Vasopressin: Behavioral Roles of an “Original” Neuropeptide

Abstract

Vasopressin (Avp) is mainly synthesized in the magnocellular cells of the hypothalamic supraoptic (SON) and paraventricular nuclei (PVN) whose axons project to the posterior pituitary. Avp is then released into the blood stream upon appropriate stimulation (e.g., hemorrhage or dehydration) to act at the kidneys and blood vessels. The brain also contains several populations of smaller, parvocellular neurons whose projections remain within the brain. These populations are located within the PVN, bed nucleus of the stria terminalis (BNST), medial amygdala (MeA) and suprachiasmatic nucleus (SCN).

Since the 1950's, research examining the roles of Avp in the brain and periphery has intensified. The development of specific agonists and antagonists for Avp receptors has allowed for a better elucidation of its contributions to physiology and behavior. Anatomical, pharmacological and transgenic, including “knockout,” animal studies, have implicated Avp in the regulation of various social behaviors across species.

Avp plays a prominent role in the regulation of aggression, generally of facilitating or promoting it. Affiliation and certain aspects of pair-bonding are also influenced by Avp. Memory, one of the first brain functions of Avp that was investigated, has been implicated especially strongly in social recognition. The roles of Avp in stress, anxiety, and depressive states are areas of active exploration. In this review, we concentrate on the scientific progress that has been made on understanding the role of Avp in regulating of these and other behaviors across species, as well as discuss the implications for human behavior.

1. Introduction

The neurohypophysial hormone arginine vasopressin (Avp) was originally detected by Oliver and Schäfer (1895) who demonstrated that extracts of the pituitary altered blood pressure. Subsequently, the antidiuretic properties of Avp were discovered (Farini, 1913; Vongraven, 1913). However, it was not until du Vigneaud and colleagues (1952) isolated the peptide that specific activities could be ascribed. Following this finding, the nine amino acid sequence and structures of Avp (Turner et al., 1951; Archer and du Vigneaud, 1953; du Vigneaud et al., 1953a) and the related hormone oxytocin (Oxt) (Tuppy, 1953; du Vigneaud et al., 1953b) were elucidated, followed shortly by their synthesis (du Vigneaud et al., 1954a; du Vigneaud et al., 1954b). In 1955, du Vigneaud won the Nobel Prize in Chemistry due, in part, to his early descriptions and syntheses of Avp and Oxt.

Avp is mainly synthesized in the magnocellular cells of the hypothalamic supraoptic (SON) and paraventricular nuclei (PVN) whose axons project to the posterior pituitary (Brownstein et al., 1980; Young and Gainer, 2003). Avp is then released into the blood stream upon appropriate stimulation (e.g., hemorrhage or dehydration) to act at the kidneys and blood vessels (Nishimura and Fan, 2003). The brain also contains several populations of smaller, parvocellular neurons whose projections remain within the brain. These populations are located within the PVN, bed nucleus of the stria terminalis (BNST), medial amygdala (MeA) and suprachiasmatic nucleus (SCN) (Buijs et al., 1978; De Vries and Buijs, 1983; De Vries et al., 1985). Other studies suggest that Avp may be more widely expressed in the brain but at significantly lower levels (Planas et al., 1995; Hallbeck et al., 1999).

Since the 1950's, research examining the roles of Avp in the brain and periphery has intensified. The development of specific agonists and antagonists for Avp receptors has allowed for a better elucidation of its contributions to physiology and behavior (Manning and Sawyer, 1991; Barberis et al., 1999; Serradeil-Le Gal et al., 2002a). Anatomical, pharmacological and transgenic, including “knockout,” animal studies, have implicated Avp in the regulation of various social behaviors across species. Enough scientific progress has been made that we are now gaining a better understanding of Avp from its transcription to its regulation of behavior.

1.1 Structure and evolution of vasopressin

The Avp gene contains three exons and two introns (see Fig. 1). It is on the same chromosome as Oxt (chromosome 2 in mice and chromosome 20 in human), but oriented in the opposite transcriptional direction (Mohr et al., 1988; Hara et al., 1990), implying that these two genes were duplicated during evolutionary development. Avp peptide has a ring structure formed by a disulfide bridge and differs from Oxt at two amino acid residues (Hruby et al., 1990). The two genes are separated by an intergenic region (IGR) which is about 11 kbp in rat and human, and 3.6 kbp in the mouse (Gainer et al., 2001; Young and Gainer, 2003). Regulatory DNA sequences exist within conserved portions of the IGR (Gainer et al., 2001; Young and Gainer, 2003). The preprohormone consists of the signal peptide, the nonapeptide Avp, and the first nine amino acid residues of the neurophysin protein (encoded by first exon); the central part of the neurophysin (encoded by the second exon); and the C-terminal part of the neurophysin as well as a glycopeptide (encoded by the last exon) (Burbach et al., 2001; Young and Gainer, 2003). Avp is evolutionarily well-conserved as even the primitive organism hydra expresses an Avp-like peptide (Grimmelikhuijzen and Spencer, 1984). The evolutionary progenitor of the vertebrate Avp, vasotocin, is found in birds and reptiles (Acher, 1990).

Figure 1.

Schematic diagram of the oxytocin and vasopressin genes (top, large arrows) preprohormones (middle boxes) and peptides (bottom). Both genes are composed of three exons (shown as small solid arrows) separated by two introns. The genes are located on the same chromosome but are transcribed in opposite directions and separated by an intergenic region (IGR). The IGR length varies across species. The preprohormones contain a signal peptide (SP), Avp or Oxt, neurophysin (NP), and a glycopeptide (GP) in the case of Avp. Protein processing signals are represented with thick lines. Cysteine residues form a disulfide bond to create a cyclic six amino acid ring in both peptides. Seven out of nine amino acids are common and two amino acids (bold) are different between Avp and Oxt.

1.2. Pharmacology of the vasopressin receptors

There are three major receptor types for Avp: Avpr1a, Avpr1b, and Avpr2. The Avp receptors have seven transmembrane domains: Avpr1a and Avpr1b couple to G_αq/11 GTP binding proteins, which along with G_βλ, activate phospholipase C activity whereas Avpr2 couples to Gs and acts through the cyclic AMP system (Michell et al., 1979; Jard et al., 1987). Although not confined to these tissues in the periphery, the Avpr1a is prominent in the liver, kidney and vasculature. The Avpr1b is prominent in the anterior pituitary in cells that make adrenocorticotropin hormone (Antoni, 1984) but is detected in many other tissues (Lolait et al., 1995). Avpr2 is mainly expressed in the kidney and the antidiuretic function of Avp is mostly transduced via this receptor in the renal collecting duct (Bankir, 2001). The Avpr1a and Avpr1b are expressed throughout the brain (see below), but evidence for the presence of the Avpr2 there is scant (Foletta et al., 2002). A review of the receptor structures and ligand interactions is available (Thibonnier et al., 2002).

The search for specific agonists and antagonists for the Avp receptors has been an active area of research (Chan et al., 2000; see Table 1). In general, agonists for the receptors have been modifications of vasopressin whereas smaller organic compounds have been developed as antagonists (Serradeil-Le Gal et al., 2002), though few practical uses have been found. Desmopressin (1-desamino-8-d-arginine vasopressin, DDAVP), which acts at the Avpr2 in the renal collecting ducts, is used in patients with central diabetes insipidus to reduce urine production (and water loss) (Robinson, 1976). Vasopressin itself is used to increase blood pressure in patients with shock (Wenzel et al., 2004; Grmec and Mally, 2006). Avpr2 antagonists alleviate the effects of hypernatremia in patients with the syndrome of inappropriate secretion of antidiuretic hormone (Saito et al., 1997; Serradeil-Le Gal et al., 2002; Wong et al., 2003).

Table 1.

Pharmacological specificity of some Avp agonists and antagonists used in behavioral and physiological studies of rats.

Agent:	Avpr1a	Avpr1b	Avpr2	Oxtr
Avp	s, 1.5 (1)	s, 0.3 (1)	s, 0.5 (1)	s, 1.2 (1)
Oxt	s, 78 (2)	s, 250 (2)	s, 370 (2)	s, 1.5 (2)
[1-deamino, 8-D]Avp (DDAVP)	s, 100 (1)	s, 9.3 (3)	s, 0.3 (1)
[d(CH2)51,Tyr(Me)2]Avp (Manning compound)	i, 0.3 (2)	313 (4)	s, 65.4 (5)	i, 20 (6)
SRX251	i, 0.66 (7)
D[Leu4,Lys8]Vp	3800 (8)	s, 0.16 (8)	100 (8)	64 (8)
SSR149415	1050 (9)	i, 2.5 (9)	2900 (9)	270 (9)

Approximate nM values for K_D (Oxt values) or K_i at the specific rat receptor (s=agonist, i=antagonist where known) except for SRX251 (human)

References:

¹Pena et al., 2007a

²Manning and Sawyer, 1993

³Saito et al., 1997

⁴Antoni, 1984

⁵Howl et al., 1993

⁶Antoni and Chadio, 1989

⁷Guillon et al., 2007

⁸Pena et al., 2007b

⁹Serradeil-Le Gal et al., 2002b

The recent development of orally active antagonists shows potential in treating some mental disorders based on animal studies. In particular, the Avpr1b antagonist, SSR149415, produces reduced aggression and anxiety-like behaviors in rodents (Griebel et al., 2002; Blanchard et al., 2005; Stemmelin et al., 2005). The Avpr1a antagonist also reduces aggression in hamsters (Ferris et al., 2006). A full description of these effects can be found in sections 2.2.1 and 2.4.

