Dynamic changes in hippocampal microglia contribute to depressive-like behavior induced by early social isolation

Neuropharmacology 135 (2018) 223e233

Contents lists available at ScienceDirect

Neuropharmacology journal homepage: www.elsevier.com/locate/neuropharm

Dynamic changes in hippocampal microglia contribute to depressivelike behavior induced by early social isolation Yu Gong a, 1, Lijuan Tong a, 1, Rongrong Yang b, **, 1, Wenfeng Hu a, Xingguo Xu b, Wenjing Wang a, Peng Wang a, Xu Lu a, Minhui Gao a, Yue Wu a, Xing Xu a, Yaru Zhang a, Zhuo Chen c, Chao Huang a, * a

Department of Pharmacology, School of Pharmacy, Nantong University, #19 Qixiu Road, Nantong, Jiangsu Province, 226001, China Department of Anesthesiology, Affiliated Hospital of Nantong University, Nantong, Jiangsu Province, 226001, China Invasive Technology Department, Nantong First People's Hospital, The Second Affiliated Hospital of Nantong University, #6 North Road Hai'er Xiang, Nantong, Jiangsu Province, 226001, China b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 25 October 2017 Received in revised form 18 March 2018 Accepted 19 March 2018 Available online 21 March 2018

Depression triggered by early-life stress has begun to attract wide attention due to its severe symptoms and poor treatment outcomes. However, the pathophysiological mechanism for this type of depression remains unclear. Recently, we and others reported that different types of chronic stress induce a significant loss of hippocampal microglia, which is mediated by an initial activation of these microglia. Since early-life stress also promotes microglial activation, we investigated the dynamic changes in hippocampal microglia in mice suffering from depression induced by early social isolation (ESI). Results showed that 8 days of ESI induced depressive-like behaviors in a tail suspension test, forced swim test, sucrose preference test, and open field test, and it also induced a loss and dystrophy of hippocampal microglia. We found that this ESI-induced loss of hippocampal microglia was mediated by both microglial activation and apoptosis. This was demonstrated by the following results: (i) 1 day of ESI induced an obvious activation of hippocampal microglia followed by their apoptosis, and (ii) the blockade of the initial activation of hippocampal microglia by minocycline pretreatment suppressed the ESI-induced apoptosis and loss as well as ESI-induced depressive-like behavior. Lipopolysaccharide (LPS) and macrophage colony-stimulating factor (M-CSF), two activators of microglia, almost completely reversed ESI-induced depressive-like behavior by promoting microglial proliferation in the hippocampus. These results reveal an etiological role of hippocampal microglial loss in ESI-induced depression and demonstrate that the restoration of microglial homeostasis in the hippocampus may serve as a therapeutic strategy for depression induced by early-life stress. © 2018 Elsevier Ltd. All rights reserved.

Keywords: Hippocampus Microglia Early social isolation Lipopolysaccharide Macrophage colony-stimulating factor

1. Introduction Depression is a common disease that leads to severe social and economic problems. In adult individuals, depression is mainly induced by genetic and/or environmental factors, while in children, different types of early-life stresses, such as traumatic events and poor parental care, are considered risk factors (Amini-Khoei et al.,

* Corresponding author. ** Corresponding author. E-mail addresses: [email protected] (R. Yang), [email protected] (C. Huang). 1 Yu Gong, Lijuan Tong, and Rongrong Yang contribute equally to this work. https://doi.org/10.1016/j.neuropharm.2018.03.023 0028-3908/© 2018 Elsevier Ltd. All rights reserved.

2017; Liu et al., 2017; Lo Iacono et al., 2015). Depression induced by early-life stress has a different clinical profile, including more severe symptoms, an earlier onset, a more prolonged course of the disease, and poorer treatment outcomes (Miniati et al., 2010; Nanni et al., 2012). Depressed patients who have experienced early-life stress constitute a distinct clinical ecophenotype that can influence the therapeutic efficacy of conventional depression treatments (Andrus et al., 2012; Heim et al., 2008). To date, the exact mechanism underlying depression induced by early-life stress remains largely unknown. In the past several years, researchers have mainly focused on studying neurons in depression and considered central monoamine dysfunction as a major factor triggering the onset of depression (Dean and Keshavan, 2017; Liu et al., 2017a). Importantly, the most

224

Y. Gong et al. / Neuropharmacology 135 (2018) 223e233

widely used antidepressants, such as selective serotonin reuptake inhibitors and monoamine oxidase inhibitors, were developed as a result of the monoamine hypothesis of depression (Blier, 2016; Fujita et al., 2017). However, these agents are no longer considered effective antidepressants because numerous clinical practices have reported that only one-third of patients taking these agents exhibit relatively good treatment outcomes, while the other two-thirds only see a therapeutic effect after several weeks of treatment (Fabbri et al., 2013; Schwartz et al., 2016). Thus, it is necessary to find new methods to treat depression. Recently, we (Tong et al., 2017) and others (Kreisel et al., 2014; Yirmiya et al., 2015) reported that the initial activation and subsequent loss of hippocampal microglia mediates the development of depression induced by chronic unpredictable stress (CUS), chronic social defeat stress (CSDS), and chronic restraint stress (CRS). This finding suggests that interrupting these dynamic changes in hippocampal microglia may be a potential therapy for depression. In previous studies, the neuroinflammatory response was found to mediate depression induced by early-life stress. For example, an increased level of pro-inflammatory cytokine has been observed in the blood of individuals who had traumatic childhoods (Miller and Cole, 2012). Maternal separation can trigger microglial activation in the hippocampus of rat pups (Gracia-Rubio et al., 2016; Roque et al., 2016). Here, we hypothesized that the changes in hippocampal microglia may play a critical role in triggering depression induced by early-life stress. To test this hypothesis, we investigated the dynamic changes in hippocampal microglia in a depression model induced by early social isolation (ESI) during the third postnatal week (postnatal days 14e21; PD 14e21) in mice. ESI consists of less maternal care and social interactions (Lo Iacono et al., 2015). The third postnatal week of brain development is characterized by the maturation of visual, motor, and social abilities (Berardi et al., 2000; Pellis and Pasztor, 1999), and exposing mice to ESI during this period can induce depressive-like behaviors (Lo Iacono et al., 2015). We showed here that ESI induced a significant loss and dystrophy of microglia in the dentate gyrus (DG) of the hippocampus, which was mediated by the initial activation and subsequent apoptosis of hippocampal microglia after short-term ESI. Inhibiting the decline in hippocampal microglia abrogated the depressive-like behavior induced by ESI. These results suggest that restoring microglial homeostasis in the hippocampus may be a novel strategy for the treatment of depression induced by early-life stress. 2. Materials and methods 2.1. Materials Lipopolysaccharide (LPS, Escherichia coli, serotype 0111: B4) and macrophage colony-stimulating factor (M-CSF) were purchased from Sigma (Saint Louis, MO, USA) and Prospec (Ness-Ziona, Israel), respectively. Minocycline is the product of MedChem Express (Monmouth, NJ, USA). 2.2. Animals Three-week-old DBA/2J@Ico (DBA) male and female mice were purchased from the Model Animal Research Center in Nanjing University (Nanjing, China). Mice were group-housed under standard conditions (12-h light/dark cycle; lights on from 07:00 to 19:00; 23 ± 1
C ambient temperature; 55 ± 10% relative humidity) with free access to food and water. For the production of pups, the DBA mice were mated at 12 weeks of age, and only pup numbers varying between four and eight were included. Animal experiments

