The other side of the coin: hypersociability

Abstract

Affiliative social motivation and behavior, i.e. sociability that includes attachment, prosocial behavior (sharing, caring, and helping), and empathy (the ability to understand and share the feelings of others), has high variability in the human population, with a portion of people outside of the normal range. While psychiatric disorders and autism spectrum disorders are typically associated with a deficit in social behavior, the opposite trait of hypersociability and indiscriminate friendliness are exhibited by individual with specific neurodevelopmental disorders and following early adverse care. Here we discuss both genetic and environmental factors that cause or increase the risk for developing pathological hypersociability from human to rodent models.

Introduction

As shown in Fig. 1, sociability is variable in a population^1–3. Similar to many continuously variable behavioral phenotypes, sociability fits a normal (Gaussian) distribution, with most individuals falling in the middle of the range and some individuals, exhibiting pathological/abnormal phenotypes, at the left and right sides of the distribution^2,4–6. Deficits in social interactions and socialization are often diagnosed in anxiety, antisocial and psychopathic personality disorders, and autism spectrum disorders⁷. The opposite traits of excessive need for social contact, emotional dependence on continuous social company and attention have also been recognized in adolescents and adults⁸. Hypersociability and indiscriminate sociability are associated with some neurodevelopmental disorders and child maltreatment. In contrast to social behavioral deficits, abnormal hypersociability in human and animals models has not been comprehensively reviewed.

Figure 1.

Variability of sociability in a population.

1. Genetic variability in sociability

Twin and family studies in human, and genetic mapping in rodent genetic reference panels showed that individual levels of sociability, both high and low on the behavioral scale, are heritable, at least partly³. Furthermore, genetic association studies (GWAS) between measures of sociability and single nucleotide polymorphisms (SNPs) identified chromosomal regions/genes that, via gene expression, may explain some of the differences in sociability between individuals in a population. Genes involved in sociability/hypersociability are listed in Table 1.

Table 1.

Sociability/hypersociability associated genes

Genes	Condition	Species	Function		haploinsifficient	reference
OXTR	rs53576GG	human	gain of	receptor	N	17,18, 19
CD38	rs12644506	human		coregulator of OTX		20
Gtf2I and Gtf2ird1	Williams syndrome	human	loss of	transcription factor	Y	44
GTF2I	rs13227433	human		transcription factor		46
Gtf2I and Gtf2ird1	structural variants	dog		transcription factor		21
from Gtf2i to Fkbp6	targeted deletion	mouse	loss of		Y	69
from Gtf2i to Limk1	targeted deletion	mouse	loss of		Y	70, 71
Gtf2ird1	targeted deletion	mouse	loss of	transcription factor	Y	73, 74
UBE3A	Angelman syndrome	human	loss of	ubiquitin-protein ligase		52
Ube3a	targeted deletion	mouse	loss of			77
KANSL1	Koolen-De Vries syndrome	human	loss of	member of HAT complex	Y	64–67
DLG4	targeted deletion	mouse	loss of	synaptic protein	Y	64
DLG2	targeted deletion	mouse	loss of	synaptic protein	N	64
nNOS	targeted deletion	mouse	loss of	synaptic regulator		79
Nrg1	targeted deletion	mouse	loss of	signaling molecule	Y	103
Erbb3	targeted deletion	mouse	loss of	tyrosine kinase receptor		103

1.1. Genetic variability in sociability in human: Twin studies

In behavioral genetics, twin studies are fundamental to establish trait heritability. A classic twin study design compares trait similarity in same sex monozygotic (MZ) twins raised in the same family to that in same sex dizygotic (DZ) twins. MZ twins share virtually all, while DZ twins approximately 50% of their genes. Although twin studies have significant limitations and confounds (e.g. reciprocal influences between twins)⁹, difference in correlations between MZ and DZ twins reflects heritability. Heritability between 56%–72% was reported in twin studies for self-reported measures of altruism, empathy, and nurturance¹⁰. In another study, genetic influences accounted for 43% and 34% of the variance in observed self-initiated and compliant pro-social behavior in children, respectively¹¹. Further, genetic influences, although not apparent at 14 and 20 months, explained 34%–47% of the variance in empathy at 24 and 36 months of age¹². Although twin studies indicate a surprisingly high heritability for social behavior, they provide no information on trait-associated chromosomal regions and genes.