1.3. Distribution (expression) of the vasopressin receptors in the brain

The distribution of Avpr1a expression within the central nervous system (CNS) has been primarily studied using receptor autoradiography and hybridization histochemistry. Receptor autoradiography identifies the locations of binding to the receptor protein, whereas hybridization histochemistry identifies the cells that transcribe the receptor gene. Prominent Avpr1a binding (Fig. 2) is present in the rat lateral septum (LS), neocortical layer IV, hippocampal formation, amygdalostriatal area, BNST, various hypothalamic areas (including SCN), ventral tegmental area, substantia nigra, superior colliculus, dorsal raphe, nucleus of the solitary tract and inferior olive (Johnson et al., 1993). Avpr1a binding is moderate throughout the spinal cord, but binding is higher in the dorsolateral motoneurons in general and all motoneurons in the lumbar 5/6 levels, where innervation to the perineal muscles originates (Tribollet et al., 1997).

Figure 2.

Distinct distributions of Avpr1a at two levels within the mouse brain. Receptor autoradiography was performed as described (Young et al., 2000) using an ¹²⁵I-labeled ligand. Abbreviations: CG, central grey; CP, caudate-putamen; Ctx, neocortex; H, hippocampal formation; LS, lateral septum; MS, medial septum; Sum, supramammillary nuclei; VP, ventral pallidum; VTA, ventral tegmental area.

Neurons containing Avpr1a transcripts are found extensively throughout the rat CNS, with prominent expression in the olfactory bulb, hippocampal formation, LS, SCN, PVN, anterior hypothalamic area, arcuate nucleus, lateral habenula, ventral tegmental area, substantia nigra (pars compacta), superior colliculus, raphe nuclei, locus coeruleus, inferior olive, area postrema and nucleus of the solitary tract (Ostrowski et al., 1994; Szot et al., 1994). Transcripts are also detected in the choroid plexus and endothelial cells. The distributions of Avp binding have been examined in a number of rodent species and they are remarkably similar. Differences in binding in selected areas may mediate important adaptations or behavioral traits (see section 2 on Behavior).

The Avpr1b was initially described in the anterior pituitary (Antoni, 1984). Avpr1b transcripts as well as Avpr1b immunoreactive cell bodies have been found in the rat brain, including the olfactory bulb, piriform cortical layer II, septum, cerebral cortex, hippocampus, PVN, SCN, cerebellum, and red nucleus (Lolait et al., 1995; Saito et al., 1995; Vaccari et al., 1998; Hernando et al., 2001; Stemmelin et al., 2005). It should be noted, however, that Avpr1b distribution has not been mapped by receptor autoradiography due to the lack of a specific radiolabeled ligand. A recent study using in situ hybridization histochemistry with specific probes found particularly prominent Avpr1b expression in the hippocampal field CA2 pyramidal neurons (Young et al., 2006).

1.4. Sex differences in the expression of vasopressin and its receptors

Behaviors often differ between the sexes, so there has been great interest in mapping sex-related differences in the Avp system. The pioneering studies of De Vries and colleagues (1983) have shown that the density of Avp fibers in the LS is elevated in males compared to females and that testosterone administered to females and castrated males increases the density of Avp fibers within the LS. Similarly, gonadectomy in males or females reduces Avp in certain regions of the brain that receive innervation from the BNST and MeA, but no reduction in regions that receive innervation from the magnocellular neurons of the PVN and SON (De Vries et al., 1984; De Vries et al., 1985). Avp mRNA levels in the BNST and MeA are also testosterone dependent (Miller et al., 1989; Miller et al., 1992; Wang and De Vries, 1995). Curiously, gonadectomy has no effect on Avp immunoreactivity in Syrian hamsters (Albers et al., 1991), but it does affect Avp immunoreactivity in Siberian hamsters (Dubois-Dauphin et al., 1994). Therefore, it is not clear how widely this gonadal steroid-dependency is conserved across mammals. Also, in chronically hyperosmolar rats there is evidence that gonadal steroids may modulate Avp expression (Crowley and Amico, 1993; O'Keefe et al., 1995). There is also evidence for an interaction between gonadal steroids and Avp, as androgen receptors are found in non-magnocellular responsive neurons (Zhou et al., 1994), and estrogen receptor beta is expressed in Avp magnocellular neurons (Alves et al., 1998). An estrogen responsive element is found over 4 kbp upstream of the transcriptional start site of the rat Avp gene (Shapiro et al., 2000), but whether this is where androgens and/or its metabolites exert their effects is unknown.

Avpr1a transcription and translation are sensitive to gonadal steroids in the brain but more research is warranted. In the photoperiodic Syrian and Siberian hamsters, exposure to short “winter-like” photoperiods results in dramatic reductions in Avpr1a binding within brain areas associated with the neural regulation of social behavior (Dubois-Dauphin et al., 1994; Caldwell and Albers, 2003; Caldwell and Albers, 2004a). Likewise, gonadectomy and lactation can decrease Avpr1a mRNA and receptor binding within the hypothalamus (Delville et al., 1995; Johnson et al., 1995; Young et al., 2000). In young rats, estrogen increases Avpr1a mRNA in the preoptic area of the hypothalamus (Funabashi et al., 2000). Castration has been reported to reduce binding in the pudendal nuclei of lumbar 5/6 levels (Tribollet et al., 1997). There is no evidence yet for any sex differences in the expression of the Avpr1b.

1.5. Gene inactivation in mice

The mouse shares many, if not most, aspects of anatomy, biochemistry and physiology with humans. This is not too surprising given that about 99% of mouse protein-coding genes have human counterparts (and vice versa) and justifies the research effort expended on generating and studying transgenic mice, including those produced through homologous recombination (“knockout”, KO, −/−) (Tecott, 2003).

A knockout of the Avp gene has not been published to our knowledge, although a natural occurring Avp mutant, the Brattleboro rat, has been extensively studied (Bohus and de Wied, 1998; Grant, 2000). The key observation in this line is diabetes insipidus that results in increased thirst-driven behavior. Studies of behavioral deficits in this line are discussed below (see Section 2).

The first of the Avp receptors to be inactivated was the Avpr2 (Yun et al., 2000). The Avpr2 gene is located on the X chromosome, so the male hemizygous and female Avpr2 KO pups cannot concentrate their urine. These pups also show enlarged renal pelvic spaces and die within the first week after birth due to hypernatremic dehydration. Female heterozygous mice display normal growth but have a moderate diabetes insipidus phenotype with polyuria, polydipsia and reduced ability to concentrate urine (Yun et al., 2000).

The initial report for Avpr1a KO mice indicated a function for the Avpr1a as a negative regulator of B cell receptor signaling, suggesting a role for Avp in stress-related immune responses (Hu et al., 2003). A different KO line of the Avpr1a demonstrates a lower basal blood pressure (Egashira et al., 2004; Koshimizu et al., 2006) and impaired lipid metabolism and glucose homeostasis (Aoyagi et al., 2007; Hiroyama et al., 2007). Both lines have changes in behavior as discussed below.

The initial descriptions of the Avpr1b KO reported reduced aggressive behavior and social motivation (Wersinger et al., 2002; Wersinger et al., 2004) as further discussed below. Tanoue and colleagues also generated Avpr1b −/− mice and showed attenuated catecholamine and hypothalamic-pituitary-adrenal (HPA) axis responses to stress (Tanoue et al., 2004) as well as reduced Avp responses in the pancreas for both insulin and glucagon (Oshikawa et al., 2004; Fujiwara et al., 2007). The HPA responses to the acute insulin-induced hypoglycemic and chronic restraint stresses are also impaired (Lolait et al., 2007).

2. Behavior

The roles of Avp in the neural regulation of various behaviors, especially social behaviors, are irrevocably intertwined. Therefore, for the purposes of this review we have separated behavior into three broad categories: social behaviors; learning and memory (social and non-social); stress; and anxiety and depression. While there will be some overlapping themes, a concerted effort has been made to draw out the similarity in the roles of Avp across these behaviors. Also, it should be noted that while Avp is often discussed in concert with Oxt, the primary focus of this paper is Avp, though it should be recognized that the Oxt literature is very complementary and there are some excellent recent reviews exploring the behavioral roles of both neuropeptides (Keverne and Curley, 2004; Caldwell and Young, 2006; Hammock and Young, 2006; Lim and Young, 2006).

2.1 Regulation of behavior by the suprachiasmatic nucleus

All mammals' behaviors are influenced by a central circadian clock, located within the SCN. The SCN, entrained through input via the retinohypothalamic pathway, controls circadian activity through feedback loops of a number of ‘clock genes’ (for review, see Albrecht and Eichele, 2003; Kalsbeek et al., 2006b).

Avp was the first neurotransmitter described in the SCN, and is one of the major neuropeptides produced there. It is the only neurotransmitter to date that has been shown to be released in a circadian rhythm in vivo. Avp displays circadian rhythmicity of release into the cerebrospinal fluid (CSF), with peak release during the day in a light-dark cycle (Reppert et al., 1983). Avp in the SCN aids in circadian control of hormone secretion. Avp-immunoreactive (Avp-ir) neurons have terminals in the PVN, which synthesizes and releases corticotrophin releasing hormone (Crh) and subsequent adrenocorticotropic hormone (ACTH) secretion from the pituitary. Avp from the SCN seems to exert a controlling effect of secretion of Crh and ACTH, with a primarily inhibitory role (for review, see Buijs et al., 2003; Perreau-Lenz et al., 2004; Kalsbeek et al., 2006a; Kalsbeek et al., 2006b). There is also a polysynaptic pathway from the SCN to the adrenal glands via the intermedio-lateral column of the spinal cord and sympathetic outflow (Buijs et al., 1999). Furthermore, Avp may aid in control of the hypothalamic-pituitary-gonad axis in females (see Buijs et al., 2003; Perreau-Lenz et al., 2004 for review).

