Abstract
Uncontrollable stressors produce behavioral changes that do not occur if the organism can exercise behavioral control over the stressor. Previous studies suggest that the behavioral consequences of uncontrollable stress depend on hypersensitivity of serotonergic neurons in the dorsal raphe nucleus (DRN), but the mechanisms involved have not been determined. We used ex vivo single-unit recording in rats to test the hypothesis that the effects of uncontrollable stress are produced by desensitization of DRN 5-HT1A autoreceptors. These studies revealed that uncontrollable, but not controllable, tail shock impaired 5-HT1A receptor-mediated inhibition of DRN neuronal firing. Moreover, this effect was observed only at time points when the behavioral effects of uncontrollable stress are present. Furthermore, temporary inactivation of the medial prefrontal cortex with the GABAA receptor agonist muscimol, which eliminates the protective effects of control on behavior, led even controllable stress to now produce functional desensitization of DRN 5-HT1A receptors. Additionally, behavioral immunization, an experience with controllable stress before uncontrollable stress that prevents the behavioral outcomes of uncontrollable stress, also blocked functional desensitization of DRN 5-HT1A receptors by uncontrollable stress. Last, Western blot analysis revealed that uncontrollable stress leads to desensitization rather than downregulation of DRN 5-HT1A receptors. Thus, treatments that prevent controllable stress from being protective led to desensitization of 5-HT1A receptors, whereas treatments that block the behavioral effects of uncontrollable stress also blocked 5-HT1A receptor desensitization. These data suggest that uncontrollable stressors produce a desensitization of DRN 5-HT1A autoreceptors and that this desensitization is responsible for the behavioral consequences of uncontrollable stress.
Introduction
The degree of behavioral control that an organism has over a stressor critically determines the behavioral and neurochemical consequences of that stressor. Behavioral control is typically studied in a paradigm in which one subject can behaviorally terminate tail shocks [escapable shock (ES)], whereas another yoked subject cannot [inescapable shock (IS)]. Many behaviors occur after IS (impaired escape learning, exaggerated anxiety, etc.) that do not occur after ES, despite identical shock delivery (Maier and Watkins, 2005). This phenomenon is termed “learned helplessness” (Maier and Seligman, 1976).
Considerable effort has been devoted to understand why IS and ES produce different behavioral outcomes. Alterations in dorsal raphe nucleus (DRN) serotonergic (5-HT) functioning are clearly involved. IS, relative to ES, intensely activates DRN 5-HT neurons (Grahn et al., 1999), leading to large accumulations of extracellular 5-HT within the DRN (Amat et al., 2005) and projection regions (Bland et al., 2003). This activation sensitizes DRN 5-HT neurons for 24–72 h. Now, after IS, stimulation of 5-HT neurons, as occurs during behavioral testing, releases exaggerated amounts of 5-HT in projection regions (Amat et al., 1998). Finally, exaggerated 5-HT in projection regions appears responsible for IS behaviors because (1) inhibition of DRN activation during either behavioral testing or IS blocks and prevents, respectively, the expression of IS-induced behaviors (Maier et al., 1995a), (2) blocking 5-HT receptors in projection regions prevents IS-induced behaviors (Christianson et al., 2010), and (3) pharmacological activation of DRN 5-HT neurons in the absence of IS produces IS-induced behaviors (Maier et al., 1995b).
Clearly, sensitization of DRN 5-HT neurons is essential in this process. However, the mechanism(s) by which IS sensitizes these neurons is unknown. One possibility is that IS desensitizes DRN inhibitory autoreceptors. The 5-HT1A receptor (5-HT1A-R) is a likely candidate. In the DRN, 5-HT1A-Rs are somatodendritically expressed and, during activation, inhibit 5-HT cell firing (Sprouse and Aghajanian, 1987) and release (Bonvento et al., 1992). Indeed, reduced 5-HT1A-R expression is associated with excessive 5-HT release in DRN projection regions during stimulation (Richardson-Jones et al., 2010). Furthermore, acute administration of a 5-HT1A-R agonist can desensitize 5-HT1A-Rs (Beer et al., 1990). As noted above, IS produces large and prolonged elevations of extracellular 5-HT within the DRN. Thus, the conditions for 5-HT1A-R desensitization are present.
