Treatments to inhibit or repair neuronal cell damage sustained during focal ischemia/reperfusion injury in stroke are largely unavailable. We demonstrate that dietary supplementation with the antioxidant di-tert-butyl-bisphenol (BP) before injury decreases infarction and vascular complications in experimental stroke in an animal model. We confirm that BP, a synthetic polyphenol with superior radical-scavenging activity than vitamin E, crosses the blood–brain barrier and accumulates in rat brain. Supplementation with BP did not affect blood pressure or endogenous vitamin E levels in plasma or cerebral tissue. Pre-treatment with BP significantly lowered lipid, protein and thiol oxidation and decreased infarct size in animals subjected to middle cerebral artery occlusion (2 h) and reperfusion (24 h) injury. This neuroprotective action was accompanied by down-regulation of hypoxia inducible factor-1α and glucose transporter-1 mRNA levels, maintenance of neuronal tissue ATP concentration and inhibition of pro-apoptotic factors that together enhanced cerebral tissue viability after injury. That pre-treatment with BP ameliorates oxidative damage and preserves cerebral tissue during focal ischemic insult indicates that oxidative stress plays at least some causal role in promoting tissue damage in experimental stroke. The data strongly suggest that inhibition of oxidative stress through BP scavenging free radicals in vivo contributes significantly to neuroprotection.
We demonstrate that pre-treatment with ditert-butyl bisphenol(Di-t-Bu-BP) inhibits lipid, protein, and total thiol oxidation and decreases caspase activation and infarct size in rats subjected to middle cerebral artery occlusion (2 h) and reperfusion (24 h) injury. These data suggest that inhibition of oxidative stress contributes significantly to neuroprotection.
Stroke is a major public health challenge that is projected to increase in parallel with aging populations (Strong et al. 2007). Stroke is a major cause of morbidity and mortality and for survivors and their families, the ongoing disability and distress can be extreme. As major risk factors for stroke are aging, hypertension, and diabetes mellitus, developing nations are experiencing increasing prevalence of stroke owing to the upward trends in these risk factors.
Strokes can be either ischemic or hemorrhagic in nature, with the majority caused when a culprit thrombus occludes a blood vessel in the brain leading to acute cerebral ischemia. At present, post-stroke thrombolysis with tissue plasminogen activator (tPA) is the only approved therapy for acute ischemic stroke. However, tPA treatment has considerable limitations that preclude its widespread use. Primarily, tPA administration is restricted to within the first three hours of stroke, which excludes patients presenting late or with a delayed diagnosis. It is used in only a minority of cases owing to clinical exclusion factors (e.g., age), and reperfusion and recovery in patients is less than 50% with increased risk of hemorrhage considered a major concern (Davalos 2005). A recent study suggests that hemorrhage risk following tPA treatment may be lower than expected (Lyerly et al. 2014), indicating that there is scope to expand the treatment window for tPA. A major consideration for thrombolytics such as tPA is the potential for damage to neuronal cells via reperfusion injury and the expansion of the primary infarct. Therefore, the development of neuroprotective pharmacological therapies that increase both the treatment window and prognosis for cerebral ischemia is warranted (Zhang et al. 2004; McMahon et al. 2006).
Cerebral ischemia reperfusion injury (Puyal et al. 2013) and associated inflammation (Bohacek et al. 2012; Liu et al. 2013) play an important role in the evolution of brain damage. A hallmark of the pathogenesis of stroke is the early generation of reactive oxygen species (ROS) (Lee et al. 2009). Increased ROS production following stroke leads to lipid and protein oxidation (Polidori et al. 1998; Kelly et al. 2008), loss of endogenous antioxidants, and damage to neuronal cells via a mechanism involving ROS-induced necrosis and apoptosis (Margaill et al. 2005). Diminished plasma antioxidant activity is linked to larger infarcts and more severe neurological impairment in stroke (Leinonen et al. 2000). This suggests that complex antioxidant defenses are overwhelmed in cerebral ischemia/reperfusion injury. Consistent with this hypothesis, there is some evidence that supplementation or pre-treatment with endogenous antioxidants (Van der Worp et al.1998; Huang et al. 2001), or other antioxidant free radical scavengers (Cao and Phillis 1994; Shichinohe et al. 2004), prior to focal cerebral ischemia can be neuroprotective in some established animal models of stroke, indicating that this may be an appropriate target area for novel and much needed therapeutic development (McMahon et al. 2006, Behl and Moosmann 2002).
The polyphenolic antioxidant di-tert-butyl-bisphenol (BP) was designed as a superior radical scavenging agent following structure/function analysis of more than 60 natural and synthetic (poly)phenols (Witting et al.1996, 1999a). BP is an efficient ROS scavenger in vivo and prevents arterial lipid oxidation in animal models of atherosclerosis (Witting et al.1999a,b). Interestingly, BP was identified as a metabolite of the lipid-lowering phenolic compound Probucol (Barnhart et al. 1989). The generation of BP from Probucol involves the oxidation of the latter by one-electron (Barnhart et al. 1989) and two-electron oxidant species (Witting et al. 2005). We have demonstrated previously that BP supplementation significantly limits ROS generation, modulates acute phase gene responses, enhances cellular energetics, and improves cell viability following hypoxia/re-oxygenation injury in cultured human neuronal cells (Rayner et al. 2006; Duong et al. 2008).
The aim of this study was to extend these cellular studies to test whether dietary supplementation of BP ameliorates damage associated with focal cerebral ischemic injury in an animal model of experimental stroke, thereby providing proof of concept that the synthetic antioxidant BP can be neuroprotective.
Materials and methods
Male Wistar rats (90–108 g) were from the Animal Resources Centre (Perth, Australia) and acclimated for 1 week before use. Animals were housed at constant temperature (20 ± 2°C) and relative humidity (38 ± 3%) with 12 h light/dark cycles and free access to chow and water. Studies were performed according to national guidelines and approved by the Sydney South West Health Service Animal Welfare Committee (approval #2006/008) and reported according to ARRIVE guidelines.