Wheel-running activity in rodents is an indicator of circadian rhythmicity, with increased activity during the dark phase of the light-dark cycle in nocturnal species. While it is only correlational, diurnal fluctuations in Avp may exert some control over this behavior. In common voles, Microtus arvalis, fewer Avp-positive neurons are found in the dorsomedial SCN of highly rhythmic animals as compared to non-rhythmic animals kept in a light-dark cycle (Gerkema et al., 1994). Similar results were found in vitro, as SCN cultures prepared from voles with rhythmic wheel-running activity patterns contain significantly fewer Avp-ir cells than cultures from non-rhythmic voles (Jansen et al., 1999). Voles lacking rhythmic activity patterns do not produce a circadian pattern of Avp release (Jansen et al., 2000), indicating that Avp release is hampered in non-rhythmic voles (the Avp release deficit hypothesis), which may account for higher Avp-ir cells in non-rhythmic voles (Gerkema et al., 1994; Jansen et al., 2000). Furthermore, in the vole Avp contributes to circadian rhythmicity throughout the lifespan. Aging leads to decreased circadian activity patterns: 75% of old voles (11.5 months old), as compared to only 33% of young voles (4.5 months old) are non-rhythmic (Van der Zee et al., 1999). Avp is likely involved, as the number of Avp-ir SCN cells decreases over 78% from young to old voles (Van der Zee et al., 1999).

Avp influences on circadian rhythmicity are not limited to voles, but can be found in other rodent species. European ground squirrels, Spermophilus citellus, have absent or suppressed circadian rhythmicity when first emerging from hibernation; spontaneous regeneration of rhythmicity correlates positively with the number of Avp-ir neurons in the SCN (Hut et al., 2002). In two mice lines selected for nest-building behavior, differences in circadian rhythms are accompanied by different number of Avp-ir neurons in the SCN, as well as different transport and release of Avp in the SCN (Van der Veen et al., 2005). Reduced diurnal rhythms of Avpr1a mRNA in the SCN are seen in aged (19-20 months old), compared to young, rats (2-3 months old) (Kalamatianos et al., 2004). Evidence in golden hamsters, Mesocricetus auratus, indicates that lesions to the SCN reduce the number of Avp-ir fibers throughout the brain, but seem to affect only circadian locomotor behavior (wheel running), and not another Avp-dependent behavior, flank marking (see section 2.2.1) (Delville et al., 1998). Thus, evidence from a variety of rodent species implicates Avp in the SCN to be important for circadian rhythmicity of locomotor behavior.

However, despite the evidence above, the complete absence of Avp in Brattleboro rats (Groblewski et al., 1981) fails to alter circadian rhythms. Even in Avpr1a −/− mice there is only a modest effect on the circadian system. Specifically, a lengthened tau has been observed, which, while affecting the period, has no impact on rhythmicity per se (Wersinger et al., 2007b). Furthermore, the diurnal rhythm of mutant Avp mRNA and of the Avpr1a mRNA in the SCN persists in the absence of Avp in rats (Young et al., 1993). It is unclear whether these latter results signify developmental compensation, the relative unimportance of Avp in regulating circadian rhythms, or that arrhythmic animals may have altered Avp due to a malfunctioning SCN, and not the other way around.

The influence of SCN-derived Avp in human and non-human primate circadian rhythms has primarily been examined in relation to aging, dementia, and depression. Changes in circadian rhythms, particularly sleep irregularities, have been found in both healthy and ailing elderly persons (for review, see Myers and Badia, 1995; Van Someren, 2000). In people under 50 years old, Avp content within the SCN shows a clear circadian pattern; this pattern is reversed in those over 50 (Hofman and Swaab, 1994). The diurnal rhythm of Avp is reported to be absent in patients with hydrocephalus (Barreca et al., 1988). Patients with Alzheimer's disease have over 30% less Avp mRNA in the SCN than do healthy age-matched persons (Liu et al., 2000; Ishunina and Swaab, 2002). Decreased Avp mRNA is seen in Alzheimer's patients both with and without depression, although other work has found greater Avp-ir neurons in the SCN of elderly patients with depression as compared to age-matched controls (Zhou et al., 2001). It is possible that synthesis and release of Avp is reduced in depressed patients and those with Alzheimer's disease, which could contribute to the disturbances in circadian rhythmicity that are prevalent in both disorders.

2.1 Social Behavior

Working in concert, affiliation and aggression are used to maintain appropriate interactions within a social group. Aggression permits better access to resources whereas affiliative interactions allow for reproductive behavior, parental behavior, and even monogamy in some mammalian species. Across species, Avp is important in the regulation of both aggression and affiliation though, in general, predominantly in males. The next sections will delve into the role of Avp in the regulation of aggression and affiliation. Table 2 summarizes the effects of Avp on the social and non-social behaviors discussed below.

Table 2.

Summary of the behavioral effects of Avp

Behavioral Classes	Behaviors	Effects of Avp in rodents	Effects of Avp in humans
Social Behaviors
Aggression	Female aggression	Maternal aggression impaired in Avpr1b knockout mice	no known effect
	Male aggression	--⇑ when injected into the AH and LS --impaired in Avpr1b KO mice	--⇑ Avp in CSF correlated with more aggressive personality
	Scent marking and grooming	⇑ when injected into the MPOA-AH of hamsters	N/A
Affiliation	Partner preference	⇑ with i.c.v. injection in male prairie voles; dependent on Avpr1a distribution	N/A
	Paternal behavior	⇑ with intraseptal or hypothalamic rise in Avp	no known effect
	Maternal behavior	⇑ hypothalamic Avp release during lactation to aid in osmotic balance	no known effect
Learning and Memory
Social memory	Social recognition (males)	--⇑ when injected into LS --impaired in Avpr1b KO mice --impaired in Avpr1a KO mice; may be due to olfactory deficit	N/A
Non-social memory	Non-spatial memory	facilitated with Avp in dorsal or ventral hippocampus, depending on task	intranasal Avp enhances verbal memory in males, possibly by ⇑ arousal
	Spatial memory	--Avp enhances reference and working memory --impaired in Avpr1a KO mice	no known effect
Stress, Anxiety and Depression
	Anxiety	--Avp both anxiolytic and anxiogenic, depending on antagonist differences --decreased anxiety in Avpr1a mice in some studies; sexually dimorphic --correlation in AVP release from PVN and ⇑ anxiety in rats	high plasma Avp correlated with high anxiety levels
	Depression	--correlation in AVP release from SON and PVN and ⇑ depressive-like behavior	high plasma and hypothalamic Avp associated with major depressive disorder

2.2.1. Aggression

Aggression is critical to an animal's survival, allowing for better access to food and mates and, in general, conveying evolutionary fitness. Often classified into 5 categories: offensive aggression, defensive aggression, maternal aggression, play fighting, and predatory aggression, aggression is an integral part of most mammalian societies (Miczek et al., 2001; Blanchard et al., 2003). Studies of the role of Avp in the regulation of aggression have focused on offensive aggression, which in rodents is examined using a resident-intruder test. From this work two important concepts have emerged: 1) Avp primarily modulates intermale aggression and 2) Avp-facilitated aggression appears to be dependent upon prior experience, suggestive of synaptic learning and/or epigenetic regulation of gene expression. Also, while key anatomical sites have been identified, their interactions with one other, as well as the interactions of Avp with other neurotransmitter systems are still poorly understood. The vasopressinergic projections from the BNST and MeA to the caudal LS are important for aggression and are androgen-dependent (De Vries et al., 1984; De Vries et al., 1985; van Leeuwen et al., 1985; Caffe et al., 1987; Compaan et al., 1993; De Vries and Miller, 1998). The anterior hypothalamus (AH) is also consistently implicated in the regulation of aggression and Avpr1a in this area demonstrates considerable plasticity, which is discussed in section 2.1.1.2 (Ferris et al., 1989; Albers et al., 2006; Askew et al., 2006). Lastly, there is emerging evidence that the Avpr1b is critical for aggression though the neural substrate(s) involved have yet to be identified (Wersinger et al., 2002; Wersinger et al., 2007a) (see Fig. 3).

Figure 3.

Maternal aggression is much lower in Avpr1b KO mice compared to heterozygous (HET) and wildtype mice. *, p<0.05. Adapted from Wersinger et al. (2007a).

2.2.1.1. Pharmacological and behavioral manipulations

The most comprehensive work describing the role of Avp in the regulation of aggression has been carried out in Syrian hamsters. As they are a solitary species, Syrian hamsters are highly aggressive towards conspecifics. They also exhibit a stereotypic form of scent marking behavior, known as flank marking, that is displayed more prominently by socially dominant animals (Siegel, 1985). The initial, seminal finding that Avp microinjected into the medial preoptic-anterior hypothalamic area (MPOA-AH) results in a dose dependent increase in flank marking behavior (Ferris et al., 1984) set the framework for future studies in this species and was one of the first to induce a complex behavior by microinjection of a single neuropeptide. These effects of Avp on flank marking are attributed specifically to Avp's action via the Avpr1a since Avpr1a antagonists administered orally or injected into the MPOA-AH block Avp-facilitated flank marking behavior (Ferris et al., 1988; Ferris et al., 1993; Caldwell and Albers, 2003; Ferris et al., 2006). Similarly, aggression can be facilitated when Avp is microinjected into the AH (Ferris et al., 1997; Caldwell and Albers, 2004b) and inhibited when Avpr1a antagonists are microinjected there (Ferris and Potegal, 1988). Superimposed on this system is the influence of gonadal steroids. While Avp innervation of the MPOA-AH is not gonadal steroid dependent (Albers et al., 1991), Avpr1a biosynthesis is (Johnson et al., 1995; Young et al., 2000).

Other brain areas are also involved in the regulation of aggression. For example, Avpr1a binding within the ventrolateral hypothalamus is gonadal steroid-dependent, and this decrease is correlated with reduced aggression, specifically the initiation of aggression (Delville et al., 1996b). Recent work has implicated the Avpr1b in the regulation of aggression in Syrian hamsters. Hamsters administered an oral Avpr1b antagonist display less aggression than untreated controls (Blanchard et al., 2005), though some of the more compelling evidence for the role of the Avpr1b in the regulation of behavior is described below in section 2.2.1.3 (Transgenics).

One of the lingering questions in this field is the interplay between flank marking and aggression. While they are complementary behaviors that help an animal attain and maintain dominance, and appear to be regulated by some of the same neuroanatomical areas, they are not necessarily behaviors that are a part of a behavioral continuum. For instance, the specific site where Avp stimulates aggression is different than the one that stimulates flank marking. Also, the doses of Avp used to induce aggression and flank marking differ (Albers et al., 1988; Ferris et al., 1997; Caldwell and Albers, 2003; Caldwell and Albers, 2004b). This does not, however, preclude the possibility that endogenous Avp release in the same brain areas is responsible for the initiation of both behaviors. However, it is likely that the neural regulation of aggression is more complex and the circuits more redundant than that of flank marking.