1.2. Genetic polymorphism in oxytocin pathway genes is associated with variability in social attachment and empathy in human

Sociability is a polygenic trait determined by the interaction of many genes, each with a relatively small contribution to the behavioral phenotype. GWAS identified SNPs in oxytocin pathway genes that influence the individual level of sociability. Oxytocin (OXT) is a neuromodulator that affects a range of social behaviors in humans and animals, including social attachment and higher order human social functions such as empathy^13,14. It is believed that OTX increases the salience and reinforcing value of social stimuli¹⁵. OTX is produced in the hypothalamus and is released via axons projecting to the pituitary, and via axon collateral to the nucleus accumbens (NAc) and other brain regions¹⁶. Beside of OTX, the genes for the OTX receptor (OTXR) and CD38, a protein that regulates OTX release (CD38), have been studied to explain variability in human social behavior. Evidence links more efficient signaling variants of OTX pathway genes to increased affiliative behavior. For example, OXTR rs53576GG homozygotes in one study self-report higher levels of sociality¹⁷, while in another study are rated as more prosocial than A carriers (AA, AG genotypes)¹⁸. Also, the OXTR rs53576GG genotype was associate with a higher level empathy in a population, as well as in a comprehensive meta-analysis of 13 independent studies¹⁹. Similarly, SNP rs12644506 in the CD38 gene was associated with social integration and social connectedness²⁰. Finally, the rs53576A allele that predisposes for low sociability was linked to morphometric alterations of the hypothalamus, specifically to a decrease in gray matter volume and to an increase in functional coupling between hypothalamus and amygdala¹⁷.

1.3. Sociability traits identified by GWAS in domestic dogs.

Dogs exhibit a heightened propensity to initiate social contact that is extended to members of other species, when compared with human-socialized gray wolves²¹. This behavior is likely driven by behavioral neoteny, which is the retention of juvenile features in the adult. A genome-wide association scan of more than 48,000 SNP genotypes from 701 dogs from 85 breeds and 92 gray wolves identified a top ranking outlier site located within the polymorphic Wbsrc17 gene, which is within the region typically deleted in Williams syndrome²². Williams syndrome patient exhibit hypersociability as discussed in section 2.1. A follow up study found that a 5-Mb genomic region around the Williams critical region was under positive selection in domestic dog breeds and that hypersociability is a core element of domestication that distinguishes dogs from wolves²¹. Structural variants in two genes in the Williams syndrome critical regions, Gtf2I and Gtf2ird1 (encoding two related transcription factors of the TFII-I family, likely contribute to the hypersociability of domestic dogs. These data indicate that human directed evolution of dogs through selective breeding retained specific structural variants of a unique set of linked genes with a large effect on sociability (i.e. survival of the friendliest²³).

1.4. Mapping of “sociability” traits in mice by using recombinant inbred strains.

Over the past decades, murine recombinant inbred (RI) strain families from naturally divergent strains have been developed via chromosome substitution to dissect complex genetic traits, including sociability^24,25. The BXD family, currently the largest and best characterized mouse RI collection, is composed of ~160 lines that descend from crosses between C57BL/6J and DBA/2J²⁶. Statistical genetic techniques in BXDs allow the identification of quantitative trait loci (QTL) that define regions of the genome (intervals or loci) and their modulating effects on the phenotype. The BXD family has been exploited to study the genetics of immune function, infectious disease, and behavior²⁷. More recently, BXDs were used to identify QTLs associated with sociability. Although the parental C57BL/6 and DBA/2 inbred strains exhibit low and intermediate levels of sociability, Knoll and coworkers reported a wide variability in social approach and vocal communication in BXDs, extending well above the average sociability of DBA mice²⁸. The quantitative trait locus for vocal communication, and associated candidate genes, were mapped to chromosome 18. However, the large size of the QTL does not allow specifying a single gene or set of genes that may directly contribute to the trait.

2. Pathological hypersociability in genetic conditions

Affiliative behavior and empathy that exceed the normal range of sociability can be a part of neurodevelopmental disorders such as Williams syndrome²⁹.

