Programmed Necrosis: A Prominent Mechanism of Cell Death following Neonatal Brain Injury

Despite the introduction of therapeutic hypothermia, neonatal hypoxic ischemic (HI) brain injury remains a common cause of developmental disability. Development of rational adjuvant therapies to hypothermia requires understanding of the pathways of cell death and survival modulated by HI. The conceptualization of the apoptosis-necrosis “continuum” in neonatal brain injury predicts mechanistic interactions between cell death and hydrid forms of cell death such as programmed or regulated necrosis. Many of the components of the signaling pathway regulating programmed necrosis have been studied previously in models of neonatal HI. In some of these investigations, they participate as part of the apoptotic pathways demonstrating clear overlap of programmed death pathways. Receptor interacting protein (RIP)-1 is at the crossroads between types of cellular death and survival and RIP-1 kinase activity triggers formation of the necrosome (in complex with RIP-3) leading to programmed necrosis. Neuroprotection afforded by the blockade of RIP-1 kinase following neonatal HI suggests a role for programmed necrosis in the HI injury to the developing brain. Here, we briefly review the state of the knowledge about the mechanisms behind programmed necrosis in neonatal brain injury recognizing that a significant proportion of these data derive from experiments in cultured cell and some from in vivo adult animal models. There are still more questions than answers, yet the fascinating new perspectives provided by the understanding of programmed necrosis in the developing brain may lay the foundation for new therapies for neonatal HI.


Introduction
Neonatal hypoxic-ischemic encephalopathy (HIE) is a significant cause of mortality and morbidity in the pediatric population [1]. The therapeutic options for neonatal HIE are limited in part because the mechanisms of cellular degeneration in the immature brain are not fully understood. These mechanisms resulting from ischemia-reperfusion, oxidative stress, excitotoxicity and inflammation among others, activate or coactivate multiple pathways of cell death. Although, necrosis was initially described as the most prominent form of cellular degeneration following neonatal hypoxiaischemia (HI) [2,3], research emphasis switched to the study of apoptosis (programmed cell death type I) and autophagy largely due to advances in cell biology and to experimental animal studies on the molecular dissection of pathways for apoptotic and autophagocytic initiation and execution. The significance of necrosis in neonatal HI has been difficult to assess because of the presumed lack of a measurable regulatory pathway; however, the pathological evidence for necrosis has been well documented following HI [4,5]. We now know that necrosis can be regulated and programmed 2 Neurology Research International and that many components of the regulatory pathways are shared between different types of cell death opening a new window of opportunity for examination/reexamination of the cell death mechanisms in the neonatal brain with the goal of finding novel targets for therapy.
Based on morphological and biochemical data, we conceptualized that neurodegeneration in the neonatal brain is best classified according to an apoptosis-necrosis cell death "continuum" [6] and proposed that programmed cell necrosis (also called necroptosis in cell cultures) has a prominent contribution to neurodegeneration following HI [7]. It is certain that neonatal HI injury evolves through many cell death chreodes influenced by the dynamic injury landscape of the developing brain [8] and the mechanisms of injury in human neonatal HI are more complex than previously anticipated from experimental animal models. The accurate identification of the various cell death chreodes including programmed necrosis and their mechanisms unfolding within the immature brain will, in all likelihood, provide fresh rationale for the development of molecularbased therapies for neonatal brain injury following HI.

Programmed Cell Necrosis in Neonatal HI
Programmed necrosis as such has only recently been recognized as an important mechanism of injury in the immature brain following HI [7], however many aspects of programmed necrosis signaling have been comprehensively analyzed by the neonatal brain injury research community over the past decade (Table 1). This work piggybacks on a tremendous body of cell culture data on the mechanisms and contributions of programmed necrosis to cell death since the publication of 3 seminal papers in 1998, 2000, and 2003 [9][10][11]. This literature has been extensively reviewed recently [12][13][14][15][16][17].