were conducted by following internationally accepted guidelines for the use of animals in toxicology as adopted by the Society of Toxicology in 1999 and approved by the University Animal Ethics Committee of Nantong University (Permit Number: 2110836). The researchers were blinded to the group allocation during the experiment and data analysis. 2.3. Social isolation procedure and pharmacological treatments Mouse pups were randomly assigned to the control or ESI group at PD 14. In the control group, mothers and offspring were left undisturbed without cage cleaning until weaning. In the ESI group, each pup was singly housed in a novel cage with clean bedding for 30 min per day from PD 14e21. All of the pups were weaned at PD 22 and were prepared to undergo behavioral testing or gene expression assays at 8e10 weeks of age. The behavioral tests were separated by 3-week intervals according to previous studies (Lo Iacono et al., 2015; Paylor et al., 2006). Since the results of the open field test (OFT) can be affected by inter-test intervals (Paylor et al., 2006), the behavioral experiments in our study were performed in the following order: OFT, tail suspension test (TST), forced swim test (FST), and sucrose preference test (SPT). A schematic diagram for these behavioral experiments is outlined in Fig. 1A. Both LPS and M-CSF were injected intraperitoneally (i.p.) at a dose of 100 mg/kg. LPS at 100 mg/kg has been confirmed to activate microglia and produce depressive-like behaviors (Yirmiya, 1996; Yirmiya et al., 2001; Tong et al., 2017). The M-CSF dose of 100 mg/kg was selected based on previous studies showing that this dosage effectively induces microglial activation and proliferation in chronically stressed mice (Boissonneault et al., 2009; Tong et al., 2017). The behavioral tests were performed 5 h after a single LPS administration or 5 days after M-CSF administration (once daily for 5 consecutive days). Minocycline was administered 2 days before the start of ESI (PD 12) via the drinking water at a dose of 40 mg/kg/ day that has been confirmed to effectively counteract chronic stress-induced microglial activation and behavioral changes (Hinwood et al., 2012, 2013). 2.4. TST and FST The TST and FST were performed according to previous studies (Porsolt et al., 1977; Steru et al., 1985). For TST, the mice in different groups (with/without ESI and/or drug treatment) were individually suspended 50 cm above the floor for 6 min by adhesive tape placed approximately 1 cm from the tip of the tail. An investigator blinded to the study then recorded the duration of immobility during the last 4 min of suspension. Mice were considered immobile only when they hung passively and were completely motionless. Any mouse that climbed its tail was excluded from further analysis. For FST, the mice (with/without ESI and/or drug treatment) were individually placed in a clear glass cylinder (25 cm in height and 10 cm in diameter) filled to 10 cm with water at 25 ± 1
C for 6 min. An investigator blinded to the study then recorded the duration of immobility during the animal's last 4 min in the water. Immobility time was defined as the time spent by the mouse floating in the water without struggling, making only those movements necessary to keep its head above the water. 2.5. SPT The SPT was performed according to previous studies (Liu et al., 2017b; Weng et al., 2016). The mice (with/without ESI and/or drug treatment) were given the choice to drink from two bottles in

Y. Gong et al. / Neuropharmacology 135 (2018) 223e233

225

Fig. 1. Effects of ESI on depressive-like behaviors and hippocampal microglial number and morphology. (A) A schematic diagram showing the timeline for ESI, OFT, TST, FST, and SPT in this experiment. (BeE) Quantitative analysis showing depressive-like behaviors following 8 days of ESI, including the increase in immobility in the TST (B, n ¼ 10, *p < 0.05 vs. Control) and FST (C, n ¼ 10, *p < 0.05 vs. Control) and the decrease in sucrose preference (D, n ¼ 10, **p < 0.01 vs. Control) and the time spent in the center region of the OFT (E, n ¼ 10, **p < 0.01 vs. Control). (F) Representative images showing the effect of ESI on the number of microglia in the DG region of the hippocampus. Scale bars for the low and high magnification images are 100 and 8 mm, respectively. (GeI) Quantitative analysis showing the effects of ESI on hippocampal microglial number (G, n ¼ 6, *p < 0.05 vs. Control), length of processes (H, n ¼ 15, **p < 0.01 vs. Control), and the area of the soma (I, n ¼ 15, *p < 0.05 vs. Control) in the DG region. (J) Quantitative analysis showing that ESI decreases the mRNA levels of Iba-1 (n ¼ 8, *p < 0.05 vs. Control) and CD11b (n ¼ 8, *p < 0.05 vs. Control) in the hippocampal DG region. All of the data are shown as mean ± SE.

individual cages, one with 1% sucrose solution and the other with water. All of the mice were acclimatized to the two-bottle choice condition for 2 days. The position of the two bottles was changed every 6 h to prevent possible side preference in drinking behavior. Then, the mice were deprived of food and water for 24 h. On the testing day, each mouse was exposed to pre-weighed bottles for 1 h with their positions interchanged. The sucrose preference was calculated as a percentage of the consumed sucrose solution

relative to the total amount of liquid intake. 2.6. OFT The spontaneous locomotor activity of the experimental mice was evaluated in the open-field paradigm over a 3-min period. The mice (with/without ESI and/or drug treatment) were placed individually in the middle of an open-field apparatus (40 cm in height,

226

Y. Gong et al. / Neuropharmacology 135 (2018) 223e233

100 cm in width, and 100 cm in length) with 25 (5  5 cm) squares delineated on the floor. The apparatus was illuminated with a red bulb (50 W) on the ceiling. The time that each mouse spent in the center region of the OFT was counted over a 3-min period under dim light conditions by an investigator blinded to the study. The open-field apparatus was thoroughly cleaned after each trial.

isolated comparisons. P < 0.05 is considered statistically significant.