Here we investigate whether stressor controllability alters DRN 5-HT1A-R function and expression. We do this by measuring both single-unit activity of 5-HT cells in midbrain slices and DRN 5-HT1A-R protein expression. We explore causality between DRN 5-HT1A-R desensitization and IS-produced behaviors by using treatments known to (1) eliminate the protective effects of ES and (2) block the behavioral effects of IS. Because inactivation of the medial prefrontal cortex (mPFC) during ES eliminates its protective effects on behavior (Amat et al., 2005), then mPFC inactivation during ES should lead ES to desensitize 5-HT1A-Rs. Conversely, because previous exposure to ES blocks the behavioral effects of IS (Amat et al., 2006), then previous ES should also prevent IS-induced 5-HT1A-R desensitization. The experiments below investigate these hypotheses.
Materials and Methods
Subjects.
Adult male Sprague Dawley rats (Harlan) aged ∼2.5–3.5 months and weighing 275–325 g were used in all experiments. Rats were singly housed and maintained on a 12 h light/dark cycle (lights on at 7:00 A.M. and off at 7:00 P.M.) with food and water provided ad libitum. Rats were allowed at least 7 d to acclimate in the colony room before any surgery or experimentation. All procedures were approved by the Institutional Animal Care and Use Committee of the University of Colorado-Boulder.
Surgery and cannulation.
Rats were anesthetized with isoflurane (Webster Veterinary), and a 26 gauge dual guide cannula (Plastics One), 1 mm center-to-center distance, was implanted. The tips of the cannulae were directed 1 mm above the border of the infralimbic (IL) and prelimbic (PL) subregions of the mPFC [anteroposterior (AP), +2.7 mm; dorsoventral (DV), −3.3 mm from dura mater; mediolateral (ML), 0.5 mm relative to midline] (Paxinos and Watson, 1998). Rats were allowed to recover at least 7 d before experimentation.
Microinjection.
Microinjections were delivered with a David Kopf Instruments model 5000 microinjector. A dual 33 gauge microinjector tip (Plastics One) was connected to two 10 μl microsyringes with polyethylene 50 tubing (BD Biosciences). Rats were gently wrapped in a towel, and the microinjector tip was inserted into the cannula guides, extending 1 mm beyond the cannula tip. Microinjections were given over the course of 1 min and remained in the cannula for 2 min to ensure drug diffusion. Rats received bilateral microinjections containing 50 ng of muscimol (Sigma) or 0.9% saline at a volume of 0.5 μl/hemisphere. Microinjections were considered successful if, after removal from the guide cannula, fluid was readily dispensable from injector tips.
Stressor controllability.
Rats were placed in clear Plexiglas boxes (11 × 14 × 17 cm) containing a wheel at the front and a Plexiglas rod extending from the rear. The rats' tails were taped to the Plexiglas rods, and two copper electrodes were affixed to the tail, augmented with electrode paste. Electric shocks were delivered to the rats with a precision-regulated animal shocker and Graphic State 3.0 software (Coulbourn Instruments). The shock protocol consisted of 80 shocks separated by an average intertrial interval of 60 s. Shock intensity increased every 30 min (1.0, 1.3, and 1.6 mA) to maintain reliable escape behavior. Shocks were given in yoked rat pairs (ES and IS); therefore, the shock terminated for both ES and IS rats when the ES rat performed the required operant escape response. The required response at the beginning of the stress session was a one-quarter turn of the wheel. The response requirement increased to a one-half wheel turn after three consecutive trials of one-quarter wheel turns that were performed within 5 s. Subsequently, the response requirement increased by 50% provided the previous response requirement was performed within 5 s. The maximum response requirement was four full wheel turns. Notably, the only difference between shock groups was that ES rats controlled the termination of shock. Rats that received no stress were left in their home cages as home-cage (HC) controls. Rats used in the time course and the 5-HT1A-R protein expression experiments received 100 tail shocks at an intensity of 1.0 mA in a Plexiglas restraint tube (17.5 cm in length and 7 cm in diameter) with an average intertrial interval of 60 s. Restraint tubes, rather than wheel-turn boxes, were used for tail shock in the time course and protein expression experiments because stressor controllability was not manipulated.
Rats in the “behavioral immunization” experiment were assigned to one of three treatment groups. The behavioral immunization group (ES/IS) received ES on day 1 as described above and 24 h later on day 2 received IS. IS on day 2 consisted of 100 tail shocks in a Plexiglas restraint tube (17.5 cm in length and 7 cm in diameter) at an intensity of 1.0 mA with an average intertrial interval of 60 s. The IS treatment group (HC/IS) received HC treatment on day 1 and IS in a restraint tube on day 2. Last, an unstressed group (HC/HC) remained in their home cages on days 1 and 2. A treatment group that received IS on both day 1 and day 2 (IS/IS group) was not included because previous comparisons with the HC/IS group revealed no differences (Amat et al., 2006). In all experiments, midbrain slices were taken 24 h after the last shock treatment.
