Molecular and Genetic Bases of Mammalian Maternal Behavior

Abstract

New mammalian mothers undergo an increase in their maternal responsiveness with the birth of their infants. Associated with changes in responsiveness are how attracted mothers are to infant cues, mothers’ affective state, and their cognitive and executive function. In comparison to nonmothers, new mothers are more attracted to infant odors and are more easily alerted to their vocalizations; they undergo a reduction in withdrawal behaviors and anxiety, but increased lability. Their maternal sensitivity (human or licking intensity, rat) is associated with higher levels of attention and working memory. Maternal responsiveness and these associated behaviors are associated with large shifts in maternal hormones across parturition. Changes in expression of neuropeptides and neurotransmitters are affected by mothers’ prior experiences, including their very early experiences in their families of origin. The present review describes the regulation of mothering and associated behaviors by the neurotransmitters, oxytocin, dopamine, and serotonin, in a rat model and in humans. Emphasis is then given to studies that focus on the role of genes and what we know about their expression in the functioning of these 3 neurochemical systems in new mothers. Studies of early experience, genetics, and human mothering show gene-by-environment interplays (interactions) for a number of DNA single-nucleotide polymorphism within both the oxytocin and serotonin systems, where associations between mothers’ early experiences and mothering/affect depend on mothers’ genotype. Studies also show associations between different dopamine genes and many aspects of both mothering and maternal affect. Where known, we also discuss evidence that the relation between early experience and mothering is often an indirect one, mediated through an effect of experience on mothers’ affect or executive function. In many cases, mothers’ genetic profile moderates these relations. Finally, preliminary evidence suggests a role of epigenetic mechanisms in these processes.

Introduction

This brief review explores the environmental and biological bases of early mothering. In the spirit of the goals of the conference, “The Environment Writes on the Body,” the present review emphasizes genetic influences on mothering. However, to understand the role of genetics in mothering, we are helped by knowing some of the physiological and behavioral factors that contribute to mothering in both nonhuman animals and/or humans. These include the hormones of parturition, the hormones of lactation, the known and relevant brain and neurotransmitter systems, and the psychological processes that are recruited when a woman becomes a mother and engages with her offspring. By looking at all these physiologic and psychological factors, we can make some educated guesses as to which candidate genes and pathways could reasonably be a focus for the study of the genetics of mothering. Our approach in this brief review is to explore the various factors that affect mothers’ motivation to care for young when they are born and during the postpartum and the role of genes in these processes. This review is selective, not exhaustive, giving undue emphasis to our own work on this topic, focusing primarily on 3 neurochemical systems: dopamine (DA), oxytocin (OT), and serotonin (5HT). ^{1 –5} However, we direct the reader to other more up-to-date studies and reviews of the epigenetics of mothering that describe the very exciting work that is presently ongoing in this field. ^{6 –8} Portions of this brief review are based on other articles and reviews we have published and to which we refer throughout. ^{1 –5,9 –11}

Levels of Analysis

Parental Psychology and Probable Genetic Foundations

The levels of analysis at which mothering has and continues to be studied begin historically with investigation of (1) nature and psychological characteristics of postpartum onset of maternal behavior and its continuation; (2) the endocrinology and neuroendocrinology of pre- and postpartum mothers that contribute to mothering and associated behaviors; (3) the autonomic/visceral and subcortical mechanisms of co (mother–infant) -conditioning based on an early Pavlovian framework ¹² ; and (4) underlying genetic and epigenetic mechanisms contributing to each of the psychoneurobiological mechanisms. We argue that the individual’s genetic endowment, as embodied in 23 pairs of chromosomes, does not determine the individual’s biological characteristics at the system level in the absence of interaction with the environment. Or, stated the other way around, the expression of the biological systems and capabilities of the individual depends on an expression of the genome in relation to the demands and opportunities presented by the physical and social environments in which the individual exists or has existed in early life.

Mother–Infant Interactions

Rat maternal behavior

Animals (including humans) exhibit natural variations in the types and quantities of species-specific maternal care they exhibit. One of the most widely studied animal models of parenting is the rat. Rats exhibit stereotyped maternal behaviors: They nurse, lick, and groom their pups, build a nest, and retrieve the pups back to the nest. The mother and the nest provide nutrition, warmth, and protection. They also provide social and other stimuli that affect pups’ neural and endocrine development and later behaviour. ⁴ Moreover, with postpartum exposure, the young learn to prefer their mother’s odor that guides subsequent social interactions and even their responses to their own offspring later on. ^{13 –16} Maternal licking has a particularly pronounced effect on offspring development and long-term effects on the quality of mothering pups adopt toward their own young. The most effective period for transmission of licking effects seems to be in the first postpartum week when the rat brain is still rapidly developing. ^{17 –19} Licking and grooming (LG) during early life thus have a nongenetic influence on the next generations and are therefore crucial for optimal development. The absence or disruption of LG, such as during maternal separation or deprivation, has documented developmental consequences as well. ^1,3

