1. Supplementary background
Maternal odor attachment learning
In order for an infant rat pup to survive, it must learn to identify its mother’s diet-dependent odor, so that it can approach and nipple attach to the mother. Natural maternal odor was originally proposed to be a pheromone. However, decades of research have shown that it is learned and odor specific (Galef and Kaner, 1980; Leon, 1992; Teicher and Blass, 1977; Hofer et al, 1976; Pedersen et al, 1982; Moriceau et al, 2009; Sullivan et al, 1990; Logan et al, 2012). Maternal odor learning begins in the womb, via learning of the mother’s amniotic fluid. However, postnatal learning of maternal odor continues in the nest, because the maternal odor is diet-dependent (Leon, 1975, 1992). Since the maternal odor in rats is diet-dependent, pups will show a clear preference for the odor of their mother and another mother equally, due to their identical diets. Published work from our laboratory also demonstrates that pups show a very strong preference for natural maternal odor, whether or not it is the odor of their own mother (Raineki et al, 2010).
If the natural maternal odor is eliminated through a special diet, so that pups never experience the odor postnatally, pups fail to respond to the natural maternal odor as assessed by behavioral and neural responses (Sullivan et al, 1990). Importantly, a novel conditioned odor can reinstate behavioral and neural responses typically induced by natural maternal odor, simply by placing the novel odor on the mother during mother-infant interactions (Sullivan et al, 1990; Teicher et al, 1978). Thus, the neonatal learning system underlying maternal odor attachment learning is so robust, that learning can occur through the simple pairing of an odor and sensory stimulation from the mother (Raineki et al, 2010). Interestingly, a conditioned odor can acquire properties similar to maternal odor, via controlled classical conditioning experiments performed outside the nest without the mother (Johanson and Teicher, 1980; Moriceau et al, 2009; Sullivan et al, 2000a, 2000b). The pairing of a novel odor with warmth, nursing, or sensory stimulation (i.e. stroking to mimic grooming) produces a learned odor preference, so that the odor guides nipple attachment and mother-pup interactions (Brake, 1981; Galef and Sherry, 1973; Pedersen et al, 1982; Wilson and Sullivan, 1994). Interestingly, odor attachment learning can occur in early infancy even when a novel odor is paired with painful stimuli. This likely occurs to ensure pup attachment and survival, even if the caregiver is a source of pain to the infant (Hofer and Sullivan, 2001; Roth and Sullivan, 2005).
The enduring value of maternal odor
Our laboratory has analyzed the role of odors learned during infant odor-shock or odor-stroke conditioning in modulating infant and adult behavior. Following this conditioning, pups display a strong odor preference for this conditioned odor, as the odor induces approach responses in pups, and can control nipple attachment and social interactions with the mother (Camp and Rudy, 1988, Raineki et al, 2010, Roth and Sullivan, 2005; Sullivan et al, 1990, 2000b). Thus, the peppermint-conditioned odor acquires comparable value to natural maternal odor. Furthermore, the conditioned odor appears to retain value into adulthood, although the behaviors it controls change from mother–infant interactions to behaviors important in adulthood. Specifically, the conditioned odor has been shown to rescue later life deficits produced by early-life abuse, such as depressive-like behaviors (forced swim test and sucrose preference test; Sevelinges et al, 2007, 2011). In these previous manuscripts, we determined that the odor must be paired with shock in order to produce an odor that is capable of controlling the infants’ behavior towards the mother and also modulate depressive-like behavior in adulthood. The peppermint odor that was used for the infant conditioning acquired maternal odor qualities only for the animals that received paired odor-shock conditioning, but not for unpaired odor-shock, or odor-only control animals. Importantly, the data also indicated that all learning controls (odor-only, shock-only, unpaired) were not different from naïve animals (no conditioning), suggesting that the factors associated with the experimental manipulations, such as the maternal separation, did not affect that parameters analyzed.
Nevertheless, we have demonstrated that receiving unpaired odor-shock conditioning during infancy, which we believe is a model for unpredictable trauma in infancy, increases anxiety-like behaviors in the adult (Sarro et al, 2014; Tyler et al, 2007). Indeed, our data indicate that infant paired odor-shock is a model that can be used to investigate how predictable early-life adversity may lead to depressive-like behaviors and that unpaired odor-shock is a model that can be used to investigate how unpredictable early-life adversity may lead to anxiety-like behaviors. However, the animals that received unpaired odor-shock conditioning in infancy show no preference for the peppermint odor in infancy. Likewise, the odor has no enduring effects on the unpaired animals’ behavior and/or neural activity, as it has not acquired the value of the maternal odor. The current manuscript demonstrates for the first time that the natural maternal odor rescues depressive-like behaviors following early-life abuse, in the same manner as odors conditioned in early infancy. Furthermore, we expand our findings in the modulation of adult behaviors by early-life abusive attachment cues to social and sexual behaviors.
