Endocannabinoids Mediate The Effects Of Acute Stress

Thread starter #1

For animals in the wild, survival depends on being able to detect and respond rapidly to danger by switching from risky (e.g. conspicuous courtship) to survival-oriented behaviors. Very little is known about the hormonal or neuroendocrine mechanisms that control the rapid switch in behavioral state that occurs when an animal detects threats or other stressors. Prior studies with rough-skinned newts (Taricha granulosa), an amphibian model, found that stress-induced suppression of male sexual behaviors (amplectic clasping) involves corticosterone (CORT) and that this steroid hormone uses a novel membrane receptor and modulates the responsiveness of medullary neurons in clasp-controlling neural circuits. We provide evidence that this rapid suppression of male sex behaviors, when induced by either acute stress or CORT administration, involves activation of endocannabinoids signaling in the hindbrain. In a series of behavioral studies, administration of a cannabinoid antagonist, AM281, blocked the suppressive effects of exposure to acute stress or an injection of CORT on the performance of clasping behaviors in sexually active males. Similarly, in electrophysiological studies, prior treatment with AM281 blocked CORT-induced suppression of spontaneous neuronal activity and sensory responsiveness of hindbrain neurons in clasp-controlling neural circuits. These data suggest that, in response to acute stress, elevated CORT concentration increases endocannabinoid signaling in the hindbrain and alters sexual behaviors by modulating the excitability of medullary circuits.

ALTHOUGH ANIMALS commonly show suppressed reproductive behaviors in response to external threats, the hormonal and neurochemical mechanisms that regulate this behavioral shift are not well understood. Numerous studies in diverse vertebrates document that concentrations of adrenal steroid hormones, namely cortisol or corticosterone (CORT), increase shortly after exposure to stressful conditions (1, 2), but surprisingly few papers provide experimental evidence supporting the common assumption that these stress-induced increases in plasma CORT regulate an animal's behavioral response to stress. In the rough-skinned newt (Taricha granulosa), courtship clasping behaviors are suppressed by brief confinement stress or an injection of corticosterone (3). Systemic CORT administration in Taricha suppresses courtship clasping behavior within 5—7 min (4) and suppresses the firing of neurons that regulate clasping behavior within 2—5 min (5, 6). This robust behavioral and physiological effect of CORT is context-specific; CORT fails to block clasping or corresponding clasp-controlling neural activity in male newts with prior experience of clasping or pretreatment with vasotocin (7).

Courtship clasping behaviors of male Taricha are robust, stereotyped, and regulated by descending input from medullary circuits (Fig. 1⇓). These behaviors involve the male grasping the female with hind- and forelimbs wrapped around her torso with his ventral surface, including the cloaca, exposed to her dorsum. The male cloaca is consequently stimulated by physical contact from the female. Somatosensory stimulation of the cloaca leads to initiation and maintenance of clasping over the hours before gamete exchange (6, 8). The neural circuitry underlying this behavior in amphibians is generated by spinal circuits that are modulated by descending input from reticulospinal neurons located in the rostromedial medulla (5, 9) (Fig. 1⇓). Electrophysiological experiments demonstrated that CORT suppresses clasping, in part, by modulating the sensory responsiveness of neurons in the clasp-controlling neural circuits in the rostromedial medulla (5, 6, 10) (Fig. 1⇓). Furthermore, CORT-induced changes in neuronal sensory responsiveness are predictive of the suppression of clasping behavior (5, 6).

Prior research in Taricha has also found that, similar to CORT, the administration of cannabinoid (CB) agonists suppresses male clasping behaviors (7, 11). CB receptors in Taricha central nervous system are structurally and pharmacologically similar to mammalian CB type 1 (CB1) receptors (11). Taricha express a functional CB1 receptor with 84% similarity with human and mouse sequence and binds ligands with a rank order affinity consistent with the profile observed for mammalian species (11). In Taricha and other vertebrates, CB1 receptors are expressed in brain regions associated with regulating social behaviors, including medullary neurons in the hindbrain (12, 13, 14) (Fig. 1⇑). Research in rodents has added evidence that endocannabinoids (eCB) are involved in modulating emotional and affective responses to stressful stimuli (15, 16, 17). In slice preparations of rat hypothalamus, the rapid suppressive effects of glucocorticoids on glutamate release involve eCBs (18). Considering these observations, we hypothesized that stress-induced suppression of clasping behavior by CORT requires eCB signaling.

