Introduction
Adolescents increasingly become better in perspective taking (Tousignant et al.
2017), empathic concern (van der Graaff et al.
2014) and show more prosocial behavior (van der Graaff et al.
2018). Environmental stress, like abusive parenting, is supposed to influence this social development (Sandi and Haller
2015). Higher levels of environmental stress are related to less empathy and increased aggression and antisocial behavior via psychobiological processes (Susman
2006). One of the biological factors that could play a role in the relation between environmental stress and the development of social behavior is the mineralocorticoid receptor (MR) (De Kloetet al.
2005) as it is closely involved in the appraisal of a stressful situation and it is one of the two receptors for the stress hormone cortisol (ter Heegde et al.
2015). Experimental studies have shown that pharmacological stimulation of MR resulted in enhanced empathic concern in clinically depressed adult patients (Wingenfeld et al.
2014) and enhanced emotion processing in a healthy sample (Schultebraucks et al.
2016). Therefore, the aim of this study was to investigate whether genetic variation in MR would moderate the effects of stressful parenting on the social development (i.e., prosocial behavior, empathic concern and perspective taking development) from adolescence to young adulthood in a community sample.
During adolescence, social behavior changes (Forbes and Dahl
2010) due to development in social information processing (Nelson et al.
2005). While certain social cognitive processes, like prosocial reasoning (Eisenberg et al.
2015) and social knowledge (Tousignant et al.
2017), are already well developed during childhood, more complex aspects of social cognition increase over the course of adolescence. Mainly adolescents’ perspective taking undergoes quick development (van der Graaff et al.
2014), which enables adolescents to attribute mental states such as beliefs, desires and intentions to others (Blakemore and Choudhury
2006). Similarly, empathic concern and emotion processing show strong development during adolescence (Tousignant et al.
2017). A meta-analysis on prosocial behavior has shown an increase only in early adolescence until the age of 16 and no development in late adolescence (Fabes et al.
1999). This is confirmed in a recent longitudinal adolescent study that even found a decreasing trend in prosocial behavior in late adolescence (van der Graaff et al.
2018), although other studies showed decreasing trends over the whole course of adolescence (e.g., Luengo Kanacri et al.
2013). Interestingly, social development seems to differ between boys and girls. van der Graaff et al. (
2014) found that girls already reached adult-levels of empathic concern in the beginning of adolescence, while boys developed empathic concern over the whole course of adolescence. Also, perspective taking and prosocial behavior seems to develop differently in boys and girls, mainly in early and mid-adolescence, with growth starting earlier for girls than for boys (van der Graaff et al.
2018).
Adolescence is a period in which the long-lasting effects of earlier exposure to physical or emotional abuse and neglect become evident (Lupien et al.
2009). Stress models suggest that early life stress, like stressful parenting, results in decreased levels of social motivation, reduced social behaviors, increased aggressiveness, and stronger development of antisocial characteristics (Sandi and Haller
2015). A form of stressful parenting that could be related to social development in adolescence is childhood trauma like sexual, physical, or emotional abuse or neglect. Indeed, stress in humans has been related to less empathy and increased aggression and antisocial behavior (Susman
2006).
Also, psychological control is a form of abusive parenting (Del Giudice et al.
2011) that could hinder the social development of adolescents. Psychological control involves attempts that intrude or manipulate the thinking processes, self-expression, and emotions of the child (Barber
1996). According to the self-determination theory (Deci and Ryan
2000), such a controlling environment with conditional love of parents lead to negative expectations of interpersonal relationships, which elicit maladaptive social behaviors and impairments in social development (Soenens and Vansteenkiste
2010). Mainly during adolescence, psychological control can be intrusive, as adolescent have an increasing need for autonomy (Lansford et al.
2014). Therefore, parental psychological control during adolescence can be considered as stressful parenting, which could affect the social development of adolescents. Indeed, parental psychological control has been related to lower levels of social behavior and higher levels of relational aggression in adolescents (Loukas et al.
2005) and emerging adults (Clark et al.
2015).
Besides differences in stressful parenting, adolescents also differ in their general stress reactivity (Ellis et al.
2013). Resiliency to the negative consequences of stress (McEwen et al.
2015) may be explained by genetic variation. Genetic variations important in this regard are the MR-2G/C (rs2070951) and MRI180V (rs5522), which are single nucleotide polymorphisms (SNPs) that both affect in vitro transactivation by altering Mineralocorticoid Receptor (MR) expression or functionality (ter Heegde et al.
2015). Thereby, these SNPs affect the stress response (van Leeuwen et al.
2011). Upon experiencing stress, the Hypothalamic-Pituitary-Adrenal axis (HPA-axis) becomes active and releases several hormones to deal with the stressor, including cortisol (Del Giudice et al.
