Increasing brain glucose metabolism by ligustrazine piperazine ameliorates cognitive deficits through PPARγ-dependent enhancement of mitophagy in APP/PS1 mice

Rosiglitazone (HY-17386, purity = 99.90%), ligustrazine (HY-N0264, purity = 99.91%), GW9662 (HY-16578, purity = 99.83%), and Mdivi-1 (HY-16578, purity = 99.73%) were supplied by MedChem Express (Shanghai, China). Bovine serum albumin (BSA), Triton X-100, isoflurane, and paraformaldehyde were purchased from Sigma-Aldrich (MO, USA). Dulbecco’s modified Eagle’s medium (DMEM), fetal bovine serum (FBS), penicillin, streptomycin, Lipofectamine 6000, RIPA buffer, phosphate-buffered saline (PBS), saline, and phospho-Parkin (Ser65) polyclonal antibody (PA5-114616), as well as human enzyme-linked immunosorbent assay (ELISA) kits for Aβ40 (KHB3481) and Aβ42 (KHB3544), were all obtained from Invitrogen (CA, USA). Anti-Aβ (ab201060), anti-PPARγ (ab178860), anti-PINK1 (ab23707), anti-LC3B (ab192890), anti-SQSTM1/p62 (ab109012), and anti-beta tubulin (ab6046) antibodies, as well as Alexa Fluor® 647-conjugated goat anti-rabbit IgG H&L (ab150083), HRP-conjugated goat anti-rabbit IgG H&L (ab6721), and HRP-conjugated goat anti-mouse IgG H&L (ab6789), were all purchased from Abcam (CA, USA). The bicinchoninic acid (BCA) kit and enhanced chemiluminescence (ECL) kit were purchased from Pierce Biotechnology (IL, USA). The 18-Fluoro-6-deoxyglucose (¹⁸F-FDG) was obtained from Union Hospital Affiliated to Tongji Medical College, Huazhong University of Science and Technology.

N-Monosubstituted piperazine (10 mmol) and 2-chloromethyl-3,5,6-trimethylpyrazine hydrochloride (10 mmol) were added to 70 mL of toluene, followed by addition of 40 mmol of NaHCO₃ and a catalytic amount of NaI. The mixture was heated and refluxed for 10 h. TLC showed that the reaction was complete. After filtering, the filter cake was washed 3 times with a small amount of toluene, combined with filtrate, vacuum distilled to obtain oil, fast column separated to produce a light yellow powder, and recrystallized with n-hexane to obtain white crystal 1-benzhydryl-3-((3,5-dimethylpyrazin-2-yl)methyl)hexahydropyrimidine (LPD).

PC12 cells were obtained from the Cell Resource Center of the Institute of Basic Medical Sciences, Peking Union Medical College and Chinese Academy of Medical Sciences (Beijing, China). PC12 cells were maintained in DMEM supplemented with 10% FBS, 100 U/mL penicillin, and 100 μg/mL streptomycin at 37 °C and 5% CO₂ in a humidified incubator (Thermo Scientific, Langenselbold, Germany).

To determine whether LPD is a novel PPARγ agonist, the PPRE-TK-luc vector (1 μg) and PPARγ expression plasmid (1 μg) were cotransfected with 20 ng of pRL-TK (Promega, WI, USA) into PC12 cells using Lipofectamine 6000 when the cells reached 80% confluence. To construct the PPARγ expression plasmid, human PPARγ (NCBI reference sequence: NM_001354666) was PCR amplified and then fused with the GV230 vector (Shanghai Genechem Technology Co., Ltd., Shanghai, China). To construct the PPRE luciferase reporter plasmid, human PERM1 (NCBI reference sequence: NM_001291366.2) was PCR amplified and subcloned into the GV238 luciferase reporter vector (Shanghai Genechem Technology Co., Ltd., Shanghai, China). pRL-TK was used to adjust for transfection efficiency. After 48 h of transfection, PC12 cells were treated with various concentrations of LPD (2.5, 5, 10, and 20 μM) or ligustrazine (2.5, 5, 10, and 20 μM) or rosiglitazone (40 μM) or cotreated with LPD (20 μM) and GW9662 (10 μM) for 12 h. Luciferase activity was measured using a dual-luciferase reporter assay system (Promega, WI, USA) with a Glomax 20/20 luminometer (Turner Designs, CA, USA). The luciferase activity was normalized to Renilla luciferase activity. Cell viability was determined using a Cell Counting Kit-8 (Dojindo Laboratories, Kumamoto, Japan), according to the manufacturer’s instructions.