Therefore, in the experiments described below, we investigated two questions: Do eCBs mediate the effects of acute stress or CORT administration on clasping behaviors? If so, do eCBs cause these effects by altering the responses of clasp-controlling medullary neurons to a clasp-triggering sensory stimulus?

Materials and Methods

Sexually active adult newts (Taricha granulosa) were collected during the active breeding season (February—April, 2002—2003) from permanent ponds in Benton County, Oregon. Males were held in community tanks supplied with continuously flowing, aerated and dechlorinated water, maintained under environmental conditions (14 h light/10 h dark; lights on 0600 h; 10 C), and fed chopped earthworms. Males weighed, on average, 12 ± 0.5 g.

Hormones administered
Preliminary dose-response studies determined the most effective doses of CORT (7) and AM281 (data not shown), a commercially available, selective CB1 receptor antagonist. The potency of CORT to suppress male clasping behaviors varies seasonally, being weakest during the height of the breeding season (Coddington unpublished observation). Thus, a higher dose of both CORT and AM281 was necessary for the behavioral studies, conducted in February, compared with the electrophysiology experiments that were performed from March—April. Separate subsets of newts were used in dose-response studies; all animals were used in only one test or experiment. CORT (Sigma-Aldrich Inc., St. Louis, MO) was solubilized in 2% DMSO (dimethylsulfoxide; Sigma-Aldrich Inc.), 8% ethanol, and 90% Amphibian Ringer's. An ip injection was used to deliver 40 μg/newt (2.6 mg/kg) for behavioral studies and 22 μg/newt (1.46 mg/kg) for electrophysiology experiments. The CB1 antagonist, AM281 (Tocris Cookson Inc., Ellisville, MO) was chosen for this study because research has shown that AM281 is a selective antagonist for CB1 type receptors (19), antagonizes CB agonist sedative-like effects on locomotor behavior (20), and is commercially available. AM281 was solubilized in 2% DMSO + 98% amphibian Ringer's solution for ip injection of 0.5 or 5 μg/0.1 ml/newt for behavioral studies. For electrophysiological studies, 100 ng/μl/newt AM281 (0.1% DMSO) was applied directly to the medullary surface by dropping AM281 or vehicle into the IVth ventricle (Fig. 2⇓). This mode of delivery is highly efficient because of the convenient anatomical organization of the newt hindbrain; the cerebellum is very small, which leaves the medullary surface completely exposed, and more importantly, the medulla is very shallow and neuronal cell bodies lie adjacent to the ependymal surface of the IVth ventricle (which is in contrast to mammals).

Behavioral studies
Experimental procedures were approved by the Institutional Animal Care and Use Committee in Oregon and Wyoming. Experiments were performed between 1400 and 2000 h. This testing period was chosen empirically based on preliminary experiments (data not reported) that revealed less variation in robustness of clasping behavior. Males were tested individually in circular tanks (d = 27 cm) containing dechlorinated water maintained at a depth of 6 cm under red lights. To induce standardized, acute stress, males were placed in a clear dry plastic container (8 cm × 3 cm × 3 cm) and tumbled by hand 180° forward and 180° sideways, one turn every 1 s for 10 min. All males were tested within 24—48 h after capture from the wild. Under our care regime, this is an appropriate period of acclimation. RIA confirmed that animals' CORT levels were equivalent to nonstressed males (see section on RIAs) (3). Furthermore, after 24—48 h, acclimation in our care, vehicle-injected males clasped females indicating that males were not stressed at the time of the experiments. A recorder blind to treatments performed behavior analysis.

