|Home | About | Journals | Submit | Contact Us | Français|
Fibrates, one group of peroxisome proliferator-activated receptor (PPAR) activators, are lipid lowering drugs. Fibrates have been shown to attenuate brain tissue injury after focal cerebral ischemia. In this study, we investigated the impact of fenofibrate on cerebral blood flow (CBF) in male wild type and PPARα-null mice. Animals were treated for 7 days with fenofibrate and subjected to 2h of filamentous middle cerebral artery occlusion and reperfusion under isoflurane anesthesia. Cortical surface CBF was measured by laser speckle imaging. Regional CBF (rCBF) in nonischemic animals was measured by 14C-iodoantipyrine autoradiography. Fenofibrate did not affect rCBF and mean arterial blood pressure in nonischemic animals. In ischemic animals, laser speckle imaging showed delayed expansions of ischemic area, which was attenuated by fenofibrate. Fenofibrate also enhanced CBF recovery after reperfusion. However, such effects of fenofibrate on CBF in the ischemic brain were not observed in PPARα-null mice. These findings show that fenofibrate improves CBF in the ischemic hemisphere. Moreover, fenofibrate requires PPARα expression for the cerebrovascular protective effects in the ischemic brain.
Fibrates are lipid lowering drugs and they are used in patients with hypercholesterolemia and hypertriglyceridemia. Fibrates activate peroxisome proliferator-activated receptors (PPARs) with a higher affinity for PPARα than that for PPARγ and PPARδ (Gibson, 1993; Dreyer et al, 1993; Green and Wahli, 1994). PPARs are ligand-activated transcription factors belonging to the nuclear receptor family. They are important in lipid and glucose metabolism (Chinetti et al, 2000). In addition, PPARs have been shown to support multiple cellular functions, such as antiinflammatory and antioxidative effects (Staels et al, 1998; Escher and Wahli, 2000; Von Knethen and Brune, 2001). Pleiotropic actions of fibrates, such as antiinflammatory action (Cunard et al, 2002; Staels et al, 1998) and antioxidative action (Escher and Wahli, 2000; Von Knethen and Brune, 2001), contribute to their protective effects against ischemia and reperfusion in peripheral organs including kidney, heart, and intestine (Sivarajah et al, 2002; Wayman et al, 2002; Yue et al, 2003; Cuzzocrea et al, 2004). Little is known about the effects of fibrates on circulation in the ischemic tissues.
Fibrates have been shown to be beneficial against inflammation-associated central nervous system injury (for review, see Bordet et al, 2006). For example, Lovett-Racke et al (2004) showed that fenofibrate and gemfibrozil improved clinical signs of experimental autoimmune encephalomyelitis in mice. Deplanque et al (2003) showed reduced expression levels of proinflammatory molecules in the ischemic brain. Moreover, fenofibrate also attenuated brain infarct size in mice subjected to focal cerebral ischemia (Inoue et al, 2003; Deplanque et al, 2003). Wy-14643, a PPARα-specific ligand, had similar effects against permanent focal cerebral ischemia (Inoue et al, 2003). The attenuation in inflammation by fenofibrate could reduce infarct size. Conversely, the attenuated inflammation could be the consequence of brain protection by fenofibrate through non-antiinflammatory actions.
Cerebral blood flow (CBF), particularly in the penumbra, is a critical determinant to outcome after ischemic stroke. Although fenofibrate does not affect CBF at the ischemic core (Inoue et al, 2003; Deplanque et al, 2003), the influence of fenofibrate on CBF in the ischemic penumbra is unknown. Moreover, the effect of fenofibrate on regional CBF (rCBF) under nonischemic conditions has not been examined. The influence on CBF could confound the interpretation of the attenuation by fenofibrate in inflammation or oxidative stress in the ischemic brain.
In this study, we measured the effects of fenofibrate on CBF in nonischemic mice and in ischemic mice that were subjected to middle cerebral artery occlusion (MCAO) and reperfusion. In addition, we examined the role of PPARα in the fenofibrate-induced increase in CBF in ischemic brain as fenofibrate failed to attenuate infarct size in PPARα-null mice after MCAO (Inoue et al, 2003; Deplanque et al, 2003). Furthermore, to gain insight into the molecular mechanism of the protection by fenofibrate, we analyzed the impact of fenofibrate on gene expression levels of PPARs and endothelial nitric oxide synthase (eNOS), which is one of the major players in CBF regulation, in the aorta and brain microvessels. We here show evidence that fenofibrate improves CBF in the ischemic brain through a PPARα-mediated mechanism.