Brain slice preparation.
Under sodium pentobarbital (60 mg/kg, i.p.), rats were decapitated and the brain was rapidly removed, blocked coronally, rostral to the DRN, and affixed to a vibratome stage (DSK Microslicer; DTK-1000; Dosaka) with cyanoacrylate glue. Brains were immersed in cold (4°C) artificial CSF (aCSF) containing the following (in mm): 124 NaCl, 3.25 KCl, 2.4 MgSO4, 2 CaCl2, 1.25 KH2PO4, 10 d-glucose, and 26 NaHCO3 (bubbled with 95% O2/5% CO2 during slicing). The time from decapitation to immersion in aCSF was never >150 s. Midbrain slices (450 μm) containing the DRN were collected starting caudally at approximately −8.76 mm from bregma (Paxinos and Watson, 1998). Three consecutive sections of the DRN were taken, with the most rostral slice ending at approximately −7.80 mm from bregma (Paxinos and Watson, 1998). Midbrain slices were placed in a Petri dish containing aCSF [room temperature (RT)] bubbled with 95% O2/5% CO2 and allowed to equilibrate for at least 1 h. After equilibration, midbrain slices were placed on a sloping liquid-gas interface perfusion chamber for recording. The slope was lined with a double layer of lens tissue paper underneath the slice and tissue paper flanking the four sides of the slice to maintain hydration. Slices in the chamber were heated at 35 ± 1°C and perfused with oxygenated aCSF containing 3 μm phenylephrine hydrochloride (an α1 adrenergic receptor agonist) at a flow rate of 750 μl/min. Phenylephrine hydrochloride was added to increase 5-HT neuronal firing rates to levels observed in vivo (Vandermaelen and Aghajanian, 1983) because noradrenergic afferents to the DRN are severed during slicing. Because the strategy of the experiments below was to assess feedback inhibition of baseline 5-HT cell firing rates, restoration of spontaneous 5-HT firing rates with an α1 adrenergic agonist was essential. Additionally, despite the fact that axonal projections of 5-HT neurons are severed in this preparation, neurons in the DRN remain viable and maintain firing properties that are characteristic of in vivo raphe cell activity (Mosko and Jacobs, 1976).
Extracellular recording.
Extracellular recordings were made with borosilicate glass pipettes filled with aCSF. The glass electrode was connected to an alternating current differential preamplifier (1000×) and visualized in a window discriminator. Units were screened for characteristics consistent with a serotonergic phenotype (Vandermaelen and Aghajanian, 1983). All cells were preferentially sampled from the mid-rostrocaudal to caudal (approximately −7.80 to −8.30 mm, relative to bregma) dorsomedial DRN, because this region has several efferents to anxiety- and fear-related structures (Lowry et al., 2008). Once isolated, activity of a unit was recorded with Spike2 software (version 5.05; Cambridge Electronics Design) for 5 min to assess the baseline firing rate. Slices were then perfused with 50 μm 5-HT for 2 min. Units that were reversibly inhibited by 5-HT and expressed characteristics consistent with the 5-HT cell phenotype, such as long-duration biphasic or triphasic action potentials, a regular firing pattern, and a firing rate approximately ranging from 0.5 to 2.5 Hz (Vandermaelen and Aghajanian, 1983; Burlhis and Aghajanian, 1987), were deemed putative 5-HT cells. After recovery from application of 50 μm 5-HT, a variety of drugs were applied to the slice and changes in firing rate were calculated.
Drugs.
For mPFC microinjections, muscimol (Sigma) was dissolved in 0.9% saline according to the required dose. The 5-HT1A-R antagonist WAY 100635 (N-[2-[4-(2-methoxyphenyl)-1-piperazinyl]ethyl]-N-(2-pyridinyl) cyclohexanecarboxamide trihydrochloride) and 5-hydroxytryptamine hydrochloride were purchased from Sigma and aliquoted with aCSF. The 5-HT1A-R agonist ipsapirone (2-[4-[4-(2-pyrimidinyl)-1-piperazinyl]butyl]-1,2-benzis othiazol-3(2H)-one-1,1-dioxide) was purchased from Tocris Bioscience and aliquoted in dimethylsulfoxide (Calbiochem). Importantly, ipsapirone is a selective 5-HT1A-R agonist in the DRN but a partial agonist in other brain regions (Glaser et al., 1985; Dong et al., 1997). Unless otherwise noted, all drugs used for ex vivo extracellular recordings were applied for 2 min and dissolved in aCSF containing phenylephrine hydrochloride (Sigma) during slice application.
