Acute trazodone treatment ameliorates behavioral abnormalities and improves subsequent short‐term memory in a lipopolysaccharide‐induced mouse model of acute brain dysfunction

Abstract

Aim

Systemic inflammation‐associated acute brain dysfunction, typified by delirium, is linked to later cognitive impairment. Although trazodone is used clinically for delirium‐related symptoms, its effects on subsequent cognition remain unclear. We examined whether acute‐phase trazodone treatment influences behavioral abnormalities, later cognitive outcomes, and inflammatory/glial‐related changes in an lipopolysaccharide (LPS)‐induced mouse model.

Methods

Adult male C57BL/6J mice received saline or LPS (3.0 mg/kg, intraperitoneally). Control mice received saline, whereas LPS‐treated mice subsequently received saline or trazodone (10 mg/kg, subcutaneously). In addition to a group receiving trazodone during the acute phase (Days 0–2), a group receiving trazodone for 7 days was also included. Behavioral assessments were performed on Day 1 (acute phase), Day 3 (recovery phase), and Day 7 (postrecovery phase) using nesting, open‐field, Y‐maze, and novel object recognition tests. Plasma cytokines and inflammatory and glial markers in the cerebral cortex and hippocampus were also assessed.

Results

During the acute phase, trazodone‐treated LPS mice showed reduced locomotor activity, whereas stimulus‐dependent exploratory behavior and nesting were relatively preserved. During the recovery phase, novel object recognition performance was better in trazodone‐treated than in saline‐treated LPS mice. Elevations of cortical Iba1 and IL‐6 and hippocampal GFAP during the acute phase were evident in saline‐treated LPS mice but not clearly in trazodone‐treated LPS mice. Direct comparison between the two trazodone regimens showed no clear additional benefit of continued treatment.

Conclusion

In this model, acute‐phase trazodone treatment was associated with milder acute behavioral abnormalities and better subsequent short‐term‐memory performance, whereas the benefit of continued treatment remained uncertain.

INTRODUCTION

Acute brain dysfunction is common in patients admitted to the intensive care unit (ICU) and those with severe physical illnesses. ¹ , ² Clinically, it encompasses a broad spectrum of manifestations, ranging from insomnia, restlessness, and agitation to somnolence, impaired consciousness, and coma. ¹ Delirium, a representative phenotype of this condition, is highly prevalent in patients in the ICU, with a reported prevalence of up to 80%. ³ Acute brain dysfunction is associated with prolonged hospitalization, increased mortality, and poor cognitive outcomes after discharge. ¹ , ² Cognitive impairment is particularly common soon after ICU discharge, reportedly affecting up to 100% of survivors in some series, and may persist in approximately 30% for up to 1 year or longer, substantially affecting activities of daily living and social reintegration. ⁴ , ⁵ Thus, interventions for acute brain dysfunction are important for short‐term symptom relief and long‐term neurological outcomes. ¹ , ²

Trazodone is used in consultation‐liaison psychiatry for acute psychiatric symptoms accompanied by delirium, insomnia, and agitation. ⁶ , ⁷ , ⁸ It is a serotonin antagonist and reuptake inhibitor, and at low doses, it exerts sedative‐hypnotic effects via the antagonism of 5‐HT2A, H1, and α1 receptors. ⁹ , ¹⁰ Unlike agents with strong anticholinergic properties, trazodone has minimal anticholinergic activity, ¹¹ which may make it less likely to exacerbate delirium, thus contributing to its clinical utility in this setting. Trazodone has been associated with lower mortality and readmission rates than antipsychotics, increasing interest in its safety profile. ¹² However, its efficacy and effects on post‐recovery cognitive outcomes remain unclear.

Recent studies suggest that neuroinflammation secondary to systemic inflammation is key in the pathophysiology of acute brain dysfunction, including delirium. ² , ¹³ Glial cells, particularly microglia and astrocytes, regulate inflammatory cytokine production and neural circuit function and may contribute to acute symptoms and subsequent cognitive impairment associated with acute brain dysfunction. ¹³ , ¹⁴ , ¹⁵ , ¹⁶ However, in humans, the temporal dynamics of these changes are difficult to assess directly, and multiple treatments are concomitantly administered, making it difficult to determine whether trazodone itself modifies neuroinflammation, glial responses, and cognitive outcomes.

Therefore, using a lipopolysaccharide (LPS)‐induced mouse model of acute brain dysfunction caused by systemic inflammation, we aimed to examine the effects of trazodone on acute behavioral abnormalities, cognitive impairment, and neuroinflammatory and glial responses.

METHODS

Animals and housing

Adult male C57BL/6J mice (11 weeks old; The Jackson Laboratory Japan, Inc.; n = 80) were used in this study. Only adult male mice were used to minimize variability associated with the estrous cycle and to enable standardized evaluation of behavioral outcomes and drug effects. Mice were housed four per cage under controlled conditions (23 ± 1°C; 12‐h light/dark cycle, lights on 08:00–20:00) with free access to food and water. The mice were acclimatized for 2 weeks before the experiments. All animal experiments were conducted in accordance with the animal ethics regulations of Fukuoka University. The study protocol was approved by the Animal Experiment Ethics Review Committee of Fukuoka University (approval no. 2313092). The study is reported in accordance with the ARRIVE 2.0 guidelines. ¹⁷

LPS‐induced acute brain dysfunction model

LPS‐EB ultrapure (InvivoGen, tlrl‐3pelps) was administered intraperitoneally at 3.0 mg/kg to induce acute brain dysfunction associated with systemic inflammation. ¹⁸ , ¹⁹ Based on previous reports on acute symptoms and subsequent neurobehavioral changes after LPS administration, Days 0–2 were defined as the acute phase, Day 3 onward as the recovery phase, and Day 7 as the postrecovery phase. ¹⁸ , ¹⁹

Trazodone administration

Trazodone hydrochloride (>98% purity; FUJIFILM Wako Pure Chemical Corporation) was dissolved in saline and subcutaneously administered at 10 mg/kg. This dose is within the range used in previous rodent studies and corresponds approximately to a human dose of 50 mg/day based on body‐surface‐area conversion. ²⁰ , ²¹

