Effects of Toll-like receptor 4 inhibition on spatial memory and cell proliferation in male and female adult and aged mice

Abstract

Toll-like receptors (TLRs) participate in the response to infection, stress, and injury by initiating an innate immune response. In addition, these receptors are expressed in many neural cell types and under physiological conditions, are implicated in modulating cognitive function and neural plasticity in the adult and aged brain. Knockout of the Toll-like receptor 4 (TLR4) subtype enhances spatial memory and adult hippocampal neurogenesis through increasing proliferation and neuronal differentiation. Currently unknown is whether pharmacological inhibition of TLR4 produces similar enhancements in cognitive function and cell proliferation. The present study evaluated water maze performance, cytokine expression, and cell proliferation in the hippocampus of young and aged male and female C57BL6/J mice following treatment with the TLR4 antagonist, TAK-242. Further, alterations in the response to an acute stressor were evaluated in TAK-242-treated mice. Results showed that TAK-242 selectively enhanced spatial learning and memory in young females. Additionally, TAK-242 treatment reduced thigmotaxis in the water maze and lowered corticosterone levels following acute stress in females. TAK-242 decreased hippocampal interleukin (IL)-1β expression but had no effect on IL-6 or tumor necrosis factor-α (TNFα). Aged mice showed decreased cell proliferation compared to young mice, but TAK-242 administration had minimal effects on estimated Ki67 positive cell numbers. Findings indicate that pharmacological inhibition of TLR4 improves cognitive function in young females likely through attenuating stress reactivity.

1. Introduction

Toll-like receptors (TLRs) constitute a vital component of the immune system, as they recognize and bind to conserved pathogen-associated molecular patterns (PAMPs) and mobilize an inflammatory defense (Barak et al., 2014; Trotta et al., 2014). The Toll-like receptor 4 (TLR4) subtype recognizes motifs from Gram-negative bacteria and through downstream signaling components, stimulates production of inflammatory molecules, including cytokines, to defend against an infection (Barak et al., 2014; Trotta et al., 2014). TLR4 also participates in the neuroinflammatory response to stress, injury, and neurodegenerative disease as pharmacological inhibition of TLR4 reduces neuroinflammation (Cui et al., 2020; Feng et al., 2017; Garate et al., 2014). One consequence of the inflammatory response induced by TLR4 activation is impaired cognitive function. For instance, administration of a TLR4 agonist such as lipopolysaccharide (LPS) impairs learning and memory processes in hippocampus-related tasks such as water maze, contextual fear conditioning, and two-way active avoidance conditioning (Kranjac et al., 2012; Pugh et al., 1998; Sparkman et al., 2005a; Sparkman et al., 2005b). Similarly, cognitive deficits are observed following a traumatic brain injury, which can be attenuated by administration of the TLR4 inhibitor, TAK-242 (Resatorvid) (Feng et al., 2017; Korgaonkar et al., 2020). TAK-242 is a small molecule antagonist that crosses the blood brain barrier (BBB) following systemic administration that selectively targets TLR4 while having minimal interaction with other TLRs (Takashima et al., 2009; Wang et al., 2013). TAK-242 inhibits both MyD88-dependent and - independent pathways following TLR4 activation by binding to the Cys747 intracellular domain (Takashima et al., 2009). While TLR4 activation elicits an immune response, regulation of its activity may prove beneficial in preserving cognitive function.

TLR4 appears to modulate cognitive processes even in the absence of an infection or active inflammation (Okun et al., 2012; Potter et al., 2019). For instance, TLR4 deficient mice show enhanced spatial learning and memory in a water maze task compared to wild-type (WT) mice (Okun et al., 2012). The enhancements in spatial memory conferred by TLR4 deficiency vary by sex and age, as the cognitive benefits are present only in young males and aged females (Okun et al., 2012; Potter et al., 2019). While the absence of TLR4 improves spatial memory, the benefits may not be widespread. For example, TLR4 knockout (KO) mice do not show enhancements when compared with WT mice in a passive avoidance task or an object recognition task (Li et al., 2016; Pascual et al., 2011). Moreover, TLR4 KO mice show reduced freezing in both contextual and cued fear conditioning compared to WT mice, suggesting impaired memory. These deficits, however, may not be specific to the hippocampus (Okun et al., 2012). Collectively, these data indicate that TLR4 modulates select cognitive processes in the healthy brain.

In the context of aging, TLR4 expression is elevated in the brain in line with a basal increase in microglia activation and neuroinflammatory molecules (Cribbs et al., 2012; Jurgens and Johnson, 2012; Letiembre et al., 2007; Ye and Johnson, 1999). The persistent neuroinflammation associated with aging contributes to physiological and functional impairments including cognitive deficits, increased anxiety- and depression-like behaviors, and impairments in neurogenesis and other measures of plasticity (Jurgens and Johnson, 2012; Kohman and Rhodes, 2013; Yirmiya and Goshen, 2011). Prior work suggests that TLR4 modifies aspects of peripheral and central inflammation. For instance, vascular cells from middle-aged TLR4 deficient mice produce lower levels of interleukin (IL)-6 compared to WT mice (Song et al., 2012). Additionally, while TLR4 deficiency does not fully eliminate age-related neuroinflammation, it appears to modulate IL-1 signaling in the aged brain (Potter et al., 2019). Thus, TLR4 appears to play a role in inflammation and cognitive function in the aged however, the precise mechanism is presently unclear.

