Neuroprotective effects of bee venom phospholipase A2 in the 3xTg AD mouse model of Alzheimer’s disease

bvPLA2 (Sigma-Aldrich, St. Louis, MO, USA) was dissolved in sterile phosphate-buffered saline (PBS) and kept at −20 °C until use. The 3xTg AD mice harboring a mutant APP (KM670/671NL), which is a human mutant PS1 (M146V) knock-in, and tau (P301L) transgenes (B6;129-Psen1 ^tm1Mpm Tg(APPSwe, tauP301L)1Lfa/J) [15] were purchased from Jackson Laboratory (Bar Harbor, ME, USA). Non-transgenic littermates were used as wild type (WT) controls. The animal procedures were approved by the University of Kyung Hee Institutional Animal Care and Usage Committee (KHUASP(SE)-13-015) and were in accordance with the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. All animals were maintained in a pathogen-free environment on a 12-h light/dark cycle and had access to food and water ad libitum. The mice were randomly assigned to five groups as follows: (1) a PBS-treated WT group (WT, n = 6); (2) a PBS-treated 3xTg group (3xTg, n = 5); (3) a 0.6 mg/kg donepezil-treated 3xTg group (3xTg+donepezil, n = 6); (4) a 0.2 mg/kg bvPLA2-treated 3xTg group (3xTg+bvPLA2 0.2, n = 6); and (5) a 1 mg/kg bvPLA2-treated 3xTg group (3xTg+bvPLA2 1, n = 4). Donepezil was used as a positive control and was purchased from Sigma (Sigma, St. Louis, MO, USA). Beginning at 3 months of age, PBS, bvPLA2, or donepezil was intraperitoneally administered for 3 months once per week until they were 6 months old. For the Treg depletion, anti-mouse CD25 rat IgG (anti-CD25, clone PC61) antibodies were generated in-house from hybridomas obtained from the American Type Culture Collection (Manassas, VA, USA). The mice received doses of 1 mg/kg of anti-CD25 rat IgG or normal anti-rat IgG1 once per week for 3 months until they were 6 months old. The efficacy of the CD4⁺ CD25⁺ Foxp3⁺ Treg cell depletion was determined by flow cytometry (FACS, FACSCalibur, BD Biosciences, San Jose, CA, USA) using anti-CD25-PE, anti-CD4-FITC, and anti-Foxp3-Cy5 PE (eBiosciences).

A modified version of the water maze procedure reported by Morris was used [25]. The water maze was a circular pool of 90 cm in diameter that was constructed of fiberglass. The pool contained water that was maintained at a temperature of 22 ± 2 °C and contained 1 kg of powdered skim milk to make the water opaque. During the water maze testing, a platform (6 cm in diameter) was fixed at 1 cm below the surface of the water at an identical location within the pool. The pool was surrounded by different extra-maze cues. Each trial was initiated at one of the different starting positions, and the swimming path of each mouse was recorded with a video camera connected to a video recorder and a tracking device (S-MART, Pan-Lab, Spain). All mice were subjected to four trials per day at intervals of 15 min for four consecutive days followed by 1 day of probe trials on the fifth day. The trials were considered to be completed when the mice found the hidden platform or the escape latency reached 60 s. For the probe trials, the platform was removed from the pool, and the mice were put in the pool for 60 s to search for the previous location of the platform. The memory retention was measured by the percentage of time spent searching for the platform in the training quadrant.

The mice were transcardially perfused with saline solution with 0.5 % sodium nitrate and heparin (10 U/ml) and then fixed with 4 % paraformaldehyde (PFA) in 0.1 M phosphate buffer (PB). Each brain was dissected from the skull, post-fixed overnight in 4 % PFA at 4 °C, stored in a 30 % sucrose solution at 4 °C until it sank, and frozen-sectioned on a sliding microtome into 30-μm-thick coronal sections. All of the sections were processed for immunohistochemical (IHC) staining as described previously [26]. The brain sections were washed with PBS and incubated overnight at room temperature with primary antibody. The next day, the brain sections were washed with PBS with 0.5 % bovine serum albumin (BSA), incubated with the biotinylated secondary antibody, and processed with an avidin-biotin complex kit (Vectastain ABC kit; Vector Laboratories, Burlingame, CA, USA). The bound anti-serum was visualized by incubation with 0.05 % diaminobenzidine-HCl (DAB) and 0.003 % hydrogen peroxide in 0.1 M PB. The DAB reaction was stopped by washing the slide section with 0.1 M PB. The primary antibodies that were used were directed against Aβ (1:100; Abcam, Cambridge, MA, USA), CD11b (1:200; Serotec, Oxford, UK), and CD4 (1:200; Serotec). The tissue sections were then mounted on gelatin-coated slides and observed under a bright-field microscope (Nikon, Tokyo, Japan).