2.7. Real-time reverse transcriptase-PCR

We first evaluated the effects of ESI on depressive-like behavior in DBA mice. In accordance with a previous study (Lo Iacono et al., 2015), our results showed that 8 days of ESI induced obvious depressive-like behaviors in DBA mice. We saw an increase in immobility in the TST (t10 ¼ 2.41, p < 0.05; Fig. 1B) and FST (t10 ¼ 2.60, p < 0.05; Fig. 1C) as well as a decrease in sucrose preference (t10 ¼ 5.49, p < 0.01; Fig. 1D) and the time spent in the center region of the OFT (t10 ¼ 3.29, p < 0.01; Fig. 1E). Similar to the effects of CUS, CRS, and CSDS on hippocampal microglial numbers (Tong et al., 2017), we found that 8 days of ESI resulted in a significant decrease in the number of microglia in the DG region of the hippocampus (t6 ¼ 2.67, p < 0.05; Fig. 1F and G). We also observed morphological changes; microglia labeled with Iba-1 in ESIexposed mice displayed significant decreases in the length of their microglial processes (t15 ¼ 4.40, p < 0.01; Fig. 1F, H) and soma size (t15 ¼ 2.33, p < 0.05; Fig. 1F, I). The mRNA levels of two microglial markers, Iba-1 and CD11b, in the DG region of the hippocampus were also reduced by 8 days of ESI (t8 ¼ 2.39, p < 0.05 and t8 ¼ 2.47, p < 0.05, respectively; Fig. 1J). These results suggest that the induction of depressive-like behaviors by ESI is accompanied by a significant loss of microglia in the hippocampal DG region.

The total RNA was isolated from the hippocampal DG region using an RNeasy mini kit (Qiagen, GmbH, Hilden, Germany). The first-strand cDNA was generated by reverse transcription of the total RNA using a reverse transcription system (Promega, Madison, WI, USA). Real-time PCR reactions were conducted with a reaction system containing 2 mL of diluted cDNA, 0.5 mM primers, 2 mM MgCl2, and 1  FastStart SYBR Green Master mix (Roche Molecular Biochemicals). The primers for Iba-1 and CD11b were cited as follows (Kreisel et al., 2014): Iba-1, 50 -GGCAATGGAGATATCGATAT-3’ (F), 50 - AGAATCATTCTCAAGATGGC-3’ (R); CD11b: 50 -CTGGTACATCGAGACTTCTC-3’ (F), R: 50 -TTGGTCTCTGTCTGAGCCTT-3’ (R). The PCR products were detected by monitoring the fluorescence increase in double-stranded DNA-binding dye SYBR Green during amplification. The expression levels of the target genes were normalized to the house-keeping gene 18S rRNA primer sequences: 50 -GTAACCCGTTGAACCCCATT-3’ (F), 50 -CCATCCAATCGGTAGTAGCG-3’ (R) (Huang et al., 2015). The fold-changes in the target gene expression in different groups were expressed as a ratio. Relative gene expression was calculated by the comparative cycle threshold (Ct) method. Melt-curve analysis and agarose gel electrophoresis were used to examine the authenticity of the PCR products. 2.8. Immunofluorescence The identification of hippocampal microglia was performed according to one of our previous studies (Huang et al., 2010). Briefly, animals were deeply anaesthetized with pentobarbital sodium and perfused transcardially with 4% paraformaldehyde in 0.01 M phosphate buffer. The brains of the mice were removed, frozen, and sectioned by a cryostat at 20 mm. Consecutive sections were collected in 24-well plates containing PBS. For immunofluorescence, the sections were permeabilized with 0.3% Triton X-100 for 30 min and incubated with 3% bovine serum albumin in PBS for another 30 min at room temperature, followed by a further incubation in PBS containing 0.3% Triton X-100, 1% BSA, and anti-Iba-1 antibody (1:200) overnight at 4
C. After that, the primary antibody was removed by washing the sections 3 times in PBS. The sections were then incubated in fluorescein isothiocyanate (FITC)-labeled horse anti-rabbit IgG (1:50) for 2 h at room temperature. After being washed, the sections were mounted on slides and coverslipped and finally examined using a Nikon Eclipse (Melville, NY, USA) and Olympus FV-500 confocal microscope and camera (Tokyo, Japan). The number of Iba-1-labeled microglia was automatically measured in a defined area exclusively containing the entire DG region using Nikon Imaging Elements Software (NIS-Elements, Melville, NY, USA). The process length was measured by manual tracing of the process of microglia using the NIS Elements software. 2.9. Statistical analysis All of the statistical analyses were performed using SPSS 13.0 software (SPSS Inc., USA), and all of the data are shown as mean ± SE. Differences between mean values were evaluated using the Student's t-test, one-way, or two-way analysis of variance (ANOVA), and the Bonferroni's post hoc test was used to assess

3. Results 3.1. ESI-induced depressive-like behaviors are accompanied by a decrease in the number of hippocampal microglia

3.2. Short-term ESI exposure induces a sequential activation and apoptosis of hippocampal microglia Since the chronic stress-induced loss of hippocampal microglia has been confirmed to be mediated by the initial activation of these microglia (Kreisel et al., 2014), we examined whether the ESIinduced loss of hippocampal microglia is also triggered by microglial activation. On day 1, but not day 3, following the start of ESI, the number of microglia in the DG region was increased markedly (F2,15 ¼ 7.39, p < 0.05; Fig. 2A) and accompanied by an obvious reduction in microglial process length (t18 ¼ 4.09, p < 0.01; Fig. 2B and C). After 24 h of ESI, the mRNA levels of Iba-1 and CD11b in the DG region were also increased substantially (t8 ¼ 2.91, p < 0.05 and t8 ¼ 3.01, p < 0.05, respectively; Fig. 2D), suggesting that short-term ESI can rapidly trigger microglial activation in the hippocampal DG region. To further determine the mechanism underlying the change in hippocampal microglial numbers in ESI-exposed mice, we investigated the effects of short-term ESI on microglial proliferation and apoptosis in the hippocampus. Our results showed that on day 1 following the start of ESI, hippocampal microglia underwent a substantial proliferation, reflected by an increase in the number of BrdU-labeled microglia in the DG region of the hippocampus (t4 ¼ 11.28, p < 0.01; Fig. 2EeG). After 4 days of ESI, we found that the number of TUNEL-labeled microglia in the DG region was increased markedly (t6 ¼ 3.31, p < 0.05; Fig. 3AeC). Taken together, these results suggest that the initial activation of hippocampal microglia triggered by short-term ESI induces their subsequent apoptosis, which contributes to the loss of microglia observed in adult depressed mice. 3.3. Inhibition of the initial activation of hippocampal microglia abrogates the effects of ESI on microglial dynamics and depressivelike behavior To further confirm the role of the initial activation of

Y. Gong et al. / Neuropharmacology 135 (2018) 223e233

227

Fig. 2. Effects of short-term ESI on microglial activation and proliferation in the DG region of the hippocampus. (A) Quantitative analysis showing the effect of 1 day and 3 days of ESI on hippocampal microglial numbers in the DG region (n ¼ 6, *p < 0.05 vs. Control). (B, C) Representative images and quantitative analysis showing the change in hippocampal microglial process length after 1 day of ESI (n ¼ 18, **p < 0.01 vs. Control). Scale bars: 8 mm. (D) Quantitative analysis showing the increased mRNA level of Iba-1 (n ¼ 8, *p < 0.05 vs. Control) and CD11b (n ¼ 8, *p < 0.05 vs. Control) in the hippocampal DG region after 1 day of ESI. (EeG) Representative images (E, F) and quantitative analysis (G) showing the effect of 1 day of ESI on the number of BrdU-labeled microglia in the DG region (n ¼ 4, **p < 0.01 vs. Control). Scale bars for the low and high magnification images are 100 and 8 mm, respectively. All of the data are shown as mean ± SE.

