since the early
1950s in the Iberian Peninsula. This is a major conservation concern because the viability of several endangered predators
endemic to this region (the Spanish imperial eagle Aquila adalberti and Iberian lynx Lynx pardina in particular) relies almost exclusively on rabbits as a prey source (Ferrer and Negro, 2004). As a consequence, translocations of wild rabbits are one of the most frequently used management tools to boost the density of natural populations in Spain (Cabezas, 2005; Calvete and Estrada, 2004; Moreno et al., 2004). In our study, wild individuals were captured, maintained in captivity during a quarantine period, and then released and monitored for survival through radio-tracking. The physiological response to stress during the quarantine period was assessed by measuring circulating corticosterone (CORT) and GC metabolites (GCM) in feces. Finally, we analyzed the association between the GC measurements and various components of individual fitness of the rabbits, including body condition and survival in the wild.
Methods
Experimental design
Between February and March 2002 we captured 44 adult wild rabbits (24 males and 20 females) in the province of Cadiz (SW Spain). After capture, rabbits were transported to experimental facilities, sexed, weighed (mean ± SE:
1062 ± 14.3 g) and identified with numbered metallic ear-tags. All rabbits were adults and sexually mature as determined by body mass and the period of year in which they were captured (Soriguer, 1981). Captures were carried out several weeks prior to the start of the seasonal peak of reproduction, which normally occurs in late spring in nearby populations (Villafuerte et al., 1997). All females were checked for signs of pregnancy and lactation by means of palpation of the abdomen and mammary glands. Only one of the 20 females was pregnant, but the fetuses were aborted few days after capture. We allowed a 3-week recovery period before release and this animal showed body mass and corticosterone values well within the recorded range for other rabbits at the end of the quarantine; consequently, we decided not to eliminate it from our experimental sample. Standardized sanitary protocols have been established for the rabbit translocation program at Doñana National Park in concert with Spanish laws (Calvete et al., 2005). We followed these protocols in consultation with the veterinary staff of the Junta de Andalucia, and prioritized ethical considerations over scientific goals. At the beginning of the quarantine period all individuals were treated for external and internal parasites and subcutaneously vaccinated against myxomatosis and rabbit hemorrhagic disease (RHD) using doses of commercial vaccines recom- mended for domestic rabbits. During the captive period, all animals were housed individually in flat-deck cages commonly used for domestic rabbits, and provided ad libitum access to water and food (commercial pelleted food and hay). Rabbits were sheltered indoors but exposed to natural photoperiod and temperature. The captive period lasted between 2 and 4 weeks to conform to a release schedule planned as part of the translocation program. Two weeks was the minimum time the rabbits spent in quarantine; this was based on the myxomatosis incubation period and the time following vaccination needed to develop immunity against myxomatosis, and RHD in domestic rabbits (Argüello, 1991).
Two days before being released, rabbits were weighed (mean ± SE: 940 ±
15.1 g) and the cubit length was measured for subsequent calculation of body mass index by means of residuals of the reduced major axis regression between log10-weight and log10-cubit length (Green, 2001). In addition, blood and fecal samples were collected to determine CORT and GCM levels, respectively. The collected fresh fecal samples corresponded to the night previous to blood extraction (i.e., between 11:00 pm and 9:00 am), and blood samples were drawn through puncture in the ear vein between 9:30 am and
2:00 pm. Serum CORT shows daily variations in rabbits (Szeto et al., 2004) and we recorded the exact time of day of blood collection for each rabbit in order to control for this potential source of variation in the statistical analyses. We also recorded the time elapsed between catching each rabbit from its cage and taking the blood sample (mean ± SE: 4.2 ± 0.2 min); this was also
considered in the statistical analyses. Serum (obtained after blood centrifu- gation) and fecal samples were stored frozen at − 20 °C until GC quantification.
After the captive period, 28 rabbits (14 males and 14 females) were equipped with a radio-collar, weighing approximately 20 g and containing an activity sensor (Biotrack, Wareham, UK), and released in the Doñana Natural Park. Survival was monitored by performing daily localization of each individual over the 30 days following release. When the activity sensor indicated lack of movement, we searched the field grounds until the animal remains were located.
Serum corticosterone measurement
Corticosterone is the major adrenal glucocorticoid secreted by the European rabbit (Szeto et al., 2004). Serum levels of CORT were analyzed following extraction with diethyl ether and utilizing techniques described fully elsewhere (Wayland et al., 2002). Extraction efficiency, measured in samples spiked with
3H-corticosterone, was 96%. Dried ether extracts were reconstituted in a small
volume of assay buffer and frozen at − 20 °C until CORT was measured by radioimmunoassay (RIA). This RIA employs an antiserum purchased from Sigma Chemicals (Oakville, Ontario) and 3H-corticosterone (Amersham Biosciences, Baie d'Urfe, Quebec). Cross-reactivity of the antiserum with other secreted steroids, including cortisol and progesterone, is low (Wayland et al., 2002). The minimum detection limit of the RIA, defined as the dose of CORT which produced a relative binding (%B/Bo) of 80% in the RIA (ED80), was 14.5 pg. Serum extracts were measured over two assays. Assay variability was determined as the %coefficient of variation (%CV) resulting from repeated measurement of samples spiked with a known amount of CORT (n = 3) in each assay. The within-assay variability in the two assays was 1.3% and 4.6%. Between assays variability was determined to be 3.0%. Serial dilutions of the rabbit serum extracts generated a displacement curve parallel to the corticosterone standard curve. All serum extracts were diluted several fold with assay buffer to correspond to a level near the midrange (ED50) of the assay.
Fecal glucocorticoid metabolites measurement
Glucocorticoid metabolites in fecal samples were extracted utilizing procedures modified from Wasser et al. (2000). Dried feces (approximately
200 mg) were placed in a glass vial with approximately 5 ml of 90% HPLC graded methanol (VWR International, Missisauga, Ontario) and homogenized at maximum speed (approximately 25,000 RPM) with a hand-held motorized homogenizer (Omni 2000, Pro Scientific, Connecticut). The homogenized samples were then placed in a shaking water bath at room temperature for
30 min, followed by centrifugation at 4000×g. The supernatant was air dried overnight, followed by reconstitution in 1.0 ml methanol. An aliquot of the reconstituted methanolic supernatant was then diluted 1:5 in assay buffer and stored at − 20 °C. Samples were further diluted in assay buffer prior to measurement of the GCM.
The GCM content of the diluted fecal extracts was determined with the same RIA used to measure serum CORT levels as described above (Sigma RIA) as well as a second RIA (ICN RIA) developed using an anti-corticosterone antiserum purchased from MP Biomedicals (Solon, Ohio). Serial dilutions of the rabbit fecal extracts were assessed for parallelism to the CORT standards in the Sigma RIA. Lower dilutions of the methanol extracts (ranging from 1:5 to 1:40) displayed significant interference and non-parallelism in the Sigma RIA. Further analysis revealed that this was due to non-specific (i.e., non-displaceable) binding between the fecal extracts and the 3H-corticosterone. The constituent(s) in the methanol extract that contributed to the interference in the Sigma RIA at these lower dilutions is not clear. However, this interference in the RIA disappeared with further dilution of the extracts. To minimize the possibility that any given sample might display a non-specific interference (i.e., non-parallelism to the standard curve) in the Sigma RIA, all fecal samples were analyzed at multiple dilutions (from 1:80 to 1:320) and were found to be parallel to the standard curve over this dilution range. Samples were analyzed in a total of three separate assays. The within-assay variability was determined to be 2.8, 5.0 and