Microglial depletion rescues spatial memory impairment caused by LPS administration in adult mice

Animal care and housing

C57BL/6J mice were chosen for this study due to their well-characterized genetic background, consistent immune response, and susceptibility to LPS-induced neuroinflammation (da Silva et al., 2024; Piirsalu et al., 2022). We obtained C57BL/6J mice at 8 weeks of age from Vital River Laboratory Animal Technology Co. (Beijing, China). Upon arrival, the animals were acclimatized for 2 weeks before the start of the experiments. Mice were housed in groups of 4 per cage under constant conditions of temperature (21 ± 2 °C) and humidity (50 ± 10%), with a 12-h light/dark cycle, and ad libitum access to food and water. All behavioral experiments were conducted using male adult mice (3–4 months old, weighing 25–30 g), during the light cycle from 9:00 am to 6:00 pm. The Chancellor’s Animal Research Committee at Qingdao University approved all animal protocols (#3207090367) used in this study, following the National Institutes of Health guidelines.

LPS administration

Intraperitoneal injections of 0.5 mg/kg LPS (Sigma, Tokyo, Japan) were given daily for 7 days, with the dosage adjusted based on previous studies (Zhao et al., 2019). The final LPS injection was done 24 h before the start of behavioral experiments. Injections were performed with a 27-gauge needle in the lower right quadrant of the abdomen to avoid vital organs, using a volume not exceeding 10 mL/kg. Mice were gently restrained during injections to minimize stress, and their condition was closely monitored afterward for any signs of distress or adverse reactions. If any animal showed discomfort, appropriate care was provided, including human euthanasia as per guidelines of the Chancellor’s Animal Research Committee at Qingdao University. No anesthesia was given for i.p. injections.

PLX5622 and minocycline treatment

PLX5622 is a highly selective CSF1R inhibitor known for its effectiveness in depleting microglia (Spangenberg et al., 2019). Mice were fed with PLX5622 (1,200 mg/kg) or control chow (Dyets Inc., Bethlehem, PA, USA) for 3 weeks. The 3-week administration period is chosen based on both previous studies (Badimon et al., 2020) and our preliminary results, which demonstrate that this duration is minimal to achieve significant microglial depletion without causing adverse effects on overall health. The intraperitoneal injections of 100 mg/kg minocycline (Sigma, Tokyo, Japan) were administered daily for 7 days (Wang et al., 2020; Dagher et al., 2015). The injection protocol and post-injection monitoring for minocycline were same as for LPS injection.

Different batches of mice were used in the following in vivo and ex vivo experiments. The hippocampal tissues used in different studies were obtained from separated batches of mice brains. Double-blind experiments were conducted by separate groups of researchers to minimize bias and ensure objective results. Specifically, JL and MD administered the drugs. NL conducted the behavioral and electrophysiological assessments. FH carried out the immunostaining, quantitative reverse transcription PCR, and enzyme-linked immunosorbent assays. Data analysis was handled by TZ, VL and MZ: TZ and VL analyzed the behavioral results, VL analyzed the electrophysiological results, and MZ analyzed the remaining results. NL, FH, TZ, VL and MZ were blinded to the treatment conditions during all assessments and data analyses.

Behavioral assessments

All behavioral assessments were done according to our previous studies (Lu et al., 2019; Liu et al., 2024). To evaluate the overall behavioral state of the mice and ensure that differences in learning and memory performance are not attributed to changes in motor function or anxiety levels, we first conducted the Elevated Plus Maze (EPM) and Open Field (OF) tests before cognitive assessment. The EPM consists of a central area, two closed arms with walls (16.5 cm height), and two open arms without walls. Each arm measures 30 cm in length and 6 cm in width, while the main frame stands at a height of 50 cm from the ground. Mice were released from the center and allowed to freely explore the maze for 10 min. The time spent in each arm and total travel distance were analyzed.

The OF test was conducted in a 28 × 28 × 35 cm square arena, divided into a 12 × 12 cm center zone and the left periphery zone. Mice were released from the center and allowed to freely explore for 10 min. The total distance traveled and time spent in both zones were analyzed.

The Morris water maze (MWM) circular water pool is 120 cm in diameter and 30 cm deep, divided into four quadrants. Mice undergo training with an invisible platform placed just below the water’s surface, aiming to climb onto it to escape the water. Training consists of 4 trials/2 blocks per day for 7 days, with an inter-block interval of 1 h and an inter-trial interval of 30 s. A trial ends when mice reach the platform or after a cut-off time of 60 s. To assess spatial memory, a probe test is conducted 1 h after training on the 3^rd, 5^th, and 7^th days respectively. In these tests, the hidden platform is removed and mice are allowed to navigate the pool for 60 s.