228

Y. Gong et al. / Neuropharmacology 135 (2018) 223e233

Fig. 3. Effects of short-term ESI on microglial apoptosis in the DG region of the hippocampus. (AeC) Representative images (A, B) and quantitative analysis (C) showing the increase in the number of TUNEL-labeled microglia in the hippocampal DG region of mice 4 days of ESI (n ¼ 6, *p < 0.05 vs. Control). Scale bars for the low and high magnification images are 100 and 8 mm, respectively. All of the data are shown as mean ± SE.

hippocampal microglia in the ESI-induced microglial loss and depressive-like behavior, we investigated the effect of minocycline pretreatment on hippocampal microglial dynamics and depressivelike behaviors. First, we evaluated the number of microglia in the hippocampus 1 day following ESI initiation with or without minocycline pretreatment (40 mg/kg/day via the drinking water). A two-way ANOVA indicated significant effects for ESI (F1,20 ¼ 7.71, p < 0.05) and minocycline pretreatment (F1,20 ¼ 9.91, p < 0.01) but not for the stress  treatment interaction (F1,20 ¼ 4.15, p ¼ 0.06) (Fig. 4A). Post hoc analysis showed that minocycline administration that was started 2 days before ESI markedly reversed the increase in the number of hippocampal microglia following 1 day of ESI (Fig. 4A). Second, we evaluated the microglial proliferation after 1 day of ESI. The two-way ANOVA indicated significant effects for ESI (F1,20 ¼ 18.33, p < 0.001), minocycline pretreatment (F1,20 ¼ 7.25, p < 0.05), and the stress  treatment interaction (F1,20 ¼ 4.49,

p < 0.05) (Fig. 4B). Post hoc analysis showed that the increase in the number of BrdU-labeled microglia in ESI-exposed hippocampi was also reversed after 2 days of minocycline pretreatment (Fig. 4B). Taken together, these results indicate that minocycline pretreatment abrogates the increase in hippocampal microglial number induced by short-term ESI. We showed using TUNEL staining to show that the subsequent decrease in the number of microglia in the hippocampus after ESI is mediated by microglial apoptosis, a process that has been previously reported to be blocked by microglial inhibition (Kreisel et al., 2014). To test if this ESI-induced apoptosis could be prevented by inhibiting microglia, we checked the status of microglial apoptosis with or without ESI and/or minocycline pretreatment (40 mg/kg/ day via the drinking water). The two-way ANOVA indicated significant effects for ESI (F1,20 ¼ 12.61, p < 0.01), minocycline pretreatment (F1,20 ¼ 15.99, p < 0.001), and the stress  treatment

Y. Gong et al. / Neuropharmacology 135 (2018) 223e233

229

Fig. 4. Effects of minocycline pretreatment (40 mg/kg/day) on ESI-induced microglial activation, microglial apoptosis, and depressive-like behaviors. (A) Quantitative analysis showing the inhibitory effect of minocycline pretreatment on the increased hippocampal microglial number observed after 1 day of ESI (n ¼ 6, *p < 0.05 vs. water þ control; #p < 0.05 vs. water þ 1 day of ESI). (B) Quantitative analysis of the hippocampal DG region showing the inhibitory effect of minocycline pretreatment on the increased number of BrdU-labeled microglia observed after 1 day of ESI (n ¼ 6, **p < 0.01 vs. water þ control; #p < 0.05 vs. water þ 1 day of ESI). (C) Quantitative analysis of the DG region showing the inhibitory effect of minocycline pretreatment on the increase in TUNEL-labeled microglia observed after 4 days of ESI (C, n ¼ 6, *p < 0.05 vs. water þ control; #p < 0.05 vs. water þ 4 days of ESI). (D) Quantitative analysis showing the inhibitory effect of minocycline pretreatment on 8 days of ESI-induced hippocampal microglial loss (n ¼ 6, **p < 0.01 vs. water þ control; #p < 0.05 vs. water þ8 days of ESI). (E, F) Quantitative analysis showing the inhibitory effect of minocycline pretreatment on the increase in immobility in the TST (E, n ¼ 10, *p < 0.05 vs. water þ control; #p < 0.05 vs. water þ 8 days of ESI) and FST (F, n ¼ 10, **p < 0.01 vs. water þ control; #p < 0.05 vs. water þ 8 days of ESI) induced by 8 days of ESI. (G, H) Quantitative analysis showing the inhibitory effect of minocycline pretreatment on the decrease in sucrose preference (G, n ¼ 10, **p < 0.01 vs. water þ control; ##p < 0.01 vs. water þ 8 days of ESI) and the decrease in the time spent in the center region of the OFT (H, n ¼ 10, **p < 0.01 vs. water þ control; ##p < 0.01 vs. water þ 8 days of ESI) induced by 8 days of ESI. All of the data are shown as mean ± SE.

interaction (F1,20 ¼ 10.83, p < 0.01) (Fig. 4C). Post hoc analysis confirmed significantly less TUNEL-labeled microglia in the hippocampal DG region after 4 days of ESI in mice pretreated with 2 days of minocycline (Fig. 4C). This suggests that minocycline pretreatment prevents ESI-induced apoptosis of hippocampal microglia. We also evaluated the change in microglial number in the hippocampus after long-term minocycline pretreatment (5 weeks). In the long-term minocycline treatment experiment, the two-way ANOVA indicated significant effects for ESI (F1,20 ¼ 8.46, p < 0.01) and the stress  treatment interaction (F1,20 ¼ 4.92, p < 0.05) but not for chronic minocycline treatment (F1,20 ¼ 2.32, p ¼ 0.14) (Fig. 4D). Post hoc analysis showed that long-term minocycline treatment beginning 2 days before ESI blocked the decrease in the number of hippocampal microglia induced by 8 days of ESI (Fig. 4D). These results provide direct evidence to show that the initial activation of hippocampal microglia by short-term ESI contributes to the subsequent apoptosis and loss of these microglia. Finally, we evaluated the influence of chronic minocycline treatment on ESI-induced depressive-like behaviors in the following tests: TST, FST, SPT, and OFT. For the TST, a two-way ANOVA indicated a significant effect for the stress  treatment interaction (F1,36 ¼ 4.12, p < 0.05) but not for ESI (F1,36 ¼ 1.71, p ¼ 0.20) or chronic minocycline treatment (F1,36 ¼ 1.30, p ¼ 0.26) (Fig. 4E). For the FST, the two-way ANOVA indicated a significant effect for ESI (F1,36 ¼ 11.48, p < 0.01) but not for chronic minocycline treatment (F1,36 ¼ 1.42, p ¼ 0.24) or the stress  treatment interaction (F1,36 ¼ 4.04, p ¼ 0.05) (Fig. 4F). For the SPT, the two-way ANOVA indicated significant effects for ESI (F1,36 ¼ 23.21, p < 0.001) and chronic minocycline treatment (F1,36 ¼ 5.28, p < 0.05) but not for the stress  treatment interaction (F1,36 ¼ 3.83, p < 0.06) (Fig. 4G). For the OFT, the two-way ANOVA indicated significant effects for ESI (F1,36 ¼ 30.49, p < 0.001) and the