2011). Cortisol binds to MR in the brain, thereby providing negative feedback on the HPA-axis which subsequently returns to the prestress state after the stressor is gone (Joëls and de Kloet
2017).
Previously, it has been found that MR haplotypes (based on two SNPs rs5522 and rs2070951) differentially affect MR activity and expression and thereby the functionality, as they affect the maximal transactivation and protein expression (ter Heegde et al.
2015). The haplotype coined “CA” is associated with increased MR expression and activity, leading to a more reactive HPA-axis with lower basal non-stress levels, and could therefore be advantageous for adolescents’ social behavior (van Leeuwen et al.
2011). Although earlier studies found no relation between MR functioning and perspective taking (Wingenfeld et al.
2014), or even diminished perspective taking (Wingenfeld et al.
2016), higher MR functioning after pharmacological stimulation resulted in enhanced empathic concern (Wingenfeld et al.
2014) in clinically depressed adults. Moreover, in healthy subjects increased emotion processing has been reported after pharmacological stimulation (Schultebraucks et al.
2016), and diminished emotion processing after pharmacological inhibition of MR functioning (Young et al.
2016).
This expected positive effect of the MR CA haplotype could even be stronger for adolescents who experienced stressful parenting. Vinkers et al. (
2015) found stronger protective effects of the MR CA haplotype on depression symptoms for adults with higher levels of childhood trauma in a population based sample. As MRs are involved in every stress response, the MR CA haplotype can result in effective termination of each stressor (ter Heegde et al.
2015). In adolescents who experience high levels of environmental stress, the MR CA haplotype can play a large role in the stress regulation of each of these stressful experiences, and thereby in the possible negative consequences of environmental stress. In contrast, in adolescents who experience only incidental daily stressors, the consequences of effective stress regulation might be smaller.
In addition, a growing body of evidence from experimental studies has shown that the effects of MR haplotypes are sex-specific, with the CA haplotype being protective in females but not in males (ter Heegde et al.
2015). The sex-specific MR haplotype effects might be explained by the female hormones progesterone and estrogen that positively affect MR functioning (Carey et al.
1995). These findings suggest that mainly female adolescents with a MR CA haplotype might be resilient to the negative consequences of stress for their social development.
Discussion
It recently has become clear that the mineralocorticoid receptor (MR) plays a role in stress responsiveness by balancing glucocorticoid levels in the brain (Joëls, Karst, DeRijk, and de Kloet
2008), resulting in changes in social behavior (van Leeuwen et al.
2011). However, a couple of limitations in this body of literature constrain its implications. Studies so far have addressed the role of MR in socio-cognitive behaviors in adult samples, clinical samples and have used experimental designs to manipulate the level of MR expression (e.g., Wingenfeld et al.
2014). However, effects of MR on social behavior might emerge in an age period when this development is strong, like adolescence, and therefore longitudinal studies are needed to examine when this association emerges. Also, for preventive purposes, consequences of natural variation in the MR gene in a community sample are important as well. Moreover, both in the MR field (ter Heegde et al.
2015) and in the social development literature (van der Graaff et al.
2018) sex-differences are pronounced, but both theoretically and empirically, it is unclear how these sex-specific factors can interact. Therefore, this longitudinal community study examined how the development of prosocial behavior, empathic concern, and perspective taking during adolescence and young adulthood was affected by stressful parenting and common genetic variation in the MR gene in boys and girls. In individuals who experienced higher levels of parental psychological control or childhood trauma, the number of CA haplotypes was positively associated with social development. This protective effect of the MR CA haplotype emerged over the course of adolescence and young adulthood for empathic concern and perspective taking. Although social development was different for boys and girls, with a later increase in prosocial behavior, empathic concern, and perspective taking for boys, there was no difference in the role of the MR CA haplotype between boys and girls in this developmental pattern.
Adolescents who likely experienced high levels of stressful parenting, as indicated by parental psychological control or childhood trauma, showed lower levels of, or a slower growth in, prosocial behavior, empathic concern, and perspective taking skills. Childhood trauma had more significant relations with the social development than parental psychological control, possibly as childhood trauma have been more extreme stressors as compared to parental psychological control. It also might be explained by the stressful event load that accumulates during lifetime (Kemner et al.