TLR4 has also been implicated in altering neural plasticity processes, specifically adult hippocampal neurogenesis (Connolly et al., 2020; Okun et al., 2011; Rolls et al., 2007). TLR4 expression on neural progenitor cells (NPCs) modulates their proliferation and differentiation (Okun et al., 2011; Rolls et al., 2007; Ye et al., 2014). Activation of TLR4 reduces survival of newly born hippocampal cells and neuronal differentiation in the young and aged brain (Bastos et al., 2008; Ekdahl et al., 2003; Littlefield et al., 2015; Monje et al., 2003). In contrast, the absence of TLR4 enhances neurogenesis. Prior work reports that adult TLR4 KO male mice show an increase in cell proliferation and the proportion of immature neurons as assessed by doublecortin (DCX) positive cells. However, survival of these immature neurons may be reduced in TLR4 KO mice as total numbers of new cells does not differ between TLR4 KO and WT mice at later time points (Rolls et al., 2007). The effects of TLR4 deficiency on neurogenesis vary by sex, as female TLR4 KOs show higher rates of neuronal differentiation compared to TLR4 KO males across the entire granular cell layer (Connolly et al., 2020). Moreover, this effect varied as a function of hippocampal subdivision, as TLR4 deficient females had higher rates of proliferation and neuronal differentiation in both the dorsal and ventral hippocampus compared to WT females. However, elevated proliferation and neurogenesis were restricted to the ventral hippocampus in TLR4 KO males. While hippocampal neurogenesis has been suggested to contribute to cognitive function and emotion-related behaviors (Deng et al., 2010; Dupret et al., 2008; Revest et al., 2009), whether these TLR4-mediated effects on neurogenesis impact learning and memory processes is currently unclear. Collectively, the data indicate that TLR4, under physiological conditions, negatively affects adult hippocampal neurogenesis and that these impacts may be greater in females (Connolly et al., 2020; Rolls et al., 2007).

Much of our understanding of TLR4’s contribution to cognitive function and hippocampal neurogenesis under physiologically normal conditions is based on TLR4 KO mice. However, in a clinical setting, modulating TLR4 would likely occur through administration of TLR4 antagonists over a select treatment period. Whether pharmacological inhibition of TLR4 produces similar effects on spatial learning and hippocampal neurogenesis is presently unknown. We predicted that administration of the TLR4 antagonist, TAK-242, would enhance cognitive function and cell proliferation and that these effects may vary by sex as observed in the TLR4 KO mice (Connolly et al., 2020; Okun et al., 2012; Potter et al., 2019; Rolls et al., 2007). Further, we expected that any effects of TAK-242 in the aged mice would relate to reductions in proinflammatory cytokine levels.

2. Materials and methods

2.1. Experimental subjects and design

A total of 148 male and female C57BL6/J mice were included in the present experiments. Eighty male and female young (<7 months) and aged (19–21 months) C57BL6/J mice were tested in the water maze and evaluated for cell proliferation. A separate group of young (4 months) and aged (20 months) female and male C57BL6/J mice were evaluated for stress reactivity as assessed by plasma corticosterone levels following an acute restraint stress protocol. Mice were produced in-house from breeding pairs purchased from The Jackson Laboratory (Bar Harbor, ME). Mice were group-housed under a reverse light–dark cycle with ad libitum access to food and water. The experiments were approved by the Institutional Animal Care and Use Committee (IACUC) at the University of North Carolina Wilmington and were carried out in alignment with the requirements of the Guide for the Care and Use of Laboratory Animals. Mice evaluated in the water maze were semi-randomly assigned to receive 14 daily intraperitoneal (ip) injections of either vehicle (1% dimethyl sulfoxide [DMSO] in saline) or TAK-242 (3 mg/kg) across ages and sexes, for eight treatment groups. The selected TAK-242 dose and treatment regime were based on prior reports that showed a 3 mg/kg dose of TAK-242 attenuated behavioral deficits and/or inflammation in models of traumatic brain injury, sepsis, and stroke when administered for 5 days prior to the insult (Takashima et al., 2009; Wang et al., 2013; Zhang et al., 2014). Based on these data, mice evaluated in the water maze received injections of TAK-242 or vehicle starting 5 days prior to behavioral testing (days 1–5), with administration continuing through behavioral testing and until tissue samples were collected (days 6–14). Groups ranged from 9 to 12 mice per condition. TAK-242 was dissolved in DMSO (final concentration 1% DMSO) and then diluted in saline. TAK-242’s low molecular weight and high lipid solubility allows the compound to cross the BBB (Wang et al., 2013). Prior work has shown that TAK-242 is detected in the brain between 30 minutes and 24 hours after a single ip injection (Hua et al., 2015; Wang et al., 2013). Body weight was assessed each day prior to injection. A difference score was calculated by subtracting the current day’s body weight from the previous day’s weight to quantify gains or losses in weight throughout the experiment. The separate groups of male and female mice tested for stress responsivity were semi-randomly assigned to the home cage or the stress condition and were given daily ip injections of TAK-242 (3 mg/kg) or vehicle for two consecutive days. Based on prior reports that showed administering a TLR4 antagonist just prior to stress exposure attenuated stress-induced changes in behavior and inflammation, injections were given 15 minutes prior to restraint stress (Fu et al., 2019; Garate et al., 2014; Wang et al., 2018).

2.2. Water maze task

The water maze consisted of a white circular tub (116.8 cm in diameter) filled with white tinted water (non-toxic paint) maintained at a temperature of 20 ± 1 °C. During training, a circular platform (11.4 cm in diameter) was submerged 1 cm under the water and remained in the same location during the five days of training. Extra-maze cues (e.g., shapes, posters, curtains, furniture) surrounded the maze. Mice received four trials per day that started from one of four start locations (semi-randomly ordered) during the five training days. Each trial lasted until the mouse located the hidden platform or 60 seconds passed, after which the mouse was gently guided to the platform. Mice were left on the platform for 10 seconds before being removed from the tub. Behavioral testing occurred during the dark cycle, approximately 2.5 hours after light offs and remained constant across testing days. Tracking software (Topscan, Clever Systems, Reston, VA) recorded each trial and was used to measure cumulative distance from the platform (mm), swim speed (mm/sec), and the proportion of time spent in the tub’s outer annulus (i.e., thigmotaxic behavior). Cumulative distance was calculated by summing one second averages of the distance from the platform (10 samples taken each second) over the 60 seconds of the trial (Gallagher et al., 1993). Data analysis was conducted on the average of the four trials for each training day.