2.1. Williams syndrome

Williams syndrome arises from a genetic deletion of 7q11.23, in 1 out of approximately 20,000 births²⁹. Williams syndrome is characterized by cognitive deficits in the IQ range of 50-60, with relative strengths in face and object recognition and in some aspects of auditory and verbal abilities, but with severe weaknesses in spatial, motor and visual-motor abilities^30–32. However, Williams syndrome is most associated with exuberant gregariousness, lack of social inhibition, and enjoyment of conversations, a distinct phenotype from most neurodevelopmental disorders that are typically associated with social anxiety. Williams syndrome affects males and females in equal numbers²⁹, in contrast to a male bias in autism spectrum disorders (characterized by social anxiety). However, hypersociability in Williams syndrome is accompanied by a poor understanding of social dynamics, facial expression, and body language and Williams patients have high rates of symptoms of generalized and anticipatory anxiety³³. Toddlers with Williams syndrome have an intense, penetrating gaze, that is the opposite of the aversion of eye contacts in fragile X syndrome and autism spectrum disorders. Although individuals with Williams syndrome can hold simple jobs, they require assistance managing their lives. Distinctive cardiovascular problems limit the average lifespan of a person with Williams to around fifty.

Because of the well-defined, albeit multigene deletion, Williams syndrome provides an opportunity to study how hemizygosity in a relatively small set of genes shape complex social behavior such as an innate drive to connect with others. Williams syndrome usually occurs sporadically, though familial cases have been reported^34,35. Due to repetitive sequences flanking the commonly deleted regions, 98% of Williams individuals studied have the same deletion breakpoints. The commonly deleted region spans 1.5 Mb and include GTF2l and NCF1 on one (the telomeric) end, and their pseudogenes (P) on the other end^30,36–38 (https://www.jci.org/articles/view/35309/figure/2). Duplication of the identical region also occurs, albeit with lower frequency, causing an opposite behavioral phenotype that falls within the autism spectrum^39,40. This indicates that the dosage of at least some of the genes within the Williams critical region is essential for the development of normal social behavior, and that lower and higher gene dosage results in behavioral abnormalities in the opposite direction.

The commonly deleted region contains approximately 25 genes, encoding transcriptional regulators (GTF2I, GTF2IRD1, BAZ1B, MLXIPL), signaling molecules (FZD9, TBL2, LIMK1) and molecules with functions in various cellular processes (STX1A, CYLN2, FKBP6, EIF4H, CLDN3, CLDN4, VPS39D and RFC2)⁴¹. Although most Williams patients have the full deletion, there are individuals with partial deletions that prompted studies to associate individual genes with specific phenotypes/behaviors. Deletion of only ELN (encoding elastin) was linked to supravalvar aortic stenosis and cutis laxa with no cognitive and behavioral abnormalities, while deletion of LIMK1 was associated with visual-spatial constructive cognition, but not with most of the Williams symptoms^29,42. However, another study found that visual-spatial construction was associated with GTF2IRD1⁴³. Still another study that analyzed partial 7q11.23 deletions implicated GTF2IRD1, as well as its close relative GTF2I, as the main genes responsible for the Williams syndrome neurocognitive profile⁴⁴. The discrepancy between these studies could be due to the difficulty to establish specific gene–phenotype correlations in Williams syndrome because of phenotypic variability, presumably due to modifier loci. Nevertheless, a report describing a deletion of all genes but GTF2I in the critical region is noteworthy because the affected individual, unlike most Williams patients, did not show increased gaze and attention to strangers^43,45. This suggests that loss of GTF2I, at least partly, is involved in the social behavioral phenotype of Williams syndrome. However, it is important to point out that some individuals with the full deletion (i.e. including GTF2I) also have fairly normal social behavior. Nonetheless, genetic variability (instead of a hemizygous deletion) in GTF2I in healthy individuals supports a possible role for this gene in social behavior. Specifically, rs13227433AA, an SNP in GTF2I showed association with low social anxiety, combined with reduced social-communication abilities⁴⁶, a phenotype reminiscent to that seen in Williams syndrome.

