Post-Treatment with Amorfrutin B Evokes PPARγ-Mediated Neuroprotection against Hypoxia and Ischemia

In this study, we demonstrate for the first time that amorfrutin B, a selective modulator of peroxisome proliferator-activated receptor gamma—PPARγ, can protect brain neurons from hypoxia- and ischemia-induced degeneration when applied at 6 h post-treatment in primary cultures. The neuroprotective effect of amorfrutin B suggests that it promotes mitochondrial integrity and is capable of inhibiting reactive oxygen species—ROS activity and ROS-mediated DNA damage. PPARγ antagonist and Pparg mRNA silencing abolished the neuroprotective effect of amorfrutin B, which points to agonistic action of the compound on the respective receptor. Interestingly, amorfrutin B stimulated the methylation of the Pparg gene, both during hypoxia and ischemia. Amorfrutin B also increased the protein level of PPARγ during hypoxia but decreased the mRNA and protein levels of PPARγ during ischemia. Under ischemic conditions, amorfrutin B-evoked hypermethylation of the Pparg gene is in line with the decrease in the mRNA and protein expression of PPARγ. However, under hypoxic conditions, amorfrutin B-dependent hypermethylation of the Pparg gene does not explain the amorfrutin B-dependent increase in receptor protein expression, which suggests other regulatory mechanisms. Other epigenetic parameters, such as HAT and/or sirtuins activities, were affected by amorfrutin B under hypoxic and ischemic conditions. These properties position the compound among the most promising anti-stroke and wide-window therapeutics.


Introduction
Cerebrovascular accidents (commonly known as strokes) are the second leading cause of death and the leading cause of disability worldwide. Patients who survive stroke incidents experience loss of vision and/or speech, paralysis, and confusion [1]. The only pharmacological treatment approved by the Food and Drug Administration for acute ischemic stroke is recombinant tissue plasminogen activator (rt-PA), which is effective only if administered until 4.5 h after stroke onset. Moreover, rt-PA has a long list of contraindications, such as excitotoxicity, hemorrhage, and cerebral edema, and for this reason, only 5% of ischemic patients can be treated with rt-PA [2,3]. Additional stroke treatment is surgical thrombectomy, which can be used only up to 8 h following the onset of symptoms located in the anterior circulation and is eligible for no more than 2% of patients [4].
Aside from the stroke, another issue is perinatal asphyxia. Oxygen deprivation is considered the most common cause of death in fetuses and newborns, i.e., 2-4 newborns die per 1000 births. Each year, one million children die due to hypoxia and neonatal oxygen deprivation, which leads to permanent brain damage and disability such as hypoxicischemic encephalopathy [5]. Oxygen therapy and moderate hypothermia occurring up to 6 h after the hypoxic episode are the gold standard treatment for neonatal asphyxia [6].

Primary Neuronal Cell Cultures
Neocortical tissues for neuronal cell cultures were obtained from Swiss CD1 mouse strain embryos (Charles River, Germany) at 15 days of gestation and cultured as described previously [16]. Briefly, neuronal cells were seeded on polyornithine-coated multiwell dishes at a density of 2.5 × 10 5 cells per cm 2 and cultivated in phenol red-free neurobasal medium containing fetal bovine serum, B27, L-glutamine, and penicillin-streptomycin antibiotics. The neuronal cell cultures were kept in a humidified incubator (37 • C with 5% (vol/vol) CO 2 ) for 7 days in vitro (DIV). After 3 days, the medium was changed to medium without FBS but supplemented with L-glutamine, B27, and penicillin-streptomycin antibiotics. The number of astrocytes did not exceed 10%, as determined by the content of the intermediate filament protein GFAP (glial fibrillary acidic protein). The experiments were conducted in compliance with European Union Legislation (Directive 2010/63/EU, amended by Regulation (EU) 2019/1010).

Experimental Models Hypoxia
To induce hypoxic conditions, the cell medium was replaced with a standard medium and placed in a prewarmed and humidified hypoxia modular incubator chamber (Billups-Rothenberg, Inc., San Diego, CA, USA) with 95% N 2 /5% CO 2 for 6 h. The O 2 level was measured with an oxygen analyzer (Greisinger, Germany) and reached less than 0.5%. After 6 h of hypoxic conditions, i.e., at the reoxygenation period, the culture medium was replaced immediately with standard medium for 18 h.

Ischemia
To induce ischemic conditions, the cell medium was replaced with medium without glucose and placed in a prewarmed and humidified hypoxia modular incubator chamber (Billups-Rothenberg, Inc., San Diego, CA, USA) with 95% N 2 /5% CO 2 for 6 h. The O 2 level was measured with an oxygen analyzer (Greisinger, Germany) and reached less than 0.5%. After 6 h of ischemic conditions, i.e., at the reoxygenation period, the culture medium was replaced immediately with standard medium for 18 h.

Treatment Co-Treatment
The co-treatment experiments occurred when the cell cultures were treated simultaneously with hypoxic and ischemic episodes for 6 h, and the culture medium was changed to standard medium for 18 h reoxygenation (Scheme 1A). Cell cultures were treated with amorfrutin B at a concentration of 0.1-10 µM. During reoxygenation, cells were cultured in a humidified incubator (New Brunswick Scientific, Edison, NJ, USA).

Post-Treatment
After 6 h of hypoxic/ischemic conditions, i.e., at the reoxygenation period, the culture medium was replaced immediately with standard medium, and the treatment occurred for the next 18 h (Scheme 1B). Cell cultures were treated with amorfrutin B at a concentration of 0.1-10 µM. During reoxygenation, cells were cultured in a humidified incubator (New Brunswick Scientific, Edison, NJ, USA).
As the post-treatment paradigm reflects much better clinical and therapeutic aspects than the co-treatment paradigm, we chose the post-treatment paradigm for majority of experiments.

Measurement of LDH Activity
To quantify hypoxia-and ischemia-induced cytotoxicity, the extent of LDH release was measured using a Cytotoxicity Detection Kit (Roche, Basil, Switzerland) as previously described [17]. After the experiment, cell-free supernatants were collected from each well and incubated with the appropriate reagent mixture for 30 min at room temperature, according to the manufacturer's protocol. The intensity of red color was measured at a wavelength of 490 nm using an Infinite M200pro microplate reader (Tecan, Männedorf, Switzerland) and was proportional to the amount of LDH in the culture, which was, in turn, directly proportional to the number of dead or damaged cells. The data were analyzed with Tecan i-control software, normalized to the color intensity from vehicle-treated cells, and expressed as a percentage of the control value ± SEM. The absorbance of the blanks, which were no-enzyme controls, was subtracted from each value.