Animal usage and all procedures were approved by the Morehouse School of Medicine Animal Care and Use Committee. Wild-type C57BL/6 and 129S6/SvEv mice (male, 6-weeks old) were purchased from Charles River Laboratories (Wilmington, MA, USA) and Taconic (Germantown, NY, USA), respectively. Homozygous PPARα-null mice (129/Sv and C57BL/6 background) with a targeted mutation in PPARα putative ligand-binding domain (Lee et al, 1995) were purchased from Jackson Laboratory (Bar Harbor, ME, USA) and our PPARα-null mouse colony was maintained by breeding homozygous males and females. Their fur coat color is agouti with white belly and they appear normal in behavior. Altered responses to peroxisome proliferators in peroxisome proliferation and hepatocarcinogenesis have been reported (Lee et al, 1995). In addition, inflammation induced by leukotriene B4 or arachidonic acid is prolonged in this mouse strain (Gonzalez, 1997). Animals had free access to water and normal diet, and were maintained under a regular day/night cycle.
Animals were treated for 7 days with fenofibrate (0, 30, or 100mg/kg) by oral administration through a feeding needle. As vehicle, 0.5% carboxymethyl cellulose (5mL/kg) was used. Animals were randomly assigned to the treatment groups.
For the study on infarct size, the type of treatment was masked from the two surgeons who performed MCAO. For laser speckle imaging, MCAO was performed by a surgeon who was aware of the treatment group. One hour after the final administration of fenofibrate, MCAO was induced. Animals were anesthetized, throughout the MCAO procedure and for 30mins after reperfusion, with 1.5% isofluorane in a gas mixture of 70% N2O and 30% O2 using a vaporizer. Rectal temperature was maintained at 37°C with a thermostat-controlled heating pad. Animals were placed in the supine position and MCAO was induced on the left side with an 8-0 nylon monofilament coated with silicone resin and hardener mixture (Heraeus, Hanau, Germany) as described earlier (Steele et al, 2008). The common carotid artery and external carotid artery were ligated with a 5-0 silk suture. A miniclip (Ohwatsusho, Tokyo, Japan) was temporally applied to the left internal carotid artery. Then, the filament was introduced into the internal carotid artery through the external carotid artery, and advanced until the tip occluded the anterior cerebral artery with confirming CBF drop by using laser speckle imaging. For reperfusion, the filament was withdrawn and the common carotid artery was re-opened; however, the external carotid artery remained permanently ligated. In the infarct size study, CBF at ischemic core was monitored by laser Doppler flowmetry using a flexible fiber glass probe (FLO-C1, Omegawave, Tokyo, Japan).
Changes in cerebral surface blood flow were monitored by using a laser speckle blood flow imaging system (Omegazone, Omegawave). Animals were placed in the prone position and the skull was exposed by a midline scalp incision. The skull surface was diffusely illuminated by 780nm laser light. The scattered light was filtered and detected by a CCD camera positioned above the head. The filter detected only scattered light that had a perpendicular polarization to the incident laser light. Raw speckle images were recorded by using a video capture card and Ultra Edit 2 software. The raw speckle images were used to compute speckle contrast, which corresponds to the number and velocity of moving red blood cells, that is CBF. Signal processing was performed by the algorithm developed by Forrester et al (2002). Color-coded blood flow images were obtained in high-resolution mode (638 pixels × 480 pixels; 1 image/sec). One blood flow image was generated by averaging numbers obtained from 20 consecutive raw speckle images. Cortical area in which CBF was either 20%, 30%, or 40% of the basal value was determined with a computerized image analysis system (MCID, Imaging Research, Cambridge, UK).
