Defining the role of corticotropin releasing factor binding protein in alcohol consumption

Abstract

The corticotropin releasing factor (CRF) exerts its effects by acting on its receptors and on the binding protein (CRFBP), and has been implicated in alcohol use disorder (AUD). Therefore, identification of the exact contribution of each protein that mediates CRF effects is necessary to design effective therapeutic strategies for AUD. A series of in vitro/in vivo experiments across different species were performed to define the biological discrete role of CRFBP in AUD. First, to establish the CRFBP role in receptor signaling, we developed a novel chimeric cell-based assay and showed that CFRBP full length can stably be expressed on the plasma membrane. We discovered that only CRFBP(10 kD) fragment is able to potentiate CRF-intracellular Ca ²⁺ release. We provide evidence that CRHBP gene loss increased ethanol consumption in mice. Then, we demonstrate that selective reduction of CRHBP expression in the center nucleus of the amygdala (CeA) decreases ethanol consumption in ethanol-dependent rats. CRFBP amygdalar downregulation, however, does not attenuate yohimbine-induced ethanol self-administration. This effect was associated with decreased hemodynamic brain activity in the CRFBP-downregulated CeA and increased hemodynamic activity in the caudate putamen during yohimbine administration. Finally, in alcohol-dependent patients, genetic variants related to the CRFBP(10 kD) fragment were associated with greater risk for alcoholism and anxiety, while other genetic variants were associated with reduced risk for anxiety. Taken together, our data provide evidence that CRFBP may possess both inhibitory and excitatory roles and may represent a novel pharmacological target for the treatment of AUD.

OPEN

Citation: Transl Psychiatry (2016) 6, e953; doi:http://dx.doi.org/10.1038/tp.2016.208

Web End =10.1038/tp.2016.208

http://www.nature.com/tp

Web End =www.nature.com/tp

CL Haass-Kofer1,2,3, AT Henry4, G Melkus5, JA Simms6, M Naemmuddin6, CK Nielsen6, AW Lasek7, M Magill3, ML Schwandt8, R Momenan9, CA Hodgkinson10, SE Bartlett11, RM Swift2, A Bonci12 and L Leggio1,3

The corticotropin releasing factor (CRF) exerts its effects by acting on its receptors and on the binding protein (CRFBP), and has been implicated in alcohol use disorder (AUD). Therefore, identication of the exact contribution of each protein that mediates CRF effects is necessary to design effective therapeutic strategies for AUD. A series of in vitro/in vivo experiments across different species were performed to dene the biological discrete role of CRFBP in AUD. First, to establish the CRFBP role in receptor signaling, we developed a novel chimeric cell-based assay and showed that CFRBP full length can stably be expressed on the plasma membrane. We discovered that only CRFBP(10 kD) fragment is able to potentiate CRF-intracellular Ca2+ release. We provide evidence that CRHBP gene loss increased ethanol consumption in mice. Then, we demonstrate that selective reduction of CRHBP expressionin the center nucleus of the amygdala (CeA) decreases ethanol consumption in ethanol-dependent rats. CRFBP amygdalar downregulation, however, does not attenuate yohimbine-induced ethanol self-administration. This effect was associated with decreased hemodynamic brain activity in the CRFBP-downregulated CeA and increased hemodynamic activity in the caudate putamen during yohimbine administration. Finally, in alcohol-dependent patients, genetic variants related to the CRFBP(10 kD) fragment were associated with greater risk for alcoholism and anxiety, while other genetic variants were associated with reduced risk for anxiety. Taken together, our data provide evidence that CRFBP may possess both inhibitory and excitatory roles and may

Translational Psychiatry (2016) 6, e953; doi:http://dx.doi.org/10.1038/tp.2016.208

Web End =10.1038/tp.2016.208 ; published online 15 November 2016

INTRODUCTION

Stress plays an important role in the development and maintenance of alcohol use disorder (AUD).1 In response to stress, corticotropin releasing factor (CRF) activates the hypothalamic pituitary adrenal axis (HPA), leading to the subsequent release of glucocorticoids.2 Additionally, CRF mediates behavioral responses to stress via extrahypothalamic regions. CRF exerts its effects via binding to its receptors (CRFR1 and CRFR2), and to the 37 kD CRF binding protein (CRFBP).3 Compared to CRF and its receptors, CRFBP has been much less investigated. This is due to the lack of control of spontaneous CRFBP proteolytic cleavage that yields a N-terminal fragment (27 kD) and a C-terminal fragment (10 kD), and the difculty of purifying large-enough quantities of CRFBP 37 kD full length (FL) for in vitro and in vivo experiments.4

Early preclinical studies showed that CRFBP is expressed in the cerebral cortex, subcortical limbic structures and anterior pituitary corticotropes, suggesting that CRFBP plays a critical role in modulating endocrine and behavioral responses to stress.5 Recent evidence shows that CRFBP and CRFR2 co-exist in rat ventral tegmental area (VTA) glutamatergic/GABA synaptosomes that originate from hypothalamic areas.6 Using microinjection of the

CRFBP antagonist CRF6 33 in the VTA, it was also demonstrated that CRFBP, via CRFR2, has a pivotal role in escalation of ethanol drinking.7 In line with this recent study, earlier work demonstrated that CRF-mediated cocaine positive reinforcement is associated with CRFBP and CRFR2 interaction in the VTA dopamine system.8 All these data support our hypothesis that CRFBP is not only a sequestering protein, but it may possess additional functions. Our previous electrophysiology work demonstrated that in the VTA CRF modulates neuronal excitability, through CRFBP and CRFR2 interactions, by potentiating N-methyl-D-aspartate-mediated excitatory postsynaptic currents.9 Therefore, elucidation of the selective molecular functional of CRFBP and its interaction with CRFR2 may help us understand the mechanisms underlying the role of CRFBP in alcohol consumption, thus facilitating the potential development of effective treatments targeting the CRFBP in AUD.

MATERIALS AND METHODS Subjects

All animal procedures were approved by the Ernest Gallo and Clinical Center Institutional Animal Care and Use Committee, were in accordance

1Section on Clinical Psychoneuroendocrinology and Neuropsychopharmacology, NIAAA DICBR and NIDA IRP, NIH, Bethesda, MD, USA; 2Center for Alcohol and Addiction Studies, Department of Psychiatry and Human Behavior, Brown University, Providence, RI, USA; 3Center for Alcohol and Addiction Studies, Department of Behavioral and Social Sciences, Brown University, Providence, RI, USA; 4Department of Psychology, University of Chicago, Chicago, IL, USA; 5Department of Radiology, University of Ottawa, Ottawa, ON, Canada;

6Department of Neurology, University of California, San Francisco, San Francisco, CA, USA; 7Department of Psychiatry, University of Illinois, Chicago, IL, USA; 8Ofce of the Clinical Director, NIAAA, NIH, Bethesda, MD, USA; 9Clinical Neuroimaging Research Core, NIAAA, NIH, Bethesda, MD, USA; 10Laboratory of Neurogenetics, NIAAA, NIH, Rockville, MD, USA;

11Queensland University of Technology, Brisbane, QLD, Australia and 12NIDA Intramural Research Program and Departments of Neuroscience and Psychiatry, Johns Hopkins University, Baltimore, MD, USA. Correspondence: Professor CL Haass-Kofer, Center for Alcohol and Addiction Studies, Department of Psychiatry and Human Behavior, Brown University, 121 South Main Street, Providence, RI 02912, USA or Professor L Leggio, Section on Clinical Psychoneuroendocrinology and Neuropsychopharmacology, NIAAA DICBR and NIDA IRP, National Institutes of Health, 10 Center Drive (10CRC/15330), Room 1-5429, Bethesda, MD 20892-1108, USAE-mail: mailto:[email protected]

Web End [email protected] or mailto:[email protected]

Web End [email protected] Received 16 July 2016; accepted 20 July 2016

ORIGINAL ARTICLE

Dening the role of corticotropin releasing factor binding protein in alcohol consumption

represent a novel pharmacological target for the treatment of AUD.