Analysis of firing rates and responses.
Unless otherwise noted, the dependent measure used throughout all experiments was mean “percentage inhibition.” Percentage inhibition was calculated by determining the mean firing rate during the 2 min before drug application, “mean baseline.” The mean firing rate during drug perfusion was calculated from the time of drug application until maximal inhibition of firing, the “mean drug firing rate.” Therefore, the percentage inhibition was calculated as [1 − (mean drug firing rate/mean baseline)] × 100. Mean percentage inhibition was chosen as the dependent measure because it accounts for the rate at which a cell becomes maximally inhibited. Other dependent measures, such as “mean maximal inhibition” do not account for the rate at which a cell reaches maximal inhibition. All drugs were applied for 2 min unless otherwise noted. Stress-induced alterations of baseline firing rates were analyzed by comparing the mean baseline firing rate 2 min before application of 50 μm 5-HT, because 50 μm 5-HT was applied to all cells as a test of a serotonergic phenotype.
Histology.
To verify cannula location, brains were frozen by placement over dry ice and then stored at −80°C. Coronal slices measuring 40 μm in thickness were taken throughout the frontal cortex and mounted on gelatin-coated glass slides. Sections were stained with cresyl violet and examined for cannula tracts under light microscopy. Rats with cannula tips outside of IL or PL subregions were excluded from analysis.
Western blot.
Rats were killed 24 h after IS or HC treatments. An ES group was not included because 5-HT1A-R-mediated inhibitions were similar between HC and ES groups. After an injection of sodium pentobarbital (60 mg/kg, i.p.), rats were decapitated, and brains were rapidly dissected, placed in ice-cold isopentane, and stored at −80°C. Frozen brains were mounted in a cryostat, and micropunches measuring ∼1 mm3 were taken from the DRN, centered on the dorsomedial DRN (AP, −7.5 to −8.5 mm; DV, −5.4 mm from dura mater; ML, 0.0 mm) and sonicated in a mixture containing extraction buffer (Invitrogen) and protease inhibitors (Sigma). Ice-cold tissue samples were centrifuged at 10,000 rpm for 5 min. The resulting supernatant was removed, and the protein concentration of each sample was quantified using the Bradford method.
Samples were heated to 70°C for 10 min, and protein samples were loaded into a standard polyacrylamide gel (Bis-Tris Gel; Invitrogen). Electrophoresis was performed in 3-(N-morpholino)-propanesulfonic acid running buffer at 150 V for 1.5 h using the XCell SureLock Mini-Cell (Invitrogen). Protein was transferred onto a nitrocellulose membrane (Bio-Rad) for 75 min at 40 V. The membrane was blocked with Odyssey blocking buffer (LI-COR Biosciences) for 1.5 h and incubated with 5-HT1A-R antibody (1:500) (ab44635; Abcam) in blocking buffer overnight at 4°C. The following day, the membrane was washed in PBS containing Tween 20 (0.1%) and then incubated in blocking buffer containing goat anti-rabbit IRDye 800CW secondary antibody (1:10,000) (926-32211; LI-COR Biosciences) for 45 min at RT. Protein expression was visualized using an Odyssey Infrared Imager (Li-COR Biosciences). After imaging of 5-HT1A-R expression, the membrane was stripped and reprobed to assess whether IS altered the expression of the Na-K ATPase (1:5000; ab7671; Abcam), a housekeeping protein. The following day, the membrane was washed in PBS containing Tween 20 (0.1%) and then incubated in blocking buffer containing goat anti-mouse IRDye 800CW secondary antibody (1:10,000) (926-32210; LI-COR Biosciences) for 45 min at RT. Protein expression was quantified using Odyssey imaging software (LI-COR Biosciences), and data were expressed as a ratio of 5-HT1A-R/Na-K ATPase expression, normalized to HC.
Statistics.
All data are expressed as mean ± SEM. In experiments assessing dose–response, stress was a between-subjects factor and drug dose was a within-subjects factor. Percentage inhibitions of firing rate were analyzed as independent observations using an unpaired t test, one-way ANOVA, or two-way ANOVA, as appropriate. Statistically significant main effects were followed by either Dunnett's multiple comparison test or Fisher's protected least significant difference post hoc analysis (two-tailed α = 0.05).