Comparable data have been found in other species, but some of the neuroanatomical areas differ from those described in hamsters. In prairie voles there is evidence that Avp may be important to increases in aggression seen following pair-bond formation. Specifically, compared to controls, pair-bonded males that show high levels of aggression towards unfamiliar conspecifics have increased c-fos activation within the AH; specifically, within cells that contain Avp and tyrosine hydroxylase. The authors suggest that Avp within the AH may be important for processing chemosensory cues or responding to them (Gobrogge et al., 2007).

There is some elegant work examining differences between mice that have been bred for either a long attack latency or short attack latency. Studies of these mice have provided valuable insights into critical differences between highly aggressive individuals and their less aggressive counterparts. For instance, short attack latency mice have less Avp-ir innervation in the LS and fewer Avp-ir neurons in the BNST than long attack latency mice (Compaan et al., 1993). The findings are similar in rats in which there is a negative correlation between Avp and aggression in the LS (Everts et al., 1997). The opposite is found in Peromyscus californicus (California mouse) which are more aggressive than Peromyscus leucopus (white-footed mouse) and have a positive correlation between aggression and Avp-ir in the BNST and LS (Bester-Meredith and Marler, 2001). In prairie voles and California mice, intracerebroventricular (i.c.v.) injections of Avpr1a antagonists inhibit aggression and in prairie voles injections of Avp increase aggression (Winslow et al., 1993; Bester-Meredith et al., 2005). However, the authors suggest that Avp antagonists do not block the expression of aggression per se, as breeder males with established aggression are unaffected; but, rather, affect the transition to aggression, i.e. initiation of aggression. The idea that the effects of Avp on aggression are dependent on an animal's history is supported by several studies and discussed in more detail below.

2.2.1.2. Experience-dependent changes

The ability to display territorial aggression is susceptible to manipulation depending on an individual animal's early life experience and/or social status. This plasticity in aggression is linked with changes in either Avp or Avp receptor distributions in a variety of mammalian species (e.g. Roche and Leshner, 1979; Ferris et al., 1988; Melloni et al., 1997; Stribley and Carter, 1999; Frazier et al., 2006). The next sections will detail the experience-dependent changes in Avp and its receptors, and how they relate to aggression.

2.2.1.2.1. Early life experience

While sexually naïve prairie voles generally do not show territorial aggression, early postnatal exposure to Avp can increase displays of aggression similar to levels seen in post-mated animals (Stribley and Carter, 1999). Newborn California mice that are retrieved more by their fathers show higher levels of aggression in adulthood than mice that are retrieved less. These animals also have increased Avp-ir fibers in the dorsal fiber tracts of the BNST (Frazier et al., 2006). Adult male rats that undergo maternal separation as pups have higher aggression, more Avp mRNA and Avp immunoreactivity in the PVN and SON, perhaps due to poor stress-coping as these animals also show more depression-like activity during, and an increased ACTH response following, forced swim tests (Veenema et al., 2006).

Some interesting work in hamsters has focused on the impact of exposure to anabolicandrogenic steroids during the androgen sensitive time of adolescence on adult aggressive behavior. Not only does exposure to anabolic-androgenic steroids during adolescence increase adult aggression, but it also increases Avpr1a-ir within the AH, though the effects appear to be reversible following cessation of treatment (Grimes et al., 2006). In anabolic-androgenic steroid-treated animals that have an Avpr1a antagonist microinjected into the AH, there is a reduction in the intensity, but not the initiation, of aggression (Melloni et al., 1997; Harrison et al., 2000).

2.2.1.2.2. Social status

Initial studies suggested that Avp might have differential effects on aggression in dominant versus subordinate animals. Mice administered lysine-vasopressin just after a social defeat, increase their submissive behavior on subsequent tests compared to saline treated animals (Roche and Leshner, 1979). Further, injections into the MPOA-AH results in transient reversals of dominate/subordinate relationships in Syrian hamsters, with subordinate animals displaying increased flank marking behavior when treated with Avp and dominant animals less flank marking behavior when treated with dPTyr(Me)AVP, an Avpr1 antagonist (Ferris et al., 1986). Consistent with these findings, subordinate hamsters have fewer Avp-ir perikarya in the magnocellular nucleus circularis compared to dominant animals (Ferris et al., 1989). Following repeated agonistic encounters, dominant hamsters have more Avpr1a binding in the ventromedial hypothalamus compared to their subordinate opponents (Cooper et al., 2005). In socially isolated hamsters, there is an increase in aggression that correlates with increased Avpr1a binding in the AH, PVN and lateral hypothalamus, whereas in socially experienced hamsters Avpr1a binding is significantly greater in the central amygdala (Albers et al., 2006). Even in Avp-facilitated aggression, several weeks of social isolation is required for Avp to be effective, suggesting that social experience, or lack there of, is critical for the effects of Avp on aggression (Ferris et al., 1997; Caldwell and Albers, 2004b). In rats, dominant animals (as determined by food competition) have more Avpr1a binding in the LS compared to subordinates. These differences in receptor binding are not testosterone- or corticosterone-dependent, suggestive of experience-dependent changes in the Avp system (Askew et al., 2006). Taken together, there is substantial evidence that the Avp system, and specifically Avpr1a biosynthesis, can be altered by social experience.

2.2.1.3. Transgenics

The generation of mice with inactivation of the either the Avpr1a or Avpr1b genes has provided insights into their roles in the regulation of aggression. While it was expected that Avpr1a KO mice would show reduced aggression due to lack of Avp signaling, this was not the case (Wersinger et al., 2007b). This is likely due to developmental compensation and is not thought to reflect the true role of Avpr1a in the modulation of aggression. Conversely, Avpr1b appears to be critical for proper expression of aggression, as Avpr1b −/− mice show significant impairments in displays of aggression as compared to wildtype controls (Wersinger et al., 2002; Wersinger et al., 2004; Wersinger et al., 2007a). In addition, an orally administered Avpr1b antagonist reduces aggression in mice and hamsters (Blanchard et al., 2005; for review see Serradeil-Le Gal et al., 2005). Avpr1b −/− mice have longer attack latencies and fewer attacks compared to wildtype controls, though the latencies and frequencies of attack can be increased with experience (Wersinger et al., 2002; Wersinger et al., 2007a). These impairments are specific to social forms of aggression, as predatory aggression remains unaffected. Furthermore, only the attack component is affected, as the Avpr1b KO display defensive postures but do not initiate defensive or “retaliatory” attacks (Wersinger et al., 2007a). The actions of Avp via Avpr1b are not sexually dimorphic since female Avpr1b −/− mice show reduced maternal aggression (Wersinger et al., 2007a). Avpr1b may not be involved in the transduction of the olfactory signals, but rather in coupling the social context with the appropriate behavioral response (Wersinger et al., 2004). Given the restricted distribution of Avpr1b mRNA within the brain, with highest expression within the CA2 field of the hippocampus, it may be that the Avpr1b is important for “social memory” for which an aggressive response would be appropriately triggered (Young et al., 2006). Future studies that focus on the specific neuroanatomical areas involved in Avpr1b's mediation of aggression will be important for understanding the role of Avp in the regulation of aggression.

2.2.1.4. Vasopressin and serotonin

Whereas Avp facilitates aggression and serotonin (5-HT) inhibits aggression, their exact relationship remains elusive. Within the ventrolateral hypothalamus, 5-HT blocks Avp-facilitated offensive aggression (Delville et al., 1996a). In the AH, a Htr1a, but not a Htr1b, serotonin receptor agonist produces a dose-dependent inhibition of Avp facilitated offensive aggression (Ferris et al., 1999). The 5-HT innervation to this area is traced to the dorsal, median, and caudal linear raphe nuclei (Ferris et al., 1999). Likewise, treatment with selective serotonin re-uptake inhibitors suppresses Avp-facilitated flank marking behavior (Ferris et al., 2001). In rats that have undergone maternal separation, and who subsequently show higher levels of aggression, 5-HT-ir in the AH and SON negatively correlates with aggression (Veenema et al., 2006). With their opposing effects on aggression, Avp and 5-HT are biologically interconnected, and studies focusing on their interactions are critical to increasing our understanding of the neural regulation of aggression.

2.2.1.5. Humans

In humans, heightened aggression is associated with increased impulsivity and is a characteristic of individuals with personality disorders. There is a positive correlation of CSF Avp concentration with a life history of non-directed general aggression (including temper tantrums and property assault) as well as aggression towards individuals (Coccaro et al., 1998). Similar to rodent models, this finding is suggestive of an enhancing effect of Avp on aggression, although it should be noted that not all studies are in agreement. A study by Virkkunen and colleagues (1994) found no differences in CSF Avp between violent offenders and controls. These discrepancies, however, may be due to differences in the patient populations being studied.

Intranasal Avp administration, at least in men, is reported to increase the emotional response to neutral stimuli resulting in an increased perception of threat (Thompson et al., 2004; Thompson et al., 2006). Although Born et al. (2002) have shown that intranasal application of Avp increases concentration of Avp in the CSF within 30 minutes of application, and remains elevated for at least 80 minutes, there is no evidence to distinguish between direct entry of Avp versus CNS release of Avp by actions in the periphery. There is also preliminary work examining the differences in levels of autoantibodies for Avp between people with conduct disorder, prisoners, and healthy controls. Individuals with conduct disorder have reduced levels of Avp-reactive antibodies (Fetissov et al., 2006). These authors hypothesize that the autoantibodies could either regulate peptides crossing the blood-brain barrier or centrally interfere with peptides, allowing for elevated Avp and a resulting increase in aggression. Alternatively, greater Avp could influence the HPA axis (through the anterior pituitary's Avpr1b), as is hypothesized for ACTH and autoantibodies (though the adrenal glands) (Fetissov et al., 2006). The effects of peptide autoantibodies on behavior are just now beginning to be explored but provide an interesting and a previously unexplored mechanism by which to regulate behavior.

