Maraviroc attenuates inflammation-exacerbated cognitive and amyloid pathology in an early-stage Alzheimer’s disease mouse model

Abstract

Alzheimer’s disease (AD) is an age-related neurodegenerative disorder characterized by progressive cognitive decline, and increasing evidence indicates that systemic inflammation can accelerate disease progression. Maraviroc, a CCR5 antagonist approved for the treatment of human immunodeficiency virus (HIV) infection, has shown neuroprotective effects in several neurological contexts, but its role in AD-related pathology remains unclear. In this study, cognitive performance was assessed in 5 × FAD mice using the Y-maze, novel object recognition, novel location recognition, and social discrimination tests. Amyloid-related changes were evaluated by hippocampal APP/Aβ immunoblotting and plaque staining using 6E10 and Thioflavin S. Glial responses were examined by IBA1 and GFAP immunostaining, and inflammatory cytokines were quantified by ELISA. We found that 5 × FAD mice exhibited age-dependent cognitive impairments, with detectable deficits emerging at 3 months of age. Systemic administration of lipopolysaccharide (LPS) further exacerbated cognitive dysfunction, amyloid-related alterations, and neuroinflammatory responses in young 5 × FAD mice. Maraviroc treatment attenuated LPS-associated cognitive impairments, reduced amyloid-related measures, and dampened pro-inflammatory cytokine responses, with a trend toward reduced microglial cell density. Collectively, these findings demonstrate that systemic inflammation accelerates Alzheimer’s-like pathology and cognitive decline, and suggest that pharmacological modulation of neuroinflammatory signaling by maraviroc may mitigate inflammation-driven disease exacerbation at early stages.

Schematic diagram illustrating the effects of maraviroc on LPS-induced cognitive deficits in 3-month-old 5 × FAD mice. In this model, maraviroc is associated with modulation of glial inflammatory responses, reduced pro-inflammatory cytokine levels, and alleviation of amyloid pathology in the hippocampus, which together coincide with improved cognitive performance. Figure created with BioRender.com.

Introduction

Alzheimer’s disease (AD) is the most common age-related neurodegenerative disease and is characterized by progressive memory decline. The pathological hallmarks of AD, including amyloid β (Aβ) deposition, neurofibrillary tangles, and excessive inflammatory response, are considered as main drivers of the neurodegeneration and cognitive impairments observed in AD [1]. Clinically, age-related inflammation and chronic infection (such as viral infection, type 2 diabetes, rheumatic diseases and obesity) have been proposed as detrimental factors contributing to the progression of AD [2]. Many literatures have reported that cognitive decline in AD patients is exacerbated following SARS-CoV-2 infection, suggesting that systemic inflammatory events may contribute to more severe cognitive impairment [3, 4]. Emerging evidence from animal models also indicates that systemic inflammatory challenges can interact with pre-existing AD pathology and exacerbate disease-related phenotypes [5–7]. However, the mechanisms by which systemic inflammation influences AD progression remain incompletely understood.

Neuroinflammation is considered a critical component of AD pathogenesis. In AD patients and animal models, Aβ can stimulate microglial responses, leading to increased production of pro-inflammatory cytokines, which contribute to neurotoxicity and neuronal dysfunction [8, 9]. Elucidating how inflammatory signaling modulates early Alzheimer’s-like pathology may facilitate the identification of disease-modifying strategies. Here, we employed lipopolysaccharide (LPS), a well-established inducer of systemic inflammation, to challenge young 5 × FAD mice at an early disease stage. Our results have shown that LPS-induced inflammation might act during the developmental stage to contribute to the behavioral abnormalities and microglial responses observed in AD. This suggests that LPS-mediated systemic inflammation can be used as a model to mimic the inflammatory processes that occur during the development and progression of AD.

Maraviroc is a chemokine receptor antagonist that used to treat human immunodeficiency virus type 1 (HIV-1) infection in combination with other antiretroviral agents. Recent studies have found that maraviroc can effectively enhance memory in elderly mice [10] and promote recovery from traumatic brain injury by suppression of neuroinflammation [11]. Notably, some cognitive benefits have been observed in HIV patients whose therapy was intensified with maraviroc [12]. Whether maraviroc can mitigate cognitive impairment and AD-related pathology under inflammatory conditions, however, remains unclear. In the present study, we demonstrate that maraviroc treatment attenuates LPS-induced cognitive deficits, is associated with reduced amyloid-related alterations, and dampens microglia-associated neuroinflammatory responses in early-stage 5 × FAD mice. These findings suggest that systemic inflammation can exacerbate Alzheimer’s-like pathology and suggest that pharmacological modulation of neuroinflammatory processes may alleviate inflammation-associated cognitive and pathological changes in this AD model.

Materials and methods

Animals

5 × FAD mice (JAX MMRRC, strain B6Cg-Tg, Stock# 034848) overexpressing both mutant human amyloid beta (A4) precursor protein 695 (APP)) with the Swedish (K670N, M671L), Florida (I716V), and London (V717I) Familial Alzheimer’s Disease (FAD) mutations and human PS1 harboring two FAD mutations, M146L and L286V were used in our study. These 5 × FAD transgenic mice rapidly recapitulate major features of amyloid pathology and memory impairment in the Y-maze test [13]. The age-matched wild-type (WT) littermates were used as controls in all the experiments. All these animals were housed under standard conditions at 22 °C and a 12 h light: dark cycle with free access to food and water. Animals were assigned to experimental groups in a manner that ensured balanced distribution of sex and genotype across groups. Both male and female mice at 3, 6, and 9 months of age were included, and sex is indicated by distinct symbols in all figures. Mice at each age were derived from independent cohorts, and experiments were conducted separately for each age group. Accordingly, the data represent independent samples, and statistical analyses were performed within each age group rather than across ages.

All experimental procedures were performed in accordance with the relevant guidelines and regulations and were approved by the Committee of Ningbo University on the Ethics of Animal Experiments (Approval No. NBU20230146, May 2023).