stress  treatment interaction (F1,36 ¼ 7.37, p < 0.05) but not for chronic minocycline treatment (F1,36 ¼ 2.85, p ¼ 0.10) (Fig. 4H). Post hoc analysis showed that chronic minocycline treatment reversed the ESI (8 days)-induced increase in immobility in the TST (Fig. 4E) and FST (Fig. 4F) as well as the ESI-induced decrease in sucrose preference (Fig. 4G) and the time spent in the center region of the OFT (Fig. 4H). Taken together, these results indicate that inhibition of the initial activation of hippocampal microglia by minocycline ameliorates the depressive-like behavior induced by 8 days of ESI. 3.4. Microglial stimulation suppresses the depressive-like behavior induced by ESI To acquire causal evidence for the involvement of hippocampal microglial loss in ESI-induced depressive-like behavior, we examined the effect of microglial stimulation on ESI (8 days)-induced depressive-like behavior. When we stimulated the microglia with LPS (100 mg/kg, i.p.) and ran the different behavioral tests, a twoway ANOVA indicated the following: a significant effect for the stress  treatment interaction (F1,74 ¼ 17.81, p < 0.001) but not for ESI (F1,74 ¼ 0.57, p ¼ 0.45) or LPS treatment (F1,74 ¼ 1.75, p ¼ 0.19) in the TST (Fig. 5A); a significant effect for the stress  treatment interaction (F1,74 ¼ 11.50, p < 0.01) but not for ESI (F1,74 ¼ 0.12, p ¼ 0.73) or LPS treatment (F1,74 ¼ 0.10, p ¼ 0.76) in the FST (Fig. 5B); a significant effect for the stress  treatment interaction (F1,74 ¼ 28.69, p < 0.001) but not for ESI (F1,74 ¼ 1.31, p ¼ 0.26) or LPS treatment (F1,74 ¼ 0.03, p ¼ 0.85) in the SPT (Fig. 5C); significant effects for LPS treatment (F1,74 ¼ 6.36, p < 0.05) and the stress  treatment interaction (F1,74 ¼ 41.12, p < 0.001) but not for ESI (F1,74 ¼ 2.28, p ¼ 0.14) in the OFT (Fig. 5D). Post hoc analysis showed that at 5 h after administration of LPS, the ESI-induced increase in immobility in the TST (Fig. 5A) and FST (Fig. 5B) was

230

Y. Gong et al. / Neuropharmacology 135 (2018) 223e233

Fig. 5. Effects of LPS and M-CSF (both administered at 100 mg/kg) on ESI-induced depressive-like behaviors and hippocampal microglial loss. (A, B) Quantitative analysis showing the inhibitory effect of LPS on the increase in immobility in the TST (A, n ¼ 18e21, *p < 0.05, **p < 0.01 vs. saline; ##p < 0.01 vs. saline þ ESI) and FST (B, n ¼ 18e21, *p < 0.05 vs. saline; ##p < 0.01 vs. saline þ ESI) induced by 8 days of ESI. (C, D) Quantitative analysis showing the inhibitory effect of LPS on the decrease in sucrose preference (C, n ¼ 18e21, **p < 0.01 vs. saline; #p < 0.05 vs. saline þ ESI) and the decrease in the time spent in the center region of the OFT (D, n ¼ 18e21, **p < 0.01 vs. saline; ##p < 0.01 vs. saline þ ESI) induced by 8 days of ESI. (E) Quantitative analysis showing the inhibitory effect of LPS on the decrease in hippocampal microglial number induced by 8 days of ESI (n ¼ 5, **p < 0.01 vs. saline; #p < 0.05 vs. saline þ ESI). (F) Quantitative analysis showing the effect of ESI and/or LPS on the number of BrdU-labeled microglia in the hippocampal DG region (n ¼ 5, *p < 0.05 vs. saline; #p < 0.05 vs. saline þ ESI). (G, H) Quantitative analysis showing the inhibitory effect of M-CSF on the increase in immobility in the TST (G, n ¼ 18e19, *p < 0.05 vs. saline; #p < 0.05 vs. saline þ ESI) and FST (H, n ¼ 18e19, *p < 0.05 vs. saline; #p < 0.05 vs. saline þ ESI) induced by 8 days of ESI. (I, J) Quantitative analysis showing the inhibitory effect of M-CSF on the decrease in sucrose preference (I, n ¼ 18e19, **p < 0.01 vs. saline; ##p < 0.01 vs. saline þ ESI) and the decrease in time spent in the center region of the OFT (J, n ¼ 18e19, **p < 0.01 vs. saline; ##p < 0.01 vs. saline þ ESI) induced by 8 days of ESI. (K) Quantitative analysis showing the inhibitory effect of M-CSF on the decrease in hippocampal microglial number induced by 8 days of ESI (n ¼ 5, *p < 0.05 vs. saline; #p < 0.05 vs. saline þ ESI). (L) Quantitative analysis showing the effect of ESI and/or M-CSF on the number of BrdU-labeled microglia in the DG region of the hippocampus (n ¼ 5, *p < 0.05 vs. saline; ##p < 0.01 vs. saline þ ESI). All of the data are shown as mean ± SE.

markedly reversed. Similarly, the ESI-induced decrease in sucrose preference (Fig. 5C) and the time spent in the center region of the OFT (Fig. 5D) were also reversed by LPS administration. The change in hippocampal microglial number after LPS stimulation was also evaluated. In this experiment, the two-way ANOVA indicated significant effects for ESI (F1,16 ¼ 45.06, p < 0.001) and LPS treatment (F1,16 ¼ 23.91, p < 0.001) but not for the stress  treatment interaction (F1,16 ¼ 3.39, p ¼ 0.08) (Fig. 5E). Post hoc analysis showed that at 5 h after administration of LPS, the decrease in microglial number

in ESI-exposed mice was reversed (Fig. 5E). The stimulating effect of LPS in the hippocampal microglia was ascertained by BrdU experiments, in which the two-way ANOVA indicated significant effects for ESI (F1,16 ¼ 21.76, p < 0.001) and the stress  treatment interaction (F1,16 ¼ 6.81, p < 0.05) but not for LPS treatment (F1,16 ¼ 0.002, p ¼ 0.96) (Fig. 5F). Post hoc analysis showed that the number of hippocampal BrdU-labeled microglia was reduced in the ESIexposed mice, and this reduction was reversed by a single LPS administration (Fig. 5F). These results demonstrate that LPS can