Animals were randomly assigned to a surgical sham group (n = 25), a vehicle-supplemented chow group (control, n = 26), and a group pre-treated with 0.12% w/w dietary BP (n = 27). This level of dietary BP was selected from pilot studies that optimized the dosage to achieve ~ 50 μM circulating BP (not shown). In rabbits a similar level of dietary BP yielded ~ 100 μM BP in the circulation (Witting et al. 1999b). Standard rat chow was soaked with vehicle (Ethanol; 99.7%) or ethanolic solutions of BP (Polysciences, Warrington, PA, USA) at 0.12% BP w/w chow. Treated chow was dried in a fume hood (7 days) to remove residual vehicle and both control or BP-treated chow was stored (4°C) until used. The surgical sham group received standard, untreated chow. Dietary supplementation continued for 4 weeks with monitoring of food and fluid intake, body weight and wellbeing; chow and water were provided ad libitum.
Animal model of experimental stroke
A surgical model of focal cerebral ischemia and reperfusion, referred to hereafter as the middle cerebral artery (MCA) occlusion, was routinely applied (Anderson and Sims 2002). The sham-operated group was used to control for the surgery to expose the carotid artery, although these animals were not exposed to MCA occlusion. The vehicle-only supplemented chow group was utilized to directly compare the effects of experimental stoke with the animal group treated with 0.12% w/w BP in the diet.
After 4 weeks feeding, animals (body weight 390–400 g) were anesthetized with 5% v/v isofluorane (Veterinary Companies Australia, Sydney, Australia), intubated and maintained with mechanical ventilation on 2% v/v isofluorane. The area around the carotid region was made sterile and systemic analgesia (Rimadyl; 1 mg/100 g weight) was administered. The common carotid artery was exposed via a midline incision and the MCA occluded using a monofilament suture (4-0 Nylene, Dynek, Sydney, Australia) coated with poly-l-lysine and blunted at the tip. The blunt-end monofilament was inserted through the external carotid artery and advanced ~ 22 mm to the MCA before being tied in place and the wound sutured. Animals were then transferred to isolation cages with heat mats and monitored for recovery and checked against benchmark criteria (anti-clockwise gait and loss of righting reflex). Animals that failed to meet these criteria (~ 12%) were excluded from the study.
After 2 h of MCA occlusion, animals were re-anesthetized with 5% v/v isofluorane and the monofilament was withdrawn to induce a cycle of ischemia reperfusion injury. The wound area was treated with Bupivicane (0.5% v/v in phosphate-buffered saline) and the animals were recovered in heated isolation cages (appropriate chow/water provided ad libitum). After 24 h, the animals were re-anesthetized with 5% v/v isofluorane and administered analgesia before organ harvest. Blood samples were collected via the tail vein into heparin-coated tubes. A thoracotomy was performed to expose the heart and the vasculature was perfused with phosphate-buffered saline (50 mM, pH 7.4; 80 mm Hg) for 5 min via the left ventricle. Brain tissue was isolated, sectioned and stored at −80°C for biochemical or molecular studies, or fixed in 10% v/v formalin (30 min), washed, stored in ethanol (99.7%) and kept at 4°C for histologic studies.
Blood was centrifuged to obtain plasma as described previously (Witting et al. 1999b). Analysis of plasma hemoglobin (tHb), bicarbonate ion (HCO3−), sodium and potassium ions, urea, and creatinine were performed in a single batch assay by the Biochemistry Department, Concord Hospital (Sydney, Australia).
For analysis of lipids, vitamin E (α-TOH), BP and its oxidation product tert-butyl diphenoquinone (DQ), frozen brain tissue from the right parietal cortex was thawed at 20°C then homogenized in phosphate buffer (50 mM, pH 7.4) containing 25 μM butylated hydroxytoluene (Merck, Sydney, Australia), 1 mM ethylenediaminetetraacetic acid (Sigma, Sydney, Australia) and a protease-inhibitor cocktail (Roche, Mannheim, Germany) using a rotating piston as described previously (Witting et al.1999b). Samples of homogenate were added to hexane:methanol (5 : 1; v/v) to extract lipid-soluble components (Witting et al.1998) that were subjected to reversed-phase (RP) HPLC to quantify non-esterified cholesterol (FC), esterified cholesterol (CE), oxidized CE (referred to as CE-O(O)H), and α-TOH (Witting et al.1998).
Assessment of BP and DQ in lipid extracts was determined by RP-HPLC using a LC18 column (Witting et al. 1999b) eluted with 100% Solvent A (methanol/acetonitrile/H2O; 10 : 10 : 3 v/v/v, 1.5 mL/min) for 10 min monitored at 270 nm, followed by 100% Solvent B (methanol/acetonitrile; 1 : 1 v/v, 1.5 mL/min) for 20 min monitored at 214 nm, BP eluted at 16 and DQ at ~ 25 min. Analytes were quantified by peak area comparison with standards. Homogenate protein levels were determined with the bicinchoninic acid protein assay (Sigma) and used to normalize all biochemical parameters.
Infarct area analysis
Brains were sectioned sequentially in 2 mm slices and stained in 3% 2,3,5-triphenyltetrazolium chloride for 15 min. The posterior surface of each triphenyltetrazolium chloride-stained section was digitally imaged and infarct area calculated using Image J (v1.4; Freeware, NIH, Bethesda, MD, USA). The right side of the brain was consistently analyzed as this represented the area predominantly affected by the MCA occlusion procedure. Infarct area was then determined in the second section from the midline and expressed as a percentage of the total area of the right side of the brain.
Direct blood pressure measurement
Blood pressure was measured using a separate group of male Wistar rats with identical mean age to animals designated for experimental stroke and fed vehicle- or BP-treated diets. Mean arterial pressure (MAP) was measured in anesthetized rats via cannulation of the common carotid artery. An incision was made to expose the left external carotid artery, a cannula inserted and a catheter secured to this cannula. A three-way valve was connected, which separated a syringe containing sterile saline with 25 unit of heparin, from a pressure transducer connected to a multi-channel Powerlab Chart with amplifier (ADInstruments, Sydney, Australia). Data were analyzed using Chart (v5.1, ADInstruments) and the MAP was determined.
