Forty-three 13-week-old pregnant female C57BL/6N mice (Koatec Inc., Seoul, Korea) were used in this study. Information on which animals were used is presented in Table 1. Pregnant female mice were housed in cages that held three to five animals each, under specific pathogen-free conditions at 22 ± 0.5°C with a 12-h light–dark cycle and supplied with food and water ad libitum at the CHA BIO COMPLEX animal facility (Seongnam, Korea). After the weaning period, a series of behavioral assessments were performed on the dams. All behavioral assessments were conducted during the light phase of the cycle. Experimental procedures were approved by the Animal Care and Use Committee of CHA University (IACUC220088).

50 μg/kg DEX (Sigma, D1756) and physiological saline as control were injected subcutaneously into pregnant mice once a day for 3 days (gestational day: GD16 to 18) to recapitulate elevation of stress hormones during the antenatal period (Rim et al., 2022).

Behavioral assessments were performed during the light cycle between 9:00 AM and 6:00 PM, with one assessment per day. Behavioral tests consisted of an open-field test (OFT), light and dark (LD) test, social interaction (SI) test, sucrose preference test (SPT), tail suspension test (TST), and forced swimming test (FST). All assessments were performed as blinded tests to minimize bias. Each behavioral assessment is described in Supplementary Data 1 and referred to our previous studies (Han et al., 2015; Park et al., 2020).

Whole blood was collected by cardiac puncture when mice were sacrificed for perfusion and dissection. Serum was isolated by centrifugation at 448 g at 4°C for 20 min and stored at−80°C until assayed. Serum glucocorticoid, IL-1β, TNF-α, and IL-10 levels were measured using corticosterone ELISA kit (Enzo, ADI-901-097), IL-1β ELISA kit (R&D Systems, MLB00C), TNF-α ELISA kit (Enzo, ADI-900-047), and IL-10 ELISA kit (Abcam, ab255729), respectively, according to the manufacturer's instructions. Enzyme-reacted serum was read in a microplate reader at 450 nm except for corticosterone, and only corticosterone was read at 405 nm.

After behavioral assessments, Western blot was performed referring to our previous study (Rim et al., 2022). Briefly, the dissected hippocampi and amygdala were homogenized in radioimmunoprecipitation assay (RIPA) buffer (Thermo Fisher Scientific, 89900). The concentration of protein from the supernatant after centrifuge was measured using a protein assay reagent (Bio-Rad, #5000006) using bovine serum albumin as a standard. Protein lysates in sample buffer (Biosolution, BS0028) were incubated at 95°C for 10 min, separated by SDS-PAGE, and transferred onto polyvinylidene fluoride (PVDF) membranes. After blocking with 5% skim milk in Tris-buffered saline and 0.1% Tween-20 (TBST) for 1 h at room temperature, membranes were incubated with the primary antibodies, including anti-glucocorticoid receptor (GR) antibody (1:1,000, Santa Cruz, sc-1004), PSD95 antibody (1:500, Abcam, ab18258), synaptophysin antibody (1:10,000, Abcam, ab14692), BDNF antibody (1:1,000, Abcam, ab108319), pCREB antibody (1:500, Cell signaling, 9,198), CREB antibody (1:500, Cell signaling, 9,197), 5-HT1A receptor antibody (1:1,000, GeneTex, gtx104703), β-actin antibody (1:10,000, Cell signaling, 4,970), GAD67 antibody (1:500, Abcam, ab26116), CX3CR1 antibody (1:1,000, ProteinTech, 13885-1-AP), Arginase1 antibody (1:10,000, Novus, NB100-59740), myelin basic protein (MBP) antibody (1:10,000, Servicebio, GB12226), and glial fibrillary acidic protein (GFAP) antibody (1:10,000, Abcam, ab7260) at 4 °C overnight. Goat anti-rabbit IgG HRP-conjugated secondary antibody (1:10,000, Bethyl, A120-101P), Goat anti-mouse IgG HRP-conjugated secondary antibody (1:10,000, Bethyl, A90-116P), and Donkey anti-goat IgG HRP-conjugated secondary antibody (1:10,000, Bethyl, A50-101P) were incubated for 1 h at room temperature. The membranes were visualized using an ECL-plus solution (GE Healthcare, RPN2106). The membranes were exposed to chemiluminescence (LAS 4000, Fujifilm) to detect light emission. Quantification of band intensities was performed using the Fusion FX software (Vilber Lourmat).