We previously established stable APP695swe-transfected PC12 cells [26]. To generate APP/PS1 double-overexpressing cells, human PSEN1 cDNA (NCBI reference sequence: NM_000021.4) was amplified through PCR. Mutant human PSEN1 (PSEN1dE9) was constructed and subcloned into the GV208 vector (Shanghai Genechem Technology Co., Ltd., Shanghai, China). The PSEN1dE9 plasmid was cotransfected with the framework plasmid vector pHelper 1.0 and pHelper 2.0 into HEK293T cells to produce Lenti-PSEN1dE9 (Shanghai Genechem Technology Co., Ltd., Shanghai, China). APP695swe stably transfected PC12 cells were seeded on six-well plates and infected with Lenti-PSEN1dE9 when the cells reached 80% confluence. Puromycin (2 μg/mL) was added, and drug-resistant cells were collected after 2 weeks for single-cell cloning. Steadily transfected cells were maintained in puromycin at a final concentration of 1 μg/mL. Resistant clones were analyzed by laser confocal microscopy and western blotting to confirm the overexpression of APP and PSEN1. The APP/PS1 cells (1.5 × 10⁴ per well) were plated in 6-well plates and cultured for another 24 h. The APP/PS1cells were incubated with ligustrazine (20 μM) or GW9662 (10 μM) or different concentrations of LPD (2.5, 5, 10, and 20 μM) or cotreated with LPD (20 μM) and GW9662 (10 μM) for 24 h. The levels of Aβ40 and Aβ42 were detected using the respective ELISA kits. Cell viability was determined using the Cell Counting Kit-8, according to the manufacturer’s instructions.

Mitochondrial oxidative phosphorylation capacity was determined as the uncoupled oxygen consumption rate (OCR) using a Seahorse XF96 extracellular flux analyzer (Seahorse Biosciences, MA, USA). Briefly, the APP/PS1 cells (5 × 10⁴ per well) were plated in XF96 extracellular flux assay plates and cultured for another 24 h. The APP/PS1 cells were incubated with LPD (20 μM) or cotreated with LPD (20 μM) and GW9662 (10 μM) or cotreated with LPD (20 μM) and Mdivi-1 (20 μM). After 24 h, the medium was replaced with XF Assay Medium (Seahorse Bioscience MA, USA). After the cells were incubated in a CO₂-free incubator at 37 °C for 30 min, basal levels were measured with no additives. For OCR detection, oligomycin, FCCP, and rotenone/antimycin A were added at final concentrations of 1 μM, 0.3 μM, and 0.1 μM, respectively. Three separate measurements were taken after each of the above reagents was added. Triplicate experimental wells were examined, and the results were plotted using Seahorse software.

All animal protocols were approved by the ethics committee of Shenzhen Second People’s Hospital. The experiments were conducted in compliance with the Guide for the Care and Use of Laboratory Animals. All efforts were made to reduce the number of animals used and minimize animal suffering in the experiments. Six-month-old male Swedish mutant APP (APP695swe)/PS1 (PSEN1dE9) transgenic mice and age-matched male C57BL/6N wild-type (WT) mice were obtained from Beijing HFK Bioscience Co., Ltd. (Beijing, China). The animals were housed in a specific pathogen-free animal facility at a constant room temperature of 22 ± 1 °C and 50 ± 10% relative humidity with 12-h light/12-h dark cycles. Access to standard rodent chow and water was available ad libitum.

After adaptation for 7 days, 60 APP/PS1 mice were randomly assigned to 6 groups for dose selection: vehicle, rosiglitazone (10 mg/kg/day), GW9662 (5 mg/kg/day), LPD (5 mg/kg/day), LPD (10 mg/kg/day), and LPD (20 mg/kg/day). Twenty age-matched C57BL/6N WT mice were randomly assigned to vehicle and LPD (20 mg/kg/day) groups. Each group consisted of 10 mice. LPD, rosiglitazone, and GW9662 were prepared in 5% DMSO and 95% saline containing 20% SBE-β-CD and stored at 4 °C until use. The mice were intragastrically administered LPD, rosiglitazone, GW9662, or an equivalent volume of vehicle for 3 months.

In the following experiments, 60 APP/PS1 mice were randomly assigned into 3 groups: vehicle, LPD (10 mg/kg/day), and GW9662 (5 mg/kg/day) + LPD (10 mg/kg/day). Each group consisted of 20 mice. Twenty age-matched C57BL/6N WT mice were used as controls. Each mouse was intragastrically administered either vehicle, LPD, or GW9662 every day from the age of 6 months for a total period of 3 months.