We proposed that this novel regulated programmed necrosis, lies along the apoptosis-necrosis "continuum" and contributes meaningfully to several forms of acute neonatal brain injury [7,18]. The death domain containing serine/ threonine kinase, receptor interacting protein (RIP)-1, is central to the most well-described forms of programmed necrosis. Its kinase activity is selectively blocked by necrostatins and this affords protection against RIP-1-dependent forms of cell death [19,20]. Blockade of RIP-1 kinase using necrostatin provides protection in adult animal models of myocardial ischemia and ischemic and traumatic brain injury [18,21,22]. Similarly in neonatal HI, blockade of RIP-1 kinase attenuates brain injury at delayed stages in forebrain, hippocampus, and thalamus [7]. The necrostatins have been a major tool for investigation of RIP-1-dependent cell death pathways, however there are other tools that are now being used to explore RIP-1-dependent pathways and these will be discussed below.
The specific allosteric blockade of the kinase activity of RIP-1 has been studied extensively in cell cultures to demonstrate distinct signaling pathways leading to morphologic necrosis; however, many forms of necrosis in cultured cells, appear to proceed with different kinetics and not all are RIP-1 kinase dependent [23]. Some of the known and suspected RIP-1-independent programmed necrosis pathways include (i) caspase recruitment domain (ASC)-mediated necrosis, that is dependent of the non-catalytic activity of caspase-1 [24]; (ii) p53-cathepsin Q-mediated necrosis, that is activated by reactive oxygen species (ROS) and deoxyribonucleic acid (DNA) damage [25]; (iii) apoptosis inducing factor (AIF) and poly(ADP-ribose)polymerase-1-(PARP-1-) dependent pathways (controversy exists over the role of RIP-1 in these forms of programmed necrosis) [26][27][28][29][30]. These pathways to necrosis will not be emphasized since RIP-1dependent pathways are the focus of this paper and have been most extensively studied.

The Many Faces of RIP-1: Making the Decision between
Living or Dying. Maximal execution of RIP-1-mediated activation of programmed necrosis occurs in the setting of caspase inhibition [20,31] which can occur as a consequence of pharmacologic inhibition or significant mitochondrial dysfunction and adenosine-5 -triphosphate (ATP) depletion [32][33][34][35]. Others and we have hypothesized that energy failure interrupts the neonatal brain's proclivity to apoptosis [6,32,33,36] resulting in the hybrid, "continuum" cell death, or programmed necrosis morphology, possibly via activation of RIP-1 kinase [7]. Following activation of tumor necrosis factor (TNF) receptor (TNFR), RIP-1 signaling leads to a variety of cell fates and has been, for the most part, studied in cell culture [16]. In the setting of energy sufficiency, activation of members of TNFR superfamily (i.e; TNFR1, Fas death receptor (Fas-DR)) by their cognate ligands (TNFα and FasL, resp.), produce a conformational change in the receptor and recruitment of RIP-1, TNFR-associated death domain (TRADD), and TNFR-associated factor (TRAF) 2 and 5 to the cell membrane. Together these components constitute complex I [32]. TRAF2 recruits the cellular inhibitor of apoptosis (cIAP) that allows polyubiquitylation of RIP-1 leading to activation of p38-mitogen-activated protein (MAP) kinase, nuclear factor-kappa B (NFκB) and cell survival [37][38][39][40] (Figure 1). In a rodent model of neonatal HI, preservation of cIAP, via blockade of Smac/DIABLO, decreases injury size and improve outcomes [41], suggesting a possible role of RIP-1 ubiquitylation in cellular survival in this model. Likewise, preservation of RIP-1 ubiquitylation by genetic deletion of cylindromatosis (CYLD, deubiquitinating enzyme) in cultured cells results in resistance to TNFinduced programmed necrosis [42,43] which persists despite zVAD-fmk treatment (pan-caspase inhibitor) [44]. The roles of caspase 8 (known to cleave CYLD [44]), CYLD, and ubiquitylation of RIP-1 in determining activation of signaling pathways for programmed necrosis or survival are entirely unexplored territory in the investigation of neonatal brain injury following HI. Furthermore, RIP-1 ubiquitylation and complex I have been recently linked to cell death via Nox1 activation suggesting that many other modulators may play an important role in the elaborate intracellular signaling leading to cell survival or death [45] (Figure 1).