Drug treatment

3-month-old 5 × FAD mice were chosen as the experimental model for the initial stage of AD. LPS (E. coli, serotype 055: B5, HY-D1056, MedChemExpress LLC, USA) was dissolved in sterile endotoxin-free 0.9% saline (10 mg/mL) and administered intraperitoneally at a dose of 1 mg/kg once daily for three consecutive days. Maraviroc (MVC; 20 mg/kg; M125486, Shanghai Aladdin Biochemical Technology Co., Ltd., China) or vehicle was administered intraperitoneally once daily for three consecutive days. The vehicle consisted of saline containing 5% dimethyl sulfoxide (DMSO), 40% polyethylene glycol (PEG-400), and 5% Tween-80. MVC was prepared by dissolving the compound in vehicle solution with brief sonication to ensure complete dissolution, yielding a final concentration of 2 mg/mL. The dosing regimen was selected based on previous studies demonstrating CNS-relevant efficacy of systemic MVC administration in rodent models [11, 14, 15]. Although MVC concentrations in brain tissue or cerebrospinal fluid (CSF) were not directly measured in the present study, prior pharmacokinetic studies have shown that systemically administered MVC can reach the central nervous system. In HIV-1–infected patients receiving maraviroc, measurable CSF concentrations have been reported, supporting CNS exposure following peripheral administration [16]. Mice were randomly assigned to six experimental groups: WT, 5 × FAD, WT + LPS, 5 × FAD + LPS, WT + LPS + MVC and 5 × FAD + LPS + MVC.

Behavioral tests

Behavioral experiments were conducted with mice at ages 3, 6, and 9 months, from 8:30 am to 8:30 pm. For each age group, the same cohort of mice was subjected to all behavioral tests. Behavioral assays were performed sequentially in a fixed order with sufficient intervals between tests to minimize potential carryover effects. The order of testing and the intervals between assays are illustrated in Fig. 1A. Prior to testing, each mouse was accustomed to one room and one experimenter, receiving daily 5-min sessions. Cleaning the inner wall and undersurface of the square box after each experiment to avoid the information (such as the urine and smell) left of the last animal. Behavioral data were recorded and analyzed using ANY-maze software (Stoelting Co., IL, USA), and behavioral testing was performed in a blinded manner.

Fig. 1. Cognitive-behavioral performance of 3-, 6-, and 9-month-old WT and 5 × FAD mice.

A Schematic diagram showing the order of behavioral tests and the intervals between assays. B–E The pictures shown four behavioral paradigms including open field test (OFT), Y-maze, novel location recognition (NLR) and novel object recognition (NOR), respectively. F-J 3-month-old WT and 5 × FAD mice exhibited comparable performance in the OFTs and Y-maze tests. However, the 5 × FAD mice showed a slight decrease in the discrimination index in the NLR and NOR compared to WT controls (*p < 0.05, unpaired t-test) (WT: n = 16; 5 × FAD: n = 13). K-O In 6-month-old mice, 5 × FAD mice showed the increased distance traveled in the OFTs, improved total arm entries but reduced spontaneous alternations in the Y-maze, as well as decreased discrimination index in the NLR and NOR tasks, compared to WT mice (**p < 0.01, ***p < 0.001, ****p < 0.0001, unpaired t-test) (WT: n = 17; 5 × FAD: n = 15). P–T In 9-month-old mice, 5 × FAD mice showed the reduced spontaneous alternations in the Y-maze (**p < 0.01, unpaired t-test), decreased discrimination index in the NLR (**p < 0.01, unpaired t-test) and NOR (***p < 0.001, unpaired t-test) compared with WT mice, while no differences were observed in the OFT distance and number of total arm entries in the Y-maze (WT: n = 11; 5 × FAD: n = 14). Each data point represents an individual mouse. Male mice are indicated by filled circles, and female mice by open circles.

Open field test (OFT)

The open field test (OFT) assesses locomotor activity using a white Plexiglas box (40 cm long × 40 cm wide × 40 cm high). Experiments were performed in quiet environments. Animals were placed in the center of the box to explore freely for 10 min with a video recording their movements. The total distance, time spent in the central area (20 cm × 20 cm), speed, and rest time were measured.

Y-Maze test

The Y-maze test was used to assess spatial working memory. The apparatus consisted of three identical arms (length: 40 cm, width: 10 cm, height: 25 cm) positioned at 120° angles from each other. Each mouse was placed at the end of one arm and allowed to freely explore the maze for 8 min. Arm entries were recorded when all four paws of the mouse entered an arm. Spontaneous alternation behavior was defined as consecutive entries into three different arms (e.g., ABC, BCA). The percentage of spontaneous alternation was calculated as: (number of alternations) / (total arm entries − 2) × 100%.

Novel location recognition (NLR) and novel object recognition (NOR)

The new object and new location recognition test is performed to test recognition memory in mice. The experimental apparatus for the test consisted of an open field box (25 cm long × 25 cm wide × 25 cm high), with one wall specially marked. Mice were acclimatized to the box for three consecutive days with daily 10-min sessions.

In the NLR test, 24 h after acclimatization, mice explored two identical objects, and their investigation time for each was recorded. After a 1-h interval, one object was moved diagonally, and mice explored the setup again for 5 min. Object exploration was defined as the time a mouse’s nose touched or faced the object within 2 cm. The NLR discrimination index was calculated as (novel location investigation time – familiar location investigation time) / (novel location investigation time + familiar location investigation time).

For the NOR test, the same open field box was used with two differently shaped objects made of the same material. The acclimatization and training phases were identical to those in the NLR test. After another 24 h, one original object was replaced with a new, unfamiliar one, and mice explored the two objects for 5 min. The investigation time for each object was recorded, and the NOR discrimination index was calculated similarly to the NLR test.

Social discrimination paradigm

The social discrimination paradigm is conducted to assess social recognition memory by measuring sociability, social novelty and memory ability. The device consisted of three interconnected plexiglas chambers (60 cm long × 20 cm wide × 25 cm high) with small openings (6 cm × 8 cm) in the dividing walls. On each side, a small round empty cylinder (8 cm diameter) was placed. The test comprised three stages: Habituation: Test mice freely explored the apparatus for 15 min. Training: A 4-week-old male C57BL/6 J mouse (juvenile 1) was placed in one cylinder, and the test mice explored for 5 min. Test: (1) 15 min later, another juvenile (juvenile 2) was placed in the other cylinder, and the test mice explored for 5 min. (2) 150 min later, the test mice explored for another 5 min. The 150-min test followed previously published protocols in which the same juveniles are used to assess long-term forgetting rather than novelty response, allowing evaluation of memory retention for the initially presented mouse over an extended interval [17, 18]. The time spent on exploring two juveniles was separately calculated to access the social recognition memory. The discrimination index (DI) was calculated by the following formula: (time exploring the novel mouse – time exploring the familiar mouse) / (time exploring the novel mouse + time exploring the familiar mouse).