Spatial memory was evaluated during a single 60 second probe trial that occurred seventy-two hours after training (Potter et al., 2019). The platform was removed during the probe trial and the tracking software recorded the distance (mm) from the platform’s location during training (sampled 10 times each second) which allowed for calculation of the average proximity to the original platform location. Averages of the 10 samples per second were calculated and then averaged over the entire 60 second trial (Gallagher et al., 1993). The proximity measure has been shown to be a more reliable and sensitive measure than other commonly employed measures in the probe trial, particularly with aged subjects (Gallagher et al., 1993; Maei et al., 2009; Tomas Pereira and Burwell, 2015). Three hours after the probe trial, a visual platform task was conducted in which the platform was made visible by raising it 1 cm above the water and covering it with a dark blue top. The visible platform was positioned in the quadrant opposite to the platform location during training. The visible platform task was conducted to monitor for potential changes in performance factors (e.g., visual impairment, variations in motivation) that could have impaired performance during the hidden platform task and the probe trial. Two visible platform trials were conducted for each subject, with each trial starting from a different location. The two trial averages of path length (mm) and swim speed (mm/s) were collected and used for analysis.

2.3. Acute restraint stress

Separate groups of young and aged female and male mice were administered TAK-242 (3 mg/kg) or vehicle via ip injections on two consecutive days. Data from males and females were collected in individual experiments and therefore the data were analyzed separately. Mice in the restraint stress condition were placed in a 50 ml conical tube for 2 hours each day on two consecutive days. The restraint tubes had holes for ventilation and to allow the mouse’s tail to protrude from the lid. Mice in the home cage condition were returned to their home cage following vehicle or TAK-242 administration and left undisturbed.

2.4. Sample collection

Twenty-four hours after the water maze task, mice were anesthetized with sodium pentobarbital (150 mg/kg) and transcardially perfused with ice-cold saline. Brains were bisected into equal left and right halves down the midline. Half of the brain was post-fixed overnight in 4% paraformaldehyde, transferred into 30% sucrose at 4°C, and stored until brains were sectioned into 40 μm coronal sections using a cryostat. Sections were stored at −20 °C in a cryoprotectant solution until staining. The hippocampus was dissected from the non-fixed hemisphere of the brain. The hippocampus and brain (i.e., everything but the hippocampus) were snap frozen on dry ice and stored at −80 °C until assayed. For male and female mice evaluated for stress reactivity, blood samples were collected in heparin-coated tubes following rapid decapitation on the second day of treatment from home cage controls and stress mice (immediately following second restraint session). To reduce the influence of diurnal changes in corticosterone levels, which peak prior to the onset of the active period (i.e., dark cycle for nocturnal rodents), blood samples were collected at the same time of day. Specifically, samples were collected between 4–5 hours after onset of the dark cycle (Dickmeis, 2009; Kakihana and Moore, 1976). Samples were centrifuged (2000 rpm for 15 minutes at 4°C), and plasma collected and stored at −80 °C until assayed.

2.5. Immunohistochemistry and image analysis: Cell Proliferation

To evaluate cell proliferation in the hippocampus, one-in-six series were stained for Ki67. Sections were initially washed in tris-buffered saline (TBS). Sections were then exposed to a 0.6% solution of hydrogen peroxide, washed again, and then DNA denatured by exposing sections first to a 50% de-ionized formamide solution in a water bath at 65 °C for 90 minutes, then a rinse with a 2X solution of saline-sodium citrate (SSC) buffer at room temperature for 15 minutes, then 2N hydrochloric acid at 37 °C for 30 minutes, and finally 0.1M borate buffer at room temperature for 10 minutes. Sections were washed again and then blocked for 30 minutes using a solution of TBS with 5% normal goat serum and 0.3% Triton-X. Sections were then incubated in rabbit anti-Ki67 (1:10,000; Abcam Cat# ab15580, RRID: AB443209) with gentle rotation at 4 °C for 72 hours. The primary antibody was washed out, sections were blocked again, and were incubated in the biotinylated goat anti-rabbit secondary antibody (1:250; Vector Laboratories Cat# BA-1000, RRID: AB2313606) for 100 minutes at room temperature. After washing, sections were exposed to avidin-biotin complex solution (ABC; Vector Laboratories, Burlingame, CA) for 60 minutes and then stained with a diaminobenzidine (DAB) solution. Sections were mounted onto slides, counterstained with methylene blue, and coverslipped. The estimated number of Ki67 positive cells was quantified in the granular cell layer throughout the entire hippocampus (i.e., dorsal, ventral, and intermediate) as well as analyzed separately for the dorsal (−1.06 to −2.06 mm to bregma) and the ventral (−3.08 to −3.88 mm to bregma) subdivisions of the hippocampus (Tanti et al., 2012; Connolly et al., 2020). Boundaries of the dorsal and the ventral hippocampus are based on prior reports (Connolly et al., 2020; Rainer et al., 2012; Tanti and Belzung, 2013; Tanti et al., 2012).

Images of a single plane selected through the z-axis were taken on a Zeiss light microscope at 20X magnification. Granular cell layer images were captured from 12–14 sections per animal with 2–4 sections for the dorsal granular cell layer analysis and 2–5 sections for the ventral granular cell layer analysis. ImageJ was used to count the Ki67 positive cells and outline the granular cell layer in every image. A validated threshold was applied in ImageJ to remove background and detect positive cells. To calculate the granular cell layer volume, the outline area (pixels) was summed, converted to micrometers, and then multiplied by the thickness of the section (40 μm). Estimates of Ki67 positive cells were calculated as a density, the number of positive cells per cubic micrometer of the granular cell layer.

2.6. RT-PCR: Hippocampal IL-1β, TNFα, and IL-6

Hippocampal samples were analyzed for changes in expression of IL-1β, IL-6, and TNFα using methods previously described (Littlefield and Kohman, 2017; Potter et al., 2019). Briefly, RNA was extracted with the RNeasy Mini Lipid Kit (Qiagen, Valencia, CA) and reverse transcription performed using the High-capacity cDNA reverse transcription kit (Applied Biosystems, Foster City, CA) by following the manufacturer’s instructions. The 260/280 purity ratio was above 2.0 for all samples. Hippocampal samples were assessed for expression of IL-1β (Mm00434228_m1), IL-6 (Mm00446190_m1), and TNFα (Mm00443260_g1) using Taqman probe/primer sets. β-actin (Mm00607939_s1) was measured as the endogenous control gene. The 2^-ΔΔCT method was used to evaluate relative changes in gene expression compared to a calibrator (i.e., young vehicle-treated female).

