Vasopressin and alcohol: A multifaceted relationship

Abstract

Background:

Arginine vasopressin (VP) has been implicated in a number of neuropsychiatric disorders with an emphasis on situations were stress increased the severity of the disorder. Based on this hypothesized role for VP in neuropsychiatric disorders, much research is currently being undertaken in humans and animals to test VP as a target for treatment of a number of these disorders including alcohol abuse.

Objectives:

To provide a summary of the literature regarding the role of VP in alcohol and stress related behaviors including the use of drugs that target VP in clinical trials.

Results:

Changes in various components of the VP system occur with alcohol and stress. Manipulating VP or its receptors can alter alcohol and stress related behaviors including tolerance to alcohol, alcohol drinking, and anxiety-like behavior. Finally, the HPA axis response to alcohol is also altered by manipulating the VP system. However, clinical trials of VP antagonists have had mixed results.

Conclusions:

A review of VP’s involvement in alcohol’s actions demonstrates that there is much to be learned about brain regions involved in VP-mediated effects on behavior. Thus, future work should focus on elucidating relevant brain regions. By using previous knowledge of the actions of VP and determining the brain regions and/or systems involved in its different behavioral effects, it may be possible to identify a specific receptor subtype target, drug treatment combination, or specific clinical contexts that may point toward a more successful treatment.

1. Introduction

Arginine vasopressin (VP; also known as antidiuretic hormone or ADH) is a nine amino acid peptide known to be involved in a number of physiological and behavioral effects. In both adult mice and rats detailed immunohistchemical studies mapped the location of VP immunopositive neurons and their fibers. VP containing neurons are found most abundantly in the paraventricular nucleus (PVN), suprachiasmatic nucleus (SCN), and supraoptic nucleus (SON) of the hypothalamus. Smaller quantities have been found in the Bed nucleus of the stria terminalis (BST), and medial amygdala (DeVries et al. 1985; Otero-Garcia et al. 2016; Rood and De Vries 2011; Sofroniew 1983; Song and Albers 2017). Colchicine, which is thought to cause cell body accumulation of proteins normally localized to fibers, has revealed staining in other brain regions including the locus coeruleus and septum (DeVries et al. 1985; Sofroniew 1985). The colchicine experiments suggest that traditional immunohistochemistry techniques alone will not identify low expression VP neurons that might have functions outside the high expression regions, e.g. hypothalamus, amygdala or BST. Other studies using a VP reporter mouse have found VP neurons in the retina and olfactory bulbs (Moritoh et al. 2011; Tobin et al. 2010) and mRNA gene arrays have reported VP expression in the nucleus accumbens (Rodriguez-Borrero et al. 2010). These later studies suggest that VP may be synthesized in the cell body of neurons and rapidly transported to terminal regions. Thus, emerging studies using a combination of immunohistochemical procedures to localize protein and RNA can be expected to find a broad brain regional distribution of VP neurons beyond the traditionally known hypothalamic and extended amygdala regions.

A map showing known projections of the VP containing neurons can be seen in figure 1, however studies vary widely on which fibers come from which brain region. This variation seems to mostly stem from the technique used to identify the source of the fibers and perhaps species as well. The VP neurons in the medial amygdala have been found to project to the hippocampus, BST, and lateral septum (Caffe et al. 1987). The VP neurons in the BST project to the lateral septum, amygdala, ventral pallidum, lateral habenular nucleus, the periventricular gray, and the locus coeruleus (De Vries and Buijs 1983; Otero-Garcia et al. 2014). The VP neurons in the SCN project to the vascular organ of the lamina terminalis, the periventricular nucleus, the PVN, and the dorsomedial hypothalamus (Buijs 1978; 1980; Hoorneman and Buijs 1982; Rood et al. 2013; Sofroniew 1980; Sofroniew and Weindl 1978). The VP neurons in the PVN project to the pituitary, the spinal cord, and amygdala (Buijs 1978; 1980; Hernandez et al. 2016; Sofroniew 1980). The projection systems described above are based primarily on lesioning and retrograde-tracing (De Vries and Buijs 1983; Hoorneman and Buijs 1982; Otero-Garcia et al. 2014; Rood et al. 2013). Early studies where individual neurons were stained and followed to other brain sites showed many more possible projection pathways (Buijs 1978; 1980; Sofroniew 1980; Sofroniew and Weindl 1978). However, lesioning the nucleus of origin did not alter the VP fibers in the proposed projection region, a finding that suggested that the contribution was minor compared to inputs from other sites. These sites were not included in the above description. Therefore more detailed lesioning studies in a region by region fashion or site by site retrograde tracing using techniques like fluorescent beads combined with immunohistochemistry might assist in confirming the existing data.

Figure 1: Summary of current evidence of VP related brain regions involved in alcohol related behaviors.

The bed nucleus of the stria terminalis has been shown to have changes in VP mRNA with alcohol consumption. The amygdala exhibits changes in the levels of V1b receptors with alcohol. The hypothalamus shows many VP related changes with alcohol consumption and likely plays a role in alcohol drinking as rats breed for alcohol preference have higher levels of VP in the PVN and SON.

VP binds to one of three G protein-coupled receptors, vasopressin receptor 1a (V1a), vasopressin receptor 1b (V1b) and vasopressin receptor 2 (V2) (Barberis and Tribollet 1996; Hasunuma et al. 2013; Song and Albers 2017). The V1a receptor regulates a phosphatidylinositol 4,5-bisphosphate (PIP2) pathway and is expressed throughout the brain in the frontal cortex, piriform cortex, anterior olfactory nucleus, SCN, hippocampus, PVN, ventromedial hypothalamic nucleus, arcuate nucleus, lateral habenular nucleus, cerebellum, BST, nucleus accumbens, and amygdala (Song and Albers 2017; Szot et al. 1994; Veinante and Freund-Mercier 1997). The V1b receptor regulates PIP2 pathway and is expressed in the hippocampus, PVN, amygdala, lateral septum, BST, frontal cortex, cingulate cortex, entorhinal cortex, parietal cortex, nucleus accumbens, and pituitary (Stevenson and Caldwell 2012; Young et al. 2006). The V2 receptor regulates an adenylate cyclase/cyclic adenosine 3’,5’-monophosphate (cAMP) pathway and by adulthood V2 receptors seem restricted to the cerebellum and outside the brain (Hirasawa et al. 1994; Kato et al. 1995). Most of the work involving vasopressin and alcohol has not focused on the V2 receptor, but rather on the V1a and V1b receptors, and given their distribution there are many potential brain regions through which VP can have effects on alcohol related behaviors.