After the behavioral tests, single-cell suspensions of splenocytes were stained with fluorescently tagged CD4, CD25, and Foxp3 (eBiosciences, San Diego, CA, USA) in accordance with the manufacturer’s instructions. All FACS data were acquired using a FACSCalibur flow cytometer (BD Biosciences) and analyzed using CellQuest Pro (BD Biosciences).

Coronal sections of the hippocampus, starting rostrally from anteroposterior −2.1 mm and continuing to anteroposterior −4.5 mm relative to bregma, were collected. The numbers of CD11b-positive microglia/macrophages and CD4⁺ T cells in the hippocampi stained with single-label immunohistochemistry were counted under a ×400 objective (250 μm) using a light microscope (Nikon).

The coronal sections of the hippocampus, starting rostrally from anteroposterior −2.1 mm and continuing to anteroposterior −4.5 mm relative to bregma, were examined at ×10 magnification using the IMAGE PRO PLUS system (version 4.0; Media Cybernetics, Silver Spring, MD, USA) on a computer attached to a light microscope (Zeiss Axioskop, Oberkochen, Germany) that was interfaced with a charge-coupled device video camera (Kodak Mega Plus model 1.4 I). To determine the density of the Aβ-immunoreactive, CD11b-positive microglia/macrophages, and CD4⁺ T cells staining in the hippocampus, a square frame of 500 × 500 μm was placed on the dorsal part of the hippocampus. A second square frame of 200 × 200 μm was placed in the region of the corpus callosum to measure the background value. To control for variations in background illumination, the average of the background density readings from the corpus callosum was subtracted from the average of the density readings from the hippocampus for each section.

All the mice were fasted for 12–15 h before the experiments to enhance the [F-18] FDG uptake in the brain [27]. Before the [F-18] FDG injection, all mice were put on a heating pad in a cage and warmed for at least 30 min. The temperatures of the cages were maintained at 30 °C throughout the uptake period in accordance with an optimized [F-18] FDG uptake protocol [28]. [F-18] FDG (500 μCi/100 g body weight) was injected through a tail vein, and the mice were anesthetized with 2 % isoflurane in 100 % oxygen (Forane solution; ChoongWae Pharma, Seoul, Korea). For the PET imaging, a Siemens Inveon PET scanner (Siemens Medical Solutions, Malvern, PA, USA) was used throughout the study. The transverse resolution was used was <1.8 mm at the center [29, 30]. Transmission PET data were acquired for 15 min using a Co-57 point source with an energy window of 120–125 keV. One mCi of [F-18] FDG was injected. Thirty min after injection of 1 mCi of [F-18] FDG, which period allowing adequate time for tracer uptake, emission PET data during 30 min were acquired within an energy window of 350-650 keV. The emission list-mode PET data were sorted into 3D sinograms and reconstructed using three DRP methods. The pixel size of the reconstructed images was 0.15 × 0.15 × 0.79 mm³. Attenuation and scatter corrections were performed for all of the datasets.

Voxel-based statistical analysis was performed to compare the cerebral glucose metabolisms of the datasets of the different groups. The procedure used for the statistical parametric mapping (SPM) analysis of the animal PET data has previously been described [31]. Briefly, for efficient spatial normalization, only the brain region was extracted. A study-specific template was then constructed using all of the datasets. The PET data were spatially normalized onto a mouse brain template and smoothed using a Gaussian kernel. Count normalization was performed. A voxel-wise t test between the groups’ datasets was performed using the Statistical Parametric Mapping 5 program (P < 0.05, K > 50) (Table 1).

The body weights of the wild type (WT), donepezil-treated, bvPLA2-treated, and vehicle-treated 3xTg AD mice were measured every week during the 3-month treatment (Fig. 1a). There were no obvious effects of bvPLA2-treatment (3 to 6 months of age) on the body weights compared to the vehicle-treated 3xTg AD mice. In contrast, the donepezil treatment group exhibited significantly decreased body weight compared to the vehicle-treated 3xTg AD group.

The latencies to escape onto the hidden platform during the acquisition trials of the water maze were recorded. The escape latencies differed between the groups when the results were averaged over all sessions. The latencies to find the hidden platform of the WT, donepezil, and bvPLA2 (0.2 and 1 mg/kg) groups were significantly shorter than those of the 3xTg AD group (Fig. 1b).

The total times spent in the quadrant that previously contained the platform were used to evaluate the spatial performances of the mice the during retention trials. The results of the retention tests on the fifth day are depicted in Fig. 1c. The bvPLA2-treated group (0.2 mg/kg) exhibited a significant increase (P < 0.05) in retention time compared to the 3xTg AD group (Fig. 1c).