2.7. ELISA: Interleukin-6 and corticosterone

The non-fixed hemisphere of the brain from the mice tested in the water maze were homogenized in the 1mM serine protease inhibitor, phenylmethylsulfonyl fluoride (PMSF). The samples were then centrifuged at 3500 rpm for 30 minutes at 4 °C. The samples were assayed for total protein using a protein assay (Pierce Thermo Scientific, Rockford, IL) and were assayed for protein levels of IL-6 by an ELISA (BD Biosciences, San Diego, CA; detection limits 15.6 to 1000 pg/ml). The results of the IL-6 ELISA are given in picograms (pg) of cytokine per milligram (mg) of total protein. Separate ELISAs (Enzo, Farmingdale, NY) were run on the plasma samples collected from male and female mice (stress and home cage control conditions) in accordance with the manufacturer’s instructions to evaluate plasma corticosterone levels (ng/ml).

2.8. Statistical analyses

Water maze training measures (i.e., thigmotaxis, cumulative distance, speed) and change in body weight were analyzed using repeated measures ANOVAs. The between-subject factors were age (young or aged), sex (male or female), and treatment (TAK-242 or vehicle). The within-subject factor was day (i.e., days 1–5 of training or days 1–14 of injections). The probe trial, visual platform, cytokine, and the cell proliferation data were analyzed using factorial ANOVAs. The between-subject factors were age, sex, and treatment. Pairwise comparisons were run using t-tests to determine whether effects were limited to one age group, when the young mice drove significant treatment interactions. Plasma corticosterone data were analyzed by factorial ANOVA with age, stress condition (home cage or stress), and treatment as the between-subject factors. Normality was tested using the Shapiro-Wilk test. Data that were not normally distributed were log transformed. When effects were found to be significant (i.e., p value less than 0.05), Fisher’s least significant difference test was used as a post hoc test.

3. Results

3.1. Body weight

Analysis of the change in body weight during the 14 days of TAK-242 or vehicle administration showed that there was a main effect of age and an age × day interaction (F(1,72)=18.32, p<0.0001, F(1,72)=2.53, p<0.005, respectively, see Figures 1a and 1b). Of note, is that all gains or losses in weight were less than 1 gram of the prior day’s weight. Post hoc testing revealed that aged mice, regardless of sex or treatment, lost more weight than young mice the day after the first injection (i.e., day 2) as well as after the sixth injection (i.e., day 7) (p<0.05). Additionally, there was a sex × day interaction (F(1,72)=3.03, p<0.0005, see Figures 1a and 1b) that showed female mice, regardless of age or treatment, lost more weight than males on days 3 and 11, whereas males lost more weight than females on days 8 and 12 (p<0.05). There were no effects of TAK-242 administration on the change in body weight, as TAK-242- and vehicle-treated mice showed similar changes in weight throughout the two weeks of treatment.

Figure 1.

Body weight: Line graphs show mean change in body weight in grams for females (A) and males (B) during 14 days of TAK-242 or vehicle administration ± standard error of the means (SEMs). Change in weight was calculated by subtracting current body weight from the previous day’s weight. Female and male data were analyzed together but graphed separately to improve visualization of the data. In response to injections, aged mice lost more weight than young mice on days 2 and 7. Males and females showed differential changes in weight on days 3, 8, 11, and 12. Body weight was unaffected by treatment, as mice in the TAK-242 and vehicle conditions showed similar alterations in body weight. Changes in weight throughout the 14 days were within one gram of the mouse’ prior weight. Symbols indicate a significant difference between young and aged (#) and between males and females (^).

3.2. Water maze training: Spatial learning

A significant age × test day interaction for cumulative distance from the platform (F(4,288)=2.61, p<0.5, see Figures 2a and 2b) showed that both young and aged mice decreased their cumulative distance from the platform across the 5 days of testing, but aged mice had longer distances than young mice on days 3–5 of testing (p<0.05). A significant treatment × sex interaction (F(1,72)=6.65, p<0.05) revealed that administration of the TLR4 antagonist, TAK-242, decreased overall cumulative distance to the platform in females compared to vehicle-treated females (p<0.05, see Figure 2a and 2c), whereas TAK-242 had no effect on cumulative distance in males (see Figure 2b and 2c). Vehicle-treated females had greater cumulative distances compared to vehicle-treated males (p<0.05). Pairwise comparison of the effects of TAK-242 within the age groups showed that TAK-242 administration only reduced cumulative distance from the platform in young females relative to vehicle-treated young females (p<0.01, see Figure 2a and 2c), whereas there was no difference between TAK-242 and vehicle-treated aged females.

Figure 2.

Spatial learning: Line graphs show the mean cumulative distance from the platform across the five days of training for females (A) and males (B) ± SEMs. Bar graphs show the mean cumulative distance (C), proportion of time in the thigmotaxis area (D), and swim speed (E) ± SEMs collapsed across the five days of training. Vehicle-treated females had increased cumulative distance to the platform and proportion of time in the thigmotaxic area relative to vehicle-treated males. TAK-242 reduced cumulative distance and percent time in thigmotaxis in young females. Aged mice differed from young mice on all measures. Symbols indicate a significant difference from young mice (#), vehicle-treated males (^), and TAK-242-treated females (*).

Analysis of percent time engaged in thigmotaxic behavior showed that aged mice spent a greater proportion of time in the thigmotaxis region relative to young mice (main effect of age, F(1,72)=13.07, p<0.01, see Figure 2d). Further, a treatment × sex interaction (F(1,72)=4.91, p<0.05, see Figure 2d) showed that vehicle-treated females spent a greater proportion of time in the thigmotaxis region than vehicle-treated males (p<0.05), whereas there were no differences between the TAK-242-treated males and females. TAK-242 administration decreased the percent time females spent in the thigmotaxis region relative to vehicle-treated females (p<0.05). Pairwise comparisons showed that TAK-242 administration only decreased percent time engaged in thigmotaxic behavior in the young female mice (p<0.05, see Figure 2d).

