Maternal separation leads to regional hippocampal microglial activation and alters the behavior in the adolescence in a sex-specific manner

Early life adversities during childhood (such as maltreatment, abuse, neglect, or parental deprivation) may increase the vulnerability to cognitive disturbances and emotional disorders in both, adolescence and adulthood. Maternal separation (MS) is a widely used model to study stress-related changes in brain and behavior in rodents. In this study, we investigated the effect of MS (postnatal day 2–14, 3 h/day) in both, female and male adolescent mice. Specifically, we evaluated (i) the spatial working memory, anxiety and depressive-like behavior, (ii) the hippocampal synaptic gene expression, and (iii) the hippocampal neuroinflammatory response. Our results show that MS significantly increased depressive-like behavior in adolescent female mice and altered the spatial memory in adolescent male mice. In addition, MS led to decreased expression of genes related to synaptic function (5ht6r, Synaptophysin, and Cox-2) and induced an exacerbated microglial activation in dentate gyrus (DG), CA1, and CA3. However, while the levels of hippocampal inflammatory cytokines were not modified by MS, they did follow a sex-specific expression in adolescent mice. Taken together, our results suggest that MS induces long-term changes in hippocampal microglia and synaptic gene expression, alters the spatial memory, and induces depressive-like behavior in the adolescent mice, in a sex-specific manner.


Introduction
Adolescence is a critical period for brain development when the brain undergoes different processes (including executive function development, synaptic stabilization and synaptic pruning (Selemon, 2013)) that induce profound emotional and cognitive changes. The hippocampus is a key brain region that regulates memory processes and emotions. Owing to its neuroanatomical connections and high expression of glucocorticoids and mineralocorticoids, the hippocampus is highly vulnerable to early environmental factors, such as stress or immune activation (Hoeijmakers et al., 2014). In fact, early life stress (ELS) exposure during childhood alters hippocampal development and/or its activation, which correlates with mood and memory disturbances found in human adolescents (Chugani et al., 2001;Carrion et al., 2010;Herringa et al., 2013) and animal models (Molet et al., 2014;Harrison et al., 2014).
Over recent years, neuroinflammation has emerged as a potential link between ELS and the emergence of neuropsychiatric disorders in adolescents. In the developing hippocampus, ELS alters microglia (proliferation, morphology and phagocytic activity (Johnson and Kaffman, 2018)) and can prime them, which has been associated with behavioral abnormalities later in life (Catale et al., 2020). Microglia also undergo sex-dependent maturation processes that affect, among other factors, cytokine release which might be linked to the differential response to ELS in adolescence (Schwarz and Bilbo, 2012;Grassi-Oliveira et al., 2016).
In this study, we hypothesized that the previously observed behavioral effects of ELS in adolescents (He et al., 2020) is mediated by increased brain inflammatory responses in mice. We used an established model of ELS, maternal separation (3 h/day, postnatal day 2-14). To study the effect of MS on the brain, specifically the hippocampal inflammatory response during adolescence, we examined: (i) the anxietyand depressive-like behavior, and the spatial memory; (ii) the expression of genes involved in inflammation-induced depression and synaptic dysfunction in hippocampus (Cox-2 (Muller, 2019), 5ht6r (Rasenick, 2016) and Synaptophysin (Cui et al., 2020)); and (iii) the regional hippocampal microglia activation (in DG, CA1 and CA3) and the hippocampal cytokine concentrations in both female and male mice at 6 weeks of age.

Animals
All the experiments were performed following the international guidelines on experimental animal research and approved by the Malm€ o-Lund Ethical Committee for Animal Research in Sweden (Dnr. 5.8. 18-01107/2018).
Four breeding cages with two C57bl/6 female mice bred with one C57bl/6 male mouse (9-12 weeks old) per cage were used in this study. Pups were weaned at P30, and age-and sex-matched wild-type littermates were group-housed (3-5 animals/cage) with bedding material, 12 h light/dark cycle, and water and food provided ad libitum.

Maternal separation
Maternal separation (MS) was performed as described by Teissier et al. (2020). Briefly, 7 female and 9 male pups were daily separated from their dams from postnatal days 2-14 (P2-P14), 3 h per day (09:00 a.m.-12:00 p.m.). MS pups were placed together into a clean cage with extra nesting material (cotton pieces) to keep them warm and with enough distance to avoid vocalized communication with their dams. After 3 h, pups were returned to their dams and kept undisturbed until the following day. Control litters (7 females and 7 males) were handled similarly to the MS pups from P2 to P14. At the end of the separation, there were no significant body weight differences between MS and control mice. At 6 weeks of age, when the experiment was concluded, there were body weight differences between sexes but not due to a MS effect.