2.2.1.6. Conclusion

Given the difficulty of studying and the complexity of aggression in humans, the concordance between the findings of the animal work and that of the human literature implicating Avp in the regulation of aggression is compelling. It appears that the role of Avp in the regulation of aggression is highly conserved. It will be interesting to see what insights are gained through the use of more focused and refined genetic tools to identify specific interactions between neuroanatomical areas and different neurotransmitter systems in the regulation of aggression.

2.2.2. Affiliation

Affiliation encompasses social bonding between individuals, including sexual partner relationships and parent-infant relationships. The formation of social bonds helps increase security thereby reducing stress and anxiety. The ways in which social bonds are formed are species-specific and show high individual variability (for review see Carter, 1998). In the sections below we will explore the contributions of Avp to the regulation of pair bonding and parental behavior.

2.2.2.1. Pair bonding

Found in fewer than 5% of mammalian species, monogamy is defined as an exclusive, long-lasting sexual relationship between partners (Kleiman, 1977). Monogamy can also be defined using a group of distinguishing characteristic behaviors described by Carter and colleagues (1995) that include: 1) cohabitation and long-term pair bonding in the breeding and non-breeding season, 2) aggressive displays towards unfamiliar conspecifics, 3) biparental care, 4) socially regulated reproduction and 5) incest avoidance. The formation of pair bonds has been studied almost exclusively in voles, as the genus Microtus contains both monogamous (prairie and pine voles) and non-monogamous (montane and meadow voles) species. These voles have been vital in furthering our understanding of social bonding since the biology and behavior of these animals is so diverse, yet can be readily compared.

In the laboratory, pair bonding is measured using a partner preference test. In this test a male and female vole are mated for 24 hours and then separated. The male is then placed into the center of a three chamber apparatus and the amount of time he spends in adjacent chambers, one of which houses the mated (partner) female and the other which houses a novel “stranger” female, is measured (Carter and Getz, 1993; Williams et al., 1994; Carter et al., 1995). A partner preference is said to have occurred when the male spends twice as much time with his partner than the novel “stranger” female.

In voles, distribution of the Avpr1a may underlie differences in social organization as reflected by performance on the partner preference test. Prairie voles have less Avp immunoreactivity and binding in the LS and more in the amygdala compared to montane voles (Insel et al., 1993; Insel et al., 1994). While this is only correlative, more substantive evidence is found in prairie voles administered Avp and Avpr1a antagonists. When administered i.c.v. prior to mating, Avp antagonists inhibit partner preference whereas Avp infusions facilitate partner preference (Winslow et al., 1993; Cho et al., 1999).

Young and colleagues (Young et al., 1999; Lim et al., 2004b) have suggested that the regulatory region 5′ to the promoter of the Avpr1a gene is responsible for the differences in its distribution. Transgenic mice that have 2.2 kbp of the 5′ flanking region, both exons with the 2.5-kbp intron, and 2.4 kbp of 3′ flanking region of a prairie vole Avpr1a mini-gene have patterns of Avpr1a binding similar to that of prairie voles and show increased affiliative behavior, as measured by the amount of time they investigate and groom an ovariectomized female (Young et al., 1999). Interestingly, non-monogamous montane voles do not show increased affiliative behaviors after Avp is centrally injected, suggesting that the species-specific distribution of Avpr1a is important for the development of affiliative behaviors within a given species (Young et al., 1999). Increased affiliative behavior, as measured by huddling, is also increased through additional Avpr1a expression delivered to the ventral forebrain via a viral vector (Lim et al., 2004b). The vole Avpr1a genes were examined and two differences found: 1) the prairie vole genome contains two copies of the Avpr1a gene and 2) the 5′ regulatory region of Avpr1a of both prairie and pine voles contain a microsatellite sequence comprised of simple sequence repeats interspersed with non-repetitive elements (Young et al., 1999). Further studies showed that, in vitro, the microsatellite differences produce differences in reporter gene expression (Hammock and Young, 2004). Also, wild-caught prairie voles show a high level of variability in Avpr1a density in certain brain regions such as the olfactory bulb, cingulate cortex, amygdala, and ventral pallidum (Phelps and Young, 2003). To test the possibility that these individual differences were due to differences in microsatellite length within the same species Hammock and Young (2005) created two prairie vole breeding lines, one that had a long microsatellite sequence and one that had a shorter microsatellite sequence (746 bp vs 727 bp, respectively). There were regional differences in Avpr1a density between these animals and the F1 generation of the line that contained the lengthened microsatellite sequence tended to show a stronger partner preference than the line with the shortened microsatellite sequence. The implications are that within a given species the microsatellite satellite sequence, which is highly susceptible to mutation, may account for variations in Avpr1a density in specific brain regions.

Recently, Fink and colleagues (2006) showed that the presence of the microsatellite in the Avpr1a gene of 21 Microtus species was not predictive of monogamy. However, these findings do not preclude the possibility that variations in Avpr1a expression patterns in vole species account for the monogamous or polygamous behaviors seen (Young and Hammock, 2007). Whether these variations are due to the small differences in microsatellite lengths or some other differences in another part of the Avpr1a gene (or other genes) also awaits further study. For a more in depth review on pair bonding and its genetic regulation, see (Insel, 2003; Aragona and Wang, 2004; Young and Wang, 2004; Lim et al., 2004a; Young et al., 2005; Hammock and Young, 2006; Lim and Young, 2006; Nair and Young, 2006).

2.2.2.2. Parental behavior

2.2.2.2.1. Paternal care

Most research on paternal behavior has focused on monogamous rodent species, particularly the prairie vole and California mouse. Male prairie voles express greater Avp-ir fibers and neurons in the BNST, LS, and Lhb as compared to females (Wang et al., 1998). This pathway is steroid hormone dependent, as castration reduces Avp expression in these brain regions (Wang and De Vries, 1993; Lonstein et al., 2005), as well as displays of paternal behavior (Wang and De Vries, 1993). In a similar study, however, castration did not reduce the number of Avp-ir fibers in the LS or lateral habenula, and yet increased displays of paternal behavior in virgins (Lonstein and De Vries, 1999). Nonetheless, Avp-ir fibers increase naturally in prairie vole males after cohabitation with a female and just prior to parturition, suggesting these fibers are involved in both sexual and paternal behavior (Bamshad et al., 1994). Indeed, Avp directly affects paternal behavior in the vole as intra-septal Avp increases and Avpr1a antagonist decreases paternal behavior in sexually naïve male prairie voles (Wang et al., 1994). Similarly, when combined, Avp and Oxt antagonists significantly reduce alloparental behavior (care of other's offspring) in sexually naïve males, who then attack pups more (Bales et al., 2004).

Non-monogamous male montane and meadow voles do not naturally demonstrate paternal care. Unlike the prairie vole, in which hypothalamic Avp mRNA increases in males and females in the postpartum period, no change in Avp is observed in either male or female montane voles (Wang et al., 2000). As noted above, Avpr1a distribution differs between the monogamous and non-monogamous vole species, suggesting that these receptors may also regulate paternal behavior (Wang et al., 1998). However, elevating Avp induces paternal behavior in meadow voles by inhibiting pup-directed aggression and increasing paternal behavior in previously aggressive meadow voles (Parker and Lee, 2001). Cohabitation may increase paternal responsiveness in meadow voles, as cohabitating males given an Avp receptor antagonist display significantly less paternal behavior than cohabitating male control voles (Parker and Lee, 2001).

Monogamous male California mice spend more time providing paternal care to their offspring than do polygamous white-footed male mice (Bester-Meredith et al., 1999). Similar to prairie voles, male California mice have higher Avp-ir staining in the BNST, and more Avpr1a in the LS (Bester-Meredith et al., 1999). California mice are also more aggressive towards intruders in the presence of pups than white-footed mice, which may be similar to maternal aggression in regards to pup care (Bester-Meredith et al., 1999; Bester-Meredith and Marler, 2001). Cross-fostering studies show that offspring aggression, paternal behavior, and Avp-ir staining in the BNST and SON is similar to that of their foster, not biological, parents (Bester-Meredith and Marler, 2001; 2003). Similarly, California mice that are retrieved more by their father have decreased attack latency and greater Avp-ir fibers in the BNST (Frazier et al., 2006). Thus, parenting style alters offspring aggression and display of paternal behavior. These effects appear to be modulated by Avp in the BNST and hypothalamus.

In non-rodent mammalian species, paternal care is found in relatively few species, most notably humans and two species of New World primates (for review see Ziegler, 2000). No studies have yet examined the role of Avp in non-human primate paternal care; however, sexually dimorphic differences in numbers and density of Avp-ir fibers in the BNST of marmosets, similar to those seen in rodents, could influence sexual and paternal behavior in non-human primates (Wang et al., 1997).

2.2.2.2.2. Maternal Care

There is some evidence that Avp aids in maternal care of offspring (see Table 2). In rats, Avp levels and/or release increase during late pregnancy, parturition and lactation in various brain regions, including the SON (Caldwell et al., 1987; Mezey and Kiss, 1991), PVN (Caldwell et al., 1987), septum (Landgraf et al., 1991; Landgraf et al., 1992), and hippocampus (Landgraf et al., 1991). Similar changes have been found in the rabbit, where the size and number of Avpir neurons significantly increase in hypothalamic nuclei from late pregnancy through parturition (Caba et al., 1996). Likely, Avp release may help females to regulate water loss and elevated osmolality that result from pup anogenital licking (and subsequent intake of pup urine) during lactation (Mezey and Kiss, 1991; Caba et al., 1996; Wang et al., 2000). Indeed, female montane voles, who engage in less anogenital licking than female prairie voles, do not display elevated Avp mRNA during lactation as do female prairie voles (Wang et al., 2000). In rats, central Avp release predominates over Oxt release during late pregnancy and parturition, but once suckling begins the pattern of release is reversed (Landgraf et al., 1992).