There are sex differences in the VP system, which are partly thought to stem from the fact that many VP neurons also contain gonadal steroid hormone receptors and gonadal steroid hormones have been shown to control the synthesis of VP (Carter 2007; Pak et al. 2009; Thomas et al. 1996; van Leeuwen et al. 1985; Zhou et al. 1994). In many species, including humans, males have more VP immunoreactive fibers, more VP immunoreactive neurons, VP neurons of larger size, and altered V1a receptor binding densities (De Vries and Panzica 2006; DiBenedictis et al. 2017; Dumais and Veenema 2016; Ishunina and Swaab 1999; Rood et al. 2013; Smith et al. 2017; Winslow et al. 1993). This upregulation of the VP system in males is thought to underlie the findings that VP exposure in human and rodent males causes a behavioral response at a dose where females show either no response or a different behavioral outcome (Chen et al. 2016; Terranova et al. 2017; Thompson et al. 2006).

VP has neuromodulatory activities in brain regions suggesting it could affect behavior. For example VP can excite gamma-Aminobutyric acid (GABA) neurons in the central nucleus of the amygdala (CeA) (Huber et al. 2005), excite serotonin neurons in the dorsal raphe (Rood and Beck 2014), excite spinal motorneurons (Raggenbass 2008), excite GABA septal neurons (Raggenbass 2008), and excite GABA neurons in the hippocampus (Ramanathan et al. 2012), all of which are thought to be mediated by V1a. Additionally, VP has effects on plasticity modulating long term potentiation and long term depression in GABA neurons in the BST (Naughton 2016). In the hippocampus VP facilitates production of long term potentiation of CA1 and dentate gyrus neurons (Chen et al. 1993; Chepkova et al. 2001). Thus, VP is an excitatory transmitter at the level of the synapse and has been known to impact plasticity in multiple neurocircuits.

Effects of VP cannot be fully understood without looking at its interaction with oxytocin (OT) and these interactions occur at multiple levels. OT shares 7 of the 9 amino acids comprising VP and their receptors show a great deal of cross reactivity (Song and Albers 2017). Additionally, there are anatomical links between the VP and OT systems. For example, OT and VP receptors exist in many of the same brain regions but in different subdivisions within those brain regions (Huber et al. 2005; Veinante and Freund-Mercier 1997). In the case of the amygdala these OT and VP subdivisions even interact such that the OT receptor containing neurons in the centrolateral nucleus of the amygdala project to the VP receptor containing neurons in the centromedial nucleus of the amygdala and alter their firing pattern (Huber et al. 2005). Such serial interactions between OT and VP can explain why they have different or even opposite effects on the same behavior through the same brain regions (Huber et al. 2005; Neumann and Landgraf 2012; Stoop 2012). Therefore, in cases were VP antagonists have been suggested as a treatment OT agonists are also sometimes suggested as a treatment. This is also true of the alcohol field (Lee and Weerts 2016).

When discussing VP, another commonly discussed peptide hormone is corticotropin-releasing factor (CRF). Like VP, CRF is also made by neurons in the PVN. CRF has two receptors CRF-R1 and CRF-R2. CRF-R1 is expressed mainly in the brain and CRF binds CRF-R1 preferentially over CRF-R2. CRF from the PVN leads to release of adrenocorticotropic hormone (ACTH) from the anterior pituitary. The released ACTH leads to secretion of cortisol/corticosterone (CORT) from the adrenal glands. Therefore, CRF is an important player in the hypothalamic pituitary adrenal (HPA) stress response, an outcome which is also true of VP. The same PVN neurons contain both CRF and VP which likely can lead to coordinated release of the two peptides (Whitnall et al. 1985). Simultaneous release of both CRF and VP enhances the effects of CRF on ACTH release (Gillies et al. 1982). Therefore, for stress-related neuropsychiatric disorders where VP antagonism is a suggested treatment, CRF antagonism is a suggested treatment as well.

2. Alcohol’s effect on the VP system

An association between VP and alcohol was initially discovered in the context of alcohol-induced changes of plasma VP. Since neurons in the hypothalamus have terminals in the posterior pituitary which release VP into circulation, a change in plasma VP can be an indicator of changes in VP in the brain or at least in those specific hypothalamic neurons. In humans, plasma VP levels often decrease during alcohol consumption and increase upon cessation of consumption (Eisenhofer and Johnson 1982; Gill et al. 1982; Helderman et al. 1978; Linkola et al. 1978; Mander et al. 1989; Taivainen et al. 1995). Interestingly, individuals who abuse alcohol seem to have differences in their VP system compared to more alcohol naïve individuals. When comparing alcoholics to more alcohol naïve individuals, alcoholics were found to have a more pronounced decrease in plasma VP levels when drinking (Collins et al. 1992; Hirschl et al. 1994), suppressed VP levels even during alcohol withdrawal (Doring et al. 2003; Hirschl et al. 1994; Jahn et al. 2004; Trabert et al. 1992), and a lack of a VP increase in response to novelty (Ehrenreich et al. 1997). Similarly, differences in urine VP (urine VP being an approximate measure of plasma VP) are found between alcohol preferring (AA rats) and alcohol non-preferring rats (ANA) (Linkola and Fyhrquist 1978; Linkola et al. 1977). However the two studies do not agree on the directionality of the difference as one study found a decrease in urine VP in AA rats compared to ANA rats and the other an increase in urine VP in AA rats compared to ANA rats. A possibly related variation found in alcohol abusers is increased promoter-related DNA methylation of VP DNA extracted from blood (Hillemacher et al. 2009). Additional studies are needed to clearly define how brain, blood and urine VP are modified during alcohol use and abstinence.

It is now known that these alcohol induced changes in VP in the periphery extend to the brain. For example, it has been shown that there is a decrease in the number of VP immunoreactive neurons in the hypothalamus with alcohol exposure in rats and humans (Harding et al. 1996; Madeira et al. 1997; Paula-Barbosa et al. 2001; Silva et al. 2002; Sousa et al. 1995). However, these reports are conflicted as to whether withdrawal reverses the decrease in the number of VP neurons. Acute alcohol increases VP mRNA in the hypothalamus in rats (Ogilvie et al. 1997; Rivier and Lee 1996) whereas chronic alcohol decreases VP mRNA in hypothalamus and BST in both rats and mice (Gulya et al. 1991; Gulya et al. 1993; Hoffman and Dave 1991; Ishizawa et al. 1990; Madeira et al. 1997; Sanna et al. 1993). In contrast, with exposure to chronic alcohol there is an accumulation of VP protein in the hypothalamic neuronal nuclei of mice (Carmona-Calero et al. 1995). Thus, the amount of alcohol exposure as well as withdrawal impact brain VP.

