[1] 8w The Evolving Landscape of Neuroinflammation After Neonatal Hypoxia-Ischemia

[1] 152w Hypoxic-ischemic brain injury remains a leading cause of mortality and morbidity in neonates. The inflammatory response, which is characterized in part by activation of local immune cells, has been implicated as a core component for the progression of damage to the immature brain following hypoxia-ischemia (HI). However, mounting evidence implicates circulating immune cells recruited to the site of damage as orchestrators of neuron-glial interactions and perpetuators of secondary brain injury. This suggests that redirecting our attention from the local inflammatory response toward the molecular mediators believed to link brainimmune cell interactions may be a more effective approach to mitigating the inflammatory sequelae of perinatal HI. In this review, we focus our attention on cyclooxygenase-2, a mediator by which peripheral immune cells may modulate signaling pathways in the brain that lead to a worsened outcome. Additionally, we present an overview of emerging therapeutic modalities that target mechanisms of neuroinflammation in the hypoxic-ischemic neonate.

[1] 76w Hypoxia-ischemia (HI) occurs in 1 to 6 per 1,000 live fullterm births [1]. Of those affected, 15-20% will die in the postnatal period, and 25% of survivors will be left with longterm neurological disabilities [2][3][4]. Intrauterine asphyxia is the underlying mechanism of hypoxic injury and is a consequence of circulatory problems, including clotting of placental arteries and placental abruption [5]. HI in the neonate is a manifestation of systemic hypoxia combined with reduced cardiac output [6].

[2] 120w Studies have shown that the pathophysiology of brain injury secondary to HI consists of a biphasic profile (Fig. 1). The initial phase of HI is characterized by brain acidosis and the depletion of high-energy phosphorylated compounds, such as adenosine triphosphate and phosphocreatine [1,7,8]. This primary energy failure leads to the loss of membrane ionic homeostasis, depolarization of the cell, osmotic dysregulation, and inhibition of protein synthesis, further leading to necrosis [9,10]. The secondary processes evolve over days after the brain insult and are also characterized by a depletion of high-energy phosphorylated compounds, however without tissue acidosis. Although the pathogenesis of secondary brain injury involves multiple pathophysiological processes, accumulating evidence implicates the inflammatory response as a core component of damage [11][12][13].

[3] 89w Neuroinflammation includes initial release of proinflammatory mediators by injured or dying cells, activation of microglia and astrocytes, and leukocyte infiltration. It is the synergistic actions of these events that potentiate brain damage and lead to neurological dysfunction [14]. However, experimental studies thus far have focused mainly on selectively targeting these mechanisms, which may explain why there are no pharmacotherapies proven clinically viable for the treatment of HI brain damage. In fact, increasing evidence suggests identifying molecular mediators responsible for orchestrating brain-immune cell interactions as a more promising approach [15][16][17].

[4] 37w This review will provide a brief overview of the current understanding of the local and peripheral inflammatory response involved in neonatal HI, and the role of cyclooxygenase-2 (COX-2) in brain-immune cell interactions and the progression of neuroinflammation.

[5] 83w Microglia cells serve as specialized sensors for brain tissue damage. In response to ischemia, microglia morphologically change from a resting ramified phenotype to a motile activated amoeboid cell able to migrate to necrotic areas to remove cellular debris [10,18]. However, in the process, these activated cells contribute to secondary brain injury by releasing a variety of pro-inflammatory mediators including cytokines, reactive oxygen species, complement factors, free radical species, and nitric oxide, which contribute to cell death, ultimately creating a vicious perpetuating cycle [2].

[6] 114w Mounting evidence suggests that infiltrating peripheral immune cells may be necessary for the activation of microglia, thereby exacerbating neurodegeneration after ischemia. In an in vitro study, microglia, when co-cultured with T-cells, become activated, thereby releasing an inflammatory cytokine [19]. In an in vivo study, removal of a population of infiltrating macrophages, neutrophils, B cells, and T cells by splenectomy appeared to reduce microglia activation and dramatically reduce brain damage [20]. Systemic inhibition of monocyte/macrophage or neutrophil populations has also been shown to reduce cerebral infarct volume after ischemic injury [21][22][23]. However, the exact mechanism by which peripheral immune cells activate and/or propagate the local inflammatory response and enhance neuronal death remains to be determined.