Y. Gong et al. / Neuropharmacology 135 (2018) 223e233

reverse ESI-induced depressive-like behaviors via promotion of hippocampal microglial proliferation. In a second experiment, we used 100 mg/kg of M-CSF (administered i.p. for consecutive 5 days) to activate the hippocampal microglia and evaluated its effect on depressive-like behavior in mice exposed to ESI for 8 days. For the TST, the two-way ANOVA indicated a significant effect for the stress  treatment interaction (F1,69 ¼ 10.51, p < 0.01) but not for ESI (F1,69 ¼ 0.02, p ¼ 0.90) or MCSF treatment (F1,69 ¼ 0.18, p ¼ 0.68) (Fig. 5G). For the FST, the twoway ANOVA indicated a significant effect for the stress  treatment interaction (F1,69 ¼ 11.41, p < 0.01) but not for ESI (F1,69 ¼ 0.27, p ¼ 0.60) or M-CSF treatment (F1,69 ¼ 0.06, p ¼ 0.80) (Fig. 5H). For the SPT, the two-way ANOVA indicated a significant effect for the stress  treatment interaction (F1,69 ¼ 29.66, p < 0.001) but not for ESI (F1,69 ¼ 0.001, p ¼ 0.97) or M-CSF treatment (F1,69 ¼ 0.81, p ¼ 0.37) (Fig. 5I). Lastly, for the OFT, the two-way ANOVA indicated a significant effect for the stress  treatment interaction (F1,69 ¼ 18.84, p < 0.001) but not for ESI (F1,69 ¼ 0.009, p ¼ 0.93) or M-CSF treatment (F1,69 ¼ 1.16, p < 0.29) (Fig. 5J). Post hoc analysis showed that the treatment of mice with M-CSF markedly reversed the ESI-induced increase in immobility in the TST (Fig. 5G) and FST (Fig. 5H) as well as the ESI-induced decrease in sucrose preference (Fig. 5I) and the time spent in the center region of the OFT (Fig. 5J). We also examined the change in hippocampal microglial number and proliferation in mice treated with M-CSF. For the hippocampal microglial number, the two-way ANOVA indicated significant effects for ESI (F1,16 ¼ 4.97, p < 0.05) and M-CSF treatment (F1,16 ¼ 9.63, p < 0.01) but not for the stress  treatment interaction (F1,16 ¼ 0.13, p ¼ 0.73) (Fig. 5K). For hippocampal microglial proliferation, the two-way ANOVA again indicated significant effects for ESI (F1,16 ¼ 8.09, p < 0.05) and M-CSF treatment (F1,16 ¼ 10.23, p < 0.01) but not for the stress  treatment interaction (F1,16 ¼ 0.22, p ¼ 0.65) (Fig. 5L). Post hoc analysis showed that 5 days of M-CSF administration reversed the ESI-induced decrease in the hippocampal microglia number (Fig. 5K) by a mechanism that increased the proliferation of hippocampal microglia (number of BrdUlabeled microglia) in the ESI-exposed mice (Fig. 5L). Taken together, these results suggest that treating mice with M-CSF also reverses ESI-induced depressive-like behavior via the promotion of hippocampal microglial proliferation. 4. Discussion One of the major findings in the present study is that ESI induced a significant loss of microglia in the DG region of the hippocampus. This result is in accordance with previous studies showing that several different types of chronic stress, including CUS, CRS, and CSDS, also induce hippocampal microglial loss (Kreisel et al., 2014; Tong et al., 2017). Reversing this loss of microglia in the hippocampus exhibited clear therapeutic effects in depression induced by CUS, CRS, and CSDS (Kreisel et al., 2014; Tong et al., 2017). We showed here that the systemic administration of mice with two microglial stimulators, LPS and M-CSF, reversed the depressive-like behavior induced by 8 days of ESI from PD 14e21. The antidepressant effects of LPS and M-CSF in this model may have great significance for the treatment of depression induced by early-life stress as this type of depression: i) has many special characteristics, such as more severe symptoms, earlier onset, and worse treatment outcomes (Miniati et al., 2010; Nanni et al., 2012); ii) cannot be treated efficiently with current clinical antidepressants (Duval et al., 2006; Niciu et al., 2015); iii) has been shown to be transmitted across generations (Colvin et al., 2017); and iv) is difficult to prevent in childhood. Thus, developing a therapeutic but not a preventive strategy for patients who suffer from early-life stress appears to be more practical. We

231

demonstrated this idea in our current study by showing that the reversal of hippocampal microglial loss almost completely abrogated ESI-induced depressive-like behavior. According to this finding, we speculate that drugs with microglia-activating effects may have therapeutic potential in treating depression induced by early-life stress. Our research group has already screened some clinically available drugs that can activate microglia, and their antidepressant efficacies in different depression models are being evaluated. In contrast with the antidepressant effects of LPS or M-CSF administration after ESI, the same dose of LPS and M-CSF was found to induce depressive-like behavior in normal animals. These effects may be associated with the pro-inflammatory activity of LPS and M-CSF in peripheral tissues and in the blood. Since proinflammatory cytokines produced in the blood and peripheral tissues have been confirmed to penetrate the brain and promote the onset and development of depression in both animals and humans (Benatti et al., 2016; Lichtblau et al., 2013; Money et al., 2016), the administration of LPS or M-CSF could plausibly have aggravated depressive-like behavior in the mice exposed to ESI. However, our results showed the opposite outcome; both LPS and M-CSF almost completely reversed the depressive-like behavior induced by ESI. The beneficial or detrimental effect of microglial activation in depressed versus normal animals may be due to the existing condition of the hippocampal microglia in the two different brain states. In normal animals, since the function of the hippocampal microglia is intact, LPS or M-CSF administration would induce € ring et al., 2015; Huang et al., 2013; microglial hyperactivation (Do Zhu et al., 2014), while in depressed animals, the hippocampal microglia are in a hypoactive state and the systemic administration of LPS or M-CSF returns them to normal function. Accordingly, the increase in microglial activity level triggered by LPS and M-CSF becomes detrimental in one case and beneficial in the other. It would be valuable to investigate the mechanism underlying the therapeutic effect of LPS and M-CSF in depression induced by earlylife stress because appropriately functioning microglia are critical for the integrity of a variety of brain functions. For example, microglial activation has been reported to enhance associative taste memory via the activation of glutamatergic neurotransmission (Delpech et al., 2015). Microglial activation also triggers excitatory neurotransmission in a purinergic receptor- and astrocytedependent manner (Pascual et al., 2012). In future studies, we will identify the signal (downstream of purinergic receptors or astrocytes) that mediates the inhibitory effect of hippocampal microglial activation on depressive-like behavior induced by different types of stress, including ESI. In our study, we also showed that the reversal of ESI-induced microglial loss, triggered by LPS or M-CSF administration, was mediated by the proliferation of hippocampal microglia. This result is in accordance with the stimulating effect of LPS and M-CSF in immune cells (Fukushima et al., 2015; Kamigaki et al., 2016). However, the mechanism underlying the stimulating effect of LPS or M-CSF on the hippocampal microglia remains to be determined. The exact source of microglial proliferation in LPS- or M-CSFtreated mice is also unclear. Could these new microglia originate from the proliferation of remnant microglia in ESI brains or perhaps from the recruitment of monocytes in the blood and peripheral tissues into the brain? A recent study by Elmore et al. (2014) showed that the microglia that disappeared following CSF-1 receptor (CSF1R) inhibition reappear after the inhibitor is discontinued, suggesting that CSF1R may mediate the LPS- or M-CSFinduced increase in hippocampal microglia. In fact, both LPS and MCSF have been reported to trigger microglial activation via CSF signals in previous studies (Kamigaki et al., 2016; Okubo et al., 2016). According to Elmore et al. (2014), the newly emerged