For swim speed, there was a main effect of age (F(1,72)=37.75, p<0.0001, see Figure 2e) and an age × sex × day interaction (F(4,288)=4.43, p<0.05) that showed aged males swam slower than young males on days 2, 4 and 5 of training (p<0.05). Aged females swam slower than young females on days 1–4 of training (p<0.05). Aged females swam faster than aged males on days 2, 4, and 5 (p<0.05). Finally, young females swam faster than young males on days 1–3 of training (p<0.05). There were no effects of treatment on swim speed.

3.3. Water maze: Spatial memory

Analysis of the average proximity to the platform during the probe trial revealed that aged mice were farther away from the platform’s training location during the probe trial compared to young mice (main effect of age, F(1,72)=8.62, p<0.05, see Figure 3). Further, a treatment × sex interaction for average proximity (F(1,72)=4.58, p<0.05, see Figure 3) showed that females treated with TAK-242 were closer than vehicle-treated females (p<0.05) to the platform’s training location during the probe trial. Vehicle-treated females were farther away from the platform than vehicle-treated males (p<0.05). Pairwise comparisons showed that only young females showed a reduction in their average proximity to the platform following TAK-242 administration (p<0.05), as aged vehicle- and TAK-242-treated females did not differ.

Figure 3.

Spatial memory: Bar graph shows the mean proximity to the training platform location ± SEMs during the probe trial. Vehicle-treated females were farther from the platform than vehicle-treated males and TAK-242-treated females. Aged mice were farther from the platform than young mice. Symbols indicate a significant difference from young mice (#), vehicle-treated males (^), and TAK-242-treated females (*).

3.4. Water maze: Visible platform

In the visible platform test, there were no significant differences in path length to locate the visible platform, as all groups swam similar distances (see Figure 4a). For swim speed, aged mice swam slower than young mice (main effect of age, F(1,72)=4.70, p<0.05, see Figure 4b) and females swam faster than males (main effect of sex, F(1,72)=6.69, p<0.05, see Figure 4b). TAK-242 administration had no effect on path length or swim speed.

Figure 4.

Visual platform task: Bar graphs show mean distance swam to locate the visible platform (A) and swim speed (B) ± SEMs. Groups did not differ in distance swam (i.e., path length) to the visible platform. Aged mice swam slower than young mice and females swam faster than males. Symbols indicate a significant difference from young mice (#) and males (^).

3.5. Cell proliferation

Assessment of the estimated number of Ki67 positive cells across the entire granular cell layer of the hippocampus showed a main effect of age (F(1,72)=177.59, p<0.001, see Figure 5a) with aged mice having fewer Ki67 positive cells than young mice. There were no effects of treatment on Ki67 positive cells in the entire hippocampus. No differences in granular cell layer volume were found across the treatment groups (data not shown). In the dorsal hippocampus there was a significant main effect of age and an age × sex interaction (F(1,72)=116.09, p<0.0001; F(1,72)=5.09, p<0.05, respectively, see Figure 5b) that showed aged male and female mice had fewer Ki67 positive cells compared to their young counterparts (p<0.0001). Young females had fewer Ki67 positive cells than young males (p<0.05), this difference was driven by the TAK-242-treated females. Pairwise comparison showed that vehicle-treated male and female mice did not differ in number of Ki67 positive cells in the dorsal granular cell layer, whereas TAK-242-treated females had fewer dorsal Ki67 positive cells compared to TAK-242-treated young males (p<0.05). No difference existed between aged male and female mice. The ventral hippocampus had a main effect of age (F(1,72)=94.20, p<0.001, see Figure 5c) with aged mice having fewer Ki67 positive cells than young mice. There were no effects of treatment or sex on Ki67 positive cells in the ventral granular cell layer.

Figure 5.

Cell proliferation: Bar graphs show estimated mean numbers of Ki67 positive cells throughout the entire granular cell layer (A), dorsal hippocampus (B), and ventral hippocampus (C) ± SEMs. Representative images of Ki67 positive cells across treatment groups (D). Aged mice showed fewer Ki67 positive cells than young mice. TAK-242 administration had no effect on Ki67 positive cells when evaluated across the entire hippocampus or within the ventral division. Young TAK-242-treated females showed fewer Ki67 positive cells in the dorsal hippocampus than young TAK-242-treated males but did not differ from vehicle-treated females. Symbols indicate a significant difference from young mice (#) and TAK-242-treated males (^).

3.6. Hippocampal cytokines expression

Aged mice showed higher expression of IL-1β, TNFα, and IL-6 compared to young mice, as shown by main effects of age (F(1,72)=25.64, p<0.0001; F(1,72)=29.59, p<0.0001; F(1,72)=11.52, p<0.001, see Figures 6a, 6b, 6c, respectively). A main effect of treatment for IL-1β (F(1,72)=18.60, p<0.0001) showed that TAK-242 administration significantly reduced IL-1β expression relative to vehicle-treated mice (see Figure 6a). TAK-242 administration had no effect on hippocampal expression of TNFα or IL-6 (see Figures 6b and 6c). For TNFα, there was a main effect of sex (F(1,72)=6.30, p<0.05, see Figure 6b) that showed females had higher hippocampal expression of TNFα compared to male mice. There were no differences in β-actin expression across the groups (data not shown).

Figure 6.

Hippocampal cytokine expression and brain protein levels of IL-6: Bar graphs show mean hippocampal fold change of IL-1β (A), IL-6 (B), TNFα (C), and protein levels of IL-6 in brain samples (D) ± SEMs. TAK-242 administration decreased hippocampal IL-1β expression relative to vehicle-treated mice but had no effect on TNFα or IL-6. Cytokines were higher in the aged mice compared to young mice. Females showed higher IL-1β and TNFα expression compared to males. Symbols indicate significant difference from young mice (#), males (^), and TAK-242-treated mice (*).

3.7. Brain IL-6 levels

Protein levels of IL-6 were elevated in the aged brain relative to young mice, as shown by a main effect of age (F(1,72)=23.90, p<0.0001, see Figure 6d). There were no differences in IL-6 levels between males and females. TAK-242 administration had no effect on central IL-6 levels.

