br Role of CRF receptors in the BNST
Role of CRF1 receptors in the BNST in stress-induced relapse to cocaine seeking The seminal work of Stewart and co-workers  showed that rats that have extinguished their auto-administration of different drugs of abuse relapse to drug-seeking behaviour when exposed to footshock, even after a 4–6 week drug-free period and this relapse depends on the CRF system. The CRF-dependence of stress-induced relapse to drug seeking was shown by administering systemically CP-154,526, a selective CRF1 receptor antagonist or intra cerebroventricularly D-Phe CRF12-41, a non-selective CRF1/CRF2 antagonist . Both antagonists blocked the relapse induced by exposure to footshock. However, the same treatment only attenuated relapse to cocaine seeking induced by a priming dose of cocaine suggesting that different or partially different mechanisms underlie drug priming and stress-induced relapse to drug seeking, consistent with the different INT-777 subcircuits related to these distinct types of relapse initiators . In addition, by infusing D-Phe CRF12-41 directly into the BNST, Erb and Stewart  showed the critical role of this nucleus in stress-induced relapse to drug seeking behaviour. Other groups have added valuable information showing that CRF/CRF1 receptors in the BNST mediate stress-induced relapse to other drugs of abuse as well , , , .
Role of CRF2 receptors in the VTA in stress-induced relapse to cocaine seeking In 2005, the laboratory of Roy Wise  showed that CRF is released in the VTA after exposure to footshock in naïve rats and in cocaine-experienced rats. Furthermore, they showed that this released CRF is responsible for inducing relapse to cocaine seeking due to a CRF-dependent sensitization of VTA glutamate release observed only in cocaine-experienced rats. At the time of these findings, the origin of VTA CRF was not known. Interestingly, it was later shown that CRF in the VTA originates from the BNST, central nucleus of the amygdala (CeA) and the paraventricular nucleus of the hypothalamus . This evidence suggests that the regulation exerted by CRF in the VTA is downstream of the effect of CRF/CRF1 receptors in the BNST. This is also consistent with the circuitry responsible for stress-induced relapse revealed by McFarland and co-workers  by means of the reversible inactivation of different brain nuclei with muscimol/baclofen. These authors suggested that footshock stress activates subregions of the central extended amygdala that, via the VTA, activates motor output circuitry responsible for reinstalling drug-seeking behaviour. In addition, using double labelling and electron microscopy, it has been shown that VTA CRF is present mainly in VTA glutamatergic axon terminals that establish glutamatergic synapses onto VTA neurons; although some symmetric inhibitory synapses were also observed . However, the precise origin of these VTA CRF/Glutamate and CRF/GABA axon terminals is presently unknown. CRF has been shown to coexist with GABA in neurons of the lateral CeA and dorsal part of the BNST  but it is not known whether these GABAergic neurons project out of the extended amygdala. In addition, there is some discrepancy in the literature regarding the putative excitatory innervation from the BNST to the VTA. On one hand, based on electrophysiological and tracing studies, it has been proposed that glutamatergic projections from the BNST to VTA exist , . On the other hand, the group of Daniel Zahm  using a rigorous anatomical approach has performed a semi-quantitative analysis of the multiple origins of VTA glutamatergic innervation. They have shown that about 20% of VTA glutamate originates from the cortex and the remainder from several subcortical nuclei. However, the BNST was not shown to be a relevant subcortical origin of VTA glutamate . More recent evidence from the same group  reveals that the BNST innervates dopaminergic neurons located in the retrorubral field (A8 dopaminergic group) and not in the VTA. Further studies should resolve the apparent contradictions regarding the connectivity between the extended amygdala and the VTA. An important feature of the CRF-dependent sensitization of VTA glutamate release observed in cocaine-experience rats is that it is long-lasting and comparable at either 1 or 21 days after cocaine withdrawal . It is tempting to suggest that this neuroadaptation is associated to what it has been called “the end-stage of addiction”  in which vulnerability to relapse becomes more permanent.