232

Y. Gong et al. / Neuropharmacology 135 (2018) 223e233

microglia after CSF1R inhibitor withdrawal are supposed to originate from some unknown precursor cells in the brain. In our study, since the proliferation of hippocampal microglia occurs rapidly after LPS or M-CSF administration and this proliferation is accompanied by the improvement of depressive-like behavior in ESI mice, we also suppose that the proliferated microglia may originate from unknown precursor cells. However, this hypothesis needs to be further investigated before any conclusions can be drawn. The finding that the activation of hippocampal microglia by LPS or M-CSF reverses the depressive-like behavior induced by ESI suggests that the elimination of hippocampal microglia may have detrimental effects on mood-associated behaviors. However, a recent study by Elmore et al. (2014) showed that complete depletion of brain microglia by a CSF1R inhibitor has no significant effects on depressive-like behavior in mice. Two possible reasons could be considered for this difference: i) the CSF1R inhibitor may have other pharmacological effects in the brain that could effectively compensate for its detrimental effects on depressive-like behavior; and ii) the depressive-like behaviors induced by hippocampal microglial loss only occur under conditions of stress and moodassociated behaviors in normal animals may not be affected by microglial depletion. In a CUS model of depression, the loss of hippocampal microglia was shown to be mediated by an initial activation and subsequent apoptosis (Kreisel et al., 2014). Similarly, the ESI-induced loss of hippocampal microglia is also associated with an initial activation phase, during which the number of BrdU-labeled hippocampal microglia were markedly increased after 1 day of ESI. This initial activation of hippocampal microglia after short-term ESI can trigger a subsequent apoptosis of the microglia as well as depressive-like behavior, which was demonstrated by the result that the prevention of the initial activation phase of the microglia by minocycline pretreatment suppressed the ESI-induced apoptosis and depressive-like behavior. These results suggest that the hippocampal microglial activation triggered by short-term early-life stress in childhood can be a critical contributor to the onset of depression in adults. In fact, microglial activation upon maternal separation has been observed in recent studies. For example, pups undergoing maternal separation have been shown to exhibit more activated microglia and more production of pro-inflammatory cytokines, such as interleukin-1b (IL-1b), tumor necrosis factor-a (TNF-a), and IL-6 (Gracia-Rubio et al., 2016; Roque et al., 2016). How different types of early-life stress triggers microglial activation remains to be determined in future studies. 5. Conclusions Our results show for the first time that ESI, a type of early-life stress, induces depressive-like behavior in DBA mice by triggering a loss of hippocampal microglia. The restoration of the hippocampal microglia reverses the depressive-like behavior induced by ESI. These findings may be of great significance for the treatment of depression induced by early-life stress. There are no effective methods to treat depressive symptoms in patients whose developing brains were affected during childhood by some early-life stress that cannot be retroactively prevented. Our results also showed that the ESI-induced loss of the hippocampal microglia was mediated by the initial activation and subsequent apoptosis of hippocampal microglia, suggesting that the inhibition of hippocampal microglial activation in childhood periods could prevent the onset of depression in adulthood. Disclosure statement The authors have no competing financial interests to declare.