Western blot

The hippocampal tissue samples were homogenized in ice-cold RIPA buffer supplemented with protease inhibitors, then centrifuged to collect the supernatants. Protein concentration was determined prior to gel loading, with 40 μg of protein loaded per sample. The membrane was blocked in 5% milk, in 0.05% Tween-20 in PBS (PBST) at RT, and then incubated overnight at 4 °C with the primary antibodies (6E10, 1:1000, 803001, BioLegend, USA; GFAP, 1:1000, 16825-1-AP, Proteintech, USA; IBA1, 1:1000, ab178846, Abcam, UK), and then reacted with the secondary antibodies (HRP-conjugated affinity purified goat anti-mouse IgG(H + L) or goat anti-rabbit IgG(H + L), 1:10000, SA00001, Proteintech, USA) for 2 h at RT. The protein bands were detected in a fluorescence scanner (Odyssey Infrared Imaging System, LI-COR Biotechnology, USA) by automatic exposure and the gray value of the immunoreactive protein band was analyzed with the Image J software (version 1.52a). GAPDH (Rabbit polyAb, 1:5000, 10494-1-AP, Proteintech, USA) served as an internal protein control. For Western blot analysis, membranes were cut horizontally based on molecular weight markers prior to antibody incubation. The 6E10 antibody was applied to the high-molecular-weight region (>70 kDa) to detect APP and amyloid-related species. The results were obtained from three independent replications of the experiment.

Thioflavin-S staining

This thioflavin-S staining allows for the visualization of amyloid deposits in brain tissue samples. Brains were perfused and fixed with 4% paraformaldehyde and 30% sucrose solution, then frozen into sections. Brain slices were mounted on glass slides and dried overnight. The sections were first soaked in a 0.25% KMnO₄ PBS solution for 10 min, followed by a 5-min incubation in a 2% K₂O₅S₂ 1% oxalic acid solution until the color changed. The slices were then stained with a 0.015% thioflavin-S (HY-D0972, MedChemExpress LLC, USA) 50% ethanol solution for 10 min. After washing and air-drying, the stained sections were sealed with a mounting agent and imaged using an upright fluorescence microscope (DM3000, Leica, Germany). Representative images were selected from male mice for consistency, as no significant sex differences were observed in the quantified outcomes.

Immunostaining and quantification

After fixation and cryoprotection, brains were cut using a Leica Cryostat in 30 μm thickness. Sections were transferred into a blocking solution containing 0.1% Triton X-100, 10% goat serum in PBS for 1 h at RT. For immunofluorescence staining, sections were incubated at 4 °C overnight with primary antibodies (6E10, 1:500, 803001, BioLegend, USA; IBA1, 1:500, ab178846, Abcam, UK) diluted in PBS, 0.1% Triton X-100, and 3% goat serum. After washing, secondary antibodies (CoraLite488-conjugated goat anti-mouse IgG(H + L) or CoraLite594-conjugated goat anti-rabbit IgG(H + L), 1:500, SA00013, Proteintech, USA) diluted in PBS were applied for 2 h at RT, followed by thorough washing and mounting with antifading medium (S2110, Solarbio, China). For immunohistochemical staining, after quenching endogenous peroxidase activity by incubation with 3% H₂O₂ in methanol, sections were incubated with the following primary antibodies (6E10, 1:500, 803001, BioLegend, USA; GFAP, 1:500, 16825-1-AP, Proteintech, USA; IBA1, 1:500, ab178846, Abcam, UK). Afterwards, the slices were incubated with SABC solution at 37 °C for 30 min and finally incubated with ABC Vector Elite kit containing avidin and biotin (SA1022, BOSTER Biological Technology Co.,Ltd. China).

For quantification of GFAP⁺ and IBA1⁺ cells, images of hippocampal slices stained with antibodies against GFAP or IBA1 were obtained with a microscope equipped with a 10× objective. The images were then imported into Image J. Three images spanning the hippocampus between bregma −1.46 mm and −2.46 mm per mouse were selected. The numbers of microglia and astrocytes in the hippocampus were manually counted with Cell Counter plugin, and the number of cells was divided by the total area of the acquired field to represent cell density (cells/mm²). Total plaque load was quantified as the percentage area of the hippocampus occupied by 6E10-immunoreactive material. For amyloid plaque analysis, only extracellular Aβ deposits were included, while intracellular Aβ immunoreactivity within neuronal cell bodies was excluded based on morphological criteria. Plaques were defined as discrete extracellular deposits with characteristic size and staining intensity, and area coverage was quantified using ImageJ. For the analysis of plaque-associated microglia, the number of IBA1-positive microglial cells surrounding individual Aβ plaques was quantified. Three coronal sections per mouse were analyzed, and for each section, one plaque from a comparable anatomical region was selected. The number of IBA1-positive cells within a defined perimeter surrounding each plaque was counted, and the mean value per mouse was used for statistical analysis. All image analyses were performed in a blinded manner. Representative images were selected from male mice for consistency, as no significant sex differences were observed in the quantified outcomes.

ELISA

To measure the levels of tumour necrosis factor-α (TNF-α) and interleukin (IL)-1β levels in the hippocampus, brain tissues were lysed using RIPA buffer, and the supernatants were collected for experiments. The mouse TNF-α and IL-1β ELISA kits were used to determine the levels of TNF-α and IL-1β according to the manufacturer’s instructions (Shanghai Enzyme-linked Biotechnology Co.,Ltd. China).

Statistical analysis

All statistical analyses were performed using GraphPad Prism (version 8.0.1). Sample size was determined based on previous studies using similar experimental paradigms and is consistent with standard practice in the field. No animals or samples were excluded from the analysis, and no pre-established exclusion criteria were applied. Data distribution was assumed to be normal based on standard practice in the field; formal normality tests were not performed. Variance was assumed to be similar between groups for the purpose of statistical testing. Comparisons between two groups were conducted using unpaired two-tailed Student’s t-tests. For comparisons involving more than two groups within a single experimental condition, data were analyzed using one-way ANOVA followed by Tukey’s post hoc test. Data involving repeated measurements over time were analyzed using three-way repeated-measures ANOVA with genotype, treatment, and time as factors. Endpoint data from experiments involving multiple genotypes and treatment conditions were analyzed using two-way ANOVA with genotype and treatment as factors, followed by Tukey’s multiple comparisons test where appropriate. As mice at different ages (3, 6, and 9 months) were derived from independent cohorts, statistical analyses were conducted within each age group rather than across ages. Sex was initially included as a biological variable; however, no significant main effects of sex or sex-related interactions were observed. Therefore, data from male and female mice were pooled for subsequent analyses. All data represent mean ± SEM, only values with p < 0.05 were considered statistically significant for all results.