Elevated plus maze (EPM)
EPM was done to evaluate anxiety-like behavior. The mouse was gently placed in a closed arm facing the wall and it allowed to explore the maze for 5 min. The number of entries into each arm and the time spent in the open arms were recorded and used to calculate the anxiety index (AI) (Cohen et al., 2008):

. Tail suspension test (TST)
To assess depressive-like behavior, TST was performed. Mice were suspended 50 cm above the floor by the tail using adhesive tape, and the total time spent immobile during the 6 min of the test was quantified. Immobility was considered to be when the mice hung passively, not moving their limbs and body, and completely motionless.

Y-maze test
The Y-maze test was done as previously described (Hansson et al., 2019). Mice were placed at the end of one arm facing the wall and allowed to explore the maze for 5 min. The number of entries was recorded, and the spontaneous alternation was defined as entries into the 3 arms on consecutive choices. The percentage of total alternations was represented.
Hippocampus was isolated from the other half of the brain and snap frozen at -80 C until further RNA or protein isolation.

Immunofluorescence, image acquisition and image analysis
Immunofluorescence was performed as previously described (Bachiller et al., 2018). Free-floating coronal sections were permeabilized using Triton X-100 (Sigma-Aldrich) 1% (v/v) in PBS (PBS-T1%) for 1 h, incubated in the blocking solution (5% Normal Donkey Serum, PBS-T1%) for 1 h and then in primary antibody, anti-Iba1 (Wako, 1:500) at 4 C overnight. Then, sections were rinsed for 1 h with PBS-T0.1% and incubated with the corresponding secondary antibody (1:500, donkey anti-rabbit 647, Invitrogen) for 1 h and finally mounted using ProLong Diamond Antifade Mountant (Invitrogen). Images were taken with a Nikon confocal A1RHD laser-scanning microscope using an 20X air objective with the same laser acquisition parameters and the same researcher, blinded to the groups. Analysis of the images was done using Fiji ImageJ software (W. Rasband, National Institutes of Health). Soma size analyses were performed as described by Hadar et al. (Hadar et al., 2017) in DG, CA1 and CA3. The image background was subtracted and then a threshold was manually set for microglial soma visualization and measurement. At least 2-3 brain sections (bregma -2.0 to -2.5 mm)/region/animal were analyzed.

Statistical analysis
All statistical analyses were performed using GraphPad Prism 8.0 Software for Macintosh (GraphPad Software, San Diego, CA, USA). For the behavioral experiments, at least 7 animals/group were included. Otherwise, at least 4 animals/group were analyzed. We used different unpaired statistical analysis based on the results of the normality and lognormality test (Shapiro-Wilk test). If data was normally distributed, two-way ANOVA followed by Tukey's test for multiple comparisons was performed. For non-parametrically distributed data, Kruskal-Wallis followed by Dunn's test for multiple comparisons was used. Data is reported as mean AE SD. P values 0.05 were considered statistically significant and are stated in the figure legends.

Results
3.1. MS mice display depressive-like behavior and spatial memory impairment in a sex-specific manner Anxiety-like behavior was evaluated using EPM. Our results did not show an effect from MS in the AI (F (1,26) ¼ 1.14; p ¼ 0.2947) (Fig. 1A) between the MS mice and controls (females: p ¼ 0. Using TST, we assessed depressive-like behavior (MS effect: F (1,26) ¼ 4.69; p ¼ 0.0397). Compared to controls, MS females had a significant increase in the percentage of immobility time (p ¼ 0.0249). However, no differences were found between MS males and controls (p ¼ 0.999) (Fig. 1B).
We also evaluated the spatial memory using the Y-maze test (Fig. 1C). Our data showed an overall effect of MS (F (1,26) ¼ 8.75; p ¼ 0.0065) and a significant decrease in spontaneous alternation in the MS male group compared to controls (p ¼ 0.0204) but no differences in MS females (p ¼ 0.1751).

