Biological Contribution to Social Influences on Alcohol Drinking: Evidence from Animal Models

2.1. Peer-Rearing in Primates

A number of studies have examined the effects of peer-rearing in rhesus macaques. These studies compared monkeys that were raised in the absence of the mother or other adults for the first six months of life, but had been housed in social groups of four peers, to monkeys that were raised by their mothers. Although peer-reared monkeys established bonds with their peers, they also exhibited more fear-related behaviors and decreased exploration compared to mother-reared monkeys. When the 50-month old animals were given one hour daily access to sweetened 7% alcohol four days a week, peer-reared monkeys drank significantly more than the mother-reared animals. However, when the animals were isolated, the mother-reared monkeys increased their alcohol intake to the level of peer-reared, which was not increased by the separation [14,15]. In addition to alcohol intake, peer-rearing affected a number of other behavioral phenotypes including aggression, impulsivity, and social behavior [16,17].

These studies indicate that early childhood social rearing experiences affect future predisposition to excessive alcohol intake. However, it should be noted that in this study, peer-reared monkeys had elevated plasma cortisol levels at baseline, and increased cortisol and ACTH concentrations during acute separations. Importantly, the peak cortisol levels during separation were positively correlated with alcohol intake [14]. Therefore, it is difficult to know whether the increased alcohol consumption in peer-reared animals was due specifically to the rearing conditions, or if it was a secondary effect, due to the elevated stress hormone levels. Non-social stressors, such as foot-shock exposure in adolescence, can affect alcohol intake in adulthood [18,19], indicating that different forms of stress may yield the same behavioral outcome. In the case of peer-rearing, this negative social environment mediates the stress-induced increase in alcohol intake. In this and other cases of negative social environments that will be presented in this review, it becomes difficult to determine what aspects of the social environment or stress response may be responsible for alterations in alcohol drinking.

Serotonin

In addition to exhibiting high levels of alcohol intake, peer-reared non-human primates had lower CSF levels of serotonin metabolite 5-hydroxyindolacetic acid (5-HIAA) [15], which is associated with high levels of alcohol intake in humans and other primates [20]. Further, the selective serotonin reuptake inhibitor sertraline decreased alcohol consumption in peer-reared monkeys, while also disrupting the increased aggression observed in these animals [20]. These studies implicate a role for serotonin in mediating the behavioral effects of the peer-rearing experience.

The peer-rearing model described above has successfully elucidated a gene by environment interaction. Female macaques possessing the short allele for a polymorphism in the promoter region of the serotonin transporter (5-HTTLPR) exhibited a greater alcohol preference than those with the long allele, but only if they were peer-reared; mother-reared monkeys showed no effect of genotype on alcohol preference [21]. This indicates that the serotonin transporter plays an important role in modulating alcohol preference, since the genotype that results in lower mRNA levels of the gene corresponds to increased preference. Importantly, this finding also indicates that the effect of genotype can be buffered by the influence of mother-rearing.

Again, it is uncertain whether the effect is in fact due to the past social experiences, or to the elevated stress levels observed in peer-reared monkeys. Since the 5-HTTLPR is responsive to glucocorticoids levels, it is likely that this plays an important role in the gene by environment interaction detected. However, Barr et al. have also shown that animals with the short allele exhibit increased sensitivity to the ataxic and sedating effects of alcohol, independent of rearing condition [22], suggesting that the polymorphism may directly impact the effects experienced from alcohol, regardless of stress levels.

3.1. Isolation in Rodents

In a complex study of changes in housing conditions affecting alcohol intake, Wolffgramm and Heyne showed an increase in alcohol intake in isolated and contact-housed Wistar rats, compared to group-housed rats, and that housing changes resulting in increased levels of isolation (i.e., grouping to contact caging or contact to isolation) corresponded with increased drinking [28]. Similarly, Hall et al. observed an increased consumption of 16% alcohol (but not lower concentrations) in adult Fawn–Hooded and Wistar rats isolated since weaning, compared to pair-housed controls, although there was no difference in alcohol preference due to the two housing conditions [29].

