The role of the father in psycho-affective development is indispensable. Yet, the neurobehavioral effects of paternal deprivation (PD) are poorly understood. Here, we examined the behavioral consequences of PD in the California mouse, a species displaying monogamous bonding and biparental care, and assessed its impact on dopamine (DA), serotonin (5-HT), and glutamate (GLU) transmission in the medial prefrontal cortex (mPFC). In adult males, deficits in social interaction were observed, when a father-deprived (PD) mouse was matched with a PD partner. In adult females, deficits were observed when matching a PD animal with a non-PD control, and when matching 2 PD animals. PD also increased aggression in females. Behavioral abnormalities in PD females were associated with a sensitized response to the locomotor-activating effect of amphetamine. Following immunocytochemical demonstration of DA, 5-HT, and GLU innervations in the mPFC, we employed in vivo electrophysiology and microiontophoresis, and found that PD attenuated the basal activity of low-spiking pyramidal neurons in females. PD decreased pyramidal responses to DA in females, while enhancing responses to NMDA in both sexes. We thus demonstrate that, during critical neurodevelopmental periods, PD leads to sex-dependent abnormalities in social and reward-related behaviors that are associated with disturbances in cortical DA and GLU neurotransmission.
Early attachment experiences with the primary caregiver forge an internal working model for subsequent adult relationships, and parental neglect is notoriously linked to neurodevelopmental and behavioral disorders that may persist through adulthood (Bowlby 1978; Tyrka et al. 2008; Murray et al. 2012). Despite emerging evidence for the impact of parental deprivation (Helmeke et al. 2001; Ziabreva et al. 2003; Bredy et al. 2004; Kaffman and Meaney 2007), the primary focus has been on mother-offspring relationships. Nevertheless, human studies have pointed out that paternal deprivation (PD) impairs psychological and mental development, and increases risk for substance abuse and personality disorders (Grossmann et al. 2002; Jablonska and Lindberg 2007; Lamb 2010). However, the neurobiological substrate underlying PD-induced behavioral and neurobiological impairments remains unknown.
Paternal care or biparental rearing patterns are found only in a minority of mammalian species (Kleiman and Malcolm 1981). California mice (Peromyscus californicus) display permanent father–mother pair bond and extensive biparental investments that have accounted for increased offspring survival and growth (Yogman 1981; Gubernick and Alberts 1987; Ribble and Salvioni 1990; Gubernick and Teferi 2000; Becker et al. 2005; Schradin and Pillay 2005). This species therefore represents a useful model for examining the behavioral and neurobiological impact of PD. Few studies in biparental rodents have suggested that PD impairs cognitive and emotional functions in offsprings, especially in low-resource availability and high-foraging demand conditions (Bredy et al. 2004, 2007). This effect of PD is not surprising as all maternal offspring-directed nurturing behaviors (huddling, licking and grooming, hover-crouching, carrying, and retrieving), except for parturition and lactation, are also performed by the father (Dudley 1974; Gubernick and Alberts 1987), and are uncompensated by the mother in the father's absence (Dudley 1974; Bester-Meredith and Marler 2003; Helmeke et al. 2009; Jia et al. 2009). Paternal behaviors also include involvement in social play and enhanced attention when juveniles are distressed (Wilson and Kleiman 1974; Becker et al. 2005; Bredy et al. 2007; Lambert et al. 2013), which are associated with lower activation of brain structures involved in emotional integration, including prefrontocortical regions (Lambert et al. 2013).
Evidence in monogamous rodents suggests that PD can compromise synaptic development in the medial prefrontal cortex (mPFC) (Helmeke et al. 2009; Pinkernelle et al. 2009). The mPFC is involved in processing socially relevant information (Insel 2003; Wolff and Sherman 2007), and activity in its dorsomedial subregions has been linked to prosocial behaviors (Waytz et al. 2012). DA is known to interact with neurobiological systems regulating social behavior, including the neurohormones oxytocin and vasopressin (Insel 2003; Baskerville and Douglas 2010). Dopamine (DA) and serotonin (5-HT) receptors are highly expressed and co-localized with NMDA receptors in the mPFC, where they regulate synaptic plasticity and neural activity, as well as emotional behavior. In addition, alterations in NMDA receptor subunits have been associated with PD (Bredy et al. 2007). Sexually dimorphic effects of NMDA antagonists have been observed on social and alloparenting behavior (Kirkpatrick and Kakoyannis 2004), and on behavioral functions associated with the reward circuit (Hönack and Löscher 1993; D'Souza et al. 1999) and modulated by gonadal steroid hormones (D'Souza et al. 2003).
In this study, we examined social interaction (SI) and aggressive behavior, as well as the reward response to amphetamine, in adult California mice raised under sole maternal care in comparison to those raised under biparental care. In addition to providing the first description of DA, 5-HT, and glutamate (GLU) innervations in the California mouse mPFC, we characterized the electrophysiological responses to DA, 5-HT, and NMDA in order to correlate behavioral consequences of PD with synaptic function in the mPFC.