[7] 19w Astrocytes, which are the predominant glial cell type in the central nervous system (CNS), have been shown to produce

[8] 164w Depolarization Free radical production Enzyme induction Lipid peroxidation Osmotic dysregulation (Edema) Mitochondrial damage Na + K + Ca 2+ Na + Ca 2+ Apoptosis Neuron Microglial activation Peripheral immune cell infiltration Astrocyte GLU Re-uptake Cerebral blood flow High-energy phosphate reserves Lactic acid GLU Release Inflammatory mediators Peripheral immune cells Red blood cell Fig. 1 Pathophysiology of a hypoxic-ischemic event. Decrease in cerebral blood flow results in a decrease in high-energy phosphate reserves (i.e., adenosine triphosphate; phosphocreatine) and a build-up of lactic acid. Loss of membrane ionic homeostasis leads to intracellular accumulation of sodium (Na + ), calcium (Ca 2+ ), and water (edema), thereby depolarizing the cell and releasing glutamate (triangle) and potassium (K + ) into the extracellular space. Intracellular calcium ion accumulation leads to enzyme induction (i.e., lipases, proteases, endonucleases) and free fatty acid elevation, which undergo peroxidation. The result is the accumulation and/or release of inflammatory mediators (i.e., cyclooxygenase-2), which can lead to apoptosis, glial activation, and peripheral immune cell infiltration

[9] 222w inflammatory mediators in a variety of brain injures including HI [24,25]. Inflammatory cytokines have been associated with neonatal HI brain damage and later development of cerebral palsy [2,26]. Specifically, elevated levels of interleukin (IL)-6 in cerebrospinal fluid of asphyxiated newborns have been correlated with an increased degree of brain damage and poor neurological outcome [27]. Additionally, recent evidence has implicated IL-15 as playing a leading role in neuroinflammation in the injured immature rat brain [28]. Importantly, astrocytes are the main source of both IL-6 and IL-15 in CNS injury and inflammation [2,29]. Astrocytes may also influence the local inflammatory response through their communicative partnership with neighboring cells [30]. Under pathological conditions, astrocytes play a critical role in the activation of microglia [31,32], and by-products of reactive astrocytes such as tumor necrosis factor-alpha (TNF-a) and IL-6 are associated with neuronal demise after HI [33]. On the other hand, astrocytes are also a source of trophic factors, such as granulocyte-colony-stimulating factor (G-CSF) [34], and are responsible for regulating neurotransmitter and ion concentrations, removing debris, and maintaining an optimal environment for neuronal function [35]. Impairment of astrocyte function during HI is thought to influence neuron viability [33]. Therefore, it is important to identify key molecular mediators responsible for initiating astrocyte signaling pathways involved in worsening brain injury, without eliminating the protective function of these cells.

[10] 231w Once thought to be passive bystanders in neuroinflammation, neurons are now known to play a more active role. As such, neurons can be a source of inflammatory mediators, including complement, COX-2, and cytokines after HI [12,36]. Neurons can express COX-2 at low levels under normal conditions; however, under pathological conditions, COX-2 is upregulated in response to mitogens, inflammatory mediators, and hormones [37]. Induction of COX-2 expression in neurons is also driven by physiological synaptic activity [38] and acute paradigms of excitotoxicity [39], thereby promoting local inflammatory reactions and injury to themselves [40][41][42]. Moreover, neurons contribute to the production of proinflammatory mediators that can alter vascular permeability and regional blood flow, and enhance chemotactic activity and thereby promote leukocyte migration [11]. Once peripheral leukocytes and monocytes enter the brain parenchyma, their actions appear to be multifaceted [43]. Certain immune cell subpopulations may directly elicit neuronal death via contact-dependent mechanisms [44] or release molecular mediators that activate resident cells, thus promoting further brain injury [45,46]. In line with this concept, recent studies have shown that T-lymphocyte-deficient mice demonstrate attenuated brain injury and neurological deficits after experimental stroke [47,48]. To make matters more complex, regulatory T-lymphocytes have been shown to have a protective role in the brain after stroke [49]. Immune cells have also been implicated in the generation of new neurons and improvement of spatial learning and memory performance in neurodegenerative disease [50][51][52].

[11] 65w Astrocytes are viewed as an active participant in synaptic transmission and processing of information -a departure from the old dogma in which astrocytes were identified as merely physical supporters for neighboring neurons [35]. Moreover, opening of gap-junctional communication channels links dying astrocytes in the ischemic core with penumbral cells [53]. Therefore, astrocytes might compromise juxtaposed cells found in salvageable tissue that otherwise may have survived.

[12] 83w Studies have suggested that inflammatory mediators might be the driving force for altering astrocyte function and thereby impacting neuron-glial signaling. For example, astrocytes undergo IL-1b-induced elevations in intracellular calcium, which may enhance glia-to-neuron signaling, leading to a reduction in neuron survival [54]. Proinflammatory cytokines may also be responsible for impairing astrocyte energy metabolism, thereby jeopardizing neuronal vulnerability [35]. Thus, it is reasonable to conceptualize that targeting cytokines may lead to a profound modulation of astrocyte function and improve neuronal survivability in this mechanism.