2.2.2.3. Conclusion

Avp is strongly implicated in affiliative behaviors in males of several species, although its role in female affiliative behavior remains less clear. There has been some recent work exploring the contributions of the dopaminergic reward system to pair-bond formation in prairie voles and its possible interactions with the Avp (and OT) systems (Aragona et al., 2003; Aragona et al., 2006). Understanding the interactions of various neurotransmitter and neuropeptide systems during different “behavior states” continues to be an exciting area of research and will likely provide insight into the regulation of context-appropriate behavioral displays.

2.3. Learning and memory

2.3.1. Social memory and social recognition

The ability to recognize individuals is important for normal affiliation and aggression, as without the capacity to distinguish family or friend from foe it is difficult for an animal to display the appropriate behavior. This social memory depends largely on visual and auditory cues in primates and on olfactory, including pheromonal, cues in rodents. In rodents, it is the processing of olfactory cues and the subsequent olfactory memories that are Avp-dependent. The form of social memory examined in rodents is referred to as social recognition.

Social recognition tests are a specific group of learning and memory tests used to examine the social memory of individual conspecifics. Animals are tested for their natural inclination to investigate a novel animal more than a familiar one. Two different types of social recognition tests are generally employed. The first is a “habituation-dishabituation” test in which a subject animal is exposed to the same “stimulus” animal for repeated trials, resulting in a decrease in investigation, i.e. “habituation”. The subject is then presented a novel animal that normally results in an increase in investigation, i.e., “dishabituation.” (Winslow and Camacho, 1995). The second is a social discrimination test whereby a subject animal is exposed to a “stimulus” animal and then after some predetermined period of time, such as 30 minutes, a second trial is performed in which an animal is re-exposed to either the same familiar animal or a novel animal. Again, animals normally spend more time investigating the novel animal as opposed to the familiar animal in the second trial (Dantzer et al., 1987).

As noted above, there is compelling evidence that Avp is important for social memory formation and, thus, social recognition. In particular, the androgen-dependent vasopressinergic projections from the MeA and BNST to the LS are important for individual recognition (De Vries et al., 1984; Mayes et al., 1988; Bluthe et al., 1990; Bluthe et al., 1993). Microinjections of Avp into the LS of normal or the Avp-deficient Brattleboro rat specifically facilitate social memory whereas Avpr1a antagonists or infusions of antisense Avpr1a oligonucleotides in normal rats impair it (Engelmann and Landgraf, 1994; Landgraf et al., 1995). Also, over-expression of vole Avpr1a within the LS of rats results in a prolonged ability to socially discriminate, 2 hours in experimental animals compared to 1 hour in controls (Landgraf et al., 2003).

The use of Avpr1a and Avpr1b KO mice has further elucidated the role of Avp in the formation of social memory. In the NIMH line of knockout mice the effects of Avpr1a (Hu et al., 2003) on social recognition have been mixed. Bielsky and colleagues (2004; 2005a) demonstrated that male Avpr1a −/− mice have impaired social recognition that can be rescued by over-expression of Avpr1a in the LS. On the other hand, Wersinger and colleagues (2007b) found no deficits in social recognition but rather a deficit in olfaction in Avpr1a −/−mice. Whether this discrepancy is due to slightly different strain backgrounds or testing procedures is unknown, but it is clear from previous work that Avp within the LS is important for social recognition. Interestingly, Avpr1b mice also show mild impairments in social recognition (Wersinger et al., 2002).

In female Avpr1a and Avpr1b KO mice, results consistent with those found by Wersinger and colleagues (2007a,b) have been found using a test of mate recognition. The Bruce effect is a pheromonally mediated response to chemosensory cues that results in a pregnancy termination when a recently mated female interacts with an unfamiliar male 24 hours after removal of the original mate (Bruce, 1959). As shown in Fig. 4, Avpr1a −/− females show a normal Bruce effect while Avpr1b −/− females fail to terminate their pregnancies in the presence of the unfamiliar male (Wersinger et al., 2007c). It is of interest to note that this represents one type of failure of the Bruce effect – Oxt KO mice display a second type in which the original mate is perceived as an unfamiliar male after 24 hours of separation (Fig. 4).

Figure 4.

Pregnancy block (Bruce effect) is normal in Avpr1a KO mice but abnormal in different ways in Avpr1b and Oxt KO mice. As expected from the severe deficit in social recognition, Oxt KO females cannot remember their mates (A) and hence do not remain pregnant. In contrast, Avpr1b KO mice do not distinguish between their mates (A) and strange males (B) and “accept” either (B). Adapted from Wersinger et al. (2007c).

These findings, along with the observations in males, suggest that the Avpr1a may be more involved in the integration of main olfactory information, whereas the Avpr1b is involved in the formation and/or retrieval of memories associated with specific accessory olfactory cues and social encounters in order to generate appropriate behavioral responses (Wersinger et al., 2004). Given the abundance of Avpr1b in the hippocampal pyramidal cells of the CA2 region, the idea of a “social” cell in the hippocampus that would associate context (i.e., accessory olfactory cues) with the social experience, analogous to how CA1 pyramidal place cells associate visual input with direction and position in space (O'Keefe and Dostrovsky, 1971), is plausible (Young et al., 2006). Future studies focusing on the contribution of Avpr1b to the social recognition will provide a more complete picture of the role of Avp in the formation of social memories. Fig. 5 illustrates the proposed roles for the Avpr1a and Avpr1b in the regulation of social behaviors.

Figure 5.

Hypothesized roles of the vasopressin 1a receptor (Avpr1a) and the vasopressin 1b receptor (Avpr1b) in the regulation of social behavior. Absence of the Avpr1a results in impaired olfaction, suggestive that the Avpr1a may be important to responses to chemosensory input (Wersinger et al., 2007b). On the other hand, absence of the Avpr1b results in reduced aggression and impaired social memory in the absence of olfactory deficits (Wersinger et al., 2004; Wersinger et al., 2007a). As the Avpr1b is found predominantly within the CA2 pyramidal neurons of the hippocampus, it has been proposed to be involved in the evaluation of accessory olfactory chemosensory cues (i.e., associating the cue with the behavioral context) (Young et al., 2006).

2.3.2. Non-social memory

In addition to its effects on social memory and interactions, the pioneering work of David de Wied in the 1960s firmly established a modulatory effect of hypothalamic Avp on non-social learning and memory (see de Wied et al., 1993 for review). Beginning in 1965, de Wied found that removal of the posterior pituitary impaired active avoidance shuttlebox performance in rats (de Wied, 1965) that could be improved with Avp administration (Bohus et al., 1972). Since then, Avp has repeatedly been shown to affect non-spatial and spatial memory, with most studies finding an enhancing effect on memory, most likely through influences on the hippocampalseptal interactions (Croiset et al., 2000; see Table 2). The following section will describe a number of non-spatial and spatial tasks used to assess memory in rodents (for a full description of the tasks see Crawley, 2000)

2.3.2.1. Non-spatial learning and memory

Avp can enhance consolidation of memories, but seems more important in memory retrieval (Alescio-Lautier et al., 2000; Gulpinar and Yegen, 2004). Passive avoidance is often used to test Avp effects on non-spatial memory. In passive avoidance, animals receive a foot shock after entering a dark compartment, which they normally prefer over light spaces. At various times after this learning trial, retention of learning is tested by latency to enter the dark compartment (Laczi et al., 1983). Administration of Avp and two fragments (Avp_4-8 and Avp_5-9) immediately after the last learning trial or just prior to the retention trial both improve passive avoidance learning (Kovacs et al., 1986). Recently, subcutaneous administration of NC-1900, an Avp_4-9 analog with a longer half-life, was shown to ameliorate deficits in passive avoidance memory produced by inhibition of the COX-2 pathway (Sato et al., 2007), indicating that Avp may affect memory through this pathway. Moreover, administration of Avp fragments can affect memory up to 72 hours after training (Gaffori and De Wied, 1986), and even months after a single training trial (Izquierdo et al., 2002). Chronic treatment, as compared to acute treatment, with the Avp analog DDAVP is effective in aiding learning and retention of avoidance behavior (Hamburger-Bar et al., 1985). Endogenously, ‘good’ passive avoidance learners have significant decreases in Avp in the hippocampus, but not the amygdala, consistent with increased secretory activity in the hippocampus (Laczi et al., 1983). Conversely, administration of Avp antiserum into dorsal hippocampus and dentate gyrus significantly impair passive avoidance behavior (Kovacs et al., 1982). Thus, Avp activity in the hippocampus seems to improve memory for an aversive shock.

A second task often used is the “Go-No Go” visual discrimination task in which animals must learn to make a visual discrimination between two alleys, one of which always contains food reinforcement (Go trials) and the other never does (No Go trials). Learning is said to occur when running time decreases on the Go trial, and increases on the No Go trial (as if searching for food). Each session consists of twelve alternating trials; the first of each trial (Go and No Go) reflects retrieval of the learned memory, while the subsequent trials reflect relearning (Alescio-Lautier et al., 1995). On the Go-No Go task, intra-hippocampal Avp significantly improves retrieval and subsequent relearning when injected just prior to the retention trial (Alescio-Lautier and Soumireu-Mourat, 1998). Performance is better with injection into the ventral, as compared to the dorsal, hippocampus and only lesions to the ventral hippocampus blunt the effects of i.c.v. Avp (Alescio-Lautier and Soumireu-Mourat, 1998; Alescio-Lautier et al., 2000). Similarly, injection of Avp antiserum into the ventral, but not the dorsal, hippocampus prevents relearning of the Go-No Go task (Alescio-Lautier and Soumireu-Mourat, 1998). Finally, blockade of Avpr1a receptors impairs retrieval and relearning, even with subsequent Avp application (Alescio-Lautier et al., 1995).