Experimental design and groups

A schematic overview of the experimental design is shown in Figure 1. The main groups were the Control group, the LPS + saline group (LPS + Sal), and the LPS + trazodone group (LPS + TZD). The primary condition of interest was trazodone administration during the acute phase. To explore the effect of continuing trazodone beyond this phase, an additional group received trazodone for 7 days (Days 0–6; LPS + TZD‐7d group). Mice were assigned to the experimental groups according to a predetermined allocation plan established before mice delivery; allocation was made without knowledge of the mice's individual characteristics, such as body weight or physical condition. No formal randomization procedure was adopted. Treatment administration, behavioral testing, and sample collection were performed with knowledge of group allocation. In contrast, outcome evaluation and quantification of behavioral and molecular measures were performed blinded to group allocation, as described for each assessment below. This study used a cross‐sectional design with independent cohorts at each time point, and a longitudinal evaluation of the same animals was not performed. All injections were administered between 06:00 and 08:00. On Day 0, mice received a single intraperitoneal injection of saline or LPS (3.0 mg/kg), followed immediately by a subcutaneous injection according to group assignment. Mice in the Control group received volume‐matched saline subcutaneously on Days 0–6. Mice in the LPS + Sal group received volume‐matched saline subcutaneously on Days 0–6. Mice in the LPS + TZD group received trazodone subcutaneously on Days 0–2 and volume‐matched saline subcutaneously on Days 3–6. Mice in the LPS + TZD‐7d group received trazodone subcutaneously on Days 0–6. Body weight was measured daily before each morning injection throughout the experimental period (Supporting Information S1: Table S1).

Figure 1.

Experimental design. On Day 0, mice received a single intraperitoneal injection of saline or LPS (3.0 mg/kg), followed immediately by a subcutaneous injection according to group assignment. The Control and LPS + Sal groups received volume‐matched saline on Days 0–6. The LPS + TZD group received trazodone on Days 0–2 and saline on Days 3–6, whereas the LPS + TZD‐7d group received trazodone on Days 0–6. Behavioral assessments and tissue collection were performed on Days 1, 3, and 7 using independent cohorts. Day 1 and Day 3 included the Control, LPS + Sal, and LPS + TZD groups, whereas Day 7 additionally included the LPS + TZD‐7d group. Each group comprised eight mice (n = 8). LPS, lipopolysaccharide; Sal, saline; TZD, trazodone.

Behavioral assessments

Behavioral assessments were performed on Days 1, 3, and 7 using independent cohorts for each assessment day. For nesting assessment, mice were transferred from group‐housed cages (four mice per cage) to individual cages immediately after the final subcutaneous injection and allowed to habituate until behavioral testing. Behavioral testing was conducted during the dark phase, when mice are normally active, and sufficiently after the morning subcutaneous injection (06:00–08:00) to minimize the potential influence of acute sedation. At the beginning of the dark phase (20:00), behavioral testing was initiated with nesting, because nesting imposed minimal task demand. The same animals then underwent the subsequent behavioral tests according to a standardized schedule. Following novel object recognition (NOR) training, the open‐field (OF) and Y‐maze (YM) tests were conducted sequentially during the 1‐h retention interval before NOR testing to maintain a fixed testing schedule and consistent timing across animals. NOR testing was performed 1 h after training. Thus, the behavioral assessments were conducted in the following order: nesting, NOR training, OF, YM, and NOR testing.

Nesting

Nesting has been evaluated as an innate, highly motivated activity of daily living and as a behavioral index reflecting executive function. ²² , ²³ , ²⁴ One Nestlet (Anamec Co., Ltd.) was placed in each cage at the beginning of the dark phase, and nest construction was scored on a 5‐point scale 3 h later (Supporting Information S1: Table S2), as previously described ²⁵ Because group differences in nest‐building behavior may be more apparent during the early phase of nest construction and diminish over time, the evaluation time point was set to 3 h. ²⁶ Two photographs of each nest, including top and lateral views, were obtained at the time of assessment, and scoring was performed by an investigator blinded to the group assignment.

OF test

The mice were placed in a gray acrylic box (30 × 30 × 30 cm) and allowed to explore freely for 10 min. The behaviors were video‐recorded and quantified using SMART 3.0 behavioral analysis software (Panlab), with group allocation concealed during analysis. As the OF test involves no specific stimulus or task demand, the total distance traveled was used as an index of basal spontaneous locomotor activity. ¹⁸

YM test

A gray acrylic YM (arm length: 20 cm, width: 5 cm, height: 15 cm) was used. The mice were placed at the end of a fixed arm and allowed to freely explore it for 5 min. The behaviors were video‐recorded and quantified using SMART 3.0 behavioral analysis software (Panlab), with group allocation concealed during analysis. Arm entry was defined as the entry of the mid‐trunk into the arm. Spontaneous alternation was defined as three consecutive entries into three different arms, and the spontaneous alternation rate (number of alternations/[total arm entries − 2] × 100) was calculated as an index of spatial working memory. ²⁷ , ²⁸ Total arm entries were recorded as an index of stimulus‐dependent exploratory behavior under task conditions that involved external stimulation.

NOR test

The mice were tested in a gray acrylic box (30 cm × 30 cm × 30 cm). During training, two identical objects were presented, and the mice were allowed to explore them for 10 min. In the test session, one object was replaced with a novel object 1 h later, and the mice were again allowed to explore it for 10 min. Their behaviors were recorded on video. Object exploration was defined as directing the nose toward, touching, or sniffing the object. Climbing onto an object was not considered. The discrimination index ([number of novel‐object explorations − number of familiar‐object explorations]/[number of novel‐object explorations + number of familiar‐object explorations]) was calculated as an index of short‐term memory. ²⁷ , ²⁹ The total exploration count during the test session was recorded as an index of stimulus‐dependent exploratory behavior under task conditions involving external stimuli. Scoring was performed from video recordings by an investigator blinded to the group assignment. The NOR discrimination index is calculated only for mice that explore at least one object during the test session; mice with no object exploration are excluded from this analysis because the index is undefined for them.

Tissue collection

After completion of behavioral testing, tissue collection was performed between 06:00 and 08:00 on the following morning. The mice were deeply anesthetized with isoflurane (FUJIFILM Wako Pure Chemical Corporation), the chest was opened, and terminal blood collection was performed during euthanasia by exsanguination via incision of the right atrial appendage. This was followed by transcardial perfusion fixation with 4% paraformaldehyde (PFA; FUJIFILM Wako Pure Chemical Corporation), after which the brains were removed. The brains were post‐fixed in 4% PFA for 24 h, immersed in 15% sucrose for 24 h, and transferred stepwise to 30% sucrose.

Enzyme‐linked immunosorbent assay (ELISA)

Blood samples were centrifuged (2500g, 20 min, 18°C) to separate the plasma and stored at −80°C until analysis. Plasma interleukin‐6 (IL‐6) and tumor necrosis factor‐alpha (TNF‐α) concentrations were measured using commercial ELISA kits according to the manufacturers' instructions (Mouse IL‐6 ELISA Kit, KE10007; Mouse TNF‐α ELISA Kit, KE10002; Proteintech Group, Inc.). Plasma samples were assayed without dilution. The reported assay sensitivity/range was 3.8 pg/mL and 15.6–1000 pg/mL for IL‐6, and 1.0 pg/mL and 7.8–500 pg/mL for TNF‐α. No values were below the lower limit of quantification. ELISA measurements were evaluated by an investigator blinded to group allocation.