3. VP and tolerance to alcohol

One of the first alcohol related behaviors to be associated with VP was tolerance (see table 1 for summary). This effect was shown by giving systemic injections of VP to C57BL/6 mice during and after alcohol treatment and seeing that tolerance to the hypothermic and sedative-hypnotic effects of alcohol was prolonged (Hoffman et al. 1978). Further support comes from studies that use DGAVP, the VP fragment (des- Gly⁹-[Arg⁸]-vasopressin dicitrate) which is thought to have reduced peripheral endocrinological activity compared to full VP (Hoffman 1982; Rigter et al. 1980a; Rigter et al. 1980b; Walter et al. 1975). Acute injection of DGAVP in mice (Rigter et al. 1980a), continuous systemic infusion of DGAVP in mice (Rigter et al. 1980b), daily systemic injections of DGAVP in mice (Hoffman 1982), and systemic daily injections of DGAVP in rats (Le et al. 1982) had similar effect on alcohol tolerance to that found with full length VP. Systemic injections of VP or DGAVP in mice also prolonged tolerance to the motor incoordinating effects of ethanol, but this effect was not prolonged as long as the hypothermic and sedative-hypnotic actions (Hoffman and Tabakoff 1984). Intraventricular injection of VP in mice also prolongs tolerance to the hypnotic effects of alcohol (Hung et al. 1984). Finally, it was found that V1 receptors are more likely to mediate this effect since intraventricular injections of the V1 agonist prolongs tolerance and V1 antagonists enhanced the rate of loss of alcohol tolerance. V2 receptor equivalents do not share these effects (Szabo et al. 1988). Intraventricular injections have also blocked the acquisition of environment dependent tolerance to the hypnotic and hypothermic effects of alcohol in a mouse learning paradigm (Mannix et al. 1986). The authors of this study highlight the use of alcohol injections in a standardized environment every day to create a specific type of tolerance, which they have demonstrated to be environment-dependent. This strategy differs from most of the other experiments which use liquid diet, a method the authors claim does not produce environment dependent tolerance (Melchior and Tabakoff 1985). Therefore, while the mechanism of VP actions in alcohol tolerance may not be clear, these effects might depend on the type of tolerance with VP prolonging environment-independent tolerance and blocking environment-dependent tolerance.

Table 1:

Summary of studies exploring the effects of VP on alcohol tolerance

Paper	Species/sex	Ethanol Paradigm	Tolerance Paradigm	Drug	Results
Hoffman et al. 1978	Mice male	7 days 7% (v/v) ethanol liquid diet and 3g/kg ip challenge	hypothermic; sedative-hypnotic	SC VP	Prolongs tolerance
Rigter et al. 1980a Rigter et al. 1980b	Mice male	3 day ethanol vapor and 3g/kg ip challenge	hypothermic	SC DGAVP	Enhanced tolerance
Lê et al. 1982	Mice male	25 days of oral 5g/kg ethanol and 3g/kg ip challenge	hypothermic; motor coordination	SC DGAVP	Prolongs tolerance
Hoffman 1982	Mice male	7 days 7% (v/v) ethanol liquid diet and 3.1g/kg ip challenge	sedative-hypnotic	SC DGAVP	Prolongs tolerance
Hoffman & Tabakoff 1984	Mice male	7 days 7% (v/v) ethanol liquid diet and 2.3g/kg (14.9% v/v ethanol) ip challenge	motor coordination	SC VP & DGAVP	Prolongs tolerance
Hung et al. 1984	Mice male	5 days 7% (v/v) ethanol liquid diet and 3.1g/kg ip challenge	sedative-hypnotic	ICV VP	Prolongs tolerance
Mannix et al. 1986	Mice male	3.5g/kg ip twice daily for two 5 days blocks separated by 2 days; 3.5g/kg ip twice daily for 4 days and 10μl icv 20% ethanol (2.0mg) challenge	sedative-hypnotic; hypothermic	ICV VP	Blocked the acquisition of tolerance
Szabo et al. 1988	Mice male	5 or 7 days 7% (v/v) ethanol liquid diet and 3.2g/kg ip challenge	sedative-hypnotic	ICV V1 agonist & antagonist or V2 agonist & antagonist	Only the V1 agonist prolonged tolerance and only the V1 antagonist enhanced the rate of loss of tolerance

IP=intraperitoneal injection, SC=subcutaneous injection, ICV=intraventricular injection

VP=arginine vasopressin

DGAVP=des-Gly⁹-[Arg⁸]-vasopressin dicitrate, VP fragment with reduced peripheral endocrinological activity (Hoffman 1982; Rigter et al. 1980; Walter et al. 1975).

Altogether the studies discussed above that used DGAVP as well as those studies that used intraventricular injections of VP, suggest VP has its effects on tolerance through actions in brain. However, what brain regions and receptor subtype (V1a vs V1b) mediate VP’s effects of tolerance have thus far not been investigated. The brain regions mediating these effects most likely vary depending on which tolerance effects are studied. Future experiments on alcohol tolerance should focus on these earlier factors.

4. VP and alcohol drinking

Another alcohol related behavior to be associated with VP is alcohol drinking (see table 2 for summary). Some evidence for the role of VP in drinking behavior comes from the fact that rat strains that were selectively bred for their alcohol preference have higher levels of VP (Hwang et al. 1998; Zhou et al. 2011). For example, both alcohol-preferring (P) and high alcohol-drinking (HAD) rats have been shown to have elevated levels of VP mRNA in the hypothalamus (Hwang et al. 1998). Sardinian alcohol-preferring rats have more VP mRNA in the medial amygdala (MeA)/CeA and hypothalamus. Drinking alcohol decreased the levels of VP mRNA (Zhou et al. 2011). In agreement with the rat data, C57BL6/J mice with the highest levels of VP mRNA in the amygdala drank the greatest amount of alcohol (Nelson et al. 2018).

Table 2:

Summary of the studies examining the effects of VP on alcohol drinking

Paper	Species/sex	Paradigm	Drug	Results
Kornet et al. 1991	Rhesus monkey male	Free choice (1% and 2% v/v) or ( 4% and 8% v/v) alcohol solution for 14 days with drug treatment, followed by 14 days without treatment	IM DGVAP	Reduced alcohol intake
Kornet et al. 1991	Rhesus monkey male	Free choice 1% and 2% v/v alcohol solution for 14 days with drug treatment followed by 14 days without treatment	IM DGVAP	Reduced alcohol intake
Hwang et al. 1998	HAD & P rats male	N/A	N/A	Increased VP mRNA in hypothalamus of HAD and P rats
Caldwell et al. 2006	V1a & V1b KO mice male & female	2-bottle choice (3 to 11% v/v alcohol vs water) for 30–36 days	N/A	V1a heterozygotes consumed more alcohol; V1b KO no difference from WT; no sex related differences
Sanbe et al. 2008	V1a KO mice male	2-bottle choice (4 to 16% v/v alcohol vs water) for 21 days	N/A	KO consumed more alcohol and showed a higher preference for alcohol
Zhou et al. 2011	sP rats male	2-bottle choice (10% v/v alcohol vs water) for 4 weeks	IP SSR149415	Increased VP mRNA in hypothalamus and amygdala of sP rats; Drug reduced alcohol intake
Maher et al. 2011	Humans	N/A	N/A	V1a receptor gene SNP had a modest association with maximum drinks per day
Edwards et al. 2012	Wistar rats male	Operant self-administration of 10% w/v alcohol solution then intermittent exposure to ethanol vapor 4–6 weeks followed by operant self-administration	IP SSR149415	Drug reduced alcohol intake; Self-administration lowered V1b in the amygdala, but restored V1b levels if rat was alcohol dependent
Ryan et al. 2017	Humans male & female	Individuals who met DSM-IV criteria for alcohol dependence received 12 weeks of drug treatment	Oral ABT-436	Drug reduced frequency of drinking; no gender related differences
Zhou et al. 2018	C57BL/6J mice male & female	Paradigm 1– 3 weeks 2‐bottle choice paradigm with chronic alcohol (7.5, 15, or 30%) exposure every other day; Paradigm 2– 3 weeks 1 bottle choice of 15% alcohol every day for 4 hours in the dark	IP SSR149415	Drug reduced alcohol intake alone or in combination with Naltrexone; no sex related differences
Nelson et al. 2018	C57BL/6J mice male	2‐bottle choice paradigm with 20 % v/v alcohol solution	N/A	VP levels in the amygdala correlated to the amount of alcohol consumed