The body weights of the mice were monitored weekly. The volumes of LPD, GW9662, and vehicle were adjusted according to the body weights of the mice. After completion of drug treatments, 10 mice were randomly selected from each group for the Morris water maze test, and the other mice (n = 10 per group) were subjected to micropositron emission tomography (microPET). All mice were sacrificed with an overdose of isoflurane anesthetic, and 14 mice were randomly selected from each group and perfused transcardially with saline using a syringe infusion pump at a 5-min/min rate for 5 min. The brain tissue was quickly removed following decapitation, and the hippocampus was then dissected and frozen for ELISA (n = 8 per group) and western blotting (n = 6 per group). The other mice (n = 6 per group) were perfused with 4% paraformaldehyde following saline perfusion, and brain tissue (n = 3 per group) was collected for immunofluorescence. The hippocampi (n = 3 per group) were then dissected and further processed with transmission electron analysis.

Brain glucose uptake was evaluated using ¹⁸F-FDG microPET imaging as described in our previous study [20]. After a 6-h fast, the body weight and blood glucose level of the mice (n = 10 per group) were measured. The mice received ¹⁸F-FDG (200 ± 10 μCi) from the tail vein. The mice were anesthetized at 60 min postinjection with 2% isoflurane using a Matrix VIP 3000 calibrated vaporizer (Midmark, OH, USA) and placed on a scanning bed. PET was performed for 10 min followed by a CT scan using a TransPET Discoverist 180 system (Raycan Technology Co., Ltd, Suzhou, China). Body temperature was maintained at 37 °C with a heating pad during anesthesia. PET image reconstruction was performed using the 3-dimensional ordered-subject expectation maximization method with a voxel size of 0.5×0.5×0.5 mm³. CT images were reconstructed using the FDK algorithm with a 256×256×256 matrix. Images were displayed with Carimas software (Turku PET Center, Turku, Finland). The mean standardized uptake value was calculated using the following formula: mean pixel value with the decay-corrected region of interest activity (μCi/kg).

The APP/PS1 cells (n = 6 per group) and the hippocampus of mice (n = 8 per group) were collected and lysed in RIPA buffer containing phosphatase inhibitor and protease inhibitor cocktail. The homogenates were centrifuged at 20,000 rpm for 10 min at 4 °C, and the supernatants were pooled for the analysis of soluble Aβ40 and Aβ42. To extract fibrillar and membrane-bound insoluble Aβ40 and Aβ42, the pellets were homogenized in 70% formic acid and centrifuged at 40,000 rpm for 10 min at 4 °C. The supernatants were neutralized with 1 M Tris-base and analyzed for insoluble Aβ40 and Aβ42.

Western blot analysis was conducted as described previously [21]. The APP/PS1 cells (1.5 × 10⁴ per well) were plated in 6-well plates and cultured for another 24 h. The APP/PS1 cells were incubated with LPD (20 μM) or cotreated with LPD (20 μM) and GW9662 (10 μM) or cotreated with LPD (20 μM) and Mdivi-1 (20 μM) for 24 h. The cells (n = 6 per group) or frozen hippocampal tissues from mice (n = 6 per group) were homogenized in cold RIPA buffer containing a protease inhibitor cocktail, phosphatase inhibitor cocktail, and phenylmethanesulfonylfluoride (Roche, IN, USA) and then centrifuged at 10,000 rpm at 4 °C for 10 min. The supernatants were collected, and the protein concentrations were determined using BCA kits. Equal amounts of protein were separated by electrophoresis in 10% sodium dodecyl sulfate-polyacrylamide gels and transferred to PVDF membranes (Bio-Rad, CA, USA). The membranes were blocked in 5% nonfat milk powder in Tris-buffered saline containing 0.1% Tween-20 (TBST) for 1 h and then incubated at 4 °C overnight with anti-PPARγ, anti-PINK1, anti-p-Parkin(ser65), anti-LC3B, anti-SQSTM1/p62, and anti-beta tubulin antibodies. After rinsing in TBST, the membranes were incubated with goat anti-rabbit IgG H&L (HRP) or goat anti-mouse IgG H&L (HRP) antibodies at room temperature for 1 h. The protein bands were visualized by a ChemiDoc Touch Imaging System (Bio-Rad, CA, USA) using ECL kits and quantified by Molecular Imager Image Lab software (Bio-Rad, CA, USA). All protein band densities were normalized relative to beta tubulin.

Data are presented as the means ± SD and were analyzed using IBM SPSS Statistics version 20 (SPSS Inc., IL, USA). The independent samples t test was used to compare data between two groups. Comparisons among three or more groups were conducted using a one-way analysis of variance (ANOVA) followed by Tukey’s post hoc test. The statistical significance of the genotype and treatment effects was assessed using two-way ANOVA followed by Tukey’s post hoc test. P < 0.05 was considered statistically significant.

