(CRFR1) Signaling in Central Amygdala

www.neuropsychopharmacology.org

Traumatic Stress Promotes Hyperalgesia via Corticotropin-Releasing Factor-1 Receptor (CRFR1) Signaling in Central Amygdala

1,7 1,5,7 1 2 1 Christy A Itoga , Emily A Roltsch Hellard , Annie M Whitaker , Yi-Ling Lu , Allyson L Schreiber , 1 1,6 3 ,1,4 Brittni B Baynes , Brandon A Baiamonte , Heather N Richardson and Nicholas W Gilpin*

1 2 Department of Physiology, Louisiana State University Health Science Center, New Orleans, LA, USA; Neuroscience and Behavior Program, 3 University of Massachusetts, Amherst, MA, USA; Department of Psychological and Brain Sciences, University of Massachusetts, Amherst, MA, USA; 4 Neuroscience Center of Excellence, Louisiana State University Health Science Center, New Orleans, LA, USA

Hyperalgesia is an exaggerated response to noxious stimuli produced by peripheral or central plasticity. Stress modifies nociception, and

humans with post-traumatic stress disorder (PTSD) exhibit co-morbid chronic pain and amygdala dysregulation. Predator odor stress

produces hyperalgesia in rodents. Systemic blockade of corticotropin-releasing factor (CRF) type 1 receptors (CRFR1s) reduces stress-

induced thermal hyperalgesia. We hypothesized that CRF-CRFR1 signaling in central amygdala (CeA) mediates stress-induced hyperalgesia

in rats with high stress reactivity. Adult male Wistar rats were exposed to predator odor stress in a conditioned place avoidance paradigm

and indexed for high (Avoiders) and low (Non-Avoiders) avoidance of predator odor-paired context, or were unstressed Controls. Rats

were tested for the latency to withdraw hindpaws from thermal stimuli (Hargreaves test). We used pharmacological, molecular, and

immunohistochemical techniques to assess the role of CRF-CRFR1 signaling in CeA in stress-induced hyperalgesia. Avoiders exhibited

higher CRF peptide levels in CeA that did not appear to be locally synthesized. Intra-CeA CRF infusion mimicked stress-induced

hyperalgesia. Avoiders exhibited thermal hyperalgesia that was reversed by systemic or intra-CeA injection of a CRFR1 antagonist. Finally,

intra-CeA infusion of tetrodotoxin produced thermal hyperalgesia in unstressed rats and blocked the anti-hyperalgesic effect of systemic

CRFR1 antagonist in stressed rats. These data suggest that rats with high stress reactivity exhibit hyperalgesia that is mediated by CRF-

CRFR1 signaling in CeA.

Neuropsychopharmacology (2016) 41, 2463–2472; doi:10.1038/npp.2016.44; published online 13 April 2016

INTRODUCTION in synaptic efficiency/coupling, and changes in neuronal connectivity (Latremoliere and Woolf, 2009). Therefore, Hyperalgesia is an exaggerated and prolonged response to altered pain perception is not contingent on changes in noxious stimuli that can be produced by plasticity at sensitivity to, or even the presence of, a noxious stimulus. peripheral sites (eg, reduced threshold and/or amplified The amygdala, and more specifically the central nucleus nociceptor response to noxious stimuli) and centrally in of the amygdala (CeA), is a convergence point for multiple the spinal cord or brain. Central sensitization is a process pathways carrying nociceptive information into the central whereby nociceptive neurons and circuits exhibit increased nervous system. Peptide-expressing nociceptors synapse on function in response to activity, inflammation, or injury spinal cord lamina I neurons that project upward and through a variety of processes that include changes in recep- synapse onto parabrachial nucleus neurons that project to tor field size, increases in neuronal excitability, increases amygdala (Jasmin et al, 1997). Peptide-deplete nociceptors synapse on spinal cord lamina II/V neurons, which send *Correspondence: Dr NW Gilpin, Department of Physiology, Room ascending projections directly to amygdala (Braz et al, 2005). 7205, 1901 Perdido Street, New Orleans, LA 70112, USA, Tel: +1 504 The cortex also sends polymodal sensory information, 568 6182, Fax: +1 504 568 6158, E-mail: [email protected] 5 including nociceptive information, to CeA via relays in Current address: Department of Neuroscience and Experimental basolateral amygdaloid complex (Shi and Davis, 1999). The Therapeutics, Texas A&M Health Sciences Center, Bryan, TX 77807, convergence of nociceptive information in CeA is particu- USA. 6Current address: Department of Psychology, Southeastern Louisiana larly interesting because CeA is the primary output nucleus University, Hammond, LA 70402, USA. of the amygdala, and sends projections to effector regions 7These authors contributed equally to this work. that initiate physiological and behavioral responses to Received 14 December 2015; revised 14 March 2016; accepted 22 external events, including the response to noxious stimuli March 2016; accepted article preview online 25 March 2016 (Gilpin et al, 2015). Amygdala peptides mediate stress hyperalgesia CA Itoga et al 2464 CeA projections to periacqueductal gray (PAG) gate through the CRF-1 receptor (CRFR1) in CeA coordinates descending pathways that modulate nociceptive afferent the emotional response to stress (Heilig et al, 1994; Regev activity in the dorsal horn of the spinal cord. Electrical et al, 2012; Sakanaka et al, 1986). CRF modulates inhibitory stimulation of CeA in rodents produces analgesic effects that and excitatory transmission in CeA, thereby gating the are blocked either by lidocaine ‘silencing’ of PAG or by activity of CeA projection neurons to downstream effector blocking opioid receptors in PAG, suggesting that CeA-to- regions (Gilpin et al, 2015; Fu and Neugebauer, 2008; Ji et al, PAG projections are critical for the role of CeA signaling in 2013). With respect to nociception, both stress and intra- modulating the nocifensive response (Oliveira and Prado, CeA corticosterone pellet implants produce somatic hyper- 2001). Therefore, under normal physiological conditions, sensitivity by increasing CRF expression in CeA, and these CeA functions as a hub between incoming nociceptive effects are blocked by antisense CRF knockdown in CeA information and nocifensive responses to noxious stimuli. As (Johnson et al, 2015). Previous work by our group showed such, it is reasonable to hypothesize that dysregulation of that predator odor exposure increases thermal nociception CeA function modifies nociceptive processing and/or and this effect is reversed by systemic blockade of CRFR1s responses to noxious stimuli. (Roltsch et al, 2014). Traumatic stress disorders such as PTSD are highly co- The purpose of these experiments was to test the overall morbid with chronic pain (Asmundson and Katz, 2009; Lew hypothesis that animals with high stress reactivity exhibit et al, 2009; Moeller-Bertram et al, 2012; Shipherd et al, 2007), sensitized CRF/CRFR1 signaling in CeA, which leads to and studies report lower experimental pain thresholds and exaggerated nociception, by testing: (1) the effect of stress on acute stress-induced hyperalgesia in PTSD humans (Defrin nociception, (2) the effect of stress on CRF peptide, CRF et al, 2008; Orr et al, 2000); but see Asmundson and Katz mRNA, CRFR1 mRNA, and CRF cell numbers in CeA, the (2009). Our laboratory uses a predator odor stress model in effect of intra-CeA CRF on (3) conditioned place avoidance which rats are divided into groups that exhibit persistently and (4) nociception in naive animals, (5) the effect of high (Avoiders) or low (Non-Avoiders) avoidance of a systemic CRF-CRFR1 antagonism on stress-induced hyper- predator odor-paired context; these animals are then com- algesia in Avoiders, and (6) the effect of CeA CRF-CRFR1 pared with each other and with unstressed Controls for post- antagonism on stress-induced hyperalgesia in Avoiders. Our stress behavioral and neural changes. In our model, predator results confirm our predictions that (1) Avoiders exhibit odor increases anxiety-like behavior and corticosterone in post-stress thermal hyperalgesia, (2) Avoiders have more both Avoiders and Non-Avoiders (Whitaker and Gilpin, CRF in CeA, (3) intra-CeA CRF produces dose-dependent 2015), but animals exhibit measurable and persistent (at least conditioned place avoidance, (4) intra-CeA CRF infusion 6 weeks) differences in their avoidance of a context paired with produces thermal hyperalgesia, (5) systemic CRFR1 antag- predator odor (Edwards et al, 2013). Because one of the major onism reverses stress-induced hyperalgesia in Avoiders in a diagnostic criteria for PTSD in humans (DSM-5) is persistent manner that is contingent on CeA function, and (6) intra- avoidance of stimuli related to traumatic stress, we use CeA CRFR1 antagonism reverses stress-induced hyperalgesia avoidance to index animals for stress reactivity. in Avoiders. Animal models of chronic pain (eg, arthritis) increase the magnitude and duration of CeA neuronal responses to innocuous mechanical stimuli, independent of altered MATERIALS AND METHODS nociceptive input or plasticity in ascending pathways Animals (Neugebauer and Li, 2003). The notion that strictly psychological factors (eg, stress) can produce central sensitization is Adult male Wistar rats (Charles River) weighing ~ 300 g at a relatively newer concept. For example, psychosocial stress start of experiments were pair-housed in groups of two in a produces allodynia in both healthy humans and fibromyalgia humidity- and temperature-controlled (22 °C) vivarium on a patients (Crettaz et al, 2013). Rodents exposed to restraint 12-h light/dark cycle. Rats were acclimated for 1 week before stress, swim stress, single prolonged stress (restraint+swim), start of experiments. Behavioral tests occurred during the unpredictable sound stress, or social defeat stress exhibit dark period. Animals had ad libitum access to food and mechanical and thermal hyperalgesia that may be mediated water throughout experiments. All procedures were ap- by a combination of brain, spinal, and nociceptor mechan- proved by the Institutional Animal Care and Use Committee isms (Dina et al, 2011; Li et al, 2014). That said, stress- of the Louisiana State University Health Sciences Center and induced analgesia is also a robust behavioral phenomenon, were in accordance with the National Institute of Health and there is a large body of basic science literature exploring guidelines. the neurobiology of this effect (eg, see Millan, 2002). Whether stress produces analgesia or hyperalgesia likely Hargreaves Method depends on factors that include, but are not limited to, post- stress time course, number of stress exposures, type and To assess thermal nociception, each hindpaw was stimulated modality of stress, modality of pain stimulus (heat vs cold or using a halogen heat source from an IITC model 309 mechanical stimuli, etc.), gender, and age. Post-stress time Hargreaves apparatus (IITC Life Sciences, Woodland Hills, course may be particularly important since most stress- CA). A 20-second cut-off was employed to prevent tissue induced analgesia experiments test nociception immediately damage in non-responsive subjects. On test days, rats were after or during the stress, whereas here we test nociception allowed 5 min to acclimate to the testing room, then 10 min several days post-stress. to acclimate to the testing apparatus, which consisted of Hypothalamic corticotropin-releasing factor (CRF) drives Plexiglas enclosures on glass floors suspended 30 cm from the neuroendocrine stress response, and CRF signaling tabletop. On each test day, each hindpaw was tested twice in