rCBF in nonischemic C57BL/6 mice was measured using quantitative autoradiography with [14C]iodoantipyrine according to Jay et al (1988) with modifications (Yamada et al, 2000). The femoral artery and vein were cannulated with a polyethylene catheter (PE-10, Becton Dickinson, Sparks, MD, USA) for continuous arterial blood pressure and heart rate monitoring and for drug infusion. Animals were under 1.5% isoflurane anesthesia with spontaneous respiration. Rectal temperature was maintained at 37°C. Arterial blood samples (50μL) were collected through the catheter and pH and blood gases were analyzed. After achieving a stable blood pressure with pO2 and pCO2 within normal range, 14C-iodoantipyrine (5μCi in 0.1mL saline) was infused into the left femoral vein at progressively increasing rates over 1min using a programmable pump. During this procedure, arterial blood samples were collected every 5secs onto preweighed filter paper disks. One minute after starting infusion, animals were decapitated, and brains were quickly frozen with 2-methylbutane chilled on dry ice at −45°C. The blood-loaded filter paper disks were treated with Scintigest (Fisher Scientific, Pittsburgh, PA, USA) overnight and the radioactivity in blood samples was measured by liquid scintillation spectrometry 2 days later. For determination of tissue 14C concentration, 20μm thick coronal brain sections were made using a cryostat, mounted on glass slides, and apposed to Kodak BioMax autoradiographic film (GE Healthcare, Arlington Heights, IL, USA) together with 14C standards (GE Healthcare). The density of the autoradiograms was measured with the MCID image analyzer. The software converts the optical density to radioactive content and to CBF using the radioactive standards and the 14C-iodoantipyrine blood concentration curve. In each treatment group, the mean rCBF value and standard deviation (s.d.) were calculated for each brain structure.
In 129/Sv mice subjected to 60mins of MCAO, infarct volume was determined at 24h after reperfusion. Brains were fixed with 10% formalin and subsequently treated with phosphate-buffered saline (PBS) containing 20% sucrose. The brains were cut into 40μm thick coronal sections with a freezing microtome. Sections from five levels with 2mm intervals were stained with cresyl violet and infarct area was measured in a treatment type-masked manner by using the MCID imaging system.
Mouse brain capillaries were isolated as reported earlier (Ospina et al, 2002) with minor modifications (Guo et al, 2009). Collected fresh brains were homogenized in ice-cold PBS (pH 7.4) with a loosely fitting Dounce homogenizer, and centrifuged at 2,000g for 5mins at 4°C. The supernatant was removed and stored on ice. The pellet was re-suspended in PBS and centrifuged at 2,000g for 5mins at 4°C. The supernatant was combined with the first supernatant and centrifuged at 3,000g for 10mins at 4°C. The resulting pellet containing the parenchymal fraction was stored at −80°C. The pellet that was obtained after the second 2,000g centrifugation was re-suspended in PBS, carefully layered over a 15% dextran density gradient (molecular weight 35,000 to 40,000kDa, Sigma Aldrich, St Louis, MO, USA), and centrifuged in a swinging-bucket rotor at 3,500g for 35mins at 4°C. The supernatant was discarded, and the pellet was re-suspended in PBS, layered over dextran again and centrifuged at 3,500g for additional 35mins. The resulting pellet was thoroughly washed with ice-cold PBS over a 70μm nylon mesh and the collected cerebral vessels were stored at −80°C.
RNA was isolated with Trizol (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions. The PCR amplifications were performed in a 50μL reaction volume containing 2μL cDNA, 25μL of iQ SYBR Green Supermix (Bio-Rad, Hercules, CA, USA), primer, and nuclease-free water. Primer sequences for the studied genes were as the follows: eNOS, forward, 5′-gaccctcaccgctacaacat-3′ and reverse, 5′-ctggccttctgctcattttc-3′ PPARα, forward, 5′-gcagctcgtacaggtcatca-3′ and reverse, 5′-ctcttcatccccaagcgtag-3′ PPARγ, forward, 5′-ttttcaagggtgccagtttc-3′ and reverse, 5′-aatccttggccctctgagat-3′ PPARβ, forward, 5′-tggagctcgatgacagtgac-3′ and reverse, 5′-gtactggctgtcagggtggt-3′ GAPDH, forward, 5′-aactttggcattgtggaagg-3′ and reverse, 5′-acacattgggggtaggaaca-3′. The PCR was performed in an iCycler iQ at 50°C for 2mins, 95°C for 10mins, followed by 35 cycles of 95°C for 10secs, 55°C for 10secs, and 72°C for 10secs. The relative gene expression of eNOS and PPARs were calculated using the comparative Ct (cross-threshold) method. Briefly, the Ct of the housekeeping gene GAPDH was subtracted from the Ct of each target molecule to get ΔCt. The ΔCt value of control group sample was then subtracted from the ΔCt of the treatments to get the ΔΔCt value. Fold differences compared with control group sample are obtained by calculating 2−ΔΔCt for each treatment group.
