Sex-dependent modulation of social distance by lipopolysaccharide-induced inflammation in mice

Abstract

Depression is a major psychiatric disorder, and accumulating evidence indicates that inflammatory processes contribute to its pathophysiology. Peripheral administration of lipopolysaccharide (LPS) reliably induces systemic inflammation and is widely used to model inflammation-related depression and sickness behavior. However, despite the strong association between social behavior and depressive states, the impact of LPS-induced inflammation on social interactions remains insufficiently understood. Here, we investigated how acute inflammatory responses influence social and non-social behaviors in C57BL/6 J mice following intraperitoneal LPS administration. Immunological analyses using ELISA and flow cytometry revealed marked increases in circulating IL-1β, IL-6, and TNF-α, accompanied by reductions in T cells, B cells, neutrophils, and monocytes. Baseline levels of B cells, neutrophils, and monocytes differed between sexes. Behavioral assessments demonstrated that LPS-treated male mice exhibited increased social contact and reduced social distance in both familiar and unfamiliar dyads, whereas these effects were absent in female pairs. In contrast, LPS induced comparable reductions in body weight, locomotor activity, and fecal output in the open-field test, as well as decreased sucrose preference and overall licking counts in the sucrose preference task, in both sexes. These findings indicate that LPS-induced inflammation modulates social behavior in a sex-dependent manner. Notably, the enhanced social contact observed in male dyads cannot be attributed to weight loss, hypoactivity, altered orofacial movements, or reduced reward sensitivity, as these physiological and motivational impairments were similarly present in both sexes. This study highlights distinct social behavioral consequences of inflammatory activation and advances our understanding of immune–behavior interactions.

Introduction

Depression is among the most prevalent and debilitating mental disorders worldwide, affecting approximately 3.8% of the adult population. The incidence is particularly high among women, with around 10% experiencing depression during or after childbirth [1]. It is characterized by a persistent low mood, a loss of interest or pleasure in nearly all activities, sleep disturbances, changes in appetite, fatigue, diminished concentration, feelings of guilt or worthlessness, and, in severe cases, suicidal ideation [2]. The impact of this disorder extends beyond the individual, influencing social and economic dimensions on a global scale. As such, understanding the underlying mechanisms of depression and developing effective treatments are of paramount importance. Despite extensive research, the precise mechanisms underlying depression remain elusive, involving a complex interplay of genetic predisposition, environmental stressors, and neurobiological factors [3]. The traditional focus on monoaminergic systems, particularly the dysregulation of serotonin, norepinephrine, and dopamine, has driven much of the pharmacological treatment approaches [4]. However, emerging evidence suggests that inflammation also plays significant roles in the pathogenesis of depression.

The association between inflammation and depression was first pointed out by Smith (1991), who suggested that cytokines released by macrophages may contribute to the development of depression [5]. This hypothesis was based on clinical observations of an increased risk of depression in patients with rheumatoid arthritis, as well as reports of lung cancer patients meeting the diagnostic criteria for depression during treatment with Tumor Necrosis Factor-α (TNFα). Recent research has highlighted the intricate relationship between the immune system and the brain, revealing that inflammation may contribute to the development of depressive symptoms. Elevated levels of pro-inflammatory cytokines, such as interleukin-6 (IL-6) and tumor necrosis factor-alpha (TNF-α), have been consistently observed in individuals with depression [6–8]. These inflammatory markers are thought to influence neural circuits involved in mood regulation, potentially disrupting neurogenesis, synaptic plasticity, and neurotransmitter function.

One of the most widely used experimental approaches to investigate the relationship between inflammation and depression is the administration of lipopolysaccharide (LPS), a potent immunogenic component derived from the outer membrane of Gram-negative bacteria. Intraperitoneal administration of LPS has been extensively utilized in preclinical animal studies investigating inflammatory depression [9, 10]. LPS, a component of the outer membrane of gram-negative bacteria, is released upon the death of these bacteria within the host organism. Once released, LPS binds to toll-like receptors, leading to the activation of immune cells and the subsequent induction of a robust immune response, characterized by the release of pro-inflammatory cytokines such as IL-1β, IL-6, and TNF-α. In animal models of LPS-induced depression, elevated levels of cytokines in both the blood and brain have been observed [11, 12]. Additionally, LPS-treated mice exhibit significant reductions in body weight [13, 14], decreased spontaneous locomotor activity [15, 16], increased immobility in forced-swim and tail-suspension tests [17–19], and reduced sucrose preference [20, 21]. These findings confirm that LPS-induced inflammatory processes elicit depressive-like and sickness behaviors in animals. However, there remains a paucity of research on the effects of LPS administration on social behavior [10].

In this study, we investigated the effects of inflammatory responses on both social and non-social behaviors in male and female C57BL/6 J mice following intraperitoneal administration of LPS at a dosage of 1.0 mg/kg. First, we conducted immunological assessments using ELISA and flow cytometry to measure LPS-induced cytokine responses, including IL-1β, IL-6, and TNF-α, as well as the responses of immune cells such as T cells, B cells, neutrophils, and monocytes. We also monitored long-term changes in body weight. Second, we evaluated social behavior by measuring contact time and interaction distance in a dyadic social interaction task, involving both familiar and unfamiliar mouse pairs. Third, we assessed LPS-induced alterations in reward sensitivity through a sucrose preference test. Finally, we examined changes in locomotor activity, time spent in the center of the open-field, and defecation in an open-field test.