Mice were acclimated to the experimental environment for at least 1 h before behavioral tests. Their behaviors were recorded and analyzed using Noldus EthoVision XT software by two independent researchers unaware of the treatment details.

Electrophysiological recordings ex vivo

Mice were deeply anesthetized with 3% isoflurane and intracardially perfused with saline. Coronal hippocampal slices (300 μm thick) were freshly prepared using a Leica VT-1000 vibratome as described previously (Cui et al., 2016). Slices were perfused with 32 °C artificial cerebrospinal fluid (ACSF) containing 120 mM NaCl, 1.25 mM NaH₂PO₄, 3.5 mM KCl, 2.5 mM CaCl₂, 26 mM NaHCO₃, 1.3 mM MgSO₄, and 10 mM glucose. Field excitatory postsynaptic potentials (fEPSPs) at the hippocampal Schaffer collateral-CA1 (SC-CA1) pathway were triggered with a FHC stimulating microelectrode (Li et al., 2022b, 2023b). The input-output (I/O), paired-pulse ratio (PPR) and long-term potential (LTP) were recorded. LTP was induced by 100 Hz stimulation. All stimulating pulses were 100 μs in duration. Data were acquired using a MultiClamp 700B amplifier and pCLAMP 10.0 software (Molecular Devices, San Jose, CA, USA). All chemicals used were purchased from Sigma.

Immunostaining

Immunostaining was done according to our previous study (Li et al., 2023b; Zhao et al., 2014). Mice were deeply anesthetized with 3% isoflurane and then intracardially perfused with saline and 4% paraformaldehyde. The brains were then post-fixed in 4% paraformaldehyde for 4–6 h, dehydrated in 30% sucrose for 48 h, and sectioned into 40 μm slices. The CA1 region of the dorsal hippocampus were identified based on anatomical landmarks. Coronal sections of the dorsal hippocampus were taken from bregma −1.34 to −2.8 mm according to mouse brain atlas by Paxinos & Franklin (2019), using the corpus callosum for orientation. The CA1 region appears as a crescent shape running parallel to the curved edge of the dorsal hippocampus, characterized by a thin band of densely packed pyramidal neurons. The border between the CA1 and CA3 regions is identifiable by changes in the density and orientation of pyramidal cells. These brain sections were incubated overnight at 4 °C with a primary anti-Iba1 antibody (1:400; Millipore, Burlington, MA, USA), followed by a 1-h incubation with FITC-conjugated donkey anti-mouse secondary antibody (Jackson Immuno Research, West Grove, PA, USA) at room temperature. Images were captured using a virtual slide microscope (VS120; Olympus, Tokyo, Japan) equipped with a 20× objective lens. Three representative coronal sections spaced equally along the anteroposterior axis of the dorsal hippocampus and three representative images taken per section were selected for quantification purposes. The density of microglia in the hippocampus was analyzed as Iba1⁺ cells/mm² using ImageJ 1.5 software.

Quantitative reverse transcription PCR (RT-qPCR)

Quantitative reverse transcription PCR was done according to our previous study (Liu et al., 2024). Mice were deeply anesthetized with 3% isoflurane. The hippocampal tissue was freshly isolated and quickly collected in enzyme-free EP tube and stored at −80 °C until use. Total RNA was extracted using the PureLink RNA Mini Kit (Thermo Fisher Scientific, Waltham, MA, USA), followed by quantity and quality measurement with a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). Single-stranded cDNA was synthesized from 1 µg of RNA with SuperScript III reverse transcriptase (Invitrogen, Waltham, MA, USA). The reaction condition was as follows: 25 °C for 10 min, 50 °C for 30 min, 85 °C for 5 min. PCR-based quantification of transcripts was performed using a Thermal Cycler Dice Real Time System (Roche, Indianapolis, IN, USA) and QuantiFast SYBR Green PCR kit (Qiagen, Hilden, Germany). The PCR cycling parameters were as follows: initial denaturation at 95 °C for 10 min followed by 40 cycles of PCR reaction at 95 °C for 15 s, 60 °C for 1 min, and 72 °C for 1 min. The primer sequences used were as follows: Iba-1-F GACGACCCTTCTTCGGGTTT, Iba-1-R GAGAGCCCACAATCTTGCCT; Il-6-F GCCTTCTTGGGACTGATGCT, Il-6-R GCCATTGCACAACTCTTTTCTCA; Cd68-F ACTTCGGGCCATGTTTCTCT, Cd68-R GGGGCTGGTAGGTTGATTGT; Tmem119-F AGCCTACTACCCATCCTCGT, Tmem119-R CTGGGTAGCAGCCAGAATGT; Bdnf-F GGCTGACACTTTTGAGCACGTC, Bdnf-R CTCCAAAGGCACTTGACTGCTG; Syp-F TCCTGCAGAACAAGTACCGAGA, Syp-R GGCCATCTTCACATCGGACAG. 2^−ΔΔCT method was used to normalize CT values against housekeeping gene Gapdh and to quantify relative expression. Triplicates were done for each sample. Assays were carried out in our own lab by investigators unaware of experimental design.