Although none of the genes deleted in Williams syndrome have been functionally connected to the behavioral phenotype, brain structural⁴⁷ and functional⁴⁸ abnormalities provide a possible link between Williams syndrome genes and social disinhibition. A large number of brain abnormalities have been described in Williams syndrome, but probably the most prominent are reduced amygdala response to fearful facial expressions^48–50 and heightened amygdala reactivity to positive social stimuli⁴⁹. Reduced amygdala response to fearful facial expressions may underlie social disinhibition (i.e. approach of strangers), while increased amygdala response to happy facial expressions could explain the high motivation of Williams patients to interact with others. Consistent with the notion that loss of function in GTF2I is a major contributor to the clinical phenotype of Williams^43,45, the GTF2I rs13227433 AA genotype, previously associated with lower social anxiety in the population⁴⁶ was also associated with reduced amygdala response to threat⁵¹, linking this gene, via amygdala function, to social behavior.

2.2. Other neurodevelopmental disorders with hypersocial behavior

Angelman syndrome.

Deletion of chromosome 15q11-q13 in Angelman syndrome causes developmental delay (which becomes apparent by 6–12 months of age), lack of vocal speech, muscular hypotonia, epilepsy, and a typical behavioral profile that includes a happy demeanor, hypermotor behavior, and low attention span⁵². Frequent smiling and laughing are considered to be part of repetitive and stereotyped behavior in Angelman patients^53–55. However, some studies showed that laughing and smiling increase during social situations and occur at low levels during non-social situations⁵⁶, raising the possibility of increased motivation to interact with others.

Angelman syndrome is typically associated with the deletion of the maternal UBE3A gene at 15q12-q13, encoding the ubiquitin-protein ligase E3A⁵⁷. Since the paternal allele is imprinted and inactivated in the developing brain, Angelman patients lack UBE3A during development, one of the enzymes required for proper ubiquitination of proteins destined for proteolysis. Typical deletion in Angelman also includes genes with biallelic expression; three gamma-aminobutyric acid (GABA) receptor subunit genes and the P gene which, when mutated in both alleles, is responsible for oculocutaneous albinism type II.

17q21.31 microdeletion syndrome (Koolen-De Vries syndrome).

17q21.31 microdeletion syndrome, characterized by developmental delay, moderate intellectual disability, facial dysmorphisms, and anomalies of the brain and multiple organ systems, is a relatively recently discovered condition^58–60. Interestingly, clinical data indicate an amiable, friendly disposition, less social fear and more approach behavior^61–63. The 17q21.31 microdeletion includes the microtubule associated protein tau (MAPT ) gene^58–60 that normally promotes microtubule assembly and stability in neurites, corticotropin releasing hormone receptor 1 (CRHR1), C17orf69 (long intergenic non-protein coding RNA 2210) and KANSL1 (member of histone acetyltransferase complex). Koolen-De Vries syndrome, including intellectual disability and happy disposition, can be caused by heterozygous truncating mutations in KANSL1^64,65, demonstrating that inactivation of this gene is central to the 17q21.31 microdeletion syndrome. Indeed, recent comparison of the phenotype of microdeletion and KANSL1 SNV (single nucleotide variant) patients did not show differences of clinical importance, stressing that haploinsufficiency of KANSL1 is sufficient to cause the full Koolen-De Vries syndrome phenotype^66,67.

2.3. Mouse models of neurodevelopmental disorders exhibiting increased social behavior

Neurodevelopmenal conditions have been generated in mice by the genetic inactivation of Williams, Angelman, and 17q21.31 microdeletion syndrome candidate genes. These mouse models can reproduce the hypersociability phenotype of the human conditions.

Williams syndrome models.