In the setting of energy insufficiency, activation of TNFR signals for cellular death via a variety of mechanisms is triggered by the degree of energy deficit. If cellular energy is only partially limited, RIP-1 polyubiquitylation declines

Continuum-programmed necrosis
Cell death phenotype Figure 1: The role of RIP-1 in programmed necrosis. Following neonatal HI, members of the TNFR superfamily (also called death receptors, DR) are activated by their ligands (DR-L) (i.e., FasL, TNF-α). In the setting of energy sufficiency and upon TNFR activation, TNFR undergoes a conformational modification of its cytoplasmic portion allowing the interaction with receptor interacting protein (RIP)-1 with the death domain (DD), TNFR-associated death domain (TRADD), and TNFR-associated factor (TRAF)-2 and -5. They in turn recruit the cellular inhibitor of apoptosis (cIAP) forming the complex I. cIAP inhibits caspase-3 activation and allows ubiquitylation of RIP-1. Next, transforming growth factor-β-activated kinase (TAK)-1/TAK-1 binding protein (TAB)-2/TAB-3 form a complex that binds to ubiquitin residues on RIP-1 and activates nuclear factor-κB (NFκB). This may occur via a p38 mitogen-activated-protein-kinase-(p38-MAPK-) dependent pathway. Complex I may interact with NADPH oxidase (NOX 1) producing ROS, also possibly triggering programmed necrosis. Deubiquitylation of RIP-1 by the enzyme cylindromatosis (CYLD) favors the transformation of complex I to complex II binding to the internalized death-inducing-signaling-complex (DISC, formed by FAS-associated protein with death domain (FADD) and procaspase-8 (Pro-C8)) and RIP-3 (Complex II). If energy is only partially insufficient, RIP-1 activates caspase-8 (C8) signaling for classical apoptosis via intrinsic (where truncated BID binds to the outer mitochondrial membrane allowing the release of cytochrome C (Cyt C) and triggering apoptosome formation) or extrinsic pathway resulting in caspase-3 activation. In this setting, caspase-8 cleaves RIP-1 and RIP-3 preserving signal for apoptosis; however, if energy failure evolves, caspase activity declines favoring (i) preservation of the RIP-1 kinase activity, (ii) decrease in RIP-3/FADD constitutive interaction, and (iii) autophosphorylation between RIP-1 with RIP-3 at the RIP homotypic interaction motif (RHIM) forming the necrosome. Necrosome induces reactive oxygen species (ROS) production via activation of NOX 1 at the cellular membrane or direct effects in the mitochondria. ROS cause DNA alkylation increasing activation of calpain-dependent poly(ADP-ribose)polymerase-1 (PARP-1) which is normally required for DNA repair. Hyperactivation of PARP-1 induces ATP depletion and apoptosis-inducing factor (AIF) translocation from the mitochondria to the nucleus which in turn produces further DNA damage and PARP-1 activation. Necrosome formation is a a potential intermediate step that follows PARP-1 activation potential intermediate steps that follow PARP-1 activation. There is some evidence that it produces ATP depletion via inhibition of adenine nucleotide translocase (ANT) in the inner mitochondrial membrane. Mitochondrial dysfunction is likely at the core of the events resulting in programmed necrosis.