[13] 160w To date, many studies have investigated the anti-inflammatory properties of COX-2 inhibition and the benefits with improving neurological outcomes after various adult brain injuries. Yet surprisingly, only one study to date has investigated the use of COX-2 inhibition on neonatal HI brain injury. The study led by Fathali et al. used postnatal day 10 rat pups to assess the effects of NS398, a known selective COX-2 inhibitor, on various neurologic outcomes after right common carotid artery occlusion followed by 2 h of hypoxia [12]. The authors first described that COX-2 inhibition limited morphologic damage, improved long-term functional deficits, reversed somatic growth retardation, and lowered mortality rates after a hypoxic-ischemic injury in neonatal rats. Of note, COX-2 inhibition significantly reduced the expression of IL-6, a proinflammatory cytokine, and in turn reduced the infiltration of inflammatory cells, such as macrophages and neutrophils, which suggests that the increased survivability and neuroprotection provided by COX-2 may be mediated by a reduction in neuroinflammation [12].

[14] 70w G-CSF is a 20-kDa protein belonging to the cytokine family of growth factors. It is responsible for stimulating the proliferation, survival, and maturation of cells committed to the neutrophil granulocyte lineage by binding to specific G-CSF receptors [72]. In addition to its role in neutropenia, G-CSF has been shown to be neuroprotective in various brain injury models through direct apoptotic inhibition, inflammatory cell modulation, and/or trophic effects on neuronal cells.

[15] 154w One of the first studies to investigate the anti-inflammatory properties of G-CSF was conducted in 1992 by Gorgen et al. who looked at the role of G-CSF in gram-negative septic rodents [73]. The study showed that G-CSF could interfere with TNF-a production through a negative feedback signal. Later, in 2005, both Gibson and Komie-Kobayashi would also demonstrate the anti-inflammatory role of G-CSF by showing that treatment could modulate the inflammatory response after injury [74,75]. Specifically, Komie-Kobayashi demonstrated G-CSF's ability to suppress inducible nitric oxide synthase (iNOS) production and decrease activation of microglial cells expressing iNOS -according to Western blot and immunohistochemistry analysis. Gibson, on the other hand, found that G-CSF treatment only suppressed the upregulation of IL-1b mRNA and had no effect on TNF-a and iNOS mRNA expression. Additionally, in models of peripheral infections, G-CSF-induced JAK-STAT signaling was found to reduce TNF-a, interleukin (IL)-1b, IL-2, IL-6, and IL-8, and elevate IL-1b receptor antagonists [76].

[16] 108w In terms of its neurotrophic capabilities, various studies have confirmed G-CSF as an essential neurotrophic factor, noting its ability to stimulate the release of stem cells from the bone marrow, promoting both neural repair and neural plasticity [72]. A study led by Shyu et al. in Circulation (2004) found that ischemic stroke rats treated with G-CSF could mobilize autologous hematopoietic stem cells into the circulation, enhance their translocation into ischemic brain, and significantly improve lesion repair [77]. Additionally, in rat ischemic models, peripherally administered G-CSF was found to enhance structural repair and function by increasing the number of newly generated neurons in both healthy and ischemic subjects [78].

[17] 142w In neonatal HI, only three studies can be found on PubMed using the search criteria "G-CSF and neonatal hypoxia ischemia." Unfortunately, none of the studies looked at the role of G-CSF as an anti-inflammatory agent. Instead, the focus of attention was on the role of G-CSF in apoptosis. The first study led by Yata et al. (2007) found that five 50 mg/kg G-CSF post-treatment injections over 4 days could reduce apoptotic neuron loss while increasing the expression of prosurvival signals [79]. Specifically, the investigators found that the anti-apoptotic protein Bcl-2 declined with injury and reversed after treatment, while the pro-apoptotic protein, Bax, increased following HI injury and again was reversed following G-CSF treatment. This is in line with another study conducted by Kim et al. (2008) that found similar neuroprotective outcomes following a single injection of 50 mg/kg G-CSF after injury [80].

[1] 153w In summary, the emerging landscape of neuroinflammation reveals highly complex interactions involving neurons, glia, and peripheral immune cells in the neonatal brain injured by HI. The mechanism of these multidirectional communications and their specific involvement in brain injury began to unveil only recently. It is however important to dissect molecular orchestrators of these interactions further in order to devise novel therapeutics with increased likelihood of success in clinical trials. Recent studies demonstrated a critical involvement of COX-2 in brain inflammation after HI and implied the use of COX inhibitors in treatment for neonatal encephalopathy in the clinical setting. While COX-2 appears to be a major neuroinflammatory mediator after HI, GCSF can negatively modulate inflammatory responses in the immature brain. Interestingly, COX-2 can also mitigate GCSF action and thereby can compromise neuronal survivability in the brain. Collectively, these finding suggest that COX-2 inhibitors and exogenous GCSF are promising treatment modalities on which to rely.