CRFBP and alcohol CL Haass-Kofer et al

2

with NIH guidelines for the Humane Care and Use of Laboratory Animals, and were conducted in agreement with the Guide for the Care and Use of Laboratory Animals, National Research Council, 1996. Human subject data were collected according to screening and assessment NIAAA protocols approved by the appropriate NIH institutional review boards. All participants provided written consent before participating. Subjects description is detailed in the Supplementary information.

ProceduresTransfection and expression of human cells with the CRFBP-CRFRs chimera.

The plasmids containing the CRFBP-CRFRs, detailed in Supplementary information, were transfected into HEK293 cells to verify expression of the protein, conrm appropriate insertion into the membrane, and determine the functionality of the chimera. Prior to transfection, human cells were placed in a six-well plate, with DMEM and 10% fetal bovine serum in each of the wells, and checked for 9095% cell conuency. The media was aspirated from the wells and replaced with fresh DMEM/10% fetal bovine serum. The nal DNA plasmid preparation was then diluted in Opti-MEM reduced serum. Lipofectamine 2000 was mixed in Opti-MEM and incubated at room temperature for 5 min, and then combined with the diluted DNA. Cell selection, ow cytometry, immunouorescence, image analysis, quantication of receptors surface expression and uorescence-based Ca2+ assay are described in the Supplementary information.

Lentivirus expressing shRNA targeting CRHBP. We designed three sets of shRNAs corresponding to different regions on the CRHBP mRNA sequence and cloned DNA oligonucleotides encoding the shRNAs into the lentiviral vector pLL3.7,10,11 which also expresses GFP. pLL3.7 is available from the plasmid repository Addgene (Cambridge, MA, USA, plasmid #11795). The shRNA sequences are designated by the number of the rst base of the 19-nucleotide targeting sequence in rat CRHBP (NCBI NM_139183.2): 1050 (5-GAAACAGCATCCCGGAGTA-3), 937 (5-GAAAATTAGCTGCGACAAT-3)

and 240 (5-GCGCCAATTTGAAGCGGAA-3). As a control, we used a sequence that did not recognize any known mammalian gene in a BLAST search (Scr, 5-GCGCTTAGCTGTAGGATTC-3). Lentivirus was produced using the ViraPower Lentiviral Packaging Mix (Thermo Fisher Scientic, Waltham, MA, USA). Viral titers were determined by p24 enzyme-linked immunosorbent assay (Zeptometrix, Buffalo, NY, USA) and were ~ 8 107 pg ml1 for viruses encoding control and CRHBP shRNAs. The knockdown efciency, intra-amygdala infusion of lentivirus, rats self-administration apparatus and training are described in the Supplementary information.

Drinking in the dark procedure. Male, mixed-genetic-background (C57BL/ 6/SV129 mix) mice were all weaned at P21 and entered in the drinking-inthe-dark (DID) cycle protocol at greater than 6 weeks of age. After a 2-week acclimation period, CRHBP / and CRHBP+/+ littermates were given access to one bottle of 20% (v/v) ethanol for 2 h, for three consecutive days; access began 3 h into the dark cycle. On the fourth consecutive day, mice were given access to one bottle of 20% (v/v) ethanol for 4 h. The procedure was repeated each week, with 3 days between cycles. This DID procedure was used due to the high intake of ethanol on the fourth day of access.12

fMRI protocol. To determine if local knockdown of amygdala CRHBP by RNA interference produces functional differences compared to control rats, we measured the hemodynamic activity related to neural activity by blood-oxygen-level dependence after pharmacological stimuli (yohimbine) using functional magnetic resonance imaging (fMRI). In brief, during the magnetic resonance imaging (MRI) scan, after intraperitoneal (i.p.) saline infusion, all animals received yohimbine (i.p. 2 mg kg1, vol. 5 ml kg1). Data acquisition and post-processing are described in the Supplementary information.

Data analysisDistributional characteristics of outcome measures were examined to evaluate similarity to the normal distribution. The Generalized Estimating Equation (GEE) model was utilized to analyze the drinking behavior of all animals and to compare the hemodynamic activity between CRHBP shRNA rats and Scr shRNA rats. GEE was chosen because it can handle unmeasured dependence when outcomes are repeated measures. We selected GEE over repeated-measures analysis of variance because it can more exibly handle a number of data needs (for example, missing data, assumed auto correlation structure).

In the GEE model, for the CRHBP-decient mice analysis was included the main effect of genotype (CRHBP / mice compared to CRHBP+/+ littermates), main effect for time, dened as the six ethanol exposures in the DID paradigm, and nally genotype time interaction. For the CRHBP knockdown rats in the ethanol self-administration procedure, the GEE model included the main effect of CRFBP downregulation by shRNA (CRHBP shRNA compared to Scr shRNA), main effect of time, dened as the 25 ethanol exposures in the operant chamber after surgery, and nally shRNA time interaction. For the fMRI procedure using CRHBP knockdown rats, the GEE model included the main effect of shRNA, main effect of yohimbine (compared to saline), main effect of time, dened as the time during the fMRI procedure comprised by saline (20.8-41.5 min) and yohimbine (41.6-62.4 min) infusion corrected from baseline (0-20.7 min), and shRNA yohimbine interaction. The same model was used to analyze center nucleus of the amygdala (CeA), paraventricular nucleus (PVN) and caudate putamen. Post hoc analysis was used to assess specic time-point differences over time and P-values were Bonferroni corrected for multiple comparisons. In CRHBP / mice and their CRHBP+/+ littermates, unpaired t-test was utilized to measure blood ethanol content level, water or sucrose consumption on day 4, the amount of pro-CRF to detect the differences among genotypes in the PVN and CeA. In the CRHBP shRNA analysis, unpaired t-test was utilized to analyze the in vitro downregulation, rats ethanol self-administration after recovery from surgery, and hemodynamic activation between saline and yohimbine infusion. Also, for Ca2+ assay signaling, comparisons between groups were performed using unpaired t-test. Pearson correlation was utilized to evaluate ethanol consumption and blood ethanol content.