Effect of AM281 on behavioral response to acute stress.
If stress-induced inhibition of amplectic clasping behavior is mediated by eCBs, we predicted that animals pretreated with a CB1 antagonist, would continue to exhibit clasping behaviors after exposure to acute stress. To test this prediction, sexually active male Taricha received an ip injection of the CB1 antagonist AM281 (5 μg/0.1 ml/newt) or vehicle (VEH) 15 min before a 10-min exposure to the standardized acute stressor (or no stressor). At 8—10 min after exposure, males were individually presented with two sexually active females and the incidence of clasping was recorded for 30 min.

Effect of AM281 on behavioral response to corticosterone.
To determine whether eCB activation is a necessary signal in the CORT-induced suppression of clasping behaviors, sexually active male Taricha received an ip injection of CB1 antagonist AM281 (0.5 or 5 μg/0.1 ml/newt) or VEH 15 min before receiving an injection of CORT (40 μg/0.1 ml/newt) or VEH. Males were individually presented with two sexually active females 8—10 min after the second injection.

RIAs of plasma corticosterone
Effect of AM281 on plasma levels of corticosterone.
It is plausible that peripheral administration of a CB antagonist might block the stress-induced inhibition of clasping by inhibiting the release of endogenous CORT. Evidence from mammals has shown that CBs can regulate systemic CORT by acting on the hypothalamic-hypophyseal-adrenal system at the level of 1) the adrenal gland (interrenal gland in amphibians) or 2) the hypothalamus (18). To test these possibilities, we measured plasma levels of CORT in males from four treatment groups: VEH/stress, VEH/nonstressed, AM281/stressed, and AM281/nonstressed. Furthermore, this paradigm mirrored the experimental design of behavioral studies (see Effects of AM281 on behavioral responses to acute stress). Immediately after the standardized acute stress, males were killed and blood collected from the carotid arteries into tubes coated with a solution of 3% heparin (Sigma-Aldrich). Samples were kept on ice until centrifuged at 12,000 g for 15 min. Plasma was collected and stored at −80 C. RIAs were performed by the Endocrine Services Laboratory (Dr. David Hess, Oregon Health and Science University) with an intraassay % coefficient of variation = 5.7 and interassay % coefficient of variation = 10.6. Plasma was extracted using an organic solvent, ethanol, and for 10-μl samples, the minimum detectable level was 0.61 ng/ml. The CORT antiserum used in both the Moore and Hess laboratories has been validated for newt plasma (3).

Effect of AM281 on neural control of clasping.
We predicted that if eCBs were involved in the neural signaling pathway used by CORT to suppress clasping then administration of a CB1 antagonist would block CORT-induced suppression of medullary neuronal spontaneous firing and responses to cloacal pressure (clasp-generating somatosensory stimuli). To test this prediction, we used nonanesthetized, immobilized male newts and recorded spontaneous activity and neuronal responsiveness of rostromedial medullary neurons to cloacal pressure. AM281 (100 ng/newt) or VEH was administered directly into the IV ventricle 15 min before a systemic injection of CORT (25 μg/newt) or VEH.

Surgery under anesthesia, recovery, and recording methods are as reported elsewhere (5, 10, 21, 22). Newts prepared this way are secreting only basal levels of CORT (10) and are functionally identifiable as nonstressed given their moderate to high level of neural responsiveness to cloacal stimulation. Furthermore, the ability to clasp and the response of medullary neurons to somatosensory stimulation of the cloaca is severely decreased in a stressed male. The locations of neural recording sites are shown in Fig. 2⇑. Most neurons showed low levels of spontaneous activity and responded to stimulation to the forelimb as well as cloaca. Sensory stimuli consisted of a 3-sec pressure applied to the cloaca followed 20 sec later by a 3-sec pressure applied to the dorsal aspect of the left forelimb. Pressure was applied with a stylus attached to a force transducer allowing us to monitor the onset, magnitude, duration, and offset of the stimulus. Each stimulation bout (cloaca then forelimb) was applied every 3 min for the duration of the experiment. Total number of stimulation bouts during pretreatment was 3—4 during CB1-antagonist treatment and up to 14 during CORT treatment. The intensity of the cloacal stimuli would have been sufficient to elicit reflexive clasping behavior in nonimmobilized newts.