Total cholesterol and triglyceride levels were, respectively, measured by using the Wako Cholesterol E assay kit and Wako L-Type TG M assay kit (Wako Diagnostics, Richmond, VA, USA) according to the manufacturer's instructions.
Data are presented as mean±s.d. Statistical analyses were made using SPSS 15 software (SPSS, Chicago, IL, USA). One-way analysis of variance (ANOVA) followed by Scheffe post hoc analysis was used for autoradiography, arterial blood pressure, arterial blood gas analysis, and real-time PCR. Two-way repeated measures ANOVA followed by Scheffe post hoc analysis was performed for hemispheric CBF in wild-type mice, and Kruskal–Wallis followed by Mann–Whitney U-test was performed for ischemic areal size in C57BL/6 and infarct volume in 129/Sv. Unpaired Student's t-test was performed for hemispheric CBF and ischemic areal size in PPARα-null mice. P<0.05 was considered statistically significant.
Laser speckle imaging through the intact skull showed CBF reductions in the left hemisphere after MCAO in C57BL/6 mice (Figure 1A). We examined the temporal changes of cortical CBF in the ischemic hemisphere in the absence and presence of fenofibrate. In the vehicle-treated animals, the cortical CBF in the ischemic hemisphere continued to decrease up to 2h after MCAO, suggesting expansion of ischemic area (Figure 1B). However, the fenofibrate-treated animals did not show a continued reduction in CBF (Figure 1B), indicating that fenofibrate limited the progression of impaired CBF. The difference, compared with vehicle, was greater with longer occlusion periods and was significant at 2h after MCAO at both 30 and 100mg/kg doses.
To obtain topographical information, the cortical surface area, in which rCBF was less than 20%, 30%, or 40% of the basal value, was determined in each animal. Figure 1C shows a representative case from the vehicle-treated group at 2h after MCAO. As the occlusion time was prolonged, the area with rCBF deficit expanded in the vehicle-treated animals (Figure 1D). Fenofibrate-treated animals showed a significantly smaller area in which rCBF was less than 20% of the basal level after 2h of MCAO (Figure 1E), suggesting that fenofibrate attenuated the expansion of ischemic core area. Similarly the area in which rCBF was less than 30% was significantly smaller in the fenofibrate-treated groups.
After reperfusion following 2h of MCAO, fenofibrate (30mg/kg)-treated group exhibited an improved restoration of rCBF toward the basal level compared with the vehicle group (Figure 1B). At 30mins after reperfusion, the area in which rCBF was less than 20%, 30%, or 40% was significantly reduced by fenofibrate (30mg/kg) compared with vehicle (Figure 1F). The higher dose (100mg/kg) did not affect such areal sizes.
To measure the basal level of rCBF in nonischemic C57BL/6 animals, we performed quantitative autoradiography with the [14C]iodoantipyrine method. Mean rCBF values in each brain area were compared among the three treatment groups. There was no significant difference in mean values in the analyzed brain structures (Table 1). In addition, there was no significant difference in arterial blood pressure measurement and arterial blood analysis (Table 2).
The current treatment protocols with fenofibrate did not affect cholesterol and triglyceride levels in young male C57BL/6 mice on normal diet (Table 3), suggesting that the CBF improvement by fenofibrate in the ischemic brain was not related with the lipid lowering effect of fenofibrate.
We investigated whether PPARα is required for the fenofibrate-induced increase of CBF in ischemic brain. PPARα-null mice were treated for 7 days with either vehicle or fenofibrate (30mg/kg). This dose was effective in wild-type animals. There was no difference between fenofibrate and vehicle in cortical CBF in PPARα-null animals during MCAO and after reperfusion (Figure 2A). No difference was detected between the two groups in the cortical surface areal size, in which CBF was less than 20%, 30%, or 40% at 2h after MCAO (Figure 2B), suggesting the possibility that the improvement by fenofibrate in rCBF is related with PPARα.