Allele frequencies and genetic association analyses in humans were conducted using PLINK v1.07.13 Association with alcohol dependence diagnosis was tested using logistic regression and an additive genetic model. Further tests were conducted to evaluate associations with alcohol drinking and anxiety phenotypes in alcohol dependent individuals using linear regression. All models were conducted in the full sample, as well as split by European and African ancestry based on subject self-report. Age, gender, and ancestry informative marker scores for European and African ancestry were included as covariates in all models, even when split by ancestry; total score from the Childhood Trauma Questionnaire was also included as a covariate in the linear regression models for drinking and anxiety outcomes.

Results are reported as mean (M) and error bars indicate standard error (s.e.m.). All statistical tests were two-sided, and statistical signicance was accepted if a *P-value o 0.05 was obtained. SPSS (v.22) (Armonk, NY, USA)

was used to conduct the analysis and GraphPad Prism (v.5) was used to generate gures (La Jolla, CA, USA).

RESULTSCRFBP(FL) and its fragments, CRFBP(27kD) and CRFBP(10 kD), are stably expressed on the plasma membrane of both CRFR2 and CRFR1

Purifying human CRFBP or its fragments in sufcient quantities for investigation has not been successful to date. The precise mechanism that promotes CRFBP(FL) cleavage into two fragments, CRFBP(27kD) and CRFBP(10 kD), is still under investigation.4 Our goal was to have CRFBP(FL) and its fragments stably expressed and maintained in close proximity to the receptors in order to determine whether they modulate the receptors signaling. To achieve this expression, we created six covalently linked polypeptides between CRFBP, and its fragments, to the receptors. We fused FLAG-tagged CRFBP(FL), CRFBP(27kD) or CRFBP(10 kD) with HA-tagged CRFR1 or HA-CRFR2 and cloned these into a pCDNA3.1-based mammalian expression vector (Figure 1a). Constructs were veried by sequencing, protein integrity was veried by western blot (Figure 1b, Supplementary Figure S1a) and protein expression was conrmed using immunouorescence: FLAG-CRFBP(10 kD)-HA-CRFR2 (Figure 1c, Supplementary Movie S1), FLAG-CRFBP(FL)-HA-CRFR1, FLAG-CRFBP (FL)-HA-CRFR2, FLAG-CRFBP(27kD)-HA-CRFR1, FLAG-CRFBP (27kD)-HA-CRFR2 and FLAG-CRFBP(10 kD)-HA-CRFR1 (Supplementary Figure S1b). We also co-transfected FLAG-CRFBP (10 kD) (Figure 1d), FLAG-CRFBP(FL) and FLAG-CRFBP(27 kD) (Supplementary Figure S1c) with HA-CRFRs into HEK293 cells as

Translational Psychiatry (2016), 1 9

CRFBP and alcohol CL Haass-Kofer et al

3

Figure 1. The CRFBP(10 kD)-CRFR2 chimeras are expressed on the plasma membrane of HEK293 cells. (a) Schematic representation of the FLAG (DYKDDDDK epitope)-tagged CRFBP with hemagglutinin (HA)-tagged corticotropin releasing factor receptors (CRFRs) cloned into pCDNA3.1-based mammalian expression vector. (b) Western blot analysis of FLAG-CRFBP(10 kD)-HA-CRFR2 chimeric constructs. The blots were probed with primary antibody: anti-CRFBP C-19 (1:100) or anti-HA.11 (1:1000), then probed with secondary antibody: donkey anti-goat (IgG-HRP) (1:5000) to visualize CRFBP, or goat anti-mouse (IgG H+L-HRP) (1:5000) to visualize the receptors. Lane 1: HEK293 cells (negative control), lane 2: FLAG-CRFBP(10 kD) (MW ~10 kD) (positive control), lane 3: FLAG-CRFBP(10 kD)-HA-CRFR2 (MW ~ 60 kD), lane 4: FLAG-CRFBP (10 kD)-HA-CRFR2 (MW ~ 60 kD). Immunohistochemical staining of HEK293 cells: (c) transfected with FLAG-CRFBP(10 kD)-HACRFR2 showing

CRFBP(10 kD) stably expressed on the cell membrane and co-expressed with CRFR2. The cells were xed and permeabilized and then probed with anti-FLAG antibodies for FLAG-CRFBP(10 kD) and visualized using AlexaFluor-488 conjugated anti-mouse (IgG2a) secondary antibody, or probed using anti-HA and visualized using AlexaFluor-594 conjugated anti-mouse (IgG1) antibody and merged uorescent image with 4`,6-diamidino-2-phenylindonle (DAPI) to visualize the nuclei; (d) co-transfected with CRFBP(10 kD) with HA-CRFR2, which does not interact with the receptor at the plasma membrane. FLAG-CRFBP(10 kD) visualized using AlexaFluor-488 conjugated anti-mouse (IgG2a) secondary antibody was co-transfected into HEK293 cells with HA-CRFR2, probed using anti-HA, and visualized using AlexaFluor-594 conjugated anti-mouse (IgG1) and merged uorescent image with DAPI to visualize the nuclei. Scale bar, 5 M.

an additional control, to verify that CRFBP and its fragments, when not tethered to either CRFRs, were expressed intracellularly in regulated secretory pathways and did not interact with the receptors on the cell membrane.

Only CRFBP(10 kD)-CRFR2 can potentiate CRF-induced CRFR2 signaling

Next, we examined the functional properties of the covalently linked polypeptides created from CRFBP. We tested the ability of the chimeric proteins to activate intracellular calcium (Ca2+)

release to evaluate if they are capable of producing CRF-induced receptor signaling. Full-dose response curves were generated for CRF-induced release of intracellular Ca2+ at all chimera proteins and compared with HA-CRFR-only transfected cells. From comparison of the Emax values, we determined that only FLAG-CRFBP (10 kD) when tethered to HA-CRFR2 potentiates CRF-induced signaling, when compared to the CRFR2-only signal (Figure 2a). The activation by CRFBP(10 kD) via CRFR2, which leads to increased signaling in response to CRF, was unique to the combined action of CRFR2 tethered to the CRFBP(10 kD) fragment, was not mimicked by the other chimeric proteins (Supplementary Figures S2ac), and it was not due to the different levels of functional protein delivered to the cell surface (Figure 2b). In particular, CRF binds both CRFR1 and CRFR2, but is more potent at CRFR1.1417 Interestingly, there is ample evidence that CRFR1 is associated with anxiety-related behaviors.18,19 In light of this, we examined whether the observed

enhancement of the CRFBP(10 kD) chimeric complex signaling was specic to CRFR2 or whether the expression of CRFR1 would also increase signaling in a chimeric protein when tethered with CRFBP(10 kD). As expected, CRF induced a potent release of intracellular Ca2+ at CRFR1; however, there was no enhancement of CRF-induced signaling at both the CRFBP(10 kD)-CRFR1 and CRFBP(27kD)-CRFR1 chimeric proteins (Supplementary Figure S2b).

In additional control experiments, CRF-mediated Ca2+ signaling was absent in any FLAG-CRFBP-only transfected cells, and in nontransfected HEK293 cells, which do not express endogenous CRFR1 or CRFR2 (Supplementary Figures S2dh). To further test the full functionality of the FLAG-CRFBP(10 kD)-HA-CRFR2 chimeric protein, we examined whether it can be also inhibited by the CRFR2 antagonist antisauvagine (AS-30). The Ca2+ assay experiment demonstrated that the increase in intracellular Ca2+

release elicited by 1 M CRF (EC80) was dose-dependently inhibited by AS-30 (IC50) (Figure 2c).