2015). Childhood traumas might have happened at a younger age and therefore had more cumulative effect, as the Childhood Trauma Questionnaire is a retrospective measure about adolescents’ whole youth until the age of 18, while parental psychological control was measured parallel to social development during adolescence from age 13. We were not able to establish time order between stressful parenting and changes in social behavior and can only conclude that reported trauma or parental psychological control is related to change in social behavior. Still, both forms of stressful parenting might have the same kind of consequences. When parents are psychologically controlling, they display disappointment and children feel pressured and guilty that they did not comply with the parent’s requests or feels anxious about losing the parent’s approval (Soenens and Vansteenkiste
2010). Also, childhood abuse is characterized by guilt and self-blame of the child (Valle and Silovsky
2002). These processes may result in a stronger focus of adolescents on the consequences of their behavior for themselves, by complying to their parents, as compared the needs of others (Hoffman
1983). Future studies might provide more insight in this process by asking participants to report the level stress associated with childhood trauma or parental psychological control, and asking the exact timing and duration of these stressful experiences.
There were no main effects of MR CA haplotype for social development, but there were interaction effects between MR and stressful parenting. Overall, the MR CA haplotype had positive effects on social development of adolescents who reported high levels of parental psychological control or childhood trauma, and even seemed to have negative effects for adolescents with low levels of stressful parenting, although these effects did not reach significance in all models. This suggests that adolescents do not differ in their general susceptibility to stress based on their MR haplotype, but that the MR CA haplotype has stronger positive effects for higher levels of stress. This pattern is in accordance with the study of Vinkers et al. (
2015) that found stronger MR CA haplotype effects for higher levels of childhood trauma, resulting in less depression symptoms. The opposite pattern for adolescents with low levels of stressful parenting in combination with MR CA haplotype, resulting in relatively less positive social development, fits within the differential susceptibility theory (Pluess
2015). This theory supposes that sensitive individuals are vulnerable for developing certain outcomes, but whether the environment is adverse or supportive results in vulnerability or enhancement. In light of this theory, fewer MR CA haplotypes can be seen as an indicator of sensitivity, resulting in less positive social development for adolescents with high levels of stressful parenting and better social development for adolescents with supportive parents. Having more MR CA haplotypes is an indicator of resiliency, making adolescents with more MR CA haplotypes equally social in both stress and non-stress environments. The good social development for adolescents with high stressful parenting and more MR CA haplotypes even seemed to outreach the social development of adolescents with low levels of stressful parenting and more MR CA haplotypes, suggesting that people with more MR CA haplotypes thrive better in a stressful environment. Important to note is that the role of the MR CA haplotype in interaction with their level of stressful parenting was less pronounced for empathic concern, and although patterns for both parental psychological control and childhood trauma were comparable, the results were not always significant for both forms of abusive parenting. Until now, too little is known about the underlying biological mechanisms of MR in positive and negative environments (see ter Horst et al.
2014). Further research is needed into the interaction between environment and MR expression to better understand the consequences of the interaction between high and low levels of stressful parenting and genetic variation in the MR gene for adolescents’ prosocial behavior, empathic concern and perspective taking.
The positive role of the MR CA haplotype (in interaction with stress) for indicators of social behavior corresponds with earlier studies in which MR was pharmacologically stimulated and resulted in increased empathic concern (Wingenfeld et al.
2014) and emotion processing (Schultebraucks et al.
2016). But earlier studies found no relation between MR CA haplotypes and perspective taking (Wingenfeld et al.
2014), or even diminished perspective taking (Wingenfeld et al.
2016). In these studies, perspective taking abilities were measured directly after the pharmacological manipulation. Until date it is unclear how direct consequences of MR genetic variation differ from long term consequences, and acute stress from chronic stress (Vogel et al.
2016). Possibly, the long-term behavioral consequences of stress and MR genetic variation are larger as compared to the short-term behavioral consequences, as stress affects the development of several brain areas, like the prefrontal cortex (McEwen et al.
2016). The prefrontal cortex is relevant in social behavior as it provides top-down control (Arnsten et al.
2015). Thereby, for adolescents who experienced stress, social behavior could be affected via a different prefrontal cortex development. To better understand the role of MR in the relation between stress and social development, more research is needed into the role of MR in prefrontal cortex functioning and into the development of the prefrontal cortex under stress.
The MR CA haplotype mainly played a role (in interaction with stress) over the course of adolescence for the development of empathic concern and perspective taking, while for prosocial behavior the effects of the MR CA haplotype were already evident at the beginning of adolescence. The period at which MR CA haplotype affected the indicators of social behaviors most, corresponds to those points during adolescence at which these behaviors showed the main development. During these so called ‘developmental switch points’ (Ellis et al.
2013), there is increased susceptibility to genes and environmental influences (Del Giudice et al.
2011).
The hypothesis that mainly women with more MR CA haplotypes were relatively more social was not confirmed. This is in contrast to rodent studies that found female-specific MR effects in social discrimination (ter Horst et al.
2014) and emotional behaviors (ter Horst et al.
2012) and a human female-specific protective effect of MR CA haplotype for the development of depression (Klok et al.
2011) even as moderator of childhood stress (Vinkers et al.