Human maternal behavior

The obvious similarities among mammalian species include nursing and a posture designed to enhance the neonates’ access to the teat, some form of communication system between mother and offspring to indicate “needs” of both, a way of transporting offspring, especially if they are altricial or immature at birth, and some form of maternal “protective” defense of offspring. Most mothers also keep their offspring clean by grooming and provide a home base or “nest site” either in the environment or on their bodies where the young can sleep. In addition to performing these functions, human mothers normally develop feelings of nurturance and warmth (or “love”) toward the baby, anxiety in response to distress or unexpected separations from the baby, and grief with his or her loss or death. ^{4,20 –23}

The differences among species in the typography, timing, and duration of the behaviors, their developmental trajectory, and the range of proximal causal mechanisms are vast. The differences emerge as a function of the developmental maturity of the young. ²⁴ In most mammalian species, the young are altricial, often born with their eyes and ears closed and with immature nervous systems; these young require extensive care and are very dependent on the mother for early survival. Other species are much more mature, or precocial, at birth and are more independent early on (as with ungulates where the young stand within minutes of birth and ambulate behind the mother within days). ²⁴

Nonhuman primates vary across species on the precocial–altricial dimension. Neonates must be able to cling to their mother using hands and feet postpartum. Most species have vision, hearing, olfactory, and somatosensory proprioceptive competence within the first month after birth. Beyond this commonality, offspring achieve independent mobility and are weaned over substantially different durations and stages of maturation. ^{22,25 –27} Humans are mostly altricial: Although they can see and hear at birth, they require a long period of care before they can fend for themselves (some would say this takes 2 decades or more!!!). More intriguing—or less well understood—than cross-species differences are individual differences within a species.

The behavior of new human mothers toward their offspring shows both marked similarities and considerable differences within cultures, across cultures, and certainly in comparison to other species. Among the modal similarities are included nursing, ²⁸ singing, ^29,30 and contingent responding to infant cues and distress. ^22,31,32 However, in the absence of explicit practice, first-time mothers exhibit a range of different responses to their infants: Some look at them directly while others gaze avert; ³³ some keep their babies unclothed and stroke their bodies; others swaddle them instead. ²² Some talk or sing to their babies; others do not. ²⁹ Some sleep with their babies while others keep their babies in a cot next to them or in their own rooms. ³⁴ Babies are also transported in different ways—some on the front of the body, others on the back; some on cradle boards and others still, in a vehicle. ^35,36

More subtly, within a culture mothers show large variations in the postpartum development of nurturant feelings, from minutes to months. ^{37 –40} Once “attached” or emotionally committed, mothers vary in the intensity with which they exhibit different caregiving behaviors. More extremely, some are motivated to provide warmth, shelter, and food to the infant while others neglect or even abuse their infants. ^{37 –40}

In many of the studies that explore genetic associations to mothering, mothers’ sensitivity while interacting with infants constitutes an important phenotype that shows large variation in most populations. Maternal sensitivity is assessed (measured) by observation and recording of mother–infant interactions, and coding of mothers’ behavior in response to infant behaviors. In North America and other industrialized cultures, mothers described as showing high sensitivity (using the Ainsworth scales and maternal Behavior Q Sort) respond promptly, appropriately, and contingently to individual infant behaviors and cues and express “warmth” toward the infant. ^32,41

Postpartum Depression and Maternal Behavior: Psychological Effects on Offspring

Parturition often brings about changes in maternal mood: Mothers become more labile, experiencing periods of elation alternating with tearfulness. ⁴² Many mothers experience a period of “postpartum blues,” which is also characterized by emotional lability, but tends to involve more dysphoria, anxiety, and depression that is time limited and usually remits by the end of the first month. ⁴³ Other mothers experience a real postpartum depression, which is similar in some—but not all respects—to depression occurring during other periods of life. Postpartum depression is far more prevalent than is appreciated, with estimates of incidence rates of between 10% and 13% in developed countries, including Canada. Depression can start during the first 6 postpartum months and can persist on and off for years. ⁴³ However, peri- and postpartum depression most often remit within the first postpartum year, although one major predictor of depression during this period is depression outside the postpartum period in which case the depression may be present at other times and deepen during either pregnancy and/or the postpartum. ⁴⁴ Depressed mothers, like depressed nonmothers, often experience extreme fatigue, dysphoric mood, anxiety, tearfulness, and feelings of low self-esteem. ⁴⁴

