Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Introduction

The peroxisome proliferator-activated receptors (PPARs), including α, γ, and δ/β, are encoded by separate genes and are members of the nuclear hormone receptor superfamily of ligand-activated nuclear transcription factors. PPARγ, also known as NR1C3 (nuclear receptor subfamily 1, group C, member 3), is a pleiotropic type II nuclear receptor, which was termed for its ability to induce proliferation of hepatic peroxisomes in response to xenobiotic stimuli in mice [1]. Three different PPARγ transcripts (PPARγ 1, 2, and 3), each a derivative of the PPARγ gene through differential promoter usage and alternative splicing, have been identified [2, 3]. While PPARγ 2 is the form primarily expressed in adipose tissue, PPARγ 1 has broader tissue distribution including presence in the brain [2]. As a transcription factor that regulates target gene expression through binding to the conserved DNA sequence termed peroxisome-proliferator response element (PPRE) [2, 4, 5], PPARγ was initially described in adipose tissue as a key regulator of metabolic processes [6–9]. Soon after, PPARγ was shown to be a unique therapeutic target for the treatment of metabolic disorders, e.g., diabetes (insulin resistance), obesity, hyperlipidemia, and hyperglycemia [10–12]. Among many compounds, ligands for PPARγ activation include fatty acids (especially the oxidized form) [13–15], cyclopentanone prostaglandins (e.g., 15-deoxy-Δ12,14-prostaglandin J2; -15d-PGJ2) [16], lipoxygenase products [17, 18], the nonsteroidal anti-inflammatory drugs (NSAIDs) [19, 20], and a class of clinically relevant compounds, the thiazolidinediones (TZDs) [10, 21]; of which pioglitazone and rosiglitazone are used to treat the type 2 diabetes mellitus [22–25]. In addition, PPARγ transactivation is regulated by its phosphorylation [26, 27]. Specifically, phosphorylation of PPARγ by the extracellular signal-regulated kinase (ERK1/2) and C-Jun N-terminal kinase (JNK) reduces PPARγ activity [26, 27]. Since JNK is activated by H2O2, oxygen–glucose deprivation (OGD), NMDA or ischemic stroke and acts as pro-death signal [28–32], the deleterious JNK functions may be secondary to the phosphorylation-mediated PPARγ inhibition.

Later studies on the mechanism of PPARγ action in other than fat tissue demonstrated its important role in regulation of anti-oxidative and anti-inflammatory processes [33–35]. It is primarily the anti-inflammatory properties of PPARγ ligands that ultimately brought the closer attention to PPARγ and PPARγ-activating agents to vascular diseases process [36–38]. PPARγ (and primarily PPARγ1) expression is ubiquitous regarding the type of tissues and cells it is expressed. In terms of neurological conditions, PPARγ in preclinical studies was shown to act as potential target for the treatment of ischemic stroke [39–51], intracerebral hemorrhage [52], neurotrauma [53–58], Alzheimer’s and neurodegenerative diseases [59–69], autoimmune encephalomyelitis (EAE), a model for multiple sclerosis [70–72]. In this chapter, our focus is mainly on the role of PPARγ in ischemic stroke, attempting to discuss the interactions of PPARγ with the NF-E2-related factor 2 (Nrf2) and the nuclear factor kappa B (NF-κB) signaling pathways in regulating pro- and anti-inflammatory responses in the brain.

Pleiotropic Effect of PPARγ Agonists in Ischemic Stroke

Based on the known function of gene targets, PPARγ acts as a key regulator in a broad range of processes virtually in all brain cells including neurons [45, 73], astroglia [74–76], oligodendroglia [77–79], microglia [54, 80, 81], and endothelial cells [82, 83]. Primarily through the use of various PPARγ agonists but also through the use of cell-specific PPARγ knockouts, PPARγ was demonstrated to protect brain from damages caused by ischemic [41, 42, 45, 84–87] and hemorrhagic stroke [53–55]. The beneficial effects of PPARγ activation was linked to (1) repression of pro-inflammatory mediators production (at least in part through inhibition of NF-κB either directly or by upregulation of endogenous NF-κB inhibitor, IκB [33, 34, 53, 88–95]), (2) upregulation of antioxidant enzymes including CuZn-superoxide dismutase (SOD) and catalase [41, 54], (3) inhibition of excitotoxicity [96, 97], and (4) activation of phagocytotic activities by microglia and macrophages via mechanism involving the PPARγ-target gene—scavenger receptor CD36, the molecule that assists in cleanup of damaged brain tissue, a process necessary for efficient recovery and the termination of deleterious pro-inflammatory cascade (Fig. 17.1) [54, 98–101].