Materials and Methods
Adult California mice (Peromyscus californicus) that were used for breeding (first batch) were kindly offered by Dr Michael Meaney (McGill University, Montreal, Canada) or obtained (second batch) from the University of South Carolina Peromyscus Genetic Stock Center. All procedures were undertaken in compliance with the standards and ethical guidelines mandated by the local facility animal care committee of McGill University, the Canadian Institutes of Health Research, and the Canadian Council on Animal Care. Breeding pairs were age-matched (within a 3-month age range), and no siblings were mated. Pairs were housed in standard polycarbonate cages (26.5 cm W × 48.5 cm D × 20.3 cm H) with corncob bedding, kept under standard vivarium conditions (12:12 h light/dark cycle, lights on at 7:30 AM and temperature at 20 ± 2°C, 50–60% relative humidity) and given ad libitum access to food and water.
On postparturition day (PND) 3, litters (1–3 pups/litter) were randomly designated to either be left with only the mother (PD) or with both parents (control, CT), as already described by Bredy et al. (2004). Fathers were then permanently removed from the PD litter. Mother and pups were thereafter left undisturbed until weaning on PND30 (de Jong et al. 2012), except for routine maintenance. CT litters were left undisturbed with both parents until weaning. After weaning, a PD mouse was housed with a PD mouse and a CT mouse with another CT, in same-sex littermate pairings. Because cognitive and behavioral effects of PD can be detected as early as PND 60 (Bredy et al. 2004), we began experiments when offsprings were at least 70 days old, just at the onset of the average age of natal dispersal (in the wild) and maturity (McCabe and Blanchard 1950; Ribble 1992; de Jong et al. 2012). A chronological representation of procedures is depicted in Figure 1.
The SI test was used to assess social behavior in same-sex dyads. The open field (OF) test was conducted for the assessment of locomotor activity and anxiety-like thigmotactic (wall-following) behavior. On the day of testing, mice were acclimated for about 60 min in the behavioral room before the procedures were initiated. The apparatus was cleaned with 70% alcohol and water after each run. The behaviors were recorded, stored, and analyzed as MPEG files using an automated tracking system (Videotrack, View Point Life Science, Montréal, QC, Canada) equipped with infrared lighting-sensitive CCD cameras. The analog signals supplied by the camera were measurements of the luminosity of each point from images scanned point-by-point and line-by-line at the rate of 25 images per second. These signals were transmitted to the videotrack system and digitized on 8 bits by digital analog conversion. Before experiments, animal/image background contrast detection thresholds were calibrated by visual inspection to distinguish different behavioral patterns. Additional offline analyses were conducted by a rater who was blind to the experimental manipulations. All behavioral experiments were carried out from 2:00 PM to 7:00 PM.
OF and SI Tests
For the OF test, each mouse was placed at the center of a white-painted OF arena (40 × 40 × 15 cm) and left to explore the whole field for 5 min of recording. Four identical arenas were used to test 4 animals at a time. Arena assignments were counterbalanced so that all experimental groups were equally distributed to each of the arenas. Each animal underwent a single OF test. The experiment took place under standard room lighting (∼350 lx) by a white lamp (100 W) suspended 2 m above the arena. Locomotor activity was measured by the total distance traveled (cm). Immediately after the OF test, the SI test was carried out to analyze social behavior using a procedure modified from Whatson et al. (1974, 1976), McFarlane et al. (2008), and Cox and Rissman (2011). The test was conducted in the same arenas previously used for the OF test at a different location within the same testing room, so that animals were all habituated to the testing room. Here, we assessed changes in the social behavior of adult California mice that have been reared by both parents, in comparison to those subjected to PD, by examining the behavior of PD mice when paired with a CT or with another PD animal. The test lasted for 20 min in order to evaluate persistent changes in social behavior. On the first session, same-sex unfamiliar conspecifics were match-paired according to rearing condition (PD–PD or CT–CT). On the second session, same-sex unfamiliar conspecifics were mix-paired (PD–CT). Each animal was paired with a CT and a PD in a fixed order, and there was a very short delay between the first and second SI test. No animals met more than once. The frequency and duration of several components of social behavior for each animal from match and mixed pairings were encoded and analyzed offline. These included social investigation, which consisted of sniffing, trailing (following the partner), mounting or crawling under or over the partner, and defensive (targeting the back or flanks) or offensive (targeting the snout) aggression. The time (latency) until the dyad first contacted each other was also measured. Avoidance (evading the partner) and passive contact or side-by-side sitting (remaining close to each other but without actively investigating the partner), as well as auto-grooming, were additionally recorded (McFarlane et al. 2008; Cox and Rissman 2011). As the SI test has also been used to evaluate social anxiety, as well as anxiogenic sequelae of chemical agents or genetic manipulations (Cassano et al. 2010), it is conceivable that PD-induced social impairments observed in this test could result from a potentiation of anxiety-like states and/or secondary to motor activity impairment. We therefore also examined anxiety-like reactivity from thigmotactic behavior measured by the frequency and total duration of central zone (30 × 30 cm) visits in the OF test (Choleris et al. 2001).