Neuropsychopharmacology Amygdala peptides mediate stress hyperalgesia CA Itoga et al 2465 an alternating order (ie, left, right, 1 min wait, left, right), and Histological Verification of Intra-CeA Cannulae these scores were averaged to yield a single score for analysis. Rats were deeply anesthetized with isoflurane and bilaterally This average hindpaw withdrawal latency was the thermal injected with Evans Blue dye using the same rate and volume nociception score for each rat (ie, lower scores indicate as drug injections (0.5 μl per side over 2.5 min). Brains were hyperalgesia). removed before being snap-frozen in isopentane (Fisher Scientific) on dry ice. Brains were stored at − 80 °C, then Predator Odor Conditioned Place Aversion coronally sectioned on a cryostat. Brain sections containing CeA (−1.2 mm to − 3.2 mm relative to bregma) were mounted As previously described (Edwards et al, 2013), rats were on slides and stained with cresyl violet before being cover assigned to control or stress groups. The stressed group was slipped. Placement of the Evens Blue dye was visually exposed to predator odor (bobcat urine) or no odor in estimated by a treatment-blind observer using a rat brain atlas contexts that differed on visual and tactile cues, in an (Paxinos and Watson, 2005). Rats that did not have bilaterally unbiased procedure, while the controls were never exposed accurate cannulae were excluded from all data analysis. to predator odor. On day 1, rats were allowed 5 min to freely explore the apparatus. On day 2, rats were placed in one = context with no odor for 15 min. On day 3, rats were placed Experiment 1: Hargreaves testing over time. Rats (N 8) in the other context with predator odor (no odor for were tested for baseline thermal nociception using the Controls) for 15 min. On day 4, rats were allowed 5 min to Hargreaves test, then periodically over 32 days to determine freely explore the apparatus. Avoidance of predator odor- whether repeated testing altered thermal nociception. paired context was quantified as post-conditioning time minus pre-conditioning time in the predator odor-paired Experiment 2: stress effects on thermal nociception. Rats = context. Stressed rats that exhibited 410 s decrease in time (N 23) were tested for baseline thermal nociception using spent in predator odor-paired context were grouped as the Hargreaves test. Upon stabilization of hindpaw with- Avoiders; all other stressed rats were grouped as Non- drawal latencies, rats were tested for conditioned place Avoiders. Controls underwent the same procedure but were aversion (CPA) using predator odor and separated into = = = never exposed to predator odor. Control (N 5), Non-Avoider (N 8), and Avoider (N 10) groups. Rats were tested for thermal nociception 48 h after predator odor exposure, then re-tested for avoidance of Drug Cannula Implantation predator odor-paired chamber 8 days post stress. Rats were anesthetized with isoflurane and mounted in a stereotaxic frame (Stoelting). An incision was made, small Experiment 3: CRF content in CeA. Radioimmunoassay bilateral holes were drilled into the skull, and stainless steel (RIA) was used to measure total CRF peptide levels in CeA of guide cannulae (26 gauge) lowered such that tips were 1 mm rats exposed to predator odor stress. Rats (N = 22) were dorsal to CeA at these coordinates: AP − 2.3, ML +4.0, DV tested for CPA using predator odor and separated into − 7.5. Guide cannulae were secured to skull with metal Control (N = 6), Non-Avoider (N = 4), and Avoider (N = 12) screws and dental cement, the incision closed, and dummy groups. Twelve days post stress, rats were deeply anesthe- cannulae inserted. Rats were monitored for 1 week or until tized with isoflurane and decapitated. Brains were extracted, they resumed normal activity. frozen in isopentane, cut into 500 μm sections, and CeA was ‘punched’ out with a 17-g needle (see Supplemental Material for location of micropunches). CeA CRF total protein Drug Microinfusions content was determined by RIA in extracted samples. Frozen Infusions were administered bilaterally at a rate of 0.2 μl/min tissue punches from the CeA were homogenized in PBS over 2.5 min via injectors (33 ga, stainless steel) that extended containing EDTA, aprotinin, and 12.5% acetic acid. Lyophi- 1 mm beyond the tip of the guide cannulae. Injection lized supernatant was stored at − 80 °C until extraction. cannulae were left in guide cannulae for additional 1 min to During the extraction procedure, lyophilized powder was allow for diffusion. Infusions were delivered via polyethylene acidified with 1% trifluoroacetic acid in water (buffer A) tubing connected to a 10-μl Hamilton syringe. Animals and centrifuged for 10 min at 10 000 r.p.m. and 4 °C. C-18 underwent sham infusions on days preceding infusion to SEP-Columns (Waters Corporation, Milford, MA) were acclimate them to the procedure. washed with 1 ml buffer B (60% acetonitrile+40% buffer A) followed by three washes with 3 ml buffer A. The acidified sample was loaded onto the pre-treated column and washed Statistical Analysis three times with buffer A. The peptides were eluted with 3 ml Data were analyzed using paired t-test, one-way and two-way of buffer B and lyophilized until dry. The residue was stored at RM ANOVAs, and Pearson correlations, except where − 80 °C until analysis. On the day of the assay, the residue was otherwise indicated. In Experiment 4, CRF cell counts failed reconstituted using the buffer provided by a rat-specific RIA the Shapiro–Wilk Normality Test; therefore, data were kit (Phoenix Pharmaceuticals, Burlingame, CA). Values were analyzed using Kruskal–Wallis ANOVA on Ranks. In normalized to total protein measured in the reconstituted Experiment 7, avoidance data were analyzed using trend sample using a BCA protein assay kit (Thermo Scientific). The analysis to test for dose-response drug effect. In the case of range of detection of the assay was 10–1280 pg/ml. significant interaction and/or main effects, pairwise compar- isons were probed with Student–Newman–Keuls (SNK) Experiment 4: intra-CeA CRF-induced conditioned place post hoc analysis. avoidance. Rats (N = 38) in this experiment were all naive