Assessment of Cell Metabolic Activity
The viability of neurons after hypoxic and ischemic conditions was determined using the colorimetric MTT assay. Mitochondrial function was assessed through MTT reduction by oxidoreductase enzymes to purple formazan. MTT solution was added to the medium, and cell cultures were incubated at 37 • C. After 1 h, the medium was replaced with 100% DMSO to dissolve formazan crystals, and the intensity of the purple color was proportional to the number of viable cells. The absorbance was measured at 570 nm using an Infinite M200pro microplate reader, and the data were analyzed with i-control software. The results were normalized to the color intensity from vehicle-treated cells and are presented as a percentage of the control ± SEM.

Measurement of the Degenerating Neurons
For the analysis of cytotoxicity in neuronal cell cultures after hypoxia and ischemia, Fluoro-Jade C (FJ-C) labeling was used, which is useful to label degenerating neurons using fluorochromatic dye. The stock solution was prepared as previously described [18] by mixing FJ-C with distilled water. Then, the culture medium was replaced with a prepared solution. After 1 h, the fluorescence of the FJ-C-labeled cells was measured using an excitation wavelength of 490 nm and an emission wavelength of 525 nm. The measurements were made using an Infinite M200PRO microplate reader, and the results were analyzed by i-control software. Data were normalized to the blank and presented as a percentage of the control ± SEM.

Measurement of ROS Formation
The levels of hydrogen peroxide (H 2 O 2 ) in neocortical neurons were measured after the experiment with the bioluminescent ROS-Glo™ H 2 O 2 Assay (Promega, Madison, WI, USA). In general, the substrate reacts directly with H 2 O 2 to generate a luciferin precursor. Adding Substrate Detection Solution causes the conversion to luciferin and provides the Ultra-Glo™ Luciferase to produce a light signal proportional to the amount of H 2 O 2 in cultured cells. The bioluminescence was measured with a GloMax ® Navigator Microplate Luminometer (Promega, Madison, WI, USA), and the data were normalized to the bioluminescent signal intensity in the vehicle-treated cells and presented as a percentage of the control ± SEM.

Estimation of DNA/RNA Oxidative Damage
The DNA/RNA Oxidative Damage (High Sensitivity) ELISA Kit (Cayman Chemical, Ann Arbor, MI, USA) was used to detect 8-OHdG in the samples after the experiment. This immunoassay consisted of binding the antibody-oxidatively damaged guanine complex to the goat polyclonal anti-mouse IgG. After adding Ellman's reagent, there was an enzymatic reaction that resulted in a product with a distinct yellow color. The absorbance of this product was determined spectrophotometrically at 412 nm, and the intensity of color was inversely proportional to the amount of free 8-OHdG. The data were normalized to the vehicle-treated cells, and the results are presented as the 8-hydroxy-2 -deoxyguanosine concentration ± SEM.

PPARγ Antagonist
To verify the involvement of PPARγ signaling in the neuroprotective action of amorfrutin B, the cultures were treated with a PPARγ antagonist (GW9662, 1 µM). To avoid nonspecific reactions, amorfrutin B and GW9662 were used at concentrations that did not affect LDH release under normoxic conditions. Amorfrutin B and PPARγ antagonist were originally dissolved in dimethyl sulfoxide (DMSO) and then diluted in a culture medium, resulting in DMSO concentrations below 0.1%.
During experiments with the use of PPARγ receptor antagonist GW9662, we decided to use both effective concentrations, i.e., 1 and 5 µM of amorfrutin B. This was because of uncertainty whether antagonist activity would dysregulate chosen parameter, i.e., LDH release in nonspecific way. To inhibit Pparg expression in neuronal cell cultures, specific Pparg siRNA (50 nM) was applied (Santa Cruz Biotechnology Inc., Santa Cruz, CA USA). siRNA was added to medium without antibiotics containing the siRNA transfection reagent INTERFERin. After 7 h of transfection, the culture medium was replaced with a standard medium, and the cells were incubated until the next day of the experiment. Then, the cells were exposed to hypoxia or ischemia for 6 h and next treated with amorfrutin B during the reoxygenation. The negative siRNA did not lead to the specific degradation of any known cellular mRNA and was used as a control.
Gene silencing may affect cell survival nonspecifically. As both the 1 and 5 µM concentrations of amorfrutin B caused neuroprotection, though 5 µM amorfrutin B was more effective, we used both these concentrations of amorfrutin B to additionally confirm the specific actions of gene silencing.
2.1.11. qPCR Analysis of mRNAs Specific to Genes Encoding Hif1a, Pparg, Pgc1a, and Adipoq Total RNA was extracted from neocortical cells at 18 h post-treatment after hypoxia and ischemia with the RNeasy Mini Kit (Qiagen, Hilden, Germany) using the spin columnbased method. The quantity of RNA was measured spectrophotometrically at 260 nm and 260/280 nm using a NanoDrop ND-1000 UV-Vis Spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA), and the RNA purity was determined when the A260/A280 ratio was approximately 2.0. Then, there was two-step qPCR involving reverse transcription and qPCR, which were run on a CFX 96 Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA). Total RNA was reverse transcribed with a High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific, Waltham, MA, USA) in accordance with the manufacturers' protocol. The collected cDNA was used as the PCR template and amplified with the TaqMan Gene Expression Master Mix kit (Thermo Fisher Scientific, Waltham, MA, USA) using TaqMan probes as primers for the specific genes encoding Hif1a, Pparg, Pgc1a, and Adipoq. The PCR process consisted of series of temperature changes: 2 min at 50 • C and 10 min at 95 • C, followed by 40 cycles of 15 s at 95 • C and 1 min at 60 • C. During the exponential phase, the threshold cycle (Ct) for each sample was set, and the delta delta Ct method was used for data analysis. The Hprt was selected as a reference gene using the following algorithms: geNorm, NormFinder, BestKeeper and delta Ct.
2.1.12. Enzyme-Linked Immunosorbent Assays for PPARγ, PGC1α and ADIPOQ Enzyme-linked immunosorbent assays (ELISAs) and the quantitative sandwich enzyme immunoassay technique were used to quantify the PPARγ, PGC1α, and ADIPOQ protein expression levels in the collected cell lysates after the experiment. The protein concentration was assessed using Bradford reagent (Bio-Rad Protein Assay) and bovine serum albumin as a standard. Each well of the 96-well plate was precoated with mouse antibodies specific for PPARγ, PGC1α, or ADIPOQ. The standards and samples were added to the wells with biotin-conjugated polyclonal antibodies specific for PPARγ, PGC1α, and ADIPOQ. Then, streptavidin-HRP (horseradish peroxidase) was added to the wells bound to the biotinylated antibodies, and during the washing step, the unbound streptavidin-HRP was washed away. The substrate solution caused a color change in proportion to the amount of proteins, and the addition of acidic stop solution terminated the reaction. The absorbance was measured at 450 nm using an Infinite M200PRO microplate reader, and the data were normalized to the color intensity from vehicle-treated cells and expressed as a percentage of the control value ± SEM.