While a majority of reports on isolation in rats confirm the increased alcohol consumption following isolation described above [30–37], decades of literature examining effects of isolation on alcohol intake include some apparent contradictions; some studies report no difference in voluntary alcohol intake in response to isolation stress, while few even report a decrease. As suggested by Schenk et al., there are a number of procedural differences that make direct comparisons between studies difficult. These authors were able to demonstrate the importance of age at isolation on subsequent alcohol intake, showing in Long-Evans rats that 12 weeks of isolation starting at weaning resulted in a considerable increase in alcohol intake compared to rats that had been housed four per cage for the 12 weeks preceding alcohol testing, while 12 weeks of isolation starting as adults did not affect alcohol intake [32].

More recent studies demonstrated the importance of the age at which social manipulations and testing are performed. Juvenile Wistar rats that had been continuously isolated exhibited a higher alcohol intake than rats that had been continuously socially-housed, socially-housed with intermittent isolation, or isolated with intermittent social contact, when tested during the same pre-pubescent period in which housing manipulations occurred. However, when the same animals were later tested in adulthood, the rats that had been socially-housed with intermittent isolation drank more than the other groups, but only when alcohol was available in social conditions. The remaining groups drank less in social conditions than in isolation, in adulthood [35]. This indicates that chronic isolation can induce higher drinking at a young age, while in adulthood drinking is increased in isolation regardless of previous social history, and social history can differentially affect drinking levels under social circumstances.

One contradictory study showed that alcohol intake was actually lower in isolated Wistar rats than pair-housed rats during forced consumption, while during subsequent two-bottle choice testing there was no difference [38]. In addition to the forced consumption, one distinction between this study and others is that the pair-housed animals were not only pair-housed preceding testing, but also throughout alcohol availability. Other studies have separated group-housed animals for testing of alcohol consumption. It is possible that the difference in social circumstances during testing can affect consumption, as demonstrated above [35]. However, the same lab tested alcohol preferring (P) and non-preferring (NP) rats in a similar manner and found that isolated preferring rats consumed more alcohol than paired preferring rats, while there was no difference in consumption between isolated and paired non-preferring rats [36], indicating that the mode of testing in pairs cannot be solely responsible for the discrepancy.

In contrast to the study described above by Ehlers et al. in selectively-bred alcohol preferring and non-preferring rats, Lodge and Lawrence reported that Fawn-Hooded rats that preferred alcohol did not differ in their alcohol consumption from isolated rats [39]. However, this could be due to a ceiling effect of alcohol consumption for this strain under these conditions, particularly since the control rats were divided into alcohol preferring and non-preferring subgroups based on their intake, and only the preferring rats were compared to isolated rats, which were not divided based on intake.

There are fewer studies examining the effects of isolation on alcohol intake in mice. Post-weaning isolation increased alcohol preference and intake in male C57BL/6J (C57) mice, tested as adults (60 days old) [40], and one week of isolation at 30 or 44 days of age increased alcohol consumption tested over the following two weeks in C57 and DBA/1 mice [41]. Together, these results indicate that isolation in adolescence or early adulthood can lead to increased alcohol intake in adult mice.

Many studies of the effects of isolation on alcohol intake have focused on the biological mechanisms contributing to the observed behavioral effects, and the serotonin, GABA, and dopamine systems, as well as HPA axis activation are each affected by the conditions of the social environment and play a role in the effect on alcohol drinking.

4.1. Role of Dominance in Established Colonies

While most studies of social dominance relationships have focused on rodents, primates also exhibit stable dominance hierarchies and can be useful for examining the effects of rank on alcohol intake. One study of social rank in squirrel monkeys living in colonies of 4–10 individuals has revealed that there is a negative correlation between the index of dominance and the amount of alcohol consumed [52]. The relatively higher consumption of alcohol in subordinates is consistent with effects observed in rodent species. In primates and rodents, there is evidence that subordination is linked with heightened HPA axis activity [53,54], further supporting the hypothesis that increased stress leads to increased alcohol intake.

One study examining social behavior in Long-Evans rats in established colonies found that in each of 10 colonies containing five males and three females, one male displayed a greater degree of aggressive behavior relative to defensive behavior in interactions with other males, and was considered dominant. These dominant rats consumed significantly less alcohol than subordinate rats did at both concentrations tested (4% and 8%) [55]. One interpretation the authors provide is that the difference in anxiety levels of dominant and subordinate rats affects alcohol intake, such that the more anxious subordinate rats consume more alcohol, to act as an anxiolytic, than dominant rats, which perceive less reward from alcohol intake. Notably, females in these colonies drank significantly more than males. This finding is in agreement with many studies in mice and rats showing that females consume more alcohol than males, and it also fits the proposed theory relating stress and anxiety to alcohol intake, since females are not the dominant animals in the colony. This group subsequently showed that levels of serotonin metabolite 5-HIAA were elevated in limbic brain regions and the spinal cord of subordinate rats compared to dominant or control (isolated) rats, indicating that increased serotonin turnover is linked with increased consumption [56], similar to findings in isolated rats described above. The negative correlation between dominance rank and alcohol intake in rats has been confirmed by others, in Wistar rats [28].