Results

The different cognitive-behavioral phenotypes were found in 3-, 6-, and 9-month-old 5 × FAD mice

The behavioral tests including Y-maze, NLR, and NOR were used to measure different aspects of cognitive function in our study (Fig. 1A-E). At 3 months of age, 5 × FAD mice showed no significant differences from age-matched WT mice in open field and Y-maze spontaneous alternation or total arm entries (Fig. 1F–H; alternations: unpaired t-test: t ₍₂₇₎ = 0.9642; p = 0.3435; total entries: unpaired t-test: t ₍₂₇₎ = 1.295; p = 0.2064). However, modest impairments were detected in both the NLR and NOR tasks (Fig. 1I, J; NLR: t ₍₂₇₎ = 2.601; p = 0.0149; NOR: t ₍₂₇₎ = 2.747; p = 0.0106). At 6 months of age, 5 × FAD mice exhibited reduced spontaneous alternation in the Y-maze (t ₍₃₀₎ = 5.019; p < 0.0001), accompanied by impaired performance in both the NLR (t ₍₃₀₎ = 4.045; p = 0.0003) and NOR tests (t ₍₃₀₎ = 4.092; p = 0.0003) (Fig. 1K–O). Increased locomotor activity was also observed, as reflected by greater distance traveled in the open-field test (t ₍₃₀₎ = 2.901; p = 0.0069) and a higher number of total arm entries in the Y-maze (t ₍₃₀₎ = 3.129; p = 0.0039). At 9 months of age, 5 × FAD mice continued to display deficits in spontaneous alternation (t ₍₂₃₎ = 3.222; p = 0.0038) as well as reduced discrimination indices in the NLR (t ₍₂₃₎ = 3.635; p = 0.0014) and NOR tests (t ₍₂₃₎ = 4.526; p = 0.0002) compared with WT mice (Fig. 1P–T). No significant differences were observed in locomotor activity measures, including total arm entries in the Y-maze and distance traveled in the open-field test. Together, these findings indicate that 5 × FAD mice exhibit age-dependent impairments in working memory and recognition memory.

To further test the social recognition memory, the social discrimination paradigm was used in our study. As shown in Fig. 2A, the experiment consisted of three stages. In the first stage, all ages of 5 × FAD mice and their WT controls spent more time interacting with the juvenile than the empty cylinder, indicating normal social abilities (Fig. 2B, E, H). During the second stage, a novel juvenile was introduced 15 min after the initial exposure. The social memory was significantly decreased in 6-month-old 5 × FAD mice (Fig. 2F; unpaired t-test: t ₍₃₀₎ = 3.793; p = 0.0007), while 3- and 9-month-old 5 × FAD mice were comparable to WT (Fig. 2C, I). In the third stage, social memory forgetting was tested 150 min after the training. Again, only the 6-month-old 5 × FAD mice showed a reduced discrimination index, indicating impaired social memory forgetting (Fig. 2G; unpaired t-test: t ₍₃₀₎ = 2.944; p = 0.0062). Interestingly, the 9-month-old 5 × FAD mice did not differ from age-matched WT controls in both social memory formation and forgetting. This may be related to the dysfunction of social recognition memory also observed in 9-month-old WT mice compared to 6-month-old WT mice.

Fig. 2. The social recognition memory in 3-, 6-, and 9-month-old WT and 5 × FAD mice.

A Diagram of the experimental process of the social discrimination paradigm test. B-D 3-month-old WT and 5 × FAD mice exhibited comparable performance in the social recognition memory tests (WT: n = 16; 5 × FAD: n = 13). E-G In 6-month-old mice, 5 × FAD mice showed significantly lower discrimination index compared to WT controls, both 15- and 150-min after the initial social interaction (**p < 0.01, ***p < 0.001, unpaired t-test) (WT: n = 17; 5 × FAD: n = 15). H–J At 9 months of age, no differences were observed in the social recognition memory tests between 5 × FAD and age-matched WT mice (WT: n = 11; 5 × FAD: n = 14). Each data point represents an individual mouse. Male mice are indicated by filled circles, and female mice by open circles.

Maraviroc alleviated the LPS-induced cognitive impairment in 5 × FAD mice at 3 months of age

To investigate the effects of superimposed systemic inflammation on AD development and the therapeutic potential of maraviroc, as shown in Fig. 3A, we treated 3-month-old 5 × FAD mice with LPS and maraviroc sequentially. LPS, also known as endotoxin, can cause a decrease in body weight in mice, depending on the dose and strain of mouse [19]. Here, we first measured the body weight and locomotor activity in WT and 5 × FAD mice following LPS administration. No significant differences in body weight were observed among groups (Fig. 3B). In contrast, LPS-exposed mice exhibited a significant reduction in locomotor activity during the early post-injection period; however, this effect was transient and returned to baseline levels within two days (Fig. 3C; three-way repeated-measures ANOVA: time: F _{(6, 96)} = 127.1, p < 0.0001; genotype: F _{(1, 16)} = 1.205, p = 0.2885; LPS: F _{(1, 16)} = 63.40, p < 0.0001). Importantly, locomotor activity had fully normalized prior to subsequent cognitive behavioral testing, indicating that these transient changes were unlikely to confound the interpretation of the behavioral outcomes.

Fig. 3. Maraviroc attenuated LPS-induced cognitive deficits in 3-month-old 5 × FAD mice.