Results
Figure 1 shows cannula placements within the mPFC for Experiment 5 and the region of the DRN where 5-HT cell recordings were preferentially taken throughout all experiments. All cannulated subjects had microinjector tips terminating in either the IL or PL regions of the mPFC.
Experiment 1: effect of stressor controllability on serotonin-mediated inhibition of DRN 5-HT cell firing
Rats received ES, IS, or HC treatment and were killed 24 h later for extracellular single-unit recording. To determine whether IS desensitized 5-HT1A-Rs in the DRN, the endogenous ligand for the 5-HT1A-R, 5-HT, was used. Varying doses of 5-HT (1, 10, 25, 50, and 100 μm) were applied for 2 min, and percentage inhibition was calculated. As shown in Figures 2 and 3, serotonin-mediated inhibition of DRN 5-HT cell firing was impaired after IS but not after ES. This result was confirmed by a two-way ANOVA that revealed significant main effects of stress (F(2,248) = 6.410; p < 0.01) and dose (F(4,248) = 38.530; p < 0.0001) but no significant interaction. Subsequent post hoc comparisons revealed that IS rats were significantly different from ES and HC rats. Importantly, the ES and HC post hoc comparison was not significant. Stress-induced changes of baseline firing rate were analyzed but no significant differences were found (F(2,58) = 3.008; p > 0.05). Indeed, we did not expect to observe stress-induced differences in firing rate because tonic 5-HT1A-R-mediated autoinhibition of 5-HT cell firing rates in an in vitro preparation is dependent on exogenous tryptophan availability (Evans et al., 2008).
Experiment 2: role of the 5-HT1A-R during serotonin-mediated inhibition of DRN 5-HT cell firing
Although 5-HT is the endogenous ligand for 5-HT receptors, the DRN expresses several 5-HT receptor classes (Barnes and Sharp, 1999). To determine the role of the 5-HT1A-R during serotonin-mediated inhibition of DRN cell firing, the 5-HT1A-R antagonist WAY 100635 was used. Slices of the midbrain were taken from experimentally naive rats for extracellular single-unit recording. After 5 min of baseline firing rate collection, 100 μm 5-HT was applied for 2 min, and percentage inhibition was calculated. After recovery from 100 μm 5-HT, 20 nm WAY 100635 was applied to the slice for 16 min. After superfusion with WAY 100635, a mixture of 20 nm WAY 100635 and 100 μm 5-HT was applied to the slice for 2 min, and percentage inhibition was calculated. This protocol has been used by others to assess the contribution of the 5-HT1A-R during serotonin-mediated inhibition in the rat DRN (Fairchild et al., 2003), although in this previous study, recordings were not targeted specifically to the dorsomedial DRN. WAY 100635 blocked serotonin-mediated inhibition of cell firing in the dorsomedial DRN (data not shown). An unpaired t test revealed a significant difference between mean percentage inhibition of the 5-HT applications (t(12) = 3.230; p < 0.01).
Experiment 3: effect of inescapable stress on ipsapirone-mediated inhibition of DRN 5-HT cell firing
Rats received either IS or HC treatment and were killed 24 h later for extracellular single-unit recording. Experiment 3 did not contain an ES group because Experiment 1 revealed no significant differences between ES and HC rats. The 5-HT1A-R agonist ipsapirone was applied at varying doses (5, 50, 100, 150, 250, and 500 nm) for 2 min, and percentage inhibition was calculated. As shown in Figure 4, ipsapirone-mediated inhibition of DRN cell firing was impaired in IS rats at a number of doses. These results were confirmed by a two-way ANOVA. The effect of stress (F(1,161) = 16.32; p < 0.001), dose (F(5,161) = 13.76; p < 0.001), and interaction of stress × dose (F(5,161) = 3.27; p < 0.01) were significant. Subsequent post hoc comparisons indicated that IS was significantly different from HC. Additionally, post hoc analysis of stress × dose indicated that IS and HC were significantly different at the 150 and 250 nm ipsapirone doses. Again, stress had no significant effect on baseline firing rate (t(68) = 0.8031; p > 0.05).