Western Blot Analysis
After 18 h of reoxygenation, the cells were lysed in ice-cold RIPA lysis buffer with a protease inhibitor cocktail. The lysates were sonicated and centrifuged at 15,000× g for 20 min at 4 • C. The supernatants were collected, and the protein concentration was assessed using Bradford reagent (Bio-Rad Protein Assay) and bovine serum albumin as a standard. The cell lysates containing 30 µg of the total protein were reduced and denatured by boiling each sample in a 2× Laemmli sample buffer at 95 • C for 5 min. The molecular weight marker and samples were loaded into the wells of a 10% SDS-polyacrylamide gel, and the proteins were electrophoretically separated using a Bio-Rad Mini-Protean II Electrophoresis Cell. After separation, the proteins were electrotransferred from the gel to polyvinylidene fluoride (PVDF) membranes (Millipore, Burlington, MA, USA) using the Bio-Rad Mini Trans-Blot apparatus. To prevent nonspecific antibody binding, the membranes were washed and blocked with 5% dried milk and 0.2% Tween-20 in 0.02 M Tris-buffered saline (TBS) for 2 h. The membranes were incubated overnight (at 4 • C) with one of the following appropriate dilutions of primary antibodies: anti-HIF1α rabbit polyclonal antibody (diluted 1:100), anti-PPARγ rabbit polyclonal antibody (diluted 1:100) or antiβ-actin mouse monoclonal antibody (diluted 1:3000) in TBS/Tween. After incubation with primary antibodies, the membranes were washed with TBS and 0.02% Tween 20 and incubated for 2 h with HRP-conjugated goat anti-rabbit IgG or goat anti-mouse IgG secondary antibodies (Santa Cruz Biotechnology, Santa Cruz, CA, USA) diluted at 1:1000 in TBS/Tween. To control the amount of denatured protein loaded onto the gel, the membranes were stripped and reprobed with an anti-β-actin antibody. The signals were developed by enhanced chemiluminescence (ECL) using BM Chemiluminescence Blotting Substrate (Roche Diagnostics, Basil, Switzerland) and visualized using a Luminescent Image Analyzer Fuji-Las 4000 (Fuji, Japan). Immunoreactive bands were quantified using a MultiGauge V3.0 image analyzer and normalized to β-actin. The data are presented as a percentage of the control value ± SEM, and pictures of exemplary stripes are shown.

Immunofluorescence Staining of PPARγ and MAP2
To visualize the cellular localization of PPARγ and confirm the neuronal nature of the neocortical cells, immunofluorescence labeling, and confocal microscopy were used. The neocortical cells were cultured on glass coverslips and subjected to immunofluorescence double labeling at 7 DIV. Cultured cells were fixed for 15 min at room temperature with a 4% paraformaldehyde solution in PBS, and then cells were incubated for 1 h in a blocking buffer containing 5% normal donkey serum and 0.3% Triton X-100 in 0.01 M PBS. After that, the neurons were incubated for 24 h at 4 • C with primary antibodies: anti-PPARγ mouse monoclonal (diluted 1:50) and anti-MAP2 rabbit polyclonal (diluted 1:100) antibodies (Santa Cruz Biotechnology, Santa Cruz, CA, USA). This step was followed by an overnight incubation in a mixture of the following secondary antibodies (Thermo Fisher Scientific, Waltham, MA, USA): Alexa Fluor Plus 488-conjugated anti-mouse IgG (1:300) and Alexa Fluor Plus 647-conjugated anti-rabbit IgG (1:300). The microscopic preparations were then washed with PBS, mounted, and cover-slipped. For viewing slides, a Leica TCS SP8 WLL confocal laser scanning microscope (DMi8-CS, Leica Microsystem, Wetzlar, Germany) was used.
To quantify the immunofluorescence signal corresponding to PPARγ expression level, the frequency of the brightest pixels in the region of interest (ROI) has been measured. To assess pixel intensity the ImageJ software has been used.