Methods

Subjects

All data were collected from adult male and female wild-type C57BL/6 J mice. The animals were maintained on a 12:12 h light/dark cycle, and all experiments were conducted during the dark phase. Mice had ad libitum access to food and water in their home cages. All animal procedures were approved by the Animal Research Committee of Keio University (A2022-103). We used separate cohorts of mice for each experiment to ensure that the experiments were conducted precisely 24 h after LPS administration. In addition, this approach allows us to prevent habituation to the open-field box, which was used for both the open-field test and the dyadic interaction test. To ensure a consistent physiological response to LPS, exclusion criteria were pre-established prior to data analysis. Mice that exhibited a body weight decrease of less than 5% following LPS injection were excluded from the analysis.

Procedure

Experiment 1: immunological investigations by ELISA and flow cytometry

To quantify peripheral cytokines, plasma was collected from 12 adult male C57BL/6 J mice aged 2.3–3.7 (mean ± SD: 3.0 ± 0.24) months and 12 adult female C57BL/6 J mice aged 2.3–3.8 (mean ± SD: 2.8 ± 0.26) months at 1.5 or 3 h after intraperitoneal LPS administration (1.0 mg/kg). Blood was obtained from the orbital vein under brief isoflurane anesthesia into heparinized hematocrit tubes, centrifuged at 4 °C for 10 min, and the supernatants were collected as plasma. Cytokine concentrations were measured using ELISA kits (R&D Systems, Minneapolis, MN, USA). The detection limits of the ELISA assays were 15.625 pg/mL for IL-6, 31.25 pg/mL for TNF-α, and 15.625 pg/mL for IL-1β. The intra-assay CV for the IL-1β ELISA was 1.3%, and the inter-assay CVs for the TNF-α and IL-6 ELISAs were 7.4 and 10.6%, respectively. For immune cell quantification, red blood cells were lysed using a hemolysis buffer (5 mM EDTA, 0.1% KHCO₃, 0.8% NH₄Cl). Leukocytes were stained with the following antibodies: FITC–anti-Ly6C, PE–anti-CD3, BV711–anti-CD11b, BV421–anti-Ly6G, PE/Cy7–anti-CD19, APC–anti-CD4, and PE/Cy5–anti-CD3 (BioLegend, San Diego, CA, USA). Live singlet leukocytes were gated by propidium iodide (PI) exclusion prior to lineage identification. Immune cell numbers were calculated using CountBright™ Absolute Counting Beads (Thermo Fisher Scientific, Waltham, MA, USA) and analyzed on a FACS Aria III (BD Biosciences, San Jose, CA, USA). Sampling time points were selected based on the well-characterized temporal dynamics of LPS-induced cytokine responses. Circulating TNF-α, IL-6, and IL-10 typically peak within 1–6 h following peripheral LPS challenge, with maximal levels around 1 h and a rapid decline thereafter. By 24 h post-injection, these cytokines are generally near or below detectable levels [22]. Accordingly, blood collection at 1.5 and 3 h post-LPS was chosen to capture the acute inflammatory phase during which cytokine concentrations are physiologically meaningful. Because cytokine elevations are not detectable at 24 h, additional sampling at behavioral testing time points was not included.

Experiment 2: dyadic social interaction test

To examine the effects of intraperitoneal LPS injection on dyadic interactions between two familiar or unfamiliar mice, we observed such interactions in an open-field box. In the familiar condition, we used 32 adult male C57BL/6 J mice aged 2.1–4.5 (mean ± SD: 2.8 ± 0.61) months and 32 adult female C57BL/6 J mice aged 2.1–4.7 (mean ± SD: 3.4 ± 0.83) months. All mice were group-housed in pairs in their home cages. In the unfamiliar condition, we used 32 adult male C57BL/6 J mice aged 2.2–4.2 (mean ± SD: 2.9 ± 0.44) months and 32 adult female C57BL/6 J mice aged 2.3–2.7 (mean ± SD: 2.5 ± 0.01) months. These mice were either single-housed or group-housed with 2–4 individuals per cage. The apparatus was a custom-made white vinyl chloride box (50 × 50 × 50 cm). Cameras (Logicool HD Webcam C922n, Logicool Co Ltd., Tokyo, Japan) were positioned 106 cm above the floor of the box. Dyadic interactions were recorded on a Windows PC for 30 min. White noise (75 dB) was presented throughout the experiment to mask external sounds. The experimental environment was maintained at 6.5 lux. Mice were not habituated to the open-field arena prior to the test. 24 h before the test, they received intraperitoneally injection of LPS (1.0 mg/kg). On the test day, body weight was measured, and two mice were inserted at the center of the open-field box.