IP=intraperitoneal injection, SC=subcutaneous injection, IM=intramuscular injection

VP=arginine vasopressin

DGAVP=des-Gly⁹-[Arg⁸]-vasopressin dicitrate, VP fragment with reduced peripheral endocrinological activity (Hoffman 1982; Rigter et al. 1980; Walter et al. 1975)

V1b antagonist=SSR149415 & ABT-436

An association between the V1 receptor subtypes and alcohol drinking has also been established. In humans a modest association was found between a single nucleotide polymorphism in the V1a receptor gene and maximum drinks per in a 24 hour period in European-American males (Maher et al. 2011). A mouse knock-out of the V1a receptor on a mixed 129Sv and C57BL/6Cr Slc background increased alcohol consumption and preference for alcohol in the two bottle choice paradigm (Sanbe et al. 2008), but in a second study using a mixed C57BL/6J and 129/Sv-CP background only the heterozygote for the V1a knock-out showed increased alcohol consumption in the two bottle choice paradigm (Caldwell et al. 2006). Difference in genetic background of the knock-outs could account for the differences in results in the homozygous mice. A relationship has also been demonstrated between the V1b receptor and alcohol dependence when it was discovered that acute alcohol consumption in nondependent Wistar rats decreased V1b receptor protein in the basolateral amygdala (BLA), but once rats become dependent, alcohol consumption restores the protein levels of the receptor (Edwards et al. 2012). The authors of this study suggest that this change in the response of the V1b receptor to alcohol might be involved in the transition to dependence. On the other hand, when tested on the two bottle choice paradigm V1b knock-out mice showed similar alcohol consumption and preference as their wild-type littermates (Caldwell et al. 2006). In the case of knock-outs of either receptor subtype there is the chance of compensatory changes in the levels of other receptors that can conserve the behavior, for example increased expression of the oxytocin receptor can compensate for the loss of the V1b receptor to allow for normal ACTH release in the V1b receptor knock-out (Nakamura et al. 2008). Thus, VP impacts drinking in complex ways needing further elucidation.

The studies discussed in the two paragraphs above use designs that mainly demonstrate a correlation, but leave cause and effect unclear. Evidence for a causal role of VP in drinking came from studies that used VP antagonists to alter drinking. When the V1b receptor antagonist is administered systemically, Sardinian alcohol-preferring rats decreased their alcohol intake (Zhou et al. 2011). This effect extends to other rats strains, as the V1b receptor antagonist given systemically also reduced drinking in alcohol dependent Wistar rats (Edwards et al. 2012). Additionally, the V1b antagonist blocked drinking in C57BL/6 mice alone at high doses and at lower doses in combination with naltrexone (Zhou et al. 2018). But perhaps, most importantly, a recent clinical trial demonstrated a significantly greater percentage of days abstinent following V1b receptor antagonism for individuals who met the criteria for DSM-IV alcohol dependence (Ryan et al. 2017). Conversely, 14 weeks of alcohol exposure with concurrent systemic DGAVP administration, the VP fragment that has reduced peripheral endocrinological activity compared to full VP, caused long lasting decreases in alcohol drinking in rhesus monkeys (Kornet et al. 1991; Kornet et al. 1992). One important difference between the Kornet et al (1991; 1992) studies and the other studies was that while the other studies initiated alcohol exposure followed by testing of the antagonist, this study did concurrent DGAVP administration with initial alcohol exposure. These results suggest that timing of the V1b antagonist might be important and therefore a V1b antagonist might have different effects before and after animals have transitioned to alcohol dependence. This interpretation is consistent with the findings of Edwards et al (2012) that there is a change in V1b receptor response after transition to alcohol dependence. Altogether these observations provide strong evidence for the role of VP in alcohol drinking behavior at least partially through the actions of the V1b receptor.

All of the pharmacological studies discussed above were done with systemic injections, and therefore do not give an indication as to which brain areas might be involved in these effects. As noted above, the amygdala is a source of VP (Otero-Garcia et al. 2016; Sofroniew 1983; Song and Albers 2017) and alcohol preferring rats have more VP in the amygdala than alcohol non-preferring rats (Zhou et al. 2011). The amygdala is a known contributor to drinking behavior (de Guglielmo et al. 2016; McBride 2002) and changes in the V1b receptor occur with drinking in the basolateral amygdala (Edwards et al. 2012). As V1b receptor antagonists have shown strong effects on drinking, these observations make the amygdala and the V1b receptor an attractive target for future work exploring the role of VP in alcohol drinking. The V1a receptor is a less attractive target as both heterozygous and homozygous V1a knock-outs increased alcohol consumption (Caldwell et al. 2006; Sanbe et al. 2008), but this work was only done in traditional knock-outs and more work using conditional knock-out approaches or pharmacological agents are warranted.

The diagnostic and statistical manual of mental disorders, 5^th edition (DSM-5) criteria for alcohol use disorders (AUD) includes drinking related criteria like “had times when you ended up drinking more, or longer, than you intended?” and “spent a lot of time drinking? Or being sick or getting over other after effects” (APA 2013), so unsurprisingly alcohol drinking paradigms like dependence induced excessive drinking have been suggested as models for testing treatments for alcoholism (Litten et al. 2012). Therefore, the effects of VP antagonists on dependence induced drinking could be a good indication for effectiveness as a treatment for AUD.

