Paternal Deprivation and Female Biparental Family Rearing Induce Dendritic and Synaptic Changes in Octodon degus: II. Nucleus Accumbens

While the majority of studies on the importance of parental caregiving on offspring behavioral and brain development focus on the role of the mother, the paternal contribution is still an understudied topic. We investigated if growing up without paternal care affects dendritic and synaptic development in the nucleus accumbens of male and female offspring and if replacement of the father by a female caregiver “compensates” the impact of paternal deprivation. We compared (a) biparental rearing by father and mother, (b) monoparental care by a single mother, and (c) biparental rearing by two female caregivers. Quantitative analysis of medium-sized neurons in the nucleus accumbens revealed that growing up without father resulted in reduced spine number in both male and female offspring in the core region, whereas spine frequency was only reduced in females. In the shell region, reduced spine frequency was only found in males growing up in a monoparental environment. Replacement of the father by a female caregiver did not “protect” against the effects of paternal deprivation, indicating a critical impact of paternal care behavior on the development and maturation of neuronal networks in the nucleus accumbens.


Introduction
A variety of studies in different species demonstrated that adverse environmental conditions such as socioemotional neglect, abuse, or disturbed family relationships during critical perinatal time periods can interfere with the maturation of brain structure and function [1][2][3][4][5][6]. One of the most important environmental impacts during early life is represented by the infant-caregiver interaction, a dynamic process that implicates specific rhythm, reciprocity, and synchrony in the relationship between both partners [7,8]. Since maternal care behavior is found in nearly all mammalian species, the vast majority of studies focus on the role of the mother in supporting infant growth and stimulating cognitive as well as socioemotional development [3,[9][10][11][12]. However, in up to 5% of mammalian species, additional paternal care and thus biparental care can be observed, mainly in monogamous species such as prairie voles, mandarin voles, karger@karger.com www.karger.com/dne marmosets, and of course humans [13][14][15]. Regarding human family settings, biparental care provides the best suited environment for the offspring, in contrast to being raised by a single caregiver, which may represent an "impoverished" environment due to lack of specific care behavior provided by a second caregiver. Thus, it is surprising that the influence of paternal care on offspring's development and maturation of brain and behavior has been less studied and BR 1692/11-1 not understood in detail. However, an increasing number of studies indicate that paternal care exerts a specific and unique impact on the child's development [13,16]. For example, growing up in a "fatherless" environment can increase an individual's risk for drug abuse, criminal activity, and for the development of mental disorders at later life periods [17][18][19][20][21]. Since there is still a profound lack of systematic analyses on brain development in father-deprived children, a number of animal models in monogamous and biparental rodents have been established to address this aspect under controlled experimental conditions. These studies revealed the important role of biparental care and specifically the critical impact of paternal care behavior on the development and maturation of functional neuronal networks in the offspring [13,14,16,[22][23][24][25]. The trumpet-tailed rat (Octodon degus), a precocious, diurnal South American rodent, has become an established animal model to study the influence of adverse environmental conditions, including disturbed family structure, on the development of prefronto-limbic brain structures and socio-emotional behavior [3,13,16,26]. In the wild, degus live in complex social family structures since they are polygynous, communal breeders [27,28]. Under controlled laboratory conditions, including biparental (mother + father) rearing, degus show high levels of both maternal as well as paternal care and degu pups develop an intense social bond with both parents [29][30][31][32][33]. Disturbance of this social bond, as induced by repeated interruptions of parent-offspring interaction, represents a stressful, adverse environmental challenge for degu pups that critically affects functional brain activity [34] and results in the dysfunctional maturation of synaptic circuits and transmitter systems in prefrontal and limbic brain areas [35][36][37]. These detrimental effects are paralleled by symptoms of mental disorders such as attention deficit hyperactivity disorder (ADHD) [26]. The important role of degu fathers for their offspring's development is revealed by studies showing that paternal care comprises up to 40% of total parent-offspring interactions during the early postnatal period [38,39], and a series of studies also showed that growing up without paternal care significantly affects the development and maturation of a number of prefrontal cortical and limbic brain areas in degus [39][40][41][42][43][44]. One of the areas that appears to be specifically sensitive to the absence of paternal care is the nucleus accumbens (NAc). This is based on our finding that father-deprived male degu offspring showed alterations of catecholaminergic fiber innervation in the NAc [43]. Dysfunctions of the NAc is critically implicated in a number of neurological and mental disorders such as addiction, drug abuse, depression, ADHD, and cognitive dysfunctions [45][46][47]. The NAc, which is subdivided into a lateral core (NAcc) and a medial shell region (NAcs), receives limbic inputs from the hippocampal formation and the amygdala, cortical input specifically from prefrontal cortical areas, and, as a major part of the mesolimbic system, prominent dopaminergic innervation, mainly from the ventral tegmental area [48,49]. However, the NAcc and NAcs display somewhat different innervation patterns and therefore are considered to play distinct roles in a number of behavioral functions. Whereas the core appears to be implicated in spatial learning, conditioning, and the response to motivational stimuli, the shell is mediating reinforcement in response to rewarding substances, novelty, or feeding behavior [45,49]. The most prominent and main projection neuron subtypes within the NAc are the GABAergic medium-sized spiny neurons (MSNs), which are characterized by their relatively high number of excitatory spine synapses [50,51]. The aims of the present study were to analyze (i) if deprivation from paternal care during early childhood may influence dendritic length, complexity, and synaptic density of MSN neurons in the NAcc and NAcs of degu offspring, (ii) if paternal deprivation affects neuronal and synaptic development in male and female offspring differently, and (iii) if the proposed effects on neuronal development, induced by the absence of the father, can be compensated by growing up in the presence of a second female caregiver (the "aunt").