Assessment of sexual motivation involved a behavioral task that our laboratory has not investigated previously, and because receiving unpaired odor-shock conditioning in infancy could affect the sexual performance in adulthood, we have added supplementary results where we compare the group that received paired odor-shock conditioning in infancy with the learning controls (unpaired and odor only; see below).
Animal models of abusive attachment
Our laboratory employs two rodent models of early-life abuse, which are used to examine the infant response to abuse within the attachment system and the development of later-life neurobehavioral deficits following abuse. The first model is a naturalistic abuse paradigm where the mother handles her pups roughly when provided with insufficient bedding for nest building (Hill et al, 2014; Ivy et al, 2008; Raineki et al, 2010, 2012; Roth and Sullivan, 2005). This impoverished environment results in frequent attempts at nest building, trampling, and rough handling of pups, as well as decreased nursing, however typical weight gain occurs (Raineki et al, 2010; Roth and Sullivan, 2005). The second model uses infant odor-shock conditioning to paradoxically produce an odor that is preferred by infant rat pups (Camp and Rudy, 1988; Haroutunian and Campbell, 1979; Roth and Sullivan, 2005). Importantly, this neutral odor paired with shock in early infancy acquires the same value of natural maternal odor, and can control mother-pup social behavior, despite the association of the odor with aversive shock presentations. Furthermore, associative learning of odor-shock pairings before postnatal day (PN) 10 uses the same neural pathway the infant rat naturally uses to learn maternal odor (Landers and Sullivan, 2012; Moriceau and Sullivan 2006; Raineki et al., 2010). Lastly, odor-shock conditioning provides a more controlled adverse environment when modeling early-life abuse, and allows assessment of changes in the brain based exclusively on aversive stimulation. The simultaneous use of our naturalistic (abusive rearing) and experimentally controlled (odor-shock) models of early-life abuse provides great insight into the mechanisms by which abuse produces enduring neurobehavioral deficits, and how early-life attachment cues acquire their enduring value.
Local field potential recordings
Local field potentials (LFPs) represent a measure of summed or cooperative synaptic activity within the region around a recording electrode. Synaptic activity within a specific region often occurs in a cooperative pattern of oscillations that can be divided into different frequency bands. These specific frequency bands – including theta, beta and gamma – are believed to reflect both different underlying cellular mechanisms and circuit functions depending on the region of interest (Buzsáki, 2006). Below is a brief description of how the specific frequency bands within the LFP oscillation are often referred, with a special focus on the higher frequency oscillations, gamma (35-90Hz), as these are specifically altered in the odor-evoked response of the amygdala in animals with early-life abuse.
Theta
Slow-wave activity or theta oscillations in adults are often associated with endogenous mechanisms of sleep and homeostasis (Steriade et al, 1993; Tononi and Cirelli, 2006), and specifically have been found in rodents during REM sleep as well as during a transient sleep state characterized by synchronized whisker twitching (Vanderwolf, 1969; Nicolelis et al, 1995; Fanselow and Nicolelis, 1999; Gervasoni et al, 2004). Slow-wave activity has also been demonstrated to be critically involved in memory consolidation and synaptic homeostasis (Tononi 2009; Diekelmann and Born, 2010).
Beta
Alongside gamma oscillations, beta oscillations are found during wake states and arousal (Steriade et al, 1993). Often these higher frequency oscillations are thought to be associated with information transfer across brain regions (Buzsáki, 2006; Engel et al, 2001). These kinds of activity are widely observed in sensorimotor regions and related to performing motor actions, such as exploratory behavior (Murthy and Fetz, 1992; Sanes and Donoghue, 1993). Notably, beta oscillations within the olfactory bulb have been associated with odor sampling in rats (Ravel et al, 2003).
Gamma
Most relevant to the present study, gamma frequency oscillations are commonly associated with reverberatory activity in local excitatory-inhibitory circuits during wake states, and are especially sensitive to GABAergic interneuron function in many brain areas (Lasztóczi and Klausberger, 2014; Traub et al, 1996; Buzsáki, 2006; Cardin et al, 2009; Volman et al, 2011; Baldauf and Desimone, 2014) including the amygdala (Sinfield and Collins, 2006). This is interesting since we show an importance of GABAergic function in the mechanisms and consequences of early abusive learning (Thompson et al, 2008). Additionally, gamma oscillations have been associated with cognitive functions such as attention, integration of sensory and multisensory signals, and memory formation (Engel et al, 2001; Jensen et al, 2007). In the amygdala, there are enhanced gamma oscillations in response to learned stimuli (Headley and Weinberger, 2013), and evidence suggests that they may coordinate local amygdala neural activity with activity in other cortico-limbic areas (Bauer et al, 2007).