To confirm brain tissue protection, we analyzed the impact of fenofibrate (30mg/kg) on infarct volume in 129/Sv mice subjected to 60mins of MCAO. A total of 20 animals were randomly assigned to either vehicle (n=10) or fenofibrate treatment (n=10) for 7 days. One animal in the fenofibrate group died because of technical problem during MCAO. Three animals in each group did not survive the reperfusion period. There was no difference in neurological score at 24h after reperfusion (Figure 3A). Fenofibrate significantly attenuated infarct volume at 24h after reperfusion: vehicle, 23.7%±5.0% contralateral hemisphere (n=7); fenofibrate, 16.9%±5.4% contralateral hemisphere (n=6) (Figure 3B). Fenofibrate did not affect CBF at the ischemic core measured by laser Doppler flowmetry (Table 4).
To analyze the impact of fenofibrate on PPAR expression, we measured mRNA levels of PPARs in brain and aorta in nonischemic animals that were treated with fenofibrate for 7 days. Fenofibrate significantly increased only PPARα in the aorta (Figure 4A) and isolated brain microvessels (Figure 4B). Fenofibrate showed a small increase in PPARγ in brain microvessels, but this increase was not statistically significant (P0.05 by ANOVA) (Figure 4B). Conversely, the brain fraction containing neurons and glia showed decreases in PPARα and PPARγ; however, PPARβ was not affected (Figure 4C). We next measured the effects of fenofibrate on eNOS level as eNOS has been shown to mediate CBF increases by statins, another type of lipid lowering drug (Endres et al, 1998). Fenofibrate did not significantly alter eNOS mRNA levels in both of the aorta and brain microvessels (Figure 4D).
In this study, laser speckle imaging documented both areal and temporal changes in cortical surface CBF in the same animal as reported by other investigators (Dunn et al, 2001; Zhou et al, 2008). We observed a progressive expansion of surface area in which CBF was severely compromised (<30%) at least up to 2h after MCAO in vehicle-treated animals (Figure 1D). In addition, even after 30mins of reperfusion, there was a substantial area with poor perfusion (Figure 1F), supporting no-reflow phenomenon (Hossmann, 1983; del Zoppo and Mabuchi, 2003). Fenofibrate improved CBF in the ischemic brain during MCAO (Figure 1E). In addition, fenofibrate (30mg/kg) enhanced CBF recovery after reperfusion (Figure 1F), suggesting that fenofibrate reduced the no-reflow phenomenon. This study also confirmed an attenuation in infarct size by fenofibrate in 129/Sv mice subjected to 60mins of MCAO. Therefore, the infarct size reduction by fenofibrate was likely attributable to the CBF improvement in the ischemic brain, at least in part.
The mechanism by which PPARα activation by fenofibrate exerts the cerebrovascular protective effects remains unknown. Deplanque et al (2003) showed that fenofibrate increased acetylcholine-induced endothelium-dependent relaxation of isolated middle cerebral artery from rats. Thus, PPARα activation may also enhance cerebral artery relaxation in the ischemic brain. In addition, antiinflammatory action of PPARα activation may prevent white blood cells or platelet adhesion/aggregation to microvessels in the ischemic area (Yamakawa et al, 1987; Kataoka et al, 2004), by doing so fenofibrate may improve CBF in the ischemic brain. Consistent with this notion, prevention of adhesion molecule expression by fenofibrate has been shown in ischemic brain (Deplanque et al, 2003). As part of such an investigation, we measured the influence of fenofibrate on eNOS mRNA levels in the brain microvessles. The effect was not significant. Conversely, the mRNA levels of eNOS in the aorta tended to decrease (Figure 4D). These findings contrast with those reported for statins, another type of lipid lowering drug, which has been shown to improve CBF by eNOS-mediated mechanisms (Endres et al, 1998). Clarifying the mechanistic linkage between PPARα activation and CBF improvement in the ischemic brain may provide insights into a novel intervention to enhance neuronal survivability after ischemic stroke.