Western blotting

The brain was separated into the cerebral cortex and hippocampus, and proteins were extracted from the fixed tissues using a previously described method. ³⁰ The fixed tissue was homogenized in lysis buffer containing 500 mM Tris‐HCl and 2% sodium dodecyl sulfate (3 µL buffer per 1 mg tissue), heated at 90°C for 120 min, and centrifuged at 10,000g for 15 min at 4°C; the supernatant was used for western blotting. Protein concentrations were determined using a Qubit 3.0 fluorometer (Invitrogen, Q33216) with the Qubit Protein Assay Kit (Invitrogen, Q33211), and 15 µg of protein per lane was subjected to SDS‐PAGE on NuPAGE 4%–12% Bis‐Tris Mini Gels (15‐well; Thermo Fisher Scientific, NP0323BOX) using NuPAGE MOPS SDS Running Buffer (Thermo Fisher Scientific, NP0001). Precision Plus Protein Dual Color Standards (Bio‐Rad, 161‐0374) were used as molecular weight markers. Proteins were transferred onto 0.2‐µm polyvinylidene difluoride membranes (Trans‐Blot Turbo Transfer Pack; Bio‐Rad, 1704156) using the Trans‐Blot Turbo system (Bio‐Rad). Membranes were blocked with 5% skim milk in phosphate‐buffered saline containing 0.1% Tween 20 (PBST) at 4°C overnight and incubated overnight at 4°C with primary antibodies diluted in Can Get Signal Solution 1 (TOYOBO, NKB‐101). In the cerebral cortex and hippocampus, the microglial marker ionized calcium‐binding adaptor molecule 1 (Iba1), the astrocyte marker glial fibrillary acidic protein (GFAP), the neuronal marker neuronal nuclei (NeuN), and the inflammatory cytokines IL‐6 and TNF‐α were quantified using primary antibodies against Iba1 (Proteintech, 81728‐1‐RR; 1:2000), GFAP (Proteintech, 16825‐1‐AP; 1:5000), NeuN (Abcam, ab128886; 1:1000), IL‐6 (Proteintech, 29444‐1‐AP; 1:1000), and TNF‐α (Proteintech, 26162‐1‐AP; 1:1000), with β‐actin (Proteintech, 20536‐1‐AP; 1:2000) used as a loading control. After washing with Tris‐buffered saline containing 0.1% Tween 20 (TBST; Bio‐Rad, 161‐0781), membranes were incubated with HRP‐conjugated anti‐rabbit IgG (Cell Signaling Technology, 7074; 1:2000) diluted in Can Get Signal Solution 2 (TOYOBO, NKB‐101) for 1 h at room temperature. Immunoreactive bands were visualized using Clarity Western ECL Substrate (Bio‐Rad, 170‐5061), and images were acquired with a MultiImager II (Biotools). When membranes were re‐probed, antibodies were stripped with a stripping solution (FUJIFILM Wako, 193‐16375) at room temperature, and complete removal of the previous signal was confirmed before subsequent antibody incubation.

Band intensities were quantified by densitometry using ImageJ (National Institutes of Health). A common reference sample—a protein extract prepared from an unfixed whole mouse brain (laboratory stock)—was loaded in a designated lane of every gel, and signal intensities were first normalized to the corresponding band of this reference sample within the same membrane to correct for inter‐membrane variation. Band intensities were then normalized to β‐actin to correct for well‐to‐well loading variation. Protein expression levels are expressed relative to the mean of the Control group. The procedures used to obtain the representative band images and to normalize signal intensities across membranes through the common reference lane are illustrated in Supporting Information S1: Figure 1 using representative Day‐1 samples. Western blot measurements and densitometric analyses were evaluated by an investigator blinded to group allocation.

Statistical analysis

This study was designed to assess multiple predefined behavioral, cytokine, and western blot outcomes rather than a single primary endpoint. Predefined behavioral outcomes included cognitive‐related measures of short‐term memory, spatial working memory, and nesting behavior, as well as complementary activity‐ and exploration‐related measures, namely OF total distance, NOR total exploration count, and YM total arm entries. No formal sample‐size calculation based on a single primary endpoint was performed. Sample sizes were determined with reference to prior studies using similar LPS‐based behavioral and molecular assessments ¹⁸ , ¹⁹ and practical feasibility.

Because independent cohorts were used at each assessment day and the primary comparisons of interest were day‐specific, the main analyses were performed separately for each assessment day. Three groups were compared on Days 1 and 3, and four groups were compared on Day 7. Because several outcomes were ordinal (e.g., nesting scores) or count‐based and the data did not satisfy the assumptions of parametric analysis of variance, nonparametric methods were used. Behavioral, ELISA, and western blotting data were analyzed using the Kruskal–Wallis test followed by the Steel–Dwass multiple‐comparison test. Effect sizes were calculated as Hedges' g with small‐sample correction (J), and 95% confidence intervals were derived from the standard error of g. For contrasts in which the standard deviation was zero (complete separation), Hedges' g was not estimable and only p‐values are reported.

In addition, as a supplementary analysis, an aligned rank transform (ART)‐based two‐way analysis of variance was performed for the across‐day three‐group comparison shown in Supporting Information S1: Table S3. Because LPS‐induced behavioral and molecular changes may vary over time, a factorial analysis of the effects of assessment day and treatment condition was considered appropriate, as in previous studies. ³¹ In this analysis, the three groups were defined according to the predefined treatment conditions at each assessment day: Control, LPS + Sal, and LPS + TZD‐7d. Because the LPS + TZD group (trazodone on Days 0–2 followed by saline) is identical to the LPS + TZD‐7d group through Day 3 and becomes a distinct treatment group only on Day 7, the across‐day factorial analysis was performed using the three groups that could be defined consistently across all assessment days. ART, which enables factorial analysis of nonparametric data, was used to evaluate the effects of assessment day, treatment condition, and their interaction, and was implemented using ARTool version 2.2.2. ³² , ³³ Statistical significance was set at p < 0.05. Analyses were performed using JMP 12.2.0 (SAS Institute Inc.) and ARTool version 2.2.2.