Immunohistochemistry was performed referring to our previous study (Rim et al., 2022). Brain sections were incubated with the primary antibodies diluted with 1% NDS in PBS-T, including anti-c-Fos antibody (1:1,000, Abcam, ab190289), Iba-1 antibody (1:500, Wako, 019-19741), Olig2 antibody (1:300, Millipore, AB9610), MBP antibody (1:800, Servicebio, GB12226), and GFAP antibody (1:500, Abcam, ab7260) at 4 °C overnight. Brain sections were washed three times and incubated with Donkey anti-rabbit Alexa 488-conjugated secondary antibody (1:200, Invitrogen, A-21206), Donkey anti-mouse Alexa 488-conjugated secondary antibody (1:200, Invitrogen, A-21202), and Donkey anti-rabbit Alexa 647-conjugated secondary antibody (1:200, Invitrogen, A-31573) diluted with PBS-T at room temperature for 1 h. Brain sections were mounted using ProLong^TM Gold Antifade mounting solution with DAPI (Invitrogen, P36931). Fluorescent images were obtained using a confocal microscope (TCS SP5 II; Leica Microsystems, Wetzlar, Germany). One field of view (FOV) was obtained per brain region per mouse for confocal imaging. In 25 μm thickness brain section, images in 1 μm units were merged into a z-stack. All intensities in the region of interest (ROI) and counting of positive cells were analyzed using ImageJ software (National Institutes of Health, Bethesda, MD, USA). Dentate gyrus was taken from the level of interaural 1.98 mm, bregma−1.82 mm to the level of interaural 1.95 mm, bregma−1.85 mm. Amygdala was taken from the level of interaural 1.77 mm, bregma−2.03 mm to the level of interaural 1.74 mm, bregma−2.06 mm. External capsule was taken from the level of interaural 1.26 mm, bregma−2.54 mm to the level of interaural 1.23 mm, bregma−2.57 mm.

To identify microglia and oligodendrocytes throughout the hippocampus, coronal brain sections were stained with Iba-1 and Olig2. To capture whole brain slides, automated slide scanning was performed using AxioScan Z1 slide scanner (Carl Zeiss Microscopy) at 20x magnification. In the middle 5 μm part of the 25 μm section, images in 1 μm units were merged into a z-stack; 405-nm and 488 nm lasers were used for the excitation of DAPI and AF488 fluorophores, respectively. Images were adjusted and obtained in Zen Blue image acquisition software (Zeiss). Counting of Iba-1- and Olig2-positive cells was analyzed with ImageJ. For counting without bias, the cells were counted by two blinded observers using ImageJ. Images by cell counting were analyzed by drawing each brain region with ImageJ to obtain unit area, and the counted cells were divided by each unit area.

The microglial morphology analysis was conducted according to our previous method (Yang et al., 2022). A total of 80–120 microglial cells per group (five mice, five fields per group) were used for skeletal analysis.

The hippocampus, hypothalamus, amygdala, and mesenteric lymph nodes were analyzed for qPCR according to our previous study (Rim et al., 2022). Primer sequences are listed in Table 2.

The experimental data are presented as the mean ± error (SEM). The statistical significance of differences between groups was assessed with unpaired Student's t-tests using GraphPad Prism version 7 (GraphPad, La Jolla, CA, USA). Statistical significance was set at P < 0.05.

To examine the effect of stress hormone elevation during pregnancy, we subcutaneously administered a synthetic glucocorticoid (dexamethasone; DEX, 50 μg/kg, GD16-18, once a day for 3 days) to pregnant mice. Behavioral assessments were performed in postpartum dams after 3 weeks, considering the weaning period. There was no difference in maternal care between the vehicle and DEX-dam groups. An overview of the experimental procedure is shown in Figure 1A. DEX-dam showed a decrease in the cumulative time of the center zone compared with the Vehicle-dam in the anxiety-related OFT (Figure 1B). DEX-dam spent a significantly longer time in the dark zone than the Vehicle-dam in the LD test (Figure 1C). In SI, which reflects sociality, there was no difference between the vehicle and DEX-dam (Figure 1D). In addition, there was no difference in the SPT, which reflects anhedonia, and sucrose consumption was not different (Figure 1E). The TST and FST were conducted to evaluate depressive-like behaviors. In the TST (Figure 1F), there was no difference between groups, but in the FST, DEX-dam showed a longer immobility time than Vehicle-dam (Figure 1G). After a series of behavioral tests, mice were sacrificed for molecular analysis. DEX-dams did not show hypothalamus–pituitary–adrenal axis (HPA axis) activation. The hypothalamic corticotropin-releasing hormone (CRH) mRNA expression and serum glucocorticoid (CORT) levels did not change in DEX-dam (Figures 1H, I).

First, we examined well-known factors related to depressive- and anxiety-like behaviors in the hippocampus and amygdala. The protein levels of GR, postsynaptic density protein 95 (PSD95), synaptophysin, BDNF, pCREB, CREB, and 5-HT1AR were not changed in the hippocampus of DEX-dams (Figures 2A, B). In the immunofluorescence study, pCREB in the dentate gyrus of hippocampus was not altered, in line with the Western blot results (Figure 2C). To examine neuronal activity in the hippocampus dentate gyrus (HPC_DG), amygdala (Amyg), and paraventricular nucleus (PVN), c-Fos, a neuronal activity marker, immunoreactivity (IR) was stained, and the number of c-Fos-positive cells did not change between the two groups. However, neuronal activity tended to increase in the paraventricular nucleus (p = 0.0656) (Figure 2D). GABAergic neurons and synapses in the amygdala are known to be involved in regulating anxiety-related behaviors (Babaev et al., 2018). Thus, we examined the protein levels of glutamate decarboxylase 67 (GAD67), a GABAergic neurotransmission marker, using Western blotting. However, there was no difference in the levels of GAD67 as well as glucocorticoid receptor (GR), postsynaptic density protein 95 (PSD95), synaptophysin, BDNF, pCREB, CREB, and 5-HT1AR in the amygdala (Figure 2E).