Finally, the complete functionality of FLAG-CRFBP(10 kD)-HACRFR2 chimera was tested by evaluating its post-endocytic behavior. A three-dimensional (3D) confocal analysis showed chimeric receptors were distributed in the plasma membrane and projected outward from nite regions of the membrane of the transfected cells (Figure 2d). When cells transfected with FLAG-CRFBP(10 kD)-HA-CRFR2 were left untreated, there was no signicant receptor internalization. CRF-induced internalization of FLAG-CRFBP(10 kD)-HA-CRFR2 was inhibited by pretreatment with the CRFR2 selective antagonist AS-30.

Translational Psychiatry (2016), 1 9

CRFBP and alcohol CL Haass-Kofer et al

4

Figure 2. The CRFBP(10 kD)-CRFR2 chimeras potentiate CRF-induced release of intracellular Ca2+. (a) Full-dose response curves were generated for CRF-induced release of intracellular Ca2+ at the CRFBP(10 kD)-CRFR2 chimera (EC50 = 385 4.0 nM) compared with CRFR2-only transfected cells (EC50 = 99 8 nM). From comparison of the Emax values, we determined that CRFBP(10 kD), when tethered to CRFR2

(Emax = 2520 85 RFU), potentiates CRF-induced signaling when compared to the CRFR2-only signal (Emax = 930 81 RFU, ***P40.05). (c) The CRFR2 antagonist, antisauvagine-30 (AS-30, 1 pm 10 M), inhibits CRF (1 m)-induced intracellular calcium release in HEK293 cells stably expressing FLAG-CRFBP(10 kD)-HA-CRFR2 (IC50 = 5 3 0.4 nM). (d) Immunocytochemical analysis of receptor trafcking. Cells were left untreated (NT); treated with the agonist (CRF) for 30 min (1 M, promoted endocytosis of CRFR2 and the extracellular receptors were contracted); pretreated with CRFR2 selective antagonist AS-30 for 30 min (1 m) and then treated with agonist (CRF). Antagonist with AS-30 reduced CRF-induced endocytosis; cells showed no signicant difference in receptor contraction compared to cells left untreated. Scale bar, 5 M. Results are expressed as M s.e.m., relative uorescence units (RFU), calculated as agonist-induced maximum calcium peak/cell number 1000. CRFR, corticotropin releasing factor receptor.

Higher ethanol consumption in CRHBP / mice compared to their CRHBP+/+ littermates

To investigate the in vivo role of CRFBP in binge-like alcohol consumption, we tested the well-validated CRHBP-decient ( / ) mouse model 20 with the DID procedure using a 6-week paradigm.12 Wild-type (CRHBP+/+) littermates, used as control, underwent the same procedures. There was a signicant main effect of genotype in increasing ethanol consumption in the CRHBP / mice, compared to CRHBP+/+ littermates during the 4-h period on day-4 of ethanol access (Figure 3a, Supplementary Table S1). There were no main effects of time; however, we found a trend of genotype time interaction. Post hoc analysis revealed that ethanol consumption increased after repeated DID exposure, with a strong effect at the 6-DID cycle.

There was no difference between genotypes in 5% (v/v) sucrose consumption in the DID model (Figure 3b), nor in total water consumption during the 21-hour period when mice were not

exposed to ethanol (Figure 3c). Furthermore, consistent with previous DID studies,12 there was a linear correlation between ethanol intake and blood ethanol content in both genotypes. At baseline, the two genotypes consumed similar amounts of ethanol and we found no difference in blood ethanol content between CRHBP / and CRHBP+/+ littermates (Supplementary Figures S3a and b). To evaluate if there were CRF compensatory effects in the CRHBP / mice, we measured the amount of CRF from brain homogenates in the PVN of the hypothalamus and in the CeA. Western blot of pro-CRF expression in the PVN, where CRF is synthesized,21 and in the CeA, where it is highly expressed,22 at week 1 of the DID procedure in ethanol naive mice showed no signicant difference between the genotypes (Supplementary Figure S3c). Taken together, these results indicate that there were no baseline compensatory effects in CRHBP / mice that might account for the difference in ethanol consumption compared to wild-type littermates.

Translational Psychiatry (2016), 1 9

CRFBP and alcohol CL Haass-Kofer et al

5

*

9

*

Ethanol[gr/kg/4hr]

CRHBP -/- 4-hrCRHBP +/+ Littermates 4-hr

6

Day4

3

*P < 0.05

1 2 3 4 5 6

DID Cycle

20

28

CRHBP-/-

CRHBP+/+ Littermates

5%Sucrose[g/kg]

15

26

10

H 20[mL/100g]

24

5

22

P > 0.05

Figure 3. Ethanol consumption is signicantly higher in CRHBP / mice compared to their CRHBP+/+ littermates in the drinking in the dark (DID) procedure. (a) During the 4-h period on day 4, there was a signicant main effect of genotype (CRHBP / vs CRHBP+/+; n = 17 16, respectively; B = 0.977, *P o 0.05), no main effect of time (n = 17 16, P 40.05), but a trend of genotype time interaction (n = 17 16, B1 = 0.289, P = 0.073) with ethanol consumption increased after repeated DID exposures and a strong effect at the sixth DID cycle. There was no difference in (b) 5% (v/v) sucrose consumption (CRHBP / vs CRHBP+/+; n = 10 4, P40.05) and (c) total water consumption (CRHBP / vs CRHBP+/+; n = 8 3, P40.05), between the CRHBP / mice and their CRHBP+/+ littermates. Results are reported as M s.e.m.; not signicant (P 4 0.05).

P > 0.05

0 1 2 3 4

20 1 2 3 4

Exposure(Days)

Exposure(Days)

CRFBP downregulation in the CeA reduces ethanol self-administration and CeA hemodynamic activity, but does not attenuate yohimbine-induced ethanol self-administration

Previous studies have reported human (h)CRFBP mRNA and protein level in rat brain5 and dense hCRFBP-stained terminal elds in the CeA of human brain.23 However, very little is known about the mechanism and molecular network that underpin the stress-related functions mediated by CRFBP in the CeA, a limbic region that controls emotionality and alcohol reinstatement.24,25

Therefore, to investigate the causal link between CRFBP in the CeA and alcohol-seeking behaviors, we utilized a rat model trained to self-administer ethanol, designed with controlled regional expression of CRFBP. Sequences encoding small hairpin RNAs (shRNAs) targeting CRHBP were cloned into a lentiviral vector. We tested for the ability of the shRNAs to knock down CRHBP expression in vitro. We found that CRFBP expression was signicantly reduced in cells infected with lentivirus expressing CRHBP shRNA, as compared to cells infected with the control non-targeting shRNA (Scr) (Supplementary Figures S4a-c). We then selected the most potent CRHBP shRNA for in vivo testing. Long-Evans rats trained to self-administer ethanol were injected bilaterally in the CeA with lentivirus expressing the CRHBP or the control Scr shRNA. We conrmed infection in the CeA in vivo by magnetic resonance imaging scan (Supplementary Figure S4d) and post-mortem by GFP visualization of infected cells in each animal.