Unlike depression outside the peripartum period, maternal depression can have profound consequences for mother–child interactions and child development. Maternal depression often impairs the mothers’ attention, sensitivity, and “bonding” with the baby. ^{44 –49} These changes may also alter maternal perceptions of infant/child cues and her attention to them, which can, in turn, influence the salience of the infant/child to the mother and hence her motivation and competence. ^{45 –49} Although these studies focus on the first postpartum year, there is a substantial literature that suggests that outside the postpartum period, depressed individuals find various social stimuli to be less rewarding and that there is reduced neural activation in the striatum among depressed patients. ⁵⁰ Indeed, depression associates with states of anhedonia and impairments in executive functions, both of which may serve as the basis for later parenting difficulties.

Cognition and maternal behavior

Mother rats that naturally show high levels of pup licking perform better on a variety of attention tasks, including attention set-shifting and prepulse inhibition (PPI) of the acoustic startle response. ⁵¹ Mothers reared apart from their own mothers, show both reduced licking and correlated decrements in attention set-shifting, PPI, and enhanced action impulsivity. ^{51 –53}

Numan and Insel ⁵⁴ argue that in primates, as compared to rodents, the balance between size and role of the medial preoptic area (MPOA) and the neocortex in regulating maternal behavior has shifted in favor of the latter, reflecting the greater importance of the prefrontal function in human parenting. This expanded role of the neocortex is consistent with the extended period of parental care in humans, as well as with the demands of parenting children at very different developmental stages. ⁵⁴ The DA neurochemical input into the prefrontal cortex (PFC) is closely associated with attentional and executive function systems critical for quality parenting. ^{55 –57}

Dyadic synchrony between a mother and her infant, which often develops over the first postpartum months, consists of crucial individual components, including positive affect, physiological synchrony, a shared focus of attention, and temporal coordination and contingency; all controlled primarily by the caregiver. Our studies with human mothers confirm the importance of cognitive flexibility, mother–child synchrony, and emotional well-being for maternal behaviour. ^{45,58 –60}

Mothers with disturbed attachments to their own caregivers (disorganized/unresolved), and to their own infants (irrational fear of loss of the infant), and mothers of disorganized infants (infants whose attachment strategies collapse under stress), show attentional difficulties when assessed with emotional Stroop tasks. ⁶¹ Mothers with slow Stroop reactivity are slower to respond to infant signals. ⁶⁰ Furthermore, we found that mothers with fewer errors on extradimensional shift and spatial working memory tasks at 2 to 6 months postpartum are more sensitive in their interactions with their infants and show more contingent responding to infant cues. ⁵⁶ Finally, a recent randomized controlled intervention study Family Nurture Intervention shows that additional guided support to promote contingent interactions between mothers and newborns, especially in the neonatal intensive care unit, enhances mothers’ self-confidence, positive touch, and interactional sensitivity and reduces the probability of PPD in these mothers; it also (and thereby?) has positive effects on the infant’s own behaviors as well as its growth and development. ^62,63

Maternal experience, early adversity, and maternal behavior

Parturitional hormones in rats are critical during the early postpartum period for attraction to pups. However, throughout lactation and the accompanying interactions with young, associative learning processes render the pups highly rewarding, a characteristic sustained following weaning as mothers remain responsive to young for many months. ²⁴ In the absence of these usual interactions with mother and siblings, as in situations of early social isolation, rat pups grow up to show reduced licking of their own young and reduced interest in them. ^18,64,65 In monkeys, maternally deprived mothers are more likely to reject infant suckling attempts and engage in more physical aggression toward their infants, are less likely to “retrieve” a crying infant. ^66,67

Human mothers who experienced early adversity may also show suboptimal mothering; they are more likely to be abusive and neglectful, and/or less sensitive and responsive to their babies. ⁶⁸ In the absence of a subsequent positive social support during development, negative parenting experiences tend to be transmitted across generations. ^18,69,70 However, along the way, supportive positive experiences including by family and/or friend support networks, or a supportive spouse may act as protective factors that ameliorate the effect of negative early experiences in the family of origin. ^{27,71 –73}

Hence, as we have seen, optimal mothering requires efficient emotion regulation, cognition, learning/experience/, and executive function. Human mothers must be sensitive to infant cues and respond appropriately and in synchrony with the needs of the infant. Infant cues must be attractive and salient for the mother, recruiting attentional systems. Mothers must be emotionally prepared and positively motivated to engage socially with the infant. They must selectively attend to the offspring in the context of competing stimuli and be consistent in their responsiveness. Finally, and not surprisingly, optimal mothering and its psychological profile are affected by mothers’ early and other life experiences and their effects can be transmitted to the next generation of mothers. ^{56,74 –76}

