Dissecting neural mechanisms of prosocial behaviors

Abstract

Prosocial behaviors are essential for group cooperation, which enrich life experience and enhance survival. These complex behaviors are governed by intricate interactions between numerous neural circuits functioning in concert. Impairments in prosocial interactions result from disruptions of this coordinated brain activity and are a prominent feature of several pathological conditions including autism spectrum disorder, depression and addiction. Here we highlight recent studies that use advanced techniques to anatomically map, monitor and manipulate neural circuits that influence prosocial behavior. These recent findings provide important clues to unravel the complexities of the neural mechanisms that mediate prosocial interactions and offer insights into new strategies for the treatment of aberrant social behavior.

Introduction

Prosocial behavior, often termed sociability, is a complex amalgamation of a variety of distinct types of social interactions. Positive prosocial interactions occur in many species ranging from insects to mammals and are critically important for development, survival, and reproduction. Thus, the neural mechanisms mediating prosocial, non-aggressive interactions have likely been evolutionarily conserved and involve circuits that play a role in a range of motivated behaviors that are critical for survival.

Pioneering studies in prairie voles demonstrated that the actions of the neuropeptide oxytocin (OXT) in the nucleus accumbens (NAc) were critical for pair bonding, providing one of the first hints of the neural circuit mechanisms regulating one particular form of sociability [1,2]. In the ensuing years, it became clear that other neuromodulators, including dopamine (DA) and serotonin (5-HT), are also important for adaptive social behaviors [3,4]. This body of work generated the hypothesis that these molecules regulate prosocial interactions by modulating neuronal activity in key nodes of the mesolimbic reward circuitry, perhaps by tuning intrinsic neuronal properties and filtering fast synaptic transmission [5,6]. However, progress in delineating more precisely the neural circuits governing prosocial behaviors did not occur until the arrival of now standard tools, which enable genetic access to discrete neuronal populations to monitor and precisely manipulate their activity patterns [7,8]. Here, we confine our discussion to recent findings on the brain regions, circuits and neuromodulators in mice implicated in non-aggressive, non-sexual prosocial interactions, which encompass social reward, social motivation, and social memory. In addition, we discuss the neural mechanisms underlying sociability impairments in rodent models of neuropsychiatric conditions.

Prosocial behaviors and social reward

The foundational investigations on the role of OXT in prairie vole pair bonding begged the question of its mechanism of action in the NAc. In mice, social reward was found to depend on OXT-induced release of 5-HT from dorsal raphe (DR) inputs in the NAc [9*]. The increase in NAc 5-HT induces long-term depression (LTD) of excitatory synaptic transmission in the NAc via activation of presynaptic 5-HT1b receptors, which are required for the development of a social conditioned place preference [9*]. While 5-HT had previously been implicated in regulating social behaviors [3,4,10,11], these findings [9*] generated the prediction that 5-HT release, specifically in the NAc would play a critical role in sociability. Recent experiments using a “circuits-first” approach have confirmed this prediction [12**]. Selective optogenetic activation of DR 5-HT inputs to the NAc enhanced prosocial behaviors while inhibiting these inputs decreased sociability. Furthermore, the prosocial action of activating DR 5-HT neurons was prevented by infusion of a 5-HT1b receptor antagonist into the NAc [12**].

Recent findings on the mechanisms of action of the recreational drug (±)3,4-methylenedioxymethamphetamine (MDMA), which is known to have powerful prosocial effects in human subjects [13], provide further support for the critical role of 5-HT release in the NAc in sociability [14**]. Specifically, in rodents, direct NAc infusion of MDMA, which causes large increases in 5-HT levels due to its potent interaction with the 5-HT transporter (SERT) [15], promoted sociability in the three chamber task, while NAc infusion of a 5-HT1b receptor antagonist prevented the prosocial effect of parenteral MDMA administration. Furthermore, MDMA application generated LTD in the NAc due to activation of 5-HT1b receptors [14**; but see 16]. MDMA may also prolong a developmental critical period for social reward learning due to OXT release [16]. A challenging question that warrants further investigation is how 5-HT induced depression of excitatory transmission at some unknown population of NAc inputs leads to the enhancement of social reward and sociability.

Unlike DA release in the NAc, 5-HT release is not inherently rewarding [7,12**,17, but see 18], suggesting that 5-HT modulation of NAc activity must differ from the modulation caused by DA. Nevertheless, presumably because of its powerful role in influencing a range of appetitive behaviors, DA is also a critical regulator of social behavior [19]. Indeed, ventral tegmental area (VTA)-to-NAc DA neurons exhibit increased activity during social interactions and optogenetic activation of this circuit enhances sociability, due to activation of D1 receptors in the NAc [20]. Consistent with a role for NAc DA release in social reward, OXT also acts in the VTA to promote social reward via enhancement of DA cell firing [21–23]. Thus, OXT plays a critical role in promoting social reward by influencing two key nodes of mesolimbic reward circuitry, the NAc and VTA.