A Experimental timeline showing LPS and maraviroc (MVC) administration and subsequent behavioral testing. B, C Body weight and locomotor activity following LPS administration (WT / Saline: n = 5; WT / LPS: n = 5; 5 × FAD / Saline: n = 5; 5 × FAD / LPS: n = 5). Transient reductions in locomotor activity were observed in LPS-treated mice but recovered prior to behavioral testing (three-way repeated-measures ANOVA). D, E The spontaneous alternations and total arm entries in the Y-maze tests. LPS decreased the percentage of spontaneous alternations in both WT (*p < 0.05, Tukey’s test following two-way ANOVA), and MVC improved in the 5 × FAD group (**p < 0.01). F, G The discrimination index in the NLR and NOR tests. LPS induced a significant decrease in discrimination index of both WT and 5 × FAD mice (****p < 0.0001, **p < 0.01; *p < 0.05), and MVC ameliorated these deficits in the LPS-treated WT and 5 × FAD mice (*p < 0.05). H, I The discrimination index in the social discrimination paradigm tests. LPS reduced the social recognition memory in 5 × FAD mice at 150-min after the initial stage (*p < 0.05), MVC did not improve this specific deficit. (D–I: WT, n = 11; 5 × FAD + LPS, n = 13; 5 × FAD + LPS + MVC, n = 10; 5 × FAD, n = 11; 5 × FAD + LPS, n = 15; 5 × FAD + LPS + MVC, n = 11). Each data point represents an individual mouse. Male mice are indicated by filled circles, and female mice by open circles.

To assess LPS-induced cognitive impairments and the effects of maraviroc (MVC), data shown in Fig. 3D–I were analyzed using two-way ANOVA with genotype and treatment as factors. In the Y-maze test, there were significant main effects of treatment (F _{(2, 65)} = 11.82, p < 0.0001) and genotype (F _{(1, 65)} = 4.030, p = 0.0489), with no significant interaction (F _{(2, 65)} = 0.9508, p = 0.3917). Post hoc analysis revealed that LPS reduced spontaneous alternation, whereas maraviroc improved performance, particularly in 5 × FAD mice (Fig. 3D). The total number of entries in the Y-maze was unaffected by either LPS or maraviroc (Fig. 3E), suggesting the observed cognitive deficits were not confounded by changes in general locomotor activity. In the NLR and NOR tasks, which assess spatial and object recognition memory, respectively, LPS significantly reduced the discrimination index in both WT and 5 × FAD mice. Two-way ANOVA revealed significant main effects of treatment (NLR: F _{(2, 65)} = 31.24, p < 0.0001; NOR: F _{(2, 65)} = 18.61, p < 0.0001) and genotype (NLR: F _{(1, 65)} = 10.56, p = 0.0018; NOR: F _{(1, 65)} = 6.938, p = 0.0105), with a significant genotype × treatment interaction in NLR (F _{(2, 65)} = 3.223, p = 0.0463). Post hoc comparisons indicated that maraviroc significantly attenuated LPS-induced deficits in both tasks (Fig. 3F-G). In the social recognition (TCT) paradigm, LPS-treated 5 × FAD mice exhibited a selective impairment at the 150-min time point (long-term social memory), but not at 15 min (short-term memory). Two-way ANOVA revealed significant main effects of treatment (F _{(2, 65)} = 6.176, p = 0.0035) and genotype (F _{(1, 65)} = 9.750, p = 0.0027), with no significant genotype × treatment interaction. Post hoc analysis indicated that LPS impaired social recognition memory at 150 min, whereas maraviroc did not significantly improve this deficit (Fig. 3H, I).

To determine whether these effects depend on inflammatory conditions, an additional cohort of 3-month-old 5 × FAD mice treated with maraviroc alone was examined. No significant differences were observed in Y-maze, NLR, NOR, or TCT performance compared with untreated 5 × FAD mice (Supplementary Fig. 1), indicating that maraviroc does not alter baseline cognitive function in early-stage 5 × FAD mice.

Taken together, these results demonstrate that LPS-induced neuroinflammation impairs spatial working memory, recognition memory, and long-term social memory. Notably, MVC selectively ameliorates inflammation-associated deficits in spatial working and recognition memory, but does not rescue social memory impairment in 5 × FAD mice.

Maraviroc reduced LPS-associated Aβ burden in the hippocampus of 5 × FAD mice

These cognitive tasks mainly reflect the ability of hippocampus-dependent learning and memory, and the results hinted at an abnormality of hippocampal function in 5 × FAD mice at 3 months of age. Western blotting with the 6E10 antibody revealed that the relative protein levels of APP / Aβ progressively increased in the hippocampus of 3-, 6-, and 9-month-old 5 × FAD mice (Fig. 4A and B; one-way ANOVA: drug treatment, F _{(2, 19)} = 22.08; p < 0.0001). Moreover, APP / Aβ levels were found to be elevated in the 5 × FAD mice following LPS administration, and maraviroc led to a reduction in the APP / Aβ levels in this 3-month-old AD model (Fig. 4C and D; one-way ANOVA: drug treatment, F _{(2, 15)} = 5.416; p = 0.0170). Similarly, the Aβ plaque load, as visualized by immunohistochemistry staining with the 6E10 antibody and Thioflavin S (Thios-S), was increased in the hippocampus of LPS-treated 5 × FAD mice. Importantly, maraviroc was able to attenuate the LPS-induced Aβ plaque accumulation (Fig. 4E and F; one-way ANOVA: drug treatment, F _{(2, 14)} = 15.60; p = 0.0003; Fig. 4G and H; one-way ANOVA: drug treatment, F _{(2, 13)} = 8.090; p = 0.0052). We did not measure Aβ levels in WT mice, as only mouse-derived Aβ is present in these animals, and human Aβ was the primary focus of our study.

Fig. 4. Maraviroc alleviated the LPS-induced Aβ pathology in hippocampus of 5 × FAD mice.

A Representative Western blot of 6E10 in the hippocampus of 3-, 6-, and 9-month-old 5 × FAD mice. GAPDH was used as a loading control. Membranes were sectioned prior to antibody incubation, and the 6E10 antibody was applied to the high-molecular-weight region (>70 kDa). B Quantification of the bands at different ages (3 months, n = 10; 6 months, n = 6; 9 months, n = 6) (*p < 0.05, Tukey’s test following one-way ANOVA). C, D Representative Western blot of 6E10 in 3-month-old 5 × FAD mice treated with LPS or LPS + maraviroc (MVC). GAPDH served as a loading control. D Quantification showing that LPS administration further elevated Aβ levels compared to 5 × FAD controls, and this LPS-induced increase was attenuated by MVC treatment (n = 6 mice for each group) (*p < 0.05). E Representative images of Thioflavin S staining in the hippocampus of 3-month-old 5 × FAD mice treated with LPS or LPS + MVC. Scale bar, 200 μm. Representative images are from male mice and reflect the overall group trends. F Quantification of Thioflavin S-positive plaques (5 × FAD, n = 6; 5 × FAD + LPS, n = 6; 5 × FAD + LPS + MVC, n = 5) (***p < 0.001). G Representative images of 6E10 immunostaining in the hippocampus of 5 × FAD mice treated with LPS or LPS + maraviroc (MVC). Scale bar, 200 μm. Representative images were selected from male mice and reflect overall group trends. H Quantification of Aβ plaque burden (area coverage) in the hippocampus (5 × FAD, n = 10; 5 × FAD + LPS, n = 5; 5 × FAD + LPS + MVC, n = 6; *p < 0.05, **p < 0.01). Quantification was restricted to extracellular amyloid plaques, with intracellular Aβ immunoreactivity within neuronal cell bodies excluded based on morphological criteria. Each data point represents an individual mouse; male mice are indicated by filled circles and female mice by open circles.