Fig. 17.1
figure 1

PPARγ regulated pathways after stroke—role of PPARγ activators. PPARγ transcriptionally controls expression of numerous genes including the anti-oxidative enzymes, such as catalase and superoxide dismutase (SOD), as well as the transcription factor Nrf2. Nrf2 plays a key role in amplifying the expression of many anti-oxidative genes including catalase and SOD, similar to PPARγ. This anti-oxidative feature of PPARγ is critical in combating oxidative damage imposed by cerebral ischemia. Importantly, since PPARγ and Nrf2 are ubiquitously expressed, this anti-oxidative mechanism may apply to all brain cell types affected by stroke. In addition, both PPARγ and Nrf2 regulate expression of CD36, a scavenger receptor that is abundant on microglia/macrophages. CD36 plays important role in endocytosis of oxidized lipids and phagocytosis of dead (including apoptotic) cells and other cellular debris, thereof aiding in cleanup—process allowing for a faster inflammation resolution and more efficient tissue repair. Another important task of PPARγ is to inhibit NF-κB, a proinflammatory transcription factor implicated in BBB disruption and brain edema formation. Ultimately, augmented PPARγ activation improves inflammation resolution, tissue repair, and functional recovery after stroke

Full size image

PPARγ and Neuroprotection

In response to the prolonged ischemia, neurons that are localized in the ischemic core die rapidly as consequence of ischemia-induced energy failure, anoxic depolarization, and excitotoxicity, which is the result of glutamate receptors overactivation, calcium overload, and a breakdown of ion homeostasis [102–109]. Using oxygen–glucose deprivation (OGD) or glutamate/NMDA toxicity (in vitro models of ischemia) to study the neuroprotective capacity of PPARγ agonists [including pioglitazone, rosiglitazone, or cyclopentanone prostaglandins (CyPG)], we and other groups demonstrated that activation of PPARγ potently reduces the neuronal death in the primary neurons [96, 97, 110], implying that PPARγ may act as pro-survival factor for neurons under the ischemic/excitotoxic stress. The anti-excitotoxic effect of PPARγ agonists was observed not only in cultured neurons but also in the animal injury model that assess the extent of brain damage caused by intracortical injection of NMDA [97]. Finally, we have established that neurons derived from animals engineered to lack PPARγ, selectively in neurons, demonstrated significantly increased susceptibility to excitotoxic damage and to OGD [84]. In agreement with the in vitro data, mice lacking PPARγ in neurons were significantly more susceptible to the ischemic damage caused by focal cerebral ischemia [84].

Reactive oxygen species (ROS) are well known to represent one of the most important components of brain injure in response to ischemia/reperfusion insult. ROS are generated by the ischemia-affected brain cells, the activated microglia, and infiltrating neutrophils that collectively impose oxidative stress to cells located in proximity to the ischemia [111–114]. To combat the oxidative stress, cells have developed a number of self-defense mechanisms including upregulation of enzymes with anti-oxidative functions. Superoxide dismutase along with catalase and glutathione peroxidase plays key roles in eliminating ROS through catalytic decomposition of superoxide or H2O2 [84, 115, 116]. Catalase is a large homotetrameric protein that is usually localized in peroxisomes (the membrane-bound organelles that house β-oxidation of very long chains of fatty acids, in which toxic peroxides are generated as side products) [117], where it acts to protect the cells from the toxic effects of H2O2 by catalyzing its decomposition. As a ubiquitous enzyme to most cells in our body including neuroglia and neurons [118], catalase expression is regulated by PPARγ and Nrf2 [115, 119]. The distribution pattern of catalase-immunopositive neurons throughout the brain inversely corresponds to increased susceptibility to damage induced by global cerebral ischemia [118], suggesting that catalase plays important role in cell survival. Overexpression of catalase in rat striatum through virus-mediated gene transfer decreases the vulnerability to ischemic stroke [120]. In response to PPARγ activation, expression of catalase rapidly increased in the ischemia-affected brain [118, 121] and in the OGD-injured neurons [122], which likely reflect an adaptive response aiming at improving the antioxidant buffering capacity under the pathological scenarios. In agreement with this notion, treatment with catalase of neurons in culture subjected to H2O2-induced injury provided a robust cytoprotection [123, 124]. Thus, catalase upregulation by PPARγ may reflect a self-protective mechanism to combat oxidative stress in stroke. It is important to point out that in addition to catalase, PPARγ regulates expression of superoxide dismutase (including in neurons), an enzyme well recognized for decades as a key player in mitigating oxidative injury and brain damage after cerebral ischemia [41, 84, 125, 126].