As DA is implicated in the adverse consequences of psychostimulant exposure on pair bonding and social behavior, and that drug reward and social behavior may recruit overlapping DA pathways (McGregor et al. 2008; Liu, et al. 2010), we ascertained whether changes in mPFC DAergic neurotransmission would translate into an enhanced long-term sensitization to the locomotor-activating effects of amphetamine. Robust sensitizing response to the locomotor-activating effects of amphetamine that is independent of age and environmental context has been demonstrated in Peromyscus mice (Tanimura et al. 2009). Therefore, d-amphetamine sulfate (Sigma Aldrich, UK) was administered intraperitoneally (1.5 mg/kg) to CT and PD males and females, daily at 2 PM for 7 days, to induce the development of locomotor sensitization. This dose is the minimal dose to optimally induce rapid-onset sensitization in both mice (Harrison and Nobrega 2009) and rats (Bhardwaj et al. 2006). Immediately following each amphetamine injection, each animal was placed in an OF arena (40 × 40 × 15 cm) to which the animals were adequately habituated (previously used for basal locomotor testing and SI test). Locomotor activity (distance traveled, cm) was recorded for 5 min, the enhancement of which was indicative of behavioral sensitization to the psychostimulant's effects. The daily progression of amphetamine-induced locomotor activation was also monitored. The effects of amphetamine in these animals were compared with CT and PD animals receiving once daily injections of saline. The experiments were carried out from 2PM–7PM.
Immunohistochemical Detection of GLU, 5-HT, and DA Fibers
Antibody staining was used to visualize, for the first time in this species, GLUergic, 5-HTergic, and DAergic innervations of prelimbic and cingulate cortices. Four adult California mice (2 males and 2 females) were deeply anesthetized with sodium pentobarbital (80 mg/kg, i.p.) and perfused through the heart with 100 mL of phosphate-buffered saline (PBS: 0.9% NaCl in 50 mM PB, pH 7.4) followed by 200 mL of fixative solution (4% paraformaldehyde in 0.1 mM phosphate buffer, PB, pH 7.4). The brain was removed, postfixed in 4% PFA (1 h, 4°C), and washed in PBS. Coronal forebrain sections (50 μm thick) were cut with a vibratome and processed free-floating as follows. Primary antibodies were monoclonal mouse anti-tyrosine hydroxylase (TH, for DA; Sigma, St Louis, MO, USA); polyclonal goat anti-serotonin transporter (SERT, for 5-HT; Santa Cruz Biotechnology, Santa Cruz, CA, USA); and rabbit antivesicular GLU transporter 2 (VGLUT2, for GLU; Synaptic Systems, Göttingen, Germany). The specificity of these commercially available antibodies has already been described in full (TH: Haycock 1993; SERT: Huang and Pickel 2002; VGLUT2: Takamori et al. 2001). Secondary antibodies used for light microscopic immunocytochemistry were biotinylated goat anti-mouse, donkey anti-goat, and goat anti-rabbit IgGs (Jackson Immunoresearch, West Grove, PA, USA), to be revealed with the 3,3-diaminobenzidine (DAB) or VIP substrate kits from Vector Labs (Burlingame, CA, USA). Species-specific IgGs conjugated to Alexa Fluor-488 or 594 (Invitrogen Corporation, Grand Island, NY, USA) were used for double immunofluorescence and confocal microscopy.
For single immunoperoxidase labeling and light microscopy, all incubations were at room temperature (RT). Sections were immersed 20 min in 3% hydrogen peroxide (H2O2), rinsed in PBS, and preincubated for 2 h in a blocking solution (BS) of PBS containing 5% normal goat serum, 5% normal donkey serum (Vector Labs), 0.5% gelatin, and 0.3% Triton X-100. Sections were then washed in PBS, incubated overnight with either anti-TH, anti-VGLUT2 (1:1000 in BS), or anti-SERT (1:700 in BS) antibody, rinsed in PBS, incubated 2 h with species-specific biotinylated IgGs (1:1000 in BS), and incubated for 1 h in a 1:1000 dilution of horseradish peroxidase-conjugated streptavidin (Jackson). After washes in PBS, labeling was revealed 1–2 min with the DAB (SERT) or VIP (TH, VGLUT2) peroxidase substrate kit. The reaction was stopped in distilled water, the sections rinsed in PBS, air-dried on gelatin-coated slides, dehydrated in ethanol, cleared in toluene, and mounted with DPX (Fluka, Sigma-Aldrich, Oakville, ON, Canada). The prefrontal cortex (PFC) was carefully examined at low (×1.6 objective) and high (×6.3 objective) magnification under a Leitz Diaplan microscope couple to an Olympus DP21 color digital camera and software (Olympus Corporation, Tokyo, Japan).