Looking to the animal work and to human studies, and by understanding the phenomenology of mothering and its neurochemistry and neurobiology, we are in a good position to choose candidate genes. We look both to the animal work on the role of gene expression and gene products on maternal and associated behaviors, especially in rodents but extending up to nonhuman primates and humans. Within the past 2 decades, an extensive literature on gene polymorphisms related to neurobiological function in animals and to maternal psychological function in humans has emerged. In general, genetic studies exploring these functions focus on their dysfunction and look for gene polymorphisms that relate to conditions such as postpartum depression, impulsivity, memory disorders, and so on. Hence, in what follows we discuss not only what is known about the genetics of mothering per se but also the genetics of behavioral maladaptation and disorders in psychobiological domains that contribute to mothering. Before addressing maternal behavior at the genetic level, the biology and neurobiology of mothering will briefly be reviewed.

Neurobiological Effects of Mothering Experience on the Mother: Hormones and Gene Expression

Genes associated with the hormones that undergo change during later pregnancy and at parturition are good candidates for investigation. These include genes associated with the ovarian or placental reproductive hormones (estrogen and progesterone), the peptide prolactin, and the neuropeptides OT and arginine vasopressin (AVP). ⁵⁴ Hormonal changes in the final trimester of pregnancy, including elevations in estradiol and a decline in progesterone, form the basis for the onset of maternal behavior following parturition in the rat. ^{77 –80} These changes have been implicated in the regulation of maternal behavior in many mammalian species. ^2,54

Three neurochemical systems and gene pathways have been associated with mothering and associated cognition, attention, and affect. These include those associated with OT, DA, and 5HT. ⁴ While no single neural system operates in isolation to regulate a complex phenotype like maternal and associated behavior, all 3 systems are likely to be implicated in different aspects of maternal behavior. Genes for the hormones, such as estrogen, prolactin, and indeed cortisol, are relevant to early licking effects in rats and probably play a role in human maternal behavior as well. They have yet to be studied—or have been studied minimally—in this context in humans and are not discussed further in the present review (except to note that an orthologue of a gene first identified in Drosophola melanogaster, the Foraging Gene, has recently been associated with maternal behavior in humans ⁸¹ ). Below we review findings relevant to the 3 neurochemical systems (OT, DA, and 5HT), and their genetic precursors, to mothering and associated behaviors.

Oxytocin

Oxytocin physiology and mothering

Oxytocin is a neuropeptide hormone involved in parturition and milk let-down across all mammals. ⁸² Estrogen enhances OT receptor (OTR) binding in the MPOA, a brain region critical to maternal behavior across mammals. ^83,84 Central administration of OT stimulates maternal behavior, while OTR blockers reduce maternal behavior in virgin rats. ^85,86 Moreover, individual differences in specific forms of maternal behavior, such as pup LG, correlate with OTR expression in the MPOA, and intracerebral infusion of an OTR antagonist eliminates individual differences in pup LG among lactating rats. ⁸⁷

The role of OT is more ambiguous in women. A number of studies show that OT is closely related to affiliative relationships, being a neuropeptide that “acts to enhance the perceptual salience of social and emotional stimuli in general.” ^88,89 However, whether OT has effects (or associations) is not always straightforward, varying as a function of the context in which their effects (or associations) are assessed as well as the person’s gender, maternal status, personality, attachment style, general stress reactivity, and early life experiences to name a few. ^{88,90 –93} Given the importance of these variables, our analysis of the role of any of the hormones, neuropeptides, or neurotransmitters in human mothering needs to keep in mind the potential role of other factors that could influence whether or not these brain chemicals have effects on maternal behavior and affect. In prospective mothers, OT levels in the first trimester predict postpartum characteristics of the expression of maternal behavior and coordination with infant state. ⁹⁴ The OT levels in mothers have also been associated with affectionate, rather than stimulatory parenting. ⁹⁵ Interestingly, OT is anxiolytic, reducing the impact of stress on emotional states, although often effects of OT on depressive symptomatology and maternal sensitivity depend on mothers’ psychosocial stress ⁹³ ; among highly stressed individuals (but not among women with low levels of psychosocial stress), plasma OT levels are associated positively with maternal sensitivity and inversely with depression. ⁹³ An intriguing analysis of OT nasal spray effects in humans suggests that OT may promote positive affective states within social interactions and the recognition of emotional states in others. ^96,97

Within the maternal context, and when it has positive effects, OT seems to be mostly associated with the onset of maternal behavior, “maternal motivation,” and warmth. It is likely to be of greatest direct relevance in the early stages of maternal responding when close contact behaviors predominate, and nursing may still be occurring. ⁹⁸ The OT may also be related to mothers’ depression after birth. For instance, low OT levels during pregnancy are associated with the development of postpartum depression. ⁹⁸ On the other hand, as Zelkowitz et al ⁹³ point out, OT levels during pregnancy are only related to maternal symptoms of depression at 2 months postpartum among mothers experiencing high levels of psychosocial stress. Further, in nonmothers, plasma OT levels are higher in patients with major depression than in healthy controls. ⁹⁹ The relationship between plasma OT measures and mood states in women is not straightforward and depends on multiple moderating influences.