Maraviroc attenuated LPS-induced glial inflammatory responses in 5 × FAD mice

Neuroinflammation is commonly associated with increased microglial and astrocytic responses. Western blot analysis showed that LPS treatment significantly increased the protein levels of the astrocytic marker GFAP and the microglial marker IBA1 in the hippocampus of 5 × FAD mice (Fig. 5A and B; one-way ANOVA: drug treatment, F _{(3, 20)} = 6.284; p = 0.0035; Fig. 5C and D; one-way ANOVA: drug treatment, F _{(3, 20)} = 9.050; p = 0.0005). Maraviroc treatment attenuated these LPS-induced increases at the protein level. In contrast, immunohistochemical analyses revealed more modest effects at the cellular level. LPS markedly increased the density of both GFAP-positive astrocytes and IBA1-positive microglia in the hippocampus. Maraviroc treatment showed a trend toward reduced IBA1-positive microglial cell density, although this effect did not reach statistical significance, while GFAP-positive astrocytic cell density remained largely unchanged (Fig. 5E and F; one-way ANOVA: drug treatment, F _{(3, 20)} = 10.93; p = 0.0002; Fig. 5G and H; one-way ANOVA: drug treatment, F _{(3, 20)} = 13.61; p < 0.0001). Taken together, these results suggest that CCR5 inhibition effectively modulates glial inflammatory signaling, as reflected by protein expression changes, but has a more limited impact on glial cell density under these conditions.

Fig. 5. Maraviroc alleviated the glial inflammatory response to LPS in 5 × FAD mice.

A Western blot analysis of GFAP in the WT and 5 × FAD mice with LPS or LPS + MVC treatment. GAPDH served as the loading control. B Quantification of the bands (n = 6 mice for each group) (*p < 0.05, **p < 0.01, Tukey’s test following one-way ANOVA). C Western blot analysis of IBA1 in the WT and 5 × FAD mice with LPS or LPS + MVC treatment. GAPDH served as the loading control. D Quantification of the bands (n = 6 mice for each group) (*p < 0.05, ***p < 0.001). E Representative images of GFAP staining in the hippocampus of WT and 5 × FAD mice treated with LPS or LPS + MVC. Scale bar, 200 μm. Representative images are from male mice and reflect the overall group trends. F Quantification of GFAP-positive cells expressed as cells/mm² (n = 6 mice for each group) (*p < 0.05, ***p < 0.001). G Representative images of IBA1 staining in the WT and 5 × FAD mice treated with LPS or LPS + MVC. Scale bar, 200 μm. Representative images are from male mice and reflect the overall group trends. H Quantification of IBA1-positive cells expressed as cells/mm² (n = 6 mice for each group) (*p < 0.05, ****p < 0.0001). Each data point represents an individual mouse. Male mice are indicated by filled circles, and female mice by open circles.

Maraviroc attenuated LPS-induced proinflammatory cytokine levels and microglial responses in 5 × FAD mice

To evaluate whether microglia were involved in the increased Aβ plaque, double immunofluorescence staining of IBA1 and 6E10 was performed in brain sections of 5 × FAD mice. We observed that LPS administration significantly increased number of IBA1-positive microglial cells surrounding Aβ plaques in the hippocampus of 5 × FAD mice as compared to the WT. This LPS-induced microglial reactivity was significantly reduced by maraviroc treatment (Fig. 6A and B; one-way ANOVA: drug treatment, F _{(2, 6)} = 4.622; p = 0.0610). Furthermore, LPS significantly elevated the production of the pro-inflammatory cytokines IL-1β and TNF-α in the hippocampus of 5 × FAD mice, which was suppressed by maraviroc (Fig. 6C; IL-1β, one-way ANOVA: drug treatment, F _{(3, 28)} = 8.018; p = 0.0005; Fig. 6D; TNF-α, one-way ANOVA: drug treatment, F _{(3, 28)} = 10.61; p < 0.0001). Collectively, these results indicated that LPS-mediated neuroinflammation can exacerbate Aβ pathology in the AD model, and that maraviroc has the potential to attenuate these LPS-induced changes by regulating microglia.

Fig. 6. Maraviroc attenuated microglia activity and proinflammatory cytokines in LPS-treated 5 × FAD mice.

A Images of 6E10-immunoreactive plaques (green) and anti IBA1-immunostained microglia (red) in the hippocampus of 5 × FAD mice treated with LPS or LPS + MVC. Representative images are from male mice and reflect the overall group trends. B Quantification of the number of 6E10 / IBA1 co-expression cells in the hippocampus of 5 × FAD mice treated with LPS or LPS + MVC (n = 3 mice for each group) (*p < 0.05, Tukey’s test following one-way ANOVA). C and D ELISA analysis of the IL-1β and TNF-α levels in the hippocampus of WT and 5 × FAD mice treated with LPS or LPS + MVC (n = 8 mice for each group) (*p < 0.05, ***p < 0.001). Each data point represents an individual mouse. Male mice are indicated by filled circles, and female mice by open circles.