RESULTS

Acute‐phase locomotor activity and novel‐object exploration differed among treatment groups, whereas few group differences were observed after recovery

During the acute phase (Day 1), total distance in the OF test was lower in both the LPS + Sal and LPS + TZD groups than in the Control group (p = 0.0027 and p = 0.0027, respectively; Hedges' g = 4.63, 95% CI [2.75, 6.51] and g = 8.16, 95% CI [5.17, 11.16]), and was also lower in the LPS + TZD group than in the LPS + Sal group (p = 0.0027; g = 3.76, 95% CI [2.13, 5.40]; Figure 2a). In the NOR test, total exploration counts were lower in the LPS + Sal group than in the Control group (p = 0.0035; g = 3.01, 95% CI [1.58, 4.44]) and the LPS + TZD group (p = 0.0085; g = 1.57, 95% CI [0.45, 2.69]), whereas no significant difference was detected between the Control and LPS + TZD groups (Figure 2d). In the YM test, total arm entries were lower in the LPS + Sal group than in the Control group (p = 0.0027; g = 4.47, 95% CI [2.64, 6.30]) and the LPS + TZD group (p = 0.0064; g = 2.32, 95% CI [1.05, 3.59]), and were also lower in the LPS + TZD group than in the Control group (p = 0.0297; g = 1.37, 95% CI [0.28, 2.46]; Figure 2g).

Figure 2.

Activity‐ and exploration‐related measures at Days 1, 3, and 7. (a–c) Total distance traveled in the open‐field (OF) test at Days 1, 3, and 7, respectively. (d–f) Total exploration counts in the novel object recognition (NOR) test at Days 1, 3, and 7, respectively. (g–i) Total arm entries in the Y‐maze (YM) test at Days 1, 3, and 7, respectively. Mice were assigned to the Control, LPS + Sal, and LPS + TZD groups for the Day 1 and Day 3 analyses, and to the Control, LPS + Sal, LPS + TZD, and LPS + TZD‐7d groups for the Day 7 analysis. Data are shown as box‐and‐whisker plots with individual data points (n = 8 per group). Statistical analyses were performed separately at each time point using the Kruskal–Wallis test followed by the Steel–Dwass post hoc test. *p < 0.05, **p < 0.01, ***p < 0.001. LPS, lipopolysaccharide; Sal, saline; TZD, trazodone.

During the recovery phase (Day 3), no significant difference was detected in OF total distance. In the NOR test, total exploration counts were lower in the LPS + TZD group than in the Control group (p = 0.0088; g = 2.17, 95% CI [0.94, 3.41]; Figure 2e). In the YM test, total arm entries were lower in the LPS + TZD group than in the Control group (p = 0.0102; g = 2.31, 95% CI [1.05, 3.58]) and the LPS + Sal group (p = 0.0089; g = 2.45, 95% CI [1.15, 3.75]; Figure 2h).

During the postrecovery phase (Day 7), no significant group differences were detected in OF total distance, NOR total exploration counts, or YM total arm entries (Figure 2c,f,l).

Group differences in cognitive‐related behavioral measures were observed from the acute phase for nesting behavior and from the recovery phase for NOR performance.

For nesting behavior, the LPS + Sal group showed lower scores than the Control group (p = 0.0004) and the LPS + TZD group (p = 0.0107; g = 1.83, 95% CI [0.66, 2.99]) on Day 1 (Figure 3a). On Day 3, nesting scores were lower in the LPS + Sal group than in the Control group (p = 0.0008; Figure 3b). On Day 7, no significant group differences were detected in nesting scores (Figure 3c).

Figure 3.

Cognitive‐related measures at Days 1, 3, and 7. (a–c) Nesting score in the nesting test at Days 1, 3, and 7, respectively. (d–f) Discrimination index in the novel object recognition (NOR) test at Days 1, 3, and 7, respectively. (g–i) Spontaneous alternation in the Y‐maze (YM) test at Days 1, 3, and 7, respectively. Mice were assigned to the Control, LPS + Sal, and LPS + TZD groups at Days 1 and 3, and to the Control, LPS + Sal, LPS + TZD, and LPS + TZD‐7d groups at Day 7. Data are shown as box‐and‐whisker plots with individual data points. Statistical analyses were performed separately at each time point using the Kruskal–Wallis test followed by the Steel–Dwass post hoc test. *p < 0.05, **p < 0.01, ***p < 0.001. For most groups, n = 8; however, for the NOR discrimination index in the LPS + Sal group on Day 1, n = 4 because four of eight mice showed no object exploration and therefore no discrimination index could be calculated. LPS, lipopolysaccharide; Sal, saline; TZD, trazodone.

For the NOR test, four of eight mice in the LPS + Sal group showed no object exploration during the Day‐1 test session; therefore, these data were treated as missing, and the discrimination index was calculated from the remaining four mice (Figure 3d). On Day 3, the discrimination index was lower in the LPS + Sal group than in the Control group (p = 0.0078; g = 1.76, 95% CI [0.60, 2.91]) and the LPS + TZD group (p = 0.0125; g = 2.00, 95% CI [0.80, 3.19]; Figure 3e). On Day 7, the discrimination index was lower in the LPS + Sal group than in the Control group (p = 0.0195; g = 2.09, 95% CI [0.87, 3.31]) and the LPS + TZD‐7d group (p = 0.0362; g = 1.73, 95% CI [0.58, 2.88]; Figure 3f).

No significant group differences were detected in spontaneous alternation in the YM at any time point (Figure 3g–i).

Plasma IL‐6 showed time‐dependent group differences, whereas plasma TNF‐α did not

Plasma IL‐6 showed group differences that varied across time points. On Day 1, IL‐6 concentrations were lower in the LPS + TZD group than in the LPS + Sal group (p = 0.0027; g = 1.86, 95% CI [0.69, 3.03]; Figure 4a). On Day 3, IL‐6 concentrations were lower in the LPS + Sal group than in the Control group (p = 0.0362; g = 1.29, 95% CI [0.21, 2.37]; Figure 4b). On Day 7, IL‐6 concentrations were lower in both the LPS + TZD and LPS + TZD‐7d groups than in the Control group (p = 0.0052 and p = 0.0074, respectively; g = 1.32, 95% CI [0.24, 2.40] and g = 1.23, 95% CI [0.16, 2.30]). IL‐6 concentrations were also lower in the LPS + TZD and LPS + TZD‐7d groups than in the LPS + Sal group (p = 0.0074 and p = 0.0105, respectively; g = 1.24, 95% CI [0.17, 2.31] and g = 1.18, 95% CI [0.12, 2.24]; Figure 4c).

Figure 4.