[2] 42w Based on the results obtained so far, it is also reasonable to anticipate that COX-2 plays a major role in mediating neuroglial interactions as well as in orchestrating immunological cell response in HI-induced brain injury. Further studies of this matter are warranted.

[3] 256w In the adult stroke model, splenectomy prior to cerebral ischemia reduced brain injury by elimination of the largest pool of immunological cells in the system. Consequently, it positively verified the involvement of peripheral immune cells in the mechanism of brain injury progression after stroke. Considering that all clinical trials of anti-inflammatory agents against stroke failed, these latest findings may point towards a reassessment as to whether anti-inflammatory therapies for stroke can reduce the peripheral immune system's involvement. These new evaluation criteria would also stand true for numerous candidate treatments of neonatal HI currently being tested in neuroscience laboratories, including G-CSF. In addition, it would be worthwhile to develop immunomodulatory therapies aimed at switching the immune response after HI to Th 2 cells. Significant progress in this field should not come as a surprise quite soon considering the latest studies employing "beneficial" subtypes of T cells (Th 2 /Th 3 ) to support neuroprotection and/or regeneration. However, the existing studies of neuro-glialimmune interactions have been conducted almost exclusively in adult stroke models. In addition, a few relevant studies of neonatal HI included unsexed animals. Thus, the peripheral immune involvement in neonatal HI awaits further investigations, considering distinct characteristics of the developing brain and immunologic immaturity of the neonate. In conclusion, the molecular circuitry of neuro-glial-immune communications in the hypoxic-ischemic neonatal brain with gender-specific investigations needs to be deciphered. It is believed that targeting master mediators of these interactions may pave the way to the first successful clinical trial with therapeutic agents that combat neuroinflammation after neonatal HI.

[1] 55w The immune response in the brain is highly complex and involves the participation of several different resident cells (Fig. 2). Microglia, astrocytes, and neurons directly react and contribute to neuroinflammation in the HI-injured neonate. The role of each of these cell types in propagating the local inflammatory response is important in understanding the dynamic microenvironment.

[2] 185w Cyclooxygenase is a rate-limiting enzyme responsible for catalyzing the synthesis of prostaglandins from arachidonic acid [55]. Cyclooxygenase possesses two catalytic sites: a COX active site responsible for the conversion of arachidonic acid into the endoperoxide, PGG2, and a peroxidase active site responsible for the rapid conversion of PGG2 into another endoperoxide, PGH2 [56]. PGH2 is further processed to form prostaglandins, prostacyclin, and thromboxane A2. To date, two COX isoforms have been identified. COX-1 is constitutively expressed in the brain, and its by-products are thought to contribute to normal physiological function [57]. COX-2 is also constitutively expressed in the brain (neurons, astrocytes, microglia, and endothelia), but can be inducible under pathological conditions [57]. In the brain, COX-2 acts as a key mediator of inflammation, orchestrating a wide spectrum of brain injuries, including excitotoxic brain injury, cerebral ischemia, traumatic brain injury, and neurodegenerative disorders [58]. COX-2 can propagate the neuroinflammatory response and contribute to tissue damage through the production of toxic prostanoids and reactive oxygen species [37]. COX-3 has also been reported in brain tissue [59], but is a splice variant of COX-1 with unknown function [60].

[3] 183w COX-2, a well-established mediator of adult brain injury [37], is emerging as a key player in neuroinflammation after hypoxic-ischemic brain damage in the neonate [12]. Peripheral immune cells such as T-lymphocytes, B-cells, and natural killer cells have the capability to upregulate COX-2 expression when activated [61]. However, immune cell infiltration into the brain parenchyma is thought to play a beneficial role also through the production of neurotrophic factors [62]. Activated monocytes, macrophages, and neutrophils are the peripheral cell source for the neurotrophic factor, G-CSF. Peripherally produced G-CSF can also enter the brain by crossing the intact blood-brain barrier and binding to its receptor on neurons and glial cells [63,64]. In the brain, G-CSF has been shown to protect neurons and trigger neurogenesis [65]. However, excessive and/or prolonged activation of inflammatory mediators can decrease neurotrophic support and neurogenesis in brain areas responsible for behavior and cognition [66][67][68]. Studies suggest COX-2 may mediate suppression of G-CSF, since inhibition of COX-2 was able to increase G-CSF production [69,70]. This downregulation of neurotrophic factors contributes to secondary brain injury and cell death after a hypoxic-ischemic insult [71].

[4] 85w There are various therapeutic modalities that have attempted to modulate the neuroinflammation that results from HI brain injury in neonates. These treatment options have targeted various stages in the inflammatory cascade, including COX-2 inhibition, and also investigated the use of growth factors such as G-CSF. In the following paragraphs, we will review these emerging therapeutic modalities and explore the various studies that have been conducted between 1970 and 2010 using relevant literature from the National Library of Medicine and National Institute of Health Database (www.pubmed.gov).