For double immunofluorescent labelings and confocal microscopy, all sections were preincubated 2 h in BS as above. For double GLU/DA immunofluorescence, sections were incubated overnight with both anti-VGLUT2 and -TH primary antibodies (1:1000 in BS). After washes in PBS, sections were further incubated for 3 h (light protected) with a mix of species-specific Alexa Fluor 488 (VGLUT2) and Alexa Fluor 594 (TH) conjugated IgGs (1:200 in BS). For double GLU/5-HT immunofluorescence, sections were incubated overnight with the anti-SERT primary antibody (1:700 in BS). Subsequent incubations and washes were light protected. After washes in PBS, sections were then incubated for 3 h with Alexa Fluor 594 conjugated anti-goat IgGs (1:200 in BS). After washes in PBS, sections were further incubated overnight with the anti-VGLUT2 antibody (1:1000 in BS), and then 3 h with Alexa Fluor 488-conjugated anti-rabbit IgGs (1:200 in BS) following washes in PBS. For both double labeling experiments, after rinses in 0.1 M PB and distilled water, sections were mounted in Vectashield medium for fluorescence (Vector) and observed with a Leica TCS-SP1 confocal microscope using sequential laser analysis (×100 magnification). Images of 1024 × 1024 pixels were produced with the Leica confocal software (v2.65) and adjusted for framing and contrast with the Adobe Photoshop software.
In Vivo Electrophysiology and Microiontophoresis
We aimed at determining whether PD could significantly influence DA, 5-HT, and GLU neurotransmission in the mPFC. To do so, we assessed the response of mPFC pyramidal neurons to these neurotransmitters, indicating the sensitivity and function of their attendant receptors. Electrophysiological experiments were conducted between PND 70 and 105, with animals weighing 30–40 g. The electrophysiological procedures were previously described in Bambico et al. (2010a, 2010b) and Gobbi et al. (2001). Briefly, animals were anesthetized (1.2 g/kg urethane, intraperitoneal). Additional doses were delivered during the procedure to maintain full anesthetic state. Animals were fixed to a stereotaxic frame (David Kopf Instruments, Tujunga, CA, USA), the scalp overlying the rostral medial skull was removed, and a burr hole was drilled through the skull around the area of bregma + 2.3 mm. A glass multibarrel recording pipette (Protech International, Inc., TX, USA) was then lowered into the mPFC (0.5–2.5 mm from the brain's surface) to record pyramidal neurons in the dorsomedial (cingulate and prelimbic) region. These neurons were identified based on their steady response to standard short pulses of NMDA (5 nA) and by large amplitudes (0.5–1.2 mV), long durations (0.8–1.2 ms), and single-action potential patterns alternating with complex spike discharges (Barthó et al. 2004; Gobbi and Janiri 2006; Labonte et al. 2009). Since there is no published stereotaxic atlas of the Peromyscus brain, which is considerably larger than that of Mus musculus, stereotaxic coordinates were approximated by integration of coordinates from the mouse (Paxinos and Franklin 2007) and the rat (Paxinos and Watson 2007) stereotaxic atlases. The location of the last recording was verified histologically in each animal by Pontamine sky blue microinjections. The recording barrel was filled with Potamine sky blue, and adjusted to a signal impedance of 0.5–4 MΩ by breaking the tip. A 2 M NaCl solution acted as an automatic balance and was leaked against a 50 MΩ impedance. Drug barrels were filled as follows: DA (50 mM in 200 mM NaCl, pH 4–5) was ejected against an impedance of 150 MΩ, retained with a −5 nA current, and ejected as a cation. NMDA (50 mM in 200 mM NaCl, pH 8) was ejected against an impedance of 125 MΩ, retained with a 10 nA current, and ejected as an anion. 5-HT (50 mM in 165 mM NaCl, pH 4) was ejected against an impedance of 150 MΩ, retained with a −5 nA current, and ejected as a cation. When a neuron displaying a stable firing activity was encountered, the spontaneous firing rate was recorded for at least 5 min, and the first minute was excluded from analysis to avoid artifacts. Each drug was then ejected for at least 30 s, and then withdrawn to re-establish the baseline. Upon drug current withdrawal, the firing rate normally reverts to baseline. Drug-induced changes in firing rate (percentage increase/decrease from predrug baseline activity) was calculated as the difference between the mean firing rate during the ejection phase and the baseline mean firing rate (at least 30 min prior to ejection) divided by the baseline mean firing rate. Neuronal responses were sampled across 4–6 drug currents. All electrophysiological recordings were conducted from 2 to 10 PM. Forty-eight animals (33 males + 15 females) underwent electrophysiological procedures. Spontaneous basal pyramidal firing rate data were obtained from all of these animals (4–7 electrode descents per animal = 307 neurons recorded from males and 196 neurons from females). Some animals representing each group (6–8 animals per group) were used for iontophoretic drug applications. Six to 19 neurons were recorded per group. Two to three transmitters (1 drug current response each) were tested per neuron. As animals' ages were from PND 70 to PND 105, a correlational analysis between firing rate and age of animals was done, as well as maximal excitatory/inhibitory response to drugs and age, without reaching any significant correlations. We therefore did not proceed running covariate analyses. Moreover, extensive studies in our laboratory suggest that monoaminergic firing activity within this age range is not significantly influenced by age in mice (Gobbi et al. 2001; Bambico et al. 2010a).