Acute phase response gene expression
Total RNA was extracted from homogenates using a GeneElute kit (Sigma). cDNA was constructed by reverse transcriptase–polymerase chain reaction (RT-PCR) using SuperScript II RT (Invitrogen, Sydney, Australia) and a GeneAmp Thermo-Cycler system (Applied Biosystems, Foster City, CA, USA). Reactions containing total RNA (2 μL), oligo (dT) (1 μL) and diethylpyrocarbonate (DEPC)-treated Nanopure water (9 μL) were denatured (70°C, 5 min), chilled (4°C 5 min) then treated with 1 μL RNaseOut, 1 μL (dNTP Mix; Proligo), 4 μL of 5x reaction buffer, 1.75 μL DEPC-treated water, and 0.25 μL Bioscript (Bioline, Sydney, Australia). Samples were heated (42°C, 60 min) and the RT-reaction was stopped by increasing temperature to 70°C (10 min). Transcribed cDNA was stored at −20°C for prior to use.
Quantitative RT-PCR was performed using the Gene2000 Real-Time PCR system (Corbett Research, Sydney, Australia) and SYBR Green PCR Master Mix (Invitrogen). cDNA samples (2 μL) were mixed with SYBR Green Master Mix (12.5 μL), primers (1.0 μL, 5 pmol/L) as described in Table 1, and RNase-free water (8.5 μL). Cycling conditions were as for RT-PCR. The average output of three independent samples was reported and gene expression was quantified by the comparative Ct method. Target gene expression was normalized to β-actin and expressed as a fold-change relative to the corresponding control (arbitrarily assigned a value of 1).
Table 1. Sequences of primers used for mRNA detection by RT-PCR
Primers were designed using a BLAST search of protein databases and converting sequence data to the primary DNA sequence. Primers were designed using this consensus sequence with specific fragments between 100 and 200 bp chosen for each of the genes of interest.
SDS–PAGE and western blot analysis
Levels of cerebral hypoxia inducible factor (HIF)-1α protein were determined with sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) combined with Western blot analysis. Briefly, homogenates were centrifuged (1500 g, 4°C, 10 min) and the supernatants normalized for protein. Samples (20 μg) were denatured at 95°C in the presence of SDS, dithiothreitol and β-mercaptoethanol, loaded onto 8% v/v SDS–PAGE and resolved by electrophoresis.
Proteins were transferred to nitrocellulose membranes and non-specific sites were blocked by incubation (4°C) in Tris-buffered saline containing 0.1% v/v Tween 20 and 5% w/v skim milk powder. Membranes were incubated with mouse anti-HIF-1α mAb (Sigma; 1 : 200 v/v dilution, 12 h, 4°C), washed (three times, 5 min), and incubated with horseradish peroxidase-linked secondary antibody (1 : 2000 v/v, Sigma). After 4 h and consecutive washes (three times Tris-buffered saline containing 0.1% v/v Tween 20), immune-reactive proteins were detected using a commercial kit (iNtRON Biotechnology, Songdo, Korea) and quantified with densitometry (BioDocAnalyzer, Biometra, Gottingen, Germany).
Tissue ATP concentration
Tissue ATP content was determined with a commercial kit (ATPlite®; Perkin Elmer, Melbourne, Australia). Briefly, tissue homogenates (20 μL) were treated with kit lysis buffer (130 μL) and gently mixed in an orbital shaker at 386x g for 5 min. Lysates were then treated with substrate solution (50 μL), agitated for 5 min and incubated in the dark for 10 min. ATP-dependent luminescence was determined using a Victor III Multi-label plate reader (Perkin Elmer) and normalized to total homogenate protein content.
Caspase 3/7 activity
The activity of caspase 3/7 in brain tissue homogenate was determined using a commercial kit (Caspase-Glo™ 3/7 Assay; Promega, Sydney, Australia). Aliquots of brain homogenate (20 μL) were added to plates, treated with Caspase-Glo™ 3/7 reagent (100 μL) and gently mixed (30 s, plate shaker). Plates were then incubated (4°C, 40 min) and caspase 3/7 activity assessed by measuring luminescence output using a Victor III Multi-label plate reader (Perkin Elmer). Caspase 3/7 activity was expressed as a fold-change compared with the activity measured in the corresponding controls.
Assessment of the thiol redox ratio
Total tissue thiol and corresponding oxidized disulfide products were determined in cerebral homogenates. Homogenates were incubated (37°C. 10 min) with 100 μM 5,5′-dithiobis(2-nitrobenzoic acid) (DTNB, Sigma). A matched sample was mixed with 10 units/mL glutathione reductase and 100 μM NADPH and incubated at 37°C to reduce tissue disulfides. After 1 h, these samples were incubated with 100 μM DTNB (Sigma) and thiol levels measured (416 nm) with a multi-label Victor III plate reader. Total thiol was normalized to cell protein, whereas levels of oxidized thiols were calculated from the difference between values of thiol in the matched samples and expressed as: [Total thiol]/([total thiol]+[oxidized thiol]).
Brain tissue was dehydrated, cleared with organic solvents, and infiltrated with paraffin using an automated tissue-processing machine (TP1020; Leica, Sydney, Australia). Serial sections (5 μm) were cut with a microtome (RM2125; Leica), dewaxed, and rehydrated (95% v/v alcohol). Sections were immersed in 0.1% w/v Luxol fast blue solution (60°C for 2 h) followed by 2 min differentiation in saturated lithium carbonate. Sections were then counterstained in 0.5% v/v cresyl violet solution (2 min), washed, and differentiated in acid alcohol for 8 s. Slides were then washed, dehydrated, cleared, and mounted in dibutyl phthalate xylene for visualization with light microscopy (Olympus microscope fitted with a digital camera; DP Controller; v184.108.40.206, Olympus, Melbourne, Australia) and converted to JPEG for handling with MS PowerPoint (v7, Microsoft, Redmond, WA, USA).