Plasma interleukin‐6 (IL‐6) concentrations at Days 1, 3, and 7. (a–c) Plasma IL‐6 concentrations at Days 1, 3, and 7, respectively. Plasma IL‐6 concentrations were measured by enzyme‐linked immunosorbent assay at Days 1, 3, and 7. Data are shown as box‐and‐whisker plots with individual data points (n = 8 per group). Statistical analyses were performed separately at each time point using the Kruskal–Wallis test followed by the Steel–Dwass post hoc test. *p < 0.05, **p < 0.01. LPS, lipopolysaccharide; Sal, saline; TZD, trazodone.

No significant group differences were detected in plasma TNF‐α on Days 1, 3, or 7 (Supporting Information S1: Figure 2A–C).

Brain glial and inflammatory marker differences varied by region and time point

During the acute phase (Day 1), cortical Iba1 expression was higher in the LPS + Sal group than in the Control group (p = 0.0027; g = 2.60, 95% CI [1.27, 3.93]; Figure 5a). Cortical IL‐6 expression and hippocampal GFAP expression were also higher in the LPS + Sal group than in the Control group (p = 0.0149 and p = 0.0272, respectively; g = 1.65, 95% CI [0.51, 2.78] and g = 1.20, 95% CI [0.13, 2.26]; Figures 5d and 6g). The corresponding measures in the LPS + TZD group did not differ significantly from those in the Control group (Figures 5A,D, and 6g). Representative cropped bands and corresponding source membranes for Figures 5A,D, and 6g are shown in Supporting Information S1: Figure 3.

Figure 5.

Protein expression levels in the cerebral cortex at Days 1, 3, and 7. (a–c) Relative protein expression levels of ionized calcium‐binding adaptor molecule 1 (Iba1) in the cerebral cortex at Days 1, 3, and 7, respectively. (d–f) Relative protein expression levels of interleukin‐6 (IL‐6) in the cerebral cortex at Days 1, 3, and 7, respectively. (g–i) Relative protein expression levels of glial fibrillary acidic protein (GFAP) in the cerebral cortex at Days 1, 3, and 7, respectively. Expression levels were normalized to the mean value of the Control group at each time point. Mice were assigned to the Control, LPS + Sal, and LPS + TZD groups at Days 1 and 3, and the LPS + TZD‐7d group was additionally included at Day 7. Data are shown as box‐and‐whisker plots with individual data points (n = 8 per group). Statistical analyses were performed separately at each time point using the Kruskal–Wallis test followed by the Steel–Dwass post hoc test. *p < 0.05, **p < 0.01. Representative cropped bands and corresponding source membranes for panels A and D are shown in Supporting Information S1: Figure 3. LPS, lipopolysaccharide; Sal, saline; TZD, trazodone.

Figure 6.

Protein expression levels in the hippocampus at Days 1, 3, and 7. (a–c) Relative protein expression levels of ionized calcium‐binding adaptor molecule 1 (Iba1) in the hippocampus at Days 1, 3, and 7, respectively. (d–f) Relative protein expression levels of interleukin‐6 (IL‐6) in the hippocampus at Days 1, 3, and 7, respectively. (g–i) Relative protein expression levels of glial fibrillary acidic protein (GFAP) in the hippocampus at Days 1, 3, and 7, respectively. Expression levels were normalized to the mean value of the Control group at each time point. Mice were assigned to the Control, LPS + Sal, and LPS + TZD groups at Days 1 and 3, and the LPS + TZD‐7d group was additionally included at Day 7. Data are shown as box‐and‐whisker plots with individual data points (n = 8 per group). Statistical analyses were performed separately at each time point using the Kruskal–Wallis test followed by the Steel–Dwass post hoc test. *p < 0.05, **p < 0.01. Representative cropped bands and corresponding source membranes for panel G are shown in Supporting Information S1: Figure 3. LPS, lipopolysaccharide; Sal, saline; TZD, trazodone.

During the recovery phase (Day 3), cortical GFAP expression was lower in the LPS + TZD group than in the Control group (p = 0.0272; g = 1.61, 95% CI [0.48, 2.73]; Figure 5b). Hippocampal IL‐6 expression was lower in the LPS + Sal and LPS + TZD groups than in the Control group (p = 0.0039 and p = 0.0055, respectively; g = 2.05, 95% CI [0.84, 3.26] and g = 1.90, 95% CI [0.72, 3.08]; Figure 6e).

During the postrecovery phase (Day 7), cortical Iba1 expression was lower in the LPS + Sal group than in the Control group (p = 0.0074; g = 1.80, 95% CI [0.64, 2.96]; Figure 5c), whereas it was higher in the LPS + TZD group than in the LPS + Sal group (p = 0.0074; g = 2.38, 95% CI [1.10, 3.66]) and lower in the LPS + TZD‐7d group than in the Control group (p = 0.0202; g = 1.54, 95% CI [0.43, 2.66]). Hippocampal Iba1 expression was also higher in the LPS + TZD group than in the LPS + Sal group (p = 0.0372; g = 1.46, 95% CI [0.34, 2.58]; Figure 6c). No other significant group differences were detected in the examined Iba1, IL‐6, or GFAP measures (Figures 5 and 6).

No significant group differences were detected in cortical or hippocampal TNF‐α or NeuN expression at any time point (Supporting Information S1: Figures 4 and 5).

Most outcomes did not differ between the two trazodone regimens on Day 7.

On Day 7, cortical Iba1 protein expression was lower in the LPS + TZD‐7d group than in the LPS + TZD group (p = 0.0105; g = 1.95, 95% CI [0.76, 3.14]; Figure 5c). No significant differences were detected between the two trazodone regimens in the other molecular markers, behavioral outcomes, or cognitive‐related behavioral measures examined in this study.

Supplementary across‐day analysis

As a supplementary analysis, an ART‐based two‐way ANOVA was performed for the across‐day three‐group comparison (Supporting Information S1: Table S3). Significant interactions between assessment day and treatment condition were observed for several activity‐ and exploration‐related behavioral measures, including OF total distance, YM total arm entries, and NOR total exploration count, as well as nesting score and plasma IL‐6. Among the brain molecular markers, significant interactions were observed for cortical Iba1, cortical IL‐6, hippocampal GFAP, and hippocampal Iba1. For the NOR discrimination index, a significant treatment effect was observed, although the interaction between assessment day and treatment condition did not reach statistical significance. Neither a significant treatment effect nor interaction was observed for YM spontaneous alternation. These supplementary across‐day analyses assessed overall effects and temporal patterns across assessment days and were complementary to the primary day‐specific comparisons.

