**1. Introduction**

Stress responses occur when an animal perceives an external noxious stimulus (stressor) such as predation, adverse weather, habitat change, or anthropogenic disturbances. In physiological terms, such stress reactions can be described as a cascade of endocrine secretions involving the

hypothalamic-pituitary-adrenocortical axis (HPA) and sympatho-adrenomedullary system (SAS), wherein the adrenal glands play an important role [1,2]. These physiological mechanisms are induced to avoid, survive, or recover from an adverse condition [3], and involve the increase of glucocorticoid and/or catecholamine secretion leading to cumulative costs defined as allostatic load. When energy requirements of the organism exceed energy reserves, allostatic overload (type 1) happens [4]. Once the threshold is reached, the emergency response can be triggered, redirecting the animal towards survival and interrupting its normal life history. At a short-term scale, the rise of glucocorticoids (GC) induces physiological effects such as the regulation of the immune system, the increase in glucose created from the body's energy stores, and contributes to memory consolidation [3,5,6]. However, long-term exposure to stressors may trigger type 2 allostatic overload, known as chronic stress [4]. At this point, there is sufficient or even excess energy consumption which can have negative consequences such as hypertension, inhibition of the immune system, promotion of neuronal cell death and reduction of body growth [3,5,6]. Additionally, with chronic, non-lethal stressors, animals can exhibit habituation to the stressor, which means that after a certain period of exposure the animals no longer perceive the stimulus as noxious. Therefore, at this point, the rise of GC is no longer triggered [2].

The quantification of glucocorticoids is an important tool to study the neuro-endocrine stress axis and can provide insights into an animal's well-being as well as a better understanding of ecological and evolutionary processes, which can be useful for conservation purposes [7–12]. The determination of cortisol or corticosterone in plasma is widely used as an indicator of stress [1,12]. This measure represents the instant stress status of an animal and can be used as a tool for experimental studies to understand the effects of specific stressors such as predators or weather conditions [12–14]. However, since this method requires the handling of animals for blood collection, which in itself represents a stressful situation for animals, new methods with non-invasive sampling have been developed. Above all, analysis of glucocorticoid metabolites in feces has been established as a validated, successful alternative to blood, in several studies [1,9,10,15,16]. In red deer, using fecal samples, maximum levels of cortisol metabolites are usually recorded with a delay of approximately 18 h after the disturbance [17]. In addition, cortisol in hair has been used as an indicator of stressful conditions [18,19]. Some recent studies sugges<sup>t</sup> that hair shows levels of GC that are accumulated for some weeks to months, depending on the species [20–23].

It has been shown that hunting is capable of causing an intense stress response in wildlife, which can significantly affect animal welfare. Bateson and Bradshaw [24] showed that hunting triggers physiological effects in red deer (*Cervus elaphus*), such as disruption of muscle tissue and alterations of concentrations of β-endorphin. Moreover, the same authors demonstrated that cortisol levels increased considerably when animals were exposed to this kind of harvesting. Although its effects are poorly explored, hunting activities, due to their seasonality, can result in repeated and prolonged stress. Recently, a study suggested that active hunting events using hounds or involving grea<sup>t</sup> densities of people can have a stronger impact with higher cortisol levels in cervids than calmer stalking [25].

In Portugal and Spain, the main hunting process used is a non-selective process called "montaria", which involves the release of packs of dogs (usually 150 to 300 dogs), that will chase the animals during a maximum of five hours, and move them toward the hunters, positioned at the edge of the hunting area [26,27]. Red deer is an important game species in Portugal, and its exploitation is mostly done with this type of hunting process [26]. Due to the effects that this hunting practice can have on wildlife, its ethical and economical effects have been debated [28]. Notwithstanding the revenue that hunting can generate, this activity can also lead to several consequences in terms of disruption of age and social structure, animal dispersion and life expectancy [29,30]. Moreover, and as suggested in some studies [31,32], in wild populations, hunting may introduce an element of artificial selection. Thus, hunting activities can induce significant changes in demography and population genetics, such as social disruption or consequences on the gene pool of hunted populations [32,33]. The sustainability and conservation implications of this practice, which include effects on demography, genetic pool, behavior and general health, highlight the importance of managemen<sup>t</sup> and regulation of hunting activities [5,27,33].

The main aim of this study was to investigate the impact of hunting activities on a wild population of red deer using measurements of glucocorticoids in plasma and hair and its metabolites in feces, all collected from hunted animals. We thus indirectly evaluated the stress levels of each hunted animal from a few months before the hunting event until its death. Given the wide e ffect that a stress response can have on the body, the physical and immunological condition of the animals was also evaluated.

#### **2. Materials and Methods**

#### *2.1. Study Area and Red Deer Population*

The study took place at Lousã Mountain (40◦3 N, 8◦15 W) and surrounding hunting areas, located in Central Portugal. The climate in this Mediterranean area is characterized by hot and dry summers and rainy winters [34]. With an altitude that ranges from 100 to 1205 m, this mountainous region is predominantly composed of plantations of coniferous and broadleaf trees, and large shrubland areas [26].