Neuropsychopharmacology Amygdala peptides mediate stress hyperalgesia CA Itoga et al 2466 to predator odor (ie, no stress). Rats were acclimated to 5′-CGCAGCCGCAAATGC-3′ (reverse). The RT2 SYBR handling before surgery. Guide cannulae were bilaterally Green Mastermixes (Qiagen) were used for real-time PCR. implanted 1 mm dorsal to CeA of rats. After 1 week of All reactions were performed on a CFX96 system (Bio-Rad recovery, rats were acclimated to handling in the experi- Laboratories, Hercules, CA). RT-qPCR data were analyzed mental rooms and sham injected for 3 days before testing. using the DDCT method. Target genes were compared with On test day 1, rats were allowed 5 min to freely explore the housekeeping gene 36B4 and normalized to control values. apparatus. Designation of vehicle vs. CRF paired chambers were assigned in an unbiased procedure. On days 2, 4, and 6, Experiment 7: CRF cell counts in medial and lateral CeA all rats received intra-CeA vehicle infusions immediately (CeM and CeL). Rats (N = 31) underwent a CPA proce- before being placed in one chamber. On days 3, 5, and 7, all dure using predator odor and were separated into Control rats received intra-CeA 0.5 μg infusions of one of four CRF (N = 9), Non-Avoider (N = 12), and Avoider (N = 10) groups. doses (0, 0.05, 0.25, and 0.5 μg/side) immediately before Nine days post stress, rats were deeply anesthetized with being placed in the second chamber. On day 8, rats were isoflurane, injected with chloral hydrate (35%, 2 ml), and allowed 5 min to freely explore both chambers and a CPA intracardially perfused with 4% paraformaldehyde/0.1 M score was calculated. Avoidance of CRF-paired context was borate buffer, pH 9.5. Brains were post-fixed in the same quantified as post-conditioning time minus pre-conditioning fixative for 4 h at 4 °C and submerged in the 20% sucrose/ time in the CRF-paired context. Immediately after the final 0.1 M phosphate buffer, pH 7.4 for 48–72 h before being test session, rats were deeply anesthetized with isoflurane, snap-frozen in isopentane on dry ice. Brains were coronally injected with Evans Blue dye (0.5 μl), and decapitated. Brains sectioned at 35 μm on a freezing microtome, and sections were extracted, frozen in isopentane, cut into 50-μm sections, were stored in cryoprotectant (50% 0.1 M phosphate- mounted on glass slides, stained with cresyl violet, cover buffered saline, 30% ethylene glycol, 30% sucrose, and slipped, and visually inspected for cannula placement. Rats 1% polyvinyl pyrrolidone) at − 20 °C until immunolabeling. that did not have bilaterally accurate intra-CeA cannulae Brain sections containing CeA (from − 2.04 mm to placements were excluded from data analysis. − 3.24 mm relative to bregma) were sorted and labeled by rabbit anti-h/rCRF antiserum (1 : 5000, generously provided Experiment 5: intra-CeA CRF effects on thermal nocicep- by Dr Wylie Vale, Salk Institute) and biotin-conjugated goat tion. Rats (N = 9) were tested for baseline thermal nocicep- anti-rabbit antiserum (1 : 200, Vector Laboratories) using a tion using the Hargreaves test. Rats in this experiment were free-floating immunohistochemistry procedure. CRF labeling all naive to predator odor (ie, no stress). Upon stabilization signal was developed with 3,3′-diaminobenzidine (DAB, of hindpaw withdrawal latencies, guide cannulae were Vector Laboratories), and sections were mounted on slides bilaterally implanted 1 mm dorsal to CeA of rats. After for microscopic analysis. 1 week of recovery, rats were re-tested for baseline thermal Five coronal sections containing CeA (each separated by nociception. Rats were tested for the effects of intra-CeA 175 μm) were collected for quantification of CRF-positive cells. CRF on thermal nociception. On day 1 of the infusion For CeM and CeL, a treatment-blind experimenter counted protocol, all rats received intra-CeA vehicle infusions before cells expressing CRF peptide using a Nikon microscope. Cells testing; on day 2, all rats received intra-CeA CRF (0.5 μg) containing labeling throughout the soma and neuronal infusions. Rats that did not have bilaterally accurate intra- processeswithaclearborderaroundthesomawereconsidered CeA cannulae placements were excluded from data analysis. CRF-immunolabeled and counted as described above. Adjust- ing the optical focus (z axis) of the microscope allowed for clearer visualization of soma depth to distinguish CRF-ir Experiment 6: CRF mRNA and CRFR1 mRNA content in somata from ambiguous particles, and of individual cell CeA. Rats (N = 16) were exposed to predator odor stress, borders when cells were clustered (Karanikas et al, 2013). indexed for avoidance, and separated into unstressed Control (N = 5), Non-Avoider (N = 6), and Avoider (N = 5) groups. Forty-eight hours post stress, rats were deeply anesthetized Experiment 8: systemic CRFR1 antagonism and intra-CeA with isoflurane and decapitated. Brains were extracted, tetrodotoxin effects on stress-induced changes in thermal frozen in isopentane, cut into 500-μm sections, and CeA nociception. Rats (N = 14) were tested for baseline thermal was ‘punched’ out with a 17-g needle. Quantitative real-time nociception using the Hargreaves test. Upon stabilization of PCR (qRT-PCR) was used to measure CRF and CRF1R gene hindpaw withdrawal latencies, cannulae were surgically expression in CeA of rats 48 h post stress. Total RNA was implanted 1 mm dorsal to CeA bilaterally and secured to extracted from brain tissue using an RNeasy Plus Universal skull. After 1 week of recovery, rats were re-tested for Mini Kit (Qiagen, Valencia, CA). Total RNA was reverse baseline thermal nociception, underwent a CPA procedure transcribed using TaqMan Reverse Transcription Reagent kit using predator odor and separated into Control (N = 3), (Life Technologies Corporation, Carlsbad, CA). The primer Non-Avoider (N = 5), and Avoider (N = 6) groups. On days concentrations used were 500 nmol. The primer sequences 2–5 post stress, rats were tested for thermal nociception. (Integrated DNA Technologies, Coralville, IA) used in this Each test was preceded by systemic (i.p.) injection of the study were as follows: CRF 5′-TGATCCGCATGGGTGAAG CRFR1 antagonist R121919 (10 mg/2 ml/kg) or vehicle (20% AATACTTCCTC-3′ (forward), 5′-CCCGATAATCTCCATC HBC, 2 ml/kg) 60 min pre-test, and intra-CeA infusion of AGTTTCCTGTTGCTG-3′ (reverse), CRFR1 5′-TCCACTA TTX (10 ng/0.5 μl) or vehicle (aCSF, 0.5 μl) 40 min pre-test. CATCTGAGACCATTCAGTACA-3′ (forward), 5′-CCTGC On day 2 post stress, all rats received systemic vehicle and CACCGGCGCCACCTCTTCCGGA-3′ (reverse), housekeep- intra-CeA vehicle. On days 3 and 4 post stress, all rats ing 36B4 5′-TTCCCACTGGCTGAAAAGGT-3′ (forward), received systemic R121919, then intra-CeA TTX or vehicle,

