Since its initial identification and characterization 1 , corticotropin-releasing factor (CRF) has been shown to have a major role in coordinating the endocrine, autonomic and behavioural response to stress 2,3 . CRF is released into the periphery from the paraventricular nucleus of the hypothalamus (PVN) and steers the acti- vation of the hypothalamic–pituitary–adrenal (HPA) axis by triggering the release of adrenocorticotropic hormone from the anterior pituitary, which in turn stim- ulates the synthesis and secretion of glucocorticoids from the adrenal gland 1 . CRF also steers the sympathetic stress response by acting on the locus coeruleus (LC), adrenal medulla and peripheral sympathetic nervous system 4,5 . CRF plays a crucial part in coordinating the periph- eral stress-response systems and the central release of noradrenaline in reaction to stressful challenges. However, centrally released CRF also contributes directly to (and may even be necessary for) a behavioural state of anxiety, independently of its effects on the pitui- tary and sympathetic systems 3,6 . In rodents, central over- expression of CRF induces an anxiogenic behavioural phenotype 2,7,8 , whereas suppressing CRF expression has anxiolytic effects in both basal anxiety 9 and stress- induced anxiety 10 . Interestingly, CRF levels are elevated in the CNS of individuals affected by stress-related psy- chiatric diseases such as major depressive disorder 11 and post-traumatic stress disorder (PTSD) 12 , and in some cases are normalized after antidepressant treatment 13,14 . The anxiogenic effects of CRF have been attributed to the activation of CRF receptor type 1 (CRFR1), one of the two G protein-coupled receptors to which CRF binds (BOX 1). CRFR1 blockade in rodents prevented the CRF-induced anxiogenic phenotype 15 , and mice lacking CRFR1 display reduced anxiety-like behaviour as assessed in a wide range of anxiety-related tests 16–18 . These findings have led to a suggested (causative) role for CRFR1 overactivation in stress-related psychopatholo- gies, triggering the development of CRFR1 antagonists as potential next-generation anxiolytics and antidepressants, with varying levels of success (BOX 2). In contrast to CRFR1, the role of CRFR2 activation in mediating anxiety and depression has been less clear, and two prominent theories circulate to explain its role. The most common view is that CRFR2 activation is responsible for ensuring physiological and psychological homeostasis and counteracts the initial stress-response-provoking effects and anxiety-like behaviours induced by CRFR1 acti- vation 3 . This view is primarily based on evidence obtained from Crfr2-knockout mice, which display an increased corticosterone response to stress 19,20 , an anxiogenic phenotype 19,21 and impaired stress recovery 22 (the latter of which was also seen in mice lacking all three urocortins (UCN1–UCN3), the primary ligands for CRFR2 (REF. 23) (BOX 1). The alternative hypothesis of the role of CRFR2 is that CRFR1 and CRFR2 contribute to opposite modes of stress-coping behaviour, with CRFR1 mediating active defensive behaviour (as triggered by controllable stress) and CRFR2 mediating passive coping behaviour and depression-like responses (such as learned helplessness, which is triggered by uncontrollable stress) 24,25 . This hypothesis is primarily based on the observation that intact CRFR2 signalling in the dorsal raphe nucleus 1 Department of Neurobiology, Weizmann Institute of Science, Rehovot 7610001, Israel. 2 Department of Stress Neurobiology and Neurogenetics, Max Planck Institute of Psychiatry, 80804 Munich, Germany. 3 Department of Cognitive Neuroscience, Donders Institute for Brain, Cognition and Behaviour, Radboud University Medical Centre, 6500 HB Nijmegen, The Netherlands. Correspondence to A.C.  alon.chen@weizmann.ac.il doi:10.1038/nrn.2016.94 Published online 02 Sep 2016 Controllable stress A stress paradigm in which exposure to a stressor (usually footshock or tailshock) can either be avoided or escaped. Learned helplessness A paradigm in which exposure to a severe inescapable stressor induces ‘helpless’ behaviour when it becomes possible to escape the stressor. Region-specific roles of the corticotropin-releasing factor– urocortin system in stress Marloes J. A. G. Henckens 1–3 , Jan M. Deussing 2 and Alon Chen 1,2 Abstract | Dysregulation of the corticotropin-releasing factor (CRF)–urocortin (UCN) system has been implicated in stress-related psychopathologies such as depression and anxiety. It has been proposed that CRF–CRF receptor type 1 (CRFR1) signalling promotes the stress response and anxiety-like behaviour, whereas UCNs and CRFR2 activation mediate stress recovery and the restoration of homeostasis. Recent findings, however, provide clear evidence that this view is overly simplistic. Instead, a more complex picture has emerged that suggests that there are brain region- and cell type-specific effects of CRFR signalling that are influenced by the individual’s prior experience and that shape molecular, cellular and ultimately behavioural responses to stressful challenges. NATURE REVIEWS | NEUROSCIENCE ADVANCE ONLINE PUBLICATION | 1 REVIEWS ©2016MacmillanPublishersLimited,partofSpringerNature.Allrightsreserved.