The term stress originated in physics to describe pressure and deformation in a system, but it has been adopted into a biological context through the work of Hans Selye [1,2,3]. He recognized in mammals, as a “general adaptation syndrome,” a similar suite of coordinated reactions to diverse noxious stimuli or “agents” [4]. Selye’s concept was at first criticized by physiologists as vague and immeasurable, but he subsequently clarified his concept defining several stress response elements, principally the hypothalamo-pituitary-adrenal (HPA) axis system. Stress is now recognized as a valid physiological concept, which allows organisms to respond to adverse environmental pressures [5]. Most studies use the word “stress” to describe negative treatments applied to organisms in an experiment, such as nutritional, heat or oxidative stress. Here, the triggering stimuli will be called “stressors” while “stress” will be considered as the response syndrome to any aversive or harmful treatment in a specific system. This understanding can be applied to different levels of organization: molecular, cellular, histological, physiological, even ecological or social, but this review will focus on physiological processes involved in an integrated response at the level of the organism. Also, the definition of stress should take into account the duration and intensity of the stressors involved, thereby the distinction between acute or chronic stress responses. Here, we review the putative elements participating in a physiological stress response, and propose an integrated model of the honey bee stress response. Although the model is based to a degree on the stress literature known from other insects [6,7,8,9], our intention is to build (as far as possible) a model that is honey bee specific. In doing so, we identify what is known about this particular species, and what may be assumed from our knowledge of other insects. Consistently, we wish to highlight the gaps in our existing knowledge to propose directions for future stress research.

The concept of stress is useful in understanding the physiological and behavioral responses of honey bees to harmful situations. This research is timely since, over the last years, beekeepers from different geographic areas have reported a marked increase in honey bee colony failure rates and in the number of stressors affecting bees, including diseases, parasites, pesticides and poor nutrition [10,11,12]. The syndrome termed “colony collapse disorder” seems to be the result of an accumulation of stressors chronically weakening honey bee colonies [13,14]. As honey bees are the most important insects in agriculture for both pollination of diverse crops and honey production, the recent decline in their populations brings an urgent need to know more about the stress response systems of this ecologically and economically important insect.

In addition, the honey bee is an ideal insect model to understand the evolution of sociality. A key feature of honey bees is their high level of social organization and their well-developed system of division of labor among workers [15]. Honey bees exhibit age polyethism; young workers perform in-hive tasks (e.g., taking care of the brood), then become guards patrolling the entrance of the hive and later become foragers. Studying differences in stress responses across behavioral castes might help elucidate how a defined division of labor has evolved.

There are three stages to an organism’s acute stress response: it first detects the stressor with sensory organs, then responds to it by defense or escape. Finally, if the stressor cannot be avoided and is sustained, the organism enters a state of exhaustion [1]. Following detection of a stressor, mammalian physiological responses are coordinated by neural activity within the autonomic nervous system and the HPA axis (Figure 1). The first immediate response is an activation of sympathetic neurons, which stimulate the adrenal medulla to release adrenaline and noradrenalin into the blood. These two catecholamines increase heart rate and vasoconstriction. In parallel, the paraventricular nucleus (PVN) neurons in the hypothalamus release corticotrophin-releasing hormone (CRH) and arginine/vasopressin (AVP), which are conveyed to the nearby anterior pituitary gland via the blood stream. In response, the pituitary secretes adrenocorticotropic hormone (ACTH), which acts on the adrenal cortex to release glucocorticoids (such as cortisol) causing mobilization of energy reserves (i.e., glucogenolysis in the liver) [16]. Glucocorticoids also potentiate catecholamine release from the adrenal medulla. In parallel, adrenaline activates a release of glucagon by the pancreas to further increase the catabolism of glycogen in the liver and raise glucose concentration in the blood. Other hormones such as cytokines or endogenous opioids may also be produced and/or released, depending on the nature, duration and intensity of the stressor, to act in diverse ways to limit the degree of tissue damage. Therefore, it is interesting to note that, additionally to the general stress response pathways described previously, certain stress responses can vary depending on the type and duration of the stressor.

Under chronic stress, the immune system, metabolic pathways and cognitive processes in the organism gradually weaken until exhaustion and failure are reached [1,17]. For example, repetitive HPA activation resulting in an excess of glucocorticoids in the blood can lead to metabolic diseases such as diabetes [2,18].