After recovery from surgery, during the subsequent 15 ethanol sessions, a signicant shRNA time interaction was found. Post hoc analysis indicated that rats infected with CRHBP shRNA showed a signicant reduction in ethanol consumption. By contrast, in the Scr shRNA control rats, there was no statistical

difference between the amount of ethanol consumed at baseline and the following sessions (Supplementary Figure S4E). Reduction of CRFBP expression in the CeA did not however attenuate yohimbine-induced ethanol drinking, as both CRHBP shRNA and Scr shRNA rats increased ethanol consumption after yohimbine i.p. administration. Taken together, these results indicate that CRHBP shRNA rats were unable to return to their baseline self-administration drinking behaviors. However, this effect disappeared after administering a pharmacological stressor via a yohimbine challenge.

To further validate the in vivo downregulation of CRFBP and the effect of yohimbine administration, we evaluated intrinsic brain excitability using blood-oxygen-level dependence, a measure of neural hemodynamic activity. During the magnetic resonance imaging scan, after a baseline assessment, all animals received a saline, followed by a yohimbine, i.p. infusion25 (Figure 4a and b). To make a mechanistic link between the effect of CRFBP knockdown in the CeA and ethanol self-administration, we analyzed other brain regions that are known to show activation during yohimbine administration. Yohimbine produced a composite pattern of blood-oxygen-level dependence % signal change between the CRHBP and the Scr shRNA rats in the limbic, hypothalamic and subcortical areas (Figure 4c and d). We found a main effect of yohimbine in the caudate putamen and a trend in the PVN. There was no main shRNA effect or yohimbine shRNA interaction in these brain regions. In the CeA, there was a lower hemodynamic activation in CRHBP shRNA rats compared to controls. Furthermore, when we compared the shRNA rats (CRHBP vs Scr) and condition (saline vs yohimbine) effects, we found a main signicant effect of yohimbine and a shRNA yohimbine

Translational Psychiatry (2016), 1 9

CRFBP and alcohol CL Haass-Kofer et al

6

Figure 4. Brain hemodynamic activity in the center nucleus of the amygdala (CeA) downregulated CRFBP in rats trained to self-administer ethanol. CRHBP short hairpin RNA (shRNA) (upper panel) and Scr shRNA (lower panel). (a, b) CeA bilateral representative functional magnetic resonance imaging scans of Scr shRNA controls and CRHBP shRNA rats. Timeline (green line) and pharmacological experimental manipulation (black: baseline; blue: saline; red: yohimbine). (c, d) We found a strong main effect of yohimbine in the caudate putamen (n = 6 each group, B1 = 0.801, ***Po0.001) and a trend of main effect of yohimbine in the PVN (n = 6 each group, B1 = 0.8876, P = 0.069). There was no main shRNA effect or yohimbine shRNA interaction for these outcomes. In the CeA, when we compared the shRNA rats and condition (saline and yohimbine) effects, we found a main signicant effect of yohimbine (n = 6 each group, B1 = 1.125, *P o 0.05) and a shRNA yohimbine interaction (n = 6 each group, B 1 = 2.372, *P o 0.05). A Bonferroni post hoc analysis revealed that (e) in the CRHBP shRNA rats there was no difference in CeA activation between saline and yohimbine infusion (Pcorrected 4 0.05), but (f) there was a stronger activation in the control Scr shRNA rats (t1499 = 2.511, ***Pcorrected o 0.001). Coronal sections are presented from their antero-posterior orientation. Results are reported as M s.e.m..; not signicant (P 4 0.05). PVN, paraventricular nucleus.

Table 1. Associations between SNPs, alcohol- and anxiety-related phenotypes in subjects with alcohol dependence

SNP Allele N P Ancestry Score Risk

rs10062367 A 467 1.310 0.02286 Full TLFB Increased

206 1.616 0.04844 African TLFB Increased 86 4.034 0.04699 European STAI anxiety Increased rs7718461 a A 220 1.457 0.01491 European CPRS anxiety Reduced rs1053989 b C 223 2.050 0.03561 European Neuroticism Reduced

221 0.298 0.03035 European CPRS anxiety Reduced

Abbreviations: CPRS, Comprehensive Psychopathological Rating Scale; SNP, single-nucleotide polymorphism; STAI, State-Trait Anxiety Inventory; TLFB, alcohol time life followback average drinks. aProxy for rs7718461. bProxy for rs10055255.

interaction. Post hoc analyses revealed that there was no difference in CeA activation between saline and yohimbine infusion in the CRHBP shRNA rats (Figure 4e). By contrast, as expected, there was a signicant difference in CeA activation between saline and yohimbine infusion in the control Scr shRNA rats (Figure 4f).

Human genetic association study of CRFBP(10 kD) candidate gene variants in alcoholic individuals

Consistent with previous rodent work, the in vivo experiments above indicate a role of CRFBP in alcohol-seeking behaviors. Furthermore, the in vitro chimeric work demonstrates a peculiar role of the CRFBP(10 kD) fragment in potentiating CRF signaling. Finally, a human genetic study was conducted in order to provide preliminary translation of these in vivo and in vitro ndings.

We a priori selected single-nucleotide polymorphisms (SNPs) of the CRHBP gene that are located in gene regions related to CRFBP (10 kD), an approach driven by the chimeric results reported above and were investigated for possible association with alcoholism in previous studies,26,27 an approach driven by the rodent results reported above. This approach led to the selection of three SNPs located in the CRFBP(10 kD) gene regions: rs10055255 and rs10062367, which were previously evaluated in stress-induced alcohol craving in heavy drinkers;26 and rs7728378, which was previously investigated in Caucasian alcohol

dependent individuals.27 Notably, the latter SNP was associated with suicide attempt in another study of patients with alcohol and substance dependence.28 Only rs10062367 was present in our dataset. For the other two SNPs, we were able to identify proxies (rs1053989 and rs7718461) in high linkage disequilibrium, in both the European and African subsamples (an important aspect given that the dataset includes approximately 60% Caucasians and 40% African Americans, Supplementary Table S2).

While there were no differences in the prevalence of these three SNPs analyzed between alcohol dependent patients and controls (data not shown), the rs10062367 A allele showed a signicant association with average drinks per drinking day, both in the full sample as well as in the African ancestry subgroup. Associations between the three SNPs and alcohol phenotypes in subjects with current alcohol dependence are summarized in Table 1 and allele frequencies in Supplementary Table S3. For the anxiety measures, there was a signicant main effect of rs10062367 A allele for the State-Trait Anxiety Inventory anxiety in European ancestry. We found a signicant effect of rs7718461 A allele for the Comprehensive Psychopathological Rating Scale anxiety in European ancestry, and a negative association with this scale anxiety in European ancestry. There was a signicant effect of rs1053989 C allele for the Comprehensive Psychopathological Rating Scale anxiety in the European ancestry. Finally, we found a signicant effect for neuroticism in the European ancestry. These

Translational Psychiatry (2016), 1 9

CRFBP and alcohol CL Haass-Kofer et al

7

results suggest the possibility that variation of the CRFBP(10 kD) protein might alter the risk of alcohol drinking and/or anxiety in alcohol-dependent patients.