Analyses of DNA Methylation Global DNA Methylation
The DNA from the cell cultures was extracted with the use of Quick-gDNA™ MicroPrep Kit (Zymo Research, Irivine, CA, USA) and quantified spectrophotometrically using the Nan-oDrop ND-1000 UV-Vis Spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). The methylation pattern was evaluated using an Imprint Methylated DNA Quantification Kit (Sigma-Aldrich, Saint Louis, MO, USA). According to the manufacturer's instructions, 50 ng of purified DNA was added to each well where the methylated DNA was detected by capture and detection antibodies. The reaction for color change was monitored using the developing solution. The absorbance of the colorimetric reaction product was measured at 450 nm using an Infinite M200PRO microplate reader.
The Pparg Gene Specific Methylation Genomic DNA was extracted using a Quick-gDNA™ MicroPrep Kit (Zymo Research, Irvine, CA, USA), and the quantity of DNA was determined spectrophotometrically at 260 nm and 260/280 nm with the use of a NanoDrop ND-1000 UV-Vis Spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). Then, denaturation and complete bisulfite conversion of GC-rich DNA were accomplished by an EZ DNA Methylation-Gold™ Kit (Zymo Research, Irvine, CA, USA). The samples were eluted in a 10 µL volume and subjected to qPCR (MethyLight) using the EpiTect MethyLight PCR Kit (Qiagen, Hilden, Germany). The methylation regions in Pparg gene were verified in a CpG hot spots in the 5' flanking sequence (2000 bp). The Methyl Primer Express Software 1.0. has been used to design methylated and unmethylated target sequences and to design the primers. Fully methylated and fully unmethylated TaqMan probes were designed for Pparg promoters, and the internal reference set for the Hprt gene was designed to control the input DNA. The unmethylation-specific TaqMan probe was bound to VIC ® , while the methylation-specific TaqMan probe was linked to FAM™ as a 5 reporter dye. The release of these indicators and determination of the ratio of measured Ct values with both fluorescence dyes enable the assessment of methylation status. To calculate the methylation degree in samples, the Ct determined with each of both dyes were used: percentage of methylation [%]: C meth = 100/[1 + 2(∆Ct meth − ∆Ct unmeth )].

Estimation of Histone Deacetylase and Acetyltransferase Activities Histone Deacetylase (HDAC) Activity
To verify HDAC activity, the Histone Deacetylase Assay Kit (Sigma-Aldrich, Saint Louis, MO, USA) was applied. The assay kit is based on a two-step enzymatic reaction: I. deacetylation of the acetylated lysine side chain by the HDAC-containing sample and II. cleavage of the deacetylated substrate by the developer solution. The fluorescence was measured with an Infinite M200PRO microplate reader at λex = 365 nm/λem = 460 nm, and the release of the free highly fluorescent group was proportional to the deacetylation activity. The kit included positive (HeLa cell lysate) and negative (trichostatin A) controls.

Sirtuins Activity
The Sirtuin Activity Assay Kit (Sigma-Aldrich, Saint Louis, MO, USA) was used to estimate the sirtuins activity in the samples. Sirtuins in the presence of NAD+ are able to form deacetylated p53-AFC substrates, nicotinamide, and O-acetyl-ADP ribose. According to the protocol, adding 'developer solution' caused the cleavage of the deacetylated p53-AFC substrate and the release of the fluorescent group, which was fluorometrically detected at λex = 400 nm/λem = 505 nm with an Infinite M200PRO microplate reader. The p53-AFC substrate can also be deacetylated by nonsirtuin HDACs; therefore, trichostatin A was added to the solution to inhibit nonsirtuins activities in the samples. The positive control has also been included.

Estimation of Histone Acetyltransferase (HAT) Activity
HAT activity was measured using the HAT Activity Fluorometric Assay Kit (Sigma-Aldrich, Saint Louis, MO, USA). According to the manufacturers' protocol, the reaction product of histone acetyltransferase activity reacts with the developer solution, which results in the fluorescent product of the reaction measured using an Infinite M200PRO microplate reader (λex = 535/λem = 587 nm). The obtained values were proportional to the HAT activity in the samples. The assay contained HeLa nuclear extract, which was used as a positive control.

Data Analysis
The statistical analysis was performed based on raw data expressed as the absorbance or fluorescence units per well containing 50,000 cells for the LDH, MTT, FJ-C, ROS formation, DNA/RNA oxidative damage assays; the fluorescence units per 1.5 million cells for qPCR, global DNA methylation, HAT, HDAC, and sirtuin activity; the mean optical density per 30 µg of protein for Western blot assays; the pg of PPARγ, PGC1α, and ADIPOQ per µg of total protein for the ELISAs. To determine overall significance, an analysis of variance (ANOVA) was used, previously preceded by Levene's test of homogeneity of variances. The differences between the control (vehicle-treated cells) and experimental groups were defined with a post hoc Newman-Keuls test. Significant differences were marked in the following ways: * p < 0.05, **p < 0.01, and *** p < 0.001 (compared to the control groups), # p < 0.05, ## p < 0.01, and ### p < 0.001 (compared to the cultures exposed to hypoxia), p < 0.05,ˆˆp < 0.01, andˆˆˆp < 0.001 (compared to the cultures exposed to ischemia). The results for DNA/RNA oxidative damage, qPCR, ELISAs, DNA methylation, and histone deacetylase activity are expressed as the mean ± SEM of three independent experiments. The results for LDH release, MTT, FJ-C, ROS activity, Western blot analysis, and HAT activity are presented as the percentage of the control ± SEM of three independent experiments. The number of replicates ranged from 6 to 12.
Under normoxic conditions, amorfrutin B did not induce changes in: MTT, Fluoro-Jade C, ROS activity, 8-OHdG level, protein and mRNA expression levels (except for slight increase in Pgc1a mRNA level), epigenetic status of neuronal cells, and immunofluorescence intensity (Supplementary Materials- Table S1). Hypoxic or ischemic conditions and exposure to amorfrutin B (1 and 5 µM) post-treatment did not contribute to changes in caspase-3 activity ( Figure S1).

The Effects of Amorfrutin B on Hypoxia-and Ischemia-Induced Lactate Dehydrogenase (LDH) Release in Neocortical Cell Cultures
In this study, a model of 6 h hypoxia or ischemia followed by 18 h of reoxygenation was applied. Hypoxia and ischemia conditions induced LDH release to 192% and 338% of the normoxic value, respectively (Figure 1a).
In the paradigm of co-treatment, i.e., when the amorfrutin B was added at the same time as the injury started, the effects were as follows: (i) in the hypoxic model, amorfrutin B (1 µM and 5 µM) co-treatment inhibited LDH release to 83% (17% decrease) and 81% (19% decrease) of the hypoxia value, respectively; (ii) in the ischemic model, amorfrutin B (1 µM and 5 µM) co-treatment also reduced ischemia-induced LDH activity to 82% (18% decrease), Figure 1b.
Under normoxic conditions, amorfrutin B used at concentrations ranging from 0.1 to 5 µM did not change LDH release, but amorfrutin B at a concentration of 10 µM was cytotoxic to neocortical cells. Therefore, for further experiments, we did not use 10 µM amorfrutin B.
We chose the post-treatment paradigm for majority of experiments because it reflects much better clinical aspects than co-treatment paradigm. As the most promising posttreatment paradigm was with the 5 µM concentration (Figure 1c), it has been used in the next experiments. The results are presented as a percentage of the control ± SEM of 3 independent experiments, consisting of 8-12 replicates per group. *** p < 0.001 compared to the control group, ### p < 0.001 compared to the cultures exposed to hypoxia,ˆˆp < 0.001 compared to the cultures exposed to ischemia.