Neuropsychopharmacology Amygdala peptides mediate stress hyperalgesia CA Itoga et al 2467

Control Avoider Non-Avoider 9 9

8 8

7 7 # * 6 6

5 5 Paw Withdrawal Latency (s) BL 5 9 12 18 22 32 PawLatency Withdrawal (s) Pre-Odor Baseline 48 Hr Post-Odor Vehicle

100 9 100 8 50 7 50 6 0 5 0 4 -50 -50 3

-100 2 -100 2 R2= 0.025 R2=0.445 R = 0.602

48 hr Post-Odor Paw 1 Chamber at 24 hrs (s) p>0.05 (s) Chamber at 8 days p<0.001 Withdrawal Latency (s) Withdrawal Latency p=0.003 Δ in Predator Odor-Paired

-150 0 Δ in Predator Odor-Paired -150 3579 -200 -100 0 100 -100 -50 0 50 Pre-Odor Hindpaw Δ in Predator Odor-Paired Δ in Predator Odor-Paired Withdrawal Latency (s) Chamber at 24 hrs (s) Chamber at 24 hrs (s) Figure 1 Repeated testing does not alter thermal nociception, and predator odor increases thermal nociception in avoiders. (a) Mean ± SEM paw withdrawal latency for experimentally naive rats. Latency remained stable over 32 days, indicating that repeated testing does not lead to sensitization nor habituation. (b) Mean ± SEM paw withdrawal latency of Control (open triangles, solid line), Non-Avoider (open circles, dotted line), and Avoider (solid circle, solid line) rats at baseline and 48 h after odor (Avoider and Non-Avoider) or air (Control) exposure. Scatter plots denote (c) individual rat baseline paw withdrawal latency before odor exposure vs avoidance of predator paired chamber 24 h after odor exposure, (d) avoidance of predator paired chamber 24 h after odor exposure vs paw withdrawal latency 48 h after odor exposure, (e) avoidance of predator paired chamber 24 h after odor exposure vs avoidance of predator odor paired chamber 8 days after odor exposure. *Denotes po0.05 when compared with pre-odor baseline. #Denotes po0.05 when compared with Control and Non-Avoider 48 h after odor exposure. in a counterbalanced order. On day 5 post stress, all rats volume), then diluted into 2-hydroxypropyl-β-cyclodextrin received systemic vehicle and intra-CeA TTX. Rats that did (HBC; Sigma-Aldrich, 20% wt/vol final concentration in not have bilaterally accurate intra-CeA cannulae placements distilled water) and back-titrated with NaOH to pH 4.5. were excluded from data analysis. Tetrodotoxin (TTX, Tocris Bioscience) and CRF (Sigma Aldrich) were solubilized in artificial cerebrospinal fluid Experiment 9: intra-CeA CRFR1 antagonism effects on (aCSF, Tocris Bioscience). stress-induced changes in thermal nociception. Rats (N = 26) were tested for baseline thermal nociception using the Hargreaves test. Upon stabilization of hindpaw with- RESULTS drawal latencies, cannulae were surgically implanted 1 mm Repeated Testing Does Not Alter Thermal Nociception dorsal to CeA bilaterally and secured to skull. After 1 week of recovery, rats were re-tested for baseline thermal nocicep- In Experiment 1, naive rats were tested for thermal tion, then underwent a CPA procedure using predator odor nociception seven times over the course of 32 days. A one- and separated into Control (N = 7), Non-Avoider (N = 11), way RM ANOVA indicated that there were no significant and Avoider (N = 8) groups. On days 2–5 post stress, rats differences across test days, indicative of the stability of were tested for thermal nociception. Each test was preceded responding and a lack of tolerance/sensitization after by intra-CeA infusion of vehicle (20% HBC, 0.5 μl/side) or repeated thermal stimulation (F[6, 55] = 0.937, p = 0.479; one of three R121919 doses (0.0625, 0.125, or 0.25 μg/0.5 μl/ Figure 1a). side) 5 min before the test. All rats received all doses in a counterbalanced order. Rats that did not have bilaterally Predator Odor Produces Persistent Avoidance and accurate intra-CeA cannulae placements were excluded from Increases Thermal Nociception in Avoiders data analysis. In Experiment 2, rats were sorted into Controls, Non- Drugs Avoiders, and Avoiders, tested for thermal nociception 48 h after predator odor exposure, and re-tested for avoidance The CRFR1 antagonist R121919 (generously supplied by 8 days post stress. Forty-eight hours after predator odor Neurocrine) was solubilized first in 1 M HCl (10% final exposure, Avoiders exhibited significantly lower nociceptive

Neuropsychopharmacology Amygdala peptides mediate stress hyperalgesia CA Itoga et al 2468

60 30 * 10 50 * 9 20 8 40 10 7 ** 6 30 0 5 4 20 -10 3 Latency (s) 10 -20 2 Paw Withdrawal

Paired Chamber (s) 1 Δ in IntraΔ in - CRF- CeA CRF (pg/mg protein) CRF (pg/mg 0 -30 0 0 0.05 0.25 0.5 Controls Non-Avoiders Avoiders Vehicle Intra-CeACRF CRF Intra-CeA CRF Dose (ug/side) Figure 2 Predator odor increases CRF in CeA of Avoiders, and Intra-CeA CRF produces CPA and hyperalgesia. (a) Mean ± SEM pg CRF/mg total protein in CeA of CeA of Control (white bar), Non-Avoider (gray bar), and Avoider (black bar) rats. *Denotes po0.05 when compared with Control and Non- Avoider. (b) Mean ± SEM change in time spent in the Intra-CeA CRF-paired chamber. *Denotes po0.05 trend analysis of CRF dose. (c) Mean ± SEM paw withdrawal latency in rats infused in CeA with vehicle or CRF (0.5 μg). *Denotes po0.05 when compared with vehicle infusion.