As depicted in Fig. 1A, solvent V ethylacetate to V cyclohexane equals 1:3 for rapid column separation, and a light yellow powder was obtained in 56% yield, m. p. 121–122 °C. LPD was obtained by recrystallization with n-hexane (purity = 98%). IR (KBr), cm⁻¹: 2 809.86 (CH); 1 598.06 (C=C); 1 585.41 (C=N). ¹HNMR (CDCl3), δ: 7.41 (d, 4H, J = 7.75 Hz, Ar-H); 7.26 (t, 4H, J = 7.70 Hz, Ar-H); 7.16 (t, 2H, J = 7.46 Hz, Ar-H); 4.22 (1H, CH); 2.30–3.62 (m, 10H, CH₂); 2.56 (s, 3H, CH₃); 2.52 (s, 3H, CH₃); 2.47 (s, 3H, CH₃). ESI-MS: 387.5 (M+1).

In this study, a dual-luciferase reporter assay system was used to determine whether LPD is a novel PPARγ agonist. As depicted in Fig. 1B, with rosiglitazone as a positive control, both ligustrazine and LPD caused a concentration-dependent increase in PPRE-driven luciferase activity in PC12 cells (P < 0.01). The PPRE-driven luciferase activity dramatically increased to sevenfold after LPD (20 μM) treatment for 12 h (P < 0.01). Conversely, a statistically significant decrease in PPRE-driven luciferase activity was observed in PPRE-transfected PC12 cells cotreated with LPD and the selective PPARγ antagonist GW9662 (P < 0.01). However, the viability of PPRE-transfected PC12 cells was not affected by ligustrazine or LPD or GW9662 (P > 0.05, Fig. 1C).

As depicted in Fig. 1D, stable APP/PS1-overexpressing cell lines were established. Treatment with ligustrazine (20 μM) for 24 h significantly reduced the levels of Aβ40 and Aβ42 in APP/PS1 cells. Moreover, incubation of APP/PS1 cells with LPD resulted in a dose-dependent decrease in the levels of Aβ40 and Aβ42 (P < 0.01, Fig. 1E, F). In contrast, a significant increase in Aβ40 and Aβ42 levels was observed in APP/PS1 cells cotreated with LPD and GW9662 (P < 0.01). However, the viability of stable APP/PS1 overexpressing cell lines was not affected by ligustrazine or LPD or GW9662 (P > 0.05, Fig. 1G).

As depicted in Fig. 2A, there was no difference seen in the body weight of mice among treatments and genotypes (P > 0.05). The effects of LPD on the spatial learning and memory of APP/PS1 mice were evaluated using the Morris water maze. At the age of 9 months, the latency to locate the hidden platform for APP/PS1 mice was significantly longer than that for WT mice (P < 0.01, Fig. 2B). However, we found that intragastric administration of LPD for 3 months significantly reduced the escape latency of APP/PS1 mice (P < 0.01, Fig. 2B). In the probe test, APP/PS1 mice demonstrated a decreased time in the target platform quadrant compared with WT mice (P < 0.01, Fig. 2C, D). Treatment with LPD for 3 months significantly increased the target quadrant time (P < 0.01, Fig. 2C, D), using rosiglitazone as a positive control. The protective effect of LPD reached a maximum at a dose of 10 mg/kg. Therefore, treatment with LPD (10 mg/kg) for 3 months was selected for further experiments. Moreover, treatment with GW9662 (5 mg/kg) alone had no significant effect on the escape latency (P > 0.05, Fig. 2A) or the target quadrant time of APP/PS1 mice (P > 0.05, Fig. 2C, D). In addition, treatment with LPD (20 mg/kg) had no significant effect on the escape latency (P > 0.05, Fig. 2A) or the target quadrant time of WT mice (P > 0.05, Fig. 2C, D). However, cotreatment with GW9662 and LPD significantly increased the escape latency (P < 0.01, Fig. 2E) and decreased the target quadrant time compared with LPD treatment alone (P < 0.01, Fig. 2G, H). These results suggested that LPD could attenuate cognitive dysfunction in APP/PS1 mice in a PPARγ-dependent manner. No difference in swimming speed was identified in the probe test among treatments and genotypes (P > 0.05, Fig. 2F). To evaluate the effect of LPD treatment on pathomorphology of WT and APP/PS1 mice, the liver sections were stained with H&E. In livers of WT mice, the cells were complete, clear, and regular (Fig. 2I). However, in livers of APP/PS1 mice, the boundary of cells was not clear, and the vacuoles of lipid droplets were observed (Fig. 2I). In addition, there was no significant change in livers of WT and APP/PS1 mice treated with LPD at a dose of 20 mg/kg (Fig. 2I).