DISCUSSION

This study provides preclinical evidence relevant to consultation‐liaison psychiatry by addressing whether symptomatic pharmacological intervention during the acute phase of acute brain dysfunction associated with systemic inflammation may influence acute behavioral abnormalities and later cognitive outcomes. In this LPS‐induced model of acute brain dysfunction associated with systemic inflammation, acute trazodone treatment was associated with relative preservation of stimulus‐dependent exploratory behavior and nesting during the acute phase and with better short‐term memory performance during the recovery phase. In contrast, extending trazodone administration through Day 7 did not yield a clear or consistent additional benefit beyond the shorter regimen. Several Day‐1 cortical and hippocampal inflammatory‐ and glia‐related markers also showed a different pattern in the trazodone‐treated group than in the LPS + Sal group.

During the acute phase, trazodone was associated with a distinctive behavioral profile under LPS‐treated conditions. On Day 1, locomotor measures were lower in the LPS + TZD group than in the LPS + Sal group, whereas stimulus‐dependent exploratory measures appeared relatively preserved. This suggests that the acute‐phase behavioral profile in the LPS + TZD group was not characterized simply by broader behavioral suppression. In relatively mild acute brain dysfunction, behavioral changes may be understood within the framework of sickness behavior, in which nonessential behaviors such as spontaneous locomotion are suppressed while responses to biologically relevant external stimuli are relatively preserved. ¹⁸ , ³⁴ The acute behavioral pattern observed in the LPS + TZD group is therefore more consistent with a sickness‐behavior‐like state than the broader behavioral suppression observed in the LPS + Sal group, and may indicate that acute brain dysfunction was expressed in a relatively milder behavioral form in this group.

In parallel with this acute behavioral profile, selected inflammatory and glial‐related differences were also observed during the acute phase, in both the cerebral cortex and the hippocampus. On Day 1, protein expression of cortical Iba1 and IL‐6 and of hippocampal GFAP was higher in the LPS + Sal group than in the Control group, whereas the corresponding comparisons in the LPS + TZD group were not clearly different from the Control group. In vitro studies have shown that trazodone can suppress inflammatory signaling and modulate inflammation‐related markers, including NF‐κB, Iba1, and IL‐6. ³⁵ , ³⁶ To the best of our knowledge, this is the first in vivo study to suggest that acute trazodone treatment may influence selected inflammatory and glial‐related responses in an LPS‐induced model of acute brain dysfunction, although the present study does not establish a direct mechanism.

On Day 3, the LPS + TZD group showed reduced NOR total exploration and YM arm entries. This pattern may reflect an effect of trazodone on activity and exploration during recovery, although its underlying basis remains unclear. By Day 7, no significant group differences were observed in activity or exploration measures.

Turning to cognitive aspects, nesting performance was higher in the LPS + TZD group than in the LPS + Sal group on Day 1. Nest building is a highly motivated activity of daily living in mice and is sensitive to deterioration in general condition and sickness. ²² , ²³ , ²⁴ Thus, the marked reduction in nesting observed in the LPS + Sal group is consistent with the acute sickness state induced by LPS, whereas the relative preservation of nesting in the LPS + TZD group provides additional behavioral support for the interpretation that trazodone mitigated the severity of acute brain dysfunction. Moreover, because nest building requires the organization and execution of a sequence of goal‐directed behaviors and has been used as a behavioral index reflecting executive function, ²² , ²³ , ²⁴ the better nesting performance in the LPS + TZD group may also suggest relative preservation of executive function during the acute phase. By Day 3, the marked locomotor suppression observed on Day 1 was no longer evident, whereas nesting scores remained lower in the LPS + Sal group than in the Control group. However, no significant difference was detected between the LPS + Sal and LPS + TZD groups at this time point. By Day 7, no significant group differences were observed in nesting scores. Thus, the nesting findings further support a beneficial effect of trazodone on both behavioral and cognitive aspects of acute brain dysfunction, whereas its contribution to subsequent recovery of executive function remains uncertain.

In contrast to nesting behavior, impairment in the NOR discrimination index, an index of short‐term memory, persisted beyond the acute phase in the LPS + Sal group. On Day 1, the discrimination index in the LPS + Sal group could be calculated for only four of eight mice, because the remaining four showed no object exploration, and no clear between‐group difference was observed at this time point. Between‐group differences emerged from the recovery phase onward. On Day 3, the NOR discrimination index was lower in the LPS + Sal group than in the LPS + TZD group, suggesting a between‐group difference in short‐term memory impairment during recovery. A related molecular finding was that the Day‐1 elevation of hippocampal GFAP observed in the LPS + Sal group was not evident in the LPS + TZD group, although the direct comparison between the two LPS‐treated groups did not reach statistical significance. Previous studies have linked hippocampal astrocytic responses to systemic LPS with short‐term memory impairment. ¹³ Thus, the present pattern is compatible with a possible astrocytic contribution to the relatively milder memory impairment in the LPS + TZD group, although the present data do not establish such a mechanism. On Day 7, the NOR discrimination index remained lower in the LPS + Sal group than in the LPS + TZD‐7d group, whereas no significant difference was detected between the LPS + Sal and LPS + TZD groups. Thus, acute‐phase trazodone treatment alone was associated with better short‐term memory performance during the recovery phase (Day 3), whereas the Day‐7 findings raise the possibility that continued treatment may provide an additional or more sustained benefit. However, this possibility remains uncertain because the two trazodone regimens did not differ significantly in the NOR discrimination index.

The spatial working memory findings were less clear, as no obvious between‐group difference was detected in spontaneous alternation. In the acute phase, marked reductions in locomotor activity and arm entries may have limited the sensitivity of this measure to detect spatial working memory impairment. From the recovery phase onward, no clear abnormality in spontaneous alternation was observed in the LPS + Sal group, consistent with a previous report showing no YM impairment by day 7 after systemic LPS exposure. ¹⁹ At least at the dose used here, trazodone was not associated with a clear worsening of spatial working memory despite its sedative effects.