Spontaneous activity of medullary neurons is a measure of neuron excitability and was defined as the total number of spontaneous spikes in a 10-sec time window at 90 sec after each stimulation bout. Stimulus-induced responses were recorded as mean peak firing rate (bin 0.2 sec) corresponding to the time of onset and offset of cloacal and forelimb stimulation.

We determined whether drug/hormone (AM/CORT) treatment altered neuron activity using two independent measurements. First, as a means of understanding the direction of change after treatment, we determined whether the activity of the neurons increased, decreased, or did not change after treatment using a stringent objective criterion. A neuron was identified as increased activity if the frequency of firing (spontaneous and in response to sensory input) increased by >50% of its own baseline. Decreased activity was designated as a drop in activity more than 50%. Any neuron that did not fall under these two categories was designated as no change. The selection of an arbitrary threshold of 50% is a standard procedure and allows for identifying major changes in activity patterns while preserving individual neuron variation. Data are presented as a percentage of neurons that increased, decreased, or did not change under a specific drug/hormone treatment regimen. Data were analyzed with χ2 z-tests and are expressed as a percentage of total number of neurons. Second, we determined the magnitude of change to understand how great the actual change was for each neuron within each treatment group. The magnitude of change was determined by calculating the fold change in activity (frequency of firing before compared with after treatment). Given that the data were not normal nor was the variation equal, the median fold change of each treatment group was compared using nonparametric Kruskal-Wallis and Dunn's multiple comparisons tests.

We recorded from multiple neurons within a subject; therefore, we first established that each neuron behaved independently of subject. To test for independence, we used a mixed model analysis of variance, including newt and neuron as covariates, which revealed that neuronal responses were best explained by drug/hormone regime and were independent of newt for all parameters analyzed. For example, 92% of the variation of firing rate observed at onset of cloacal stimulus were explained by drug/hormone treatment (F2,7 = 6.60, P = 0.0245), confirming that neurons recorded from within one newt were responding independently of one another and can be treated as such in the subsequent analyses.


Effect of AM281 on behavioral response to acute stress
The proportion of stressed males observed clasping was significantly affected by drug pretreatment (Kruskal-Wallis test, H = 23.7, P < 0.0001; Fig. 3A⇓). In VEH-injected males, the incidence of clasping was significantly lower in stressed than nonstressed males (Dunn's, P < 0.001). In contrast, stress-induced suppression of clasping was abolished in AM281-injected males (Dunn's multiple comparison among medians, P < 0.001). The behavioral effect of AM281 is not due to effects on plasma CORT concentrations, because basal and stress-induced plasma CORT concentrations did not differ between VEH- and AM281-treated males (F2,42 = 14.9593, P < 0.001; Fig. 3C⇓).

Effect of AM281 on behavioral response to CORT
The proportion of hormone-injected males observed clasping was significantly affected by drug pretreatment (H = 45.6, P < 0.0001; Fig. 3B⇑). In CORT-injected males, the incidence of clasping was higher in males pretreated with AM281 (AM/CORT) compared with VEH/CORT-injected males, especially at the higher dose of AM281 (Dunn's, P < 0.001). In contrast, in males injected with VEH and then CORT (VEH/CORT), the incidence of clasping was significantly lower than in males given two VEH injections (VEH/VEH) (Dunn's, P < 0.0001).

Effect of AM281 on neural control of clasping
Comparison of VEH/CORT and VEH/VEH treatment groups reveals that CORT administration rapidly and robustly suppressed spontaneous and sensory stimulus-induced activity of medullary neurons (Mann-Whitney U test, CORT vs. VEH, Pspont = 0.0248, Pcloaca = 0.0114; Fig. 4A⇓), which is consistent with previous research (5). The suppressive effects of CORT were apparent within 2—6 min and lasted at least 30 min, producing diminished neuronal excitability and responsiveness to clasp-triggering stimuli.

Prior treatment with the CB antagonist, AM281, completely blocked the CORT-induced suppression of spontaneous and stimulus-induced medullary neuron firing (Fig. 4B⇑). Analysis of the direction of change in the percentage of neurons that showed an increase or decrease in spontaneous activity demonstrated that treatment produced pronounced shifts in the overall patterns of excitability and responsiveness to sexually relevant sensory stimulation. A significantly higher percentage of neurons showed an increase in spontaneous activity when treated with AM/CORT (50%; AM281 preceding CORT) compared with VEH/CORT (0%; z = 2.476, P = 0.013; white bars in Fig. 5A⇓).