CBF-independent mechanisms may also contribute to the fenofibrate-induced attenuation in infarct size. Our recent study with another type of PPARα agonist gemfibrozil suggested CBF-independent neuroprotection against ischemia (Guo et al, 2009). Deplanque et al (2003) showed that chronic treatment with fenofibrate increased major antioxidant enzymes including copper/zinc superoxide dismutase in the nonischemic mouse brain, which is likely to contribute to the neuroprotection in the brain after ischemia. Compared with the antiinflammatory action of PPARα agonists applied to microglia (Jana et al, 2007), the consequence of their direct action on neurons is less clear. Santos et al (2005) showed that a selective PPARα agonist Wy-14643 (100μmol/L) increased survivability of rat hippocampal dissociated neurons subjected to β-amyloid peptide. Contrary, Bento-Abreu et al (2007) observed that Wy-14643 (>50μmol/L) was neurotoxic. We recently found that a higher dose of gemfibrozil (120mg/kg) tended to aggravate brain swelling after permanent MCAO (Guo et al, 2009). The neurotoxicity of PPARα agonists at their high concentrations might cause the loss of effect by fenofibrate at 100mg/kg on the CBF after reperfusion (Figure 1F). Developing a novel PPARα agonist with minimal neurotoxicity will help to further improve stroke outcome by PPARα activation.
The beneficial effect by fenofibrate was not seen in PPARα-null mice (Figure 2). In addition, we found that mRNA levels of PPARα were significantly increased by fenofibrate in the brain microvessels and aorta (Figures 4A and 4B). Thus, PPARα was likely a mediator of the cerebrovascular protective effects of fenofibrate. The selective effects of fenofibrate on the mRNA levels of PPARα in the aorta and brain microvessels may be related with its higher affinity as a ligand for PPARα over PPARγ and PPARβ. Our recent study with another fibrate gemfibrozil showed similar findings in the aorta. However, gemfibrozil increased not only PPARα but also PPARγ and PPARβ to a lesser extent (Guo et al, 2009), which may also reflect the binding ability of gemfibrozil to PPARγ and PPARβ. In addition, the increase in mRNA levels of PPARα by both fenofibrate and gemfobrozil may suggest a positive feedback loop as a regulatory mechanism of PPARα gene expression.
Gemfibrozil has been shown to decrease PPARβ in both brain microvessels and neuroglial fractions (Guo et al, 2009); however, fenofibrate did not significantly reduce PPARβ in this study (Figures 4B and 4C). As exacerbated infarct formation after permanent focal cerebral ischemia has been shown in PPARβ-null mice (Arsenijevic et al, 2006; Pialat et al, 2007), the lack of reduction in PPARβ expression levels in the fenofibrate-treated brains may explain the more prominent infarct reduction by fenofibrate compared with that by gemfibrozil (Inoue et al, 2003; Guo et al, 2009). In contrast, both PPARα and PPARγ were reduced by fenofibrate in the neuroglial fraction. These findings suggest that gene expression of PPAR is regulated in an organ-dependent manner. The findings also raise the possibility that the decrease in PPARα and PPARγ by fenofibrate in the neuroglial fraction might counteract the beneficial effect of fenofibrate against ischemic stroke, as selective neuronal PPARγ gene disruption has been shown to increase susceptibility to brain damage after MCAO (Zhao et al, 2009), and because magnetic resonance imaging showed increased infarct volume in PPARα-null mice compared with their wild-type controls (Pialat et al, 2007).
In summary, we showed that pretreatment with fenofibrate improved the penumbral CBF during MCAO. In addition, fenofibrate enhanced CBF recovery after reperfusion. These effects by fenofibrate required PPARα. The mechanism by which fenofibrate exerts cerebrovascular protection needs to be determined. As gemfibrozil has been show to reduce the incidence of stroke in patients (Bloomfield Rubins et al, 2001), further investigation of the impact of prophylactic usage of fibrates on stroke outcome may be meritorious.
Supported in part by NS034194, NS048532, and NS060659 from NIH/NINDS and S21MD000101 from NIH/NCMHHD. The study was conducted in a facility constructed with support from Research Facilities Improvement Grant 1 C06 RR-07571 from NIH/NCRR. We thank Dr Zhihong Huang at Johnson and Johnson Pharmaceutical Research and Development and Dr Christian Waeber at Massachusetts General Hospital for their advice on the iodoantipyrine autoradiography method. We thank Drs Peter R MacLeish and Benveniste Morris at Morehouse School of Medicine for critical reading and comments.
Conflict of interest
The authors declare no conflict of interest.