The entire Williams syndrome critical region is conserved in the mouse⁶⁸ and its genetic manipulation provided further insight into the possible contribution of individual genes to the phenotypes of Williams syndrome. Heterozygous deletion of the mouse Williams critical region between Gtf2i and Fkbp6 (that includes Gtf2ird1) resulted in increased social interactions and less habituation in a repeated social approach task⁶⁹. Mutants with a partial deletion of the Williams critical region have also been generated⁷⁰. Proximal deletion (PD) mice lack Gtf2i to Limk1 (that includes Gtf2ird1), distal deletion (DD) mice lack Limk1 to Fkbp6, and double heterozygotes (PD/DD) model the complete human deletion. In the three chamber social test⁷¹, male PD and PD/DD mice spent significantly more time interacting with s stranger mouse, while in the tube test⁷² that assess social dominance, PD and PD/DD animals were less competitive than WT and DD mice against a control C57BL/6 mouse. This suggests that haploinsufficiency in the proximal part of the Williams critical region, that includes Limk1, Eif4h, Rfc2, Cyln2, Gtf2ird1, and Gtf2i, is associated with increased sociability and subordination. However, only PD but not PD/DD mice exhibited increased activity in direct social interaction suggesting a more complex gene-phenotype association in mouse models of Williams syndrome. Nevertheless, knockout mice heterozygous for Gtf2ird1 (a close relative of Gtf2i) was shown to exhibit an increase in the number and length of social interactions and a decrease in the number of aggressive interactions in the resident intruder test⁷³. Similarly, a more recent study reported that mice with a heterozygous deletion of Gtf2ird1 have increased social interaction with unfamiliar mice and do not show typical social habituation⁷⁴. Other genes within the proximal region of the Williams syndrome critical region may contribute to phenotypes other than hypersociability. Cyln2 heterozygote knockout mice exhibit mild growth retardation, mild brain abnormality, attenuated contextual fear, and reduced synaptic plasticity and motor function⁷⁵. No apparent behavioral phenotype was found in LimK1 heterozygote knockouts⁷⁶. Taken together, haploinsufficiency in GTF2I (human studies, see 2.1.) and/or in the related GTF2IRD1 (human and mouse studies) is probably critical in the development of the hypersociable phenotype of Williams patients.

Angelman syndrome model.

Consistent with the central role of UBE3A deletion on the maternal allele in the development of the Angelman syndrome behavioral phenotype, including typical happy demeanor, genetic deletion of Ube3a in mice was reported to increase preference for social stimuli in the three chamber social approach task and physical contacts of both males and females paired with unfamiliar genotype-matched females⁷⁷.

17q21.31 microdeletion syndrome model.

Human studies indicated that haploinsufficiency in KANSL1 causes 17q21.31 microdeletion syndrome. However, heterozygous ablation of Kansl1 caused no apparent change in social behavior in the three-chamber sociability test, albeit exhibited other behaviors typically seen in 17q21.31 microdeletion syndrome (i.e. impaired recognition memory and associative learning and memory and brain malformations)⁷⁸. A larger deletion that included Mapt and Crhr1 and mimicked the 17q21.31 microdeletion however resulted in increased sociability, similar to the human disease phenotype. The authors speculate that the discrepancy between the human and mouse phenotypes is due to mice requiring more reduction in Kansl1 gene dosage to exhibit increased sociability and that haploinsufficiency in another gene(s) from the larger deletion could contribute to lower KANSL1 levels in heterozygotes.

2.4. Genetic inactivation of synaptic genes in mice exhibiting increased social behavior

Behavioral testing of genetically engineered mice revealed genes whose deletion causes hypersociability. Interestingly, all of these genes are associated, directly or indirectly, with synaptic functions.

Hypersociability of PSD95 null mice.

PSD-95 (postsynaptic density-95, DLG4), a member of the membrane-associated guanylate kinase family of synaptic molecules, is located in the postsynaptic density⁷⁹ and is involved in anchoring synaptic proteins at excitatory synapses, including neuroligins, NMDA receptors, AMPA receptors, and potassium channels⁸⁰. PSD-93 (DLG2) is a similar protein, forming heterotrimers with PSD95⁸¹.

Dlg4^−/− mice were reported to interact with unfamiliar conspecifics more than control wild-type mice in the three chamber social approach and social novelty tests⁸². Null mice were indistinguishable from controls in investigating nonsocial objects, indicating that increased sociability was not due to a general preference for novelty.