The magnitude of change is a measure of how great the actual change was for each neuron within each treatment group. The magnitude of the change in spontaneous activity of medullary neurons treated with AM/CORT was significantly higher than VEH/CORT and AM/VEH (H = 15.9, P = 0.0004, Dunn's P < 0.01; Fig. 5C⇑), suggesting the presence of a tonic background influence of eCBs on these neurons.

A significantly higher percentage of neurons showed an increased response to cloacal stimulation when treated with AM/CORT (56.5%) compared with VEH/CORT (0%; z = 2.745, P = 0.006; white bars in Fig. 5B⇑). Furthermore, the magnitude of neuronal response to cloacal stimulation was significantly higher in males treated with AM/CORT than in males treated with VEH/CORT (H = 14.94, P = 0.0006; Figs. 3⇑ and 5D⇑). Overall, these data reveal that blockade of eCB signaling prevents CORT-induced suppression of neuronal activity and sensory responsiveness.


These data support the hypothesis that rapid suppression of consummatory sexual behaviors by exposure to stress or injection of CORT requires the activation of CB1 receptors. Thus, the present findings might indicate a potentially conserved mechanism through which the classic stress steroid, CORT, exerts its diverse neurobehavioral effects. It is known that the patterns and mechanisms of stress responses are conserved among vertebrates. For example, stress typically stimulates the secretion of corticosteroids and favors rapid changes in behavior in many species from fish to mammals (7, 23, 24, 25, 26, 27). Furthermore, a comparison of the hypothalamic-hypophyseal-interrenal axis that exists in adult amphibians with the mammalian hypothalamic-hypophyseal-adrenal axis reveals conservation of hormones, function, and regulation (28, 29), including circadian rhythms of CORT that peak at onset of activity period (30). Studies in rodents suggest that the rapid inhibitory effects of CORT on hormones in the hypothalamopituitary-adrenocortical axis involve the activation of eCB signaling (16, 18, 31). Our findings extend the role of eCBs in rapid responses mediated by CORT beyond the physiology of the hypothalamopituitary-adrenocortical axis to include stress-induced changes in behavior.

Collectively, our data suggest that the mechanism through which stress and CORT suppress sexual performance in males involves eCB signaling. Consistent with previous research (5, 6), CORT administration rapidly and robustly suppressed the clasping reflex and functionally important activity of medullary neurons that control the clasping behavior. Prior treatment with a CB antagonist completely blocked the CORT-induced suppression of clasping behaviors and the corresponding depression of clasp-related functions of medullary neurons. Studies using rodents have shown that administration of CB agonists reduces the performance of sexual behaviors by males, increasing latencies to mount or intromit (32, 33). Neural pathways that control penile erections in mammals (34, 35) and clasping behaviors in Taricha (36) both involve spinal motor pattern generators that are activated by somatosensory stimulation and influenced by descending input from medullary sites. Considering the present data, showing that eCB signaling suppresses the performance of clasping behaviors by acting on medullary neurons, a similar mechanism might be acting when exposure to stress suppresses the performance of consummatory sex-related behaviors of other species.

We have considered and eliminated alternative interpretations of the data. We considered that eCBs might bind membrane CORT receptors in this system because eCBs were found to bind to hippocampal glucocorticoid receptor and to alter their function allosterically (37). However, this was not the case when examined using Taricha brain membrane preparations (38). Alternatively, CORT may act on corticotrophin-releasing neurons located in the paraventricular nucleus of the hypothalamus and by up-regulating retrograde signaling of eCBs remove excitatory glutamatergic input to the releasing hormone (18). If this was occurring in our in vivo behavioral model, then we would expect to observe higher CORT levels in a stressed animal that was pretreated with AM281 compared with a stressed animal that was VEH-pretreated. We did not observe this. Under our experimental paradigm, a male receiving AM281 and then stressed has the same CORT levels as a male receiving VEH and then stressed. These results imply that plasma concentrations of CORT are elevated by stress and AM281, at the doses given in an in vivo behaving animal, does not alter the endocrine response to our stressor. However, pretreatment with a CB antagonist AM281 does alter the behavioral response to CORT elevation. These results are consistent with the notion that CORT might act on behavior by increasing the local release of eCBs, a notion supported by the work of Di et al. (18).