Experiment 4: time course of IS effects on ipsapirone-mediated inhibition of DRN 5-HT cell firing
Rats received IS in a Plexiglas restraint tube or remained in the colony as HCs. Rats were killed immediately after IS (0 h), 1 d later (24 h), or 1 week later (7 d) for extracellular single-unit recording. These time points were chosen because the behavioral effects of IS are observed between 24 and 72 h after IS exposure (Glazer and Weiss, 1976; Grau et al., 1981; Maier, 1990). Additionally, because Experiment 3 revealed a significant difference between IS and HC rats at the 150 nm dose of ipsapirone, only the 150 nm dose was used as it was the lowest dose that produced a maximal difference between IS and HC rats. As shown in Figure 5, ipsapirone-mediated inhibitions were significantly impaired 24 h after IS. A one-way ANOVA revealed (F(3,48) = 3.977; p < 0.05) a significant stress group effect and a subsequent Dunnett's multiple comparison test revealed a significant difference between the IS 24 h time point and HC. The 0 h and 7 d time points were not significantly different from HC. No stress-induced changes in baseline firing rate were observed (F(3,48) = 0.8020; p > 0.05).
Experiment 5: effect of mPFC inactivation during stress on ipsapirone-mediated inhibition of DRN 5-HT cell firing
Rats received intra-mPFC saline or 50 ng of muscimol 45 min before ES, IS, or HC treatment, and 24 h later midbrain slices were taken for single-unit recording. Importantly, intra-mPFC muscimol does not interfere with escape learning during ES (Amat et al., 2005). As shown in Figures 6 and 7, impairment of 5-HT1A-R-mediated inhibition was found in IS rats compared with HC at 150 nm ipsapirone. Intra-mPFC muscimol had no effect on IS subjects. However, intra-mPFC muscimol led ES to now impair ipsapirone-mediated inhibition of firing. A two-way ANOVA revealed a significant effect of stress (F(2,57) = 8.08; p < 0.001) and stress × muscimol interaction (F(2,57) = 4.57; p < 0.05). The effect of muscimol (F(1,57) = 0.26; p > 0.05) was not significant. Subsequent post hoc comparisons revealed a significant difference between muscimol- and saline-ES rats. Importantly, post hoc comparisons also revealed (1) no difference between muscimol–ES and muscimol–IS groups, (2) no difference between saline–ES and saline–HC groups, (3) a significant difference between muscimol–ES and muscimol–HC groups, and (4) a significant difference between IS and HC groups treated with either saline or muscimol. All comparisons between stress groups revealed no significant differences in baseline firing rate.
Experiment 6: the effect of behavioral immunization on ipsapirone-mediated inhibition of DRN 5-HT cell firing
Because previous studies have shown that an experience with ES before IS (so-called “behavioral immunization”) blocks the behavioral and neurochemical consequences of IS (Williams and Maier, 1977; Amat et al., 2006), we tested whether ES before IS would block IS-induced functional desensitization of DRN 5-HT1A-Rs. Rats received behavioral immunization or experimental control treatments (for details, see Materials and Methods), and midbrain slices were taken 24 h later. Again, only the 150 nm dose of ipsapirone was used. As shown in Figures 8 and 9, IS again reduced ipsapirone-mediated inhibition of cell firing, but this reduction was blocked by previous ES. This conclusion was confirmed by a one-way ANOVA (F(2,41) = 4.497; p < 0.05). Subsequent post hoc comparisons revealed that the HC/IS group was significantly different from the ES/IS and HC/HC groups; however, no significant difference was found between the ES/IS and HC/HC groups. Last, stress-induced changes in baseline firing rate were analyzed, and no significant differences were found (F(2,41) = 0.06547; p > 0.05).
Experiment 7: the effect of IS on the expression of 5-HT1A-R protein in the DRN
The mechanism by which IS impairs 5-HT1A-R-mediated inhibitions was explored by assessing whether IS selectively downregulates 5-HT1A-R protein expression in the DRN 24 h after IS and HC. The 24 h time point was chosen because this is when the electrophysiological differences between HC and IS are observed. Figure 10 is a representative Western blot showing protein expression of the 5-HT1A-R and Na-K ATPase. There was no significant difference in 5-HT1A-R protein expression after IS and HC treatments (t(30) = 0.4207; p > 0.05).