Experiment 3: sucrose preference test

To assess reward sensitivity under conditions that minimize stress and metabolic confounds, we conducted a sucrose preference test without food or water deprivation. Sixteen adult male C57BL/6 J mice aged 2.1–7.1 (mean ± SD: 3.7 ± 2.1) month and sixteen adult female C57BL/6 J mice aged 2.4–5.8 (mean ± SD: 4.4 ± 1.6) month were used. Mice were habituated to the sucrose preference apparatus for two days prior to LPS injection (1.0 mg/kg). During both the habituation and test phases, standard food pellets were continuously available within the apparatus, allowing ad libitum feeding. During habituation, a single bottle containing 1% sucrose solution was placed in the apparatus, with bottle position counterbalanced across days. This procedure ensured familiarity with the sucrose solution and minimized the possibility that mice would fail to detect the subsequent choice. In the test phases, two bottles were placed in the apparatus, one containing 1% sucrose solution and the other containing water, with positions counterbalanced. Following intraperitoneal injection of LPS (1.0 mg/kg) or saline, mice were placed in the apparatus, and licking responses were recorded continuously for 24 h. Sucrose preference was calculated as the number of licks on the sucrose bottle divided by the total number of licks on both bottles.

Experiment 4: open-field test

To examine the effects of LPS on spontaneous locomotor activity, we conducted an open-field test using 8 adult male C57BL/6 J mice aged 2.2–5.5 (mean: 3.2, SD: 1.5) months and 8 adult female C57BL/6 J mice aged 2.2–4.3 (mean: 2.7, SD: 0.43) months. No habituation was conducted prior to the experiment. The apparatus was a custom-made white vinyl chloride box (50 × 50 × 50 cm). Cameras (Logicool HD Webcam C922n, Logicool Co Ltd., Tokyo, Japan) were positioned 106 cm above the floor of the box [23, 24], and video recordings were acquired on a Windows PC. Animals were placed in the box 24 h after intraperitoneal injection of LPS (1.0 mg/kg) and allowed to explore freely for 30 min. White noise (75 dB) was presented throughout the experiment to mask external sounds. After each trial, the interior of the box was wiped with 70% ethanol.

Drug

Lipopolysaccharide (LPS; CAS L2630, Sigma-Aldrich, Gillingham, UK) was dissolved in saline, divided into small aliquots, and stored at −20 °C. LPS was administered intraperitoneally at a dose of 1.0 mg/kg (0.01 mL/g) in all experiments. For control conditions, saline was administered intraperitoneally in the same manner as in the LPS condition. The LPS dose used in this study (1.0 mg/kg) was selected based on doses frequently used in previous studies (0.5, 0.83, and 1.0 mg/kg), which together accounted for 73.19% of studies in which LPS was administered intraperitoneally to mice [10]. These doses have been reported to increase immobility time in the forced swim test in C57BL/6 mice [10].

Analysis

In all experiments, data analysis was performed using RStudio (version 2022.02.0; RStudio PBC, MA, USA) and GraphPad Prism (version 10.3.1; GraphPad, CA, USA). An open-source visual programming framework, Bonsai, was used to track mouse activity via computer vision analysis [25]. Video recordings were converted to grayscale, smoothed, and monochrome-inverted. Mice were detected by applying a contrast threshold, and locomotor activity in the open-field box was quantified by measuring changes in the animals’ central coordinates. In the dyadic interaction test, the distance between two male mice was calculated [26, 27]. A distance of 0 cm was defined as physical contact between the two mice.

Sample sizes were determined based on previous studies using LPS-induced inflammatory models [22] and behavioral assays [28, 29], as well as our prior experience, which consistently detected robust effects. Mice were assigned to the saline or LPS group in a randomized manner after matching for body weight and age to minimize potential confounding effects. The investigator was not blinded during drug administration. Data analysis was performed using automated analysis, with the investigator blinded to group allocation. Data in all figures are presented as the mean ± the standard error of the mean (SEM).

Results

Experiment 1: effects of LPS injection on immunological reactions and body weight

To investigate the effects of LPS on inflammatory responses, we measured cytokine levels using ELISA and quantified immune cell numbers by flow cytometry at 1.5 and 3 h after LPS injection (Fig. 1).

Fig. 1. Immunological investigations.

A Experimental procedure for ELISA and flow cytometry. Mice received an intraperitoneal injection of LPS (1.0 mg/kg). At 1.5 or 3.0 h post-injection, blood samples were collected from the eyelid. Cytokine levels (IL-1β, IL-6, TNF-α) were quantified using ELISA. Immune cell populations (CD4⁺ T cells, CD8⁺ T cells, B cells, neutrophils, and monocytes) were quantified using flow cytometry. B Gating strategies for peripheral blood immune cells. B cells (CD19⁺), CD4⁺ T cells (CD3⁺CD4⁺CD8⁻), CD8⁺ T cells (CD3⁺CD4⁻CD8⁺), neutrophils (CD11b⁺Ly6C⁺Ly6G⁺), and monocytes (CD11b⁺Ly6C^hiLy6G⁻) were identified by flow cytometry.

