The Peroxisomal Proliferator-Activated Receptor (PPAR) α Agonist, Fenofibrate, Prevents Fractionated Whole-Brain Irradiation-Induced Cognitive Impairment

Animals

Eighty 10–12-week-old young adult male Fischer 344 × Brown Norway (F344×BN) rats were obtained from Harlan Laboratories, Inc. (Indianapolis, IN) and housed in pairs on a 12:12 h light-dark schedule with free access to food and water. All animal handling and experiments were performed in strict accordance with the NIH Guide for Care and Use of Laboratory Animals as approved by the Wake Forest School of Medicine Institutional Animal Care and Use Committee. After an acclimation period of 2 weeks, rats were randomized to 4 experimental groups (n = 20 rats/group), Group 1: sham irradiated; Group 2: fWBI; Group 3: sham irradiated + 0.2% w/w of fenofibrate (Sigma-Aldrich, St. Louis, MO) in the diet; and Group 4: fWBI + 0.2% w/w of fenofibrate in the diet. Rats received fenofibrate beginning 3 days before the start of fWBI and continuously until the end of the experiment. All rats were weighed weekly to assess their overall health.

fWBI Procedures

A 40 Gy total dose of fWBI was delivered in 8 fractions of 5 Gy, twice/week, for 4 weeks as described previously (25). Briefly, irradiations were performed on lightly anesthetized [Ketamine (75 mg/kg)/xylazine (7 mg/kg)] rats in a 267 TBq (7,214 Ci) self-shielded ¹³⁷Cs irradiator using lead and Cerrobend devices to collimate the beam so that the whole brain, including the brain stem, was irradiated. The average dose rate to the midline of the brain was ~4 Gy/min; the eyes and body received ~15% and ~3% of the brain dose, respectively. Doses were delivered to opposite sides of the head on alternate days to ensure that each rat brain received the same midline dose. Sham-irradiated rats were also anesthetized twice weekly for 4 weeks and handled similarly to the irradiated rats.

NOR Test Procedures

The cognitive function of each rat was assessed 26 weeks after the completion of fWBI using a PRh-dependent version of the NOR task as previously described (27, 28). The NOR task is comprised of three phases: sample, delay and test phases. Before testing, each animal was habituated to the testing arena (an opaque rectangular container: 24 in × 18 in × 20 in) by allowing exploration of the empty arena for 5 min a day for 5 days. After habituation, the following trials were divided into a sample phase and a test phase, with a delay occurring between the two phases. During the sample phase, the rat was allowed to explore and become familiar with two identical objects (A₁ and A₂), placed in two adjacent corners of the testing arena, for 3 min. Following the sample phase, a 1 min delay was introduced during which the animal was removed from the arena and returned to its home cage. A test phase followed in which two objects (A₃ and B₁) were placed in the arena, and the rat was allowed to explore the arena for 3 min. Object A₃ was identical to A₁ and A_2, while B₁ was a novel object; activity of the rat was recorded using an automated tracking system (Ethovision, Nodulus, Leesburg, VA). The objects were secured to the arena floor with Velcro patches so that the subject could not displace them during exploration. Recognition memory was measured as the time spent exploring the novel object compared to the time spent exploring the familiar object. The position (left or right) of the novel object in the test phase was balanced between sessions to avoid any spatial preference.

The basic measurement was the time the rat spent exploring an object, defined as placing his nose within ≤2 cm of the object and actively exploring it. The following parameters were obtained: E₁, the total time spent exploring the identical objects A₁ and A₂ in the sample phase (3 min); E₂, the total time spent exploring object A₃ and the novel object B₁ in the test phase (3 min); and D₁, the index of discrimination defined as the difference in time spent exploring objects A₃ and B₁ in the test phase (i.e., B₁ – A₃). The discrimination ratio (D₁/E₂), a measure of the rat’s recognition memory, was calculated and the average discrimination ratio for each treatment group compared.

MWM Test Procedures

Spatial learning, reference memory and spatial reversal learning were evaluated 27–29 weeks after completion of fWBI using the MWM. As previously described (34), rats were placed in a circular white plastic tank filled with opaque water and surrounded by dark geometric cues affixed to white curtains. The tank was divided into 4 imaginary quadrants with an escape platform 2 cm under the water surface in the middle of one quadrant.