This study is approved by the University Animal Ethics Committee of Nantong University (Permit Number: 2110836). Conflicts of interest The authors declare no competing financial interests. Acknowledgments This work is supported by the Natural Science Foundation of China (81571323, 81771467, and 81701286), the Natural Science Foundation of Jiangsu Province (BK20141240), and the scientific research project of Nantong First People's Hospital (YJ201601). References Amini-Khoei, H., et al., 2017. Oxytocin mitigated the depressive-like behaviors of maternal separation stress through modulating mitochondrial function and neuroinflammation. Prog. Neuro-Psychopharmacol. Biol. Psychiatry 76, 169e178. Andrus, B.M., et al., 2012. Gene expression patterns in the hippocampus and amygdala of endogenous depression and chronic stress models. Mol. Psychiatr. 17, 49e61. Benatti, C., et al., 2016. Disease-induced neuroinflammation and depression. CNS Neurol. Disord. - Drug Targets 15, 414e433. Berardi, N., et al., 2000. Critical periods during sensory development. Curr. Opin. Neurobiol. 10, 138e145. Blier, P., 2016. Neurobiology of depression and mechanism of action of depression treatments. J. Clin. Psychiatr. 77, e319. Boissonneault, V., et al., 2009. Powerful beneficial effects of macrophage colonystimulating factor on beta-amyloid deposition and cognitive impairment in Alzheimer's disease. Brain 132, 1078e1092. Colvin, A., et al., 2017. The role of family history of depression and the menopausal transition in the development of major depression in midlife women: study of women's health across the nation mental health study (SWAN MHS). Depress. Anxiety 34, 826e835. Dean, J., Keshavan, M., 2017. The neurobiology of depression: an integrated view. Asian J. Psychiatr 27, 101e111. Delpech, J.C., et al., 2015. Microglial activation enhances associative taste memory through purinergic modulation of glutamatergic neurotransmission. J. Neurosci. 35, 3022e3033. €ring, A., et al., 2015. Stimulation of monocytes, macrophages, and microglia by Do amphotericin B and macrophage colony-stimulating factor promotes remyelination. J. Neurosci. 35, 1136e1148. Duval, F., et al., 2006. Treatments in depression. Dialogues Clin. Neurosci. 8, 191e206. Elmore, M.R., et al., 2014. Colony-stimulating factor 1 receptor signaling is necessary for microglia viability, unmasking amicroglia progenitor cell in the adult brain. Neuron 82, 380e397. Fabbri, C., et al., 2013. Clinical features and drug induced side effects in early versus late antidepressant responders. J. Psychiatr. Res. 47, 1309e1318. Fujita, M., et al., 2017. cAMP signaling in brain is decreased in unmedicated depressed patients and increased by treatment with a selective serotonin reuptake inhibitor. Mol. Psychiatr. 22, 754e759. Fukushima, S., et al., 2015. Robust increase of microglia proliferation in the fornix of hippocampal axonal pathway after a single LPS stimulation. J. Neuroimmunol. 285, 31e40. Gracia-Rubio, I., et al., 2016. Maternal separation induces neuroinflammation and long-lasting emotional alterations in mice. Pro. Neuropsychopharmacol. Biol. Psychiatry 65, 104e117. Heim, C., et al., 2008. The dexamethasone/corticotropin-releasing factor test in men with major depression: role of childhood trauma. Biol. Psychiatr. 63, 398e405. Hinwood, M., 2012. Evidence that microglia mediate the neurobiological effects of chronic psychological stress on the medial prefrontal cortex. Cerebr. Cortex 2012 (22), 1442e1454. Hinwood, M., 2013. Chronic stress induced remodeling of the prefrontal cortex: structural re-organization of microglia and the inhibitory effect of minocycline. Cerebr. Cortex 23, 1784e1797. Huang, C., et al., 2010. Existence and distinction of acid-evoked currents in rat astrocytes. Glia 58, 1415e1424. Huang, C., et al., 2013. TSG (2,3,4' ,5-tetrahydroxystilbene 2-O-b-D-glucoside) suppresses induction of pro-inflammatory factors by attenuating the binding activity of nuclear factor-kB in microglia. J. Neuroinflammation 10, 129. Huang, C., et al., 2015. Requirement for endogenous heat shock factor 1 in inducible nitric oxide synthase induction in murine microglia. J. Neuroinflammation 12, 189. Kamigaki, M., et al., 2016. The Toll-like receptor 4-activated neuroprotective microglia subpopulation survives via granulocytemacrophage colonystimulating factor and JAK2/STAT5 signaling. Neurochem. Int. 93, 82e94. Kreisel, T., et al., 2014. Dynamic microglial alterations underlie stress-induced

Y. Gong et al. / Neuropharmacology 135 (2018) 223e233 depressive-like behavior and suppressed neurogenesis. Mol. Psychiatr. 19, 699e709. Lichtblau, N., et al., 2013. Cytokines as biomarkers in depressive disorder: current standing and prospects. Int. Rev. Psychiatr. 25, 592e603. Liu, F., et al., 2017b. Harmine produces antidepressant-like effects via restoration of astrocytic functions. Prog. Neuro-Psychopharmacol. Biol. Psychiatry 79, 258e267. Liu, F.G., et al., 2017a. Z-Guggulsterone produces antidepressant-like effects in mice through activation of the BDNF signaling pathway. Int. J. Neuropsychopharmacol. 20, 485e497. Liu, H., et al., 2017. Behavioral and cognitive impact of early life stress: insights from an animal model. Prog. Neuro-Psychopharmacol. Biol. Psychiatry 78, 88e95. Lo Iacono, L., et al., 2015. Adversity in childhood and depression: linked through SIRT1. Transl. Psychiatry 5, e629. Miller, G.E., Cole, S.W., 2012. Clustering of depression and inflammation in adolescents previously exposed to childhood adversity. Biol. Psychiatr. 72, 34e40. Miniati, M., et al., 2010. Clinical characteristics and treatment outcome of depression in patients with and without a history of emotional and physical abuse. J. Psychiatr. Res. 44, 302e309. Money, K.M., et al., 2016. An altered peripheral IL6 response in major depressive disorder. Neurobiol. Dis. 89, 46e54. Nanni, V., et al., 2012. Childhood maltreatment predicts unfavorable course of illness and treatment outcome in depression: a meta-analysis. Am. J. Psychiatr. 169, 141e151. Niciu, M.J., et al., 2015. A history of early life parental loss or separation is associated with successful cognitive-behavioraltherapy in major depressive disorder. J. Affect. Disord. 187, 241e244. Okubo, M., et al., 2016. Macrophage-colony stimulating factor derived from injured primary afferent induces proliferation of spinal microglia and neuropathic pain in rats. PLoS One 11 e0153375. Pascual, O., et al., 2012. Microglia activation triggers astrocyte-mediated modulation

233

of excitatory neurotransmission. Proc. Natl. Acad. Sci. U. S. A 109, E197eE205. Paylor, R., et al., 2006. The use of behavioral test batteries, II: effect of test interval. Physiol. Behav. 87, 95e102. Pellis, S.M., Pasztor, T.J., 1999. The developmental onset of a rudimentary form of play fighting in C57 mice. Dev. Psychobiol. 34, 175e182. Porsolt, R.D., et al., 1977. Behavioral despair in mice: a primary screening test for antidepressants. Arch. Int. Pharmacodyn. Ther. 229, 327e336. Roque, A., et al., 2016. Maternal separation activates microglial cells and induces an inflammatory response in the hippocampus of male rat pups, independently of hypothalamic and peripheral cytokine levels. Brain Behav. Immun. 55, 39e48. Schwartz, J., et al., 2016. Ketamine for treatment- resistant depression: recent developments and clinical applications. Evid. Base Ment. Health 19, 35e38. Steru, L., et al., 1985. The tail suspension test: a new method for screening antidepressants in mice. Psychopharmacology 85, 367e370. Tong, L., et al., 2017. Microglia loss contributes to the development of major depression induced by different types of chronic stresses. Neurochem. Res. 42, 2698e2711. Weng, L., et al., 2016. Apigenin reverses depression-like behavior induced by chronic corticosterone treatment in mice. Eur. J. Pharmacol. 774, 50e54. Yirmiya, R., 1996. Endotoxin produces a depressive-like episode in rats. Brain Res. 711, 163e174. Yirmiya, R., et al., 2001. Effects of antidepressant drugs on the behavioral and physiological responses to lipopolysaccharide (LPS) in rodents. Neuropsychopharmacology 24, 531e544. Yirmiya, R., et al., 2015. Depression as a microglial disease. Trends Neurosci. 38, 637e658. Zhu, F., et al., 2014. Minocycline alleviates behavioral deficits and inhibits microglial activation induced by intrahippocampal administration of GranulocyteMacrophage Colony-Stimulating Factor in adult rats. Neuroscience 266, 275e281.