Enzyme-linked immunosorbent assay (ELISA)

Enzyme-linked immunosorbent assays (ELISAs) were done according to our previous study (Guo et al., 2019). The hippocampal tissue was freshly isolated and homogenized in 0.5 ml of ELISA buffer and centrifuged at 3,500× g for 10 min. The resulting supernatant was collected and stored at −80 °C until use. Concentrations of IL-6, Aβ1-40, and Aβ1-42 in the hippocampus were measured using mouse ELISA kits from Wuhan Colorful Gene Biological Technology Co., China. Absorbance values were measured at 450 nm using a 96-well microplate spectrophotometer, with triplicates performed for each sample.

Euthanasia

Humane endpoints were set to ensure animal welfare. Mice displaying severe distress (e.g., over 20% weight loss, inability to eat or drink, severe lethargy, or unresponsiveness) were immediately euthanized using carbon dioxide (CO₂) asphyxiation for 5 min followed by cervical dislocation. No mice met these criteria. At the study’s end, any remaining mice not needed for further research were humanely euthanized using CO₂ for 5 min to ensure minimal stress and suffering.

Statistical analyses

Data were expressed as mean ± S.E.M. and analyzed with GraphPad Prism 9.0 software. Sample size, ANOVAs or t-tests used for statistical analyses were described, with normal distributions and equal variances confirmed before performing parametric statistical analyses. A significance level of P < 0.05 was considered statistically significant.

Adult depletion of microglia with PLX5622 causes mild spatial memory impairment

Consistent with previous reports (López-Aranda et al., 2023), we initially demonstrated that treatment with PLX5622, a highly selective brain-penetrant inhibitor of the CSF1R, for 3 weeks resulted in widespread depletion of microglia throughout the brain, including approximately 83% reduction in the hippocampus (Figs. 1A–1C; Unpaired t test, PLX5622 chow vs. control chow, P < 0.0001). Under physiological conditions, microglial depletion in adult mice did not impact locomotor activity or anxiety-related behaviors (Figs. 1E and 1F), nor did it affect spatial learning (Fig. 1G) or swimming speed (Fig. 1I) in the WMW task. However, during the probe test, it was found that while both groups of mice were able to form long-term spatial memories (Fig. 1H; one sample t test, P < 0.001 in comparison to random 50%), PLX5622-treated mice spent less time searching in the training quadrant than controls (Fig. 1H; Unpaired t test, PLX5622 vs. CON, t = 2.61, P < 0.05), indicating mild impairment in spatial memory.

Depletion or inactivation of microglia blocks LPS-induced spatial memory impairment

In previous studies, we have demonstrated that systemic administration of LPS (0.5 mg/kg) daily for 7 days impairs spatial learning and memory in C57BL6 mice (Liu et al., 2024). To investigate the effect of microglial depletion on LPS-induced memory impairment, we initiated PLX5622 treatment 2 weeks prior to LPS administration. MWM training started 24 h after the conclusion of LPS injection (Jung et al., 2023) (Fig. 2A). Our results indicate that microglial depletion prevented LPS-induced spatial learning and memory impairment (Figs. 2B–2D). Specifically, PLX5622-pretreated mice (PLX5662+LPS) exhibited reduced latency to platform compared to controls (CON+LPS) over the course of 7 training days (Fig. 2B; Two-way repeated measure ANOVA with Sidak’s multiple comparisons test, P < 0.05 to P < 0.001 from training day 5 to day 7). During the probe test on day 7, PLX5622+LPS mice spent a significantly higher percentage of time navigating the training quadrant than CON+LPS mice (Fig. 2C; unpaired t test, t = 3.74, P < 0.01). Both groups of mice displayed similar swimming speed during the probe test (Fig. 2D). Therefore, our findings demonstrate that adult depletion of microglia prevents LPS-induced impairment in both spatial learning and spatial memory.