For immune cells, LPS injection significantly reduced immune cell counts (Fig. 2A–J). A significant main effect of LPS treatment was detected for CD4⁺ T cells at 1.5 h post-injection (F(1,19) = 137.8, p < 0.0001, two-way ANOVA), CD4⁺ T cells at 3.0 h (F(1,18) = 61.80, two-way ANOVA), CD8⁺ T cells at 1.5 h (F(1,19) = 99.78, two-way ANOVA), CD8⁺ T cells at 3.0 h (F(1,18) = 53.33, two-way ANOVA), neutrophils at 3.0 h (F(1,18) = 6.814, p = 0.0177, two-way ANOVA), and monocytes at 3.0 h (F(1,19) = 19.83, p = 0.0003, two-way ANOVA). For B cells, significant main effects of LPS treatment and sex were observed at 1.5 h (LPS: F(1,19) = 43.05, p < 0.0001; sex: F(1,19) = 11.07, p = 0.0035, two-way ANOVA) and at 3.0 h (LPS: F(1,18) = 28.71, p < 0.0001; sex: F(1,18) = 7.096, p = 0.0158, two-way ANOVA). Significant main effects of LPS treatment, sex, and LPS treatment × sex interactions were found for neutrophils at 1.5 h (LPS: F(1,19) = 55.89, p < 0.0001; sex: F(1,19) = 11.42, p = 0.0031; interaction: F(1,19) = 8.700, p = 0.0082, two-way ANOVA) and monocytes at 1.5 h (LPS: F(1,19) = 29.44, p < 0.0001; sex: F(1,19) = 9.587, p = 0.0059; interaction: F(1,19) = 7.635, p = 0.0124, two-way ANOVA). To explore this interaction, pairwise comparisons were conducted using Fisher’s Least Significant Difference (LSD) test. The result showed a significant difference between male-Saline and male-LPS (p <0.0001, Fisher’s LSD test), female-Saline and female-LPS (p =0.0056, Fisher’s LSD test), and Saline-male and Saline-female (p = 0.0002, Fisher’s LSD test) in neutrophil at 1.5 h, male-Saline and male-LPS (p <0.0001, Fisher’s LSD test), and Saline-male and Saline-female (p = 0.0004, Fisher’s LSD test) in monocytes at 1.5 h.

Fig. 2. Effects of LPS injection on immune responses and body weight.

A–E Effects of intraperitoneal LPS injection on immune cell numbers at 1.5 h post-injection (Male-Saline: N = 6; Male-LPS: N = 6; Female-Saline: N = 6; Female-LPS: N = 5). A CD4⁺ T cells, (B) CD8⁺ T cells, (C) B cells, (D) neutrophils, (E) monocytes. F–J. Effects of LPS injection on immune cell numbers at 3.0 h post-injection (Male-Saline: N = 6; Male-LPS: N = 6; Female-Saline: N = 6; Female-LPS: N = 4). F CD4⁺ T cells, (G) CD8⁺ T cells, (H) B cells, (I) neutrophils, (J) monocytes. K–M. Effects of LPS injection on cytokine levels at 1.5 h post-injection (Male-Saline: N = 6; Male-LPS: N = 6; Female-Saline: N = 6; Female-LPS: N = 5). (K) IL-6, (L) IL-1β, (M) TNF-α. N–P. Effects of LPS injection on cytokine levels at 3.0 h post-injection (Male-Saline: N = 6; Male-LPS: N = 6; Female-Saline: N = 6; Female-LPS: N = 4). N IL-6, (O) IL-1β, (P) TNF-α. Statistical significance is indicated as *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Q Effects of LPS injection on body weight over 7 days (Male-Saline: N = 12; Male-LPS: N = 12; Female-Saline: N = 12; Female-LPS: N = 9). Error bars represent SEM.

LPS injection significantly increased cytokine levels (Fig. 2K–P). Significant main effects of LPS treatment were found for IL-6 at 1.5 h (F(1,19) = 127.1, p < 0.0001, two-way ANOVA), IL-1β at 1.5 h (F(1,19) = 73.13, p < 0.0001, two-way ANOVA), and TNF-α at 3.0 h (F(1,18) = 23.38, p = 0.0001, two-way ANOVA). Significant main effects of LPS treatment, sex, and LPS treatment × sex interactions were also observed for IL-6 at 3.0 h (LPS: F(1,18) = 532.4, p < 0.0001; sex: F(1,18) = 19.97, p = 0.0003; interaction: F(1,18) = 19.97, p = 0.0003, two-way ANOVA), IL-1β at 3.0 h (LPS: F(1,18) = 193.4, p < 0.0001; sex: F(1,18) = 5.004, p = 0.0375; interaction: F(1,18) = 5.004, p = 0.0375, two-way ANOVA), and TNF-α at 1.5 h (LPS: F(1,19) = 78.38, p < 0.0001; sex: F(1,19) = 8.308, p = 0.0095; interaction: F(1,19) = 8.252 p = 0.0082, two-way ANOVA). To explore this interaction, pairwise comparisons were conducted using Fisher’s LSD test. The result showed a significant difference between male-Saline and male-LPS (p <0.0001, Fisher’s LSD test), female-Saline and female-LPS (p <0.0001, Fisher’s LSD test), and LPS-male and LPS-female (p = 0.0001, Fisher’s LSD test) in IL-6 at 3.0 h, male-Saline and male-LPS (p <0.0001, Fisher’s LSD test), female-Saline and female-LPS (p <0.0001, Fisher’s LSD test), and LPS-male and LPS-female (p = 0.0075, Fisher’s LSD test) in IL-1β at 3.0 h, male-Saline and male-LPS (p =0.0004, Fisher’s LSD test), female-Saline and female-LPS (p <0.0001, Fisher’s LSD test), and LPS-male and LPS-female (p = 0.0008, Fisher’s LSD test) in TNF-α at 1.5 h. Body weight was monitored for seven days following LPS injection (Fig. 2Q). Both male and female mice exhibited significant weight loss. A three-way ANOVA (LPS treatment x days from injection x sex) revealed significant main effects of treatment (F(1, 328) = 335.5, p < 0.0001, three-way ANOVA), days post-injection (F(7, 328) = 41.05, p < 0.0001, three-way ANOVA), and sex (F(1, 328) = 25.43, p < 0.0001). A significant interaction was observed between LPS treatment and days post-injection (F(7, 328) = 36.75, p < 0.0001, three-way ANOVA), whereas the other two-way interactions and the three-way interaction were not significant (all p > 0.05, three-way ANOVA). Post hoc analyses with Tukey’s multiple comparison test indicated that LPS-treated female mice had significantly lower body weight than saline-treated females on days 1–3 post-injection (day 1: p < 0.001; day 2: p < 0.001; day 3: p = 0.007; Tukey’s multiple comparison test). In contrast, LPS-treated male mice exhibited significant weight loss relative to saline-treated males on days 1–4 post-injection (days 1–3: p < 0.001; day 4: p = 0.0054; Tukey’s multiple comparison test). Although both sexes showed comparable initial weight loss following LPS injection, female mice appeared to recover slightly earlier than male mice.