In terms of OT and early-life experience, OT levels in cerebrospinal fluid (CSF) are negatively correlated with anxiety in women who have experienced early childhood trauma. ¹⁰⁰ In fact, there are differences among women in their susceptibility to early childhood experiences, which may be passed on across generations and which may also relate to individual differences in genetic profile. ^{101 –107}

Oxytocin genetics and mothering

Very few studies have explored the association between genetics within the OT system and mothering and they are predominantly human studies. ^{11,98,108 –110} Hence, we know more about the role of OT neuroendocrinology in mothering from the animal work and about OT genetics and mothering from human studies. Among new human mothers, evidence shows that the presence of a C allele on 1 OT gene single-nucleotide polymorphism (SNP) is associated with higher stress-induced anxiety and lower emotional well-being. ¹¹ Variations in OT-related genes modulate both non-postpartum depression and anxiety and, possibly, maternal mood and sensitivity in mother–infant interactions. ¹¹ Genetic variations within the OT system also interact with mothers’ own childhood experiences of adverse rearing and current stressors to influence mothering. ¹¹ These effects are often mediated through alterations in mothers’ affective state.

Specifically, Mileva-Seitz et al ¹¹ examined 2 polymorphisms in the OT peptide gene SNP, OT (rs2740210 and rs4813627), and one polymorphism in the oxytocin receptor gene OTR (rs237885) in 187 Caucasian mothers at 6 months postpartum. For OT, both rs2740210 and rs4813627 were significantly associated with maternal vocalizing to the infant. These polymorphisms also interacted with the quality of care mothers experienced in early life to predict variation in maternal instrumental care and postpartum depression. However, postpartum depression did not mediate the gene–environment (G×E) effects of the OT SNPs on instrumental care. In contrast, the OTR SNP rs237885 did not associate with maternal behavior, but it did associate with prenatal (but not postnatal) depression scores. These findings illustrate the importance of variation in OT genes as predictors of individual differences in human mothering, both alone and in interaction with early environment. Depression does not appear to have a causal role in the variation we report in instrumental care of the infant. This suggests that variation in instrumental care varies in association with an effect of gene by early environment interaction regardless of the magnitude of ongoing depressive symptomatology.

Finally, our findings highlight the importance of examining multiple dimensions of human maternal behavior in studies of genetic associations. Mothers who experienced early adversity in their family of origin and who showed increased depression, also showed reduced length of breastfeeding if they possessed the CC genotype of OT rs2740210 but not if they possessed the A allele (AA/AC genotypes). ⁹⁸ In the case of breastfeeding, then, depression apparently does play a role in mediating the relationship between early adversity and breastfeeding outcome, but only among women who are homozygous for the C allele. ⁹⁸ In this case, the gene variants have a moderating effect on the association between adversity and breastfeeding, reflected in a clear G×E interaction (OT genotype × level of early adversity). However, in a meta-analysis of the role of genes in the OTR system (as opposed to the OT ligand system, including AVP, for example), Bakermans-Kranenburg and van Ijzendoorn ¹¹¹ found that there was no evidence to support an association between 2 of the most frequently studied OTR coding region SNPs (rs53576, rs2254298) and 5 domains of human behavior. This meta-analysis was limited, however, by its focus on only 2 SNPs and a broad definition of human behaviors that included personality, psychopathology, and autism.

In a more recent analyses of the Maternal Adversity Vulnerability and Neurodevelopment sample, we tested and tagged 2 SNPs, OTR rs237885, which was not explored by the Bakermans-Kranenburg group ¹¹¹ and OTR rs2254298 (which was) against executive function at 48 months postpartum. ⁵⁷ Executive functions were assessed using the Cambridge Neuropsychological Test Automated Battery (CANTAB; Attentional set-shifting, Spatial working memory, Stop signal task, and Decision-making). A mother–child interpersonal interaction at 48 months postpartum, the Etch-a-Sketch task, was coded and factored into Physical Controlling Parenting and Positive Parenting. OTR rs2254298A had an indirect effect on Positive Parenting mediated through parental performance on Decision-making. Further, OTR rs2254298A had both direct and indirect effects on Physically Controlling parenting, in this case, also mediated through enhanced performance on CANTAB Decision-making. ⁵⁷

Although OT has previously been associated with only the onset of maternal behavior, “maternal motivation,” and warmth, involved in bonding, close contact, and nursing during the early postpartum period, we now know that OTR polymorphisms are associated with maternal behavior at 48 months postpartum via associations with the mothers’ executive functions. These studies highlight the importance of the OT system to maternal sensitivity at birth and beyond infancy and the role in these associations of mothers’ executive function.