We investigated the number and morphology of microglia in the hippocampus of DEX-dam. There was no difference in the number of microglia (Iba-1-positive cells) in the hippocampus between the vehicle and DEX-dam groups (Figure 3A). Still, the average branch length and endpoints of DEX-dam were significantly increased in morphology analysis (Figure 3B), indicating a hyper-ramified form. Next, we examined the microglial functional phenotype-related factors. The mRNA levels of microglia homeostatic genes, cytokines, phagocytosis-related genes, and kynurenine pathway-related genes were investigated using qRT-PCR. In the hippocampus, P2ry12 and Cd206 were lower in DEX-dams than in Vehicle dams, and Tgf-β showed a tendency to decrease (Figure 3C). The protein levels of CX3CR1 and Arginase1 (Arg1), which were related to stress vulnerability and depression in our previous studies, were not changed in the hippocampus (Figure 3D). In the hypothalamus and amygdala, which are regions related to depression and anxiety, there was no difference in microglia-related genes, except for C3ar1 in the hypothalamus. C3ar1 was decreased in DEX-dams (Figures 3E, F).

Along with the involvement of microglia in DEX-dams, oligodendrocytes and astrocytes were also analyzed. First, to examine the change in the number of oligodendrocytes in the hippocampus, immunofluorescence was performed with Olig2, an oligodendrocyte marker, and olig2-positive cells per ROI were counted using ImageJ. This showed that there was no difference compared with Vehicle-dam (Figure 4A). White matter abnormalities have been implicated in the pathogenesis of major depressive disorder (Tham et al., 2011; Cole et al., 2012; Zhou et al., 2021). Therefore, we conducted Western blotting to measure MBP, the most common cellular marker for myelin, in the hippocampus, but there was no difference (Figure 4B). Immunofluorescence studies also confirmed the results of Western blotting. MBP intensity was not altered in the external capsule (Figure 4C). As DEX-dam showed anxiety-like behaviors, oligodendrocytes were also analyzed in the amygdala, a brain region associated with anxiety. There was no change in the number of Olig2-positive cells and MBP intensity in the amygdala (Figures 4D, E). Next, we examined changes in astrocytes using GFAP as an astrocyte marker in the hippocampus. There was no difference in the intensity of GFAP in the hippocampus dentate gyrus (Figure 4F) between immunofluorescence and Western blots (Figure 4G). Taken together, except for microglia, other glial cells were not altered by DEX-dam.

To examine the possible involvement of the peripheral immune system in anxiety-/depressive-like behaviors shown in DEX-dams (Han et al., 2015; Kohler et al., 2017), we measured the mRNA expression of cytokines in lymph nodes. We found that the mRNA expression levels of IL-10, an anti-inflammatory cytokine, were decreased in the mesenteric lymph nodes of DEX-dam. In contrast, other cytokines (e.g., IL-1β, IL-4, IL-6, IL-17, TGF-β, TNF-α) and T-cell subtype markers (e.g., Tbx21, Foxp3, gata3, RORγ) did not change (Figure 5A). Circulating IL-1β, TNF-α, and IL-10 levels in the serum did not change (Figures 5B–D).

A series of behavioral assessments were conducted at 10 weeks postpartum to determine whether anxiety-/depressive-like behaviors were maintained for a longer period in DEX-dam (Figure 6A). In the OFT, there was no difference in the cumulative time of the center zone compared with the Vehicle-dam, and the LD test showed no difference in time spent in the dark zone compared with the Vehicle-dam (Figures 6B, C). This finding suggests that anxiety-like behaviors were restored over time. In SI, there was no difference in DEX-dam compared with Vehicle-dam consistent with the 3-week results (Figure 6D). In addition, there was no difference between vehicle and DEX-dam in the SPT and TST (Figures 6E, F). In the FST, the immobility that increased in 3-week DEX-dams was recovered at 10 weeks (Figure 6G). The mRNA expression levels of P2ry12 and Cd206, which decreased in the hippocampus at 3 weeks, were normalized at 10 weeks (Figure 6H). IL-6, IL-10, and IFN-γ mRNA levels were not different from those of the Vehicle-dam in the lymph nodes (Figure 6I). At 10 weeks, there was no difference in serum IL-10 level as at 3 weeks (Figure 6J).