For an individual to mount a successful defense against an acutely stressful event, the brain must coordinate numerous and diverse physiological and behavioral functions. Studies are revealing that eCB signaling plays a critical role in a variety of stress-related responses across diverse brain regions. We suggest that the coordinated release of eCBs might be initiated by the hormonal signal and nongenomic actions of CORT. The results from our study point to a behavioral function performed by eCBs during an acutely stressful event. Another behavioral function for eCBs is the proposed role of lowering anxiety-like behavior in mice performing in plus mazes (39), light-dark box (40), or in vocalization tests (41). A related and important role also performed by eCBs during an acutely stressful event is opioid-independent antinociception (42). Thus, eCBs regulate a variety of stress-related behaviors at distinct locations of the brain: sex behaviors at the level of the hindbrain, nociceptive-induced behaviors at the level of the midbrain, and anxiety-like behaviors at the level of the forebrain. We hypothesize that eCBs might be involved in coordinating multiple physiological and behavioral functions during acutely stressful events. All available evidence shows that eCBs are locally released and signal in a paracrine fashion. Thus, if eCBs are coordinating multiple functions in multiple brain regions then a more global signal such as a hormone would be required to facilitate and coordinate eCB release at multiple sites. In both the Hohmann et al. study and the present work, the physiological (antinociception) and behavioral (blocking clasping) regulation by eCBs occurred within minutes. A likely candidate hormone for triggering this response is CORT. We could find no reports of studies that tested whether CORT might be involved in the nonopioid antinociception observed in the midbrain of rats, but this effect of CORT seems likely and presents a testable hypothesis. Therefore, we hypothesize that diverse functions necessary to mount a coordinated and successful defense during an acutely stressful event might involve eCB signaling at diverse sites and the coordinated release of eCBs be initiated by the hormonal signal and nongenomic actions of CORT.

Although our data support the hypothesis that CORT initiates the release of eCBs, our study does not directly test this hypothesis nor does it elucidate the cellular signaling mechanism by which CORT might initiate the release of CBs. Prior research suggested that eCBs are released from postsynaptic neurons in an activity-dependent manner and, acting as retrograde signals, bind to receptors on presynaptic axon terminals suppressing release of glutamate or GABA (18, 43, 44). Importantly, glutamate release and subsequent excitatory postsynaptic potentials recorded from magnocellular neurons (paraventricular nucleus, hypothalamus) of rats are suppressed rapidly by corticosteroid-induced release of eCBs (18). Our data are consistent with this mechanism given that CORT suppresses neuronal responses to sensory input and this CORT-mediated effect is blocked by a CB1 antagonist.

The current data also supports the hypothesis that eCBs play a key role in the selective regulation of behavioral output. By regulating the activity of eCB signaling, regulation of multiple potential behavioral responses to acute stress can be made in a way that is appropriate for the animal's current context. Behavioral biologists have long known that behavioral responses to environmental stress are context-specific. Given that the state of neural a system will vary with the behavioral state of the animal, it follows that synaptic events mediated by eCB retrograde signaling might contribute to context-specific behaviors. Expression of sex-related behavior by Taricha occurs in a context-specific manner; namely clasping behaviors are not suppressed by acute stress or CORT if the male has had immediate prior experience of clasping or received an injection of a clasp-facilitating peptide, vasotocin (7). We expect that the precise manner in which eCBs affect behavior will vary among species, but we suggest that the mechanism by which activity of eCB signaling controls the expression of context-appropriate behaviors might be a general phenomenon occurring in all vertebrates.

Source, Graphs and Figures: Endocannabinoids Mediate the Effects of Acute Stress and Corticosterone on Sex Behavior