The red deer population in this region is the result of a reintroduction program that occurred between 1995 and 1999 with the release of 96 animals. Currently, this species occupies around 435 km<sup>2</sup> of the Lousã Mountain and surrounding areas, with an estimated density of 5.6 deer/km<sup>2</sup> during the rut season [35,36]. The calving season of this population occurs in May/June and rutting in mid-September to the end of October. With females being 37 ± 3% (mean ± SE) smaller than males, this species shows a very marked sexual dimorphism in body size. In addition, the sexual segregation outside the breeding season results in a matriarchal society wherein the females adopt a more philopatric behavior and males tend to disperse [35,37]. Since there are no natural predators, red deer are mainly preyed upon by feral dogs. This species is one of the most hunted big game species of the Lousã Mountain region, which includes 12 hunting areas, with red deer hunted in seven of them since the start of hunting in the region in 2006/2007 [26].

## *2.2. Data Collection*

The study was performed during two hunting seasons (2013/2014 and 2014/2015), enabling the collection of samples from 80 red deer (38 adults: 26 females and 12 males; 30 sub-adults: 14 females and 16 males; and 12 young). The samples were collected during autumn (October and November) and winter (January and February) from red deer hunted in six "montarias" in three contiguous hunting areas: ZCM (Municipal Hunting Area) of Lousã (three hunting events—34 animals), ZCM of Vila Nova (one hunting event—7 animals) and ZCA (Associative Hunting Area) of Miranda do Corvo (two hunting events—39 animals) located in the Lousã Mountain area. No hunting events were conducted during the month of December. Post-mortem examination was made in situ, one to four hours after the animals were killed, and included the collection of blood, feces directly from the rectum, hair from the dorsal region and the metatarsus, and the recording of the sex and age class (based on animal size, body conformation and characteristics of antlers) of each animal. In addition, in November, fecal samples (*n* = 11) were also collected non-invasively during behavioral observations in the non-hunting area (i.e., area where hunting is not allowed) located in the central part of the Lousã Mountain, to use as the control.

Blood samples were taken directly from the heart into EDTA-tubes and centrifuged at 2000× *g* for 5 min. Plasma was collected and frozen at −20 ◦C in multiple aliquots. Feces, hair, and the metatarsus were frozen and stored at −20 ◦C for subsequent analyses.

#### *2.3. Steroid Extraction and Quantification*

Five ml of diethyl ether was added to each plasma sample (0.5 mL). After being shaken and centrifuged (2500× *g*, 15 min), the samples were frozen. Afterwards, the liquid component was transferred to a glass vial and dried under a stream of nitrogen (40 ◦C). This procedure was performed twice to increase the recovery of the extraction process. The combined and dried down extracts of each sample were dissolved in 0.5 mL of assay bu ffer and an aliquot analyzed by a cortisol enzyme immunoassay (EIA), as described in detail before [38]).

From each dried fecal sample 0.2 g were taken and mixed with 4 mL of methanol (100%) and 1 mL of water. The samples were shaken for 30 min and centrifuged at 2500× *g* for 15 min [39]. A group (with a 3 α-11-one structure) of fecal cortisol metabolites (FCMs) was measured using an 11-oxoetiocholanonolone EIA. The utilized biotinylated label and the antibody (including its cross-reactivity) have been previously described in detail [40]. The assay has been successfully validated for use in red deer [41].

Hair samples were cut into fragments (<0.5 cm), washed with 5 mL 100% n-hexane to remove any lipids and potential external contamination, and air dried. Hair samples (0.05 g) were mixed with 5 mL of 100% methanol and incubated for 72 h for glucocorticoid extraction. After transferring the supernatant into a new glass vial, the organic solvent was evaporated at 40 ◦C using a stream of nitrogen. The extracts were dissolved in 0.5 mL of assay bu ffer and analyzed with a cortisol EIA [38,42].

#### *2.4. Physical and Immunological Conditions*

The physical condition of animals was assessed using bone marrow fat (BMF). The BMF was determined using the metatarsus, from where the bone marrow was extracted and weighed (± 0.0001 g). The bone marrow samples were then oven-dried at 60 ◦C and reweighed. BMF was determined as BMF = (the weight of oven-dried marrow/fresh marrow) × 100 [43].

To evaluate the immunological state of an individual, we used the blood collected in the field to make blood smears in the lab, followed by staining [44]. White blood cell counts and identification (100 cells per smears) were performed using a Nikon Eclipse Ni microscope (Nikon Instruments Europe B.V., Amsterdam, Netherlands). The examination and classification of blood cells were made based on morphological criteria and properties of staining, allowing the identification of five types of white blood cells, namely lymphocytes, neutrophils, eosinophils, monocytes and basophils, respectively [45].