In 3xTg-AD mice, Aβ deposits are present in the hippocampus by 6 months of age [16, 17]. Quantifications of the Aβ burden revealed significantly increased burden in the hippocampal CA1 and CA3 regions of 3xTg AD group at the age of 6 months compared to the age-matched WT group (100 ± 4.72, P < 0.001, 100 ± 7.6, P < 0.001, respectively, Fig. 2c, d). Compared to the vehicle-treated 3xTg AD group, the Aβ burden in CA1 was dramatically decreased following bvPLA2 treatment (0.2 mg group 69.7 ± 5.84 %, P < 0.001 and 1 mg group 54.2 ± 4.79 %, P < 0.001, Fig. 2a, c). Similarly, donepezil treatment also decreased the Aβ burden in the CA1 region (60.9 ± 5.84, P < 0.001, Fig. 2). In contrast, only the CA3 regions of the bvPLA2-treated groups exhibited significant decreases in the amounts of Aβ burden (0.2 mg group 76.4 ± 5.42 %, P < 0.05 and 1 mg group 70.3 ± 3.72 %, P < 0.01, Fig. 2d), and there were no significant differences between the donepezil treatment group and the vehicle-treated 3xTg AD group.

In vivo analyses of the brain glucose metabolism of live mice were performed using positron emission tomography. FDG-PET imaging scans indicated differences in the cerebral glucose metabolism rates from the hippocampi to the prefrontal cortices between the WT and 3xTg-AD groups (Fig. 3a), the 3xTg-AD and 3xTg-AD+donepezil groups (Fig. 3b), and the 3xTg-AD and 3xTg-AD+bvPLA2 groups (Fig. 3c). The SPM analysis revealed that the glucose metabolism in the diencephalon of the WT mice was significantly increased compared to that of the 3xTg-AD mice (P < 0.05). The glucose metabolism of the donepezil treatment group in the cortex was significantly increased compared to that of the 3xTg-AD group (P < 0.05). The glucose metabolism of the bvPLA2-treated group in the diencephalon was significantly increased compared to that of the 3xTg-AD group (P < 0.05).

To determine whether the neuroprotective effects of bvPLA2 resulted from the inhibition of microglial activation in the hippocampus, immunostaining of brain sections anti-CD11b was performed (Fig. 4). In the WT group, only a few CD11b⁺ microglia/macrophages with resting morphologies (i.e., small cell bodies and thin processes) were observed in the hippocampi. In contrast, numerous CD11b⁺ microglia/macrophages with activated morphologies (i.e., larger cell bodies and thick processes) were apparent in the hippocampi of the 3xTg AD group. However, the majority of the CD11b⁺ microglia/macrophages in the bvPLA2-treated group had returned to a resting morphology in the hippocampi.

To determine whether bvPLA2 modulated the infiltration of CD4⁺ T cells into the hippocampi of the 3xTg-AD mice, we performed immunostaining of the hippocampi with an anti-CD4 antibody. The bvPLA2 treatment groups exhibited reduced numbers of infiltrating CD4⁺ T cells in the hippocampus compared to the 3xTg AD mice (Fig. 4).

To ascertain whether the increase in Tregs caused by bvPLA2 treatment was directly involved in the Aβ neuroprotective actions, the mice received PC61 anti-CD25 mAb to deplete Tregs or normal anti-rat IgG1 as an isotype control. PC61 anti-CD25 mAb is generally used to deplete of Tregs, and this depletion is mediated by the antibody-dependent cellular-phagocytosis of FcγRIII⁺ macrophages in vivo [32]. Consistent with our previous data [33], the injection of the PC61 anti-CD25 mAb resulted in the depletion of over 90 % of the Foxp3⁺ cells in the CD4⁺-gated splenocytes (data not shown). The brain sections were immunostained with an anti-Aβ antibody. Aβ burden did not differ following Treg depletion in the 3xTg AD mice, and the neuroprotective effects of bvPLA2 were abolished in the Treg-depleted 3xTg AD mice (Fig. 5). Neither PC61 anti-CD25 mAb nor normal anti-rat IgG alone had any effect on Aβ burden in the hippocampus (Fig. 5). Next, we performed Morris water maze to measure cognition dysfunctions. The results demonstrated that there were no significant differences in the latency between bvPLA2 and PBS group after Treg depletion by PC61 anti-CD25 mAb injections (Fig. 5). These results strongly indicate that the neuroprotective effects of bvPLA2 were closely associated with Tregs in the 3xTg AD mice.

To determine whether bvPLA2 possessed immune tolerance-enhancing activity that was mediated by Treg induction, T cell sub-populations in splenocytes were analyzed in splenocytes. The percentages of CD4⁺ CD25⁺ Foxp3⁺ T cells in the splenocytes were reduced in the 3xTg AD group compared to the WT group, while CD4⁺, CD8⁺, and total lymphocytes were not altered. Treatment with bvPLA2 significantly increased the Treg population without altering the other lymphocytes, which suggests that bvPLA2 might suppress neuroinflammatory responses via Treg induction in 3xTg AD mice (Fig. 5).