2.3.2.2. Spatial learning

Spatial memory is memory of the spatial aspects of the surrounding environment. One of the oldest tests used for spatial memory is the radial arm maze in which food pellets are placed at the end of eight radiating arms. Optimal performance requires the animal visit each arm only once, and rodents quickly learn to do so by selecting unvisited arms (working memory) on the basis of their spatial location relative to extra-maze cues (in Packard and Ettenberg, 1985). Working memory in the radial arm maze is enhanced by Avp, as administration significantly enhances learned extinction behavior, i.e. learning that food was no longer present in the arms (Packard and Ettenberg, 1985). Avp_4-9 also enhances working and reference memory on the radial arm maze (Dietrich and Allen, 1997a). Scopolamine, an anticholinergic, produces deficits in reference and working memory in the radial arm maze. These memory deficits are reversed with administration of Avp (Taga et al., 2001), Avp_4-9 (Fujiwara et al., 1997; Mishima et al., 2001), and NC-1900 (Mishima et al., 2003).

Avp or its active fragments aid performance on other spatial memory tasks as well. In the hole board search task, animals receive a food reward in the same 4 (of 16) holes through multiple trials per day over many days. Reference memory consists of only going to those holes in which food has been found during the learning trials. Subcutaneous injection of Avp_4-9 significantly improves reference memory and reduces the number of errors on this task (Vawter et al., 1997). The Hebb-Williams mazes consist of a battery of 12 mazes, and with moveable barriers the mazes can be made easy, moderate, or difficult (see Paban et al., 2003). Subcutaneous (s.c.), i.p., and dorsal hippocampal administration of Avp all significantly increase consolidation of a Hebb-Williams maze, but do not affect transfer of learning, i.e. learning times on a similarly-configured maze (Paban et al., 2003). Combined, the above studies indicate a general beneficial effect of Avp administration on spatial memory. Endogenous Avp is likely necessary for spatial memory performance, as animals with partial vasopressin insufficiency (Brattleboro heterozygotes: +/di) demonstrate poorer working memory than wildtype controls on a delayed-non-match-to-sample task in which the animal has to respond to an illuminated aperture different from the one on which it had been trained (Aarde and Jentsch, 2006).

Not all studies agree on the effect within the brain of Avp or its fragments on spatial memory. Early work found intra-septal administration of Avp significantly impairs learning of the Morris water maze (Engelmann et al., 1992). Injection of an Avpr1a antagonist suppressed the NC-1900 improvement on scopolamine-induced spatial memory deficits (Mishima et al., 2003), contrary to prior studies in which an Avpr1a antagonist did not impair performance on the radial arm maze (Dietrich and Allen, 1997a) or the Morris water maze (Everts and Koolhaas, 1999), in which an animal is trained to swim to a hidden escape platform using extra-maze spatial cues. However, Avpr1a KO mice have impaired working memory in the radial arm maze as compared to controls (Egashira et al., 2004; Egashira et al., 2006), indicating a role for the Avpr1a receptor in spatial task performance. Conversely, Avpr1b KO mice do not have impaired spatial memory on the Morris water maze (Wersinger et al., 2002; Egashira et al., 2004). Contrary to results of Paban et al (2003), i.c.v. injection of the active fragment Avp_4-9 into the ventral, but not dorsal, hippocampus reverses scopolamine-induced spatial memory deficits (Fujiwara et al., 1997). Finally, Avp activity within the hippocampus seems necessary only for certain aspects of spatial memory, as rats with hippocampal lesions injected with Avp_4-9 demonstrate deficits in working memory, but not reference memory, in the radial arm maze indicating effects outside of the hippocampus (Dietrich and Allen, 1997b). More study is needed to identify the receptor and its location in the brain upon which Avp acts to affect spatial memory.

2.3.2.3. Clinical studies

Some research indicates that Avp enhances human memory (Born et al., 1998). Early work has indicated a possible sexually-dimorphic effect of Avp, as an Avp analog improves memory for sentences (Beckwith et al., 1984; Beckwith et al., 1984) and enhances word recall (Pietrowsky et al., 1988) in males, but not females. Similarly, acute and chronic Avp treatment in boys with learning disorders enhances memory of stories (Hamburger-Bar et al., 1987). As post-trial administration of an Avp analog does not enhance recall of passages, any effects of Avp on memory would be on acquisition only (Beckwith et al., 1995; see also Born et al., 1998).

Other studies indicate that Avp does not affect memory directly, but may enhance attention and arousal. DDAVP improves reaction times in males, but not memory itself (Beckwith et al., 1983). Avp in young and old subjects significantly elevates event-related potentials in the brain, which are associated with attentional mechanisms, but does not improve age-related cognitive impairments (Dodt et al., 1994). Similarly, repeated intranasal administration of Avp to healthy elderly persons does not significantly improve general long-term memory, although certain aspects of word recall are enhanced (Perras et al., 1997). Avp given to males does not affect consolidation of declarative memories, but there is a significant arousing effect of Avp as measured by electroencephalographic activity (Gais et al., 2002).

2.3.3. Conclusion

Administration of Avp agonists and antagonists, as well as data from Avpr1a and Avpr1b knockout mice, indicates a positive relationship between Avp and memory, both social and non-social. It is also possible that some of the effects of peripheral administration of Avp on behavior reflect changes in arousal, perhaps after an increase in blood pressure (see Koob et al., 1989, for review). Use of conditional knockouts for both the Avpr1a and Avpr1b are likely to be particularly useful in evaluating Avp's role in memory.

2.4. Stress, Anxiety and Depression

Anxiety and depression are behavioral states associated with stress. Pharmacological and transgenic studies in rodents allow the modeling of anxiety- and depression-related behaviors. In this section, the role of Avp and its two main receptors in the brain, Avpr1a and Avpr1b, will be discussed in relation to modulating anxiety and depression, as well as to actions on the HPA stress axis (see Table 2).

2.4.1. Anxiety

High densities of vasopressinergic fibers (De Vries and Buijs, 1983) and Avpr1a (Ostrowski et al., 1992) are found within the septum. Anxiety-related behavior, as measured on the elevated plus maze (EPM), is significantly reduced with septal infusion of antisense oligodeoxynucleotide to the Avpr1a mRNA (Landgraf et al., 1995), as well as with the Avpr1a antagonist d(Ch₂)₅Tyr(Me)AVP (Liebsch et al., 1996; Engelmann et al., 2000), in the absence of altered locomotor activity. However, Appenrodt et al (1998) found both intraseptal and intraperitoneal (i.p.) application of Avp is also anxiolytic, indicating possible actions outside of the septum. Additionally, blocking Avpr1a in the septum via the antagonist d(Ch₂)₅Tyr(Et)VAVP either fails to affect anxiety (Appenrodt et al., 1998) or increases anxiety (Everts and Koolhaas, 1999). Technical and antagonist differences may account for some of the differing results.

Some transgenic mice studies indicate a role for Avpr1a in modulating anxiety. Avpr1a KO males display significantly less anxiety-related behavior than wildtype males on a variety of tasks (Bielsky et al., 2004; Egashira et al., 2007). Additionally, over-expression of the Avpr1a gene in the LS significantly increases anxiety-related behavior (Bielsky et al., 2005a), further implicating Avpr1a in controlling the anxiety response. However, in another lab, no anxiolytic phenotype has been found in this line (Caldwell et al., 2006). Differences in strain background and technique may account for these differences. Interestingly, Avpr1a KO females do not display anxiolytic behaviors (Bielsky et al., 2005b; Caldwell et al., 2006). Sexual dimorphism in the extra-hypothalamic Avp system that ultimately projects to the septum (Wang et al., 1993) may underlie the lack of an Avp-related anxiety phenotype in Avpr1a KO females. Current work has implicated oxytocin pathways in the amygdala as underlying anxiety in females (Amico et al., 2004).

Avp and Crh, both made in the PVN, synergize to increase release of ACTH from anterior pituitary corticotrophs (for review see Aguilera and Rabadan-Diehl, 2000b; Carrasco and Van de Kar, 2003). Avp release significantly increases in the PVN in response to acute (Wotjak et al., 1996) and chronic (Aguilera and Rabadan-Diehl, 2000b) stress. Stress can also increase Avpr1b in the pituitary (Aguilera et al., 1994; Aguilera and Rabadan-Diehl, 2000a,b). As expected, antagonism of Avpr1b reduces ACTH secretion (Serradeil-Le Gal et al., 2002b), and Avpr1b −/− mice have significantly lower ACTH and corticosterone levels compared to wildtype mice after certain insulin-induced hypoglycemic and restraint stresses (Lolait et al., 2007a).

The creation of an antagonist specific for the Avpr1b (SSR149415) has allowed greater analysis of the role Avpr1b plays in regulation of the stress response (see Serradeil-Le Gal et al., 2005 for review). Specifically, oral administration of SSR149415 prevents the Avp-induced rise in plasma ACTH levels, confirming a specific relationship between Avpr1b and the stress response (Serradeil-Le Gal et al., 2002b). Additionally, oral and i.p. administration of SSR149415 is anxiolytic in various classical (e.g., EPM) and atypical (mouse defensive test battery and social defeat) models of anxiety (Griebel et al., 2002; Serradeil-Le Gal et al., 2002b; Serradeil-Le Gal et al., 2003). Effects of hypophysectomy have been mixed; in one study, hypophysectomy did not alter the anxiolytic effects of the Avpr1b antagonist (Griebel et al., 2002), while in another study hypophysectomy blocked those effects (Shimazaki et al., 2006).

SSR149415 has no anxiolytic effects in either punished drinking behavior or exploration of the EPM when administered into the LS (Stemmelin et al., 2005). However, infusion of SSR149415 into the basolateral amygdala, but not central or medial nuclei, produces anxiolytic-like effects in performance on the EPM (Salome et al., 2006). All three nuclei contain Avpr1b receptors (Stemmelin et al., 2005) and likely mediate anxiety-related behaviors in both non-humans (Menard and Treit, 1999) and humans (Sheline et al., 1998).