PPARγ-Induced CD36 Expression on Phagocytes and the Endogenous Cleanup Mechanism

After cerebral ischemia, the infarcted/dead tissue not only acts as a reservoir of various cytotoxic and pro-inflammatory molecules that harm the adjacent healthy brain tissue, but it also forms a biological and physical barrier hampering neural reorganization, repair, and ultimately, neurological recovery. Thus, in order to minimize such detrimental effects, infarcted tissue needs to be removed to facilitate recovery. Microglia and hematogenous macrophages (MMΦ) are the cells primarily responsible for such cleanup and repair processes. Successful removal of the disintegrated and apoptotic brain cells or debris (including the neutrophils that accumulate in brain in response to injury and consequently die through apoptosis) by MMΦ is also essential in achieving resolution of inflammation. While apoptotic cells appear to be considerably benign to the surrounding brain tissue, an apoptotic cells non-phagocytosed in a timely manner may undergo secondary necrosis causing spill of the intracellular toxic content, leading to the damage to the neighboring cells and causing inflammation. Several macrophage scavenger receptors that mediate cleanup process have been identified. These include not only CD36 but also CD91, SR-A, and several others [54, 127–133]. Regarding apoptotic cell efferocytosis by macrophages, the phosphatidyl serine on the sickle red blood cells, symmetric red cell ghosts [134–136], or apoptotic neutrophils was suggested to act as the recognition molecule for CD36, a class II scavenger receptor on macrophages [137–139]. Expression of CD36 on macrophages (MΦ) is transcriptionally regulated by both PPARγ [98, 140, 141] and Nrf2 [142–145]. Although CD36 has various functions, one of its primary roles is to mediate endocytosis of (oxidized) fatty acids and phagocytosis of dead/apoptotic cells [129, 137, 146–148]. Deficiency of CD36 in macrophages due to genetic deletion of PPARγ leads to delayed uptake of oxidized LDL by macrophages and aggravation of atherosclerotic lesions [149]. In CD36-KO mice, aberrant phagocytotic capacity of macrophages was proposed to explain the deficiency in remyelination in response to sciatic nerve crush injury [150]. In addition, transfection of non-phagocytic cells with CD36 renders these cells capable of ingesting apoptotic neutrophils, lymphocytes, and fibroblasts [138], further confirming the important role of CD36 in phagocytosis. As pointed above, since CD36 transcription is under control of Nrf2 and PPARγ, the upregulation of CD36 by MMΦ in response to Nrf2 and/or PPARγ activators may ensure a more efficient interaction between the MMΦ and their targets for phagocytosis. This may allow for more efficient phagocytosis-mediated clearance of dead cells/tissues from the ischemic brain. However, despite its beneficial role in the cleanup process, CD36 may have detrimental effect which is normally characterized by increased oxidative stress and pro-inflammatory responses, as adult animals deficient in CD36 suffer from the less profound damage in response to cerebral ischemia [151, 152]. The nature of these responses is not known; however, the likelihood is that upon engulfment of cellular debris including oxidized lipids, the MMΦ generate damaging levels of oxidative stress during degradation of debris in the phagolysosomes. Interestingly, CD36 knockout neonates subjected to cerebral ischemia experienced more damage (suggesting beneficial function of CD36), which was suggested to be in part due to the impaired cleanup mechanism [153]. Independent of the natural responses that were tested in experiments using CD36 knockout mice, we suggest that under conditions using pharmacologic agents to activate PPARγ, MMΦ not only express higher levels of CD36 for a more efficient phagocytosis but also produce more anti-oxidative enzymes (e.g., catalase) that are regulated by PPARγ. Recently, we provided the evidence that MMΦ in culture challenged with PPARγ or Nrf2 activators, despite expressing CD36 at much higher level and demonstrating the augmented phagocytosis, experienced less oxidative damage and showed reduced pro-inflammatory gene expression [54].