3.8. Plasma corticosterone levels

For female mice, there was a significant treatment × stress interaction (F(1,30)=5.67, p<0.05, see Figure 7a) for plasma corticosterone levels. Female mice exposed to acute restraint stress had higher corticosterone levels than home cage controls (p<0.005). Treatment with TAK-242 reduced the stress-induced increase in plasma corticosterone compared to vehicle-treated mice (p<0.05). TAK-242 had no effect on corticosterone levels in female home cage controls. Age had no effect on plasma corticosterone levels in either the stress or the home cage female mice. For male mice, there was a main effect of stress (F(1,22)=106.55, p<.0001, see Figure 7b) that showed acute restraint increased plasma corticosterone levels relative to home cage controls. Administration of TAK-242 had no effect on corticosterone levels in young and aged males, as there was no significant main effect or interaction of treatment. Similar to females, age had no effects on plasma corticosterone levels in males.

Figure 7.

Plasma corticosterone: Bar graphs show mean plasma corticosterone levels ± SEMs. Female (A) and male (B) mice in the stress condition had higher corticosterone levels than home cage mice. TAK-242 administration decreased corticosterone levels in response to stress selectively in females (A) as TAK-242-treated males showed similar corticosterone levels as vehicle-treated males. Due to variability in select groups, box plots of corticosterone data were included for females (C) and males (D) which show the range of individual values, median (solid line), and mean (dashed line). Symbols indicate a significant difference from stress group (^) and TAK-242-treated female mice in the stress group (*).

4. Discussion

In the absence of an inflammatory stimulus, TLR4 activation plays a physiological role in cognitive function and adult hippocampal neurogenesis, as TLR4 deficient mice show enhanced spatial memory, cell proliferation, and neurogenesis (Connolly et al., 2020; Okun et al., 2012; Potter et al., 2019; Rolls et al., 2007). The present study evaluated whether pharmacological inhibition of TLR4 would produce comparable effects on hippocampal cell proliferation and spatial memory. Moreover, the potential influence of age and sex were assessed by comparing young and aged male and female mice. Our findings demonstrate that administration of the TLR4 antagonist, TAK-242, selectively enhanced performance in the water maze in young females. This apparent improvement in spatial learning and memory coincided with reductions in emotionality. TAK-242 administration selectively attenuated hippocampal IL-1β expression across all groups. Assessment of hippocampal cell proliferation demonstrated minimal effects of TLR4 inhibition, as TAK-242- and vehicle-treated mice showed similar estimated numbers of Ki67 positive cells in the hippocampal subdivisions and throughout the entire granular cell layer. Aging resulted in the expected decrease in Ki67 positive cells relative to young mice. Collectively, these data extend our understanding of the physiological function(s) of TLR4 by supporting a modulatory role for TLR4 in emotionality as well as differential responsivity of males and females to TLR4 inhibition.

Activation of TLR4 through injury or TLR4 agonist (e.g., LPS) negatively affects cognitive function through the induction of a neuroinflammatory response (Kohman et al., 2013; Kranjac et al., 2012; Pugh et al., 1998; Sparkman et al., 2005a). Even in the healthy brain, TLR4 may negatively affect select cognitive processes as evidenced by cognitive benefits seen in TLR4 deficient mice (Okun et al., 2012; Potter et al., 2019). While a valuable model, the absence of TLR4 from birth in the global KOs can result in compensatory adaptations as well as alterations in developmental processes (Nelson, 1997). Thus, the present study evaluated whether water maze performance was affected at select stages of adulthood following pharmacological inhibition of TLR4. Assessment of spatial learning over five days of training, demonstrated that TLR4 inhibition enhanced acquisition in young females. Treatment with TAK-242 reduced young females’ cumulative distance to the hidden platform relative to vehicle-treated young females. In addition, TLR4 inhibition enhanced spatial memory in young females as measured by proximity to the original platform location during the probe trial. The observed differences were not the result of altered motor function or motivation, as swim speed and performance in the visible platform task were unaffected by TAK-242 administration. Whether these cognitive benefits result from TAK-242 acting within the periphery or central nervous system cannot be determined from the present data, as TAK-242 crosses the BBB and may be acting in both areas (Wang et al., 2013). Future work is needed to dissociate the whether the beneficial effects of TLR4 inhibition relate to a central or peripheral site of action.

While mice administered a TLR4 antagonist and TLR4 KO mice both show enhanced performance in the water maze the effects vary across age and sex. Prior evidence indicates that young TLR4 KO males, but not females, show enhanced spatial memory (Okun et al., 2012; Potter et al., 2019). Whereas the present data indicate that the pro-cognitive effects of the TLR4 antagonist are limited to young females. Potential contributing factors are the timing and length of TLR4 suppression between these approaches, as the TLR4 KO mice lack the receptor throughout their life while the TLR4 antagonist was only given during adulthood. Okun et al. (2012) concluded that the spatial learning enhancements seen in male TLR4 KO mice related to developmental alterations, potentially in neural circuits, given that central administration of a TLR4 antagonist had no effect on memory in adult males. TLR4 shows differential expression during pregnancy and has been implicated in modulating fetal growth as well as inducing labor (Firmal et al., 2020). Further, prior work has shown that blocking TLR4 during embryonic development has differential effects on males and females (Chin et al., 2019). Specifically, administration of a TLR4 antagonist during late gestation selectively increased growth in male offspring as assessed by muscle and fat mass at 20 weeks of age, but no differences were found in females (Chin et al., 2019). Similarly, adult male TLR4 KOs have been reported to show higher body weights compared to WT mice (Okun et al., 2014). Collectively, these data indicate that the absence of TLR4 differentially affects males and females during development and these developmental alterations may account, in part, for the differences between TLR4 KOs and mice administered a TLR4 antagonist in adulthood.