DISCUSSIONUsing a combination of different in vitro, in vivo, rodent and human studies, our work represents a comprehensive translational effort to dene the discrete biological role of CRFBP. Our results provide converging evidence supporting the role of CRFBP in alcohol consumption.

First, we sought to establish a CRFBP role in CRF-receptor signaling. Here we describe a novel approach where, by tethering human CRFBP with its receptors, we were able to express it at the cell surface and evaluate if CRFBP participates in CRF signaling. We discovered that only CRFR2 is involved in the CRF-mediated receptor activation in the presence of CRFBP. This observation is consistent with our previous electrophysiological work showing that CRF selectively increases N-methyl-D-aspartate currents via CRF interaction with CRFR2 and CRFBP.9 Furthermore, our results are in line with more recent work demonstrating that CRFBP interacts with CRFR2 and increases receptor expression on the cell surface.29 The chimeric receptor approach allowed us to study not only the native protein, but also its fragments. Receptor signal response was detected only when CRFBP(10 kD) was tethered to CRFR2. By creating a series of chimeric receptors tethered to individual CRFBP fragments, we discovered a potential activating role of the CRFBP(10 kD), specically via CRFR2 signaling, that would not have been discovered without this approach. The fact that CRFBP(10 kD) tethered to CRFR2 was able to potentiate receptor signaling indicates that CRFBP(10 kD) may play a role in vivo. We propose that CRF interacts with CRFBP(10 kD) and CRFR2 in an allosteric manner since CRFBP(27kD) retains the active picomolar afnity binding site for CRF.30 As such, our results suggest a potential dual action of CRFR2 and CRFR1 during stress maladaptation, as was previously postulated.31 Genetic and pharmacological manipulations targeting CRFR2 have shown that CRF may lead to either increased or decreased response to stressors.32 In normal conditions, CRF signal via CRFR133 and CRFBP, in its protecting role, binds the extra unbound CRF.30

However, during maladaptive stress induction, CRFBP(27kD) cannot sequester the excess of free CRF, while CRFBP(10 kD) allosterically potentiates CRFR2 signaling. CRFBP is a critical component of the CRF system, yet its specic physiological role has not been elucidated, mostly because of the difculty in purifying a large-enough quantity for experimental design.4 Our approach is not intended to show that the covalently linked CRFBP-CRFR2 mimics in vivo physiological conditions. Rather, this translational approach provides, to the best of our knowledge, the rst novel screening assay to identify non-peptidic chemical inhibitors that disrupt the CRFR2 activation in the presence of CRFBP. Our data show an enhancing role of CRFBP and CRFR2 interaction in the central nervous system, a nding consistent with recently reported molecular,29 and our original electrophysiology,9 data. The coexistence of CRFBP and CRFR2 in glutamatergic/ GABA synaptosomes6 also supports our ndings, as well as independent work suggesting that CRFBP may possess a facilitatory effect on administration of cocaine8 and ethanol.7 Our chimeric results provide additional evidence to previous work indicating that (1) CRFBP possesses both a sequestering and a potentiating role for CRF;34 (2) CRFR2 may play a role in the regulation of stress response;35 and (3) free CRF alone does not necessarily predict stress response.23

Previous work showed that CRF is a modulator of ethanol drinking within the DID paradigm.3638 In the present work, we evaluated the in vivo role of CRFBP in binge-like ethanol consumption. For this purpose, we utilized the CRHBP / mouse integrated in the DID procedure, a validated procedure to study

the transition to ethanol dependence. Consistent with the previous characterization of CRHBP / mice,3942 our results indicated that, in the absence of the CRFBP, higher levels of unbound CRF increase CRFR activation, leading to increased ethanol consumption in the CRHBP / mice. Our results extend previous ndings demonstrating an important role of CRFBP in attenuating the neural responses produced by continuous exposure to stress insults.30 These observations are further corroborated, rst by a signal detected in the genotype time interaction. Then, a post hoc analysis revealed that CRHBP / mice compared to the CRHBP+/+ littermates showed increased ethanol consumption after 6-DID exposures. While CRF involvement was not tested directly and is a recognized limitation of this study, we speculate that our results are CRF-mediated as they corroborate previous work using the DID paradigm, where ethanol intake enhances CRF function in the CeA, and systemic, intracerebroventricular and intra-amygdalar administration of CRFR1 antagonist reduces DID intake.32,36,37,43,44

In line with these results, we found here the increase of ethanol drinking behavior observed in CRHBP / mice after the repeated DID cycle is most likely the result of increased unbound CRF in the brain of CRHBP-decient animals, rather than a basal down-regulation of CRF synthesis measured in the CeA and PVN, or an alteration of ethanol metabolism. Our western blot analysis, in ethanol naive mice, are consistent with the recent published data on CRF mRNA levels in CRHBP / mice and their littermates.45

Together, these results support a contribution of CRFBP in ethanol consumption in mice, and raise the possibility that CRFBP imbalance contributes to binge drinking.

CRF mediates behavioral responses to stress and ethanol consumption in extra-hypothalamic regions, including the CeA.24

The in vivo function of CRFBP in regulating neuronal activity in circuits involved in addiction has been poorly understood due to the instability of CRFBP in vitro and the lack of pharmacological probes. Thus, in order to determine the mechanistic role of CRFBP in extrahypothalamic regions and its effect on ethanol-seeking behavior, we hypothesized that downregulation of CRHBP in the CeA modulates excitatory behavioral outcomes. This was based on the observation that the CeA is a limbic brain region known to be important for modulation of stress activation where CRHBP is highly expressed.22 Reduced CRHBP levels in the CeA resulted in lower drinking in the ethanol operant self-administration procedure; however, the reduction of ethanol consumption did not persist after a yohimbine challenge. Our results suggest that downregulating CRFBP regional expression, by knocking down CRFBP in the CeA, was sufcient to blunt the ethanol drinking-behavioral phenotype per se. However, it is not sufcient to control drinking during activation of the brain norepinephrine system. Consistent with this hypothesis, our fMRI data showed that other known brain regions activated by yohimbine and sensitive to norepinephrine stimulation (caudate putamen and PVN studied here) are activated, and they may be responsible for the increase of ethanol consumption. Viral infusion into the CeA selectively altered the reactivity to yohimbine in this area, but not in other brain structures in CRHBP knockdown rats, probably because shRNA CRHBP only downregulated CRFBP expression (by 50%), resulting in conned reduction of metabolic activation in the CeA without affecting other brain areas with strong connectivity. These observations are in line with recent results demonstrating that microinjection of CRFBP antagonist CRF6-33

into the CeA does not change alcohol intake, while CRFBP in the VTA may be responsible for escalation of ethanol drinking.7

Our nal goal was to provide initial translational evidence of our ndings. While the covalently linked CRFBP-CRFR2 does not represent an in vivo physiological condition, our in vitro chimeric work guided us in narrowing down the SNP selection from 322 to 88 amino-acid residues. We showed that rs10062367 A seems to be a risk allele, as it was associated with increased average drinks/ day and an increased anxiety trait. These data are not consistent