In male C57 mice exposed to seven consecutive days of confrontations by different pairings of dominant and subordinate animals, subordinate mice consumed more 20% alcohol than aggressive mice, particularly in the second of two weeks of alcohol testing, while CBA/Lac mice, a low-preferring strain, showed no effect of social role or experience on alcohol intake [57]. These results are repeatable for C57s [58] and are consistent with findings in rats, but they also show that the effect of social defeat or subordinance may not universally increase alcohol intake, since low-preferring animals did not alter their intake.

5.1. Demonstrator-Observer Rat Models of Alcohol Acceptance

Rats will exhibit a greater preference for a novel substance when they are allowed to observe another rat that has been exposed to the substance [66–68]. This demonstrator-observer paradigm has been extended to assess effects related to alcohol preference, but interestingly, alcohol odor preference was only increased in adolescent rats that had been able to interact with an alcohol-intoxicated peer, not in those that were exposed to an anesthetized rat that had also received alcohol [69]. Furthermore, alcohol preference was increased in adolescent male Sprague-Dawley rats that had been allowed to observe and interact with an intoxicated familiar cagemate, whereas alcohol preference was decreased in rats that had observed and interacted with an intoxicated unfamiliar peer [70]. This indicates that the relationship with the demonstrator is an important factor in interpretation of the stimulus substance. Interestingly, the relationship was not important for female adolescent rats, which exhibited an increased preference for alcohol after exposure to either a familiar cagemate or an unfamiliar peer. While in some cases familiarity may be important, these demonstrator-observer studies performed in rats have the disadvantages of, first, eliminating the possibility of interactions during drinking that may affect alcohol intake, and second, not allowing study of the effects of specific social affiliations on alcohol drinking.

6.1. Overlap in the Neurobiology of Affiliation and Drug Reward

The partner preference test has been developed to assess the strength of a pair bond in voles in the laboratory. This test compares the time the animal spends with a partner versus time spent with a stranger after a period of cohabitation with the partner [84]. Studies using this test have examined the effects of pharmacological manipulations on pair bond formation to elucidate a variety of neurobiological substrates involved in the social bond, which have been recently reviewed [85]. Importantly, many of these molecules are known to play a role in addiction as well. For example, the neuropeptide arginine vasopressin acting on central V1a receptors facilitates pair bond formation in male prairie voles [86], and is implicated in addiction, particularly alcohol intake [87]. Alcohol directly affects the release of vasopressin [88], and vasopressin is thought to play a part in modulating alcohol preference [89] and tolerance [90,91]. Oxytocin facilitates partner preference formation in female prairie voles [92], and has a number of addiction-related effects similar to vasopressin, including modulation of alcohol tolerance and dependence [93]. Together with oxytocin, dopamine is required for pair bond formation [94–96]. Activation of D1 receptors in the nucleus accumbens blocks partner preference formation, while activation of D2 receptors facilitates this formation. A large number of studies have demonstrated that the dopamine system is one of the neurotransmitter systems critically regulating addiction to various drugs [97–101], including alcohol [102]. The CRF system, a system regulating the HPA axis, is also implicated in regulation of pair bond formation. Importantly, central administration of CRF, in doses not affecting anxiety levels, promotes pair bond formation, while central administration of CRF antagonists attenuates these effects [103,104]. It is also well known that components of the CRF system regulate alcohol intake [105,106].

Preliminary evidence has shown that the opioid system may also be involved in pair bond formation despite relative resistance to manipulations of the opioid system in prairie voles [107]. Thus, female prairie voles injected with 7.5 mg/kg naltrexone, an opioid receptor antagonist, three times throughout the 18-hour cohabitation period did not exhibit a partner preference in the test following cohabitation, and exhibited a preference for the stranger, while the majority of vehicle-treated voles displayed a partner preference [108]. These recent studies are significant because the opioid system is important for regulation of alcohol-associated reward, and naltrexone is the first centrally-acting treatment approved for alcoholism.