Animals
For the experiments, we used degus (Octodon degus) that were bred in our colony at the Institute of Biology, Otto von Guericke University Magdeburg. Degu families were housed in wire cages (l/ w/h: 50 cm/42 cm/67 cm) on a 12/12-h light/dark cycle. Temperature was set at 22°C (±2°) with humidity at 55% (±5%). Fresh water and rodent diet pellets were accessible ad libitum; vegetables were added occasionally. The animals lived in a complex environment, with hiding places and material for nest building; a running wheel was accessible after the pups in the respective families reached the age of 21 days.

Experimental Groups
Four weeks after mating a male and female breeding pair, an additional female, the sister of the female (aunt), was introduced and housed in the same cage. On the day of delivery (postnatal day [PND] 0), the families were assigned to three experimental groups.
Control Family (+F Group = Male and Female Biparental) On the day of delivery (from the mother), the aunt was removed, resulting in an intact biparental family in which the pups were raised by the mother and father (N offspring = 7 males, 5 females, each from different families to prevent litter effects). The data presented for this group include a total of 56 neurons for males and a total of 40 neurons for females.
Single Mother Family (−F Group = Father-Deprived Monoparental) On the day of delivery, the father and the aunt were removed, resulting in a monoparental family in which the pups were raised by the mother only (N offspring = 6 males, 6 females, each from different families to prevent litter effects). The data presented for this group include a total of 48 neurons for males and a total of 48 neurons for females.
Mother-Aunt Family (MA Group = Female Biparental) On the day of delivery, the father was removed, resulting in a biparental family in which the pups were raised by the mother and her aunt (N offspring = 6 males, 5 females, each from different families to prevent litter effects). The data presented for this group include a total of 48 neurons for males and also a total of 40 neurons for females. All families were group-housed in their home cage until the pups reached adolescence (PND 45).