Translational Psychiatry (2016), 1 9

CRFBP and alcohol CL Haass-Kofer et al

8 with previous work26 where rs10055255, but not rs10062367, was associated with alcohol- and stress-related phenotypes. However, several factors may explain the differences in these results, for example, differences in the size of the two samples, race- and ethnicity-related differences, and different instruments used to assess the outcomes of the two studies. On the other hand, consistent with the other previous studies,26,27 we found that

rs7718461 A and rs1053989 C alleles were associated with lower neuroticism and anxiety (reduced risk). However, it is important to note that the SNPs analyzed here are all located in the intronic portion of the gene, which suggests that they might not have functional signicance. Although the selected SNPs were all intronic, there are no known SNPs in the gene region encoding for the CRFBP(10 kD) fragment; some are very rare while others have extremely low minor allele to allow for properly powered genetic analyses. Nonetheless, our approach was based on a strong a priori hypothesis and our results suggest that CRFBP may play a role in alcohol- and anxiety-related behaviors in alcohol-dependent patients. Although no case control differences were found and the exact functionality of these SNPs remain uncertain, the present results, together with previous work, suggest a role of CRFBP in alcohol- and anxiety-related phenotypes in a clinically relevant sample of individuals with alcohol dependence. Indeed, the present data represent the largest human study looking at the genetic variant of the CRFBP gene in alcohol dependence.

In summary, CRFBP has been studied for more than two decades, and a variety of hypotheses on its physiological role have been developed. Our ndings suggest that CRFBP may possess a dual role: CRFBP(27kD) is responsible for neutralizing CRF effects, while CRFBP(10 kD) has a potential excitatory function. Evaluating the individual functional interaction of each CRFBP fragment with each receptor may elucidate many aspects of stress and AUD, and lead to the development of novel effective treatments.

CONFLICT OF INTEREST

Dr Swift has received travel grants and honorarium from D&A Pharma, Lundbeck, and consultant fees from CT Laboratories. The remaining authors declare no competing nancial interests.

ACKNOWLEDGMENTS

The authors would like to thank their funding sources: K01AA023867, T-32AA007459, UCSF Schools of Pharmacy and Medicine, and Strategic Opportunities Exploratory Award (CLH-K); NIAAA Division of Intramural Clinical and Biological Research (MLS, RMS, CAH, LL) and NIDA Intramural Research Program (AB, LL); State of California and NIH Fast Track Award (SEB). The authors also would like to thank Dr Audrey Seasholtz, University of Michigan, for providing the breeding pairs of the CRHBP KO mice; the NIAAA and Clinical Center staff involved in the human data collection at the NIH Intramural Research Program; and Ms. Karen Smith, NIH Library for assistance.

REFERENCES

1 Sarnyai Z, Shaham Y, Heinrichs SC. The role of corticotropin-releasing factor in drug addiction. Pharmacol Rev 2001; 53: 209243.

2 Shaham Y, Erb S, Stewart J. Stress-induced relapse to heroin and cocaine seeking in rats: a review. Brain Res Rev 2000; 33: 1333.

3 Haass-Kofer CL, Bartlett SE. Stress and addiction: contribution of the corticotropin releasing factor (CRF) system in neuroplasticity. Front Mol Neurosci 2012; 5: 91.

4 Woods RJ, Kemp CF, David J, Sumner IG, Lowry PJ. Cleavage of recombinant human corticotropin-releasing factor (CRF)-binding protein produces a 27-kilodalton fragment capable of binding CRF. J Clin Endocr Metab 1999; 84: 27882794.

5 Potter E, Behan DP, Linton EA, Lowry PJ, Sawchenko PE, Vale WW. The central distribution of a corticotropin-releasing factor (CRF)-binding protein predicts multiple sites and modes of interaction with CRF. Proc Natl Acad Sci USA 1992; 89: 41924196.

6 Slater PG, Noches V, Gysling K. Corticotropin-releasing factor type-2 receptor and corticotropin-releasing factor-binding protein coexist in rat ventral tegmental area nerve terminals originated in the lateral hypothalamic area. Eur J Neurosci 2016; 43: 220229.

7 Albrechet-Souza L, Hwa LS, Han X, Zhang EY, DeBold JF, Miczek KA. Corticotropin releasing factor binding protein and CRF2 receptors in the ventral tegmental area: modulation of ethanol binge drinking in C57BL/6 J mice. Alcohol Clin Exp Res 2015; 39: 16091618.

8 Wang B, You ZB, Rice KC, Wise RA. Stress-induced relapse to cocaine seeking: roles for the CRF(2) receptor and CRF-binding protein in the ventral tegmental area of the rat. Psychopharmacology 2007; 193: 283294.

9 Ungless MA, Singh V, Crowder TL, Yaka R, Ron D, Bonci A. Corticotropin-releasing factor requires CRF binding protein to potentiate NMDA receptors via CRF receptor 2 in dopamine neurons. Neuron 2003; 39: 401407.

10 Rubinson DA, Dillon CP, Kwiatkowski AV, Sievers C, Yang L, Kopinja J et al. A lentivirus-based system to functionally silence genes in primary mammalian cells, stem cells and transgenic mice by RNA interference. Nat Genet 2003; 33: 401406.

11 Lasek AW, Janak PH, He L, Whistler JL, Heberlein U. Downregulation of mu opioid receptor by RNA interference in the ventral tegmental area reduces ethanol consumption in mice. Genes Brain Behav 2007; 6: 728735.

12 Rhodes JS, Best K, Belknap JK, Finn DA, Crabbe JC. Evaluation of a simple model of ethanol drinking to intoxication in C57BL/6 J mice. Physiol Behav 2005; 84: 5363.

13 Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D et al. PLINK: a tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet 2007; 81: 559575.

14 Lovenberg TW, Liaw CW, Grigoriadis DE, Clevenger W, Chalmers DT, De Souza EB et al. Cloning and characterization of a functionally distinct corticotropin-releasing factor receptor subtype from rat brain. Proc Natl Acad Sci USA 1995; 92: 836840.

15 Liaw CW, Grigoriadis DE, Lovenberg TW, De Souza EB, Maki RA. Localization of ligand-binding domains of human corticotropin-releasing factor receptor: a chimeric receptor approach. Mol Endocrinol 1997; 11: 980985.

16 Zhao Y, Valdez GR, Fekete EM, Rivier JE, Vale WW, Rice KC et al. Subtype-selective corticotropin-releasing factor receptor agonists exert contrasting, but not opposite, effects on anxiety-related behavior in rats. J Pharmacol Exp Ther 2007; 323: 846854.

17 Fekete EM, Zorrilla EP. Physiology, pharmacology, and therapeutic relevance of urocortins in mammals: ancient CRF paralogs. Front Neuroendocrinol 2007; 28: 127.

18 Magalhaes AC, Holmes KD, Dale LB, Comps-Agrar L, Lee D, Yadav PN et al. CRF receptor 1 regulates anxiety behavior via sensitization of 5-HT2 receptor signaling. Nat Neurosci 2010; 13: 622629.