The Impact of Amorfrutin B on the Viability of Neuronal Cells under Hypoxic and Ischemic Conditions
Six hours of hypoxic or ischemic conditions reduced the viability of neuronal cells from the control value (100%) to 88% and 85%, respectively, as measured using MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide). Exposure to 5 µM amorfrutin B for 18 h of reoxygenation after hypoxia and ischemia resulted in an improvement in mitochondrial activity and an increased number of viable cells. Amorfrutin B administration increased cell survival after hypoxia and ischemia to 97% (9% increase) and 100% (15% increase) of the control value, respectively (Figure 2a). The results are presented as a percentage of the control ± SEM of 3 independent experiments, consisting of 8-12 replicates per group. ** p < 0.01 and *** p < 0.001 compared to the control group, # p < 0.05 and ### p < 0.001 compared to the cultures exposed to hypoxia,ˆˆˆp < 0.001 compared to the cultures exposed to ischemia.

Amorfrutin B Reduced the Degeneration of Neuronal Cells Caused by Hypoxia and Ischemia
Fluoro-Jade C (FJ-C) labeling showed that cell degeneration levels due to hypoxia and ischemia reached 119% and 124% of the control value, respectively. In neocortical cells, amorfrutin B (5 µM) used in the post-treatment paradigm reduced the extent of neurodegeneration in both the hypoxic and ischemic models. After amorfrutin B application, the FJ-C indicator decreased 18% compared to hypoxia (101% of the control value) and 30% compared to ischemia (94% of the control value), Figure 2b.

The Effects of Amorfrutin B on ROS Formation under Hypoxic and Ischemic Conditions
In both the hypoxic and ischemic models, there were statistically significant increases in ROS activity. ROS formation reached 207% and 245% of the control value during hypoxia or ischemia, respectively. Amorfrutin B (5 µM) reduced oxidative stress markers in both models; there was a reduction in ROS activity to 122% of the control value (Figure 3a). There was an increase in reactive oxygen species formation and 8-OHdG in the hypoxic and ischemic models, and amorfrutin B diminished the activity of these markers in both cases (a,b). The ROS activity results are presented as a percentage of the control ± SEM, and the 8-OHdG level results are presented as the mean ± SEM. There were 3 independent experiments, consisting of 10-12 replicates per group. *** p < 0.001 compared to the control group, ### p < 0.001 compared to the cultures exposed to hypoxia,ˆˆˆp < 0.001 compared to the cultures exposed to ischemia.

Influence of PPARγ Antagonist on the Effect of Amorfrutin B in Neuronal Cells Exposed to Hypoxia and Ischemia
A PPARγ antagonist (GW9662) was used at a 1 µM concentration that did not alter LDH release under normoxic conditions. In this study, GW9662 reduced the neuroprotective effects of 1 and 5 µM amorfrutin B in the model of hypoxia, which was manifested by elevated levels of LDH release to 94% and 134% of the control value, respectively, i.e., 20% and 100% increases. Under ischemic conditions, amorfrutin B reduced LDH release to 79% (1 µM) and 76% (5 µM) of the normoxic value. GW9662 reversed these protective effects by enhancing LDH release up to 91-106% of the control value ( Figure 4). The results are presented as a percentage of the control ± SEM of 3 independent experiments, consisting of 6-8 replicates per group. *** p < 0.001 compared to the control group, ### p < 0.001 compared to the cultures exposed to hypoxia,ˆˆˆp < 0.001 compared to the cultures exposed to ischemia.

Effect of Amorfrutin B on Hypoxia-and Ischemia-Induced LDH Release in Pparg siRNA-Transfected Neocortical Cells
In Pparg siRNA-transfected cells, amorfrutin B lost its neuroprotective potential. After transfection and treatment with amorfrutin B (1 or 5 µM), LDH reached 113% of the control value under hypoxic conditions. Moreover, in Pparg siRNA-transfected cells in the ischemic model, LDH activity reached 119%-121% (both concentrations of amorfrutin B) of the control value ( Figure 5). The effectiveness of mRNA silencing was verified through the measurement of Pparg mRNAs using qPCR. Based on our previous study, after the silencing of Pparg with the specific siRNAs, the Pparg mRNA concentration has been reduced by 31% compared to the non-transfected wild-type cells [19].

Effects of Amorfrutin B on the mRNA Expression Levels of Hif1a, Pparg, Pgc1a, and Adipoq in Models of Hypoxia and Ischemia
The mRNA expression levels were measured with quantitative polymerase chain reaction (qPCR). In our study, a 6 h exposure to hypoxia enhanced the hypoxia-inducible factor 1 alpha (Hif1a) mRNA expression level to 1.26-fold than under normoxic conditions. Pparg and PPARγ-dependent adiponectin (Adipoq) mRNA expression levels were also stimulated to 2.43-fold and 3.68-fold, respectively. Hypoxia did not change the expression level of peroxisome proliferator activated receptor gamma coactivator 1 alpha (Pgc1a) mRNA in neocortical cells. During hypoxia, amorfrutin B (5 µM) post-treatment decreased the expression levels of Hif1a mRNA to 0.97-fold (23% decrease) and Pgc1a mRNA to 0.72fold (28% decrease). In turn, amorfrutin B (5 µM) increased the mRNA expression level of Pparg to 1.84-fold compared to normoxia, but it did not change the mRNA expression level of Adipoq mRNA.