Quantitative Neuromorphology
Male and female offspring were decapitated on PND 45, brains were dissected out, and unfixed brains were impregnated in Golgi-Cox solution for 14 days. After embedding in celloidin, 150-µm sections were developed by using a modified Golgi-Cox technique (for methodological details, see [52]). For each brain and hemisphere, a total of four MSN neurons were analyzed, two in the NAcc and two in the NAcs, respectively, resulting in an overall of eight neurons per animal. Brain regions were defined according to the rat brain atlas [53] and the degu brain atlas [54] (shown in Fig. 1a). All neurons were reconstructed at a final magnification of ×1,000 using a computer-based neuron tracing system (NEU-ROLUCIDA ® , MicroBright-Field, Williston, VT, USA), allowing quantitative three-dimensional analysis of complete dendritic trees. Neurons selected for analysis had to fulfill the following criteria (representative example shown in Fig. 1b): (1) localization within the boundaries of the NAcc or NAcs, (2) uniform and complete staining of dendritic trees within the 150-μm section, (3) dendrites had to branch regularly into a series of bifurcating branches divided into primary, secondary, tertiary, etc., (4) sufficient distance from neighboring neurons, glia, or blood vessels, which could obscure their morphology. All protrusions, thin, stubby, or mushroom type, were counted as spines if they were in direct continuity with the dendritic shaft. No attempt was made to correct for hidden spines [55] as the use of visible spine counts for comparison between different experimental conditions had been validated previously [56]. The following parameters for each reconstructed neuron were quantified: (i) dendritic length in μm; (ii) spine frequency representing the number of visible spines per 10 μm dendritic length; (iii) number of visible spines; and (iv) dendritic complexity. The length of the dendritic trees was measured by tracing the entire dendrite in three dimensions while counting dendritic spines. For a more detailed analysis, three parallel analysis strategies were performed. (1) To assess whether spine changes are confined to specific areas of the dendritic field, the natural branches (segments) of the dendritic trees were numbered consecutively (primary, secondary, tertiary, etc.) from proximal to distal [44,52] shown in Figure 1c. Spine frequency was calculated (i) as average of the complete dendrite and (ii) as average of the individual branching orders. (2) Since it turned out that the most distal dendritic segments showed the most pronounced differences between the rearing groups, we applied an additional analysis: the values for each parameter (dendritic length, spine frequency, spine number) were pooled for the 2nd-4th branch order. (3) To obtain more detailed information about changes of dendritic complexity, a three-dimensional version of the Sholl analysis [57] was performed in which concentric spheres at 50 μm intervals were placed around the soma and the number of intersections between dendrite and Sholl sphere was calculated (shown in Fig. 1d). All analyses were conducted doubleblind by an experimenter who was unaware of the experimental conditions of the animals.

Statistical Analysis
A two-way ANOVA was performed to test for potential differences with the main factors rearing condition (+F, −F, MA) and sex and for a potential interaction of these factors. For a more detailed comparison of the individual experimental groups, a oneway ANOVA for male and female groups was conducted separately and in case of significance, a Student-Newman-Keuls (SNK) post hoc test was applied, with a significance level set to p ≤ 0.05. As described above, we compared values of spine frequency, spine number, and dendritic length for (i) complete neurons and (ii) 2nd-4th order segments for dendrites. In addition, values for comparison of dendritic complexity for complete neurons were analyzed with a Sholl analysis.

Results
Rearing-Induced Differences in Spine Frequency, Spine Number, Dendritic Length, and Complexity in the NAcc and NAcs Two-way ANOVA for factors rearing condition and sex revealed several significant effects or strong tendencies for the factor rearing condition and interaction of rearing condition × sex, whereas no significant effects were detected for the factor sex (for details, see Table 1). Specifically, in the NAcc, we found that rearing condition significantly affected spine frequency (p = 0.016) and dendritic length (p = 0.05) on average over the entire dendrite and spine number on the pooled 2nd-4th branch order segments (p ≤ 0.01). In the NAcs, significant effects were only found for rearing condition and the parameter spine frequency (p = 0.05). Also, in the NAcs, a Paternal Deprivation and Brain Development a b c d Fig. 1. a Schematic illustration of a coronal section indicating the nucleus accumbens (NAc) with subregions NAc core and NAc shell (modified after Paxinos and Watson, 1982). b Photomicrograph of an MSN in the NAc; inset shows a dendritic segment with dendritic spines. c Schematic illustration of dendritic branching with numbers indicating individual branch orders ("natural" dendritic segments). d Schematic illustration of Sholl analysis with concentric spheres around the soma and respective radii in µm used for analysis of dendritic complexity; arrows indicate intersections of dendrites ("artificial" dendritic segments). significant interaction between rearing condition and sex was found for dendritic length on pooled 2nd-4th branch order segments (p = 0.049) and a tendency for interaction of the two main factors for the parameters dendritic length (p = 0.096) and dendritic complexity (p = 0.055) of entire dendrites. Subsequently, to test for more detailed differences between the individual experimental groups, a one-way ANOVA with an SNK post hoc test was conducted for male and female groups separately. This analysis revealed that rearing conditions affect neuromorphological parameters of MSNs in the NAc and that these effects were similar for males and females in the NAcc. In contrast, in the NAcs, significant effects were only found in male but not in female offspring. The results are described in more detail in the following section.