Significance of finding a specific difference within the Gamma band
While we obtained and compared the full spectrum of oscillation frequencies across the animal conditions, we focused our discussion and presentation of the data on the higher frequency gamma frequency oscillations (35-90Hz) because this was where the animal conditions differed. Thus, the odor-specific enhancement of amygdala gamma oscillations to the odor learned during abusive experience in infancy may reflect long-lasting changes in amygdala GABAergic function. In fact, previous work has demonstrated changes in amygdala paired-pulse inhibition following early-life abusive experience, which also implicates a change in GABAergic circuitry (Sevelinges et al, 2007, 2011; Rincón-Cortés et al, unpublished observations). Furthermore, amygdala GABAergic function undergoes dramatic developmental changes during the period in which the animals used in the present study were exposed to early-life abuse (Thompson et al, 2008; Ehrlich et al, 2013), suggesting that these may have been particularly vulnerable during the manipulation.
2. Supplementary materials and methods
Subjects
Male Long-Evans rats (Harlan Labs) born and bred in our colony were used in the experiments. The animals were housed (polypropylene cages 34 x 29 x 17 cm, wood shavings, ad libitum food and water) in a temperature (20±1°C) and light (6:00-18:00 hours) controlled room. The day of birth was considered PN0 and litters were culled to 12 pups (6 males, 6 females) on PN1. Procedures were approved by the Institutional Animal Care and Use Committee, which follow National Institutes of Health guidelines.
Infant abuse paradigms
Naturalistic abusive mother paradigm. The mother and her pups were housed in a cage with limited nesting/bedding material from PN8-12. Specifically, on the morning of PN8 all pups and the mother were transferred to a clean cage with limited nesting/bedding material that consisted of a 1.2 cm layer of wood shavings. The animals remained in this limited bedding environment until the afternoon of PN12. During this period, the maternal behavior was observed daily for 30 min. The behaviors observed included the time that the mother spent in the nest and nursing (nipple attached, but not necessarily feeding), the frequency of rough handling (i.e. mother aggressively grooming pups, transporting pups by limb), stepping or jumping on the pups, and nest building. Additionally, the frequency of the pups’ vocalizations was also recorded. Similar to our previous data (Raineki et al, 2010, 2012), this limited bedding environment (Table 1) decreased the mothers’ abilities to construct nests, which resulted in frequent attempts at nest building (t(11)=3.91 p<0.003), more time spent away from the nest (t(11)=3.91 p<0.003), an increased frequency of stepping or jumping on the pups (t(11)=2.95 p<0.05), and rough handling of pups (t(11)=2.04 p=0.06). Consequently, pups spent less time nursing (t(11)=4.61 p<0.001) and had increased vocalizations (t(11)=2.61 p<0.03). Despite the reduction in the time nipple attached, being reared by an abusive mother did not lead to a reduction in pups’ body weight at PN12 (t(10)=0.01 p=0.99). We have not yet assessed if animals reared by an abusive mother show a difference in feeding bouts; however, since no reduction in body weight is found, it seems that the pups are not malnourished.
Olfactory classical conditioning paradigm. Beginning at PN8, pups were odor-shock conditioned daily for 5 consecutive days. Pups were removed from the mother, who stayed in the home cage, and were transferred to a different room where they were placed in individual 600 mL beakers and given a 10 min acclimation period. During conditioning sessions, pups received 11 pairings of a 30 sec peppermint odor with a 0.5 mA hindlimb shock during the last 1 sec of odor, with an intertrial interval (ITI) of 4 min. The odor (peppermint, McCormick & Co Inc.) was delivered by a flow dilution olfactometer (2 liters/min flow rate) at a concentration of 1:10 peppermint to air vapor.
Control group. The mother and her pups were housed in a cage with abundant (5-7 cm layer) nesting/bedding material from PN8-12, during which time they were not disturbed. This environment permits the mother to build a nest and spend most of her time inside the nest caring for pups (Table 1).
Infant Y-maze test
At PN13, pups were assessed with a 5-trial Y-maze (start box: 8.5 x 10 X 8 cm; choice arms: 8.5 x 24 x 8 cm) to measure approach responses to the natural maternal odor or conditioned peppermint odor. After 5 sec in the start box, the alley doors were opened and pups were given 60 sec to choose an arm. A response was considered a choice when a pup’s entire body moved past the entrance to the alley.