Formalin-fixed, paraffin embedded specimens were sectioned (thickness 5 μm) and labeled with a rabbit anti-mouse 4-Hydroxynonenal (4-HNE) antibody (BIOSS, Boston, MA, USA). The 4-HNE antibody detects proteins that have been chemically modified by aldehydic products of lipid peroxidation. Briefly, specimens were dewaxed, rehydrated and treated by heat antigen retrieval (pH 9) using a commercial retrieval solution (Dako, Glostrup, Denmark). After peroxidase blockade, slides were incubated with anti-4-HNE (1 : 100 v/v) or an isotype control (rabbit IgG, Dako) for 1 h. Sections were labeled with an anti-rabbit secondary antibody (Sigma) and developed with diaminobenzidine (Dako, Sydney, Australia). Slides were dehydrated, cleared, and mounted in dibutyl phthalate xylene for visualization with an Olympus light microscope fitted with a digital camera (DP Controller; v220.127.116.11, Olympus) and converted to JPEG for handling with MS PowerPoint (v7, Microsoft).
Apoptotic cell death was detected using DeadEndTM Colorimetric TUNEL system (Promega) according to the manufacturer's instructions. Briefly, slide sections (5 μm) were deparaffinized in Histoclear, rehydrated and treated with NaCl (0.85% w/v, 5 min) then 4% w/v paraformaldehyde, washed, and permeabilized with Proteinase K. Sections were then equilibrated and incubated with Biotinylated Nucleotide mix and recombinant Terminal Deoxynucleotidyl Transferase (rTdT) at 37°C. After 1 h, the reaction was stopped and endogenous peroxidases blocked with H2O2. Sections were incubated with Streptavidin-horseradish peroxidase conjugate solution and color developed with a diaminobenzidine substrate. Alternate sections were processed in parallel without rTdT as controls. Sections were visualized with light microscopy and images captured digitally and analyzed using Image J software as described above.
Aortae were harvested from male Wistar rats (~ 0.20–0.3 kg) fed normal rodent chow via a thoracotomy procedure to expose the chest cavity. Prior to removal, the aortae were perfused (phosphate buffer pH 7.4, 80 mmHg), carefully dissected and placed in modified Krebs–Henseleit solution (in mM: 11 D-glucose, 1.2 MgSO4, 12 KH2PO4, 4.7 KCl, 120 NaCl, 25 NaHCO3, 100 L–arginine, and 2.5 CaCl2·2H2O) and cut into 3 mm ring segments, mounted in a myobath system (World Precision Instruments, Inc., Sarasota, FL, USA) bathed in 20 mL of modified Krebs–Henseleit solution aerated at 37°C with Carbogen (5% CO2) and contracted by titrating with phenylepherine as described in detail elsewhere (Rayner et al. 2005). A dose that caused half maximal contraction was selected for all further studies using the same ring segment.
A concentration–response curve to acetylcholine (ACh) was constructed, with relaxation expressed as percentage of initial constriction. Prior to use, rings were pre-treated with H2O2 (50 μM) or 200 μM 1H-[1,2,4]oxadiazole[4,3-a]quinoxalin-1-one (ODQ, an inhibitor of soluble guanylyl cyclase that prevents vascular relaxation) for 60 min followed by washing, pre-constriction (with phenylepherine, at the dose that yielded half-maximal constriction) and monitoring relaxation to added ACh. Other rings were pre-treated with BP (final concentration 50 μM) before addition of H2O2 (50 μM) followed by pre-constriction and assessment of vascular relaxation. Under pathophysiological conditions concentrations of H2O2 up to 100 μM are reported in rat striatal tissue following forebrain ischemic injury (Hyslop et al. 1995) therefore, the dose used here is within the pathologically relevant range.
Differences between groups were assessed with one-way anova employing the Newman–Keuls multiple comparisons post hoc test. Significance was accepted at the 95% confidence level (p < 0.05).
Dietary supplementation with BP
Pilot studies in rats pre-treated with 0.12% wt/wt BP in the diet showed circulating BP concentrations of 43.3 ± 15.4 μM (mean ± SD; n = 3 rats) while its oxidation product DQ was undetected. These data indicate that 100% of the circulating polyphenol was available in the antioxidant active form. A similar analysis of brain homogenates revealed that BP was present at 130 ± 50 pmol/mg protein, whereas the yield of DQ was 60 ± 20 pmol/mg protein indicating that the drug entered cerebral tissues and that ~ 70% of the polyphenol was present in the antioxidant active form.
The dietary uptake of BP was initially examined in the absence of experimental stroke to observe any adverse physiological reactions to the drug. Overall, animals tolerated 0.12% wt/wt BP in the diet with no marked change in body weight, plasma biochemistry, or renal function compared with animals receiving the control diet (Table 2). Overall, there were no significant changes in the plasma levels of total hemoglobin (tHb), bicarbonate (HCO3−), sodium (Na+), and potassium (K+) ions (Table 2). Only creatinine levels were significantly decreased in BP-treated animals (Table 2). Although not significant, urea levels tended to be lower in the BP pre-treated animals, however, the urea:creatinine ratio was similar in both groups (Table 2) suggesting that renal function was largely unaffected in BP supplemented animals.
There was no significant difference in MAP between the BP-supplemented and control animals (Fig. 1). These data indicated that BP was unlikely to confound through a mechanism involving blood pressure lowering.
Table 2. Dietary supplementation with BP (0.12% w/w diet) in rats does not affect renal physiologya
Animals supplemented with vehicle-treated chow (Vehicle control) or chow-supplemented with BP (0.12% w/w in the diet) for 4 weeks were killed and samples of whole blood were obtained. These isolated samples were immediately centrifuged to obtain plasma as described previously (Witting et al. 1999a) and the biochemical parameters listed were measured as described in detail in the Methods section. All work with animals was performed with approval of the local Ethics Committee. Data presented as mean ± (SD); n = 4 except weight where n = 31 (control) or 35 (BP group).