During the standard version of the MWM, spatial learning and reference memory were measured. The escape platform was placed in quadrant 4, the rats were introduced into the quadrants in a systematically random pattern, and each rat’s performance was recorded using an automated tracking system (Ethovision). For each block (Supplementary Table S1; http://dx.doi.org/10.1667/RR13202.1.S3) the first two days consisted of two training trials each day and the third day consisted of one training trial and one probe trial. During each training trial, the rat was allowed to search for 90 s to locate the platform and the total distance to the platform, the path length to the platform and the escape latency were measured. During the probe trial, the platform was lowered beyond reach, the rat was allowed to search for 30 s, and the mean distance and latency to the previous platform site was measured. This schedule was repeated for 4 blocks (Supplementary Table S1; http://dx.doi.org/10.1667/RR13202.1.S3). In the reversal version of the MWM, the platform location was moved to the opposite quadrant with the visual cues remaining in the same positions. This task required the rats to inhibit their previously learned response and remember a new location to find the escape platform.

Spatial working memory was also assessed using a DMTP procedure that started on the last day of the standard MWM (Supplementary Table S1; http://dx.doi.org/10.1667/RR13202.1.S3). The reversal day of the standard MWM was the first day of the DMTP procedure that lasted for a total of 4 days. Each DMTP testing day consisted of 4 trials, a sample trial (trial 1) and three match trials (trial 2–4). The escape platform in the sample phase each day was located 2 cm under the water surface, in a unique location (Supplementary Fig. S2; http://dx.doi.org/10.1667/RR13202.1.S2) and remained in the same location for the remaining match trials for that day. The delays between daily trials were 20 min, 15 s and 15 s.

Strategy preference was assessed 24 h after the last day of DMTP testing. A visible platform was placed in the opposite quadrant to where the hidden platform was previously located. The rats were given two 60 s strategy probe trials. Start locations were on either side of the tank, equidistant from the visible cue platform and the prior hidden platform location. A “place strategy” was recorded if the rat crossed within 5 cm of the prior hidden platform location before escaping to the visible platform. A “cue strategy” was recorded if the rat did not cross the prior hidden platform before swimming to the visible platform. Data were scored by a blinded observer.

Visual acuity was assessed after the cue strategy task was performed. The task consisted of 4 consecutive trials for each rat. A visible platform was placed in each quadrant for each trial excluding the quadrant where the rat entered the maze. The rats were given 60 s to locate the platform. Locomotor behavior was assessed using swim velocity during the visual acuity trials in order to reduce confounding factors of trial acquisition on motor behavior. Data were scored by an automated tracking system (Ethovision, Nodulus, Leesburg, VA).

Tissue Processing

Following completion of the cognitive function testing at 29 weeks post-fWBI, rats were euthanized by decapitation. The brains were removed rapidly and hemisected. The right hemisphere was flash frozen in liquid nitrogen, the left hemisphere was immersion fixed in phosphate-buffered 4% paraformaldehyde for 24 h. The left hemispheres were then cryoprotected for 24 h in 10, 20 and 30% sucrose, frozen in embedding medium, cryosectioned at a thickness of 40 μm in the coronal plane, placed in an antifreeze solution (1:1:2 ethylene glycol, glycerol and 0.1 M of phosphate-buffered saline) and stored at −20°C.

Immunohistochemistry

Based on previously published studies, immunohistochemistry was performed on the left hemispheres of 4 brains selected at random from each experimental group (19, 26). For each brain, the first section was randomly selected from the first 12 sections through the DG (bregma −1.8 to bregma −6.8). Subsequently, every 12th section was then washed in tris-buffered saline (TBS; pH 7.4) to remove the cryoprotectant. For immunohistochemical staining, endogenous per-oxidase activity was reduced by 30 min incubation in 0.6% hydrogen peroxide in TBS. To label immature neurons, sections were incubated in blocking solution (5% normal serum + 0.2% Triton X-100 in TBS) and then overnight at 4°C with primary antibody. The primary antibodies used were a goat polyclonal anti-rat doublecortin (DCX⁺) antibody [sc-8066 (C18), Santa Cruz Biotechnology, Inc., Santa Cruz, CA; 1 μg/mL; 1:200] and mouse monoclonal anti-rat CD68 (ED1 clone), which labels activated macrophages/microglia (MCA341R, AbD Serotec, Raleigh, NC; 1:400). The primary antibody was detected using biotinylated secondary antibodies [Dcx: horse anti-goat, ED1: horse anti-mouse (Santa Cruz Biotechnology, Inc.; 1:300)] and visualized using peroxidase-conjugated avidin-biotin complex (ABC Elite kit) with nickel enhanced diaminobenzidine (DAB) as substrate (Vector Laboratories, Burlingame, CA). Sections were counterstained with the nuclear binding dye, 4^′, 6-diamidino-2-phenylindole dihydrochloride (DAPI, Sigma Aldrich, St. Louis, MO), to facilitate recognition of anatomical landmarks.