Discussion
Behavioral and neurochemical outcomes after ES and IS are often quite different. In several fear- and anxiety-related tests, ES subjects resemble HC controls, whereas IS subjects exhibit a fear/anxiety phenotype (Maier and Watkins, 2005). Similarly, IS induces several neurochemical changes that do not occur after ES (Maswood et al., 1998). Several studies have shown that IS, relative to ES, activates 5-HT neurons in the DRN (Grahn et al., 1999; Amat et al., 2005), resulting in the sensitization of these neurons to subsequent input (Amat et al., 1998; Bland et al., 2003; Christianson et al., 2010). Moreover, a number of studies have shown that this process is necessary and sufficient to produce typical IS behaviors (Maier et al., 1995a,b; Will et al., 2004; Christianson et al., 2008). However, the mechanism(s) responsible for IS-induced DRN 5-HT sensitization is unexplored. The present experiments examined whether IS, relative to ES, reduces 5-HT feedback inhibition on 5-HT cells, an outcome that would sensitize these cells. Sensitization would occur if IS reduced DRN 5-HT1A-R function. However, reduced 5-HT1A-R function would not, by itself, indicate that this consequence of IS causally mediates IS behaviors.
To investigate the possibility that 5-HT1A-R changes are casual, two strategies were adopted. The first was to determine whether a manipulation that eliminates the protective effects of control on behavior, i.e., a manipulation that produces IS-like behavior in an ES rat, would now lead ES to also reduce inhibition of DRN cells, as does IS. The second was to determine whether a manipulation known to block the behavioral effects of IS would also block the effects of IS on 5-HT1A-R inhibition of 5-HT activity. This type of covariation is necessary for implicating a mediational role. Ex vivo extracellular single-unit recording was used because it directly assesses function, and the effects of drug application with multiple doses can be readily examined.
In Experiment 1, we demonstrated that IS, but not ES, impairs serotonin-mediated inhibition of 5-HT cell firing. Although Experiment 1 has limitations because 5-HT is not selective for the 5-HT1A-R, the use of 5-HT rather than a selective 5-HT receptor agonist most closely mimics in vivo 5-HT release. For this reason, Experiment 2 was designed to determine whether 5-HT inhibits DRN 5-HT cell firing primarily via the 5-HT1A-R. Here, a selective 5-HT1A-R antagonist blocked the inhibitory effects of 5-HT on DRN cell firing, suggesting that 5-HT inhibition is exerted primarily via the 5-HT1A-R. Furthermore, Experiment 3 used the 5-HT1A-R agonist ipsapirone to inhibit cell firing. Again, previous IS interfered with the neuronal inhibition. Last, the time course in Experiment 4 demonstrated that impaired 5-HT1A-R-mediated inhibition follows the same time course as the behavioral effects of IS. Together, these experiments indicate that IS, relative to ES, interferes with serotonin-mediated inhibition of DRN neuronal activity 24 h later, an effect likely mediated by reduced 5-HT1A-R function.
However, the occurrence of IS-induced reductions in 5-HT inhibition of DRN activity does not indicate that this change is responsible for the behavioral effects of IS. If this change is causal, then manipulations that block the protective effects of control on behavior should eliminate the lack of effect of ES on DRN 5-HT1A-R sensitivity. Several studies demonstrate that the mPFC inhibits DRN 5-HT activity during ES but not during IS. Thus, inactivation of the mPFC with muscimol during ES eliminates the protective effects of ES; consequently, ES now produces the same behavioral outcomes as does IS (Amat et al., 2005; Rozeske et al., 2009). If reduced 5-HT1A-R sensitivity is causal, then inhibition of the mPFC during stress should lead both IS and ES to now interfere with DRN unit inhibition. Experiment 5 explored exactly this prediction. Muscimol was microinjected at the IL–PL border as in the study by Amat et al. (2005), and intra-mPFC muscimol during stress led ES to reduce ipsapirone-mediated inhibition to the same degree as IS.
A causal role for reduced 5-HT inhibition of DRN neurons would also predict the converse. Namely, manipulations known to prevent the behavioral effects of IS should also eliminate the IS-induced impairment of 5-HT1A-R inhibition. A number of studies (Williams and Maier, 1977; Amat et al., 2006) have shown that previous experience with ES blocks the behavioral effects of subsequent IS, a phenomenon labeled behavioral immunization. As a causal role for reduced inhibition of DRN neurons would predict, previous ES eliminated IS-induced reduction in ipsapirone-mediated inhibition of unit activity. Using a similar strategy, voluntary exercise was shown to also block the behavioral effects of IS by increasing DRN 5-HT1A-R mRNA (Greenwood et al., 2003). These data together support the idea that reduced 5-HT1A-R function is part of the causal network that mediates the behavioral effects of IS.