Experimental stroke induction in BP-supplemented animals
To study the effects of BP supplementation on experimental stroke, a MCA occlusion model was employed to mimic focal cerebral ischemia. To test the neuroprotective efficacy of the polyphenolic antioxidant, BP- or vehicle-supplemented (control) animals were subjected to MCA occlusion for 2 h followed by removal of the occluding monofilament and 24 h reperfusion. Brain infarcts in these two animal groups were compared with a sham-operated group (Fig. 2a). MCA occlusion followed by 24 h reperfusion caused focal cerebral damage that extended to the caudate putamen and the parietal cortex area 1 and 2. In the absence of BP, the MCA occlusion/reperfusion insult induced a ~ 5-fold increase in infarct area compared with the sham (infarct area 6.7 ± 2.4 vs. 34.5 ± 6.7%; mean ± SD, respectively) (Fig. 2b). Despite the observation that the sham-operation caused some minor cerebral damage (Fig. 2b), these animals exhibited no obvious neurological symptoms post-surgery. In contrast, the size of the infarct in the brain of BP-supplemented animals decreased by 27% (p < 0.05) compared with vehicle-treated controls (Fig. 2b). Although, BP-treatment did not diminish the infarct to the level observed in the sham-operation group (Fig. 2b).
Oxidation in brain tissue after MCA occlusion
The effect of oxidative stress in the ischemic brain following MCA occlusion and reperfusion was examined initially by measuring tissue antioxidant and oxidized lipid content (Table 3). Compared with the sham-operated group, the concentrations of α-TOH, free cholesterol and CE were similar in brain tissue from the control animals after MCA occlusion and reperfusion (Table 3). Importantly, the formation of oxidized CE (CE-O(O)H), and also its ratio relative to the total CE (native parent lipid), increased significantly, up to 10-fold, after experimental stroke (Table 3). The increased cerebral lipid oxidation in control animals was reduced significantly in BP-supplemented animals. In contrast to CE-O(O)H, α-TOH levels did not alter following MCA occlusion procedure or with BP supplementation and were similar across all animal groups (Table 3).
Table 3. Dietary supplementation with BP affects the level of oxidative damage in the ischemic braina
Animals supplemented with normal chow (Sham-operated), vehicle-treated chow (Vehicle control) or chow supplemented with BP (0.12% w/w in the diet) for 4 weeks were subjected to experimental stroke (2 h MCA occlusion, 24 h reperfusion) or sham-operation then euthanized and samples of brain tissues were obtained. Isolated tissue samples were processed as described previously (Witting et al. 1999a,b) and the biochemical parameters listed were analyzed by liquid chromatography as described in detail in the Methods section. Data are expressed as mean ± (SD). FC = free cholesterol; α-TOH = α-tocopherol (biologically active vitamin E); CE = cholesteryl esters (the sum of the polyunsaturated lipids: cholesteryl linoleate, C18:2 and cholesteryl arachidonate, C20:4); CE-O(O)H = CE-derived lipid hydroperoxides and hydroxides. The ratio of CE-O(O)H/CE represents the proportion of oxidized lipid per parent lipid.
Units of measurement and the numbers of samples tested (n) for all parameters are as indicated.
Different from the Sham-operated group; p < 0.05.
Different from the control group, p < 0.001. ND = not detected.
To assess whether oxidant sensitive water-soluble substrates were susceptible to oxidation we compared the homogenate levels of reduced and oxidized thiols in the tissue samples and expressed this as the total thiol redox ratio (Fig. 3a). After MCA occlusion the total thiol redox ratio was significantly lower than the corresponding value in the sham-operated group indicative of increased oxidative stress in these tissues. Supplementing BP prior to experimental stroke markedly improved the total thiol redox ratio returning this index of oxidative stress to levels similar to the sham-operated group.
The presence of oxidized proteins was determined to complement the assessment of lipid and thiol oxidation after MCA occlusion (Fig. 3b). Low background levels of 4-HNE modified proteins were detected in the sham-operated group when imaging the same cortical region where focal infarcts were detected in vehicle-treated controls (Fig. 3b, panel i). This data indicated that the surgical procedure alone stimulated oxidative stress, which resulted in lipid oxidation (Table 3) and the formation of 4-HNE modified proteins. However, the level of immune-positive staining increased markedly in the vehicle-treated controls particularly in the penumbral region surrounding the infarct core (Fig. 3b, panel ii). Supplementation of BP before MCA occlusion decreased immune-positive staining in this peri-infarct region compared with the vehicle-treated control although this remained slightly higher than in the sham samples (Fig. 3b, compare panels i and iii). Notably, no staining was apparent in the isotype control (Fig. 3b, panel iv).
These data are consistent with focal cerebral ischemia enhancing oxidative damage in the affected brain region. Furthermore, the data indicate that oxidation of substrates in both lipid and water-soluble environments can be effectively ameliorated with BP and that dietary supplementation with BP augments the antioxidant capacity of brain without altering vitamin E or lipid status.
Gene expression in response to focal ischemia reperfusion injury
Cerebral ischemia adversely affects oxygen homeostasis and induces a hypoxic challenge in tissue. An important and early response gene activated by hypoxia encodes the transcription factor, HIF-1α. Expression of HIF-1α was examined in brain tissue following MCA occlusion/reperfusion (Fig. 4). HIF-1α gene expression increased 2-fold in control animals compared with the sham-operated group (Fig. 4a). A detectable amount of HIF-1α mRNA and protein occurred even in the surgical sham group (Fig. 4a–c), suggesting that either a basal expression of HIF-1α or that the surgical procedure to isolate MCA without occlusion/reperfusion induced HIF-1α. Notably, BP-supplementation restored HIF-1α mRNA levels to those observed in the surgical sham group (Fig. 4a). Consistent with the mRNA data, HIF-1α protein tended to increase in control animals following MCA occlusion/reperfusion (Fig. 4b). However, the increase in the level of HIF-1α protein was not significant and was unaffected by BP supplementation (Fig. 4c). The data indicate that changes in HIF-1α gene expression did not result in significant concomitant changes in its protein level over the time range tested.