In accordance with prior results, pharmacological inhibition of TLR4 had no effect on spatial learning or memory in young males (Okun et al., 2012). Okun et al. (2012) centrally administered a TLR4 antagonist into the lateral ventricles via an osmotic pump to young male mice and reported no difference in spatial memory. The present data, in agreement, demonstrate that systemic administration of TAK-242 did not influence water maze performance in male mice. Our data in combination with Okun et al. (2012) indicate that TLR4 has sexually dimorphic effects on spatial memory performance. Prior work has reported sex-related differences in the response to TLR4 activation (Loram et al., 2012; Roberts et al., 2013). For instance, males show greater deficits following early-life exposure to TLR4 agonist, whereas females are more resilient to TLR4 mediated insults, potentially indicating differential importance of the TLR4 receptor in development across males and females (Chin et al., 2019; Nway et al., 2017). Microglia and astrocytes isolated from neonatal male and female rats show differential expression of the inflammatory cytokine IL-1β in response to LPS exposure (Loram et al., 2012). Further, Doyle et al. (2017) found that differences in activation of TLR4 on microglia mediate the sex-related differences in the analgesic effects of morphine. We reported that while TLR4 deficiency enhanced hippocampal neurogenesis in both sexes, the increase was greater in females and was region-specific in males (Connolly et al., 2020). The sexually dimorphic response to TLR4 inhibition may relate to sex hormones. Prior work has shown that estrogens have immunomodulatory effects, for instance estrogens modulate surface expression of TLR4 on macrophages and microglia (Rettew et al., 2009; Zhang et al., 2018). Further, recent work by Bonet et al. (2021) demonstrated that estrogens mediated the sex-dependent response to a TLR4 antagonist in a pain model. Specifically, males and females showed the same response to a TLR4 antagonist if females were ovariectomized, however differential effects were seen between the sexes if females were gonad-intact or if 17β-estradiol replacement was given (Bonet et al., 2021). In addition, differences in TLR4 activation of downstream signaling pathways in males and females may contribute to the differential response across the sexes (Fields et al., 2018). Dissociation of the specific contribution of these potential mechanisms to the host of sex differences reported following TLR4 inhibition and activation is required. The present data extend the existence of sexually dimorphic effects of TLR4 to cognitive function, as pharmacological inhibition of TLR4 selectively enhanced spatial learning and memory in females.

The enhanced water maze performance in young females following TLR4 inhibition may relate to reductions in stress responsivity. An abundance of research demonstrates that elevated anxiety or stress reactivity impairs learning and memory (Maloney et al., 2014; McEwen and Sapolsky, 1995). The water maze may be particularly sensitive to deficits related to altered emotionality, as the water maze is a more stressful task when examining plasma corticosterone levels relative to other spatial learning tests (Harrison et al., 2009). The stress response to the water maze varies in males and females. While both males and females show elevated plasma corticosterone in response to water maze training, corticosterone levels are higher in females and inversely correlate with their performance as higher levels of corticosterone were associated with poorer cognitive ability (Beiko et al., 2004). Further, adrenalectomy selectively enhances water maze performance in females but not males, indicating that the stress response has greater influence on females’ performance in the water maze (Beiko et al., 2004). Our data indicate an increase in anxiety-like behavior in females, as vehicle-treated females spent a greater proportion of time engaged in thigmotaxic behavior during water maze training compared to vehicle-treated males. Thigmotaxic behavior has been reported to positively correlate with anxiety-like behavior in the elevated plus maze (EPM), as rats that spent a lower proportion of time in the open arms of the EPM spent a higher proportion of time engaged in thigmotaxic behavior in a water maze (Herrero et al., 2006). Pretreatment with the TLR4 antagonist decreased thigmotaxis in young females, indicating reduced anxiety-like behavior. Further, in separate groups of young and aged male and female mice, reactivity to acute restraint stress was evaluated by measuring plasma corticosterone levels in vehicle- and TAK-242-treated mice. Administration of TAK-242 attenuated stress-induced plasma corticosterone levels in females. In contrast, when evaluated in young and aged males TAK-242 administration had no effects on the corticosterone response to an acute stressor as TAK-242- and vehicle-treated males had similar corticosterone levels. These data show that TLR4 inhibition reduced the stress-induced increase in corticosterone levels selectively in female mice. TLR4 inhibition may have attenuated the corticosterone response to water maze training in young females, which has been reported to enhance their spatial learning and memory capabilities (Beiko et al., 2004). Prior work indicates that TLR4 participates in the response to stress, as TLR4 deficient mice show an attenuated neuroinflammatory response to stress exposure and are less susceptible to develop learned helplessness (Cheng et al., 2016; Garate et al., 2013; Medina-Rodriguez et al., 2020). Similarly, administration of TAK-242 reduces stress-induced neuroinflammation and depression-like behaviors (Garate et al., 2014; Zhang et al., 2020). The engagement of TLR4 in response to stress may result from release of danger-associated molecular patterns (DAMPs) and/or translocation of bacteria from the intestines (Cheng et al., 2016; Garate et al., 2014). Even in the absence of a stressor, TLR4 deficient mice show altered anxiety under some circumstances, though these effects appear to be dependent on the animal’s age, sex, and the testing parameters (Femenia et al., 2017; Li et al., 2016; Okun et al., 2012; Potter et al., 2019). More work is needed to dissociate the modulatory effects of TLR4 on emotionality under basal and stress conditions. Collectively, these data indicate that TLR4 inhibition may enhance water maze performance in young females potentially by reducing stress reactivity.