thresholds (ie, hyperalgesia) than Non-Avoiders and intra-CeA vehicle (saline, 0.5 μl/side) with the other context. Controls. A two-way RM ANOVA yielded a significant A trend analysis revealed a dose-dependent avoidance of the interaction effect (F[2, 21] = 4.124, p = 0.031; Figure 1b). context paired with Intra-CeA CRF infusions (p = 0.017; Post hoc analyses revealed that Avoiders exhibited signifi- Figure 2b). cantly lower post-stress nociceptive thresholds relative to = pre-odor baseline (p 0.006; Figure 1b), and also relative to Intra-CeA CRF Produces Thermal Hyperalgesia post-stress Non-Avoiders and Controls (p = 0.037 and p = 0.023; Figure 1b). There was no difference in nociceptive In Experiment 5, stress-naive rats were tested for thermal thresholds between groups before stress, suggesting that nociception after receiving intra-CeA vehicle and again 24 h higher thermal nociception in Avoiders 48 h post stress was later after receiving intra-CeA CRF (0.5 μg/side) infusions. A elicited by stress (ie, not pre-existing). There was no paired t-test revealed that intra-CeA CRF infusion reduced significant correlation between pre-odor exposure hindpaw nociceptive thresholds (ie, produced hyperalgesia) relative to withdrawal latencies and 24-h post-stress avoidance vehicle in predator odor naive rats (p = 0.009; Figure 2c). (p40.05; Figure 1c), but there was a significant correlation between 24-h post-stress avoidance of predator odor-paired Predator Odor Stress Does Not Increase CRF mRNA, chamber and 48-h post-stress hindpaw withdrawal latency 2 = = CRFR1 mRNA, or CRF-Peptide Labeled Cell Counts (R 0.445; p 0.003; Figure 1d), suggesting that avoidance in CeA by individual rats predicts post-stress thermal sensitivity. Consistent with previous work from our laboratory Experiments 6 and 7 measured CRF mRNA, CRFR1 mRNA, (Edwards et al, 2013), avoidance persisted over time, as and CRF-immunoreactive cell body counts in CeA. CeA evidenced by the significant correlation between avoidance punches were taken from brains harvested 48 h (Experiment of predator odor-paired chamber 24 h and 8 days post odor 6) or 8 days (Experiment 7) post stress. Separate one-way exposure (R2 = 0.626; po0.001; Figure 1e). ANOVAs indicated no effect of stress history on CRF mRNA (F[2, 15] = 3.805, p = 0.44; Figure 3a) or CRFR1 mRNA = = Predator Odor Stress Increases Total CRF Peptide (F[2, 15] 3.805, p 0.85; Figure 3b) in the CeA 48 h post Content in CeA of Avoiders stress. Across all groups, CRF-immunoreactive cell counts were much higher in lateral CeA than in medial CeA (Figure In Experiment 3, rats were sorted into Controls, Non- 3e–f). Because the results of CRF-positive cell counts were Avoiders, and Avoiders, brains were extracted and frozen not normally distributed (Shapiro–Wilk Normality Test, 12 days post stress, tissue was ‘punched’ to isolate CeA lateral CeA: W = 0.88, po0.01; medial CeA: W = 0.86, samples, and RIA was used to determine CeA CRF total po0.001), a Kruskal–Wallis Test on ranks was used, and protein content. There was a significant main effect of stress revealed no effect on CRF-positive cell number in CeA of group on CRF peptide levels, as measured by RIA, in CeA Avoiders 10 days post stress (CeL, H(2) = 3.5, p = 0.17; CeM, 12 days post stress (F[2] = 6.018, p = 0.009; Figure 2a). H(2) = 1, p = 0.60; Figure 3d–f). These data suggest that Avoiders exhibited significantly higher CRF levels than lasting increases in CRF peptide in the CeA of Avoiders may unstressed Controls and Non-Avoiders (p = 0.032 and not be synthesized in the CeA. p = 0.010, respectively). A correlation between Avoidance and CRF levels yielded a weak negative correlation (R = − = Intra-CeA TTX Blocks Anti-Hyperalgesic Effect of 0.306) that did not reach statistical significance (p 0.250). Systemic CRFR1 Antagonist in Avoiders

Intra-CeA CRF Produces Conditioned Place Avoidance In Experiment 8, rats were sorted into Controls, Non- Avoiders, and Avoiders, then tested for thermal nociception In Experiment 4, stress-naive rats received three pairings of 2–5 days post stress. All rats received all combinations of one of four CRF doses (0, 0.05, 0.25, 0.5 μg/0.5 μl/side) with systemic vehicle/R121919+intra-CeA vehicle/TTX across 4 one of two contexts in a CPA chamber, and three pairings of test days. A two-way ANOVA yielded a significant effect of

Neuropsychopharmacology Amygdala peptides mediate stress hyperalgesia CA Itoga et al 2469

1.4 1.4 1.2 1.2

1 1 CeL 0.8 0.8 CeM 0.6 0.6 CeC 0.4 0.4 change from control) change from control) 0.2 0.2

CRF/Cyp mRNA (fold BLA 0 (fold CRFR1/Cyp mRNA 0 Control Non-Avoider Avoider Control Non-Avoider Avoider

4 80 80 3 60 60

40 40 2

CeM section 1 CeA section 20 20 CeL section CRF-ir cells per CRF-ir cells CRF-ir cells per CRF-ir cells per CRF-ir cells 0 0 0 Control Non-Avoider Avoider Control Non-Avoider Avoider Control Non-Avoiders Avoiders Figure 3 Predator odor does not increase CRF mRNA, CRFR1 mRNA, or CRF-positive cell counts. (a) Mean ± SEM fold change of CRF mRNA level in CeA of Control (white bar), Non-Avoider (gray bar), and Avoider (black bar) rats. (b) Mean ± SEM fold change of CRFR1 mRNA level in CeA of Control (white bar), Non-Avoider (gray bar), and Avoider (black bar) rats. (c) Photomicrograph shows representative CRF labeling in CeA under × 5 objective. Scale bar, 200 μm. White box indicates area of inset figure. The inset shows the individual CRF-ir cells under × 100 objective. Scale bar, 20 μm. (d) Mean ± SEM CRF cell count per CeA section in Control (white bar), Non-Avoider (gray bar), and Avoider (black bar) rats. (e) Mean ± SEM CRF cell count per CeL section in Control (white bar), Non-Avoider (gray bar), and Avoider (black bar) rats. (f) Mean ± SEM CRF cell count per CeM section in Control (white bar), Non-Avoider (gray bar), and Avoider (black bar) rats.