Dopamine

Dopamine physiology and mothering

Oxytocin does not act alone. Dopamine is also a central neurochemical for the expression of mothering in the rodent model. In fact, OT is one regulator of DA release and activity in the nucleus accumbens (nACC), and nACC and DA are important for the maternal attraction to young ^112,113 and to consolidation of maternal experience. ^114,115 Over the first postpartum week, rat mothers become increasingly attracted to the young. ¹¹⁶ This effect is dependent on the interactions of the OT and DA systems. The OT infused into the midbrain region stimulates DA release in a forebrain region (the nACC) involved in experience-enhanced stimulus salience and reward. ¹¹⁷ Forebrain DA levels increase during both nursing bouts and licking or sniffing interactions with pups not only in newly parturient female rats but also in response to foster pups in multiparous experienced animals and virgins induced to become maternal by continuous exposure to pups. ^{113,117 –121}

Chemical lesions of DA neurons or infusions of DA receptor (D1, D2) or DA transporter (DAT) antagonists, profoundly disrupt maternal behavior in the mother. ^{113,117 –124} Agonists produce the opposite effect. ²⁴ Moreover, the magnitude of DA release in the forebrain in the mother is directly related to the level of pup LG by rat mothers and to the experience level of mothers. ^119,120,125 The DA release in the cortex also occurs and is implicated in the mediation of cortical executive function. ^126,127 Microarray studies show that genes for DA receptors show differential expression in the brains of virgin and postpartum animals. ¹²⁵ Together, these studies suggest that the mesocorticolimbic DA system regulates maternal behavior in part by acting on the forebrain to influence incentive motivation, reward, and lower level attentional processes, and in part by acting on the PFC to influence executive function (eg, working memory, attention). Lesions to brain regions that receive DA input disrupt approach responding and sequential organization of behavior, respectively. ¹²⁸

There is no direct evidence relating intracerebral levels of DA to human mothering, or mothers’ responses to infant cues. However, functional magnetic resonance imaging and positron emission tomography (PET) studies reveal that activity along DA pathways in forebrain, limbic system, and cortex is associated with individual differences in responsivity to infant stimuli assessed through blood-oxygen-level-dependent contrasts. Increased activity within these systems is associated with enhanced executive function and selective attention functions necessary for maternal responsiveness. ^{23,129 –133} Not surprisingly, these same regions are also activated when women view pictures of their own versus other children or hear infant cries. ^{134 –137} The DA system may differ from the OT system in becoming more involved when the reward, attentional systems, and memory come to be more strongly involved in maternal behavior as mothers gain experience with infants. This starts soon after birth but demands on these capacities increase over the ensuing months and years (6-60 months).

Dopamine genetics and mothering

What is the role of the DA-related genes to mothering? One of the first DA genetic association studies of maternal behavior in humans found no direct association between the DA D2 receptor Taq1 polymorphism and maternal sensitivity, parenting stress, negativity, or depressive symptoms. ¹³⁸ When environmental influences were also considered, however, it was found that DA D4 receptor and catechol-O-methyltransferase (COMT) Val158 Met polymorphisms in mothers were found to interact with daily stresses, or “hassles,” to predict maternal sensitivity. ¹³⁸ Consistent with these findings, mothers with a DA D4 receptor 7-repeat allele behaved more sensitively to fussier infants and less sensitively to less fussy infants compared to mothers without the 7-repeat allele, suggesting that mothers are differentially sensitive to their infants’ fussiness dependent on their own DA genotypes. ¹³⁹ As well, DAT polymorphisms in mothers were associated with “negative” maternal behaviors during a structured mother–child interaction task; an effect that was particularly strong if the infants were being disruptive during the interactions. ¹⁴⁰ Finally, in a recent series of studies relating DA genes to mothering at 6 months postpartum, we found that 2 DA D1 receptor SNPs were significantly associated with time mothers spent attending to the infant. ⁹

These findings are consistent with the idea that DA acts at the D1 receptor to enhance the salience and thus attention toward infant stimuli. Mothers who are heterozygous for the DA D1 receptor SNP oriented away from their infants significantly less than either homozygote group, showing evidence of heterosis for the DA D1 receptor haplotypes. In contrast, 2 DA D2 receptor haplotypes were significantly associated with maternal vocalizing/speech to the infant. Our findings provide important evidence that genetic variation in receptors critical for mothering in nonhuman species also affects human maternal behaviors. This also highlights the importance of exploring multiple dimensions of the complex human mothering phenotype. In contrast, in terms of incentive salience or “reward,” no studies to date have explored genetic influences on mothers’ attraction to infants or positive stimulus salience in general. Most of the genetic studies have examined genetic predictors of neuropsychopathologies in the reward system, including impulsivity, addiction, and gambling, where polymorphisms within the DA receptor systems have been associated with variations in a variety of addictive traits or reward deficiency conditions. ^{141 –143} In terms of depression, to our knowledge, there are no studies on genetic variation in DA systems relating to postpartum depression. However, genetic variants in the DA-related genes, COMT, and monoamine-oxidase A all appear to be important predictors of depression. ^{144 –147}

Serotonin

Serotonin Physiology and mothering

Serotonin influences mood, emotion, and cognition, as well as regulatory functions like sleep. ¹⁴⁸ Central 5HT release occurs primarily in the hindbrain of the brainstem, with axons projecting to the entire brain. The 5HT has multiple effects on maternal behaviors, depending on which receptor systems and brain regions are affected. ¹¹⁵ The 5HT brain levels in rats are associated with maternal aggression and maternal behavior; different 5HT receptor agonists increase maternal aggression and disrupt expression of many maternal behaviors, without affecting maternal motivation, affect, or stimulus salience per se, ¹⁴⁹ although some 5HT agonists seem to disrupt aspects of maternal startle and attention. ^{150 –152} In primates, 5HT is implicated in anxiety arousal of mothers ¹⁵³ and lower CSF 5HIAA levels associate with more restrictive and rejecting maternal behaviours. ¹⁵⁴ Female monkeys raised by abusive mothers have lower levels of 5HIAA and are more likely to abuse their own offspring. ¹⁵⁴

There are no studies of 5HT and mothering in humans, although 5HT dysregulation is thought to be related to postpartum depression and other affective disorders and is associated with increased aggressive behaviours. ^148,155 There is also a substantial literature that suggests that outside the postpartum period, depressed individuals find various social stimuli to be less rewarding and that there is reduced neural activation in the striatum among depressed patients. ⁵⁰ Indeed, depression associates with states of anhedonia and impairments in executive functions, both of which serve as the basis for later parenting difficulties. Although DA, OT, and 5HT are implicated in major depression, the 5HT system seems to take precedence, likely in conjunction with the other systems in different aspects or symptoms of depression. ^{155 –158} In fact, DAergic agents appear particularly useful as adjunct treatments targeting hedonic processes and the quality of life. ¹⁵⁹ The PET studies show decreased brain 5HT2A receptor density in drug-naive depressed patients. ¹⁶⁰ Therapeutic action of antidepressants is associated with the increase and/or normalization in brain 5HT2A receptor density. ^161,162

In terms of 5HT gene polymorphisms and mothering, one gene has received wide attention in the developmental literature and that is the serotonin transporter (5HTT) gene. ^163,164 We examined the influences of 5HTT gene polymorphism and early-life experience on the expression of individual differences in 3 dimensions of maternal responsiveness in 6-month postpartum mothers: These included maternal sensitivity, maternal behavior (orienting away from the infant), and maternal attitudes (perceived attachment). ¹⁰ Mother’s 5HTT gene polymorphisms are characterized by 1 of 3 allele types designated S (short), L (Long), and Lg (Long as well but Lg is functionally similar to S allele in terms of 5HTT expression). ¹⁶⁵ We found that mothers carrying an S, or the functionally similar Lg allele, were more sensitive than those lacking these alleles. We also found a highly significant G×E interaction effect on 2 other mothering dimensions. With increasing quality of care, mothers with an S, or Lg allele, tended to orient away from their babies less often and score higher on ratings of perceived attachment whereas mothers lacking an S (or Lg) allele tend to exhibit increased frequency of orienting away and lower ratings on perceived attachment with increased early care quality. ¹⁰

These results are not consistent with an earlier study by Bakermans-Kranenburg and van Ijzendoorn ¹⁰⁸ who report that independent of daily stress levels, S-carrying mothers are less maternally sensitive, which is opposite to the present findings. However, there were many differences between these 2 studies, including the age of the young at time of assessment (6 months vs 2 years) and the location of the assessment (less stressful home environment vs more stressful laboratory setting). It is possible that S-carrying mothers are more sensitive than non-S-carrying mothers under conditions of low stress, but less sensitive under conditions of high stress. This hypothesis is consistent with arguments that genetic influences depend on environmental conditions ¹⁰³ and fits well with the emerging theory that the S allele confers greater sensitivity to environmental signals, both good and bad, rather than acting only as a “vulnerability” allele. ¹⁶⁶ Rather than a susceptibility (or vulnerability) allele, the S allele may be viewed as a plasticity allele. ^167,168

These interactions between early experience and genotype may be related to underlying neural and cognitive evidence that the S allele predicts improved attentional processes ¹⁶⁹ and social cognition, ¹⁷⁰ which as indicated above are important components of mothering. ^23,61,171 In terms of physiological mechanisms, the 5HTTLPR serotonin transporter genotype in humans is also directly associated with differences in 5HTT binding in a variety of brain sites important for DA and mothering. ^{172 –175} This is consistent with the theory that the neurotransmitter 5HT inhibits dopaminergic neuron signalling. ¹⁷⁶

Epigenetic Mechanisms Relating to Mothering

The environment could affect later mothering through direct effects on gene expression. This has been demonstrated in the cortisol–hypothalamic–pituitary–adrenal system rats. ^{7,177 –179} Unfortunately to date, we have no information about epigenetic processes acting on the DA or 5HT genes that we have argued are important for mothering. However, such effects have been found for other gene products that affect mothering, especially for some of the steroid receptors. As earlier described for the rat model, in comparison to high-licking mother rats, mothers who lick their pups less have female offspring who as adults lick their offspring less. ^{17,18,180,181} In the MPOA (the area crucial to mothering), these female offspring show both reduced expression of the estrogen receptor α gene (and hence are less sensitive to estrogen action) and an increased DNA methylation of its promoter region, which functionally reduces gene expression. ¹⁸² A similar epigenetic process may also affect the BDNF gene under adverse early experiences. ¹⁸³ The reduction in the demethylation (or a retention of methylation) is a mechanism by which the early adverse experience, as opposed to a positive and healthy experience, could have a long-term impact on mechanisms through which mothering behavior is mediated. Finally, as early adversity effects on methylation patterns in the human (postmortem) brains have also been reported, ¹⁸⁴ it would be important to know whether the gene-by-early experience interactions reported in relation to mothering involve differential methylation patterns of the relevant OT, DA, and 5HT genes. These studies are ongoing in the rat model. ^177,178

Summary and Conclusions

We know considerably more about the endocrine and neurochemical bases of mothering in nonhuman mammals than in humans. In rodents, we have determined their effects on brain mechanisms of maternal behavior directly and on mechanisms regulating other behavioral systems that affect mothering. In both nonhuman and human studies, many behavioral systems and experiences affect how mothers interact with their offspring. In a number of mammalian species, the quality of mothering is influenced by the mother’s affective state, attention, impulse control, cognitive learning function, executive function, and the salience of stimuli mothers attribute to their environments including infant cues and related stimuli. We also know from both animal and human studies that early-life experiences, current stressors, and surrounding social and physical environments impact the quality of mothers’ behavior and the effects of hormones and neurochemicals on mothering affect and behavior. What we have tried to briefly review but clearly know less about is the role of genetics in this process and how the environment affects the expression of potential gene products. What is the role of genes in mothers’ transition to parenting and in the quality of mothering shown? How do early stressors affect mothers with different genetic profiles? These are questions that will guide future research in this area.

What we have attempted to do is describe studies that show, primarily in human research, a potential role for certain gene systems in the regulation of early postpartum mothering, choosing those systems assiduously, and based on our knowledge of the relevant biochemistries shown to be important for nonhuman rat mothering. What is missing from our review is an extensive description of the multiple genes involved in the regulation of mothering, which can occur now through use of gene-wide arrays and predicted signaling pathways (as in DA, OT, and 5HT pathways) and the role of epigenetics in their action. Also understudied are the myriad of environmental and experiential influences that moderate the effects of these signaling pathways and their genetic bases. Clearly, one mechanism for the G×E interaction or interplay is through processes of epigenetics-methylation, histone acetylation, chromatin modification—which we know occur when the environment impinges on the genes and which regulates gene expression. ⁷ Clearly, an epigenetic analysis would provide us with a mechanism through which G×Es could occur, that is, how experiences, both early and recent, could exert their effects on gene expression to affect the brain mechanisms of mothering. We look forward to future explorations involving very large longitudinal samples of mothers (and fathers!) who have been deeply phenotyped (assessed for their lifetime family and social experiences, their own and child characteristics, their demographics, and their behavioral interactions with offspring). This program of research would focus on hypothesized gene systems that are derived from our knowledge of the underlying physiology of mothering/parenting and associated neurobiological affective and cognitive characteristics.