Data are presented as mean ± standard error of the mean (SEM). All data were analyzed using Sigma Plot version 11 (Systat Software, Inc., San Jose, CA, USA) and SPSS version 17 (SPSS, Inc., Chicago, IL, USA) softwares. Following tests for assumptions of data distribution normality (Shapiro–Wilk and Kolmogorov–Smirnov tests) and for variance homogeneity, datasets were accordingly submitted to one-way or two-way analyses of variance (ANOVAs) with pairing scheme (match-paired vs. mix-paired) and rearing condition (PD vs. CT) as between-group factors, when assumptions were satisfied. Otherwise, the nonparametric counterpart of these tests were carried out. Electrophysiological data were submitted to Student's t-test or Mann–Whitney U-test for between-group comparisons or to one-way or two-way mixed-design ANOVA (rearing condition × current) when accounting for repeated microiontophoretic current ejections. Tukey's honestly significant difference test was used for multiple post hoc comparisons. Probability value of P ≤ 0.05 was considered statistically significant.
PD Impairs Dyadic Social Interactions in Both Sexes
Figure 2 shows that among males, a dramatic reduction of social investigation was observed only when a PD male was matched with another PD male (P < 0.01), but not when matched with a CT male (two-way ANOVA: pairing, F1,64 = 6.574, P = 0.013; rearing, F1,64 = 16.240, P < 0.001; interaction, F1,64 = 10.570, P = 0.002; Fig. 2A). The first mutual contact of PD–PD pairs had significantly longer latencies in comparison to CT–CT or CT–PD pairs (P < 0.05) (Kruskal–Wallis, followed by Mann–Whitney U-test, P = 0.011; Fig. 2C). In females, however, a reduction of the total time spent in social investigation was measured in all PD animals, regardless of whether they were paired with another PD female (PD–PD) or a non-deprived control (PD–CT) (two-way ANOVA: pairing, F1,74 = 0.376, P = 0.542; rearing, F1,74 = 176.188, P < 0.001; interaction, F1,74 = 1.358, P = 0.258; Fig. 2B). Furthermore, the latency to the first mutual contact was prolonged in both CT–PD and PD–PD female pairs in comparison to CT–CT pairs, although achieving significance in PD–PD pairs only (P < 0.05) (Kruskal–Wallis, df = 2, P = 0.028; Fig. 2D).
The reductions in the duration of social investigation observed in all affected animals were not due to increased avoidance or active movement away from the partner (Fig. 2E,F). Rather, passive contact with the partner (physical proximity devoid of any active, investigational behavior) was particularly increased in males (Fig. 2G,H) (PD–PD vs. CT–CT or vs. PD–CT, P < 0.001; two-way ANOVA: pairing, F1,64 = 7.842, P = 007; rearing, F1,64 = 19.997, P < 0.001; treatment, F1,64 = 19.042, P < 0.001) and females (PD–PD vs. CT–CT or vs. PD–CT, P < 0.05; two-way ANOVA: pairing F1,74 = 0.619, P = 0.434; rearing, F1,74 = 116.821, P < 0.001; interaction, F1,74 = 1.008, P = 0.319).
PD Increases Aggressive Behavior in PD Female Dyads
Assessment of offensive and defensive behaviors in antagonistic encounters revealed that PD females in a matched PD–PD pair exhibited dramatically longer duration episodes (P < 0.001) of aggressive behavior than any other pairs (Fig. 3; two-way ANOVA: pairing, F1,74 = 3.584, P = 0.062; rearing, F1,74 = 11.399, P = 0.001; interaction, F1,74 = 6.893, P = 0.011).
PD does not Affect Anxiety-Like Thigmotaxis and Locomotor Activity in the Open Field, but Induces Sensitization to the Locomotor-Activating Effect of Amphetamine in Females
Contrary to the hypothesis that social behavioral impairments in the SI test may be confounded by anxiety-like reactivity, we found no significant effects of PD in the anxiety-like thigmotactic behavior and locomotor or ambulatory activity (distance traveled) in the OF test (Fig. 4A,B). At best, there was a trend toward a lower locomotor activity in PD males compared with CT males (Tukey's test, P = 0.068; Fig. 4A) and an attenuation of the central zone duration (thigmotaxis) in PD compared with CT mice (two-way ANOVA: rearing, F1,40 = 3.66, P < 0.063; Fig. 4B). However, when PD was associated with psychostimulant amphetamine treatment (Fig. 4C,D), a significant enhancement of ambulatory activity was observed only in females most detectable on the seventh day of locomotor testing (P = 0.029), indicating a sensitized response of the DA system in a sex-specific manner (Fig. 4D).
GLU, DA, and 5-HT Fibers Coexist in the California Mouse mPFC
Following single immunoperoxidase labeling for photon microscopy and double immunofluorescence for confocal microscopy, the 3 transmitter systems (GLU, DA, and 5-HT) were detected in the California mouse forebrain (Fig. 5). As illustrated in Figure 5A (left), we demonstrate that VGLUT2 immunoreactivity is strong throughout the forebrain and denser in neocortical layers III–VIa, a distribution also observed in the mPFC (Fig. 5A, middle). At high magnification, VGLUT2 immunofluorescence appears punctate, which is characteristic of axon terminal labeling (Fig. 5A, right). TH immunoreactivity appeared particularly strong in the striatum, the major target of mesencephalic DAergic projections, and much more diffuse in the cerebral cortex (Fig. 5B, left). Despite being present in all cortical layers, TH-positive fibers appeared denser in neocortical layer VI, particularly in the mPFC (Fig. 5B, middle). As detected following anti-SERT immunohistochemistry, 5-HT fibers were found throughout the forebrain (Fig. 5C, left), particularly in the neocortex where they formed a denser network within layer IV. In the mPFC, however, SERT-positive axons were uniformly distributed (Fig. 5C, middle). At high magnification, both DA and 5-HT fibers (Fig. 5B,C, right) appeared as varicose axons intermingled with GLUergic nerve endings. Noteworthy is the fact that no co-localization was observed between the three markers in the mPFC.