While Dlg4^−/− mice, alongside of an increased social interaction phenotype, exhibit impaired motor coordination, increased stress reactivity, and anxiety-related responses⁸², heterozygote Dlg4^+/− mice have only hypersociability. Heterozygote mice of both sexes exhibited increased interest towards an unfamiliar conspecific in dyadic social interactions in neutral testing cages⁸³, and males had enhanced aggression as measured in the resident-intruder paradigm. This suggests that Dlg4, similar to some of the Williams genes, is haploinsufficient. Interestingly, Dlg2 null mice (PSD93) displayed a similar, increased sociability phenotype in dyadic social interaction tests, while heterozygote mice were not different from controls, indicating that Dlg2 is not a haploinsufficient gene⁸³. These data suggest that PSD95 and PSD93 are both involved in controlling social behavior, presumably in a redundant manner. The authors speculate that the increased social phenotype of PSD95 and PSD93 mutant mice could be the consequence of ‘social blindness’, i.e. the inability to adequately process and respond to social signals⁸³. Indeed Dlg4^+/− mice exhibited impaired social recognition while nonsocial cognitive functioning was retained⁸⁴.

It is surprising that genetic deletion of PSD95 interacting proteins, including the transynaptic proteins of neuroligins, neurexins and shanks, results in reduced sociability, and that their loss of function has been implicated in monogenetic forms of autism^85–88. However, PSDs have a much wider function in organizing postsynaptic complexes than the individual transynaptic proteins.

Hypersociability in nitric oxide synthase (NOS) null mice.

NO, a highly diffusible gas with function in neurotransmission, synaptic plasticity, gene expression and neurotoxicity, is produced by neuronal (n) NOS from L-arginine in the brain^89,90. nNOS is expressed in the hippocampus, cortex, striatum, cerebellum, olfactory bulb, and brain stem^91,92. Unlike traditional neurotransmitters, NO levels are not regulated by release but rather via production by the phosphorylation dependent catalytic activity of nNOS⁹³. nNOS is part of a macromolecular signaling complex at excitatory synapses that consists of PSD95 and the NMDA receptor, rendering nNOS activity dependent on Ca²⁺⁹⁴. In turn, NO acts via retrograde signaling to activate soluble guanylyl cyclase in the presynaptic terminal and regulate neurotransmitter release^95,96. At higher concentrations, NO is covalently attached to a cysteine residue to form an S-nitrosothiol in various proteins that include NMDA subunits⁹⁷.

Mice with targeted disruption in nNOS⁹⁸ (partially backcrossed to the C57BL/6 background), was reported to exhibit increased social interaction in their home cage under familiar conditions, but less sociability (i.e. time spent around a stranger mouse) in the three chamber social test, suggesting a different response with a familiar and unfamiliar conspecific⁹⁹. In another report, genetic and pharmacological inhibition of nNOS in mice on mixed genetic background was reported to have reduced social investigation and increased aggression toward an unfamiliar mouse in a modified resident–intruder partition test¹⁰⁰. These data indicate that the social behavior of nNOS null mice is altered, with higher and lower sociability depending on familiarity with a conspecific.

Neuregulin (NRG1) and its receptors ERBB3 and ERBB4 are linked to sociability in mice.

NRG1 is an EGF-like signaling molecule, which, by interacting with transmembrane tyrosine kinase receptors of the ErbB family, is involved in cell-cell communication, both during development and in the adult. In the brain, it regulates neuron migration and progenitor cell proliferation, but also glutamate release, NMDA and GABA receptor subunit expression and synaptic plasticity^101,102. Moy and coworkers reported that Nrg1^+/− mice spent significantly more time in the side containing an unfamiliar stranger mouse, in comparison to the wild-type mice, in a three chamber social test¹⁰³. A similar pattern was found with CNS specific Erbb3 mutants (Erbb3^{tm2Dwt/tm2Dwt}, Nestin-Cre) and a milder phenotype was apparent in CNS specific Erbb4 mutants (Erbb4^lox/-,hGFAP-Cre). NRG1 is a schizophrenia candidate gene^104,105, and several studies reported increased expression of NRG1 mRNA isoforms in frontal cortex and hippocampus of schizophrenia patients^106,107 and in neurons derived from induced pluripotent stem cells of patients¹⁰⁸. Given the presumably higher NRG1-EERB signaling in schizophrenia, and the association of schizophrenia with impaired social functioning, the higher sociability of Nrg1, Erbb3, and Erbb4 deficient mice indicates a bidirectional modulation of social behavior by the NRG1-EERB pathway.