favoring the transition of complex I to cytosolic complex II via internalization of activated TNFR and formation of the death-inducing signaling complex (DISC) containing TRADD, Fas-associated protein (FADD) and procaspase 8 [32,70,71]. When RIP-1 kinase is active, caspase-8 is cleaved and activated, initiating the intrinsic and extrinsic apoptotic cascades [72]. Activated caspase 8 can then cleave RIP-1 and RIP-3 and consequently limit programmed necrosis [73,74] (Figure 1). However, in the setting of more severe ATP depletion, caspase activity is inhibited allowing the Neurology Research International 5 formation of the RIP-1/RIP-3 complex, the necrosome, and cell death proceeds via programmed necrosis [10,11,75]. Interaction between RIP-1 and RIP-3 occurs at the RIP homotypic interaction motif (RHIM) which is the site of mutual phosphorylation [76]. Other RIP-1-dependent pathways do not require kinase activity as suggested by the lack of modulation of NFκB following RIP-1 kinase blockade with necrostatin in cell culture [19]. Once again, no studies have addressed the formation of complex II in vivo following neonatal HI. The interaction between FADD, RIP-1, and RIP-3 appears to be critical following TNFR activation [77]. RIP-1 is recruited to FADD in a TNF-dependent manner, while RIP-3 is more constitutively associated with FADD [78]. Following TNF exposure of cell cultures, FADD-deficient cells undergo RIP-3-and CYLD-dependent programmed necrosis with prominent inflammation, suggesting that FADD may prevent formation of the necrosome [79]. In addition to FADD, caspase 8 also seems to be necessary for survival of cultured cells due to its role in modulating CYLD activity and perhaps other functions [80]. In the developing mouse brain, there is abundant expression of caspase 8, TNFR, FAS death receptor, FADD, RIP-1, and RIP-3 [6,7,63]. In the normal developing brain, RIP-3 and FADD coimmunoprecipitate; following HI, RIP-1 is recruited to complex with RIP-3 disrupting RIP3's association with FADD [7]. These events are RIP-1 kinase dependent as proven by the partial restoration of RIP-3 and FADD association following treatment with necrostatin [7].
In the neonatal HI model, necrostatin not only provides neuroprotection but also partially shifts the death phenotype from necrosis to apoptosis validating the reality of the cell death continuum and providing insights into mechanisms that may drive the cell death continuum [6,7]. A similar finding has been reported in cell culture; knockdown of RIP-1 prior to TNFα exposure switches cell death from necroptosis to apoptosis [42]. Some factors that may permit a switch from necrosis to apoptosis in mice treated with necrostatin early after HI are (i) preservation of the mitochondrial function and consequently ATP production, (ii) inhibition of FLIP ((Fas-associated death-domain-like IL-1β converting enzyme)-inhibitory protein) gene and protein expression [7,81]; (iii) the fact that RIP-1 pathways leading to survival and apoptotic cell death are not kinase dependent [10,19,82]. We suspect that necrostatin-1, by blocking programmed necrosis, may allow a "cleaner" and less inflammatory form of cell death, similar to what is described for therapeutic hypothermia [83]. This possibility has not yet been explored.

Energy: The Driving
Force. Mitochondrial dysfunction and energy failure is a hallmark in necrotic cell death following neonatal HI [6,[84][85][86][87][88]. RIP-1-dependent necroptosis evolves with increased reactive oxygen species (ROS) production, decreased ATP production, and decreased mitochondrial membrane potential [89]. In cultured cells, nitric oxide inhibits NADH dehydrogenase (mitochondrial complex I) causing depletion of intracellular ATP and promoting a switch from apoptosis to necrosis [33,90,91]. Nitricoxide-(NO-) induced inhibition of mitochondrial complex I is reversible at low concentrations [91][92][93] but irreversible at high concentrations resulting in additional free radical production [94,95]. After neonatal HI, inducible nitric oxide synthase (iNOS) expression and NO accumulation increase, events that are followed by a progressive decline in complex I activity in forebrain during the first 24 h (unpublished data, Pediatric Academic Society Meeting 2011 abstract 2170.2; Neuroscience 2012, submitted). This decline in complex I activity results in a significant impairment in ATP production at early stages following HI that is also prevented by blockade of RIP-1 kinase [96]. Blockade of RIP-1/RIP-3 complex formation in cell culture using necrostatin or RIP-1 siRNA prevents 3-nitrotyrosine accumulation and nitrosylation of complex I and attenuates NO-dependent necrosis [95] similar to findings in the neonatal in vivo HI model. These data are consistent with the hypothesis that an intact mitochondrion is initially required to produce physiological superoxide (O 2 − ) that will react with NO to generate peroxynitrite (ONOO − ) resulting in mitochondrial membrane potential loss [97,98].