Systemic administration of LPS has been reported to induce significant microglial activation in the hippocampus (Zhao et al., 2019; Jung et al., 2023). We sought to inhibit microglial activation by administering minocycline during LPS insult. Previous research has demonstrated that minocycline, the broad-spectrum tetracycline antibiotic with pleiotropic effects, is capable of inhibiting microglial activation and M1 polarization (Schmidtner et al., 2019; Kobayashi et al., 2013). We found that pretreatment with minocycline for 1 week replicated the blocking effect of PLX5622 on LPS-induced cognitive dysfunction (Figs. 2E–2G). Similar to PLX5622, minocycline administration reduced latency to platform during MWM training (Fig. 2E; Two-way repeated measure ANOVA with Sidak’s multiple comparisons test, minocycline+LPS vs. vehicle+LPS, P < 0.05 to P < 0.01), and increased training quadrant searching time during the probe test (Fig. 2F; Unpaired t test, minocycline+LPS vs. vehicle+LPS, t = 3.40, P < 0.01). Notably, minocycline treatment did not affect swimming speed (Fig. 2G). Overall, our results demonstrate that either depletion or inactivation of microglia can protect against LPS-induced spatial learning and memory impairment, highlighting the crucial role of proinflammatory activation of microglia in mediating LPS-induced cognitive deficits.

PLX5622 treatment protects against LPS-induced synaptic dysfunction in the hippocampus

Next, we examined the possible impact of microglial depletion on LPS-induced synaptic dysfunction in hippocampal SC-CA1 pathway. Mice were fed with PLX5622 or control chow for 3 weeks prior to ex vivo field recordings in acute hippocampal slices isolated from LPS-insulted mice (Fig. 3A). Our results revealed that microglial depletion led to a slight increase in basal synaptic transmission (Fig. 3B; Two-way repeated measure ANOVA with Sidak’s multiple comparisons test, PLX5622+LPS vs. CON+VEH, P < 0.0001 at 100 µA stimulation intensity) and a slight facilitation of paired-pulse ratio (Fig. 3C; PLX5622+LPS vs. CON+VEH, P < 0.05 at ISI of 50 ms) in SC-CA1 synapses. Notably, microglial depletion ameliorated LPS-induced LTP deficit (Figs. 3D–3F). Specifically, both the initial 5-min post-tetanic potentiation (PTP) and the last 20-min LTP were greater in PLX5622+LPS mice than in CON+LPS mice (Fig. 3B; Unpaired t test, PTP: t = 3.64, P < 0.01; LTP: t = 3.26, P < 0.01). Therefore, our findings indicate that microglial depletion protects against LPS-induced synaptic dysfunction in the hippocampus, contributing to memory improvement.

PLX5622 treatment suppresses LPS-induced proinflammatory activation of microglia in the hippocampus

Consistent with previous findings (Skrzypczak-Wiercioch & Sałat, 2022; López-Aranda et al., 2023), our RT-qPCR assays revealed upregulation of Iba-1, Il-6, and cd68, as well as downregulation of Tmem119 in the hippocampus following systemic LPS administration (Figs. 4A–4D; one sample t test, CON+LPS mice vs. control naïve mice, P < 0.01), indicating proinflammatory activation of microglia. Additionally, we observed reduced expression of Bdnf and Synaptophysin (Syp) in the hippocampus of LPS-treated mice (Figs. 4E and 4F; one sample t test, CON+LPS vs. control naïve mice, P < 0.01), suggesting synaptic deficit due to LPS insult. Notably, PLX5622 pretreatment inhibited LPS-induced microglial activation. Specifically, PLX5622+LPS mice exhibited significantly reduced expression of Iba-1, Tmem119, Il-6, and cd68 in the hippocampus (Figs. 4A–4D; Unpaired t test, PLX5622+LPS vs. CON+VEH, P < 0.01 to P < 0.0001), along with a dramatic increase in Syp expression (Fig. 4F; Unpaired t test, t = 8.01, P < 0.0001). Moreover, our ELISA assays showed decreased levels of IL-6 (Fig. 4G) and Aβ1-40 (Fig. 4I) in the hippocampus of PLX5622+LPS mice (Unpaired t test, PLX5622+LPS vs. CON+LPS, IL-6: t = 4.20, P < 0.01; Aβ1-40: t = 4.61, P < 0.01), indicating that microglial depletion can mitigate LPS-induced microglial activation, which may contribute to improved synaptic and cognitive dysfunction.