Cellular stress responses described in various models (bacteria, yeast, worms and flies) include the increased production or activation of antioxidant proteins and heat shock proteins (HSP) when facing high metabolic load or environmental stressors [19,20]. Such proteins may be called “stress proteins” [21] and used as cellular stress biomarkers [22,23]. These factors are induced by a variety of stressors such as extreme temperatures, elevated ion concentrations or toxic substances, all usually resulting in excessive amounts of denatured proteins [24]. Their actions are principally intracellular and hence we do not focus on them in this review that considers instead more integrated elements of a general stress response.

Multiple aspects of stress responses have been used to evaluate stress in honey bees, including behavioral, physiological or cellular stress responses. The many parameters used are listed in Table 1. Behavioral stress responses, usually immediate responses, characterized early on by Cannon [26] as the “fight-of-flight” responses, are easy to observe in the honey bee. For example, extension of the sting (or stinging of a target) has been used to evaluate sensitivity to stressors, and is widely considered as indicative of stress in honey bees, as well as an aggressive response. Physiological measures of stress responses in honey bees include hormonal titers and neurotransmitter levels—these parameters have been integrated into our model (see Table 1 and Table 2, Section 2.2 and Section 3 and Figure 2). Honeybee stress studies usually use acute stressors but the nature and duration of the stressors could sometimes be qualified as chronic (Table 1). Cellular stress responses have also been used in honey bees [27,28,29,30,31,32], and Duell et al. [33] even suggest some cellular stress biomarkers as elements for a diagnostic of general stress in honey bees. Also, since exhaustion is the final stage of chronic stress response described by Selye [1], survival rate has been used to assess the degree of stress.

Based on the data available for honey bees (Table 1) and other insect species we have tried to synthesize a model of a general stress response pathway specific to the honey bee. It should be kept in mind that many proteins or genes of unknown function may be affected by stressors; we will only focus on a few of them, for which sequence homologies and/or functional data suggest a potential role in a physiological stress response.

When faced with a stressor such as predation, robbing or adverse climatic conditions, a honey bee will need to increase her mobility and mobilize energy reserves to cope with the sudden increase of metabolic demand. We suggest here that this rapid change in physiology is achieved by a coordinated endocrine and neuroendocrine response (Figure 2).

As soon as the stressor is detected via appropriate receptors (e.g., olfactory, mechanosensory or visual), our model proposes that there will be release of octopamine and dopamine within the brain, thus increasing arousal [49] (see Section 3.1). Other signals like cortico-releasing hormone-binding protein (CRH-BP) might also participate in the brain stress response [37]. Octopamine is also released into the hemolymph [57] from neurohemal cells [58] to act on many organs and coordinate a physiological response to the stressor (see Section 3.2.1). Peripheral octopamine increases heart rate, and may modulate ventilation and stimulate mobilization from muscles [51,59]. The activation of the neurosecretory cells of the corpora cardiaca (CC), the major brain neurosecretory organ, stimulates the release of several neurohormones: adipokinetic hormones (AKH) and possibly corazonin, into the hemolymph [6,54,56] to mobilize energy from body stores (see Section 3.2). Finally, we suggest here (from honey bee physiology studies) that hormonal factors including AKH and candidates like allatostatin-A (AST-A), diuretic hormones (DHs) and tachykinins [56], could reinforce the liberation of trehalose and glucose from the fat body, but also from the main energy store of the honey bee—the crop [56,60,61] (see Section 3.3 and Figure 3).

In the periphery, the immediate physiological stress response might be coordinated by nerve signals allowing a very fast reaction, but (as in vertebrates) neuroendocrine systems seem to also play a major role in honey bees and other insects. Important components are neurosecretory cells of the CC, which integrate neuronal signals and may trigger broad effects in a variety of target cells through endocrine signals in the hemolymph [89]. Like the vertebrate pituitary gland, the insect CC houses many neuroendocrine cells that play a central role in the regulation of diverse metabolic functions [90].

Increased energy mobilization triggered by hormonal signals plays a very important role in stress responses. In insects, carbohydrates (especially trehalose) are the most important energy source [164]. However, available trehalose is rapidly depleted, and in several insects, a sustained effort such as a long flight requires the release of trehalose from the main energy store—the fat body [109,165]. Consequently, mobilization of energy from energy stores by hormonal factors is an essential part of the stress response. By contrast, honey bees show specific mechanisms for energy mobilization which appear related to their social organization. Strikingly, forager honey bees have fat bodies almost entirely depleted and seem to use the sugar energy reserve carried in their crops to sustain the energetic demands of flight [60,61]. This observation would also explain why foragers express almost no AKH in their CC, and have lower abdominal glycogen stores [166,167].