19 Muller MB, Zimmermann S, Sillaber I, Hagemeyer TP, Deussing JM, Timpl P et al. Limbic corticotropin-releasing hormone receptor 1 mediates anxiety-related behavior and hormonal adaptation to stress. Nat Neurosci 2003; 6: 11001107.

20 Karolyi IJ, Burrows HL, Ramesh TM, Nakajima M, Lesh JS, Seong E et al. Altered anxiety and weight gain in corticotropin-releasing hormone-binding protein-decient mice. Proc Natl Acad Sci USA 1999; 96: 1159511600.

21 Haas DA, George SR. Estradiol or ovariectomy decreases CRF synthesis in hypothalamus. Brain Res Bull 1989; 23: 215218.

22 Beckerman MA, Van Kempen TA, Justice NJ, Milner TA, Glass MJ. Corticotropin-releasing factor in the mouse central nucleus of the amygdala: ultrastructural distribution in NMDA-NR1 receptor subunit expressing neurons as well as projection neurons to the bed nucleus of the stria terminalis. Exp Neurol 2013; 239: 120132.

23 Behan DP, Khongsaly O, Owens MJ, Chung HD, Nemeroff CB, De Souza EB. Corticotropin-releasing factor (CRF), CRF-binding protein (CRF-BP), and CRF/ CRF-BP complex in Alzheimer's disease and control postmortem human brain. J Neurochem 1997; 68: 20532060.

24 Roberto M, Cruz MT, Gilpin NW, Sabino V, Schweitzer P, Bajo M et al. Corticotropin releasing factor-induced amygdala gamma-aminobutyric acid release plays a key role in alcohol dependence. Biol Psychiatry 2010; 67: 831839.

25 Simms JA, Haass-Kofer CL, Bito-Onon J, Li R, Bartlett SE. Mifepristone in the central nucleus of the amygdala reduces yohimbine stress-induced reinstatement of ethanol-seeking. Neuropsychopharmacology 2012; 37: 906918.

26 Ray LA. Stress-induced and cue-induced craving for alcohol in heavy drinkers: preliminary evidence of genetic moderation by the OPRM1 and CRH-BP genes. Alcohol Clin Exp Res 2011; 35: 166174.

27 Enoch MA, Shen PH, Ducci F, Yuan Q, Liu J, White KV et al. Common genetic origins for EEG, alcoholism and anxiety: the role of CRH-BP. PLoS One 2008; 3: e3620.

28 Roy A, Hodgkinson CA, Deluca V, Goldman D, Enoch MA. Two HPA axis genes, CRHBP and FKBP5, interact with childhood trauma to increase the risk for suicidal behavior. J Psychiatr Res 2012; 46: 7279.

29 Slater PG, Cerda CA, Pereira LA, Andres ME, Gysling K. CRF binding protein facilitates the presence of CRF type 2alpha receptor on the cell surface. Proc Natl Acad Sci USA 2016; 113: 40754080.

30 Behan DP, De Souza EB, Lowry PJ, Potter E, Sawchenko P, Vale WW. Corticotropin releasing factor (CRF) binding protein: a novel regulator of CRF and related peptides. Front Neuroendocrinol 1995; 16: 362382.

Translational Psychiatry (2016), 1 9

CRFBP and alcohol CL Haass-Kofer et al

9

31 Janssen D, Kozicz T. Is it really a matter of simple dualism? Corticotropin-releasing factor receptors in body and mental health. Front Endocrinol (Lausanne) 2013; 4: 28.

32 Kaur S, Li J, Stenzel-Poore MP, Ryabinin AE. Corticotropin-releasing factor acting on corticotropin-releasing factor receptor type 1 is critical for binge alcohol drinking in mice. Alcohol Clin Exp Res 2012; 36: 369376.

33 Heilig M, Koob GF. A key role for corticotropin-releasing factor in alcohol dependence. Trends Neurosci 2007; 30: 399406.

34 Ketchesin KD, Seasholtz AF. Novel roles for CRF-binding protein and CRF receptor 2 in binge drinking. Alcohol Clin Exp Res 2015; 39: 22962298.

35 Ryabinin AE, Tsoory MM, Kozicz T, Thiele TE, Neufeld-Cohen A, Chen A et al. Urocortins: CRF's siblings and their potential role in anxiety, depression and alcohol drinking behavior. Alcohol 2012; 46: 349357.

36 Sparta DR, Hopf FW, Gibb SL, Cho SL, Stuber GD, Messing RO et al. Binge ethanol-drinking potentiates corticotropin releasing factor R1 receptor activity in the ventral tegmental area. Alcohol Clin Exp Res 2013; 37: 16801687.

37 Lowery EG, Spanos M, Navarro M, Lyons AM, Hodge CW, Thiele TE. CRF-1 antagonist and CRF-2 agonist decrease binge-like ethanol drinking in C57BL/6 J mice independent of the HPA axis. Neuropsychopharmacology 2010; 35: 12411252.

38 Pleil KE, Rinker JA, Lowery-Gionta EG, Mazzone CM, McCall NM, Kendra AM et al. NPY signaling inhibits extended amygdala CRF neurons to suppress binge alcohol drinking. Nat Neurosci 2015; 18: 545552.

39 Westphal NJ, Seasholtz AF. CRH-BP: the regulation and function of a phylogenetically conserved binding protein. Front Biosci 2006; 11: 18781891.

40 Gammie SC, Seasholtz AF, Stevenson SA. Deletion of corticotropin-releasing factor binding protein selectively impairs maternal, but not intermale aggression. Neuro-science 2008; 157: 502512.

41 Seasholtz AF, Burrows HL, Karolyi IJ, Camper SA. Mouse models of altered CRH-binding protein expression. Peptides 2001; 22: 743751.

42 Seasholtz AF, Valverde RA, Denver RJ. Corticotropin-releasing hormone-binding protein: biochemistry and function from shes to mammals. J Endocrinol 2002; 175: 8997.

43 Sparta DR, Sparrow AM, Lowery EG, Fee JR, Knapp DJ, Thiele TE. Blockade of the corticotropin releasing factor type 1 receptor attenuates elevated ethanol drinking associated with drinking in the dark procedures. Alcohol Clin Exp Res 2008; 32: 259265.

44 Lowery-Gionta EG, Navarro M, Li C, Pleil KE, Rinker JA, Cox BR et al. Corticotropin releasing factor signaling in the central amygdala is recruited during binge-like ethanol consumption in C57BL/6J mice. J Neurosci 2012; 32: 34053413.

45 Ketchesin KD, Stinnett GS, Seasholtz AF. Binge drinking decreases corticotropin-releasing factor-binding protein expression in the medial prefrontal cortex of mice. Alcohol Clin Exp Res 2016; 40: 16411650.

This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the articles Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

Web End =http://creativecommons.org/licenses/ http://creativecommons.org/licenses/by/4.0/

Web End =by/4.0/

The Author(s) 2016

Supplementary Information accompanies the paper on the Translational Psychiatry website (http://www.nature.com/tp)

Translational Psychiatry (2016), 1 9