Thus, in response to PPARγ in activated microglia, the upregulation of the antioxidant enzymes (in addition to CD36) may play a protective role allowing for effective and safe phagocytosis. Consequently, cleaning the apoptotic/dislocated/damaged cells or debris will help to reestablish the nurturing environment necessary for restoring tissue structure and neurological function recovery [154, 155].

PPARγ Activation and the Interaction of PPARγ and RXR

PPARγ regulates target gene expression by binding to PPRE as heterodimers with the retinoic acid receptor (RXR). Interestingly, existing studies indicate that activation of PPARγ–RXR complex can be achieved with either PPARγ and/or by RXR ligand (e.g., 9-cis retinoic acid), indicating some level of the promiscuity in activation of PPARγ [156, 157]. Although each ligand can initiate transactivation independently, the effect of co-activation appears to be stronger [9], suggesting that the occupancy of both PPARγ and RXR ligand (e.g., 15d-PGJ2 plus 9-cis retinoic acid) is needed for the maximal receptor activity [9, 158–160]. In agreement with this notion, we found that co-treatment of cultured neurons with 15d-PGJ2 and 9-cis retinoic acid was more effective in reducing the OGD-induced damage, as compared to each ligand alone [53]. This beneficial interaction between PPARγ and RXR ligands in our neuroprotection assay is consistent with an earlier report showing that combination use of 15d-PGJ2 and 9-cis retinoic acid was superior to each drug alone in reducing behavioral dysfunction in a mouse model of experimental autoimmune encephalomyelitis [161].

Interaction of PPARγ and Nrf2 and NF-κB

The pro-survival role of PPARγ includes the non-genomic inhibition of deleterious pro-inflammatory transcription factor, nuclear factor kappa B, NF-κB. In the ischemia-injured brain, the delayed cell death is in part triggered by the overproduction of pro-inflammatory molecules including pro-inflammatory cytokines (such as tumor necrosis factor alpha, TNF-α or interleukin-1 beta, IL-1β), adhesion molecules (such as intercellular adhesion molecule 1, ICAM-1 or vascular cell adhesion molecule, VCAM), matrix metalloproteinases (including MMP9) or the pro-oxidative inducible form of nitric oxide synthase (iNOS) capable of generating large quantities of nitric oxide, that in presence of superoxide generated by NADPH oxidase is converted to a highly cytotoxic peroxynitrites [109, 162–165]. Once perpetuated by ischemia, these potentially deleterious factors act in concert to damage blood–brain barrier (BBB) and cause edema and/or hemorrhage [166–168]. Interestingly, the expression of all these factors is tightly regulated by NF-κB. The activation of PPARγ can antagonize these harmful effects through inhibition of NF-κB [33, 34], which may be achieved by at least three independent mechanisms (Fig. 17.1) [33, 34, 53, 88–95]. First, PPARγ may directly bind to the NF-κB subunits, p50 and p65, resulting in NF-κB inactivation [169]; second, PPARγ may indirectly inhibit NF-κB by sequestering the common transcription co-activators such as SRC-1 [170] and p300/CBP (CREB-binding protein) [88–90]; and third, PPARγ may upregulate the production of inhibitor kappa B (IκB) [91, 93–95], the protein that directly inhibit NF-κB activation. Inhibition of NF-κB by PPARγ agonists may reduce generation of pro-inflammatory mediators involved in the secondary brain damage.

Nrf2 is a ubiquitous pleiotropic transcription factor and a key genomic homeostatic regulator of intracellular stress [171]. By combining with Mif family proteins, Nrf2 forms heterodimeric complexes capable of transactivating the antioxidant response elements (ARE) within the regulatory region of many cytoprotective target genes including catalase, superoxide dismutase, glutathione-S-transferase, thioredoxin, NQO1, and many other proteins with important role in neutralization of oxidative stress and detoxification [172]. In most cells, Nrf2 is present at low concentrations due to continuous Nrf2 degradation through the proteasome pathway [173, 174]. Nrf2 contributes to cytoprotection and amelioration of tissue damage through reducing the oxidative stress in many pathogenic conditions including cerebral ischemia [175–181], neurodegenerative diseases [182], and mitochondrial metabolic stress [183]. The growing body of evidence suggests that PPARγ may play important role in regulation of Nrf2 and thus Nrf2 target genes (Fig. 17.1). The interaction between PPARγ and Nrf2 may involve several layers of interaction. Most importantly, PPARγ was demonstrated to regulate Nrf2 gene expression and Nrf2-regulated genes containing putative PPREs [184]. Interestingly, it appears that Nrf2 also regulates PPARγ and PPARγ-regulated genes containing the ARE [185]. Next, PPRE and ARE coexist in the same genes, such as CD36 and catalase, suggesting an interactive function of Nrf2 and PPARγ in expression of these genes. Finally, an interaction between PPARγ and Nrf2 may be through NF-κB inhibition. Since NF-κB activation requires the presence of oxidative stress [186], the effect of Nrf2 in ameliorating oxidative stress was proposed to inhibit NF-κB [187]. As different mechanisms are used by Nrf2 and PPARγ in inhibiting NF-κB, it is likely that the mutual effect may lead to a synergistic role [188–190].