Experiment 2. effects of LPS injection on dyadic social interactions

To investigate the effects of intraperitoneal LPS injection on direct social interactions, we examined dyadic interactions in the open-field box (Fig. 3A). In this test, pairs of mice received either LPS or saline, and their inter-individual distance and contact time were quantified (Fig. 3B). LPS administration decreased body weight (Fig. 3C-F) and increased dyadic interactions between male mice (Fig. 3G, I, K, M, O, Q, S, U), but not female mice (Fig. 3H, J, L, N, P, R, T, V). Inter-individual distance was reduced in both familiar and unfamiliar male mice (familiar: t(14) = 3.166, p = 0.0069; unfamiliar: t(14) = 3.204, p = 0.0064, unpaired t-tests), whereas no significant differences were observed in female mice (familiar: t(14) = 0.801, p = 0.437; unfamiliar: t(14) = 1.249, p = 0.232, unpaired t-tests). Dyadic contact time was also increased in both familiar and unfamiliar male mice (familiar: t(14) = 3.059, p = 0.0085; unfamiliar: t(14) = 2.430, p = 0.0291, unpaired t-tests), with no significant change in female mice (familiar: t(14) = 0.794, p = 0.441; unfamiliar: t(14) = 1.789, p = 0.095, unpaired t-tests). Although interactive behaviors such as anogenital sniffing, grooming, and general contact were not classified in detail, no aggressive behavior was observed in any of the dyads.

Fig. 3. Effects of intraperitoneal LPS injection on dyadic interactions in male and female mice.

A Experimental timeline of the dyadic interaction test (30 min), conducted 24 h after saline or LPS injection (1.0 mg/kg). B Schematic illustrating measurement of inter-individual distance. C–F. Body weight changes 24 h after injection in familiar and unfamiliar male (C, E) and female (D, F) dyads. G–J Mean dyadic distance in familiar and unfamiliar male (G, I) and female (H, J) dyads. K–N Time course of dyadic distance during the 30-min test in familiar and unfamiliar male (K, M) and female (L, N) dyads. O–R Total dyadic contact time in familiar and unfamiliar male (O, Q) and female (P, R) dyads. S–V Time course of dyadic contact time in familiar and unfamiliar male (S, U) and female (T, V) dyads. Error bars indicate SEM. Statistical significance is indicated as *p < 0.05, **p < 0.01, ****p < 0.0001; ns not significant. N = 8 per group.

Experiment 3: effects of LPS injection on a sucrose preference test

To investigate the effect of intraperitoneal LPS injection on hedonic behavior, we assessed preference for a 1% sucrose solution over water (Fig. 4A). Sucrose preference rate and total number of licks were quantified over 24 h. The sucrose preference rate was reduced in the LPS condition in both males and females (males: Fig. 4B, t(14) = 3.390, p = 0.0067; females: Fig. 4C, t(14) = 3.169, p = 0.0068, unpaired t-tests). Additionally, LPS decreased the total number of licks (males: Fig. 4F, t(14) = 25.67, p < 0.0001; females: Fig. 4G, t(14) = 9.070, p < 0.0001, unpaired t-tests), and licking behavior was suppressed following LPS injection (Fig. 4D, E).

Fig. 4. Effects of intraperitoneal LPS injection on sucrose preference in male and female mice.

A Schematic diagram of the sucrose preference test procedure. Mice were first habituated to drink 1% sucrose solution from either side of the bottle in the experimental box for two days, with the side of the sucrose solution counterbalanced. Following habituation, mice received an injection of LPS or saline and were placed in the experimental box equipped with a touch sensor for lick detection. B Effects of LPS injection on sucrose preference rate in male mice. C Effects of LPS injection on sucrose preference rate in female mice. D Time course of licking responses over 24 h in male mice after LPS injection. E Time course of licking responses over 24 h in female mice after LPS injection. F Effects of LPS injection on total number of licks in male mice. G Effects of LPS injection on total number of licks in female mice. Error bars represent SEM. Statistical significance is indicated as **p < 0.01, ****p < 0.0001. N = 8 per group.

