I VISITED THE TORONTO ZOO FOR THE FIRST TIME ON A BRISKLY beautiful day in October, 2008. One of my first stops was the gorilla exhibit, where I remained entranced for over an hour. The social dynamics were fascinating, and there was a docent present, who augmented what we were seeing with much useful information about the relationships among the individuals, their life stories and personality quirks. There were several adult females, a couple of juveniles, two youngsters—one born fairly recently—and a silverback male, much like a gorilla group in the wild. Physically, the silverback, Charles, was the most riveting figure, with his huge head and neck and an upper body that would put any steroid-enhanced body builder to shame. Charles was nearly thirty-five years old at the time, past his prime for a gorilla in the wild but still good to go in the cage. He had sired numerous progeny, including the youngest juvenile, during his many years at the zoo.
Charles’s personality, though, was much less compelling than his physique. Older silverbacks in general tend to be dour and not much prone to initiate interactions with the rest of the troop. But Charles is extreme in this respect, even for a silverback. For example, he has a strong-seeming aversion to his own young offspring, which is not normal. The females, both mothers and aunts, were acutely aware of this and tried to steer the ebullient youngsters away from papa, lest, presumably, they get a swat. This was quite entertaining to watch, as the youngest in particular was not easily deterred from interacting with the imposing fellow. The mother was often forced to physically divert her youngster, warily watching papa all the while.
Charles could perhaps be excused for his social infelicities, given his own traumatic early life. He was found in the wild beside the body of his dead mother, who had been shot by poachers. Soon thereafter he was sent to the Toronto Zoo, where he was raised by humans. He never had a chance at normal gorilla socialization. In retrospect, it is impressive that Charles was such a successful sire. Many human-raised male gorillas are sexually incompetent.1 Some who do engage in sexual activity can’t even do it right. In general, human-raised male gorillas don’t make good sires. There are a lot of data on this because a considerable percentage of captive gorillas are rejected or neglected by their mothers and must be hand raised as a last resort, for both ethical and conservation purposes. (All three subspecies of gorilla are threatened with extinction.)
For hand-raised females, the consequences are more far-reaching. Social interactions are much more important for adult females than males; it is their interactions that provide the social glue for a group. Human-raised females often have all sorts of issues as a result of improper socialization. And one of the main issues is mothering, which is why there are so many neglected captive gorillas that need to be human nurtured. This creates a vicious cycle: human-nurtured gorillas make inadequate mothers, which results in more human-nurtured gorillas and hence more inadequate mothers.
This problem is exacerbated by the fact that for conservation purposes, the females are often moved around from zoo to zoo for pairings designed to minimize inbreeding. This makes captive female groups more unstable and hence more stressful for the constituents. It takes a group to raise a gorilla and stable groups are always better. The effect of confinement itself must also have some effect, if only because any disputes are less easily meliorated by, for example, avoidance. In any case, there is a huge mothering issue among captive gorillas.
Clearly, mothering is not an instinct in gorillas; it is a learned skill. But the motherless mothers problem also points to the fact that to be a good gorilla mother, she must be in the proper emotional state, and not having had a gorilla mother herself makes it less likely that she will be in that state when confronted with her own baby.
For the purposes of this chapter, I will focus on the fact that gorilla mothering also represents a form of inheritance, a social inheritance.2 For gorillas, proper mothering requires a well-functioning social structure, which zoos have been unable to adequately replicate. The effects of these deviations from the social norm influence neural development and other physiological processes. These physiological changes are often mediated by epigenetic processes. In this chapter, we will explore epigenetically mediated social inheritance in gorillas and in other social animals from rodents to humans.
From Motherless Monkeys to Poorly Mothered Rats
In the 1950s, Harry Harlow conducted some pioneering experiments at the University of Wisconsin concerning the emotional attachment of an infant to its mother.3 The experiments were conducted on rhesus monkeys, a favorite primate model for many scientific investigations. I remember watching a film on the experiments as an undergraduate and feeling both enthralled and repulsed, for certainly by today’s standards, Harlow’s experiments seemed to cross an ethical line in his treatment of his subjects. Harlow himself did not come across as a particularly sympathetic figure but as an almost Dr. Strangelove caricature of an experimental psychologist. But the claims of his most outspoken detractors—that his sometimes sadistic seeming experiments yielded nothing of scientific value—are complete nonsense.
Harlow’s initial experiments were designed to address a fundamental question about the mother-infant bond from the infant’s perspective: is it sustained by the sustenance the infant receives (initially milk), or by some other, less obviously life sustaining qualities that the mother has to offer? The answer may seem obvious now but it wasn’t at the time. According to the then prevailing view of the behaviorist school, infants clung to their mothers only because of the food rewards she offered. Harlow doubted the behaviorist line and decided to test it. He constructed wire-monkey mothers, some equipped with nipples through which milk could flow, others lacking nipples but clothed in terry cloth. Neonates that had been recently removed from their biological mothers could choose which artificial mother to cling to. All the neonates quickly learned to get their milk from the bare-wire surrogate but spent the rest of the time, including sleep, clinging to the terry-cloth surrogate. The tactile stimulation provided by the terry-cloth surrogate—though a poor approximation of a mother’s fur—was more compelling to these infants than the milk provided by the bare-wire version.
Obviously, an inanimate wire surrogate, even covered in terry cloth, is no substitute for an actively grooming and solicitous mother. So the infants “nurtured” by the surrogates exhibited high levels of stress and deep psychosocial problems. Nor could they ever be properly resocialized with other monkeys, no matter what methods were tried.4 In the 1960s, one of Harlow’s students, Steven Suomi, investigated the behavior of surrogate-raised females when they became mothers themselves—in other words, the motherless mothers. He found that they were at best neglectful and at worse abusive toward their own infants.5 There are obvious parallels here with the captive gorilla situation. Moreover, these studies provided some important insights regarding the human condition, particularly the fact that child abuse and neglect tends to run in families over several generations. Neglect begets neglect. Abuse begets abuse.
But how does neglect beget neglect? What happens in the brain of a neglected infant to make it a neglectful parent? For an approach to answering that question, we turn again to the rat studies by Michael Meaney and his associates. Recall that mother rats vary in the degree to which they tactilely stimulate their offspring through licking, and that rats not receiving adequate licking tend to become stressed-out adults as a result of epigenetic changes in the NGF gene. What happens when these stressed-out adults become mothers? Pretty much the same thing, it turns out, as occurs in captive gorillas and Harlow’s rhesus monkeys. Neglected (lick-deprived) female rats become neglectful mothers.6
Through the rat model, we can delve a bit deeper into the mechanism for this social inheritance. One possible explanation is that stressed-out mothers are neglectful mothers: A lick-deprived mouse becomes a stressed-out mother as a result of epigenetic alterations to the NGF gene; because of her stress, she is a neglectful parent; as a result, her female offspring experience the same epigenetic alterations to their NGF genes and become stressed-out mothers too—and so the cycle continues. That seems to be part of the story, but not the whole story.
Mother rats, like all mammal mothers, including humans, undergo a suite of hormonal changes just prior to and after giving birth. Levels of oxytocin rise, as do levels of estrogen and estrogen receptors. Estrogen receptor levels seem to play a particularly important role in maternal behavior. Levels of this receptor are reduced in the female offspring of poor lickers relative to those of good lickers.7 One consequence of the low estrogen receptor levels is that when the female pup becomes a mother, she doesn’t respond normally to the elevated estrogen levels she experiences when she gives birth. One consequence of this dampened response is a reduction in the binding of oxytocin in the hypothalamus, a region of the brain that is crucial for maternal behavior. Oxytocin, especially in its actions in the hypothalamus, promotes social or affiliative behavior.8 (Some researchers, reductively, refer to it as the “love hormone.”) This effect of the estrogen receptor on oxytocin occurs because the estrogen–estrogen receptor complex stimulates the expression of the oxytocin receptor gene in the hypothalamus by binding directly to the oxytocin receptor gene’s control panel.
The reduced expression of the estrogen receptor gene in female pups tends to persist into adulthood, hence making it more likely that she will be a less devoted licker when she becomes a mother, perpetuating for another generation the effects of inadequate maternal care. As you might expect by now, the long-term effects of lick deprivation on the expression of the estrogen receptor gene are epigenetic in nature. Pups born to good lickers but raised by poor lickers have lower levels of the estrogen receptor in the hypothalamus than their siblings that remained with the biological mother.9 The reverse is also true: pups born to poor lickers but raised by good lickers have elevated levels of the estrogen receptor in the hypothalamus. In both cases, the alterations in estrogen receptor levels in the hypothalamus persist as a result of methylation of the control panel for the estrogen receptor gene. Pups raised by poor lickers evidence more methylation of this control panel than do pups raised by good lickers. (Recall that high methylation levels generally result in lower expression levels for a given gene.)
Social Inheritance
Female offspring of poor lickers experience an epigenetic double whammy that affects their own mothering. First, as we saw in Chapter 4, their stress response is elevated as a result of epigenetic alterations in the NGF gene in the hippocampus. Since the presence of neonates is stressful in and of itself, they are distractible and unsolicitous in the presence of their own young. Second, as a result of epigenetic alterations of the estrogen receptor gene in the hypothalamus, they are less likely to lick their own offspring even when they are relatively unstressed.
Therefore, poor mothering tends to perpetuate itself in a vicious cycle; conversely, high-quality mothering tends to perpetuate itself in a virtuous cycle, from one generation to the next. This is a form of social inheritance mediated by epigenetic processes. Though most research on maternally based social inheritance has been conducted on rodents, there is substantial evidence for similar processes in primates, including humans. In rhesus monkeys, the species studied by Harlow, maternal rejection and abuse during the first three months causes a multitude of pathologies in brain and behavior, including the stress response.10 Similar effects have been observed in other primates.11 In rhesus monkeys, under conditions much less severe than those of Harlow’s experiments, maternal behavior, and hence its effects on offspring, tends to “run in families.”12
Humans have an especially protracted infancy and childhood. Children subject to poor parenting, including both psychological and physical abuse, suffer reduced mental health.13 As in rats and monkeys, these effects are associated with an altered stress response.14 Moreover, also as in rats and monkeys, poorly parented children tend to grow up to be poor parents.15 Recall from Chapter 4 that lick-deprived rats have an elevated stress response as a result of epigenetic changes to the NGF gene in the hippocampus. There is now evidence for a similar effect in humans abused during childhood.16
But parenting that falls far short of the abuse threshold can also have lifelong effects on behavior, much of it mediated by the stress response. The best documented of these effects in humans concern the quality of maternal care as measured by an index called the Parent Bonding Instrument, or PBI. Low scores for maternal care are often, somewhat paradoxically, associated with high levels of maternal control. The combination is referred to as affectionless control, which is a risk factor for depression, anxiety, antisocial personality disorder, obsessive-compulsive disorder, drug abuse, and a reactive stress response.17In contrast, high levels of maternal care, as measured by the PBI, are associated with high self-esteem, low anxiety, and a dampened stress response.18
“Maternal style,” is the term sometimes used to describe the suite of behavioral responses of a mother to her offspring.19 The term encompasses not only abuse and neglect but also what would be considered the normal range of maternal behavior, from affectionless control to an affectionate hands-off approach, and all manner of intermediate conditions. In both rats and humans, maternal style within the normal range can be transmitted transgenerationally.20 In humans though—in contrast to rats and most other mammals, including primates such as rhesus monkeys and gorillas—fathers, as well as mothers, play an important role in parenting. Paternal style has not been much studied to date, nor its effects on the emotional behavior and stress response of the next generation. But evidence from studies of child abuse and its social transmission suggest an important role for the male parent. Moreover, one recent study found a correlation between levels of the master stress hormone CRH (see Chapter 4) and reported levels of parental, not just maternal care.21 The social inheritance of paternal style clearly requires investigation.
Bent Twigs
There is much truth to the old adage that as the twig is bent, so the tree inclines. Things that happen early in life have long-lasting consequences. As we have seen, one mechanism behind this tendency is environmentally induced epigenetic change. But in my many walks in the forest, I have noticed that some trees start growing in one direction, then change direction quite dramatically, sometimes up to ninety degrees. This usually occurs when the tree begins life by growing sideways due to some environmental impediment, like a rock or other trees, then makes a turn toward the vertical and the sunlight. There are numerous analogous cases in human psychological development. Many who get off to a bad start make midcourse corrections. Most victims of child abuse do not become child abusers. That cycle can be broken.
Infant-parent interactions are the foundation of the process of socialization but subsequent events, especially interactions with peers, also figure prominently in social and hence emotional development. This is true even in rats. Michael Meaney and his colleagues were able to reverse many of the adverse effects of poor mothering by providing his subjects an enriched social environment after weaning.22 After time spent with well-adjusted same-sex peers, there were noticeable changes in the stress response. Significantly, this change was accompanied by changes in the methylation of the NGF gene. Though these epigenetic attachments tend to persist, they are not irreversible.
Primates such as rhesus monkeys can also be rehabilitated in similar ways. There are limits, of course. Harlow’s motherless monkeys could not be rehabilitated, nor could Charles the gorilla ever be properly socialized beyond sexual intercourse. And Charles was fortunate in that respect, compared with many other motherless male gorillas. Sometimes the twig is bent too much to be remedied.
In humans, given our protracted period of socialization, the opportunities for overcoming an adverse childhood seem especially promising. To the extent that children at risk experience successful remediation, we would expect to find epigenetic reversals of the sort identified by the Meaney lab, as well as perhaps new epigenetic alterations in other genes. (Epigenetic processes do not end or begin with childhood.) For intractable cases where pharmaceutical intervention is deemed necessary, the most effective drugs will be those that epigenetically alter gene expression. The Meaney lab successfully used such pharmaceutical treatments to remedy the stress response of rats that were victims of poor mothering.23
Those badly bent twigs that are not remedied by subsequent alterations to their social environment develop into prostrate trees that bear poor fruit. This is the pathological dimension of social inheritance, the dimension of social inheritance that is easiest to identify, but only the tiny, exposed portion of the iceberg.
Expanding Our Notion of Inheritance
We inherit more than our genes. Among the extragenetic things we inherit is a social environment that begins with our parents but can extend well beyond that, up to and including a whole culture. Gorillas inherit their social environment as well. Charles and other captive gorillas dramatically evidence what can go wrong in the socialization process when this social environment is aberrant. Harlow created even more extreme pathological social conditions in his maternal deprivation experiments. His motherless mothers could not begin to adequately care for their offspring. Charles, at least, had bonded with his mother before the poachers killed her. Many captive-born gorillas aren’t so lucky. While not subjected to the extreme deprivation of Harlow’s monkeys, they are raised by poor gorilla-mother substitutes called humans, and become inept fathers or neglectful mothers as a result, in a perpetual cycle of pathological socialization.
Social inheritance of a less pathological sort was long ago demonstrated in rats. A female rat subjected to stress transmits her elevated stress response to her daughters and her daughters’ daughters. But even unmanipulated rats that exhibit the normal range of maternal behavior (as evidenced by high or low grooming rates) tend to transmit their maternal style to their female offspring. In this case, though, scientists have uncovered the mechanism, which involves epigenetic changes in two genes: NGF and an estrogen receptor. It will be worth exploring whether the same epigenetic processes underlie the transmission of maternal styles in nonhuman primates and humans. In humans, unlike in most mammals, the transmission of paternal style is worth exploring as well.
The Meaney lab has shown that the effects of poor parenting can be reversed through an enriched social environment for the offspring of rats within the normal range of maternal behavior. Not surprisingly, the behavioral changes are associated with epigenetic changes in the NGF gene. The pathologies induced by Harlow on his rhesus monkeys and by captivity on gorillas are more refractive. But though human child abuse runs in families, most of those abused do not become abusers, suggesting an important role for subsequent socialization.
Though parenting style, including child abuse, is not transmitted from one generation to another with the same fidelity as classical genetic inheritance of, say, eye color, it is an important form of inheritance nonetheless, especially with regard to psychosocial behavior. In fact, social inheritance can be easily mistaken for classical genetic inheritance involving “abuse genes” and such. Genes are involved, to be sure, but as effects rather than causes. To the extent that genes play a causal role in the social inheritance described here, it is through their environmentally induced epigenetic modifications. Is this, then, a form of epigenetic inheritance? A case could be made that it is an indirect form for epigenetic inheritance. The difference between this indirect epigenetic inheritance and direct epigenetic inheritance will be explored in the next chapter.