Maintenance of oxygen homeostasis by HIF-1α is thought to involve activation of other acute response genes against hypoxia. We therefore examined the changes in expression of the acute response genes; hemeoxygenase-1 (HO-1), glucose transporter-1 (Glut-1), superoxide dismutase (SOD1), and glutathione peroxidase (GPx-1) (Fig. 5), as the products of these genes are thought to mediate adaptive responses to oxygen imbalances in tissue. The gene expression of HO-1 (Fig. 5a), Glut-1 (Fig. 5b), and SOD1 (Fig. 5c) in brain homogenates all significantly increased (up to 10 times) following experimental stroke. The increases in Glut-1 and SOD1, but not HO-1 gene expression, were significantly abated by BP supplementation (Fig. 5a–c). Notably, GPx-1 gene expression decreased significantly following experimental stroke and this decrease was reversed with BP supplementation (Fig. 5d). The data suggest that hypoxia-related acute response gene expression induced in brain tissue is modulated by BP supplementation.
Attenuation of hypoxia-induced low energy levels and neuronal cell apoptosis
After ischemia reperfusion insult, mitochondria may become dysfunctional which can adversely affect ATP bioavailability. Compared with the sham-operated group, the concentration of protein-normalized ATP decreased markedly in damaged rat brain following MCA occlusion (Fig. 6a). However, ATP levels improved significantly with BP pre-treatment indicating that BP is protective against hypoxia-induced tissue energy dysfunction (Fig. 6a).
Both dysfunctional energy metabolism, and enhanced oxidative stress, can facilitate neuronal cell apoptosis after transient focal cerebral ischemia. The activity of pro-apoptopic caspases (3/7) in brain tissue were examined as an indication of increased apoptosis in tissue following MCA occlusion. Caspase 3/7 activity was markedly enhanced in the control animal group after cerebral ischemia compared with the sham (Fig. 6b). Significantly, BP pre-treatment effectively inhibited this activation of pro-apoptopic caspases in experimental stoke suggesting that the polyphenol may be useful in enhancing tissue viability in the area-at-risk by limiting apoptosis in the peri-infarct zone surrounding the primary infarct (Fig. 6b).
Brain tissue was further examined histochemically for apoptosis and necrosis in neuronal cells following MCA occlusion (Fig. 6c). Co-staining with Luxol fast blue and Cresyl violet depicted edema, increased neuronal pyknosis with cellular debris and focal areas of myelin aggregation in the border zone of peri-infarcted tissues from the vehicle-treated control compared to the corresponding tissue region in brains from the sham-operated group (Fig. 6c). Also, Cresyl violet-positive neurons tended to contain fewer nucleoli in the vehicle-treated control group than the sham-operated group. In the corresponding tissues from BP pre-treated animals less myelin aggregation was evident in the border zone of the peri-infarcted tissue and more neurons tended to contain neucloli, although the extent of edema was unaffected (Fig. 6c).
The presence of apoptopic nuclei in cells, detected by TUNEL labeling, was evident in the vehicle-treated control group after MCA occlusion/reperfusion and this was markedly ameliorated in the BP-supplemented group (Fig. 6c). Only weak TUNEL labeling was detected in the surgical sham group (Fig. 6c). Together the data indicate that development of neuronal cell damage and apoptosis following experimental stroke was significantly inhibited by pre-treatment with BP.
To assess whether the polyphenolic antioxidant BP protects the vascular endothelium from oxidative insult, segments from isolated rat aortic vessels were exposed to H2O2, in the presence and absence of BP, and vascular response to vasodilating ACh monitored. Consistent with previous reports (Thomas et al. 2006; Witting et al. 2007), exposing vessels to H2O2 in the absence of BP diminished vascular relaxation to ACh (Figure S1). As expected vascular rings incubated with the soluble guanylyl cyclase inhibitor ODQ showed little response to ACh. In contrast, vessels pre-treated with BP showed a similar vaso-response to the control suggesting that BP, added at a concentration that matched plasma levels in BP-supplemented animals, inhibited H2O2-mediated oxidative insult and preserved vascular relaxation to ACh stimulus.
In this study, we demonstrated that BP, a synthetic antioxidant with greater radical scavenging activity than vitamin E [as judged by its lower oxidation potential (Shanu et al. 2010)], permeates the blood–brain barrier and is predominantly in its antioxidant (phenolic) active form. Dietary supplementation with BP did not affect blood pressure or the levels of endogenous vitamin E in either plasma or cerebral tissue. Increasing the antioxidant capacity of cerebral tissue with BP prior to experimental stroke significantly inhibited oxidative stress and the accumulation of lipid- and water-soluble biomarkers of oxidative damage. Infarct size decreased in the brain following MCA occlusion as a model of focal ischemic stroke. Inhibition of oxidative stress and infarct size by BP supplemented before experimental stroke was accompanied by a down-regulation of hypoxia-sensitive gene expression, preservation of cerebral ATP levels and attenuated neuronal apoptosis. In addition, BP reversed H2O2-mediated endothelial dysfunction in an ex vivo bioassay for vascular reactivity. The cumulative data suggest that BP has the potential to enhance the viability of cerebral tissue in the setting of experimental stroke through increasing neuronal tissue viability, as suggested by our earlier studies with neuronal cells in vitro (Rayner et al. 2006; Duong et al. 2008), and potentially maintaining vascular perfusion. Overall, our data suggest that supplementation with BP may provide neuroprotection through multiple mechanisms.
Our data showing that BP pre-treatment ameliorates oxidation during focal ischemia and reperfusion insult and affords neuroprotection are consistent with the notion that oxidative stress promotes brain damage following ischemic stroke (for review see Sugawara and Chan 2003). Oxidative stress may be a key factor underlying cerebral damage in stroke, and reperfusion after ischemia has the potential to increase ROS generation and aggravate tissue damage. There is substantial evidence for the production of ROS in the brain during acute cerebral ischemia/reperfusion (Kawase et al. 1999; Morita-Fujimura et al. 2001) leading to the expansion of damage from the initial focal event (Chen et al. 2009). On balance the majority of literature suggests that modulation of antioxidant genes that enhance antioxidant capacity is linked to neuroprotection in animal models of stroke (Kinouchi et al. 1991; Kondo et al. 1997; Doré et al. 1999; Davis et al. 2007).