3. Effect of the environment on social behavior

3.1. Hypersociability (indiscriminate friendliness) in adopted children with early adverse care experiences

In addition to genetic factors, the environment has a significant effect on shaping social behavior. After the overthrow of the Ceausescu regime in December 1989, the world discovered that 170,000 children were being raised in Romania’s impoverished institutions. It was noted that adopted children, who have had early adverse care experiences during the first two years of life, approached and interacted with unfamiliar adults in a way expected and typical with familiar individuals¹⁰⁹. The prominent social abnormality of children adopted following institutional rearing has been variably called indiscriminate friendliness¹⁰⁹, disinhibited social behavior¹¹⁰, disinhibited attachment behavior¹¹¹, or disinhibited social engagement disorder (DSM-5 313.89). No sex difference was found in indiscriminate friendliness¹⁰⁹. Beside social disinhibition, children adopted following institutional rearing tend to have difficulty with executive functions such as cognitive flexibility, inhibitory control and working memory. They struggle to regulate their emotions and suffer from anxiety. These children can be exposed to safety risks¹¹², have later difficulties in peer relationships^113,114 and have impaired attention and impulse control^115,116.

A similar clinical phenotype is also apparent in maltreated foster children¹¹⁷. Moreover, disinhibited social engagement disorder can develop in noninstitutionalized children due to parental adjustment problems that include poverty, teen parenting, substance abuse and mental health problems. Children, once removed from abusive or neglectful environments tend to recover physically but can still present the symptoms of disinhibited social engagement disorder¹¹⁸.

Although the overall behavioral phenotype of individuals with prior childhood maltreatment and Williams syndrome are different, they share increased and indiscriminate sociability suggesting a common neurobiological basis of hypersociability. Indeed, similar to Williams syndrome, amygdala activity in previously institutionalized youths is altered. It was reported that, while controls showed an amygdala response that clearly discriminated mother versus stranger stimuli, previously institutionalized individuals exhibited reduced amygdala discrimination between mothers and strangers¹¹⁹. Reduced amygdala differentiation correlated with greater reports of indiscriminate friendliness that in turn correlated with age-at-adoption. Lack of discrimination was the result of atypically high amygdala response to strangers in previously institutionalized youth, whereas responses to mother stimuli were equivalent across groups.

An animal model of early life neglect or early adverse care would be an invaluable tool for understanding the potential mechanisms underlying increased sociability of post-institutionalized children. Although daily separation of pups from their mothers during the early postnatal period has been extensively used to study persistent alterations in emotional and anxiety-related behaviors, less is known about the impact of maternal separation on social behavior. The limited literature suggests that current maternal separation models and even maternal separation combined with early weaning (to mimic early life neglect) may not be able to reproduce the increased sociability phenotype seen in post-institutionalized children^120–122. Development of animal models of early life neglect that recapitulate indiscriminate sociability would help to understand the underlying neurobiological mechanism.

3.3. Parental genotype as an environmental effects in programming hypersociability

Genetic variants in parents can affect the behavior of their offspring, even if the child does not carry the allele. For example, the maternal genotype can influence the in utero environment that in turn, could alter prenatal development and the long-term health of the offspring¹²³. The parental genotype may also alter the postnatal development of the offspring, a phenomenon recently named genetic nurturing¹²⁴. Although no study, to the best of our knowledge, has linked a specific parental genotype (i.e. SNP) to measures of offspring social behavior in human, parental serotonin and dopamine transporter and brain-derived neurotrophic factor (BDNF) polymorphisms have been reported to influence parenting behaviors^125–127 that in turn, could modulate offspring social behavior.

The animal literature regarding parental genotype effects on offspring social behavior is also unexplored, but we reported that heterozygous deletion of the Fmr1 gene (encoding the fragile X mental retardation protein FMRP) in dams programs higher sociability, specifically reduced avoidance and increased approach in both a three-chamber sociability assay⁷¹ and a one-chamber social interaction procedure¹²⁸, in genetically wild-type male FVB offspring¹²⁹. In contrast to this maternal Fmr1 genotype effect, absence of FMRP in male offspring had no significant effect on their social interaction (i.e. no offspring genotype effect)¹²⁹. This was surprising because lack of FMRP in fragile X syndrome is typically associated with a high degree of gaze aversion, interpreted as social anxiety¹³⁰. Nevertheless, similar to our results, Spencer and coworkers found no effect of deleting Fmr1 in C57BL/6 males either on social behavior using a “partitioned” social interaction test apparatus ^131,132. However, more in line with human studies, others reported social anxiety in FMRP deficient mice in the presence of a stranger in various tests^133–136, suggesting that offspring Fmr1 genotype has no or variable effect on social behavior, depending on testing conditions and the genetic background. In summary, FMRP seems to have a dual, genetic and non-genetic role in social behavior; inactivating Fmr1 in mothers increases sociability of their genetically normal male offspring, in contrast to increased social anxiety in fragile X syndrome and in some mouse models of the disease.