PD-Induced Behavioral Impairments are Associated with Altered Glutamatergic Transmission in the mPFC
The basal discharge activity of mPFC (cingulate and prelimbic) pyramidal neurons is illustrated in Figure 6 (left). The recorded neurons were subdivided into low-spiking (firing rates < 20 Hz) and high-spiking (firing rates > 20 Hz) according to previously established criteria (Barthó et al. 2004; Gobbi and Janiri 2006). The high-spiking group corresponds to putative interneurons; the low-spiking group corresponds to putative pyramidal neurons (Swadlow 2003; Barthó et al. 2004; Gobbi and Janiri 2006) that were further tested in the microiontophoresis experiments. In male PD mice, when compared with CTs, no significant alterations in basal firing rates of both low-spiking and high-spiking neurons were found (Fig. 6). However, low-spiking pyramidal neurons from PD females showed significantly lower firing rates than in CT females (t = 329, P = 0.008).
In order to assess the functional integrity of GLU neurotransmission in the mPFC, we evaluated the sensitivity of pyramidal neurons to NMDA that was administered microiontophoretically (Gobbi and Janiri 2006). Under control conditions, the responses to increasing microiontophoretic infusion of NMDA were higher in female than in males (Fig. 7, upper panel). However, in comparison to their respective CTs, the excitatory response of mPFC pyramidal neurons to increasing NMDA currents was significantly enhanced in both PD males (two-way ANOVA: rearing F1,100 = 9.680, P = 0.002; current, F3,100 = 0.595, P = 0.620; interaction, F3,100 = 0.453, P = 0.982) and PD females (two-way ANOVA: rearing F1,36 = 18.661, P < 0.001; current, F3,36 = 7.234, P < 0.001; interaction, F3,36 = 0.066, P = 0.977) (Fig. 7, lower panel).
PD does not Affect Serotonergic Neurotransmission in the mPFC
In order to further explore the neurobiological substrates of the above-described behavioral deficits, we investigated whether 5-HT transmission in the mPFC was altered by PD. Microiontophoretic application of 5-HT in the mPFC mostly produces inhibitory neuronal responses that rely on 5-HT1A receptor activation (Labonte et al. 2009). We assessed the inhibitory responses of 5-HT and found neither between-group (CT vs. PD) nor sex differences (Fig. 8), suggesting that PD-induced behavioral impairments do not depend on 5-HT transmission in the mPFC.
PD-Induced Behavioral Deficits are Associated with Abnormal Synaptic Processing of Dopamine in the mPFC
In males, low DA currents (5–30 nA) slightly decreased baseline mPFC pyramidal activity, but with no significant difference between CT and PD males (Fig. 9). In females, however, microiontophoretic ejections of DA produced a current-dependent increase in pyramidal activity. Most interestingly, significantly lower excitatory responses were recorded in mPFC neurons of PD females compared with CT females (two-way ANOVA: rearing, F1,42 = 6.825, P = 0.012; current, F3,42 = 4.272, P = 0.01; interaction, F3,42 = 0.849, P = 0.475; Fig. 9).
In the present study, we demonstrate that PD in California mice, particularly during critical developmental periods, leads to impaired social and behavioral functions in adults. PD-induced behavioral impairments were more pronounced in females and were associated with modifications in DA and GLU neurotransmission in the mPFC, an area long propounded as essential to normal psycho-affective and social development.
PD particularly increased initial mutual contact latencies and aggressive episodes in females. Social investigation was also impaired, as indicated by significant reductions in body sniffing, trailing, and crawling-over. Our data suggest that these social impairments were neither a result of alterations in impulse control nor from increased anxiety, because anxiety-like behaviors and locomotor activation in the OF test were unaffected by PD. Social deficits were also unlikely to have resulted from a dysfunctional switching to evasive behavior because PD failed to influence active social avoidance, while dramatically increasing passive contact. PD animals, notably females, were also sensitized to the locomotor-activating effects of amphetamine, suggesting that DA neurotransmission might contribute to impairments in social behavior in female PD mice. These data are of particular interest since, in contrast, repeated maternal separation, which mimics early-life deprivation, has shown controversial and equivocal outcomes on different measures of stress, cognition, social behavior, and hypothalamic–pituitary–adrenal axis regulation (Millstein and Holmes 2007; Hulshof et al. 2011).