2015). These effects are generally explained by the female hormones, estrogen and progesterone that affect MR functioning, although the exact underlying mechanism is unclear (Carey et al.
1995). Evidence for a sex-specific MR effect did not come from adolescent studies. Adolescent girls may respond differently to stress as compared to woman, as during adolescence, the increasing levels of female hormones levels influence neurotransmitter systems which affect the maturing HPA axis (Naninck et al.
2011). More research is needed about the interaction between MR and female hormones during adolescence versus adulthood to better understand the sex-specific role of MR for different behaviors.
Looking at the development of the different indicators of social behavior, whereas perspective taking showed a relatively gradual increase over the whole course of adolescence and young adulthood, prosocial behavior mainly increased during early adolescence, and empathic concern increased during late adolescence and young adulthood. These findings are consistent with earlier studies on social development. Earlier longitudinal studies of prosocial behavior have also found an increase in early adolescence (Fabes et al.
1999), and no development in late adolescence (Flynn et al.
2015) or young adulthood (Eisenberg et al.
2005). Studies on empathic concern (Brouns et al.
2013) and perspective taking mainly showed an increase during late adolescence (Eisenberg et al.
2005). These differences in developmental pattern for prosocial behavior versus empathic concern and perspective taking are likely due to prefrontal cortex functioning. During adolescence, the prefrontal cortex rapidly develops (Blakemore
2008), which is supportive for taking one’s own and others’ perspective, and regulation of behavior, thought and emotion (Arnsten et al.
2015). Perspective taking and empathic concern rely more on these skills as compared to prosocial behavior (Shamay-Tsoory et al.
2009), which might explain why perspective taking and empathic concern develop until later as compared to prosocial behavior.
In line with other studies, sex differences were found in social development. Boys showed lower levels of social skills as compared to girls, and a later increase in all social skills as compared to girls. Other studies have also found higher levels of social skills in girls (Miklikowska et al.
2011) and an earlier development of empathic concern for girls during adolescence (Carlo et al.
2015). But developmental differences in prosocial behavior between boys and girls have rarely been studied and there are also studies showing a more rapid decline for the prosocial behavior of boys (Carlo et al.
2007). The current pattern of findings, with an earlier growth for girls as compared to boys, has already been reported by van der Graaff et al. (
2018), conclusions whom were derived from the first 6 waves of the current data set. A process that may account for these sex differences is gender role expectations. Due to gender role identification, which tends to be strongest during early adolescence (Eagly
2009), girls show higher levels of social behaviors because they are stimulated to show nurturance and care.
This study had some limitations. First, the sole focus on hypothesis-driven MR as a CA haplotype, instead of MR expression, MR functioning, or even the broader stress system, gives only a limited view of how the stress system performs in relation to social behavior. More research that examines the whole stress-axis is needed to better understand the biological mechanism behind the role of MR for social behavior. Second, we included genetic influences and parental influences on social development, although parenting might be confounded with heritability of the MR genes (i.e. passive gene-environment correlation). This might suggest that parents who have better stress regulation show relatively less stressful parenting towards their child, and at the same time have children with a MR haplotype that makes them better able to regulate the stressful parenting. However, in our study we did not find a relation between MR haplotype and stressful parenting, which makes it unlikely that heritability of MR-genes would have confounded our findings. Moreover, given the hypothesis driven set-up of this study, there was no independent sample taken into account to replicate the findings, while this is currently standard in gene-environment studies. It is of note that one of the SNPs, rs5522, was not in HWE and that we cannot exclude possible genotyping errors. Importantly, however, a previous study showed that approximately 10% of genotype-phenotype associations deviated from HWE (Trikalinos et al.
2006). Another limitation is the use of questionnaires to let adolescents report on indicators of social behavior, with possible differences between their actual social behaviors and the reports of their social behaviors. It would have been better if experimental tasks were added to measure their actual prosocial behavior and cognitive and empathic concern. Moreover, the questionnaire on childhood trauma was retrospective, which may limit the validity of the reports as it may have been subject to recall bias. Still the questionnaire resulted in reliable reports in multiple studies, amongst others in a community based sample from 18-65 years, with even reliable reports in the age group of 45-65 years (Scher et al.
2001). Also, a composite score for psychological control was calculated over fathers, mothers and seven waves, with possibly interesting information being lost. Finally, a limitation is the use of six different growth models, resulting in multiple testing and possibly false positives. On the other hand, the effect sizes are quite large, and by estimating separate models the results can be better related to previous studies on these different aspects of social functioning. Further research into the consequences of stressful parenting on social behavior, and the biological underlying mechanism, with larger sample sizes, is needed to better understand the exact role of MR for social development in highly and lowly stressed adolescents.
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.