Stereological Analyses of Tissue Sections

The estimated total number of Dcx⁺ cells in the GCL and SGZ of the DG and ED1⁺ cells in the DG/hilus were estimated by the optical fractionator method using Stereo Investigator software (Microbright-field, Inc., Colchester, VT). The counting parameters for Dcx⁺ cells in the GCL/SGZ were: sampling grid size (100 × 100 μm), counting frame size (100 × 100 μm), disector height (10 μm) and guard-zone thickness (2 μm). The counting parameters for ED1⁺ cells in the DG/hilus were: sampling grid size (150 × 150 μm), counting frame size (75 × 75 μm), disector height (10 μm) and guard-zone thickness (5 μm). The precision of the stereological counts was measured using the coefficient of error (CE) by the Gundersen-Jensen CE estimator. The variance introduced by the stereological analysis was never more than 50% of the observed group variance (33, 34).

Statistical Analysis

All analyses were performed using GraphPad Prism (La Jolla, CA), SPSS v. 21 (Chicago, IL) and/or SAS (Cary, NC) software. P < 0.05 were considered significant.

Cognitive Studies

A two-way ANOVA was used to analyze the NOR. For all MWM data, the escape latency, distance to platform and path length, over the course of training and probe trials were analyzed using repeated measures ANOVA; variables were corrected for sphericity using Greenhouse-Geisser correction. Additionally, probe trial data were analyzed using a mixed effects model to determine if the rats exhibited a preference for the target quadrant (defined as percentage time (%) in the target quadrant > 25%), and to determine if there were any drug, radiation, probe or interaction effects. MWM visual acuity and swim velocity were analyzed using a two-way ANOVA with Bonferroni corrected pair-wise post-test comparisons. Strategy selection was assessed using χ² analysis.

Immunohistochemical Studies

Data are presented as the mean ± SEM and include CE calculations. A two-way ANOVA was used to determine the effect of fenofibrate, fWBI or any interaction effects. Bonferroni post-tests were used for pair-wise comparisons.

NOR Testing Results

Consistent with previously published results using the PRh-dependent NOR task with this rat model (10, 20), fWBI produced a significant radiation effect at 26 weeks post-fWBI [Fig. 1A; fWBI effect (F_1.68 = 8.64, P < 0.005)]. There was also a significant dietary fenofibrate effect (F_1.68 = 4.92, P < 0.030), but no significant interaction between dietary fenofibrate and fWBI (F_1.68 = 0.46, P > 0.5). Administration of dietary fenofibrate before, during and after fWBI prevented this PRh-dependent cognitive impairment without any significant effect on the cognitive function of the sham-irradiated controls (Fig. 1A). Data analysis using two-way ANOVA showed that there was no significant difference in the total time spent exploring both objects in the test phase (E2) between any of the groups (Fig. 1B).

Standard MWM Testing Results

No overall deficits in spatial learning, reference memory or spatial reversal learning were measured 27–29 weeks post-fWBI using the standard MWM (Figs. 2–4).

Training Trials

Probe Trials

Reversal Trials

During the reversal trial of the MWM, there were no group differences in escape latency (Fig. 4A) or total distance to the platform (Fig. 4B).