Experiment 4: effects of LPS injection on locomotor activity

To examine the effect of intraperitoneal LPS injection on whole-body motor function, we assessed spontaneous locomotor activity in the open-field box (Fig. 5). We verified that body weight decreased after LPS injection but not after saline injection in both male and female mice (males: Fig. 5C, t(14) = 13.94, p < 0.0001; females: Fig. 5D, t(14) = 11.94, p < 0.0001, unpaired t-tests). Time-course analysis of spontaneous locomotor activity during the open-field test indicated that LPS affected activity throughout the experiment in both sexes (males: Fig. 5E; females: Fig. 5F). Overall locomotor activity was significantly reduced in the LPS condition compared to saline in both male and female mice (males: Fig. 5G, t(14) = 3.539, p = 0.0033; females: Fig. 5H, t(14) = 5.741, p < 0.0001, unpaired t-tests). We also examined the time spent in the center of the box during the open-field task. LPS injection had no significant effect on center time in either sex (males: Fig. 5I, t(14) = 2.108, p = 0.0536; females: Fig. 5J, t(14) = 1.127, p = 0.279, unpaired t-tests).

Fig. 5. Effects of intraperitoneal LPS injection on the open-field test in male and female mice.

A Experimental timeline of the open-field test, conducted for 30 min, 24 h after LPS injection (1.0 mg/kg). B Definition of the center arena. C Body weight 24 h after saline or LPS injection in male mice. D Body weight 24 h after saline or LPS injection in female mice. E Time course of locomotor activity in male mice following saline or LPS injection. F Time course of locomotor activity in female mice following saline or LPS injection. G Overall locomotor activity in male mice after saline or LPS injection. H Overall locomotor activity in female mice after saline or LPS injection. I Time spent in the center area of the box in male mice. J Time spent in the center area of the box in female mice. K Fecal output after the open-field test in male mice. L Fecal output after the open-field test in female mice. M Urinary output after the open-field test in male mice. N Urinary output after the open-field test in female mice. Error bars represent SEM. Statistical significance is indicated as *p < 0.05, **p < 0.01, ****p < 0.0001; ns not significant. N = 8 per group.

To assess the impact of LPS on autonomic function, we quantified fecal and urinary output after the open-field test. LPS injection significantly reduced fecal output (males: Fig. 5K, t(14) = 3.714, p = 0.0023; females: Fig. 5L, t(14) = 2.580, p = 0.0218, unpaired t-tests), whereas urinary output was unaffected (males: Fig. 5M, t(14) = 0.500, p = 0.625; females: Fig. 5N, t(14) = 1.055, p = 0.309, unpaired t-tests).

Discussions

In this study, we examined the effects of intraperitoneal injection of LPS on behavior in male and female mice. Immunological examinations using ELISA and flow cytometry revealed prominent increases of IL-1β, IL-6, and TNF-α and decreases of T cells, B cells, neutrophils, and monocytes. These results confirm that inflammatory responses were caused by the LPS injection. There was a sex difference in the baseline number of B cells, neutrophils, and monocytes. Through the behavioral assessments, we found that LPS-induced inflammatory responses increased contact time and decreased dyadic distance in familiar and unfamiliar males. Meanwhile, these differences were not found in female dyads. LPS injection caused decreased body weight, locomotor activity, and fecal outputs in an open-field test and decreased not only sucrose preference but also the overall number of licking in a sucrose preference task in both male and female mice. Our results suggest that social behavior is modulated by LPS-induced inflammatory responses in a sex-dependent manner. The increased contact time of dyadic interaction in male dyads could not be explained by weight loss or decreased locomotion, orofacial movements, and reward sensitivities because these results were also observed in LPS-treated females. Our results suggest that social behavior is modulated by LPS-induced inflammatory responses in a sex-dependent manner.

Immunological analyses using ELISA and flow cytometry demonstrated that LPS injection significantly elevated levels of proinflammatory cytokines IL-1β, IL-6, and TNF-α, confirming the induction of a systemic inflammatory response. Additionally, flow cytometry revealed substantial reductions in circulating T cells, B cells, neutrophils, and monocytes following LPS administration. This decrease likely reflects the migration of immune cells from the bloodstream to inflamed regions inside the body and/or inflammation-induced cell death [30, 31]. In addition, we revealed a sex difference in the baseline number of B cells, neutrophils, and monocytes. Due to technical challenges and a lack of translational approaches, investigations into the relationship between peripheral immune cell dynamics and behavior remain limited. Our study thus provides a foundation for further research into the complex interactions between inflammation and behavioral performance.

The present study demonstrates that acute systemic inflammation differentially alters specific components of social interaction in male and female mice. In a dyadic interaction context, male dyads exhibited increased physical contact and reduced interpersonal distance following LPS administration, whereas female dyads showed no comparable change. Although sex differences in post-inflammatory social behavior have been documented previously [28, 29], the current findings refine this literature by showing that alterations in social proximity and contact can emerge independently of canonical sickness behaviors such as reduced locomotion, body weight loss, or diminished reward sensitivity. This dissociation suggests that social interaction constitutes a partially distinct behavioral domain that is not invariably suppressed during inflammatory states.

Early work on inflammation-induced behavioral changes emphasized a generalized suppression of motivated behaviors, largely based on reduced exploratory activity following LPS treatment [32]. Subsequent studies, however, have revealed more nuanced and context-dependent effects of immune activation on social behavior. For example, Yee & Prendergast (2010) reported that LPS decreased certain interactive behaviors such as grooming while paradoxically increasing overall contact time in home-cage settings [29]. Similarly, mild inflammatory challenges enhanced social contact among familiar individuals in rhesus monkeys [33], and LPS increased male preference in female prairie voles [34]. Viewed in this context, the present findings highlight that inflammation can selectively modulate social proximity and contact, rather than uniformly suppressing social engagement, and that these effects are expressed in a sex-dependent manner.

