Differential activation of CRF receptor subtypes removes stress-induced memory deficit and anxiety Cedomir Todorovic, 1,2 Jelena Radulovic, 2 Olaf Jahn, 2 Marko Radulovic, 2 Tessi Sherrin, 1 Cathrin Hippel 2 and Joachim Spiess 1,2 1 John A Burns School of Medicine, SNRP2, 651, Ilalo St, Honolulu, Hawaii-96813, USA 2 Molecular Neuroendocrinology Laboratory, Max Planck Institute for Experimental Medicine, Hermann Rein Str. 3, 37075 Goettingen, Germany Keywords: anxiety, C57BL ⁄ 6J mice, corticotropin-releasing factor receptors, fear conditioning, hippocampus, lateral septum Abstract The objective of this study was to investigate the role of corticotropin-releasing factor receptors 1 (CRF 1 ) and 2 (CRF 2 ) in anxiety-like behavior and learning of C57BL ⁄ 6J mice after exposure to a stressful stimulus. When C57BL ⁄ 6J mice were exposed to immobilization (1 h) serving as stressful stimulus, context- and tone-dependent fear conditioning were impaired if the training followed immediately after immobilization. The stress-induced impairment of context-dependent fear conditioning was prevented by specific blockade of CRF 2 of the lateral septum (LS) with anti-sauvagine-30. Immobilization did not only affect conditioned fear, but also enhanced, through CRF 2 of the LS, anxiety-like behavior determined with the elevated plus maze. Recovery from stress-induced anxiety and impairment of context-dependent fear conditioning was observed after 1 h delay of training and required hippocampal CRF 1 , as indicated by the finding that this recovery was prevented by blockade of intrahippocampal CRF 1 . It was concluded that exposure to a stressor initially affected both anxiety-like behavior and contextual conditioned fear through septal CRF 2 , while the later activation of hippocampal CRF 1 resulted in the return to baseline levels of both processes. Intraventricular injection of mouse urocortin 2, a CRF 2 -selective agonist, removed the stress-induced anxiety and learning impairment, but did not reduce the activation of the hypothalamic pituitary adrenal axis indicative of the hormonal stress response. We propose that the enhanced anxiety is the component of the stress response responsible for the memory deficit. Introduction Corticotropin-releasing factor (CRF), a 41-residue neuropeptide (Spiess et al., 1981), mediates many neuroendocrine and behavioral responses to stress (Vale et al., 1981; Koob & Heinrichs, 1999). CRF exhibits its actions through two distinctly distributed, G-protein- coupled CRF receptor subtypes, CRF 1 and CRF 2 (Van Pett et al., 2000). CRF 1 and CRF 2 are differentially involved in the modulation of fear and anxiety formation. Our previous results demonstrated that injection of human ⁄ rat CRF (h ⁄ rCRF) into the dorsal hippocampus (i.h.) enhances conditioned fear by activation of CRF 1 . In contrast, conditioned fear is reduced by h ⁄ rCRF acting through CRF 2 of the lateral septum (LS; Radulovic et al., 1999). Thus, depending on the brain region and the receptor subtype involved, CRF enhances or reduces conditioned fear. Similarly, CRF can be anxiogenic or anxiolytic. Mice lacking either the CRF 1 (Smith et al., 1998; Timpl et al., 1998) or CRF 2 gene (Bale et al., 2000; Kishimoto et al., 2000) display reduced or heightened anxiety-like behavior, respectively, suggesting that CRF 1 mediates while CRF 2 predominantly attenuates anxiety-like behavioral responses. In addition, it has been described that urocortin 2 (Ucn2), a CRF 2 -selective agonist (Reyes et al., 2001), exhibits delayed anxiolytic-like effects in the elevated plus maze (EPM; Valdez et al., 2002). However, other evidence indicates that LS CRF 2 is capable of inducing anxiety-like behavior (Radulovic et al., 1999) and increasing certain defensive behaviors, such as stress- induced freezing (Bakshi et al., 2002). Based on these results, it was hypothesized that during the early phase of the stress response CRF plays a stimulatory role in stress responsiveness through activation of CRF 1 and septal CRF 2 , whereas a delayed activation of non-septal CRF 2 by Ucn2 and possibly urocortin 3 (Ucn3), another CRF 2 -select- ive agonist (Lewis et al., 2001), may participate in reduction of the behavioral responsiveness to stress (Reul & Holsboer, 2002; Bale & Vale, 2004). Our previous analysis of the time courses of stress-induced changes of anxiety-like behavior in the EPM and fear conditioning suggested a complex involvement of CRF 1 and CRF 2 in the stress response. In particular, exposure of Balb ⁄ c mice to 1-h immobilization results in an immediate LS CRF 2 -mediated increase of anxiety measures in the EPM, whereas hippocampal CRF 1 -mediated enhancement of condi- tioned fear occurs when training is delayed by 3 h after immobiliza- tion (Radulovic et al., 1999). These observations raise two important issues. Firstly, they disagree, at least on the level of CRF-mediated regulation of anxiety-like behavior and conditioned fear, with the general hypothesis that CRF 1 and CRF 2 act in an antagonistic manner, such that CRF 1 initially activates and CRF 2 later attenuates the stress response (Reul & Holsboer, 2002; Bale & Vale, 2004). Secondly, they raise the question whether the initial anxiety response was responsible for the subsequent modulation of conditioned fear, or whether these responses occurred independently from each other (Davis, 1998). Correspondence: Dr C. Todorovic, as above. 1 E-mail: cedomir@hawaii.edu Received 5 February 2007, revised 27 March 2007, accepted 17 April 2007 European Journal of Neuroscience, Vol. 25, pp. 3385–3397, 2007 doi:10.1111/j.1460-9568.2007.05592.x ª The Authors (2007). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd