Since the early 1990s there has been a steady increase in prescription rates for opiates (Substance Abuse and Mental Health Services Administration [SAMHSA], 2009), including prescribing for conditions ranging from routine dental procedures to menstrual cramps. This escalation in prescribing, coupled with increased availability of these substances online (Forman et al., 2006), and the advent of long-acting forms of opiate analgesics such as OxyContin®, have combined to create a dangerous upsurge in both the medical and non-medical use of prescription pain medications (Paulozzi et al., 2006; Cai et al., 2010). Of great concern is the increased use of these potent opiates in adolescent populations (Sung et al., 2005), with 60.6% of respondents in a recent survey reporting initiation of use before the age of 15 (Wu et al., 2008). Misuse in younger populations is likely due to the decreased risk perception associated with prescription opiates (e.g., as compared to heroin), in conjunction with their increased availability. Indeed, overall, prescription drugs are reported as the “drug of choice” in 12- and 13-year-old populations (SAMHSA, 2007) and unlike other drugs of abuse, they are used at higher rates in young female populations (Sung et al., 2005; Alemagno et al., 2009). Currently, the long-term impact of this increased use of opiates in adolescent female populations remains unknown.

Given that adolescence represents a period of significant brain maturation, perturbation of the endogenous opioid system during this period may induce significant long-term effects. Endogenous opioids (beta-endorphin, dynorphin, and enkephalin) and their receptor targets (mu, kappa, and delta) are ubiquitous, serving as critical modulators of neural, endocrine, and immune function. For example, during adolescence, endogenous opioids modulate the timing of sexual maturation (Cicero et al., 1986; Reiter, 1987; Sizonenko, 1987). Opioids also regulate stress responsiveness as well as modulating numerous cognitive and reward-related processes (Zager and Black, 1985; McCubbin, 1993; Van Ree et al., 2000; Drolet et al., 2001; Kreek, 2007). Thus, adaptations in response to high levels of opioids during adolescent development could impact a wide-range of opioid-mediated functions.

A significant body of literature indicates that endogenous opioids play a role in maternal behavior. Initial studies conducted with morphine, showed a disruption of postpartum maternal behavior (Bridges and Grimm, 1982; Rubin and Bridges, 1984; Kinsley and Bridges, 1986, 1990; Mann et al., 1990; Kinsley et al., 1995; Sukikara et al., 2007). These disruptive effects were even more pronounced when the female was pre-exposed to morphine during pregnancy (Bridges and Grimm, 1982; Miranda-Paiva et al., 2001; Slamberova et al., 2001). In a number of species, including humans and non-human primates, endogenous opioids modulate affiliative behaviors, including mother–infant attachment (Panksepp et al., 1994; Kalin et al., 1995; Nelson and Panksepp, 1998; Saltzman and Maestripieri, 2010). Indeed, there is evidence in women, that opioid use (e.g., methadone) can induce subtle alterations in mother–infant contact and attachment (Goodman et al., 1999). Moreover, studies indicate that endogenous opioids in both the mother, and the infant, are important for infant emotional regulation and attachment (Schino and Troisi, 1992; Weller and Feldman, 2003; Barr et al., 2008). If this system is altered by prior exposure to opiates, then postpartum maternal–offspring interactions could be affected.

We have previously documented significant effects of adolescent morphine exposure on the expression of opioid-related genes, including increased expression of mu- and kappa-opioid receptor genes and decreased expression of the proopiomelanocortin gene in the mediobasal hypothalamus (Byrnes, 2008). In addition, we observed attenuated suckling-stimulated prolactin secretion during early lactation in these females (Byrnes, 2005b, 2008). These findings demonstrate long-term changes in the endogenous opioid system of adolescent-exposed females, as well as alterations in physiological parameters that can influence maternal care. One additional component of these studies was an examination of rudimentary aspects of maternal behavior, such as the latency to retrieve and crouch over pups following a brief separation. No differences in maternal behavior latencies were observed, however, such measures do not assess quantitative or qualitative aspects of maternal care.

In addition to observing direct effects of adolescent morphine exposure on the female rat, we have also demonstrated transgenerational effects in both male and female offspring. These offspring effects include differences in anxiety-like behavior, shifts in morphine sensitization and alterations in morphine analgesia (Byrnes, 2005a; Byrnes et al., 2011). It is important to note that all adolescent-exposed females in our studies are drug-free for several weeks prior to mating. Thus, the developing embryo/fetus is never directly exposed to morphine. These findings indicate that even when morphine exposure is confined to the adolescent period, there can be significant effects on future offspring. Moreover, the nature of these transgenerational effects suggests an alteration in the endogenous opioid system of the offspring. Variations in maternal care can significantly impact offspring neurodevelopment, including the development of the endogenous opioid system (Weaver et al., 2007; Gustafsson et al., 2008; Michaels and Holtzman, 2008). Thus, one potential mechanism underlying transgenerational effects of adolescent morphine exposure may be altered maternal–offspring interactions.

The current study was designed to examine maternal behavior in females exposed to morphine during adolescent development. In addition, we also investigated social play behavior in their male and female offspring. The choice of social play behavior was based upon studies indicating that play behavior is modulated by opioids (Niesink and Van Ree, 1989; Vanderschuren et al., 1995; Van den Berg et al., 2000) and can be altered by changes in the postnatal environment (Janus, 1987; Veenema and Neumann, 2009). Our working hypothesis is that adolescent morphine exposure induces significant changes in the endogenous opioid system of both the female and her offspring.

The current findings demonstrate that exposure to escalating doses of morphine, confined to the adolescent period, can induce subtle modifications in subsequent maternal care and can alter the behavioral phenotype of subsequent offspring. These effects were largely expressed as differences in frequency of nursing and contact time during early lactation. In addition, the offspring of MOR-F0 mothers showed sex-specific differences in rough and tumble play. These results indicate that even when opiates are withdrawn several weeks prior to mating, a history of opiate exposure can influence both maternal care and offspring development.

A number of animal models have documented the effects of changes in maternal care on developing offspring. For example, the amount of licking and grooming that a female exhibits in the first week postpartum, can modify the behavioral phenotype of her offspring (Caldji et al., 1998). These offspring effects include shifts in the maternal behavior of adult female offspring (Champagne et al., 2003; Kikusui et al., 2005), as well as changes in stress responsiveness (Fish et al., 2004), play behavior (Moore and Power, 1992), and cognition (Liu et al., 2000; Champagne et al., 2008). Other models, focusing on the effects of either brief or prolonged maternal separation, have also illustrated the importance of mother–offspring interactions (D'Amato et al., 1998), with significant effects on the regulation of fear and anxiety, stress responsiveness, and motivated behaviors observed in adult offspring (Romeo et al., 2003; Lee et al., 2007; Michaels et al., 2007; George et al., 2010; Skripuletz et al., 2010; Macri et al., 2011). Often, these effects are sex-specific (Slotten et al., 2006). It is not clear whether all of these alterations in offspring development are directly related to maternal behavior. It is certainly possible that a number of these effects are mediated by factors present in the milk (e.g., corticosterone or prolactin) or represent some interplay between maternal care, maternal endocrine milieu, and the offspring's own physiology. When considering how these findings might translate to human mothers and their infants, it is important to remember that there are significant developmental differences between rodents and humans. Indeed, neurodevelopment in the postnatal rat is comparable to that observed in second and third trimester infants (Bayer et al., 1993). By determining what mechanisms underlie changes in adult phenotype that are induced by alterations in maternal care, we may gain significant insight into how early life experience, both in utero and during the early postpartum period, may alter neurodevelopment. Overall, these findings on maternal care in rodents clearly indicate that even ostensibly modest changes in the postpartum environment can significantly impact offspring development.

How then might adolescent morphine exposure induce alterations in maternal care? One possible mechanism may be a shift in the endogenous opioid system. Indeed, we previously observed significant shifts in the regulation hypothalamic, opioid-related gene transcription following adolescent opioid exposure (Byrnes, 2008). These effects on gene transcription persisted for at least 10 weeks following cessation of morphine administration. Thus, a shift in endogenous opioid-mediated regulation of maternal behavior may be one consequence of adolescent morphine exposure.

The importance of endogenous opioids in both the pre- and postnatal period has been well documented. Opioids directly regulate numerous aspects of embryonic and fetal development (Kar and Quirion, 1995; Leslie et al., 1998; Zagon et al., 1999; Kivell et al., 2004; Cooney et al., 2009). Moreover, during pregnancy and parturition, endogenous opioids modulate both the maternal and fetal hypothalamic–pituitary–adrenal axis (Taylor et al., 1997; Douglas et al., 1998), and regulate maternal central oxytocin activity (Douglas et al., 1995; Douglas and Russell, 2001; Kutlu et al., 2004). The role of endogenous opioids in the regulation of specific aspects of maternal care in the rat is not well-defined. There is evidence, however, that the administration of an opioid antagonist increases the duration of nursing bouts and mother–offspring contact time during early lactation (Byrnes et al., 2000). These data fit well with the hypothesis that opioids regulate maternal–offspring attachment processes (Nelson and Panksepp, 1998; Weller and Feldman, 2003). Thus, in the face of opioid receptor blockade, the female may prolong contact and/or nursing bouts to achieve a similar level of reward associated with pup contact. In this context, one would postulate that MOR-F0 females may have more sensitive opioid receptors and therefore may terminate their nursing bouts sooner. While we observed significantly increased mu- and kappa-opioid receptor mRNA in non-lactating MOR-F0 females, no differences in the expression of these receptor subtypes were observed during lactation (Byrnes, 2008). Of course, protein expression and/or postsynaptic response of these receptors could be modified in MOR-F0 females. Moreover, our previous studies only examined changes in the mediobasal hypothalamus. Certainly, significant differences in opioid receptor number and/or function in any number of brain regions could underlie changes in nursing behavior during early lactation. This would not, however, explain the increased contact time (i.e., hovering) observed in MOR-F0 mother. Perhaps then, some other change in the female's behavior might underlie the observed effects on nursing frequency.

Examination of the differences in nursing behavior and contact time reveal a significant influence of both the time of day and the testing method. MOR-F0 females only spent more time away from the nest during the dark and these differences were only significant when tested using frequency data. Similarly, when recorded continuously, no decrease in the duration of nursing was observed during early lactation. One obvious difference between our frequency and duration data was the presence of an observer during frequency data collection. We have noticed that qualitatively, MOR-F0 females appear to be more sensitive to any disturbance in their environment. Thus, while all of our observations were conducted in the home cage, and care was taken not to disturb the female and her litter, MOR-F0 females would often come off the nest and rear up toward the front of the cage during the 30-min observation session. In addition, this behavioral pattern appeared more robust during the dark phase. Thus, the simple act of observing MOR-F0 females may have altered their nursing frequency. These effects would be more robust during early lactation when immature pups are unable to maintain nipple contact when the dam rears or changes positions. If a more general, non-specific, increase in “distractibility” underlies these changes in maternal care, this would suggest that MOR-F0 mothers may be more likely to decrease the care of their offspring in the presence of substantial environmental distracters.

The data on the duration of nursing suggests that during later postnatal time points, MOR-F0 mothers have more prolonged nursing bouts. As mentioned previously, during these later time points pups are more able to maintain nursing contact even when the female moves either within or even off of the nest. Thus, rearing up in the presence of any distracter would not be as detrimental to nursing when pups are older. However, this would not explain why the female nurses for longer periods when compared to SAL-F0 females. One possibility is that the female and/or offspring are compensating for decreased nutrition during early lactation with more sustained lactation during later development. Indeed, the bodyweight data demonstrate that MOR-F1 weanlings are heavier than their SAL-F1 counterparts. To what extent maternal nursing behavior, as opposed to some other metabolic factor, induces this difference in body weight is unknown. However, we have previous data suggesting differences in the lactogenic hormone prolactin in MOR-F1 mothers, with lower levels observed during early lactation and higher levels observed during late lactation (Byrnes, 2005b). Thus, MOR-F0 females may not necessarily be deficient in their nursing behavior, but rather may demonstrate a shift in the developmental profile of this behavior over the course of lactation. Overall, these data suggest that subtle modifications in maternal care, especially relating to nursing and non-nursing contact, are a consequence of adolescent morphine exposure. To what extent these differences play a role in the behavioral phenotype of their offspring remains to be determined.

In addition to modifying maternal care, adolescent morphine exposure induced transgenerational effects on juvenile play behavior. MOR-F1 males demonstrated a significant reduction in rough and tumble play, with no change in other aspects of social behavior (social grooming or exploration). In addition, there was a trend toward increased rough and tumble play behavior in MOR-F1 females, although these effects did not achieve significance. Rough and tumble play is displayed by a wide-range of species and is an important developmental marker (Auger and Olesen, 2009; Auger et al., 2011). It has been suggested that low levels of rough and tumble play may indicate vulnerability toward reduced motivated behavior in adulthood (Trezza et al., 2010). For example, animal models have demonstrated a relationship between rough and tumble play and future sexual and aggressive behaviors (van den Berg et al., 1999a; Cervantes et al., 2007; Wommack and Delville, 2007). Thus, decreased rough and tumble play in MOR-F1 males may suggest an increased risk for deficits in other motivated behavior in adulthood.

The neural systems underlying rough and tumble play have been fairly well elucidated (Gordon et al., 2002), with opioids playing a significant role in the regulation of this behavioral repertoire. Specifically, administration of morphine enhances play behavior, while administration of the mu-opiate receptor antagonist naloxone, reduces rough and tumble play (Panksepp et al., 1985; Vanderschuren et al., 1995; Guard et al., 2002; Trezza and Vanderschuren, 2008). Moreover, when juveniles are socially isolated, thereby eliminating all experiences of play, both mu and kappa receptors are significantly up-regulation in several nuclei related to emotional regulation (Van den Berg et al., 1999b). Thus, the reduction in rough and tumble play behavior could be a symptom of a down-regulation of functional mu-opiate receptors in MOR-F1 males or conversely, their low levels of play could induce alterations in opiate receptors. In line with these data, previous findings in adult MOR-F1 males, demonstrate significant alterations in their response to opiates (Byrnes, 2005a; Byrnes et al., 2011). Thus, one intriguing possibility is that MOR-F0 females transfer modifications in neural opioid systems to their offspring. The mechanism underlying this type of epigenetic effect is unknown, but could certainly involve alterations in maternal care or maternal endocrine milieu.

Finally, consideration of the current findings in the context of effects following prenatal morphine administration is warranted. Certainly, morphine exposure during the prenatal period has been shown to significantly impact offspring development (Sobrian, 1977; Vathy and Katay, 1992; Lesage et al., 1996; Vathy et al., 2000; Slamberova et al., 2005), with many of these effects involving changes in opioidergic function (O'Callaghan and Holtzman, 1976; Ramsey et al., 1993; Gagin et al., 1997; Chiou et al., 2003; Villarreal et al., 2008). Interestingly, the effects we observed in MOR-F1 animals are in the opposite direction of those observed in the offspring of females exposed to morphine in utero, with rough and tumble play behavior found to be increased in the offspring of females exposed to morphine during gestation (Hol et al., 1996; Niesink et al., 1996). In fact, our effects are more similar to those observed following administration of opioid antagonists during fetal development (Shepanek et al., 1995; Medina Jimenez et al., 1997). Thus, the transgenerational effects observed in the offspring of adolescent morphine-exposed females may be indicative of a down-regulation of the endogenous opioid system in their mothers, perhaps both pre- and postnatally, which is then transmitted to their offspring via currently unidentified, epigenetic processes. Such processes could include modifications in maternal–offspring interactions.

Adolescent use of prescription pain relievers has increased dramatically in the past decade, especially in young females. Beyond the risks of overdose or addiction, the long-term effects of exposure to such powerful opiates during a critical period of neurodevelopment are unknown. As endogenous opioids play such a significant role in reproductive function, prior opiate use could influence pre- and/or postnatal factors. Given the importance of opioids in maternal–offspring interactions, and the critical role that mothers play in the healthy development of their children, it is possible that opiate use in adolescent girls could have repercussions for future generations. To begin to elucidate the possible long-term effects of adolescent opiate use, animal models examining the impact of adolescent opiate exposure on both the female and her future offspring are required. The current findings demonstrate that exposure to increasing doses of morphine during adolescent development can induce subtle changes in maternal care and offspring development. While the neural and/or endocrine mechanisms underlying these effects remain to be determined, a shift in the endogenous opioid system of both mother and offspring seems likely. These findings strongly suggest that adolescent female opiate use, occurring prior to mating, and in the absence of any further use pre- or postnatally, can impact maternal–offspring interactions and the behavioral phenotype of their offspring. Thus, concerns about the impact of maternal drug use on children's health, should not only include consideration of in utero exposure, but prior drug history as well.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.