The consequences of activating excitatory inputs to the NAc fit well with the putative role of depressing NAc excitatory synaptic transmission in promoting sociability. Activation of a subset of NAc projecting neurons in the prelimbic cortex (PL) decreased the preference for a social target, while their activity increased during social investigation, but only in specific locations [24**]. Activation of basolateral amygdala (BLA) inputs in the NAc also decreased sociability and increased social avoidance, but did not reduce palatable food seeking [25]. Collectively, these studies provide compelling evidence for a critical role of the NAc in social reward and prosocial behaviors. A major challenge will be to elucidate how the different sets of excitatory and modulatory NAc inputs modify NAc activity in a coordinated fashion to robustly and perhaps specifically influence the rewarding aspects of social interactions.

Social Memory

Social motivation is influenced by social cognition and social memory, the ability to recognize and remember conspecifics, respectively. Older studies revealed that OXT acting in the medial amygdala and septum is necessary for the social preference of a novel, rather than a familiar, conspecific [2,26,27]. More recently, via the use of transgenic mouse lines, specific sub-regions of the hippocampus have been implicated in the storing and processing of memories associated with social interactions. Selective inactivation of the dorsal CA2 (dCA2) region using the Amigo2-Cre driver line reduced social memory, but surprisingly did not influence sociability per se nor other hippocampal-dependent forms of memory [28*]. Similarly, excitotoxic lesion of the CA2 abolished social memory, but not olfactory memory [29]. Furthermore, single-unit recordings revealed that unlike CA1 neurons, CA2 neuron firing patterns remapped during social encounters and exhibited reduced response to spatial stimuli compared to CA1 neuron firing [30].

A heterosynaptic form of input-timing-dependent inhibitory LTD (iLTD) in CA2 parvalbumin-positive (PV+) interneurons provides one plausible physiological substrate for social learning and memory [31] in that blockade of this iLTD in PV+ interneurons impairs social memory [31,32]. Consistent with this proposal, 22q11.2 deletion mice exhibit social memory deficits associated with reduction of CA2 PV+ interneuron plasticity [29]. Modulation of CA2 pyramidal neuron firing by OXT and/or vasopressin may also be important for social memory. The CA2 region expresses a very high density of OXT receptors (OXTRs) [33–35] and vasopressin 1b receptors (Avpr1bRs) [36] compared to other hippocampal sub-regions. OXTR activation increases CA2 pyramidal neuron burst firing [33], while conditional deletion of OXTRs from the CA2 and CA3 sub-regions prevents social memory [34,35]. Furthermore, activating vasopressin inputs in the CA2 enhances social memory due to actions at Avpr1bRs [36].

Recent studies have also revealed the importance of the ventral CA1 (vCA1) in social memory [37**–40]. Leveraging the specificity of transgenic lines for vCA1 and dorsal CA1 (dCA1), selective inhibition of the two populations of pyramidal neurons demonstrated that only vCA1 neurons participate in encoding social memory [37**]. Projections from vCA1 to both the NAc and medial prefrontal cortex (mPFC) appear to be specifically important in regulating social memory as are inputs from dCA2 to vCA1 [37**–39]. Collectively, these studies suggest a circuit substrate that is critical for social memory, involving OXT and AVP modulatory inputs to dCA2, which relays social information to vCA1 that in turn influences extrahippocampal regions including the NAc and mPFC.

Pathological social motivation

Impairments in adaptive prosocial behaviors (e.g. social amotivation or avoidance) are common debilitating features of many neuropsychiatric conditions, including autism spectrum disorder (ASD), depression, and addiction [4]. Here, we briefly review recent findings on some of the neural mechanisms that contribute to sociability impairments in rodent models of these disorders.

Autism Spectrum Disorders

Rodent models of ASDs based on causal genetic variants have been invaluable tools for advancing the understanding of ASD pathophysiology. One common cause of ASD is a copy number variation on chromosome 16p11.2. Selective deletion of the syntenic region of chromosome 16p11.2 from DR 5-HT neurons caused significant deficits in prosocial behaviors, which were associated with reductions in the activity of these neurons during social interactions, as well as decreases in their intrinsic excitability [12**]. Optogenetic activation of DR 5-HT inputs in the NAc restored sociability in these mutant mice to normal levels and this rescue was dependent upon 5-HT1b receptors [12**]. Impaired social interactions were also observed in mice expressing a gain-of-function SERT variant that decreased 5-HT levels [41]. Collectively, these findings provide further support for the hypothesis that 5-HT release in the NAc is critical for sociability.

Alterations in the balance of excitatory and inhibitory synaptic transmission, so called E/I balance, in specific brain regions are found in several genetic ASD models. In the CNTNAP2 deletion mouse, correction of E/I imbalance by optogenetic activation of inhibitory PV neurons in mPFC rescued social deficits [42*]. Similarly, in a 16p11.2 duplication mouse model, restoring mPFC inhibitory synapse function reversed social and cognitive impairments [43*]. Surprisingly, E/I imbalance in the anterior cingulate cortex, but not the adjacent mPFC, contributes to social deficits in Shank3 deletion mice [44*]. These mice also exhibit aberrant BLA-to-NAc activity, modulation of which via endocannabinoids [25] restored adaptive social interaction, as did activation of DR 5-HT neurons [45].