Importantly, the observed increase in contact among male dyads does not necessarily imply enhanced affiliative motivation. Social proximity can reflect multiple underlying processes, including affiliative interaction, stress coping, or physiological regulation. One plausible alternative interpretation is that increased contact in LPS-treated males reflects thermoregulatory huddling rather than altered social motivation per se. Previous studies have shown that males exhibit a stronger pyrogenic response to LPS than females [35], and that huddling behavior itself can be sex dependent [36]. Because core body temperature was not assessed in the present study, the relative contribution of fever-driven thermoregulation to increased proximity cannot be excluded. This possibility underscores the importance of interpreting social distance measures within a broader physiological context.

At the neurobiological level, prior work indicates that proinflammatory cytokines such as IL-1β, IL-6, and TNF-α can influence neurotransmission and plasticity in brain regions implicated in social behavior, including the prefrontal cortex, amygdala, and hippocampus [37–39]. Moreover, microglial activation and cytokine signaling have been linked to sex-dependent regulation of social and affective behaviors [40, 41]. For instance, Decker Ramirez et al. [28] reported reduced social interaction in female dyads following LPS administration, accompanied by elevated IL-6 levels in the ventral striatum [28], while Hodes et al. [42] demonstrated that IL-6 signaling is required for stress-induced social avoidance [42]. While the present study does not directly assess central cytokine activity or neural mechanisms, the behavioral dissociation observed here is consistent with the notion that inflammatory signals may engage sex-dependent regulatory pathways influencing social behavior.

Finally, sex hormones likely modulate immune–behavior interactions and may contribute to the differential sensitivity to inflammation observed between males and females. Estrogen and testosterone are known to influence immune responsiveness and cytokine signaling, and fluctuations in estrogen levels are associated with altered inflammatory states and mood regulation in females [43, 44]. Although hormonal status was not controlled in the present study, these factors represent important variables for future investigations aimed at disentangling hormonal, immune, and behavioral contributions to sex-specific social outcomes during inflammation.

Intraperitoneal injection of LPS significantly reduced locomotor activity in both male and female mice, consistent with previous reports demonstrating hypoactivity following systemic inflammatory challenge [16, 20]. At the same time, other studies have reported minimal or no locomotor suppression at comparable post-injection intervals [15, 45], highlighting the variability of motor outcomes depending on dose, timing, and experimental context. As discussed by Remus and Dantzer (2016), acute sickness behaviors typically peak within hours after LPS administration and begin to resolve by 24 h, whereas depressive-like behaviors may emerge or persist beyond this period [9]. Nevertheless, the temporal overlap between residual sickness and depression-like states complicates behavioral interpretation in acute inflammatory models. In this context, the present study does not seek to establish the novelty of LPS-induced depressive-like or immunological phenotypes per se, as these effects are well documented [46, 47]. Rather, our behavioral and immunological assessments serve to characterize the physiological and motivational state of the animals at the time of social interaction testing, thereby providing an essential interpretive framework for understanding inflammation-induced alterations in social behavior. This distinction is particularly important given the “Janus-faced” nature of inflammation, which can simultaneously promote adaptive sickness responses and contribute to maladaptive affective states [47]. Consistent with prior studies, we observed a significant reduction in sucrose preference following LPS administration [19, 20]. However, the pronounced decrease in total licking behavior suggests that reduced sucrose preference may partly reflect hypoactivity or diminished behavioral engagement rather than a selective deficit in reward valuation. This observation underscores a key limitation of sucrose-based assays in acute inflammatory settings, where motor and motivational impairments can confound the interpretation of anhedonia. Importantly, acknowledging these constraints allowed us to avoid overinterpreting depressive-like phenotypes and instead focus on social and communicative behaviors that may be differentially regulated under inflammatory conditions.

Intraperitoneal injection of LPS induced a significant decrease in body weight in both male and female mice, with gradual recovery observed over the course of a week. Notably, female mice exhibited less pronounced weight loss and a faster rate of recovery compared to males. This weight loss is partially attributable to reduced food and water intake, as evidenced by our sucrose preference test results, which showed marked decreases in both consumption and licking behavior following LPS administration. However, Morita-Takemura et al. [48] reported that while food intake returned to baseline within 24 h post-LPS injection, body weight recovery was delayed, occurring approximately five days later [48]. This discrepancy suggests that weight loss following LPS is not solely driven by decreased consumption but also involves inflammation-related metabolic changes. Thus, systemic inflammatory responses likely contribute directly to the observed alterations in body weight beyond the effects of reduced nutrient intake.