Interestingly, Avpr1b −/− males do not have an anxiolytic phenotype (Wersinger et al., 2002; Caldwell et al., 2006). The normal catecholamine release in response to stress is decreased in Avpr1b −/− males, particularly release of epinephrine (Itoh et al., 2006). As noted above, no differences in circulating ACTH and corticosterone levels between Avpr1b KO and wildtype mice have been found (Wersinger et al., 2002; Lolait et al., 2007a). Using the resident-intruder paradigm, stress-induced changes in corticosterone did not differ between Avpr1b KO and wildtype mice (Wersinger et al., 2002). However, after chronic, but not acute, restraint stress, significantly less ACTH is released in Avpr1b −/− mice as compared to wildtype mice (Lolait et al., 2007a). Furthermore, acute insulin-induced hypoglycemic stress and acute ethanol or lipopolysaccharide administration elevate both ACTH and corticosterone to lesser extents in the Avpr1b KO (Lolait et al., 2007a; Lolait et al., 2007b). These studies indicate that Avpr1b contributes to normal stress responses in a context-dependent manner.

2.4.2. Trait-dependent anxiety

Rats bred for high or low anxiety (HAB and LAB, respectively) provide an animal model for studying the relationship between abnormal HPA activity and neuroendocrine regulatory factors. HAB rats display greater depressive- and anxiety-like behaviors that are accompanied by enhanced stress-induced secretion of ACTH and corticosterone than LAB rats (Landgraf et al., 1999; Landgraf, 2006). The HPA axis of male HAB rats is incompletely suppressed by the synthetic corticosteroid dexamethasone (DEX): treatment with Crh after DEX pretreatment stimulates ACTH secretion, indicating abnormal HPA activation (Keck et al., 2002). Additionally, Avp is significantly elevated and released within the PVN of the hypothalamus in HAB, but not LAB, male rats, both with (Keck et al., 2002) and without stress (Wigger et al., 2004; Bosch et al., 2006). Administration of the Avpr1a antagonist d(Ch₂)₅Tyr(Me)AVP suppresses the ACTH secretion normally seen in HAB rats in response to DEX/Crh administration (Keck et al., 2002), as well as decreases anxiety- and depression-related behaviors (Wigger et al., 2004). Interestingly, prenatal stress increases Avp mRNA expression in the PVN of LAB rats and subsequent anxiety in the EPM and hole-board tests of anxiety (Bosch et al., 2006).

Single nucleotide polymorphisms (SNPs) have been studied as potential causes of the different HAB and LAB phenotypes (Murgatroyd et al., 2004; Landgraf et al., 2007). One specific SNP [A(-1276)G] results in reduced binding of the transcriptional repressor CArG binding factor A (CBF-A), which is highly co-expressed in Avp-containing neurons in the PVN (Murgatroyd et al., 2004). Reduction of CBF-A binding causes increased transcription of Avp in the PVN of HAB animals and, perhaps, the subsequent high-anxiety phenotype. Conversely, an SNP leading to an amino acid change in the signal peptide of Avp may lead to the decrease in Avp in the LAB rats (see Landgraf, 2006; Landgraf et al., 2007 for review).

2.4.3. Depression

Ample evidence indicates that, in addition to affecting anxiety-related behavior, Avp affects depression-related behavior. In rodents, the forced swim task produces immobility that can be countered with antidepressants (Porsolt et al., 1977; West and Weiss, 2005). Acute forced swimming induces a significant increase in Avp release in both the SON and PVN that increases further in the PVN with repeated swimming (Wotjak et al., 1998). Similarly, training in the Morris water maze for three days significantly increases release of Avp in the PVN after each test session, and Avp release tends to rise each day (Engelmann et al., 2006). Forced swim stress also increases Avp release in the mediolateral and ventral septum (Ebner et al., 1999), the amygdala (Ebner et al., 2002), and the SCN (Engelmann et al., 1998). Brattleboro rats show less immobility on the forced swim task (Mlynarik et al., 2007). Forced swimming does not elevate plasma Avp levels, indicating dissociation between central and peripheral Avp release with an applied stressor (Wotjak et al., 1998).

Applications of the Avpr1a antagonist d(Ch₂)₅Tyr(Me)AVP into both the mediolateral septum and the amygdala have antidepressant-like effects (Ebner et al., 1999; Ebner et al., 2002). Antidepressant-like effects are observed with oral and i.p. administration of the Avpr1b antagonist SSR149415, even after hypophysectomy (Griebel et al., 2002). Infusion of SSR149415 into the LS (Stemmelin et al., 2005) and into three nuclei of the amygdala also decreases immobility time in the forced swim task (Salome et al., 2006). Finally, ACTH increases after forced swim in wildtype, but not Avpr1b KO, mice (Tanoue et al., 2004). However, wildtype, Avpr1a −/− and Avpr1b −/− mice do not differ in this swimming behavior (Bielsky et al., 2004; Tanoue et al., 2004; Caldwell et al., 2006; Egashira et al., 2007). Nevertheless, Avpr1b in general appears to modulate depression-related behaviors and their associated HPA responses.

2.4.4. Clinical studies

A large body of evidence indicates involvement of Avp in human stress-related diseases, such as anxiety disorders and major depression. Gold et al (1978) were the first to propose a role for vasopressin in mood disorders; studies since then have upheld this idea. Patients with major depression have significantly elevated plasma Avp compared to healthy controls (van Londen et al., 1997), as well as elevated ACTH and cortisol levels after Avp administration (Dinan et al., 2004). An increase in Avp mRNA in the SON of depressed patients compared with controls could explain the elevated plasma Avp levels (Meynen et al., 2006). Interestingly, depressed patients with the melancholic subtype have significantly greater Avp mRNA in both SON and PVN, indicating the increase in Avp may be more pronounced with this type of depression (Meynen et al., 2006). Similarly, depressed patients who attempted or committed suicide have higher plasma Avp (Inder et al., 1997) or Avp-ir neurons in the PVN (Merali et al., 2006), respectively, than depressed patients who did not. Thus, higher Avp expression in the hypothalamus may be associated with more severe types and symptoms of depression, reflected in alteration of the HPA axis (for review, see Landgraf, 2006).

The likelihood of developing a mood disorder increases with a family history of the disease (Sullivan et al., 2000). A recent study indicates that above-normal plasma Avp (relative to healthy controls) is highly correlated to higher anxiety and psychomotor retardation levels, and is associated with family history of depression (Goekoop et al., 2006). Additionally, an SNP in the Avpr1b gene may protect against development of major depression, as it is significantly over-represented in healthy controls as compared to depressed patients (van West et al., 2004). Thus, a fair amount of evidence exists to implicate Avp in anxiety and depression.

2.4.5. Conclusion

Avp, particularly within the LS, seems to aid in the regulation of anxiety- and depression-related behavior. However, the precise contributions of each Avp receptor remain unclear, in part due to differing results using receptor antagonists and transgenic mice. Ample evidence exists indicating a role for Avp in regulation of the HPA axis in both rodents and humans, which could account for some of the anxiolytic or antidepressant effects of decreased Avp. Future studies using conditional knockouts of the Avpr1a and Avpr1b will allow for greater insight into the relationship between Avp and stress-related states.

3. Implications for Human Behavior

Many of the defining characteristics of autism spectrum disorders (ASD) such as atypical social behavior, repetitive and stereotypic behavioral patterns, and increased anxiety and emotionality (for review see Baron-Cohen and Belmonte, 2005; Lord et al., 2000), are affected by Avp. Interestingly, ASD are more prevalent in males (Fombonne, 2003; Chakrabarti and Fombonne, 2005); likewise the effects of Avp on these behaviors are most pronounced in male animals (Bluthe et al., 1990; Bielsky et al., 2005b; for review see Carter, 2007). While the above observations may simply be correlative, there have been a few studies of human sibling pairs (only one affected) and child-parents trios hinting that the Avpr1a is an autism susceptibility gene (Kim et al., 2002; Wassink et al., 2004; Yirmiya et al., 2006). A more powerful approach using more extended families found a significant linkage to chromosome 112q14 where, among other genes, the Avpr1a gene is located (Ma et al., 2007). While the 5′ flanking region of the Avpr1a is not thought to cause autism, because ASD likely has diverse etiologies (Muller, 2007; Sebat et al., 2007; Yang and Gill, 2007), it has been proposed that its variation could contribute, along with other genetic and environmental factors, to affect the onset and severity of the disease (for review see Hammock and Young, 2006; Carter, 2007).

The microsatellite sequences of the 5′ flanking region of the Avpr1a have been examined for their influences on various behaviors. A very strong linkage to creativity in dance was found by Bachner-Melman et al. (2005), with a weaker correlation between the microsatellite sequences and eating behavior (Bachner-Melman et al., 2004), and none for hypersexuality (Geller et al., 2005). These linkage studies, taken together, suggest a possible role for these polymorphic microsatellites in influencing Avpr1a's role in social communication and/or the interpretation of social context. Also, as noted above, a particular SNP in the Avpr1b is associated with reduced susceptibility to recurrent major depression (van West et al., 2004). These genetic approaches, coupled with the development of better behavioral animal models, suggest that the next decade should be a rewarding one for investigators of Avp.

Abbreviations

ACTH

adrenocorticotropic hormone

AH

anterior hypothalamus

Avp

arginine vasopressin

ASD

autism spectrum disorders

BNST

bed nucleus of the stria terminalis

CNS

central nervous system

CSF

cerebrospinal fluid

Crh

corticotrophin releasing hormone

DDAVP

Desmopressin

DEX

dexamethasone

HPA

hypothalamic-pituitary-adrenal

ir

immunoreactive

IGR

intergenic region

i.c.v.

intracerebroventricular

i.p.

intraperitoneal

KO, −/−

knockout

LS

lateral septum

MeA

medial amygdala

MPOA-AH

medial preoptic-anterior hypothalamic area

Oxt

oxytocin

Oxtr

oxytocin receptor

PVN

paraventricular nucleus

5-HT

serotonin

SNPs

Single nucleotide polymorphisms

SCN

suprachiasmatic nucleus

SON

supraoptic nucleus

Avpr1a

vasopressin 1a receptor

Avpr1b

vasopressin 1b receptor

Avpr2

vasopressin 2 receptor