Male Offspring
NAc Core In the NAcc of male offspring, one-way ANOVA revealed significant differences between the three experimental groups for spine number on the 2nd-4th branch order segments (p = 0.003). Post hoc analysis revealed a significantly lower spine number in offspring of the −F group compared to +F offspring (p = 0.042, 17.1% decrease, shown in Fig. 2e, f). Also, a significantly lower spine number was found in the mother-aunt (MA) group compared to the offspring of the +F group (p = 0.033, 15% decrease, shown in Fig. 2e,f). No significant difference between the experimental groups was detected for spine frequency (shown in Fig. 2a-c), dendritic length, and dendritic complexity (shown in Fig. 2g-k).

NAc Shell
In the NAcs, one-way ANOVA revealed significant differences between the three experimental groups for the averaged spine frequency over the entire dendrite (p = 0.045). Post hoc test revealed a significantly lower spine frequency of offspring from the single mother (−F) group compared to offspring from the biparental (+F) group (p = 0.03, 12.8% reduced, shown in Fig. 3a). No further differences regarding spine frequency or number could be found in the NAcs (shown in Fig. 3b-f). One-way ANOVA also indicated a tendency for differences between the experimental groups for dendritic length of the combined 2nd-4th branch order segments (p = 0.073). Post hoc test revealed that this was due to reduced dendritic length on these segments in animals from the MA group compared to the −F group (p = 0.058, 13.9% decrease, shown in Fig. 3h, i). No further effects on dendritic length or complexity could be found (shown in Fig. 3g, j, k).

Female Offspring
NAc Core In the NAcc of female offspring, one-way ANOVA revealed significant differences between the three experimental groups for spine frequency (p = 0.035) and spine number (p = 0.0015) of the combined 2nd-4th branch order segments. Post hoc analysis revealed a significantly lower spine frequency in the MA offspring compared to the +F group (p = 0.024, 12.1% decrease, shown in Fig. 4b, c) and a strong tendency toward reduced spine frequency in the −F group compared to +F offspring (p = 0.059, 12.8% decrease, shown in Fig. 4b, c). Also, a significantly lower spine number for the −F group compared to the +F group (p = 0.003, 24.7% decrease) as well as for the MA group compared to the +F group (p = 0.002, 22.2% decrease, shown in Fig. 4e, f) was detected. No significant differences between the experimental groups were detected for spine frequency and spine number over the averaged entire dendrite (shown in Fig. 4a, d) and for dendritic length and dendritic complexity (shown in Fig. 4g-k).

NAc Shell
In contrast to the findings in male offspring, no significant difference between the experimental groups could be detected for the analyzed neuromorphological parameters in female offspring (shown in Fig. 5a-k).