Despite the comprehensive assessment of LPS-induced peripheral immune activation and behavioral alterations, several limitations should be acknowledged. First, female mice were not stratified according to estrous cycle stage, which may have introduced variability in immune and behavioral responses due to hormonal fluctuations. Although female behavioral responses to LPS appeared relatively stable in the present study, immune–behavioral interactions are known to vary across the estrous cycle, limiting the interpretation of sex differences. Second, the mice included in this study spanned a relatively broad age range (2–6 months). Given that aging influences immune responsiveness, cytokine production, and microglial phenotype, future studies employing more narrowly age-matched cohorts will be important to reduce potential maturational confounds. Third, while peripheral cytokine levels and immune cell profiles were systematically quantified, the present study did not directly assess neuroimmune activity within the brain, such as region-specific cytokine expression, microglial activation, or inflammatory signaling in discrete neural circuits. Accordingly, we do not claim direct evidence for central neuroimmune mechanisms underlying the observed behavioral changes. Rather, our findings establish a behavioral dissociation under equivalent peripheral inflammatory conditions, identifying a sex-specific alteration in social interaction that cannot be readily explained by generalized sickness behavior alone. Importantly, this dissociation provides a critical empirical foundation for future studies aimed at elucidating how peripheral immune activation is translated into circuit-specific neural changes. Targeted analyses of cytokine signaling, glial activation, and neural activity within social behavior–relevant regions such as the medial preoptic area, amygdala, and prefrontal cortex will be necessary to determine whether and how systemic inflammation differentially engages neuroimmune pathways in males and females. Fourth, some behavioral measures, particularly sucrose preference, may have been influenced by LPS-induced hypoactivity or motor impairments, complicating the distinction between motivational deficits and generalized sickness behavior. Although steps were taken to minimize metabolic and stress-related confounds, this limitation is inherent to acute inflammation-based behavioral paradigms. Finally, immune and behavioral assessments were restricted to acute post-injection time points, precluding conclusions regarding subacute or chronic neuroimmune adaptations. Longitudinal studies will be required to determine whether the observed sex-dependent behavioral alterations persist or undergo further modulation over time.

Future studies should aim to systematically address these limitations and extend the present findings. In particular, incorporating precise control of the estrous cycle in female subjects will be important for disentangling the contributions of ovarian hormones to immune–behavior interactions under inflammatory conditions. Such approaches may clarify whether the relative stability of female social behavior observed here reflects true sex-specific resilience or stage-dependent hormonal modulation. To elucidate the neural substrates through which peripheral inflammation influences social behavior, brain region–specific neuroimmune analyses will be essential. Targeted assessments of cytokine signaling, glial activation, and inflammatory transcriptional responses within regions implicated in social motivation and affective processing—such as the medial preoptic area, amygdala, and prefrontal cortex—may help bridge the gap between systemic immune activation and behavioral outcomes, particularly with respect to sex-dependent effects. Moreover, the use of behavioral paradigms that better dissociate sickness-related hypoactivity from motivational or affective deficits will strengthen interpretability. Combining standard assays with ethologically relevant or effort-independent measures may allow more precise attribution of observed behavioral changes to altered social motivation rather than generalized malaise. Longitudinal experimental designs tracking both immune parameters and behavioral phenotypes across acute, subacute, and later post-inflammatory phases will further clarify how transient systemic inflammation may give rise to persistent or evolving behavioral alterations. Such approaches will be critical for distinguishing reversible sickness responses from longer-lasting neurobehavioral adaptations. Finally, investigating the role of sex hormones and their receptors in shaping immune signaling and social behavior may yield mechanistic insights into the biological basis of sex differences observed in inflammation-associated behavioral disturbances. These efforts may ultimately contribute to a better understanding of sex-biased vulnerability and resilience in inflammation-related neuropsychiatric disorders.

In summary, this study demonstrates that acute systemic inflammation induced by LPS differentially modulates social behavior in male and female mice, with males exhibiting decreased social distance and females showing resilience to such effects. These sex-specific behavioral alterations occur alongside robust immune activation characterized by elevated proinflammatory cytokines and dynamic shifts in peripheral immune cell populations. The findings highlight the complex interplay between immune signaling, sex hormones, and neural circuits governing social interaction. Our results emphasize the necessity of considering sex as a biological variable in neuroimmune research and lay the groundwork for future investigations into the mechanisms by which inflammation influences behavior and contributes to sex differences in neuropsychiatric vulnerability.

Author contributions

Mizuki Yamamoto: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Project administration; Software; Validation; Visualization; Roles/Writing - original draft; and Writing - review & editing. Kazuko Hayashi: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Project administration; Supervision; Validation; and Writing - review & editing. Masashi Kanayama: Data curation; Formal analysis; Investigation; Methodology; Resources; Software; Supervision; Validation; Visualization; Roles/Writing - original draft; and Writing - review & editing. Hiroto Inoue: Data curation; Formal analysis; Software; Validation; Visualization; and Writing - review & editing. Shuntaro Matsushima: Investigation; Methodology; Validation; and Writing - review & editing. Koji Toda: Conceptualization; Data curation; Formal analysis; Funding acquisition; Investigation; Methodology; Project administration; Resources; Software; Supervision; Validation; Visualization; Roles/Writing - original draft; and Writing - review & editing.

Funding

This research was supported by JSPS KAKENHI 22H01105 (KT), 23H02787 (KT), 23K27478 (KT), 23K22376 (KT), 24H00729 (KT), 24K06626 (KH), and 25KJ0306 (KH), Keio Academic Development Fund (KT), Keio Gijuku Fukuzawa Memorial Fund (KT), Smoking Research Foundation (KT), and HOKUTO Foundation for the Promotion of Biological Science (KT).

Data availability

The data supporting the findings of this study are available from the corresponding author upon reasonable request.

Code availability

The original codes written for the analyses are available from the corresponding author upon reasonable request.

Competing interests

The authors declare no competing interests.