Honey bees appear to have a quite specific mechanism for regulating hemolymph sugar levels (Figure 3), according to a model proposed by Blatt and Roces [61]. In an energy-demanding situation, trehalose synthesis by the fat body is not fast enough to match rates of trehalose consumption, so circulating trehalose levels decline. This stimulates the passage of nectar from the impermeable crop storage organ to the midgut by the contraction of the proventriculus (gut muscle between the crop and midgut) [62]. From the midgut, sugars are digested, absorbed and more glucose and fructose are transported into the hemolymph. Therefore, upon high metabolic demand, while the trehalose level decreases in the hemolymph, those of fructose and glucose increase, maintaining a stable sugar concentration (Figure 3). Moreover, the honey bee genome sequence suggests the loss of two important insect enzymes converting gluconeogenic substrates to trehalose and glycogen [168], which are both stored in fat body and considered as the primary energy storage molecules in insects. This implies that in the honey bee, regulation of sugar transport from the gut probably plays a more important role in energy balance than the regulation of trehalose release from the fat body. Interestingly, in honey bees injection of CC extracts into the hemolymph has a hypertrehalosemic effect [167], thus candidates for this function are expected to be found in this gland. As discussed above, hormonal candidates performing this role in honey bees might include corazonin, DH-I and possibly AKH (Figure 3).

Hormones activating the mobilization of glucose from the crop by stimulating the proventriculus and midgut remain to be identified. In Drosophila [56], it has been suggested that the (neuro)hormones: AST-A and tachykinin, released from secretory cells of the midgut or the nerve cord, might play this role [56] (Figure 3). Tachykinin and tachykinin-related peptides (TRPs) are known in insects to be myostimulatory of the insect midgut muscle [169], and therefore would be good candidates for modulating release of nectar from the honey bee crop. One “tachykinin-like receptor” has been identified in honey bees from sequence homology with Drosophila [170], but its function remains unclear, especially as many TRPs have been described and seem to have diverse functions in insects [169]. Nine different TRPs have been localized all along the nervous system of the honey bee [88], but none of them was detected in CC neurosecretory cells. If TRPs act to upregulate metabolism in stress, they would need to be released from the efferent peripheral nerves directly to the proventriculus.

It should be noted that insulin and vitellogenin pathways have also been linked to energy store mobilization and oxidative stress in insects [47,171]. Insulin/insulin-like growth factor signaling (IIS) pathways appears to regulate fat stores in Drosophila [171] and confer oxidative stress resistance in Drosophila and honey bees [48]. Vitellogenin expression seems to be triggered by IIS pathways [48]. In reproductive females, vitellogenin is a glycolipoprotein stored in the fat body and usually released into the hemolymph before being stored within oocytes. In the sterile honey bee worker, vitellogenin expression is inhibited by JH [172], thus giving foragers a lower amount of vitellogenin (perhaps reinforced by their decrease of fat body mass). However, vitellogenin seems to be protective against oxidative stress, perhaps via certain antioxidant properties, and may account for longevity in reproductive insects, e.g., queen bees [47,48]. Presently, the metabolic role of these molecules needs to be clarified in honey bees in order to be integrated within our model of endocrine regulation of energy sources.

JH acts on development and sexual maturation in insects; it is produced in the CC and released under the neural control of brain peptidergic innervation. JH is typically described as the master larval developmental hormone, boosting growth of insects, inhibiting metamorphosis and initiating reproductive traits in adults [173]. Still, JH has been described more recently as a stress hormone in Drosophila, as JH levels drop after exposure to various stressors [174]. As JH tends to have long-lasting effects, this hormone may be more likely to be involved in chronic than acute stress response. Whether JH acts as a stress hormone in honey bees is less clear, particularly since, in the adult worker bee, it has a species-specific function: that of a regulator of division of labor. JH titers are low in young nurse bees but higher in foragers. Indeed, pharmacological elevation of JH levels or injection of JH analogs accelerates the onset of foraging of young bees [175,176]. Perhaps because of this, studies of the possible role of JH in stress have thus far given confusing results. Lin et al. [34] could not find a consistent change in JH levels after application of various stressors in honey bees, and found a response only if JH levels were initially low (in which case JH levels increased after caging or cold anesthesia). If JH levels were already high, stress seemed to decrease JH levels. These differences probably result from the dynamics of mechanisms for metabolism and recycling of JH when JH levels are very low or very high. As a consequence of this additional complexity, the precise roles of JH in the honey bee stress response are presently unclear, but given the importance of this hormone system, this is certainly an area demanding further study [34].