We have demonstrated that BP supplemented prior to stroke successfully inhibits markers of oxidative damage. For example, the extent of lipid oxidation, the depletion of tissue thiol content and the formation of protein carbonyl products (generated by reaction between oxidation susceptible amino acid residues in proteins and 4-hydroxynonenal) were all ameliorated in animals pre-treated with BP. Similarly, the endogenous antioxidant gene response to oxidative insult was blunted in the presence of BP. Pre-treatment with BP also significantly decreased infarct size consistent with other studies that assessed the neuroprotective efficacy of antioxidant proteins (reviewed in Sugawara and Chan 2003) and free radical scavengers (reviewed in Margaill et al. 2005) in animal models of MCA occlusion. However, it remains possible that oxidative stress may not be the sole contributor to cerebral damage in stroke. This may explain the discrepancy between inhibition of oxidative stress and a relatively modest translation of antioxidant action to preventing neuronal tissue damage. In addition, BP may be differentially effective against the types of oxidants involved in generating oxidative stress that may be pertinent in stroke.
The types and extent of ROS generated in stroke are not fully understood though there is a consensus that superoxide radical anion (O2˙−) is involved together with oxidants linked to its metabolism such as H2O2. Mitochondria are a significant source of these ROS especially in reperfusion after ischemia where oxygen intake is high. Even under physiological conditions, respiratory chain generated O2˙− dismutation to H2O2 (via SOD2) may be as high as 2% of electron flow (Dikalov 2011). In addition to mitochondria, NADPH oxidases (NOX) are a significant source of O2˙− in stroke (Radermacher et al. 2013) and a recent study shows reduced stroke mortality and neurodegeneration in NOX4 gene knockout animals (Kleinschnitz et al. 2010). Interestingly, there may be significant interplay between the ROS-generating systems of mitochondria and NOX with inter-activation of these systems leading to uncontrolled ROS production (Dikalov 2011).
Whether BP can inhibit mitochondrial and/or NOX-generated ROS and whether this leads to decreased oxidative stress is not clear, although here we have demonstrated a substantial decrease in lipid and thiol oxidation (that translates to decreased protein oxidation and activation of pro-apoptotic proteases) in the brains of BP-treated animals. In addition, BP supplementation inhibits the accumulation of O2˙− in cultured neuronal cells exposed to hypoxia re-oxygenation injury (Rayner et al. 2006; Duong et al. 2008), suggesting that this polyphenol may regulate cellular pathways involved in O2˙− generation.
Despite the prevailing view that prevention of oxidative injury by quenching ROS may be therapeutically beneficial, the efficacy of antioxidant therapy remains unsubstantiated in stroke. While antioxidant therapy for stroke in animal models has been largely successful, translation to large-scale human clinical trials has been disappointing (Margaill et al. 2005; Shuaib et al. 2007). The success of some antioxidant therapies in stroke models suggests that these compounds may have neuroprotective activities that are unrelated or additive to antioxidant activity. For example, in addition to its antioxidant activity in the ischemic brain (Yamagata et al. 2010), vitamin E has been reported to have anti-inflammatory (Guo et al. 2007) and anti-proliferative activities (Azzi 2007) that may be important in suppressing cerebral tissue damage in stroke.
Regulation of redox sensitive transcription factors and signal transduction pathways may be an important consideration in limiting neuronal damage after stroke. For example, activation of the HIF pathway is generally considered to be neuroprotective as the expression of genes linked to this pathway, including Glut-1, act to restore glucose and oxygen delivery to optimize cell survival following ischemic insult. Inactivation of HIF-1α in neurons exacerbates transient focal ischemic injury in the mouse brain (Baranova et al. 2007) and stabilization of HIF-1α signaling affords protection against cerebral ischemia and enhances cell viability (Siddiq et al. 2007). Also, hypoxic pre-conditioning significantly enhances neuroprotection via induction of HIF-1α and downstream genes (Liu et al. 2005). That BP supplementation effectively reduced HIF-1α in brain tissue following experimental stroke and normalized hypoxia-related gene expression of Glut-1, GPx-1 and SOD, suggests that ROS are also involved in promoting HIF-1α expression and that scavenging of ROS attenuates the response to hypoxia. Our results with BP are similar to another study showing a SOD-induced reduction in HIF-1α expression in hypoxic pre-conditioned neurons (Liu et al. 2005).
It is evident that HO-1-mediated degradation of protein-free heme to yield products with antioxidant (bilirubin) and vasodilatory (carbon monoxide) properties is linked to neuroprotection in the ischemic brain (Zeynalov et al. 2009). In contrast to HIF-1α, BP supplementation had little effect on HO-1 gene expression in experimental stroke. While HO-1 expression can be linked to HIF-1α, there are other independent, antioxidant response pathways that induce HO-1, such as the Nrf2-KEAP system (Satoh et al. 2006). It is also possible that BP stimulates HO-1 expression via the Nrf2-KEAP system and that this overrides any dampening in HO-1 by ROS attenuation in the presence of BP. This is supported by studies that show HO-1 is fundamentally required for neuroprotection against cerebral ischemia afforded by Gingko biloba (EGb 761), a compound that has ROS scavenging properties (Shah et al. 2011).
Cumulative evidence implicates oxidative stress in promoting cell apoptosis responsible for neuronal cell death and expansion of damage to the peri-infarct zone after focal ischemia (Matsuda et al. 2009). For example, decreases in normal-nucleated neurons and increased levels of shrunken/atrophic neurons and significant DNA strand breakage in apoptopic neurons are observed after focal ischemia. This is accompanied by decreases in SOD1 (Ji et al. 2012) and SOD2 (Bonova et al. 2013) protein and activity within the infarct core. A depletion of SOD activity in the infarct core suggests that an imbalance in O2˙− clearance leading to oxidative injury may promote DNA oxidation, neuronal cell apoptosis and expansion of cerebral damage (Nagayama et al. 2000). By contrast, regional differences in the expression of SOD2 are evident between the infarct core and peri-infarct zone with the penumbra showing marked (~9-fold) increases in SOD2 content after experimental MCA occlusion (Bonova et al. 2013). Our data demonstrating increased SOD1 mRNA in the homogenized brain samples likely reflect changes in the penumbra region surrounding the primary infarct.