While aged mice showed the expected deficits in spatial learning and memory compared to young adults, administration of the TLR4 antagonist did not produce any recovery in cognitive function in aged males or females. Normal aging is associated with an increase in the inflammatory tone of the brain, which contributes to age-related cognitive decline (Braida et al., 2004; Gemma et al., 2005; Kohman et al., 2013). To determine whether TLR4 inhibition affected the low-grade chronic neuroinflammatory state in the aged brain, hippocampal and brain cytokine levels were evaluated. Findings showed that aged mice had higher central levels of IL-6 as well as increased hippocampal expression of IL-1β, IL-6, and TNFα compared to young mice in agreement with prior reports (Potter et al., 2019; Ye and Johnson, 1999). Administration of TAK-242 had selective effects on IL-1β expression in the hippocampus, as TAK-242-treated mice, regardless of age or sex, showed reduced IL-1β compared to vehicle-treated mice. No differences were seen in hippocampal expression of IL-6 or TNFα following TAK-242 treatment. Similarly, TAK-242 had no effect on brain IL-6 levels in the aged or young mice. TLRs and IL-1 receptors engage common adaptor proteins when driving an inflammatory response, which may contribute to the selective changes in hippocampal IL-1β (Loiarro et al., 2010). We have previously found that TLR4 KO mice show select changes in IL-1β and related signaling molecules (Potter et al., 2019). IL-1β during an inflammatory response as well as under physiological conditions can influence learning and memory processes, as elevating or depleting IL-1β can impair cognitive function indicating an optimal level of IL-1β that supports memory processes (Donzis and Tronson, 2014; Yirmiya and Goshen, 2011). The observed reductions in IL-1β are unlikely to drive the cognitive benefits seen in the young female mice, as the reductions in hippocampal IL-1β were seen across all groups that did not show alterations in water maze performance. These data indicate that transient inhibition of TLR4 does not globally attenuate the age-related increase in proinflammatory cytokines, as IL-6 and TNFα levels were unaffected.

Previous work with aged TLR4 deficient mice showed that aged females, but not aged males, had enhanced spatial memory relative to aged WT females in a water maze (Potter et al., 2019). This memory enhancement was accompanied by a reduction in thigmotaxic behavior as well as decreased anxiety-like behavior in an open field, indicating that TLR4 deficiency attenuated the age-related increase in anxiety (Potter et al., 2019). In the present study, TLR4 inhibition attenuated stress-induced increases in plasma corticosterone in aged females. In the water maze, TAK-242-treated aged females showed similar patterns of behavioral change as the young females in some parameters, but the differences were not significant compared to vehicle-treated aged females. Given that normal aging increases expression of TLR4 in the brain (Berchtold et al., 2008; Letiembre et al., 2007) accommodations in the administration protocol may need to be made for the aged. The present study administered TAK-242 for 2 weeks however, in a diet-induced obesity model TAK-242 was given for 12 weeks and was found to attenuate neuroinflammation (Moser et al., 2018). Thus, future investigations should determine whether increasing the treatment duration and/or dose of the TLR4 antagonist would produce beneficial effects on anxiety-like behavior, neuroinflammation, and cognitive function in aged females. The current data indicate that transient inhibition of TLR4 did not alter cognitive function in aged mice and demonstrates that pharmacological inhibition of TLR4 parallels some changes in emotionality observed in aged TLR4 deficient mice though to a lesser degree.

TLR4 in its basal and activated state negatively affects hippocampal neurogenesis in the adult brain (Bastos et al., 2008; Connolly et al., 2020; Ekdahl et al., 2003; Littlefield et al., 2015; Monje et al., 2003; Rolls et al., 2007). TLR4 and other TLRs are expressed on neural progenitor cells (NPCs) and modulate their proliferation under physiological conditions, as inhibition of TLR4 increased proliferation as assessed by sphere formation (Rolls et al., 2007). Further, TLR4 deficiency in young mice enhances cell proliferation and neuronal differentiation (Connolly et al., 2020; Rolls et al., 2007). When evaluated in the dorsal and ventral subdivisions of the hippocampus, TLR4 deficiency enhanced cell proliferation in both the dorsal and ventral hippocampus for females, but only in the ventral hippocampus for males. Further, female TLR4 deficient mice showed selective increases in new neurons in the dorsal hippocampus relative to WT females, but both males and females showed enhanced neurogenesis in the ventral hippocampus (Connolly et al., 2020). Given that TLR4 deficiency enhanced cell proliferation, the present study evaluated estimated numbers of Ki67 positive cells in the granular cell layer of the hippocampus in response to pharmacological inhibition of TLR4. Results showed that aged mice had reduced estimated numbers of Ki67 positive cells compared to young mice in accordance with previous finding (Heine et al., 2004; Potter et al., 2019). The age-related reduction in cell proliferation was apparent within the dorsal and ventral subdivisions and throughout the entire granular cell layer. Administration of the TLR4 antagonist had minimal effects on cell proliferation, as the estimated numbers of Ki67 positive cells were similar between the TAK-242- and vehicle-treated mice. This lack of an effect was seen in both the adult and aged mice. In the dorsal hippocampus, young TAK-242-treated females did show reduced proliferation compared to TAK-242-treated males, but they did not differ from vehicle-treated females. One limitation of the present study is the lack of assessment of neuronal differentiation. This was not assessed in the present study given the lack of an effect of TAK-242 on proliferation as well as prior results that showed mice given TAK-242 for 12 weeks showed no changes in doublecortin (DCX) positive cells compared to vehicle-treated mice (Moser et al., 2018). Prior work has demonstrated that TAK-242 administration protects against inflammation-associated changes in neurogenesis, supporting its utility to modulate neurogenesis under inflammatory conditions (Lei et al., 2016; Moser et al., 2018). However, no effects were observed on proliferation under basal conditions in either young or aged mice. While neurogenesis has been reported to contribute to learning and memory processes (Deng et al., 2010; Dupret et al., 2008; Kohman and Rhodes, 2013), the present data support an alternative mechanism for the cognitive benefits observed in the young females.

While the physiological function(s) of TLR4 have yet to be fully elucidated, the current data, in combination with prior reports, support a role for TLR4 in modulating emotionality. The present study is the first to demonstrate that pharmacological inhibition of TLR4 has beneficial effects on water maze performance in a sex-specific manner that coincide with reductions in stress reactivity. Further, findings demonstrate that the enhancements in spatial learning and memory may be restricted to a select stage of adulthood, as aged females did not show altered water maze performance in response to TLR4 inhibition. Both TAK-242-treated young and aged females, but not males, show attenuated corticosterone production in response to acute stress, indicating that TLR4 inhibition modulates the stress response in females. While TLR4 antagonism attenuated hippocampal IL-1β expression, these changes were not specific to the females that showed spatial memory improvements. The present findings establish that the cognitive benefits of TLR4 antagonism are sex-dependent and coincide with changes in emotionality.