Figure 5.
In Pparg siRNA-transfected cells, amorfrutin B (1 or 5 µM) lost the neuroprotective potential that was manifested by increased LDH release levels in both hypoxic and ischemic models. Neuronal cell cultures were transfected with 50 nM Pparg siRNA in INTERFERin-containing medium without antibiotics for 7 h. Then, the cells were exposed to hypoxia or ischemia for 6 h and next treated with amorfrutin B during the reoxygenation. The results were normalized to the absorbance in negative control siRNAs. The results are presented as a percentage of the control ± SEM of 3 independent experiments, consisting of 5-8 replicates per group. # p < 0.05 and ## p < 0.01 compared to the cultures exposed to hypoxia,ˆp < 0.01 compared to the cultures exposed to ischemia.
A 6 h exposure to ischemia enhanced the expression level of all studied genes, i.e., Hif1a mRNA to 1.72-fold, Pparg mRNA to 3.06-fold, Pgc1a mRNA to 1.32-fold, and Adipoq mRNA to 2.76-fold compared to the normoxia control group (1.00-fold). Under ischemic conditions, amorfrutin B (5 µM) post-treatment changed only the Pparg mRNA expression level to 2.20-fold, which was a decrease of 28% compared with the ischemic group. Amorfrutin B treatment did not significantly change the expression levels of Pgc1a, Hif1a, and Adipoq mRNA. All values were normalized to hypoxanthine-guanine phosphoribosyltransferase (Hprt), Figure 6. The results are presented as a mean ± SEM. There were 3 independent experiments, consisting of 6 replicates per group. * p < 0.05, ** p < 0.01, and *** p < 0.001 compared to the control group, ## p < 0.01, ### p < 0.001 compared to the cultures exposed to hypoxia,ˆp < 0.05 compared to the cultures exposed to ischemia.

Effects of Amorfrutin B on the Protein Expression Levels of HIF1α, PPARγ, PGC1α, and ADIPOQ in Models of Hypoxia and Ischemia
Enzyme-linked immunosorbent assays (ELISAs) showed that the protein levels of PPARγ, PGC1α, and ADIPOQ in control neuronal cells reached 0.0095, 0.0020, and 2.29 pg per µg of total protein, respectively. After hypoxia, the level of PPARγ reached 0.0115 pg, and amorfrutin B (5 µM) increased this value to 0.0166 pg per µg of total protein (44% increase). In response to ischemic conditions, the PPARγ protein expression level increased from 0.0095 pg to 0.0175 pg (84% increase), and exposure to amorfrutin B decreased the PPARγ level to 0.014 pg (20% decrease compared to ischemia). There were no statistically significant changes in PGC1α and ADIPOQ protein expression levels under hypoxic and ischemic conditions, and amorfrutin B post-treatment also did not affect the levels of these proteins (Figure 7a).
Western blot analysis showed that hypoxic conditions induced an increase in HIF1α protein expression levels from 100% to 124% of the normoxic value (24% increase) and amorfrutin B (5 µM) post-treatment decreased this value to 94% (25% decrease). In this study paradigm, hypoxia did not affect the protein expression of PPARγ; however, amorfrutin B treatment increased the PPARγ level to 124% of the normoxic value (24% increase). Under ischemic conditions, there was also an increase in HIF1α protein expression to 160% of the normoxic value (60% increase), but exposure to amorfrutin B did not significantly change this value. The same experimental conditions caused an increase in the PPARγ protein level to 142% (42% increase), and amorfrutin B reduced this level to 114% of the normoxic value (Figure 7b).

Confocal Microscopic Analysis of PPARγ and MAP2 Localization in Neuronal Cells
Immunofluorescence staining and confocal microscopy showed that PPARγ receptors (green staining) were localized in neuronal cells at 7 days in vitro (DIV). Exposure to hypoxic conditions did not change the PPARγ specific staining, but it was enhanced after amorfrutin B (5 µM) post-treatment. Moreover, PPARγ-specific labeling showed an increase in staining after ischemic conditions, which was reduced after exposure to amorfrutin B. Microtubule-associated protein 2 (MAP2) (red staining) indicated the location of neuronal cells and showed that in comparison to the control, there was hypoxia-induced inhibition of neurite outgrowth, which was even more advanced under ischemic conditions. In both cases, amorfrutin B significantly improved cell viability (Figure 8).
The expression levels of PPARγ have been quantified using basic intensity quantification method. According to these data, hypoxia did not affect PPARγ expression level (97% of the control) while post-treatment with amorfrutin B increased its expression to 125% of the control value. In the ischemic condition, PPARγ expression level enhanced by 53% (153% of the control value) and amorfrutin B restored its expression to over 90% of the control (Figure 8).  showed that hypoxia or ischemia and amorfrutin B (5 μM) post-treatment changed the protein expression levels of HIF1α and PPARγ but did not affect the protein expression levels of PGC1α and ADIPOQ. The results are presented as a mean ± SEM. There were 3 independent experiments, consisting of 5-6 replicates per group. ** p < 0.01, and *** p < 0.001 compared to the control group, ## p < 0.01, ### p < 0.001 compared to the cultures exposed to hypoxia, ^p < 0.05, ^^^p < 0.001 compared to the cultures exposed to ischemia.  post-treatment changed the protein expression levels of HIF1α and PPARγ but did not affect the protein expression levels of PGC1α and ADIPOQ. The results are presented as a mean ± SEM. There were 3 independent experiments, consisting of 5-6 replicates per group. ** p < 0.01, and *** p < 0.001 compared to the control group, ## p < 0.01, ### p < 0.001 compared to the cultures exposed to hypoxia, p < 0.05,ˆˆˆp < 0.001 compared to the cultures exposed to ischemia. 152.702 ± 9.341*** . The results are presented as a percentage of the control ± SEM. *** p < 0.001 compared to the control group, # p < 0.05 compared to the cultures exposed to hypoxia,ˆˆˆp < 0.001 compared to the cultures exposed to ischemia. The number of replicates in each group was 5. . The results are presented as a mean ± SEM. There were 3 independent experiments, consisting of 6 replicates per group. * p < 0.05, ** p < 0.01, and *** p < 0.001 compared to the control group, ### p < 0.001 compared to the cultures exposed to hypoxia,ˆˆˆp < 0.001 compared to the cultures exposed to ischemia.