Because increased social approach could be elicited by the FMRP-deficient maternal environment both pre- and postnatally, mechanisms that may mediate this maternal effect include maternal bioactive molecules that cross the placenta or are transferred to the offspring via milk, or behavioral interactions between the mother and offspring¹³⁷. The relevance of the maternal FMRP genotype effect, if present in human, is potentially broad because it includes the genetically unaffected children of mothers with full mutation, as well as mothers with the far more prevalent premutation (which is also associated with deficits in FMRP¹³⁸).

Brain-wide imaging of neuronal activity (via the expression of the immediate early gene c-fos) upon social interaction with a stranger showed that, while control male offspring (of WT mothers) exhibit activation in the ventral tegmental area (VTA) and its connected region, the NAc shell, the hypersocial male offspring (of FMRP deficient mothers) exhibit suppression of activity in these regions¹²⁹. This result is consistent with the model that medium spiny neurons in the NAc shell encode, via activation and suppression, aversion and reward, respectively, first proposed by Carlezon and Thomas¹³⁹ and also shown by others¹⁴⁰. Therefore, our interpretation is that programmed hypersociability is associated with a failure to recognize social cues and activate NAc, resulting in a failure to avoid the unfamiliar mouse.

Discussion/Synthesis

Both genetic (Table 1) and environmental effects (early life neglect) contribute to the development of hypersociability. While genetic effects can be directly linked to specific genes and, via gene expression, to neurobiological processes, this association is more difficult to establish with environmental effects. However, early life insults produce persistent changes in DNA methylation and chromatic modifications^141–144, and epigenetic changes, similar to genetic polymorphisms and mutations, can alter gene expression^145–147.

Although the precise functions of sociability-related genetic and epigenetic effects are not understood, they seem to converge on two neuronal functions that regulate social behavior (Fig. 2). The first is the ability of the amygdala to discriminate between fear-inducing and friendly social clues and to produce an appropriate behavioral response. The second is the dopaminergic reward/aversion system that assesses the salience of social situations and initiates approach (social reward) or avoidance (social aversion) behavior.

Figure 2.

Convergence of genes and early adverse care experience on neuronal networks to produce hypersocial behavior.

Hypersociability can be conceptualized as a developmental abnormality leading to a failure in discrimination between familiar and stranger in the amygdala and/or increased reward or reduced aversion in the VTA-NAc pathway. For example, Williams patients were reported to exhibit reduced amygdala response to fearful facial expressions that may explain their social disinhibition ^48–50. In addition, these patients tend to show heightened amygdala reactivity to positive social stimuli that could underlie their high motivation to interact with others⁴⁹. Similarly, previously institutionalized individuals were reported to exhibit atypically high amygdala response to strangers¹¹⁹. Regarding the reward pathway, its possible role in social behavior, specifically in pair bonding, was first suggested by higher OTXR densities (i.e. the receptor for the prosocial OTX) in the NAc of monogamous, compared to nonmonogamous, prairie voles¹⁴⁸. The involvement of VTA-NAc system in social reward was also demonstrated in mice¹⁴⁹, and a more efficient signaling variants of OTXR has been linked to prosocial behavior^17,18. However, the VTA-NAc pathway also codes for aversion^139,140 that seems to play a role in the hypersocial behavior of male offspring of FMRP deficient mothers¹²⁹.

To better understand the neurobiological basis of hypersociability, it will be important to link hypersociability-related genes and environment to specific neuronal functions. Because of the diversity of conditions that lead to hypersociability, multiple neuronal mechanisms may be involved. Nevertheless, abnormalities at the neuronal level can facilitate the understanding of neuronal network abnormalities that ultimately may explain how hypersocial behavior is generated in these circuits.