10 Control Non-Avoider Avoider These data suggest that intact CeA neurotransmission is 9 necessary for CRFR1 antagonist reversal of stress-induced 8 # thermal hyperalgesia in Avoiders, and that blockade of 7 CeA action potentials is sufficient to produce thermal #$$ 6 * #$ #$ $ hyperalgesia. 5 #$ 4 3 Intra-CeA CRFR1 Antagonist Blocks Hyperalgesia in 2 Avoiders 1 In Experiment 9, rats were sorted into Controls, Non- Paw Withdrawal Latency (s) 0 Vehicle (i.p.) + R121919 (i.p.) + R121919 (i.p.) + Vehicle (i.p.) + Avoiders, and Avoiders, then tested for thermal nociception Vehicle (CeA) Vehicle (CeA) TTX (CeA) TTX (CeA) 2 − 5 days post stress. All rats received all four doses of Intra- Figure 4 Intra-CeA tetrodotoxin blocks anti-hyperalgesia effects of CeA R121919 (0, 0.0625, 0.125, or 0.25 μg/0.5 μl/side) across systemic R121919. Mean ± SEM percent baseline paw withdrawal latency 4 test days. A two-way RM ANOVA yielded a significant of Control (white bars), Non-Avoider (gray bars), and Avoider (black bars) effect of stress group (F[2, 103] = 12.51, po0.001) and a rats with vehicle or R121919 systemic treatment and vehicle or = o stress group × drug treatment interaction effect (F[6, 51] tetrodotoxin infusion into CeA. *Denotes p 0.05 when compared with o Control and Non-Avoider at the same treatment. #Denotes po0.05 when 6.90, p 0.001; Figure 5). Post hoc analyses revealed that compared with vehicle/vehicle treatment within a group. $Denotes po0.05 Avoiders exhibited significantly lower nociceptive thresholds when compared with R121919/vehicle treatment within a group. than Controls and Non-Avoiders under vehicle and low dose R121919 conditions (po0.001 in all cases), but not after infusion with middle and high R121919 doses. Furthermore, drug treatment (F[3, 32] = 11.487, po0.001; Figure 4) and a within Avoiders, the high R121919 dose increased nocicep- group × drug interaction effect (F[6,32] = 2.498, p = 0.043). tive thresholds relative to all three other drug treatment As in Experiment 2, Avoiders exhibited significantly lower conditions (po0.05 for all conditions). Within Avoiders, the nociceptive thresholds than Controls and Non-Avoiders middle R121919 dose also increased nociceptive thresholds (p = 0.034 and p = 0.032, respectively) 48 h post stress under relative to vehicle and low dose conditions (po0.01 for vehicle (i.p.)+vehicle (CeA) conditions. Systemic R121919 both). These data suggest that CRFR1 antagonism in CeA is normalized thermal nociception in Avoiders, as evidenced by sufficient to reverse stress-induced thermal hyperalgesia in a significant increase in hindpaw withdrawal latency relative Avoiders. to vehicle (p = 0.011). Regardless of whether animals were injected systemically with R121919 or vehicle, intra-CeA DISCUSSION TTX significantly reduced hindpaw withdrawal latencies (ie, produced hyperalgesia) in Avoiders, Non-Avoiders, and Using a predator odor conditioning model that allows unstressed Controls relative to vehicle (po0.05 in all cases). for the identification of rats that exhibit high (Avoiders) or

Neuropsychopharmacology Amygdala peptides mediate stress hyperalgesia CA Itoga et al 2470 Controls Non-Avoiders Avoiders excitatory transmission. We also report that inactivation of 10 9 # CeA with TTX produces hyperalgesia in control rats, and abolishes the ability of a systemic CRFR1 antagonist to 8 # 7 reverse stress-induced hyperalgesia. We interpret this effect 6 * * to be attributable to blockade of action potentials in CeA 5 4 projection neurons (along with upstream CeA interneurons), 3 and this result may suggest that behavioral effects of CeA 2 CRF are attributable to CRF-induced increases in GABAergic 1 transmission, rather than CRF-induced decreases in gluta- 0 Paw Withdrawal Latency (s) 0 0.0625 0.125 0.25 matergic transmission. The direction of our effect is not Intra-CeA R121919 dose (μg) consistent with a past report that bilateral CeA lesion with ibotenic acid reverses footshock-induced hyperalgesia, Figure 5 Inhibition of CRFR1s in CeA blocks predator odor-induced although that study measured tail-flick during footshock increases in thermal nociception in avoiders. Mean ± SEM paw withdrawal latency of Control (white bars), Non-Avoider (gray bars), and Avoider exposure (Crown et al, 2000). The effects of CeA manipula- (black bars) rats infused in CeA with vehicle or one of three R121919 doses tions on nociception are likely dictated by whether these (0.0625, 0.125, and 0.25 μg). *Denotes po0.05 when compared with manipulations affect interneurons or projection neurons or Controls and Non-Avoider rats. #Denotes po0.05 when compared with all both, and also by the extent and laterality of the lesion. For other drug doses within Avoider rats. example, one prior study showed that unilateral optogenetic activation of right CeA promotes nociception in a mouse model of chronic bladder pain (Crock et al, 2012), whereas low (Non-Avoiders) stress reactivity, we show here that other work shows that bilateral CeA lesion blocks anti- Avoiders exhibit thermal hyperalgesia that is predicted by nociception (ie, promotes nociception; Werka, 1994; Werka post-stress avoidance but is not present before the stress and Marek, 1990). In humans with PTSD, neuroimaging event. We also show that Avoiders have higher total CRF studies report higher amygdala activity at rest (Semple et al, peptide content in CeA following predator odor stress 2000), and hyper-reactivity of the amygdala to trauma- relative to Non-Avoiders or unstressed Controls, and that related stimuli (Dickie et al, 2008; Morey et al, 2009; Rauch intra-CeA CRF infusion mimics stress-induced thermal et al, 2000), although those studies lacked the resolution to hyperalgesia. We also report that systemic antagonism of isolate the CeA. CRFR1 reverses stress-induced thermal hyperalgesia in PAG projections to rostral ventromedial medulla (RVM) Avoiders without affecting thermal nociception in Non- initiate descending modulation of incoming pain signals via Avoiders or unstressed Controls, and that this effect is RVM projections to spinal cord (Lau and Vaughan, 2014). contingent on intact CeA neurotransmission. Furthermore, This PAG-RVM-spinal pathway mediates the analgesic inactivation of CeA with TTX produced hyperalgesia in all effects of opioids and cannabinoids, as well as stress- stress groups, regardless of R121919 treatment. Intra-CeA induced analgesia. Here, we show that stress-induced TTX may have produced a floor (ie, maximal) hyperalgesic hyperalgesia is modulated by CRF-CRFR1 signaling in effect on hindpaw withdrawal latency. This floor effect may CeA, a region that projects heavily to PAG (Gray and explain the lack of a stress and/or R121919 effects in animals Magnuson, 1992). Collectively, our data and prior work by treated with intra-CeA TTX. We also found that intra-CeA other laboratories suggest that stress bi-directionally mod- antagonism of CRFR1 reverses stress-induced thermal ulates nociception, and CeA projections to PAG are a strong hyperalgesia in Avoiders without affecting thermal nocicep- candidate pathway for this bi-directional modulation (Butler tion in Non-Avoiders or unstressed Controls. These results and Finn, 2009). implicate CRF-CRFR1 signaling in CeA as a potential Our results suggest that post-stress increases in CRF- mediator of hyperalgesia in humans diagnosed with PTSD. CRFR1 signaling in CeA mediate thermal hyperalgesia in The CeA is uniquely situated to mediate the cross-talk Avoider animals. Because CRF is produced in high quantities between pain and negative affective states. The CeA sends in CeA (McDonald, 1989; Justice et al, 2008) and is also descending GABAergic projections to effector regions that imported from other regions (Uryu et al, 1992), it is not mediate physiological and behavioral responses to noxious known whether local or distal CRF sources (or both) are the stimuli (Neugebauer et al, 2004; Bourgeais et al, 2001). Our critical modulator of CeA projection neurons. Stress can current understanding of CeA circuitry suggests that trigger descending inhibition to induce anti-nociception, in increased activity of GABAergic projections to downstream order to enhance performance under conditions of actual or effector regions would reduce physiological and behavioral impending tissue damage (Millan, 2002). Our current data responses to stress, whereas lower activity of these neurons show that Avoiders exhibit higher total CRF peptide content would disinhibit neuronal firing in effector regions and in CeA, but not increases in CRF mRNA, CRFR1 mRNA, thereby facilitate physiological and behavioral stress and CRF-positive cell counts in CeA, suggesting that stress responses. may increase CRF release into CeA from distal afferents. CRF signaling through CRFR1 increases CeA GABAergic However, unchanged mRNA levels do not necessarily rule transmission by facilitating pre-synaptic GABA release out a role for locally synthesized CRF, as stress may alter (Bourgeais et al, 2001), and also facilitates NMDA mRNA stability. Another alternative explanation is that receptor-mediated EPSCs in laterocapuslar CeA (Fu and Avoiders may exhibit higher CRF content per neuron. Future Neugebauer, 2008; Ji et al, 2013). Therefore, our observation studies will use anatomical and genetic strategies to identify that intra-CeA CRF injection produces hyperalgesia may be the CeA CRF source (local or distal) that is altered by due to CRF-CRFR1 modulation of inhibitory and/or predator odor stress and important for hyperalgesia.