The data presented herein agree with evidence that apoptosis induced by oxidative stress plays a key role in neuronal cell death (Kondo et al. 1997; Morita-Fujimura et al. 2001; Noshita et al. 2003). The hallmarks of apoptosis are nuclear condensation, cell shrinkage, and chromosomal DNA fragmentation. Oxidative stress may act directly on cells to initiate apoptosis or indirectly via enhancing the production of pro-apoptopic factors such as caspase-3 and -7 within mammalian cells. The MCA occlusion model used herein induced a significant increase in cerebral caspase 3/7 activity that was accompanied by shrinkage of, and shrunken nuclei in, neuronal cells with evidence of tissue edema. The parallel increase in TUNEL staining is consistent with DNA fragmentation and induction of apoptosis. Dietary supplementation with BP attenuated tissue caspase 3/7 activity and TUNEL staining indicating the polyphenol inhibited apoptosis after experimental stroke. Increases in viable cerebral tissue in animals pre-treated with BP may be important in limiting expansion of the primary infarct.
In addition to attenuating apoptosis, BP-treatment had a positive effect on cellular energetics. Thus, BP supplementation restored cerebral tissue ATP levels following MCA occlusion and reperfusion, indicating either mitochondrial function was preserved or ATP consumption was abated. Preservation of mitochondria function in both the infarct core and surrounding penumbra is important for limiting damage because mitochondrial dysfunction can exacerbate oxidative damage and/or influence cell death pathways (Sims and Muyderman 2010). Enhanced ATP production stimulated by BP likely provides a mechanism for neuroprotection. Also, our data demonstrating that BP can inhibit vascular dysfunction induced by H2O2ex vivo suggest that BP may provide additional protective pathways involved in re-establishing the flow of oxygenated blood to the affected ischemic tissues.
There are no available treatments to inhibit or repair neuronal cell damage sustained during stroke, and the currently available tPA treatment has severe practical restrictions limiting its usage. Pre-treatment with BP was used in this model as proof of concept that a synthetic polyphenol could abrogate focal ischemic brain damage, although such prophylactic treatment does not reflect the clinical situation of a suitable post-stroke therapy. Thus, pre-treatment with BP over 4 weeks to achieve a stable pharmacologically active level of this antioxidant in the brain is a clear limitation of this study. To address this limitation further studies with BP administered after experimental stroke as well as investigations using drug delivery vehicles to increase the rate of antioxidant uptake into the brain are warranted.
Despite the limitations of our study, BP could possibly be considered as a preventative for patients who suffer a transient ischemic attack, where the risk of future stroke episodes in the ensuing weeks is high (Wu et al. 2007). Also, BP may have some potential in combination therapy with tPA to address ROS production upon re-canalization and reduce the risk of infarct expansion to the penumbra similar to the antioxidants Ebselen (Lapchak and Zivin 2003) and Edaravone (Lapchak and Zivin 2009; Yamashita et al. 2009). Neuroprotection with antioxidants such as BP may also be beneficial to elective neurosurgery where there is a significant risk of ischemia owing to micro-vessel occlusion and possibly in neurodegenerative pathologies such as in Alzheimer's disease where oxidative stress has been implicated (reviewed in Sutherland et al. 2013). Despite these possibilities, long-term antioxidant therapy with BP is unlikely given the poor outcomes documented for vitamin E in large-scale randomized control studies (Lonn et al. 2005; Sesso et al. 2008). Although it is noteworthy, that the antioxidant activity of BP is superior to vitamin E by virtue of a lower one-electron oxidation potential ~0.31 vs. 0.48 V, respectively (Shanu et al. 2010).
Natural and synthetic polyphenols may cause oxidative damage that is toxic to cells, where the measure of cytotoxicity for a given polyphenol is related to its phenoxyl radical/phenol one-electron oxidation potential (Sergediene et al. 1999). The mechanism of cytotoxicity may involve polyphenol metabolism by cytochrome P450 (Phase 1) enzymes. Inhibitors of cytochrome P450-hydroxylase partially protect cells from polyphenol cytotoxicity (Nemeikaite-Ceniene et al. 2005). However, our assessment of renal toxicity, blood biochemistry, and animal weight indicates that the di-tert-butyl-bisphenol is well tolerated (Table 2). This lack of toxicity may be because of some polyphenols exhibiting lower than anticipated cytotoxic (LD50) dose despite one-electron oxidation potentials being < 0.4 V (Nemeikaite-Ceniene et al. 2005).
There is an urgent need for new stroke therapies. Targeting oxidative stress is promising as a mechanism-based therapy against ischemia reperfusion injury sustained during stroke and antioxidant therapies have shown some potential to be neuroprotective. This is in contrast to tPA which deals only with acute stroke treatment and potentially generates ROS and aggravates cell damage. That pre-treatment with BP attenuates oxidative damage, improves mitochondrial function and limits cell apoptosis indicates that oxidative stress plays a causal role in promoting tissue damage initiated by MCA occlusion. Together, this strongly suggests that inhibiting oxidative stress through scavenging free radicals in vivo will improve cell viability, enhance vascular function and importantly provide neuroprotection. There is certainly some scope to increase the effective dose for BP in animals and future studies will employ higher dietary intake of BP and investigate modes of rapid administration so that dosage and timing may be optimized. Such pre-clinical studies would also allow investigation of the distribution of BP within different brain regions. Whether the neuroprotective activities of BP are sufficient to inhibit functional deficits when administered in the post-stroke period is not clear. Therefore, an assessment of whether this synthetic polyphenol can limit reperfusion damage is warranted, whereas long-term follow-up of neurological outcomes are also necessary.
This work was supported by grants the Heart Foundation of Australia (G 07S30435), the Australian Research Council (DP130103711) and a University of Sydney, Medical Faculty support package. We thank Professor Neil Sims (Flinders University, School of Medicine) for his assistance in training TTH Duong and AC McMahon in the surgical procedure for MCA occlusion. The authors have no conflicts of interest to declare.