Discussion

AD is the most prevalent age-related neurodegenerative disorder, and accumulating evidence indicates that its progression is closely associated with dysregulated systemic and central inflammatory responses, which can promote microglial over-activation and accelerate neurodegenerative processes [20]. In the present study, the 5 × FAD mouse model exhibited age-dependent cognitive deficits across multiple behavioral domains, recapitulating key features of AD progression. Notably, systemic administration of LPS induced long-lasting cognitive impairments in both WT and 5 × FAD mice, supporting the concept that inflammatory insults can exacerbate brain dysfunction. In 5 × FAD mice, these behavioral deficits were accompanied by pronounced microglial responses and elevated pro-inflammatory cytokine levels, indicating that LPS-induced neuroinflammation aggravates Alzheimer’s-like pathology. Although inflammatory markers were not directly assessed in WT mice, the observed behavioral impairments suggest that LPS exerts a broadly detrimental effect on cognitive function. Within this inflammation-driven context, treatment with maraviroc attenuated neuroinflammatory responses and associated cognitive deficits, supporting the notion that inflammation-related mechanisms contribute substantially to disease exacerbation rather than basal AD pathology per se. Importantly, maraviroc did not significantly improve cognitive performance in 5 × FAD mice in the absence of LPS, indicating that its beneficial effects are context-dependent and primarily emerge under conditions of heightened inflammation, rather than reflecting a general enhancement of baseline cognitive function.

Cognitive decline is a core clinical feature of AD and typically emerges after a prodromal stage of mild cognitive impairment, during which a substantial proportion of individuals progress to dementia [21, 22]. Consistent with previous studies [23–25], our behavioral analyses revealed a progressive, age-dependent decline in working memory, recognition memory, and social cognition in 5 × FAD mice. Subtle deficits were already detectable at 3 months of age, indicating an early stage of cognitive vulnerability, while impairments were most pronounced at 6 months. Genotype-dependent differences became less evident at 9 months, which does not indicate functional recovery in 5 × FAD mice but is more likely attributable to emerging age-related cognitive changes in WT mice that reduce behavioral contrast at later time points. Supporting this interpretation, C57BL/6J mice, the genetic background strain of the WT controls used in this study, have been reported to exhibit subtle age-associated declines in social recognition and other cognitive functions beginning as early as 8–12 months of age, particularly in tasks with higher mnemonic demand or repeated testing [26–28]. Such early, task-sensitive changes can lead to ceiling or floor effects, thereby diminishing genotype-dependent differences. Accordingly, the lack of significant differences between WT and 5 × FAD mice at 9 months in the 15-min and 150-min social discrimination tests is most likely explained by reduced behavioral resolution rather than true normalization of cognitive function. Within this early vulnerability window, systemic inflammatory challenge further unmasked latent cognitive deficits. LPS exposure selectively exacerbated cognitive impairment in 3-month-old 5 × FAD mice, supporting the notion that early Alzheimer’s-like pathology is particularly susceptible to inflammatory insults.

Neuroinflammation is widely implicated as a critical contributor to neurodegenerative processes and cognitive dysfunction [29, 30]. LPS is commonly employed to model systemic inflammatory challenges, with both acute and chronic exposure shown to induce neuroinflammation and impair cognition in rodents [31, 32]. In the present study, repeated systemic LPS administration elicited prolonged neuroinflammatory responses in young 5 × FAD mice, leading to persistent cognitive deficits. Mechanistically, LPS activates innate immune signaling cascades, primarily through toll-like receptor 4 (TLR4), triggering downstream NF-κB–dependent transcription of pro-inflammatory cytokines. This results in sustained microglial abundance and amplification of inflammatory signaling, which in turn exacerbates amyloid pathology and synaptic dysfunction. These findings are consistent with previous studies showing that inflammatory insults can accelerate disease progression, particularly at early stages when baseline deficits remain modest [7, 33, 34]. The chemokine co-receptor (CCR5) plays an important role in coordinating immune cell recruitment and inflammatory signaling in the central nervous system [35]. Increasing evidence suggests that CCR5 signaling not only facilitates microglial recruitment to sites of pathology but also shapes their activation state, thereby influencing the balance between protective and detrimental immune responses. In the context of LPS exposure, CCR5 signaling may act in concert with TLR4-dependent pathways to sustain and amplify inflammatory cascades [36]. Although the precise molecular crosstalk remains to be fully elucidated, it is plausible that CCR5 contributes to the maintenance of a pro-inflammatory microenvironment downstream of TLR4 activation, thereby promoting chronic neuroinflammation. Consistent with this notion, pharmacological inhibition of CCR5 with maraviroc showed a trend toward reduced microglial cell density, reduced pro-inflammatory cytokine production, and limited amyloid deposition in LPS-treated 5 × FAD mice. These effects were accompanied by improvements in spatial working and recognition memory. Notably, CCR5 signaling has also been implicated in neuronal plasticity and cognitive function, with receptor activation reported to negatively regulate learning and memory processes [37, 38]. In line with previous reports demonstrating cognitive benefits of maraviroc in conditions such as HIV-associated neurocognitive disorders, aging, and traumatic brain injury[10, 11, 39]. our findings further support a role for CCR5 inhibition in modulating both neuroinflammatory and cognitive outcomes. Importantly, maraviroc did not significantly affect cognitive performance in 5 × FAD mice under basal conditions, indicating that its effects are context-dependent and primarily evident in the presence of heightened inflammatory signaling. This suggests that CCR5 blockade mitigates inflammation-driven exacerbation of pathology rather than reversing established neurodegenerative changes. While the precise mechanisms remain to be fully defined, our data support a model in which CCR5 inhibition dampens microglia-mediated inflammatory amplification, potentially by interfering with CCR5–TLR4 signaling interactions and downstream cytokine cascades. Together, our results support the view that targeting neuroinflammatory processes may represent a viable strategy for alleviating cognitive dysfunction associated with AD, particularly during early disease stages characterized by heightened sensitivity to inflammatory challenges.