The link between programmed necrosis and opening of the mitochondrial permeability transition pore (MPTP) complex is controversial [22,99]. However, RIP-1 appears to have direct effects in cellular energy production by translocating to the mitochondria and suppressing ADP/ATP exchange [20,100] in cell culture. In concert with these findings, necrostatin also prevents the reduction in mitochondrial membrane potential caused by excitotoxic stimuli [101].

Free Radicals
Targeting the Mitochondria. RIP-1 kinase activity is essential for cell death to proceed via the most well-recognized form of programmed necrosis. RIP-1 kinase activity mediates the formation of the necrosome (RIP-1/RIP-3 complex) which induces ROS production via effects on (i) Nox 1 nicotinamide adenine dinucleotide phosphate (NADPH) oxidase and (ii) the mitochondria [23,45,102]. Nevertheless, necrostatin is not a direct antioxidant and does not prevent cell death caused by hydrogen peroxide in culture [12,103]. However, much like hypothermia, inhibition of RIP-1 kinase activity attenuates oxidative injury to proteins following neonatal HI in the mouse and piglet [7,83]. Similarly, genetic deletion of RIP-3 gene or treatment with RIP-3 silencing RNA (siRNA) in cultured cells prevents increase in ROS and programmed cell necrosis [78]. Potential oxidative injury mechanisms targeted by the blockade of programmed necrosis include (i) blockade of nitric-oxide-mediated mitochondrial dysfunction caused by lipopolysaccharides (LPS) stimulation of macrophages [95], (ii) inhibition of glutamate excitotoxicity [103], (iii) increased glutathione levels [103], and (iv) decreased ROS production [103].
Glutathione (GSH) levels decrease following both excitotoxic and HI insults but blockade of RIP-1 kinase with necrostatin increases GSH production in HT-22 cells following glutamate exposure [45,103]. In the neonatal HI mouse model, treatment with necrostatin appears to prevent glutathione oxidation rather than increasing GSH production per se [96]. This finding may reflect an indirect effect of the prevention of early protein carbonyl formation afforded by necrostatin-1 after neonatal HI [7] or it may simply be an indirect consequence of neural cell protection.
Recently, a role for Bcl-2/adenovirus E1B 19 kDa-interacting protein 3 (BNip3) has been described in a programmed necrotic-like cell death [104]. This BH3-only protein subfamily includes two members: BNip3 (also called NIP3) and BNip3L (also called NIX or BNip3-like) each with different recognized functions [105,106]. BNip3 (30 kDa monomer) binds loosely to the outer mitochondria membrane (OMM) [107]. Free radical accumulation induces BNip3 dimerization and insertion into the OMM triggering necrotic-like cell death [104,108]. In models of neonatal HI, necrostatin prevents early iNOS expression and NO accumulation and blocks hypoxia-inducible factor (HIF)-1α expression (unpublished data), a transcription factor that binds to the hypoxia response element (HRE) at the BNip3 promoter [109,110]. Because NO modulates HIF-1α expression via Ras modification and phosphorylated extracellular-signal-regulated kinase (ERK) nuclear accumulation [109], it is possible that by preventing NO accumulation, necrostatin could indirectly decrease HIF-1α and consequently BNip3 expression following neonatal HI, protecting the mitochondria and preventing the progress of programmed necrosis. The second member of the BNip3 subfamily, BNip3L, has dual, but distinct, actions depending on the targeted organelle, mitochondria, or endoplasmic reticulum [106]. Although BNip3L has not been studied in models of neonatal HI, there is data from cellular cultures. At the mitochondria, BNip3L induces Bax/Bakdependent OMM permeabilization, cytochrome c release, caspase activation and apoptosis, while, at the endoplasmic reticulum, BNip3L induces acute release of luminal Ca 2+ that triggers cyclophilin-D-dependent MPTP complex opening, mitochondria swelling, mitochondrial membrane potential loss, ATP depletion, release of free radicals, and cellular necrosis [106]. Conversely, Bax/Bak has been also associated with programmed necrosis via release of AIF and mitochondrial depolarization [89,111]. Therefore, both members of the BNip3 subfamily can be classified as sensors of mitochondrial stress as suggested previously [112] and because its expression is modulated by stimuli that are very well-recognized in association with HI, it is possible that both, BNip3 and BNip3L, are linked with the mitochondrial dysfunction seen following HI.