DMTP Testing Results

There were no significant differences in spatial working memory between the groups using the DMTP version of the MWM (Figs. 5 and 6). Escape latency decreased over the successive training days and trials (Fig. 5A; day effect (F_3,213 = 11.0, P < 0.0001) and trial effect (F_3,213 = 100.1, P < 0.001), indicating that all the rats were able to learn the task. There were no group differences in escape latency (Fig. 5A) or total distance to the platform (Fig. 5B). The percentage of time the rat spent where the platform was the previous day is an indication of how much the previous MWM training interfered with learning the new platform location; there was no significant difference among the groups (Fig. 6A). Additionally, the rats exhibited no difference in savings, (i.e., the time difference between escape latency of trial 1–trial 2), indicating that all groups were equally able to learn the new platform location regardless of fWBI or fenofibrate treatment (Fig. 6B).

Cue vs. Place Strategy Testing Results

There was no significant difference among the groups in the cue strategy task for test 1 (χ²; P < 0.91) or test 2 (χ²; P < 0.701, Table 1). More rats (73%) tended to use a spatial, hippocampal-based strategy rather than a cue, striatal-based strategy during test 1. Significant preference for the spatial strategy was seen in the sham irradiated (P < 0.02) and fWBI + fenofibrate groups (P < 0.03). A marginal preference for the spatial strategy was seen in the fWBI group (P < 0.09), but no preference was observed in the sham irradiated + fenofibrate group (P < 0.2). For test 2, 44% of rats performed using a spatial strategy however no significant preference was observed in any group (sham: P < 0.7, sham + fenofibrate: P < 0.2, fWBI: P < 0.9, fWBI + fenofibrate: P < 0.9).

Immunohistochemistry Results

Immature neurons (DCX⁺ cells) were quantified within the GCL and SGZ of the DG as an indirect marker of neurogenesis, since not all newborn neurons are incorporated into functional networks (36). There was a highly significant reduction in the total number of DCX⁺ cells at 29 weeks after fWBI (F_1.12 = 18.8, P < 0.001; Fig. 7). Continuous administration of dietary fenofibrate did not prevent this fWBI-induced reduction in the total number of DCX⁺ cells. CE values were as follows: sham =0.05, fWBI = 0.27, sham + fenofibrate = 0.06, fWBI + fenofibrate = 0.19.

The total number of activated microglia (ED1⁺ cells) were quantified within the DG/hilus as an indicator of neuro-inflammation. There was a significant increase in the total number of ED1⁺ cells at 29 weeks after fWBI (F_1.12 = 14.5, P < 0.003; Fig. 8). Continuous administration of dietary fenofibrate did not prevent this fWBI-induced increase. CE values were as follows: sham =0.15, fWBI = 0.09, sham + fenofibrate = 0.15, fWBI + fenofibrate = 0.07.

Body Weight

There was a significant decrease in body weight with radiation (F_1,69 = 22.2, P < 0.001), drug (F_1,69 = 82.8, P < 0.001) and interaction between radiation and drug (F_1,69 = 8.7, P < 0.01). Bonferroni corrected comparisons to explore these main effects indicated that sham-irradiated animals were significantly heavier than all other treatment groups (P < 0.001), while fWBI were significantly heavier than fWBI + fenofibrate rats (P < 0.01). Sham + fenofibrate rats did not significantly differ in weight from either fWBI (P < 0.2) or fWBI + fenofibrate rats (P < 0.5). There were significant effects of week (F_2.7,189.5 = 2272.0, P < 0.001), week by radiation (F_2.7,189.5 = 25.9, P < 0.001), week by drug (F_2.7,189.5 = 97.3, P < 0.001) and week by radiation by drug (F_2.7,189.5 = 13.5, P < 0.001; Supplementary Fig. S1; http://dx.doi.org/10.1667/RR13202.1.S1).

Visual Acuity and Locomotor Behavior

The average swimming distance to the visible platform over 4 trials was unaffected by fWBI or fenofibrate treatment (Supplementary Fig. S2A; http://dx.doi.org/10.1667/RR13202.1.S2). Swim velocity during visual acuity trials was assessed as an indicator of locomotor behavior. Swim velocity was significantly slower with fenofibrate (drug: F_1,68 = 11.2, P < 0.01) though not irradiation (F_1,68 = 3.0, P < 0.1). The drug by radiation interaction was significant (F_1,68 = 6.3, P < 0.02) with post-hoc testing revealing that the fWBI + fenofibrate group was significantly slower than the sham-irradiated group (Bonferroni, P < 0.01; Supplementary Fig. S2B; http://dx.doi.org/10.1667/RR13202.1.S2). All other post-hoc comparisons were not significant.