Adverse Effects of PPARγ Agonists

There is a small number of observations reporting the dose-dependent neurotoxic effects of the endogenous PPARγ ligand 15d-PGJ2 in cerebellar granule cells [191], primary cortical neurons [192], and spinal cord motor neurons [193]. The mechanism that underlies this neurotoxicity is unclear and some reports indicate that these harmful actions are probably not directly linked to PPARγ [191]. In our studies using mouse and rat neurons in culture, we have not observed neurotoxicity using PPARγ activating ligands to date. In fact, all the tested PPARγ agonists including 15d-PGJ2, 15d-PGD2, ciglitazone, rosiglitazone, and pioglitazone demonstrated potent cytoprotective effects in models of OGD and excitotoxicity [45, 50, 97]. The only instance showing toxicity was when the doses of the agonists were higher than these needed for the cytoprotection. Unlike synthetic TZDs that display rather significant levels of PPARγ specificity, prostaglandin D2 derivatives, including 15d-PGJ2, have a limited selectivity toward PPARγ and many of their biological activities are independent of PPARγ [92, 194–198]. However, the clinical use of PPARγ ligands, and primarily rosiglitazone, was associated with hemodilution, peripheral edema, increase in body weight, as well as cardiomyopathies and heart failure [46, 199–201]. Again, these are the known side effects of long-term use of these medications and as such should not necessarily influence the safety of patients subjected to short-term treatment. The study evaluating the safety of pioglitazone in patients with hemorrhagic stroke is currently ongoing [52].

PPARγ Agonists and Clinical Trials

Two of the thiazolidinediones (TZDs), pioglitazone and rosiglitazone, are currently approved by the FDA for treatment of type 2 diabetes mellitus. These insulin-sensitizing PPARγ agonists are unique among all the glucose-lowering agents as they act independent of secretion of insulin from pancreas (TZDs do not change blood insulin levels, rather make cells more sensitive to its effect) [22, 202]. The glucose-lowering effect of TZDs is of clinical importance since hyperglycemia during ischemia/reperfusion may worsens the brain damage and neurological outcome, including by increasing incidence of hemorrhage in patients subjected to thrombolysis with rt-PA [203–206]. A first case-matched controlled study reporting improved functional recovery in stroke patients with type 2 diabetes receiving pioglitazone or rosiglitazone (vs. control type 2 diabetes patients not receiving TZDs) yields a promising outlook [207]. Subsequently, PROACTIVE (PROspective pioglitAzone Clinical Trial In macroVascular Events; NCT00174993), a randomized, double-blinded, placebo-controlled study looked at the impact of pioglitazone on total mortality and macrovascular morbidity in 5,238 patients with diabetes and macrovascular disease. This secondary prevention study showed safety and a macrovascular benefit with pioglitazone in terms of major adverse cardiovascular events including all-cause mortality, nonfatal myocardial infarction, acute coronary syndrome, cardiac intervention (including coronary artery bypass graft or percutaneous coronary intervention), and stroke [208–210]. The higher beneficial rates were observed in patients with prior stroke compared with those without prior stroke [211, 212]. A meta-analysis of 19 randomized clinical trials with pioglitazone revealed a statistical difference regarding the favorable outcome including mortality, nonfatal MI, and stroke when using pioglitazone [201]. However, a recent study suggests that use of rosiglitazone may impose 1.4-fold increase in risk of acute MI and death from cardiovascular diseases compared with non-TZDs therapies [213]. As compared to pioglitazone, rosiglitazone significantly increased the risk of stroke, heart failure, and death in elderly patients [214]. In contrast, from the stroke prevention point, pioglitazone has shown significant protection from both micro- and macrovascular cardiovascular events and plaque progression [215–217].