The Pparg Gene Specific Methylation
In neuronal cell cultures, the methylation rate in the control group under normoxic conditions reached 50%. Our study showed that in response to hypoxia, there was no statistically significant change in Pparg methylation, but amorfrutin B (5 µM) post-treatment caused an increase in methylation from 42% (after hypoxia) to 62% (hypermethylation). In turn, Pparg methylation decreased from 50% in the control group to 15% (hypomethylation) under ischemic conditions, and after amorfrutin B treatment, there was an increase to 32% (hypermethylation), Figure 9b In this study, hypoxic and ischemic conditions did not significantly affect HDAC activity, reaching approximately 1.4 µM/µg. Additionally, in response to amorfrutin B exposure, there were no changes in this parameter in the studied models. Positive and negative controls were estimated at levels of 0.58 µM/µg (HeLa cell lysate) and 0.01 µM/µg (HDAC inhibitor-trichostatin A), respectively (Figure 10a). Figure 10. Hypoxic or ischemic conditions and exposure to amorfrutin B (5 µM) post-treatment did not contribute to changes in histone deacetylase (HDAC) activity (a) but altered sirtuins and nonsirtuins action (b) and histone acetyltransferase (HAT) activity (c). A positive control contained HeLa cell lysate for HDAC, an undefined sample for sirtuins, and a HeLa nuclear extract for HAT. Negative control had no cells, and an HDAC inhibitor (trichostatin A) was added to inhibit nonsirituin activities in the samples. The results are presented as the mean ± SEM (HDAC and sirtuins activity) or as a percentage of the control ± SEM (HAT activity). There were 3 independent experiments, consisting of 5-6 replicates per group. ** p < 0.01, and *** p < 0.001 compared to the control group, # p < 0.05, ## p < 0.01, ### p < 0.001 compared to the cultures exposed to hypoxia,ˆˆˆp < 0.001 compared to the cultures exposed to ischemia.

Sirtuins Activity
In response to hypoxia, sirtuins activity increased from a control value of 796 pM (100%) to 893 pM (12% increase), and amorfrutin B post-treatment decreased this value to 727 pM (19% decrease). Although hypoxic conditions did not affect nonsirituin HDAC activity (trichostatin A-treated cells), the addition of amorfrutin B decreased this parameter from 293 pM to 203 pM (31% decrease). In ischemia, treatment with amorfrutin B did not affect the activities of sirtuins and nonsirituin HDACs. A positive control was provided by the manufacturer and displayed a level of 250 pM (Figure 10b).

HAT Activity
Hypoxia and ischemia decreased HAT activity, which was reduced to 35% and 54% of the control value, respectively. Amorfrutin B (5 µM) treatment increased the HAT activity to 62% under hypoxic conditions and 209% under ischemic conditions of the control value. In the HeLa nuclear extract used as a positive control, HAT activity was estimated at 56% (Figure 10c).