Neuropsychopharmacology Amygdala peptides mediate stress hyperalgesia CA Itoga et al 2471 Interestingly, the CeA receives CRF afferents from various GABA release in central amygdala. Proc Natl Acad Sci USA 105: brain regions (Uryu et al, 1992; Swanson et al, 1983) that 8410–8415. modulate pain sensitivity (Dafny et al, 1996) and/or in which Bale TL, Vale WW (2004). CRF and CRF receptors: role in stress responsivity and other behaviors. Annu Rev Pharmacol Toxicol neurotransmission is altered by painful stimuli (eg, lateral – hypothalamus, dorsal raphe, and bed nucleus of stria 44: 525 557. terminalis). We also report that the CeL contains many Bourgeais L, Gauriau C, Bernard JF (2001). Projections from the nociceptive area of the central nucleus of the amygdala to the more CRF-positive cells than the CeM, but stress did not forebrain: a PHA-L study in the rat. Eur J Neurosci 14: 229–255. affect CRF-positive cell counts in these two sub-nuclei Braz JM, Nassar MA, Wood JN, Basbaum AI (2005). Parallel CRFR1 antagonism inhibits neuronal excitability in the ‘pain’ pathways arise from subpopulations of primary afferent laterocapsular CeA (CeLC), and reduces pain-related beha- nociceptor. Neuron 47: 787–793. vior in rats with arthritic pain (Fu and Neugebauer, 2008). Butler RK, Finn DP (2009). Stress-induced analgesia. Prog Neurobiol Facilitation of NMDA receptor-mediated excitatory trans- 88: 184–202. mission in CeLC is blocked by CRFR1 antagonism but not by Crettaz B, Marziniak M, Willeke P, Young P, Hellhammer D, – CRFR2 antagonism, and also by a PKA but not by a PKC Stumpf A et al (2013). Stress-induced allodynia evidence of inhibitor (Ji et al, 2013), although another study shows that increased pain sensitivity in healthy humans and patients with chronic pain after experimentally induced psychosocial stress. PKCε inhibitor blocks CRF-induced GABA release (Bajo PLoS One 8: e69460. et al, 2008). Crock LW, Kolber BJ, Morgan CD, Sadler KE, Vogt SK, Bruchas MR Increases in CRF-CRFR1 signaling may reflect increases in et al (2012). Central amygdala metabotropic glutamate receptor 5 CRF peptide and/or increases in CRFR1 sensitivity. Here, we in the modulation of visceral pain. J Neurosci 32: 14217–14226. present data to support both of these effects, as Avoiders had Crown ED, King TE, Meagher MW, Grau JW (2000). Shock- higher CRF peptide content in CeA and exhibited higher induced hyperalgesia: III. Role of the bed nucleus of the stria behavioral sensitivity to CRFR1 antagonism. The observed terminalis and amygdaloid nuclei. Behav Neurosci 114: 561–573. increase in CeA CRF appears to be behaviorally relevant as Dafny N, Dong WQ, Prieto-Gomez C, Reyes-Vazquez C, Stanford J, Qiao JT (1996). Lateral hypothalamus: site involved in pain intra-CeA CRF infusions produce conditioned place avoid- – ance and thermal hyperalgesia in otherwise naive rats. modulation. Neuroscience 70: 449 460. Neither systemically administered CRFR1 antagonist nor Defrin R, Ginzburg K, Solomon Z, Polad E, Bloch M, Govezensky M et al (2008). Quantitative testing of pain perception in subjects intra-CeA CRFR1 antagonist affected thermal nociception in with PTSD–implications for the mechanism of the coexistence Non-Avoiders or unstressed Controls, which suggests that between PTSD and chronic pain. Pain 138: 450–459. stress increases tonic CRF-CRFR1 signaling specifically in Dickie EW, Brunet A, Akerib V, Armony JL (2008). An fMRI Avoiders. Our observation that Avoiders exhibit increased investigation of memory encoding in PTSD: influence of sensitivity to a CRFR1 antagonist is consistent with previous symptom severity. Neuropsychologia 46: 1522–1531. reports that animals with arthritic pain exhibit increased Dina OA, Levine JD, Green PG (2011). Enhanced cytokine-induced sensitivity to CRFR1 antagonists, and that CRFR1 antagon- mechanical hyperalgesia in skeletal muscle produced by a novel mechanism in rats exposed to unpredictable sound stress. Eur J ism reduces pain-related behaviors in animals with chronic – pain (Fu and Neugebauer, 2008). In summary, we conclude Pain 15: 796 800. that traumatic stress produces hyperalgesia in some but not Edwards S, Baynes BB, Carmichael CY, Zamora-Martinez ER, Barrus M, Koob GF et al (2013). Traumatic stress reactivity in all animals and humans, and that this effect is mediated by promotes excessive alcohol drinking and alters the balance of increases in CRF-CRFR1 signaling in CeA. prefrontal cortex-amygdala activity. Transl Psychiatry 3: e296. Fu Y, Neugebauer V (2008). Differential mechanisms of CRF1 and CRF2 receptor functions in the amygdala in pain-related synaptic FUNDING AND DISCLOSURE facilitation and behavior. J Neurosci 28: 3861–3876. Gilpin NW, Herman MA, Roberto M (2015). The central amygdala This work was supported by NIH grants AA018400, as an integrative hub for anxiety and alcohol use disorders. AA023305, and AA021013. The authors declare no conflict Biol Psychiatry 77: 859–869. of interest. Gray TS, Magnuson DJ (1992). Peptide immunoreactive neurons in the amygdala and the bed nucleus of the stria terminalis project to the midbrain central gray in the rat. Peptides 13: 451–460. ACKNOWLEDGMENTS Heilig M, Koob GF, Ekman R, Britton KT (1994). Corticotropin- releasing factor and neuropeptide Y: role in emotional integra- We thank Dimitri Grigoriadis of Neurocrine for generous tion. Trends Neurosci 17:80–85. donation of R121919 and Dr Wylie Vale of the Salk Institute Huang MM, Overstreet DH, Knapp DJ, Angel R, Wills TA, Navarro M for his generous donation of the CRF antisera. We thank et al (2010). Corticotropin-releasing factor (CRF) sensitization of Marina Shirokova and Dimitra Gomes for their technical ethanol withdrawal-induced anxiety-like behavior is brain site assistance with these studies. Dr Gilpin is a consultant for specific and mediated by CRF-1 receptors: relation to stress- Glauser Life Sciences. induced sensitization. J Pharmacol Exp Ther 332:298–307. Jasmin L, Burkey AR, Card JP, Basbaum AI (1997). Transneuronal labeling of a nociceptive pathway, the spino-(trigemino-)para- – REFERENCES brachio-amygdaloid, in the rat. J Neurosci 17: 3751 3765. Ji G, Fu Y, Adwanikar H, Neugebauer V (2013). Non-pain-related Asmundson GJ, Katz J (2009). Understanding the co-occurrence of CRF1 activation in the amygdala facilitates synaptic transmission anxiety disorders and chronic pain: state-of-the-art. Depress and pain responses. Mol Pain 9:1–16. Anxiety 26: 888–901. Johnson AC, Tran L, Greenwood-Van Meerveld B (2015). Knockdown Bajo M, Cruz MT, Siggins GR, Messing R, Roberto M (2008). of corticotropin-releasing factor in the central amygdala reverses Protein kinase C epsilon mediation of CRF- and ethanol-induced persistent viscerosomatic hyperalgesia. Transl Psychiatry 5: e517.