Taken together, the present findings suggest that the principal effects of trazodone in this model emerged during the acute phase of brain dysfunction. Acute‐phase trazodone treatment was associated with a milder acute behavioral presentation, relative preservation of cognitive function during the acute phase, and better subsequent short‐term memory performance. On Day 1, elevations of cortical Iba1 and IL‐6 and hippocampal GFAP were evident in the LPS + Sal group but not clearly in the LPS + TZD group, whereas no clear change in NeuN was detected at any time point. Thus, the Day‐1 findings were more readily interpretable in the context of an acute inflammatory and glial response than of overt neuronal loss. Plasma cytokine changes were not consistently detected across the examined time points. Plasma IL‐6 did not show a significant acute‐phase elevation in the LPS + Sal group relative to the Control group, and TNF‐α showed no significant between‐group differences at any of the examined time points. Similarly, the western blot findings at Days 3 and 7 were less consistent across markers and regions. Although several significant group differences were observed at these later time points, their direction and timing varied across the molecular measures examined. These heterogeneous findings may reflect differences in the temporal dynamics of inflammatory and glial responses during the recovery process, although their functional significance remains unclear. Importantly, the two trazodone regimens did not differ significantly in the NOR discrimination index or in any other behavioral or cognitive measure, and the later molecular findings likewise did not provide a consistent basis for concluding that continued treatment yielded a clear additional benefit. Although a potential additional or more sustained benefit of continued trazodone treatment cannot be excluded, the present findings do not provide clear evidence that treatment beyond the acute phase is necessary to maintain the observed behavioral benefits. This interpretation is broadly compatible with current approaches to delirium pharmacotherapy, in which the need for continued pharmacological treatment is reassessed according to the subsequent clinical course. ³⁷ , ³⁸

LIMITATIONS

This study has several limitations. First, because different cohorts of mice were used at each assessment day, the study was not longitudinal at the level of individual mice. Accordingly, differences across days should be interpreted as group‐level temporal patterns rather than direct evidence of within‐animal recovery trajectories. Second, the evaluation of brain inflammatory‐ and glia‐related changes relied mainly on western blotting and did not directly assess cellular localization, morphology, or functional phenotypes. Third, the assessment time points were limited to Days 1, 3, and 7, and therefore may not have adequately captured very early transient cytokine responses such as TNF‐α. Fourth, because no LPS‐free trazodone group was included, the intrinsic effects of trazodone on locomotor activity, exploration, arousal, or task performance could not be clearly separated from its effects on the LPS‐induced phenotype. Fifth, because the OF and YM tests were performed during the 1‐h retention interval of the NOR test, possible effects of intervening behavioral testing, including fatigue or altered arousal, on NOR retention cannot be excluded. In addition, on Day 1 the NOR discrimination index in the LPS + Sal group could be calculated for only four of eight mice because the remaining mice showed no object exploration, introducing informative missingness into the acute NOR assessment. Sixth, the causal relationships between the behavioral and cognitive findings and the observed molecular changes were not directly demonstrated. In addition, given the multiple outcomes examined without adjustment across outcomes, the findings should be interpreted as exploratory. Finally, the LPS‐induced systemic inflammation model used here does not fully recapitulate the clinical complexity of acute brain dysfunction in humans.

CONCLUSION

In this LPS‐induced mouse model of acute brain dysfunction, acute trazodone treatment was associated with a milder acute behavioral presentation and relative preservation of cognitive function during the acute phase, as well as better subsequent short‐term‐memory‐related outcomes. Selected inflammatory and glial‐related markers also showed different acute‐phase patterns between the LPS + Sal and LPS + TZD groups. These findings suggest that the effects of acute‐phase trazodone treatment may extend beyond the acute phase to subsequent short‐term‐memory outcomes, whereas any additional benefit of continued treatment remains uncertain.

AUTHOR CONTRIBUTIONS

Hideki Matsuoka: Investigation, data curation, visualization, writing‐original draft. Leo Gotoh: Conceptualization, methodology, formal analysis, investigation, data curation, and writing—review and editing. Hikaru Hori: Conceptualization, methodology, supervision, and writing—review and editing. All authors approved the final manuscript.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

ETHICS APPROVAL STATEMENT

The study protocol was approved by the Animal Experiment Ethics Review Committee of Fukuoka University (approval no. 2313092).

PATIENT CONSENT STATEMENT

N/A.

CLINICAL TRIAL REGISTRATION

N/A.

Supporting information

Supplementary Figure 1: Representative illustration of the procedures used to obtain cropped band images and to normalize signals across membranes in the western blot analyses. Abbreviations: LPS, lipopolysaccharide; Sal, saline; TZD, trazodone; Iba1, ionized calcium‐binding adaptor molecule 1; IL‐6, interleukin‐6; TNF‐α, tumor necrosis factor‐alpha; GFAP, glial fibrillary acidic protein; NeuN, neuronal nuclei; β‐actin, beta‐actin; M, molecular weight marker lane on the membrane; R, designated reference lane containing a protein extract prepared from an unfixed whole mouse brain and used for inter‐membrane normalization; kDa, kilodaltons. This figure illustrates, using representative day‐1 samples of the LPS + Sal group (cerebral cortex), how the band images used for quantification were obtained and cropped, and how the measured signals were normalized to β‐actin. The same procedure was applied to all groups, regions, and time points. β‐actin was used as the loading control, and molecular weights are indicated in kDa. (A) Final cropped band images used for the analysis of four of the eight mice in the LPS + Sal group on day 1 (lanes A–D). The panels show GFAP, NeuN, TNF‐α, and IL‐6 together with the β‐actin signals used for their normalization; Iba1 and its corresponding β‐actin signal are shown below. (B–G) Source (uncropped) membranes from which the bands shown in (A) were cropped. These membranes contained the Control, LPS + Sal, and LPS + TZD groups (lanes A–D per group on day 1), a molecular weight marker lane (M), and the common reference lane (R). (B) Uncropped membrane used for the detection of NeuN and TNF‐α, containing the day‐1 LPS + Sal lanes (A–D) used in (A). (C) Marker reference image corresponding to panel B. This panel shows a separately imaged reference of the same molecular weight standard used to assign molecular weight positions for the NeuN and TNF‐α blots. (D) Uncropped membrane used for the detection of GFAP, containing the same day‐1 LPS + Sal lanes (A–D) and the β‐actin signal used for normalization. (E) Marker reference image corresponding to panel D. This panel shows a separately imaged reference of the same molecular weight standard used to assign molecular weight positions for the GFAP and β‐actin blots. (F) Uncropped membrane used for the detection of IL‐6, containing the same day‐1 LPS + Sal lanes (A–D). (G) Marker reference image corresponding to panel F. This panel shows a separately imaged reference of the same molecular weight standard used to assign molecular weight positions for the IL‐6 blot. Normalization across membranes: the membranes used for the detection of NeuN, TNF‐α, and IL‐6 did not contain a β‐actin signal, whereas GFAP and β‐actin were detected in the same image (D). To normalize proteins detected on different membranes against the same β‐actin reference, the reference lane (R), which was loaded with an extract of the same unfixed whole mouse brain and was common to all membranes, was used to connect the signal values obtained on the different membranes; the values were then normalized to the β‐actin signal detected in the same image as GFAP (D). (H) Uncropped membrane used for the detection of Iba1. In contrast to the membranes in (B), (D), and (F), the lane layout of this membrane differs. Because Iba1 and β‐actin were detected in the same image, Iba1 was normalized directly to β‐actin without inter‐membrane connection through R. (I) Marker reference image corresponding to panel H. This panel shows a separately imaged reference of the same molecular weight standard used to assign molecular weight positions for the Iba1 and β‐actin blots.