An interesting facet of our data is that, under the testing conditions adopted (same-sex dyad; neutral, nonhome cage environment), the observed social impairments were more profound in females. PD females exhibited most behavioral abnormalities in the SI test regardless of their pairing with another PD or CT female. In contrast, PD males showed reduced social investigation and increased passive contact only when matched with another PD mouse. These observations appear to be consistent with those observed among human patients with social cognitive and personality disorders (such as antisocial personality disorder), whereby symptoms may be maintained by mutually reinforcing SIs (Russell and Hersov 1983). We also observed dramatically greater episodes of aggressive behavior in match-paired PD–PD females than in other pairings, suggesting that females are more sensitive to PD. It has been reported that paternal behaviors occurring later during the preweaning stage, such as pup retrieving and grabbing but not huddling and grooming, positively correlate with aggressiveness (Bester-Meredith et al. 1999; Bester-Meredith and Marler 2001, 2003; Frazier et al. 2006). The fact that the observed increase in offensive behavior was exhibited only by PD–PD matched females suggests that these behaviors were most likely aggressive rather than play-fighting. In keeping with these findings, Trainor et al. (2010, 2011, 2013) showed that, under a significant stress load, female California mice unlike their male counterparts show a decrease in SI and an increase in aggressive behavior. On the other hand, Frazier et al. (2006) found that experimentally increasing paternal retrieval behavior leads to increased territorial aggression, with shorter attack latencies in both male and female offsprings in the resident-intruder paradigm. However, this effect was within a normal range of variation and was not observed in a neutral environment. Decreased paternal grooming also led to enhanced levels of corticosterone.
It is also conceivable that PD-induced female aggression and social deficits are exacerbated by a lack of social play stimulation in the absence of the father. Social play, including play-fighting (boxing/wresting and pinning), which peaks before early puberty, contributes to the development of adult social competence in rodents (Bell et al. 2009, 2010). Although, both the father and the mother may participate in social play (Wilson and Kleiman 1974), fathers exhibit greater propensities for it than mothers, as demonstrated in rodents (Guerra et al. 1999; Becker et al. 2005) and primates (Redican and Mitchell 1973). Parental absence or low parental licking and grooming experience results in altered play behavior throughout the weaning period, but appears to drive male rodent offsprings to compensate in social play by increasing initiations and thus solicit more playful interactions (Becker et al. 2005; Parent and Meaney 2008).
Some effects observed in offsprings may not necessarily be directly related to the absence of the father, but indirectly to the stress imposed on the mother by the father's absence (Gubernick and Alberts 1987; Gubernick and Teferi 2000). Our data do not rule out this possibility. We note, however, that from the delivery to the weaning of pups we did not observe any alteration in the mothers' overt behavior that might be consistent with stress-induced effects (data not published, behavior was observed daily for 30 min, comparing gross behavior in single mothers vs. coupled mothers). These effects may therefore occur under conditions of low resource availability and high-foraging demand, which have been shown to affect parenting strategies and offspring development (Bredy et al. 2007). Remarkably, as mentioned before, California mouse females do not typically compensate for mate (father) absence with increased maternal care (Dudley 1974; Bester-Meredith and Marler 2003), suggesting that the impact of altered maternal behavior over PD animals is minimal.
Overall, the observed PD-induced behavioral deficits are consistent with epidemiological studies in children raised without a father, highlighting an increased risk for deviant behavior and criminal activity, substance abuse, impoverished educational performance and mental illness. Accordingly, the direct engagement of the father with the child, including playful interaction, has been found to reliably predict later positive childhood outcomes, intellectual and linguistic competence, or decreases in adolescent delinquency even after accounting for the mother's contributions (Coley and Medeiros 2007; Sarkadi et al. 2008; Cabrera et al. 2009; Lamb 2010). A positive father–child interaction in early life or during adolescence has been documented to influence children's prosocial development, predicting later popularity, social skills, peer relationship competence, and spousal relationship adjustment (Gottman et al. 1997; Lieberman et al. 1999; Rah and Parke 2008; Lamb 2010). The notion that females, compared with males, are more profoundly affected by low paternal care has also been supported by a stronger association with adult antisocial personality traits (Reti et al. 2002).
Our anatomical investigations revealed dense innervations of GLUergic, 5-HTergic, and DAergic fibers in the prelimbic and cingulate regions of the mPFC. 5-HTergic fibers are uniformly dense throughout the mPFC, while DAergic fibers are most dense in layer VI. This suggests that the activity of mPFC pyramidal neurons is greatly regulated by GLU, DA, and 5-HT transmission via their postsynaptic cognate receptors expressed on these neurons. A significant perturbation in the signaling of these transmitters is a parsimonious explanation for the observed aberrations in the sensitivity of these receptors examined here. Indeed, our electrophysiological findings suggest that impairments in social behavior and potentiation of aggressive behavior are underpinned by modifications in the function of both DA and GLU, but not of 5-HT, synapses in the mPFC. The lack of effect on 5-HT is consistent with the lack of effect on anxiety-like behaviors, in which 5-HT plays a major modulatory role. On the other hand, cortical DAergic and GLUergic systems are engaged in a close functional crosstalk, and evidence for a role of DA on social behavior and aggression has accumulated over the years (Baskerville and Douglas 2010). As DA and NMDA receptors also exert organizational functions in brain development, aberrations of these transmitter systems at any critical time point, following early-life stress and/or parental neglect, could persistently perturb their integrity in adulthood and consequently their behavioral response to social stimuli (Frederick and Stanwood 2009; Baskerville and Douglas 2010).