5. VP and stress-potential relationship to anxiety associated with alcohol

Interestingly, stress also alters the VP system. The forced swim test has been shown to cause release of VP in the amygdala (Ebner et al. 2002) and the septum (Ebner et al. 1999). Acute immobilization stress, repeated immobilization stress, chronic foot shock or long term social isolation all cause an increase in VP mRNA and VP immunoreactive neurons in the hypothalamus (Bartanusz et al. 1993; de Goeij et al. 1992; Ma and Lightman 1998; Makino et al. 1995; Pan et al. 2009; Pinnock and Herbert 2001; Sawchenko et al. 1993). Stress also increased the levels of V1b receptor mRNA (Aguilera and Rabadan-Diehl 2000; Litvin et al. 2011; Rabadan-Diehl et al. 1995) as well as V1a receptor binding and mRNA (Duque-Wilckens et al. 2016; Gray et al. 2012; Milik et al. 2014). Stress is also associated with alcohol withdrawal induced anxiety and with relapse in alcoholics (Breese et al. 2005a; Breese et al. 2011; Sinha 2007). The stress related changes in VP in alcoholics could lead to changes in the HPA and potentially impact alcohol related behaviors like drinking and alcohol withdrawal induced anxiety. Drugs that treat stress related disorders like anxiety could then potentially have an impact on treating alcoholics during stressful situations. In fact, the VP response to stress is related to the ability of VP to share a role with CRF in regulating the HPA axis response to stressors including alcohol. Blocking VP using either a passive immunoneutralization of VP or a V1 antagonist diminishes ACTH secretion caused by alcohol (Ogilvie et al. 1997; Rivier and Lee 1996), but other studies using the same immunoneutralization did not block alcohol-induced CORT or ACTH (Thiagarajan et al. 1989). Since using a CRF or VP antagonist alone could only blunt the alcohol induced ACTH secretion, one group gave both a CRF and VP antagonist and was able to abolish the alcohol induced ACTH section (Rivier and Lee 1996). The ability of VP to regulate release of ACTH and CORT in response to alcohol might also be partially mediated through the V1b receptor as the V1b receptor knock-out has reduced ACTH and CORT plasma levels compared to wild-types in response to acute alcohol (Lolait et al. 2007b). The V1b receptor knock-out mouse also has a reduced ACTH and CORT response to physical stressors such as restraint stress, the forced swim test, and insulin administration (Lolait et al. 2007a; Roper et al. 2010). Thus, stress and alcohol share adaptations in V1b regulation of the HPA axis.

Stressors like social isolation and social defeat that cause changes in the VP system also cause anxiety-like behaviors. Table 3 summarizes studies discussing the relation between VP and anxiety. The V1b receptor antagonist given systemically during stress blocks the effects of stressors on anxiety tasks like social defeat stress-induced anxiety in the elevated plus maze, chronic mild stress-induced anxiety in the elevated plus maze, and social defeat stress-induced anxiety on the social interaction task (Griebel et al. 2002; Litvin et al. 2011). Systemic administration of V1b receptor antagonists also alters nonstress related performance on elevated plus maze, light/dark box, punished drinking conflict test, conditioned lick suppression, social interaction test, and four-plate test (Griebel et al. 2002; Hodgson et al. 2007; Hodgson et al. 2014; Iijima et al. 2014; Overstreet and Griebel 2005; Serradeil-Le Gal et al. 2002; Shimazaki et al. 2006). Systemic administration of a V1a receptor antagonist reduced anxiety on the elevated plus maze, elevated zero maze, marble burying test, and conditioned lick suppression (Bleickardt et al. 2009). Systemic administration of a VP agonist causes anxiety on the elevated plus maze, which is thought to be mediated through the V1a receptor (Mak et al. 2012). Mice bred for high anxiety were found to have a polymorphism in their VP gene leading to higher expression of VP while the reverse is true of the mice bred for low anxiety (Bunck et al. 2009; Murgatroyd et al. 2004). However, as with alcohol drinking, V1b knock-out mice show similar anxiety-like behavior on the open field, elevated plus maze and light/dark box as compared to their wild-type littermates (Egashira et al. 2005; Wersinger et al. 2002). On the other hand, male, but not female, V1a knock-out mice demonstrated reduced anxiety in the elevated open maze, light/dark box, open field, and marble burying test (Bielsky et al. 2004; Bielsky et al. 2005; Egashira et al. 2007). As mentioned above, interpretations of results of studies with knock-out animals are confounded by developmental shifts, but in general support VP involvement in anxiety.

Table 3:

Summary of the studies examining the effects of VP on anxiety

Paper	Species/sex	Paradigm	Drug	Results
Landgraf et al. 1995	Wistar rats male	On the 4th day of infusion rats were tested in the elevated plus maze	Intra-LS V1a antisense	Loss of V1a receptor reduced anxiety on the elevated plus maze
Liebsch et al. 1996	Wistar rats male	Infusions via microdialysis were performed over 30 min and immediately following infusions rats were tested on the elevated plus maze	Intra-LS VP or manning compound	VP into the LS did not alter anxiety on the elevated plus maze; V1a antagonism decreased anxiety on the elevated plus maze
Appenrodt et al. 1998	Wistar rats male	Infusions via microdialysis were performed over 30 min, 25 min into infusion rats were tested for 5 min on the elevated plus maze; For IP injections elevated plus maze testing was done 30 min post-injection	Intra-LS VP or VP antagonist or IP VP	Both injections of VP into the LS and IP caused decreased anxiety in the elevated plus maze; General VP antagonism had no effect on anxiety on the elevated plus maze
Everts and Koolhaas 1999	Wistar rats male	On the 11th day of drug infusion rats were tested in the shock probe burying test and elevated plus maze	Intra-LS VP antagonist	General VP receptor antagonism increased anxiety on the elevated plus maze, but no increase noted in the shock probe burying test
Griebel et al. 2002	SpragueDawley or Wistar rats male	Punished drinking test 30 mins post-drug injection; elevated plus-maze 1, 1.4, 3 or 6 hrs post-drug injection	IP or PO SSR149415	V1b antagonism decreased anxiety-like behavior on both tests
Griebel et al. 2002	CD1, OF1, and BALB/c mice male	Paradigm 1- light-dark test 30 min post-drug injection; Paradigm 2– 60 mins post-drug administration mice exposed for 60 min in residents cage prior to testing on the elevated plus maze; Paradigm 3– 7 weeks chronic mild stress treatment once a day and beginning at 4 weeks mice received drug treatment that continued through testing on the elevated plus maze	IP or PO SSR149415	V1b antagonism decreased anxiety-like behavior on light-dark test; V1b antagonism partially reversed the anxiety-like behaviors caused by social defeat and chronic mild stress
Serradeil-Le et a.l 2002	NMRI mice male	Four plate test was done. 5, 1, 2, 4, or 6 hr after drug was given	IP or PO SSR149415	V1b antagonism decreased anxiety in the four plate test
Wersinger et al. 2002	V1b KO mice male	Mice were tested on the elevated plus maze	N/A	Loss of V1b receptor did not alter anxiety on the elevated plus maze
Bielsky et al. 2004	V1a receptor KO mice male	KOs were tested on the elevated plus maze, open field, and light-dark box	N/A	Loss of V1a receptor reduced anxiety on the elevated plus maze, open field and the light-dark box
Murgatroyd et al. 2004	Wistar rats male	Wistar rats were bred for high (HAB) or low (LAB) anxiety-like behavior on the elevated plus maze	N/A	HAB rats have increased expression of VP gene in the PVN compared to LAB rats
Bielsky et al. 2004	V1a KO mice female	Mice were tested in the elevated plus maze, open field and the light-dark box	N/A	Loss of V1a receptor did not change anxiety on the elevated plus maze, open field or light-dark box
Egashira et al. 2005	V1b KO mice male	Mice were tested on the elevated plus maze and light-dark box	N/A	Loss of V1b receptor did not change anxiety on the elevated plus maze or light-dark box
Stemmelin et al. 2005	SpragueDawley rats male	10 min post-injections rats were tested on the punished drinking test and the elevated plus maze	Intra-LS SSR149415	V1b antagonism in the LS did not change anxiety on the punished drinking test or the elevated plus maze
Overstreet & Griebel 2005	Flinders Sensitive Line rats	Rats injected once daily for 14 days, rats were then tested on the social interaction test 20 hrs after the last injection	IP SSR149415	V1b antagonisms decreased anxiety behavior in the social interaction test
Shimazaki et al. 2006	SpragueDawley rats male	Following 7 days of isolate housing rats were administered drug followed by social interaction test 1 hr post-drug administration	PO SSR149415	V1b antagonism decreased anxiety in the social interaction test
Salome et al. 2006	SpragueDawley rats male	10 min post-injection rats were tested on the elevated plus maze	Intra-CeA & Intra-BLA SSR149415	V1b antagonism only in the BLA reduced anxiety on the elevated plus maze
Caldwell et al. 2006	V1a & V1b KO mice male & female	Alcohol 20%v/v in.9% saline IP at 1.0, 1.25 or 1.5 g/kg immediately before being placed in the elevated plus maze	N/A	V1a and V1b KO show no difference in anxiety on the elevated plus maze compared to WT; no sex related differences
Egashira et al. 2007	V1a KO mice male	Mice were tested on the elevated plus maze and marble burying test	N/A	Loss of V1a receptor reduced anxiety on the elevated plus maze and marble burying test
Hodgson et al. 2007	CD rats & CD1 mice male	Tests were performed 15 min post-injection of drug; Rats were tested in the elevated plus maze & conditioned lick suppression; Mice were tested in the marble burying task	IP SSR149415	V1b antagonism decreased anxiety in the elevated plus maze and lick suppression task in rats, but not the marble burying in mice
Bleickardt et al. 2009	CD or Sprague-Dawley rats	Drug was injected 30 min before all behavioral tests were given; Rats performed	IP JNJ-17308616	V1a antagonism decreased anxiety in all tests of rats and mice
	male & CD1 mice male	elevated plus-maze, elevated zero-maze and conditioned lick and mice performed the marble burying task
Bunck et al. 2009	Swiss CD1 mice male & female	Mice were selectively bred for high (HAB) and low (LAB) anxiety-like behavior on the elevated plus maze	N/A	LAB mice have reduced expression of VP gene in the PVN and SON; no sex related differences
Litvin et al. 2011	Swiss-Webster mice male	For 10 days mice received 10 min exposure to a novel resident male followed by 10 min separation by partition for 10 days. On the 11th day mice received 24 hours single housing followed by a drug injection one hour prior to the social investigation paradigm	IP SSR149415	Socially defeated mice showed anxiety-like behaviors in the social investigation paradigm which was partly reversed by V1b antagonist
Mak et al. 2012	SpragueDawley rats male	Rats were tested on the elevated plus maze 15 min post-drug injection	IV desmopressin	VP agonist causes anxiety on the elevated plus maze
Iijima et al. 2014	SpragueDawley rats male	Paradigm 1- Social interaction test was performed 1 hr post-drug administration; Paradigm 2– 1 hr post-drug administration the elevated plus maze measured stress induced by swim stress which was performed 5 min post-swim	PO TASP0233278	V1b antagonism decreased anxiety in the social interaction test and reversed anxiety caused by swim stress
Hodgson et al. 2014	CD rats male	Conditioned lick suppression 30 min postdrug injection	IP V1B-30N	V1b antagonism decreased anxiety in the conditioned lick suppression test
Duque-Wilckens et al. 2016	California mice male & female	30 min post-injection mice were tested in the social interaction test	Intra-BST manning compound	V1a antagonism in the BNST caused anxiety on the social interaction test; no sex related differences
Hernandez-Perez et al. 2018	Wistar rats male	Immediately following microinjections rats were tested on either the light-dark box or the shock probe burying test	Intra-CeA VP, VP/SSR149415 or VP/manning compound	VP alone caused anxiety only in the shock probe burying test; V1b antagonisms blocked the effects of VP on the shock probe burying test
Dannenhoffer et al. 2018	SpragueDawley rats male & female	4g/kg every 48 h from postnatal day 25–45. At least 25 days after last exposure rats were tested on the social interaction test	IP SR-49059 SSR149415	Only males showed anxiety in the social interaction test and it was blocked by the V1b antagonist

Studies in bold are related to alcohol associated anxiety.

IP=intraperitoneal injection, IV=intravenous injection, PO=oral

VP=arginine vasopressin

V1a antagonist=manning compound, SR-49059 & JNJ-17308616 V1b antagonist=SSR149415, TASP0233278 & V1B-30N desmopressin=nonselective VP agonist

BLA=basolateral amygdala, BST=bed nucleus of the stria terminalis, CeA=central nucleus of the amygdala, LS=lateral septum, PVN=Paraventricular nucleus

A V1a antagonist administered via microdialysis probe (Liebsch et al. 1996) or antisense to the V1a receptor mRNA administered via osmotic minipump (Landgraf et al. 1995) in the Septum decreased anxiety related behaviors on the elevated plus maze, but microinjection of the V1b antagonist into the Septum had no effect on the elevated plus maze or the punished drinking conflict test (Stemmelin et al. 2005). Surprisingly injections of VP into the septum and intraperitoneally also decreased anxiety-like behavior in the elevated plus maze (Appenrodt et al. 1998) while general VP antagonism in the lateral septum increased anxiety in the elevated plus maze (Everts and Koolhaas 1999). As the studies overlap in anxiety test and rat strains used, one author suggested that the opposite findings between studies might be methodical variations that can affect behavior such as when microdialysis was done in relation to when behavior was tested. It should also be noted that in the case of Liebsch et al (1996) and Landgraf et al (1995) the rats were purchased from Charles River in Germany while Everts & Koolhaas (1999) breed their rats in house. Interestingly, microinjections of the V1b antagonist into the basolateral amygdala, but not the CeA, decreased anxiety on the elevated plus maze (Salome et al. 2006), but only microinfusion of the V1b antagonist and not the V1a antagonist into the CeA blocked the effects of VP in Wistar rats in the shock-probe burying test and the light-dark box (Hernandez-Perez et al. 2018). V1a receptor antagonists in the BST cause anxiety in the open field and social interaction test (Duque-Wilckens et al. 2016). Altogether there appears to be a complex relationship between brain region, receptor subtype and type of anxiety tested.