Supplementary Figure 2: Plasma tumor necrosis factor‐alpha (TNF‐α) concentrations at days 1, 3, and 7. Abbreviations: LPS, lipopolysaccharide; Sal, saline; TZD, trazodone. (A–C) Plasma TNF‐α concentrations at days 1, 3, and 7, respectively. Plasma TNF‐α concentrations were measured by enzyme‐linked immunosorbent assay at days 1, 3, and 7. Data are shown as box‐and‐whisker plots with individual data points (n = 8 per group). Statistical analyses were performed separately at each time point using the Kruskal–Wallis test followed by the Steel–Dwass post hoc test. No significant differences were observed among groups at any time point.

Supplementary Figure 3: Representative cropped bands and source membranes corresponding to the day‐1 findings central to the present discussion (Figures 5A, 5D, and 6g). Abbreviations: LPS, lipopolysaccharide; Sal, saline; TZD, trazodone; Iba1, ionized calcium‐binding adaptor molecule 1; IL‐6, interleukin‐6; GFAP, glial fibrillary acidic protein; M, molecular weight marker lane on the membrane; R, designated reference lane containing a protein extract prepared from an unfixed whole mouse brain and used for inter‐membrane normalization; β‐actin, beta‐actin; kDa, kilodaltons. β‐actin was used as the loading control. Molecular weights are indicated in kDa. (A) Representative cropped bands corresponding to Figure 5a (cortical Iba1 and β‐actin), Figure 5d (cortical IL‐6 and β‐actin), and Figure 6g (hippocampal GFAP and β‐actin) on day 1. Four representative lanes are shown for each group (Control, LPS + Sal, and LPS + TZD). (B) Source Iba1 blot corresponding to Figure 5a for the LPS + Sal group. This source membrane contains the LPS + Sal day‐1 lanes used in panel A and also includes additional lanes, including LPS + Sal day 3. The image shown is a uniformly contrast‐adjusted view to facilitate visualization of the Iba1 band. (C) Source β‐actin blot corresponding to Figure 5a for the LPS + Sal group. This source membrane contains the LPS + Sal day‐1 lanes used in panel A and also includes additional lanes, including LPS + Sal day 3. (D) Marker reference image corresponding to panels B and C. This panel shows a separately imaged reference of the same molecular weight standard used to assign molecular weight positions for the source Iba1 and β‐actin blots. (E) Source Iba1 blot corresponding to Figure 5a for the Control and LPS + TZD groups. This source membrane contains the day‐1 lanes used in panel A for these two groups. The image shown is a uniformly contrast‐adjusted view to facilitate visualization of the Iba1 band. (F) Source β‐actin blot corresponding to Figure 5a for the Control and LPS + TZD groups. This source membrane contains the day‐1 lanes used in panel A for these two groups. (G) Marker reference image corresponding to panels E and F. This panel shows a separately imaged reference of the same molecular weight standard used to assign molecular weight positions for the source Iba1 and β‐actin blots. (H) Source IL‐6 blot corresponding to Figure 5d. This source membrane contains the Control, LPS + Sal, and LPS + TZD groups on day 1. The image shown is a uniformly contrast‐adjusted view to facilitate visualization of the IL‐6 band. (I) Marker reference image corresponding to panel H. This panel shows a separately imaged reference of the same molecular weight standard used to assign molecular weight positions for the source IL‐6 blot. (J) Source β‐actin blot used for normalization of the IL‐6 blot corresponding to Figure 5d. A designated reference lane containing a protein extract prepared from an unfixed whole mouse brain maintained in the laboratory was used for inter‐membrane normalization. (K) Marker reference image corresponding to panel J. This panel shows a separately imaged reference of the same molecular weight standard used to assign molecular weight positions for the source β‐actin blot. (L) Source GFAP and β‐actin blot corresponding to Figure 6g. This hippocampal source membrane contains the Control, LPS + Sal, and LPS + TZD groups on day 1. The image shown is a uniformly contrast‐adjusted view to facilitate visualization of the GFAP band. (M) Marker reference image corresponding to panel L. This panel shows a separately imaged reference of the same molecular weight standard used to assign molecular weight positions for the source GFAP and β‐actin blot.

Supplementary Figure 4: Additional protein expression levels in the cerebral cortex at days 1, 3, and 7. Abbreviations: LPS, lipopolysaccharide; Sal, saline; TZD, trazodone. (A–C) Relative protein expression levels of neuronal nuclei (NeuN) in the cerebral cortex at days 1, 3, and 7, respectively. (D–F) Relative protein expression levels of tumor necrosis factor‐alpha (TNF‐α) in the cerebral cortex at days 1, 3, and 7, respectively. Expression levels were normalized to the mean value of the Control group at each time point. Mice were assigned to the Control, LPS + Sal, and LPS + TZD groups at days 1 and 3, and to the Control, LPS + Sal, LPS + TZD, and LPS + TZD‐7d groups at day 7. Data are shown as box‐and‐whisker plots with individual data points (n = 8 per group). Statistical analyses were performed separately at each time point using the Kruskal–Wallis test followed by the Steel–Dwass post hoc test. No significant differences were observed among groups at any time point.

Supplementary Figure 5: Additional protein expression levels in the hippocampus at days 1, 3, and 7. Abbreviations: LPS, lipopolysaccharide; Sal, saline; TZD, trazodone. (A–C) Relative protein expression levels of neuronal nuclei (NeuN) in the hippocampus at days 1, 3, and 7, respectively. (D–F) Relative protein expression levels of tumor necrosis factor‐alpha (TNF‐α) in the hippocampus at days 1, 3, and 7, respectively. Expression levels were normalized to the mean value of the Control group at each time point. Mice were assigned to the Control, LPS + Sal, and LPS + TZD groups at days 1 and 3, and to the Control, LPS + Sal, LPS + TZD, and LPS + TZD‐7d groups at day 7. Data are shown as box‐and‐whisker plots with individual data points (n = 8 per group). Statistical analyses were performed separately at each time point using the Kruskal–Wallis test followed by the Steel–Dwass post hoc test. No significant differences were observed among groups at any time point.

Supporting File 1:

Supporting File 2:

Supporting File 3:

DATA AVAILABILITY STATEMENT

The data supporting the findings of this study are available within the article and its Supplementary Information files. Additional data are available from the corresponding author upon reasonable request.