PD-induced reductions in mPFC pyramidal response to DA, which were evident in females, occurred in parallel with significant deficits in social behavior and increases in aggressive behavior, and more notably with an enhanced sensitization to amphetamine. The concordance of these behavioral and neurobiological abnormalities supports the fact that the mPFC may serve as a central junction where DA signals conveying stress adaptation, social motivation and reward are simultaneously processed, and that disruption of one pathway may disturb the others (Aragona et al. 2007). The decreased basal and DA-evoked pyramidal excitability in PD females points to a functionally hypoactive mPFC that may be argued to undermine both prefrontal facilitation of prosocial behavior and inhibitory control over drug-seeking behavior. Such hypoactive mPFC may have been brought about by a desensitization/downregulation of pyramidal D1 receptors and/or upregulation of pyramidal D2 receptors, thus disturbing the balance of D1 and D2 activity that exert opposite effects on neural excitability (Tseng and O'Donnell 2004). Indeed, pharmacological activation of D1 receptors in the PFC, but not in the striatum, enhances social cognition (Di Cara et al. 2007; Loiseau and Millan 2009), while intra-mPFC infusion of D1 antagonists increases nucleus accumbens-dependent stress response (Doherty and Gratton 1992). Likewise, early social isolation, as well as psychostimulant challenge, have been found to increase 1) D2 receptor expression in mPFC (Han et al. 2012), 2) D2 intracellular activation with greater effects in females (Sun et al. 2010), and 3) impairments in D2 function as observed in PFC of psychostimulant-sensitized rats (Kroener and Lavin 2010). In studies employing the social defeat paradigm, in vivo electrophysiological recordings revealed an increased discharge activity in DAergic neurons of the ventral tegmental area from stress-vulnerable animals, an effect also associated with a socially withdrawn phenotype (Krishnan et al. 2007; Cao et al. 2010). Interestingly, PFC DAergic transmission between male and female CTs is clearly different, and that PD in females decreases DA receptor sensitivity to the level of male CTs. This difference between males and females might be explained by sex-dependent variations in D1 and D2 receptor density in the mPFC, as suggested by D1-like receptor binding being greater in females than in males in the mPFC of monogamous prairie voles, while D2-like receptor binding being greater in males (Smeltzer et al. 2006).
Finally, we noted that the excitatory response of mPFC neurons to NMDA was significantly enhanced in PD animals. This response profile likely reflects a compromised integrity of pyramidal neuronal synapses because the basal firing rate of the low-firing subpopulation of mPFC pyramidal neurons were significantly reduced in females, which appeared particularly vulnerable to PD in the SI test. Indeed, sensitized GLUergic activity or upregulation of NMDA receptor subunits in the mPFC are induced after glucocorticoid administration, following chronic stress (Martin and Wellman 2011) or by acute or chronic psychostimulant exposure (Russo et al. 2010). Changes in NMDA receptors also correlate with synaptic remodeling and dendritic atrophy in the mPFC of stressed animals (Martin and Wellman 2011). The hypothesis that neuroplastic changes may result from PD is additionally supported by findings of significant reductions in apical spine length and number of pyramidal neurons in the PFC of Octodon degus, another monogamous murid species (Ovtscharoff et al. 2006; Helmeke et al. 2009). The sex-dependent impact of PD on mPFC behavioral and neurophysiological integrity seems to diverge from recent findings regarding the lack of sexual dimorphism in the PFC of monogamous rodents (Kingsbury et al. 2012), although dimorphism in chemoarchitecture cannot be ruled out. As such, California mice may be far from displaying sex differences in emotional reactivity and in response to stress (Kingsbury et al. 2012). However, some monogamous species like the prairie vole do exhibit sexually dimorphic structural, hodological, or functional features of the PFC (Kingsbury et al. 2012), as may be the case with California mice. Indeed, female California mice relative to males exhibit glucocorticoid hyper-responsiveness (Trainor et al. 2010, 2011, 2013). This could be attributed to extrahypothalamic structures modulating the neuroendocrine response to stress. Indeed, along with the hippocampus, the prelimbic and cingulate mPFC are known to inhibit the hypothalamic–pituitary adrenal axis and negatively regulate corticosterone release (Radley et al. 2006), and that this brain stress axis is particularly sensitive in female California mice, an area that could be a subject of further investigation.
In summary, our results emphasize the importance of the father during critical neurodevelopmental periods, and that father absence induces impairments in social behavior that persist into adulthood, deficits which are associated with sex-dependent dysregulation of neurotransmission in the mPFC.
This work was supported by salaries from the Fonds de la Recherche en Santé du Québec (FRSQ) (G.G., F.R.B.) and fellowships and fund from the McGill University Health Center, McGill University Faculty of Medicine (G.G., F.R.B.).
A special thanks to Dr Michael Meaney who offered the California mouse colony to us, and Dr Guy Debonnel for his encouragement and scientific support. We extend our gratitude to Dr Sergio Dominguez-Lopez and Ms Rebecca Howell for their assistance during the immunohistochemical experiment, and to Dr Jose N. Nobrega and Mr Jean-Philippe Garant for their valuable feedback and suggestions for the manuscript. Conflict of Interest: None declared.