Discussion
Research on the impact of the early familial environment on neuronal and behavioral development has traditionally focused on the impact of maternal care [3,[9][10][11][12]. In contrast, the particular role of a second caregiver, which in human family structures traditionally is represented by the biological father, is still just poorly understood. Using the biparental rodent model, Octodon degus, we addressed the question if growing up monoparental without paternal influence may influence the development of MSNs in the NAc of male and female offspring. Moreover, we analyzed if the replacement of the father by a second female caregiver may protect from the expected neuromorphological effects of paternal deprivation. Our results reveal that growing up in a monoparental environment results in reduced dendritic spine number and frequency in the NAc. Moreover, we found that this presumably detrimental effect is not "buffered" by a second female caregiver, indicating a specific role of paternal care behavior for the development of neuronal connectivity in the NAc. Paternal Care Influences Synaptic Development of MSNs in the NAc The NAc, with its two subregions core (NAcc) and shell (NAcs), is a prominent part of the mesocorticolimbic system and thus also a central part of complex forebrain networks that are implicated in diverse functions such as motivation, reward, affect control, stress response, learning, memory, and cognition [49,50]. Approximately 95% of neurons in the NAcc are GABAergic MSNs which primarily express either D1-type or D2-type receptors. Since these receptor-defined subpopulations of MSNs cannot be morphologically differentiated in Golgi-impregnated material, both subtypes were included in our analysis. MSNs are characterized by their relatively high number of excitatory synapses on dendritic spines, indicating that glutamatergic input is an important mediator of the integrative role of the NAc in emotionally modulated behaviors. The reduced dendritic spine density and spine number in the NAc observed in paternally deprived degu offspring most likely reflects reduced excitatory input to the NAc arising from the medial prefrontal cortex (prelimbic, infralimbic), the anterior cingulate cortex, limbic areas such as the (dorsal) hippocampus and amygdala, thalamus, and lateral hypothalamus [45,50,58]. With respect to potential behavioral consequences of these synaptic changes seen in paternally deprived animals, it is interesting to note that the excitatory synapse hypothesis of depression states that changes in the strength of glutamatergic synapses within the corticolimbic reward pathway and specifically on MSNs of the NAc may underlie symptoms of mental disorders, specifically depression [58][59][60][61]. Within the reward circuit, the NAc receives prominent excitatory input from the mPFC and the hippocampus, and specifically, the mPFC exerts top-down control of reward-related behaviors. Our analysis in the same animals revealed significantly reduced spine densities on pyramidal neurons in the mPFC of father-deprived male offspring [44], which together with the reduced synaptic densities in the NAc indicates that paternal care significantly promotes the development of functional neuronal networks within reward pathways.
Major efferent connections of MSNs in the NAc include projections back to limbic areas such as the (ventral) hippocampus, amygdala, BNST, and lateral septum and to diencephalic structures such as the lateral hypothalamus, subparts of the thalamus, habenula, and globus pallidus. Thus, the reduced excitatory input on the inhibitory MSNs should lead to a disinhibition of some of these downstream efferent target regions. This may result, among other symptoms, in impaired locomotor control, a symptom typically associated with ADHD. This view is supported by a recent study in rats, which showed that the NAc directly controls motor behavior and that focal disinhibition of the NAc core results in hyperactivity [62]. In degus, the contribution of parental care to the etiology of ADHD-like behavioral symptoms was observed in a study on the behavioral and brain functional consequences of repeated parental deprivation [26].