The role of VP in alcohol associated anxiety is being investigated as well. Intermittent alcohol exposure causes long term changes after exposure to alcohol has ended including increased drinking and anxiety on the social interaction test (Crews et al. 2016; Dannenhoffer et al. 2018). Twenty-five to twenty-seven days after the end of alcohol exposure anxiety is still apparent in males only on the social interaction test and this anxiety can be blocked by systemic V1b antagonists, but not V1a antagonist (Dannenhoffer et al. 2018). In contrast to this study, V1a and V1b knock-out mice of both sexes show no difference in time in the open arm of the elevated plus maze in response to acute alcohol (Caldwell et al. 2006). As the intermittent alcohol exposure paradigm takes place in adolescence, e.g. ages P25 to P45, the developmental period as well as the length of alcohol exposure could cause changes in the VP system, such as increased V1b receptors (Dannenhoffer et al. 2018), that lead to a different VP related anxiety pathway than anxiety related to acute alcohol exposure. Adolescent alcohol has been found to have long lasting effects not found with similar adult alcohol exposure (Crews et al. 2016).

A second chronic intermittent alcohol paradigm has a relationship to both alcohol withdrawal and stress, both of which have profound effects on the VP system. This paradigm requires three withdrawals to obtain robust anxiety-like behavior, a mild stressor can substitute for the first two withdrawals to cause anxiety-like behavior (Breese et al. 2005a; Overstreet et al. 2002). Additionally, in an extended withdrawal which alone no longer causes anxiety, a mild stressor can reinstate the anxiety (Breese et al. 2005b). Different pharmacological agents, like CRF, that activate systems associated with stress and alcohol withdrawal have also been shown to substitute for alcohol/stress in these paradigms (Breese and Knapp 2016). We predict that VP would have a similar effect, and that blocking V1b receptors during withdrawal could block alcohol withdrawal induced anxiety-like behavior (see Dannenhoffer et al. 2018). Along with examining alcohol withdrawal induced anxiety, more work needs to be done to examine brain regions involved in the role in VP’s effects on alcohol associated anxiety. The adolescent study that used an alcohol paradigm to cause anxiety found changes in the V1b receptor in the hypothalamus suggesting a role for this brain region while also showing that the anxiety caused by this paradigm could be reversed by a systemically injected V1b antagonist. Thus, this finding suggests, but does not confirm, that the V1b antagonist might have its effects through the hypothalamus (Dannenhoffer et al. 2018). Studies showing anxiety not associated with alcohol as discussed above would suggest the amygdala and septum as strong targets as well (Everts and Koolhaas 1999; Hernandez-Perez et al. 2018; Landgraf et al. 1995; Liebsch et al. 1996; Salome et al. 2006; Stemmelin et al. 2005).

6. VP and clinical trials

The evidence discussed above and evidence from animal models of other neuropsychiatric disorders has led to four human drug trials for VP antagonists in stress related neuropsychiatric disorders. Unfortunately only one of these antagonists has been tested for treatment of alcohol abuse. It is important to note two things: first that the four drugs are orally active and second that they passed safety trials with mild side effects at most (Fabio et al. 2013; Griebel et al. 2012; Katz et al. 2016). The clinical trials are summarized in table 4.

Table 4:

Summary of VP antagonist clinical trials

Drug	Receptor target	Trial	Disorder	Status
ABT-436	V1b	NCT01613014 NCT01741142 NCT01380704	-Alcohol Abuse -Major Depressive Disorder	Completed
SSR149415	V1b	NCT00374166 NCT00358631 NCT00361491 NCT01606384	-Anxiety Disorders -Major Depressive Disorder	Completed
TS-121	V1b	NCT03093025	-Major Depressive Disorder	Ongoing
SRX246	V1a	NCT02055638 NCT02733614 NCT03036397 NCT02922166	-Intermittent Explosive Disorder -PTSD -Anxiety Disorders/Fear	Ongoing

ABT-436 is a V1b receptor antagonist that has reached phase II clinical trials for major depressive disorder and has thus far been the only VP antagonist tested against alcohol abuse. In trials for patients with major depressive disorder, patients were excluded if they had current or recent drug or alcohol abuse, or a psychotic disorder. After one week ABT-436 significantly improved performance on the Mood and Anxiety Symptom Questionnaire, but not the 17-item Hamilton Depression Rating Scale (Katz et al. 2017). In trials to test the effects of ABT-436 on alcohol abuse, participants were chosen that were heavy drinkers that had been diagnosed with alcohol dependence in the past year. Participants were excluded if they were diagnosed with a DSM axis I disorder or were currently abusing or dependent on a psychoactive drug other than alcohol. When tested using the Spielberger trait anxiety index (STAI) and profile of mood states (POMS) total mood disturbance, participants selected for the study had near normal anxiety with slightly elevated mood disturbance. ABT-436 reduced the frequency of drinking without reducing the amount of alcohol consumed (Ryan et al. 2017). It should also be noted that the drug seemed to have its greatest effect on patients with high stress levels as measured by the POMS tension-anxiety, the STAI, and peak cortisol levels during an ACTH stimulation test. This finding lends support for the authors’ idea that ABT-436 might have its effects through the VP stress response. Despite these somewhat promising outcomes, AbbVie has discontinued the development of ABT-436.

SSR149415 is a V1b receptor antagonist that has undergone phase II clinical trials for generalized anxiety disorder and major depressive disorder. In the clinical trial for generalized anxiety disorder the drug caused no significant changes on the Hamilton Anxiety Rating Scale or the Clinical Global Impressions-Severity of Illness score (Griebel et al. 2012). Two of the clinical trials in depressed patients showed no improvement on the Hamilton Depression Rating Scale or the Clinical Global Impressions-Severity of Illness score, although a third trial in depressed patients showed improvement on the Hamilton Depression Rating Scale without effects on Montgomery-Asberg Depression Rating Scale (Griebel et al. 2012). Sanofi has halted further drug development, but the drug continues to be used for animal experiments including many of the experiments discussed above, which show strong effects on behavior. These clinical trials have not directly tested alcohol induced anxiety, but would appear to provide safe antagonists for such a clinical trial

Two other VP antagonists that have been pursued are TS-121 and SRX46. TS-121 is a V1b antagonist developed by Taisho Pharmaceutical. As TS-121 has only begin phase II clinical trials for major depressive disorder- it is too early to draw any conclusions about this drug. SRX246 was developed by Azevan Pharmaceuticals. SRX246 is unique in that it is the only V1a antagonist that has gone to phase II clinical trials for these stress related neuropsychiatric disorders. These phase II clinical trials for anxiety disorder and post-traumatic stress disorder are currently ongoing and therefore we cannot draw conclusions. However, these results will be interesting in that V1a receptors are expressed throughout the brain while the V1b receptor has a more limited expression.