A potential role of hemal dopamine in stress is suggested by work on Drosophila as dopamine increased heart rate, while mutations impairing dopamine synthesis had the opposite effect [102]. In cockroaches, one study also found dopamine in the CC [177], thus suggesting a neurohormonal role. In the hemolymph dopamine levels have been rarely quantified in honey bees, but Bateson et al. [178], found a decrease in dopamine levels in the head hemolymph of honey bees 30 minutes after a strong vibration stressor. Clearly more work is needed here to explore the possible role of dopamine in endocrine acute stress response.

Neuropeptides are an emerging area of research, and many members of this large family have never been studied. The neurosecretory cells of the CC contain numerous unstudied peptides [89], which would be good candidates as stress neurohormones. Corazonin, DH, AKH, tachykinins and AST-A have been mentioned previously, but many other peptides might play metabolic roles in the honey bee [167]. In addition, looking at the location of neurohormone receptors is important to understand how the stress system operates. More details on the distribution of neuropeptides and their receptors might highlight their targets, the responses they elicit, as well as the feedback loops regulating the system.

As in vertebrates, the immune response can be affected negatively by stressors. This may be a direct effect of stress hormones (biogenic amines and AKH), as shown in some insect species [179,180]. Depending on the context and the stressor characteristics, the immune response has been shown to be boosted by stressful events or stress hormones [181,182]. This can be understood as a way of maintaining immune equilibrium in a harmful environment [179,183]. In honey bees, stress and immune responses do not seem to have been considered together yet, but recently several detrimental synergistic effects of various combinations of stressors suggest a link between them [184,185,186,187]. As proposed by some of the authors of those studies, such a link may be highly relevant to understand the recent decrease of honey bee populations [186,187].

The high level of sociality and the complex system of division of labor are essential characteristics of honey bees [15]. Here, we propose that at least some elements of the stress response may have been adapted in specific ways to contribute to the evolution of division of labor. During its lifetime, an individual honey bee progresses through a succession of specialized behavioral states whose sequence follows an internal developmental program modulated by various social signals (pheromones) emitted from the colony [188,189]. Honey bees typically begin their adult life undertaking in-hive activities such as brood nursing, cleaning and food storing, then guarding the hive entrance against intruders and predators and finally foraging for food sources (mostly pollen and nectar). The transition to foraging is a major event is a honey bee’s life, and corresponds to multiple changes in hormonal activity, brain circuits, and physiology [188,190].

There has been a lot of research on the mechanisms underlying and organizing this division of labor, which have been shown to involve octopamine, JH and vitellogenin [175,188,191,192]. We suggest here that modulation of stress reactivity may be linked to the evolution of task specialization in honey bees. The fact that octopamine is two to four times more abundant in brains of foragers than those of nurse bees [36] may predispose foragers to attain more easily or rapidly a state of higher energy mobilization. Indeed foragers appear to be the colony members most exposed to stressors: foraging is energetically demanding, and exposes honey bees to more adverse environments (e.g., predators, insecticides) than working within the hive [193]. Elevated brain octopamine levels, as a potential result of chronic stressor exposure [114], may prepare honey bees to cope with the higher stress levels caused by foraging, and the hormonal state of a forager bee may resemble that of nurses bees under chronic stress.

Further, chronic stressors applied at the colony level and experimental elevation of brain octopamine levels both accelerate the onset of foraging in the honey bee [14,194,195]. A high brain level of octopamine may also make foragers more sensitive to hunger, which could motivate them to gather food.

The forager’s state and number appear to be as a response to colony stress, and foragers are also the behavioral caste exposed to the greatest stress. Therefore, knowing the molecular pathways and physiological mechanisms that regulate chronic and acute stress responses at the individual level are of great interest for developing strategies to improve the health and longevity of honey bee colonies.