Aβ accumulation has been widely recognized as a central driver in the pathogenesis and progression of AD [40]. It has been reported that Aβ promotes the release of inflammatory factors, and inflammation can, in turn, affect Aβ levels [41]. Repeated injection of LPS has been shown to result in an accumulation of Aβ in the hippocampus of mice brains, through increased β- and γ-secretase activities and increased expression of amyloid precursor protein [42, 43]. LPS exposure promoted neuronal damage and Aβ deposition in mouse and cell line models [44], and to increase soluble Aβ and Aβ diffuse plaques in the brain [45]. However, no changes were observed in the Aβ levels in LPS-treated APP/PS1 transgenic mice [7], and a single intraperitoneal injection of LPS in 13-month-old transgenic mice resulted in a decrease in insoluble Aβ and an increase in soluble Aβ [46]. These discrepancies may be related to factors such as the timing and duration of LPS treatment, as well as the specific transgenic mouse model used. Here, we observed that LPS-induced neuroinflammation was associated with increased Aβ burden in the hippocampus of 3-month-old 5 × FAD mice. At this early stage, previous studies have shown that amyloid deposition initially emerges in the dentate gyrus (DG) and subsequently spreads to other hippocampal subregions, including CA1 [47]. Consistent with this progression, Thioflavin S staining revealed robust amyloid deposition in the DG but minimal plaque burden in CA1. This regional pattern aligns well with the recognition memory deficits observed in the NLR and NOR tasks, which critically depend on DG-mediated novelty detection and pattern separation [48, 49]. Together, systemic inflammation induced by LPS was associated with enhanced amyloidogenic processing in 5 × FAD mice. Notably, maraviroc treatment attenuated these LPS-associated pathological alterations, suggesting that suppression of inflammation-driven exacerbation of Aβ generation and plaque accumulation may contribute to the observed behavioral benefits.

Microglia and astrocytes, as key components of the brain’s innate immune system, play central roles in AD-associated neuroinflammation and amyloid pathology [50]. In this study, LPS induced robust responses of both glial populations, with microglia displaying pronounced clustering around Aβ plaques and elevated cytokine production. Although the increased cytokine levels likely reflect the combined responses of both glial cell types, microglia appear to be the predominant contributors in this context, given their marked responses and close spatial association with Aβ plaques. Microglia can exert both protective and detrimental roles during AD progression. Plaque-associated microglia can restrict plaque expansion and limit local neurotoxicity. [51, 52]. In the present study, LPS treatment increased microglial clustering around amyloid plaques, which may represent a compensatory response to contain pathology. However, this was accompanied by enhanced inflammatory signaling, suggesting a shift toward a context-dependent, pro-inflammatory activation state. Notably, this response was not associated with efficient Aβ clearance, indicating a transition toward a disease-exacerbating phenotype. CCR5 is a key regulator of microglial chemotaxis and inflammatory responses [53]. Correspondingly, maraviroc did not significantly reduce overall microglial abundance but showed a trend toward decreased microglial immunoreactivity. Despite this modest effect on microglial area, maraviroc modulated inflammatory responses, as reflected by reduced cytokine levels. These findings, together with its limited effects on astrocytic responses, likely reflect a greater functional dependence of microglia on CCR5-mediated signaling, whereas astrocytes engage multiple parallel inflammatory pathways [54, 55]. Importantly, the present data do not allow definitive conclusions as to whether the reduction in Aβ burden reflects a direct effect of CCR5 blockade on amyloidogenic processing or occurs secondary to suppression of neuroinflammation. Nonetheless, extensive evidence indicates that elevated pro-inflammatory cytokines, such as IL-1β and TNF-α, promote amyloidogenic APP processing by upregulating β- and γ-secretase activity and impairing Aβ clearance mechanisms [56–58]. In addition, changes in microglial functional states influence Aβ handling, including phagocytosis and degradation [59]. Therefore, the reduced amyloid burden observed following maraviroc treatment is more plausibly attributed to partial restoration of microglial functional homeostasis, rather than direct suppression of Aβ production.

Several limitations of the present study should be acknowledged. First, behavioral changes at 3 months of age were modest, reflecting an early disease stage that may limit detection of robust therapeutic effects. Future studies examining older, more symptomatic AD models will be necessary to determine whether maraviroc confers benefit beyond inflammation-driven acceleration. Second, the current work focused on Aβ pathology, whereas previous studies indicate that LPS-induced inflammation can also exacerbate tau pathology in models such as 3 × Tg-AD mice [6]. Evaluation of maraviroc in AD models that recapitulate both amyloid and tau pathology will be essential for a more comprehensive assessment of its therapeutic relevance. Importantly, the present study was designed to investigate inflammation-driven exacerbation of Alzheimer’s-like pathology. Accordingly, maraviroc was evaluated under conditions of LPS-induced neuroinflammation rather than basal disease states. Consistent with this design, we did not observe significant effects of maraviroc on baseline cognitive performance in the absence of LPS (Supplementary Fig. 1), suggesting that its benefits are context-dependent. However, whether maraviroc influences cognitive or pathological outcomes under non-inflammatory conditions requires further investigation. Finally, although behavioral assessments were conducted in WT mice, histological and biochemical analyses were not performed in WT mice treated with maraviroc. As a result, the effects of maraviroc under non-pathological conditions remain incompletely characterized and warrant further study.

Conclusions

Collectively, our findings demonstrate that systemic inflammatory challenges can exacerbate cognitive impairment and amyloid pathology in the 5 × FAD mouse model, highlighting the vulnerability of early-stage Alzheimer’s-like pathology to neuroinflammatory insults. Treatment with maraviroc attenuated LPS-induced pro-inflammatory cytokine production and the associated cognitive and amyloid alterations, with a trend toward reduced microglial cell density. These findings indicate that modulation of inflammation-related signaling pathways can mitigate inflammation-driven worsening of Alzheimer’s-like pathology. Together, our results underscore the importance of neuroinflammatory processes in shaping AD progression and support further investigation of inflammation-targeted strategies across different disease stages and pathological contexts.

Author contributions

CL and XZ conceived and designed the study. CL acquired the data and conducted data curation. All authors participated in formal analysis and investigation. TZ, EE, TX, JL and FY made the visualization. CL and QS made the validation. CL and XZ wrote the original draft. ZZ, HS, and XZ reviewed, supervised, and edited the manuscript. XZ provided funding and conducted scientific review. All authors have read and agreed to the publication version of the manuscript, consenting to take individual responsibility for their contributions and ensuring the accuracy or completeness of any questions related to any part of the work. All authors have read and approved the final manuscript.

Funding

This work was supported by grants from the National Natural Science Foundation of China (82201322), the Natural Science Foundation of Zhejiang Province (LY24H090001) and the Municipal Key R&D Program of Ningbo (2023Z175).

Data availability

All the data used and/or analyzed in this study are presented in the main text and additional files. Additional data and materials are available from the corresponding author upon reasonable request.

Competing interests

The authors declare no competing financial interests.

Ethics approval and consent to participate

All experimental procedures were performed in accordance with the relevant guidelines and regulations and were approved by the Committee of Ningbo University on the Ethics of Animal Experiments (Approval No. NBU20230146, May 2023).

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41398-026-04215-y.