Immunofluorescence was used to assess the effects of LPD on the distribution and morphology of Aβ plaques in brain sections. We observed that the number of Aβ plaques in the hippocampus of APP/PS1 mice was significantly increased compared to that in WT mice (Fig. 3A). In contrast, the area occupied by the Aβ plaques was significantly reduced in the hippocampus from LPD-treated APP/PS1 mice (P < 0.01, Fig. 3A, B). As shown in Fig. 3B, APP/PS1 mice exhibited a significant increase in the levels of insoluble and soluble forms of Aβ40 and Aβ42 in the hippocampus compared to levels in WT mice (P < 0.01). In contrast, we observed a significant reduction in the levels of insoluble and soluble forms of Aβ40 and Aβ42 in the hippocampus of APP/PS1 mice compared with vehicle-treated mice (P < 0.01, Fig. 3B). However, cotreatment with LPD and GW9662 markedly increased the area of Aβ plaques and the levels of insoluble and soluble forms of Aβ40 and Aβ42 in the hippocampus of APP/PS1 mice compared with LPD treatment alone (P < 0.01, Fig. 3A, B). These results suggested that LPD could attenuate amyloid pathologies in APP/PS1 mice in a PPARγ-dependent manner.

There is some evidence that cerebral glucose hypometabolism is associated with an increased risk of AD [1, 4]. Based on ¹⁸F-FDG microPET imaging, brain glucose uptake was significantly decreased in APP/PS1 mice compared with WT mice (P < 0.01, Fig. 3C, D), and this difference was reversed by LPD treatment (P < 0.01, Fig. 3C, D). However, GW9662 cotreatment alleviated the elevated ¹⁸F-FDG uptake in the brains of LPD-treated APP/PS1 mice (P < 0.01, Fig. 3C, D).

Transmission electron microscopy (TEM) was used to examine the effect of LPD on mitophagy in the hippocampi of APP/PS1 mice. Hippocampal neurons from APP/PS1 mice displayed altered mitochondrial morphology characterized by excessive mitochondrial damage in comparison to those from WT mice (Fig. 4A). Induction of mitophagy by LPD resulted in the clearance of damaged mitochondria in APP/PS1 mice (P < 0.01, Fig. 4A, B). However, blockade of PPARγ with GW9662 markedly decreased the number of mitophagy events in hippocampal neurons (P < 0.01, Fig. 4A, B).

To understand the molecular mechanisms responsible for the neuroprotective effects of LPD, PPARγ and potential downstream signaling pathways were investigated. Protein expression of PPARγ was significantly decreased in the hippocampus in APP/PS1 mice compared with WT mice (P < 0.01, Fig. 4B, C). However, protein levels of PPARγ, PINK1, and p-Parkin (Ser65) were markedly increased in hippocampal lysates from LPD-treated APP/PS1 mice (P < 0.01, Fig. 4B, C). Moreover, LPD treatment significantly increased the LC3-II/LC3-I ratio and decreased SQSTM1/p62 expression (P < 0.01, Fig. 4B, C). However, coadministration with GW9662 abolished the effects of LPD on the PPARγ and PINK1/Parkin signaling pathways (P < 0.01, Fig. 4B, C).

Western blot analysis showed significant increases in the protein expression of PPARγ, PINK1, and p-Parkin (Ser65) in APP/PS1 cells after LPD treatment (P < 0.01, Fig. 5A, B). A significant increase in the ratio of LC3-II/LC3-I and a significant decrease in SQSTM1/p62 expression was also observed in APP/PS1 cells treated with LPD (P < 0.01, Fig. 5A, B). In contrast, significant decreases in the expression levels of PPARγ, PINK1, and p-Parkin (Ser65) and the ratio of LC3-II/LC3-I, as well as a significant increase in SQSTM1/p62 expression, were observed in APP/PS1 cells cotreated with LPD and GW9662 or the mitophagy inhibitor Mdivi-1 (P < 0.01, Fig. 5A, B). Moreover, APP/PS1 cells exposed to LPD had significantly higher OCRs than control cells (P < 0.01, Fig. 5C). LPD treatment markedly increased the basal respiration, spare respiratory capacity, maximal respiration, and ATP production in APP/PS1 cells (P < 0.01, Fig. 5C). However, cotreatment with LPD and GW9662 or Mdivi-1 abrogated the effects of LPD on mitochondrial function in APP/PS1 cells (P < 0.01, Fig. 5C).