Direct modulation of GABA receptors may serve as an alternative strategy to remedy deficits in E/I balance as evidenced by the finding that systemic administration of the GABA_B receptor agonist, R-Baclofen, reversed social deficits in two different variants of the 16p11.2 deletion model [46]. However, altered E/I balance in the somatosensory cortex was observed in four different genetic ASD mouse models without any corresponding changes in overall circuit excitability, suggesting that altered E/I balance may be a homeostatic compensation rather than a mechanism for impaired sociability [47**].

Major Depressive Disorder

Sociability deficits are a key symptom of major depressive disorder and have been attributed to abnormalities in the mesolimbic DA system [48]. The specific form of stress used to generate depression symptoms appears to influence the pathophysiological mechanisms that mediate the sociability deficits. Chronic mild stress may influence subpopulations of VTA DA neurons differentially with a reduction in DA release in target regions presumably contributing to the behavioral deficits [48]. In contrast, the sociability deficits observed in susceptible mice following chronic social defeat stress (CSDS) appear to require the release of brain-derived neurotrophic factor (BDNF) from VTA DA terminals in the NAc [48]. Early life stress, on the other hand, induces social impairments, at least in part by altering the transcriptional profile of VTA DA neurons [49*] and reducing DA receptor 3 signaling in the lateral septum [50,51].

Changes in inputs to the VTA also regulate CSDS-induced social deficits. Specifically, blocking the stress-induced increase in ventral pallidum (VP)-to-VTA inhibitory transmission reverses social avoidance, whereas a distinct population of VP inputs to the lateral habenula (LHb), a potent regulator of DA neuron activity, mediate passive coping [52**]. Interestingly, a di-synaptic inhibitory circuit from the retina to the LHb mediates the antidepressant and prosocial effects of light therapy in the CSDS model [53].

As a key node of mesolimbic reward circuitry, modulation of NAc function has long been thought to play a critical role in depression [54]. Recent work suggests that CSDS differentially alters excitatory inputs onto NAc medium spiny neuron subtypes with enhanced synaptic transmission for one thalamic input being critical for social avoidance [54]. NAc cholinergic interneuron activity is also impacted by stress and modulation of ion channels in this cell population normalized stress-induced decreases in social behavior [55].

Of course, the symptoms of depression including sociability deficits involve circuit modifications beyond those occurring in reward circuitry. Interrogation of network activity using multi-circuit in vivo recordings coupled with machine learning revealed that chronic stress disrupts synchronous activity in the mPFC, amygdala, and VTA mesocorticolimbic network. Restoration of mPFC activity normalized network dynamics and sociability in susceptible mice [56]. Additionally, pharmacological manipulations of the mPFC, via the novel antidepressant ketamine, improves social behavior in part by stimulating descending PFC inputs to the dorsal periaqueductal gray [57].

Addiction

Drug addiction and withdrawal are commonly associated with impaired sociability. While little is known about the precise neural mechanisms governing these impairments, recent work implicates alterations in cytokine signaling in the LHb and changes in opioid receptor activity [58,59]. Furthermore, the notion that increasing social support could be a means for attenuating addictive behaviors is supported by recent findings where rats reduced drug intake when provided with the choice to socialize [60*]. Socializing also reduced drug craving which was mediated by a discrete micro-circuit in the central amygdala [61*]. These recent studies suggest that, similar to depression, maladaptive changes in corticostriatal circuitry and key modulatory inputs to the NAc contribute to abnormal social motivation in addiction.

Conclusions

What is more important in today’s world than developing interventions that will promote empathic and compassionate positive, prosocial interactions? As neuroscientists, we can hopefully contribute to this effort by delineating the complex neural mechanisms underlying social reward and social motivation. Given that sociability deficits are present in a range of neuropsychiatric conditions, a more sophisticated and comprehensive understanding of the pathophysiological circuit activity that generates these deficits will also aid in the development of improved treatments. To date, much of the research focus on these topics has appropriately been on specific circuits and cell types known to play a role in many different types of motivated behaviors. Future studies will need to assess how social cues engage these separate circuits in a manner to modify social interactions and how they work in concert to regulate and promote prosocial interactions. Perhaps with sufficient knowledge of neural mechanisms, we can help our species promote prosocial behaviors and simultaneously reduce the aggressive, self-destructive social behaviors that threaten our very survival.

Highlights.

• Neural circuits mediating prosocial behaviors are being defined using modern methods

• Modulation of mesolimbic reward circuitry plays a key role in promoting sociability

• Social memory involves specific subregions of the hippocampus

• Prosocial behavior is deficient in rodent models of autism, depression, and addiction