Sex-Specific Effects of Paternal Deprivation
Most studies investigating the effects of paternal care on the brain and behavioral development focus on male offspring. However, the present study on MSN neurons in the NAc as well as our previous study on pyramidal neurons in the mPF [44] reveal sex-specific differences of the impact of paternal care on the establishment and refinement of dendrites and spine synapses. Although in the NAcc, the lack of paternal care resulted in reduced spine number in both male and female offspring, spine frequency was only reduced in females. Moreover, growing up in a monoparental environment without father resulted in reduced spine frequency in the NAcs only in males but not in females. This is in line with findings indicating that MSNs in the NAc show sex differences specifically in excitatory synapse number and electrophysiological properties [63,64]. These studies also indicate that sex-specific functions and presumably also sex-specific development of MSNs are affected by sex hormones, specifically estradiol. However, most of the described effects are found in the NAcc. In contrast, our study found the most prominent sex-specific effects in the NAcs. It is tempting to speculate that these sex-specific effects may also be related to differences in the behavioral consequences in male and female offspring due to a suggested dichotomy in the functions of the NAcc and NAcs. Although a number of studies indicate some overlap in the functions of the NAcc and NAcs, it is generally suggested that the subregions of the NAc may have different behavioral functions [45,49,51]. The NAcc is implicated in spatial learning, conditioned responses, and goal-directed behaviors to motivationally relevant stimuli and within this context in the control of impulsive choices. In contrast, the NAcs appears to be more relevant for reward prediction and affective processing, for example, by suppressing actions in response to irrelevant or nonrewarded situations, leading to enhanced efficacy of goal-directed actions. Consequently, the NAcs is also implicated in mediating feeding behavior and responses to rewarding substances and drugs. Detailed behavioral analysis is required in our animals to unveil possible sexspecific behavioral consequences of paternal deprivation.
The impact of paternal deprivation on brain development is not confined to degus but has also been demonstrated in other biparental rodent models. However, most of these studies focus on areas such as the mPFC or the hippocampus, while only few studies analyzed the NAc. For example, a neonatal parental deprivation paradigm in the prairie vole (Microtus ochrogaster) induced a reduction in dopamine receptor 1 and dopamine receptor 2 mRNA expression in the NAc of female offspring, whereas in males the opposite was observed, enhanced dopamine receptor 1 and dopamine receptor 2 expression [65].
A Female Caregiver Cannot "Buffer" the Neuronal Consequences of Paternal Deprivation In human societies, absence of the father or growing up in a monoparental environment without a paternal caregiver often co-occurs with other risk conditions such as poverty, low social class, maternal depression, nutritional deficits, etc. Thus, animal models of paternal deprivation are required to investigate paternal influences on offspring development under controlled experimental conditions. Most of these experimental studies are conducted in biparental or monogamous rodent species by removing the father from the family before birth of the offspring or at varying time points in the early postnatal development. Overall, the outcome of these studies indicates that the absence or loss of the father increases the risk for the development of behavioral dysfunctions including anxiety, aggression, social behavior, and response to reward [65][66][67][68][69]. In the present study, we extend the knowledge about the effects of paternal care by using a well-established animal model, the degu (Octodon degus).
Observations of natural populations of this species revealed that they live in complex social family structures. They are forming colonies that consist of up to 3 males and up to 8 females and exhibit communal rearing and nursing of the offspring [32,70]. Mothers contribute to the rearing of all pups in the colony and express maternal care behavior; however, individual variations are reported since degu mothers in part show discriminative behaviors such as providing more milk to their own offspring [71]. For male degus, it is known that they contribute to nest building and defend their harem against intruders; however, the details of parental behavior in natural populations in the wild are unclear. In contrast, under controlled experimental conditions, it has been repeatedly reported that male degus exhibit paternal care behavior such as huddling and play behavior [31,32,38]. A recent study showed that quantity and quality of paternal care was associated with maternal care; while quantity was negatively correlated, quality was positively correlated [33].
Another aim of our study was to address the view that the observed consequences on neuronal and behavioral development are not specifically caused by the absence of paternal care but rather are the consequence of monoparental rearing. In our experiment, the male degu partner was removed from his pregnant "wive," which reflects an unstable family condition that may influence maternal care to the offspring. Indeed, a recent study in degus revealed that developing females, which experienced unstable social conditions, intensify offspring care, potentially to protect the offspring from negative effects of unstable environments [72]. Enhanced offspring care does not necessarily reflect improved or "better" maternal care but may indicate a more unstructured care behavior with a negative impact on the developing offspring. Interestingly, it has been shown that reduced rates of mother-initiated contacts negatively impact the development of the stress response in degu pups [73]. Moreover, this study also indicated that plural breeding with communal care could buffer some of the negative effects. In line with these findings, we experimentally addressed this issue in more detail by replacing the father with a second female caregiver (the "aunt"), resulting in a female biparental family. The results of the present study on MSNs in the NAc and of our previous study on neurons in the mPFC [44] show that in our degu model, a female caregiver cannot "protect" from or compensate for the neurostructural and synaptic changes induced by growing up in a fatherless environment. Instead, the presence of a second female caregiver induces different changes particularly in the vmPFC of female offspring but fails to compensate for the effects in the NAc in both sexes. Although the socio-emotional environment provided by a female biparental family remains to be investigated in more detail, previous observations of parental care behavior in degus revealed an increase in specific play behaviors between the mother and the pups as well as between the father and the pups over the first 3 weeks of life [38]. In recent (unpublished) observations, we found no evidence for play behavior between the aunt and the pups. Thus, it is tempting to speculate that such an "impoverished" play environment may contribute to the observed neuronal changes in the mPFC and NAc of father-deprived degu offspring.

Conclusion
Together with our previous studies, which revealed that monoparental, father-deprived rearing conditions interfere with synaptic and dendritic development in various prefronto-limbic regions [13,44] the present study provide further support for the hypothesis that paternal care is an important stimulator for developing functional neuronal networks within the reward pathway.