To establish causality, it would also be desirable to determine whether manipulations that reduce DRN 5-HT1A-R inhibitory function, in the absence of IS treatment, also produce typical IS-induced behaviors. Indeed, acute administration of fluoxetine produces typical IS-induced behaviors, 24 h after injection (Greenwood et al., 2008). Although not measured by Greenwood et al. (2008), acute fluoxetine is known to cause 5-HT1A-R internalization in the DRN (Riad et al., 2004). Finally, it would be desirable to determine whether preventing 5-HT1A-R desensitization during IS blocks the behavioral effects of IS. Essentially, the behavioral immunization experiment above provides such data. In addition, it should be noted that pharmacological blockade of DRN 5-HT activation during IS, which would prevent 5-HT1A-R desensitization because extracellular DRN 5-HT would not rise, blocks the behavioral effects of IS (Maier et al., 1995a).
The present experiments used ex vivo extracellular single-unit recording in the DRN to investigate the relationship between stressor controllability and 5-HT receptor sensitivity. Others have also used this approach to assess the consequences of a stress experience on receptor function (Laaris et al., 1999; Froger et al., 2004). Although it is remarkable that the effects of stressor controllability carried into the slice preparation, it should be noted that single-unit recording has limitations. Although 33–66% of cells in the DRN are serotonergic, extracellular single-unit recording cannot definitively identify cell type. Furthermore, despite the well-documented firing characteristics of 5-HT cells (Vandermaelen and Aghajanian, 1983), these characteristics may be insufficient for serotonergic cell identification (Kirby et al., 2003).
The present experiments also demonstrate that impaired ipsapirone-mediated inhibition 24 h after IS is not attributable to degradation of 5-HT1A-Rs in the DRN because 5-HT1A-R protein expression was similar 24 h after IS and HC. Functional desensitization of 5-HT1A-Rs could occur via multiple mechanisms, including G-protein uncoupling (Flügge, 1995; Laaris et al., 1999; Hensler, 2002), receptor phosphorylation (Lembo and Albert, 1995), reduced K+ channel function (Kobayashi et al., 2006), and reduced G-protein expression (Li et al., 1996). However, the Western blot data presented in Experiment 7 cannot dissociate these possibilities.
One mechanisms by which stress can produce functional desensitization of DRN 5-HT1A-Rs is via glucocorticoid receptor stimulation (Laaris et al., 1995). This is an unlikely mechanism for the stress-group differences observed here, because the ES and IS treatments used cause similar release of glucocorticoids (Maier et al., 1986). However, IS and ES do lead to very different levels of extracellular 5-HT within the DRN (Maswood et al., 1998; Amat et al., 2005), with IS producing large and sustained elevations. Because agonism of the 5-HT1A-R has been shown to desensitize 5-HT1A-Rs (Beer et al., 1990; van Huizen et al., 1993; Harrington et al., 1994; Riad et al., 2004), this would be a likely cause. Furthermore, inhibition of the mPFC with muscimol during the stressor not only leads ES to produce the behavioral outcomes normally produced by only IS but also leads ES to produce the same high levels of extracellular 5-HT within the DRN as does IS (Amat et al., 2005). Conversely, a previous experience with ES not only blocks the behavioral effects of later IS but also prevents the IS-induced increase in extracellular 5-HT within the DRN (Amat et al., 2006). Thus, whether 5-HT1A-R desensitization was or was not produced by a given manipulation in the present studies is perfectly predicted by whether the manipulation does or does not lead to elevated levels of extracellular 5-HT within the DRN.
The functional desensitization of DRN 5-HT1A-Rs after IS resembles clinical findings reporting a relationship between raphe 5-HT1A-R alterations and anxiety- and depression-related psychopathologies. Indeed, patients with depression show reduced binding of the 5-HT1A-R in the raphe (Drevets et al., 1999, 2007; Meltzer et al., 2004). Additionally, reduced binding of 5-HT1A-Rs in the raphe is also observed in social anxiety disorder (Lanzenberger et al., 2007) and panic disorder (Nash et al., 2008). These clinical findings encourage the notion that IS may capitulate some of the endophenotypes associated with depression and anxiety.
Footnotes
This work was supported by National Institutes of Health Grants DA023329 (R.R.R.), MH086539 (C.A.L.), and MH050479 (S.F.M.). We thank J. Christianson and T. Doyle for commentary and technical assistance.
- Correspondence should be addressed to Robert R. Rozeske, Department of Psychology and Neuroscience, Muenzinger Building UCB 345, Boulder, CO 80309-0345. rozeske{at}colorado.edu