7. Conclusions and future directions

It is clear from the evidence provided that VP plays a role in many alcohol related behaviors and might therefore make an excellent target for treating alcohol abuse disorders. Evidence from human studies examining the effects of the V1b antagonists on drinking have shown promising results (Ryan et al. 2017) and more VP antagonists are currently being tested for other neuropsychiatric disorders. Nonetheless, future testing of VP antagonists in a variety of human disorders is needed. Additionally, there is still much to be accomplished to identify which brain regions are involved in the effects of VP on alcohol related behaviors. It is clear that alcohol changes the levels of VP in the hypothalamus and perhaps other brain regions (Carmona-Calero et al. 1995; Gulya et al. 1991; Ogilvie et al. 1997; Sanna et al. 1993; Zhou et al. 2011), a finding which implicates these brain regions in alcohol related behaviors associated with release of VP (summarized in figure 1). What is less clear is in which brain regions the V1b receptor antagonist is having its effects as most studies give the antagonist systemically and whether V1a antagonism would be effective on any of these alcohol related behaviors as it has been effective in nonalcohol related anxiety.

Despite the existence of sex differences in the VP system, few alcohol studies have included both sexes. Most of these studies have not found any differences between the sexes though these studies were mostly targeting the V1b system (Caldwell et al. 2006; Ryan et al. 2017; Zhou et al. 2018). The data from Dannenhoffer et al. (2018) are harder to interpret as their alcohol paradigm caused a behavioral and molecular changes in males that were not found in females, and those changes could relate to the differential effect of the V1b antagonist on males and females. However, the data from the other studies examining both sexes would suggest that VP, or at least the V1b receptor, would be a valid treatment option for both sexes, but more work needs to be done including both sexes.

There are likely multiple factors that contribute to drinking in alcoholics as suggested by the study showing that a V1b antagonist decreased the number of drinking days, but not the total alcohol intake in humans (Ryan et al. 2017). It is reasonable that variables controlling an abstinent alcoholic taking their first drink since becoming abstinent (number of drinking days) are different than those responsible for continued drinking after that first drink (total alcohol intake) and that separate treatments are needed for each. Thus a multidrug program may be needed to effectively treat alcoholism as suggested by the work of Zhou et al (2018) showing a decrease in alcohol intake in mice with a combination of naltrexone and a V1b antagonist. Given the results in the human trials of the V1b antagonist, it might be best to combine it with a drug that affects the total alcohol intake. However, in order to successfully employ the correct combination of drugs, a better understanding is needed of how VP affects alcohol related behaviors. As mentioned in Zhou et al (2018), naltrexone and the V1b antagonist seem to effect different systems, which likely contributes to their combined effectiveness.

An additional future direction of this research would be to focus on the stress related aspects of drinking based on the results in the human trials showing the greatest effects of the V1b antagonists in humans with high stress (Ryan et al. 2017). Further support for this research direction comes from studies showing a correlation between response to stress, the measurement of VP levels in the amygdala and alcohol consumption (Nelson et al. 2018). Others have suggested that VP effects alcohol drinking through its effects on the HPA axis, and therefore plays a role in the relationship between stress-reactivity and alcoholism (Zhou and Kreek 2018), which is supported by evidence that blocking the VP system decreases the HPA response to alcohol (Lolait et al. 2007b; Ogilvie et al. 1997; Rivier and Lee 1996). Altogether these ideas suggest that perhaps VP antagonists might be most beneficial in situations with a stress aspect and that feature might be the best target for future human trails.

Finally, looking to OT and CRF human drug trial failures and successes can help strengthen the agreement for further avenues for VP trials. For example, OT has been successful in treating the symptoms of withdrawal perhaps due to its rapid reversal of tolerance (Pedersen et al. 2013), therefore given the ability of a V1 antagonist to enhance the rate of loss of tolerance maybe V1 antagonists should be tested for use during alcohol withdrawal as well. It has been suggested that there are many reasons for the failure of the CRF antagonist human trials among them are timing (the drug should have been giving during withdrawal) and selective treatment populations (for example high stress individuals) both of which are most likely relevant to the VP antagonists as well (Schreiber and Gilpin 2018; Spierling and Zorrilla 2017). Therefore, using VP antagonists during withdrawal as well as in high stress patients should be considered as future direction in human drug trials.

In conclusion, VP is an attractive target for treatment of alcohol disorders, but more basic research needs to be done to understand how VP effects alcohol related behaviors, so that the correct application for VP treatment might be found.

List of Abbreviations

AA

alcohol preferring

ACTH

Adrenocorticotropic hormone

AMY

Amygdala

ANA

alcohol non-preferring rats

AUD

alcohol use disorders

BST

Bed nucleus of the stria terminalis

cAMP

cyclic adenosine 3’,5’-monophosphate

CeA

central nucleus of the amygdala

CORT

corticosterone

CRF

corticotropin- releasing factor

CRF-R1

corticotropin-releasing factor receptor 1

CRF-R2

corticotropin-releasing factor receptor 2

DGAVP

des- Gly⁹-[Arg⁸]-vasopressin dicitrate

DMH

Dorsomedial hypothalamic nucleus

DR

Dorsal raphe

DSM-5

diagnostic and statistical manual of mental disorders, 5^th edition

GABA

gamma-Aminobutyric acid

HAD

high alcohol-drinking rats

HPA

hypothalamic pituitary adrenal

HPC

Hippocampus

LBH

Lateral habenular nucleus

LC

Locus coeruleus

LS

Lateral septum

MeA

medial amygdala

MPO AH

Medial preoptic area – anterior hypothalamus

OT

oxytocin

*OT

Olfactory tubercle

VOLT

vascular organ of the lamina terminalis

P

alcohol-preferring rats

PAG

Periaqueductal grey

PIP2

phosphatidylinositol 4,5-bisphosphate

POMS

profile of mood states (POMS) total mood disturbance

PV

Periventricular nucleus hypothalamus

PVN

Paraventricular nucleus

SCN

Suprachiasmatic nucleus

SON

Supraoptic nucleus

STAI

Spielberger trait anxiety index

V1a

vasopressin receptor 1a

V1b

vasopressin receptor 1b

V2

vasopressin receptor 2

VOLT

vascular organ of the lamina terminalis

VP

Arginine vasopressin

*VP

Ventral pallidum