Discussion
There is an urgent need to establish a new pharmacotherapy that would have a wide window, i.e., wider than 4.5 h for rt-PA, to effectively treat hypoxia-and ischemia-induced brain damage. In this study, we demonstrated for the first time that amorfrutin B, a selective modulator of PPARγ, can protect mouse brain neurons from hypoxia-and ischemiainduced degeneration when applied at 18 h post-treatment, which started at a 6 h delay from hypoxia and ischemia. Furthermore, PPARγ antagonist and Pparg mRNA silencing with specific siRNA abolished the neuroprotective effect of amorfrutin B, which points to agonistic action of the compound on the respective receptor. Previously, PPARγ agonists such as ciglitazone, mifepristone, 15d-PGJ2/15-PGJ(2), and pioglitazone were found to effectively alleviate cerebral ischemia-reperfusion injury in rats after MCAO, i.e., middle carotid artery occlusion [20][21][22] and bilateral common carotid artery occlusion [23]. Only a few studies have shown the neuroprotective potential of PPARγ agonists in post-treatment paradigms. These include thiazolidinediones such as rosiglitazone, which decrease the infarct volume and neurological deficits in rodents when applied until 6 h of reperfusion after MCAO [24,25]. Pioglitazone was found to evoke neuroprotection in terms of reduced brain infarction and neurological deficits when applied 2-5 days after MCAO [26]. Unfortunately, thiazolidinediones appeared hepatotoxic, which is why their clinical use is controversial. Moreover, newer thiazolidinediones, i.e., rosiglitazone and pioglitazone, have been linked to increased heart failure [27].
Compared to PPARγ agonists, selective PPARγ modulators such as amorfrutin B have safer pharmacological profiles and exhibit partial agonist or partial antagonist properties [12]. However, the neuroprotective capacity of selective PPARγ modulators in cellular and animal models of stroke is much less recognized than that of PPARγ agonists. The neuroprotection against hypoxia and ischemia attributed to selective PPARγ modulators is almost exclusively assessed on an example of telmisartan, which binds to the receptor in a different way than thiazolidinediones. Telmisartan, which was used as a pretreatment, suppressed cerebral injury in a murine model of transient focal ischemia but appeared ineffective in reducing stroke volume due to permanent ischemia [28]. The effect of pretreatment with telmisartan was reversed by antagonizing PPARγ with GW9662, thus confirming the agonizing action of the receptor modulator [29].
In the present study, we have provided evidence on the neuroprotective effect of post-treatment with amorfrutin B, the compound with proven properties of a selective PPARγ modulator [14,15]. The effect of amorfrutin B was estimated in terms of LDH, MTT, ROS activity, ROS-related 8-OHdG, and FJ-C, which suggests that amorfrutin B promotes mitochondrial integrity and is capable of inhibiting ROS activity and ROS-mediated DNA damage to prevent hypoxia-and ischemia-induced neural degeneration. Mitochondrial dysfunction is a major feature of degenerating neurons, and this phenomenon is connected with oxidative stress and PPARγ signaling. Recently, Xia et al. (2018) demonstrated that pretreatment with pioglitazone significantly alleviated ROS generation in cellular and animal models of cerebral ischemia, which is in line with the PPARγ-mediated neuroprotection evoked by amorfrutin B in our experiments [30]. A population study showed that blood concentrations of 8-OHdG are higher in patients with cerebral infarction than in healthy subjects [31]. In our study, this parameter was inhibited by a 6 h delayed post-treatment with amorfrutin B, as evidenced by the reduced level of 8-OHdG, which positions the compound among the most promising anti-stroke and wide-window therapeutics.
In our studies, hypoxia and ischemia evoked hypomethylation of global DNA. Similar to our studies, using next generation sequencing, Meller et al. [32] also showed that global DNA hypomethylation predominates as a response to preconditioning ischemia/harmful ischemia in the primary neuronal cultures subjected to oxygen and glucose deprivation. Although amorfrutin B did not affect hypoxia-and ischemia-evoked hypomethylation of global DNA in our study, it stimulated the methylation of the Pparg gene under hypoxic and ischemic conditions. Furthermore, amorfrutin B increased the protein level of PPARγ during hypoxia but decreased the mRNA and protein levels of PPARγ during ischemia, as evidenced by qPCR, Western blot, ELISA, and confocal microscopy. The expression of Pparg/PPARγ has been shown to depend on the duration of the injury and reoxygenation/reperfusion, as well as Hif1a/HIF-1α, which is known to activate the expression of PPARγ in response to hypoxia [33]. In our experimental models, the expression of Pparg/PPARγ was upregulated by hypoxia and ischemia, except for the protein level of PPARγ, which was not changed by hypoxia. Because amorfrutin B strongly inhibits Hif1a expression during hypoxia, we suggest that the amorfrutin B-evoked increase in the protein level of PPARγ during hypoxia does not involve a stimulatory effect of HIF-1α. However, the involvement of HIF-1α in the amorfrutin B-evoked decrease in PPARγ expression during ischemia cannot be excluded as amorfrutin B does not affect this factor during ischemia. It should also be stressed that, intriguingly, in the neurons of the ischemic hemisphere, despite increased PPARγ expression, PPARγ DNA binding activity is decreased, thus suggesting reduced transcriptional activity of the receptor [34,35]. This could explain our results, where no correlation was found between PPARγ expression and the expression of PPARγ-regulated factors, i.e., Pgc1a/PGC1α and Adipoq/ADIPOQ. In our study, the decrease in Pgc1a has occurred only in respect to mRNA expression level, but the level of PGC1α protein remained unchanged. The downregulation of Pgc1a gene may be related to complex and still not fully recognized role of sirtuins activity in PGC1α expression. Sirtuins activity is one of the possible way how the transcriptional pathway of Pgc1a can be regulated because sirtuins (SIRT1) are able to form a negative-feedback loop with PPARγ-related pathway [36]. In our studies, the sirtuins activity decreased after the treatment with amorfrutin B in hypoxic condition. We can speculate that this process initiated a negative-feedback loop with Pgc1a mRNA expression. What is more, RNA material to qPCR analyses has been collected after 18 h of reoxygenation; therefore, the observed changes in the expression level of Pgc1a may be secondary.
We demonstrated that under ischemic conditions, amorfrutin B-evoked hypermethylation of the Pparg gene was in line with the amorfrutin B-evoked decrease in the mRNA and protein expression of PPARγ. However, under hypoxic conditions, the amorfrutin B-dependent hypermethylation of the Pparg gene does not explain the amorfrutin B-dependent increase in receptor protein expression, suggesting other regulatory mechanisms. Even though hypermethylation is mostly related to a decrease in gene expression levels, recent studies shown that DNA hypermethylation may also be associated with upregulated gene expression [37,38]. We also hypothesize that this might be related to sirtuins because amorfrutin B normalized sirtuins activity in neuronal cells undergoing hypoxia but not ischemia. Sirtuins are known to be involved in various cellular processes, including hypoxic and ischemic injury; however, their specific roles are only partially recognized. For example, Sirt3 knockout male mice are less vulnerable to ischemia/reperfusion or stroke injury [39], and upregulation of SIRT3 inhibits apoptosis in ischemia-challenged PC12 cells [40]. Intriguingly, in patients with acute cerebrovascular stroke, blood levels of SIRT1 are lower than those in control volunteers [41]. We postulate that hypoxia affects the specific set of sirtuins that differs from that affected by ischemia, and this specific set of sirtuins is vulnerable to post-treatment with amorfrutin B.
Regarding other epigenetic parameters, amorfrutin B did not change HDAC activity but stimulated HAT activity, both in hypoxic and ischemic conditions, even far above the level that has been detected in mouse neurons during normoxia. Enhanced HAT activity has been observed during adipogenesis when PPARγ expression was epigenetically stimulated [42]. Recent studies showed that change in histone acetylation homeostasis is common feature in neurodegenerative diseases including stroke. Both acetylation and deacetylation represent interesting targets for potential therapies in the treatment of hypoxic and ischemic diseases. In 2013, Lanzillotta et al. [43] using OGD model in primary cultures of mouse cortical neurons noticed that histone H3 acetylation was drastically decreased, while there were no change in HDAC activity that is similar to our observation in cell-based model of brain ischemia. According to the cited paper, these changes resulted from massive energy disturbances accompanying ischemia, reduction of pyruvate dehydrogenase activity and decreased generation of acetyl-CoA, an important cofactor for HAT activity. Therefore, we postulate that the increase in HAT activity after amorfrutin B post-treatment is the result of improving the cellular energy balance due to inhibition of oxidative stress and restoration of mitochondrial function. Furthermore, in our hypoxic model, amorfrutin B post-treatment reduced sirtuins activity which is known to be involved in epigenetic modifications by deacetylation. Previously, an inhibition of deacetylation was shown to partially prevent an impairment of histone acetylation [44] that is similar to the action of amorfrutin B in our study and suggests the HAT and sirtuins activities as potential therapeutic targets. According to our study, amorfrutin B appears to be a neuroprotective compound that prevents hypoxia-and ischemia-related epigenetic modifications, and in this way, it could recognize attractive targets that would serve to improve stroke pharmacotherapy.

Conclusions
In summary, 18 h post-treatment with amorfrutin B, which started at a 6 h delay from hypoxia and ischemia, evokes a strong neuroprotective effect against hypoxic and ischemic damage that involves PPARγ activation, hypermethylation of the Pparg gene, and dysregulation of Pparg/PPARγ expression, which would account for mitochondrial integrity maintenance and inhibition of oxidative stress and related DNA damage.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/biomedicines9080854/s1, Figure S1: Effects of amorfrutin B on caspase-3 activity, Table S1: Effects of amorfrutin B on studied parameters under the normoxic conditions.

Data Availability Statement:
The data that support the findings of this study are available from the corresponding author upon reasonable request. Some data may not be made available because of privacy or ethical restrictions.