Neuropsychopharmacology Amygdala peptides mediate stress hyperalgesia CA Itoga et al 2472 Justice NJ, Yuan ZF, Sawchenko PE, Vale W (2008). Type 1 Paxinos G, Watson C (2005). The Rat Brain in Stereotaxic corticotropin-releasing factor receptor expression reported in Coordinates. Elsevier Academic Press: San Diego, CA, USA. BAC transgenic mice: implications for reconciling ligand-receptor Rauch SL, Whalen PJ, Shin LM, McInerney SC, Macklin ML, mismatch in the central corticotropin-releasing factor system. Lasko NB et al (2000). Exaggerated amygdala response to masked J Comp Neurol 511: 479–496. facial stimuli in posttraumatic stress disorder: a functional Karanikas CA, Lu Y-L, Richardson HN (2013). Adolescent drinking MRI study. Biol Psychiatry 47: 769–776. targets corticotropin-releasing factor peptide-labeled cells in the Regev L, Tsoory M, Gil S, Chen A (2012). Site-specific genetic central amygdala of male and female rats. Neuroscience 249:98–105. manipulation of amygdala corticotropin-releasing factor reveals Latremoliere A, Woolf CJ (2009). Central sensitization: a generator its imperative role in mediating behavioral response to challenge. of pain hypersensitivity by central neural plasticity. J Pain 10: Biol Psychiatry 71: 317–326. 895–926. Roltsch EA, Baynes BB, Mayeux JP, Whitaker AM, Baiamonte BA, Lau BK, Vaughan CW (2014). Descending modulation of pain: the Gilpin NW (2014). Predator odor stress alters corticotropin- GABA disinhibition hypothesis of analgesia. Curr Opin Neurobiol releasing factor-1 receptor (CRF1R)-dependent behaviors in rats. 29: 159–164. Neuropharmacology 79:83–89. Lew HL, Otis JD, Tun C, Kerns RD, Clark ME, Cifu DX (2009). Sakanaka M, Shibasaki T, Lederis K (1986). Distribution and efferent Prevalence of chronic pain, posttraumatic stress disorder, and projections of corticotropin-releasing factor-like immunoreactivity persistent postconcussive symptoms in OIF/OEF veterans: in the rat amygdaloid complex. Brain Res 382:213–238. polytrauma clinical triad. J Rehabil Res Dev 46: 697–702. Semple WE, Goyer PF, McCormick R, Donovan B, Muzic RF Jr, Li C, Yang Y, Liu S, Fang H, Zhang Y, Furmanski O et al (2014). Rugle L et al (2000). Higher brain blood flow at amygdala and Stress induces pain transition by potentiation of AMPA receptor lower frontal cortex blood flow in PTSD patients with comorbid phosphorylation. J Neurosci 34: 13737–13746. cocaine and alcohol abuse compared with normals. Psychiatry 63: McDonald AJ (1989). Coexistence of somatostatin with neuropep- 65–74. tide Y, but not with cholecystokinin or vasoactive intestinal Shi C, Davis M (1999). Pain pathways involved in fear conditioning peptide, in neurons of the rat amygdala. Brain Res 500:37–45. measured with fear-potentiated startle: lesion studies. J Neurosci Millan MJ (2002). Descending control of pain. Prog Neurobiol 66: 19: 420–430. 355–474. Shipherd JC, Keyes M, Jovanovic T, Ready DJ, Baltzell D, Worley V Moeller-Bertram T, Keltner J, Strigo IA (2012). Pain and post et al (2007). Veterans seeking treatment for posttraumatic stress traumatic stress disorder - review of clinical and experimental disorder: what about comorbid chronic pain? J Rehabil Res Dev evidence. Neuropharmacology 62: 586–597. 44: 153–166. Morey RA, Dolcos F, Petty CM, Cooper DA, Hayes JP, LaBar KS Swanson LW, Sawchenko PE, Rivier J, Vale WW (1983). Organi- et al (2009). The role of trauma-related distractors on neural zation of ovine corticotropin-releasing factor immunoreactive systems for working memory and emotion processing in cells and fibers in the rat brain: an immunohistochemical study. posttraumatic stress disorder. J Psychiatr Res 43: 809–817. Neuroendocrinology 36: 165–186. Neugebauer V, Li W (2003). Differential sensitization of amygdala Werka T (1994). Post-stress analgesia in rats with partial amygdala neurons to afferent inputs in a model of arthritic pain. lesions. Acta Neurobiol Exp 54: 127–132. J Neurophysiol 89: 716–727. Werka T, Marek P (1990). Post-stress analgesia after lesions to the Neugebauer V, Li W, Bird GC, Han JS (2004). The amygdala and central nucleus of the amygdala in rats. Acta Neurobiol Exp persistent pain. Neuroscientist 10: 221–234. (Wars) 50:13–22. Oliveira MA, Prado WA (2001). Role of PAG in the antinociception Whitaker AM, Gilpin NW (2015). Blunted hypothalamo-pituitary evoked from the medial or central amygdala in rats. Brain Res adrenal axis response to predator odor predicts high stress Bull 54:55–63. reactivity. Physiol Behav 147:16–22. Orr SP, Metzger LJ, Lasko NB, Macklin ML, Peri T, Pitman RK Uryu K, Okumura T, Shibasaki T, Sakanaka M (1992). Fine (2000). De novo conditioning in trauma-exposed individuals with structure and possible origins of nerve fibers with corticotropin- and without posttraumatic stress disorder. J Abnorm Psychol 109: releasing factor-like immunoreactivity in the rat central amygda- 290–298. loid nucleus. Brain Res 577: 175–179.

Supplementary Information accompanies the paper on the Neuropsychopharmacology website (http://www.nature.com/npp)

Neuropsychopharmacology