Discussion
Adolescence is a particularly important period when cognitive functions undergo final development and disturbances in the emotional regulation processes potentially predispose the individual to psychiatric disorders.
Our results showed no MS effect on anxiety, whereas depressive-like behavior was increased in MS females but not MS males. Similar results have been described in adolescent female rats but not male rats exposed to chronic social stress (McCormick et al., 2008). However, inconsistent data has been published regarding MS and anxiety-and depressive-like behavior in adolescents (Shin et al., 2016;Banqueri et al., 2017). This can be explained by (i) the use of different MS protocols, which can affect different stages of the brain development; (ii) the lack of comparisons between sexes; (iii) the strain of rodents; and (iv) the context-dependent variability in these behavioral tests.
Nevertheless, our spatial memory results showed an impairment in MS males but not MS females, which is consistent with previous reports (do Prado et al., 2016;Brenhouse and Andersen, 2011). Moreover, it has been demonstrated that the exposure to MS in male juvenile rats but not female juvenile rats alters hippocampal-prefrontal cortex networks, which correlates with spatial memory dysfunction (Reincke and Hanganu-Opatz, 2017). Altogether, our behavioral results reveal a sex-dependent effect. Similar sex-dependent results have also appeared in adolescent human studies where the exposure to ELS contributes to hippocampal dysfunction with deficits in episodic memory and the development of depression (Barch et al., 2019;Colle et al., 2017).
In hippocampus, ELS alters dendritic branching and synaptic plasticity, and decreases hippocampal volume in both adolescents and adults (Vythilingam et al., 2002;McEwen, 2000;van der Kooij et al., 2015). Syp expression is closely related to synaptic plasticity in hippocampus, and downregulation of this protein has been found to be associated with stress-induced depressive-like behavior in rodent models (Reines et al., 2008;Thome et al., 2001) independently of sex (Cui et al., 2020). Likewise, 5ht6r is a widely and highly expressed gene in cognitive regions (Woolley et al., 2004), and its pharmacological inhibition enhances glutamatergic neurotransmission in hippocampus (Dawson et al., 2001), suggesting a role in memory processes. On the other hand, Cox-2 is localized to glutamatergic neurons, and its inhibition alters spatial memory as measured by the Water Morris Maze Test in rats (Teather et al., 2002). Hence, we speculate that the downregulation of 5ht6r, Syp and Cox-2 found in this study might be due to postnatal MS exposition and, in the long-term, alters the spatial memory in adolescent male mice. However, the effects of MS on hippocampal neurobiology in females remain to be determined.
Microglia are central in shaping/pruning neural networks during development. In the adult mouse hippocampus, microglia follow a regional distribution (Choi and Won, 2011;Long et al., 1998), suggesting a role in the regional-specific vulnerability of this area (Jinno et al., 2007). Therefore, we evaluated if the microglial activation followed a region-and sex-specific pattern in the hippocampus of MS adolescent mice. Our results revealed higher microglial activation in the MS mice in DG, CA1 and CA3. Moreover, our analysis showed a sex-dependent difference in microglial activation in DG and CA1 but not in CA3. Anatomically, the hippocampus is affected differentially by ELS; for example, in DG, ELS alters neurogenesis, decreases the number of granule cells (Kozareva et al., 2019), and induces an exacerbated glial activation (Diz-Chaves et al., 2012Reus et al., 2017), whereas in CA1 and CA3, it alters the LTP and the dendritic branching of pyramidal neurons (Shin et al., 2016;Heydari et al., 2019). In our study, the largest effect of MS on microglia activation was found in CA3, a hippocampal region that is crucial to the stress response and that acts as a link to stress-induced neuronal plasticity and memory function (Lisman, 1999). These changes suggest that microglia could be participate in the synaptic remodeling even during adolescence.
On the other hand, the hippocampal cytokine levels were not affected by MS, but sex-differences were observed with overall higher concentrations in male mice. Growing evidence suggests that MS produces longterm effects on the immune system (Wieck et al., 2013;Danese et al., 2007), but only few studies address the importance of sex (Grassi-Oliveira et al., 2016;Avitsur et al., 2013). In adolescent rats, maternal deprivation induced an increase of TNF-α and IL6 in the hippocampus in both sexes, but those authors did not find differences in the IL10 (Stroher et al., 2020), consistent with our results. Contrarily to our results, Grassi-Oliviera et al. (Grassi-Oliveira et al., 2016) described a sex-dependent increase of inflammation in adolescent MS rats. However, they analyzed peripheral cytokines instead of hippocampal ones, and additionally, they looked at different ages compared to our study. Still, unpublished results from our lab show that, in adult mice, MS induces differential cytokine release both peripherally and in the hippocampus, suggesting that age might be an important factor to consider in these approaches.
The developing male brain has more microglia and mast cells and higher levels of inflammatory molecules (Lenz et al., 2018), which may contribute to the sex-related differences observed in the stress response of adolescents and adults (Meagher et al., 2010;Rana et al., 2012). Recently, Speirs and Tronson (2018) have shown sex differences in hippocampal cytokines after LPS administration in adult mice (bioRxiv, non-peer reviewed). They describe a differential cytokine response to LPS with a faster response in females than in males. We speculate that the lack of differences in the hippocampal cytokines between our MS mice and controls might be explained by the age of analysis as the cytokine activation may already have been resolved. Likewise, since the resolution may be faster in females than in males, this could explain the reduction of cytokines in females compared to the overall increase in males.
In summary, this study found that MS long-term strongly impacts changes in hippocampal microglia activation and synaptic gene expression in the adolescence. Also, sex plays a crucial role in the ELS-induced behavioral variations highlighting the significance of including males and females in the analyses to achieve better translational results. Further studies are needed to understand how ELS can affect the male and female brain differently and how this pathogenesis contribute to psychiatric diseases.

Funding
This work was supported by the Olle Engkvist Foundation, the Strategic Research Area MultiPark at Lund University, the Royal Physiographic Society, the Swedish Alzheimer Foundation, the Swedish Brain Foundation, the Crafoord Foundation, the A.E. Berger Foundation, the Swedish Parkinson Foundation and the Swedish Medical Research Council.

Declaration of competing interest
None.