The pathways linking RIP-1 activity and RNS production are mostly unknown. Increased NO accumulation and iNOS expression potentiates glutamate release, N-methyl D-aspartate receptor (NMDAR) activity, necrotic neuronal death, and progression of excitotoxic injury in cell cultures [33,113,114]. Allosteric inhibition of RIP-1 kinase prevents the RNS formation as evidence by the decreased nitration of the NDUFB8 subunit preventing mitochondrial complex I dysfunction and depolarization [95]. Unpublished experiments from our laboratory are in agreement with these finding suggesting that blockade of RIP-1 kinase activity following neonatal HI decreases NO accumulation by 70% coincidently with a decrease in iNOS expression (unpublished data, Pediatric Academic Society Meeting 2011 abstract 2170.2). It remains unknown which mechanisms are operative and if they are directly linked to the inhibition of programmed necrosis. Anti-iNOS/NO effects of necrostatin may involve modulation of inflammatory mediators since cytokines are primary activators of iNOS production by astrocytes and necrostatin decreases cytokine expression [7,115].
Ultimately, overproduced ROS and RNS attack the mitochondria, depleting ATP production and allowing programmed necrosis to proceed. ROS induces DNA alkylation, an event that increases the levels of calpain-dependent PARP-1 required for DNA repair [27,28] in the setting of caspase 8 inhibition. Hyperactivity of PARP-1 following glutamate excitotoxicity produces poly-ADP-ribose (PAR) accumulation and ATP depletion inducing translocation of AIF from the mitochondria to the nucleus via a c-Jun-Nterminal-kinase-(JNK)-1-mediated mechanism resulting in chromatin condensation and DNA fragmentation [29,30]. The importance of PARP-1 activation and AIF translocation in the neonatal brain after HI appears to be gender specific [26,50]. PARP-1 level peaks at 30 min and again at 12 h following neonatal HI [68] along with an early decrease in nicotinamide adenine dinucleotide (NAD + ) in male mice [26]. Furthermore, PARP-1 genetic deletion [26] or inhibition [67] provides neuroprotection following neonatal HI in male but not female mice. Blockade of calpains, required for PARP-1 activation, using MDL28170 [53] or hypothermia [54] or blocking JNK pathway [41] also decreases necrotic injury after HI. The degree of AIF translocation to the nucleus, also greater in male mice [50], correlates with the infarct size following neonatal HI [46] and its inhibition by heat shock protein (Hsp)-70 [48] TAT-Bcl-xL [49] or hypothermia [51] provides neuroprotection. Although still unclear, steps following PARP-1 activation may include RIP-1 activation as evidenced by the protection against DNA alkylation in RIP-1 knockdown mouse embryonic fibroblast [29]. Altogether, these data suggest an important role of a PARP-1-AIF feedback cycle in the events leading to brain injury following neonatal HI, direct evidence of interaction of AIF with RIP-1 (or the necrosome) has yet to be reported in the immature brain.

Inflammation and Programmed
Necrosis. The importance of inflammation following HI has been extensively studied in the immature brain [116][117][118]. In normal physiology, a primary function of RIP-1 is to transduce the NFκB signal leading to survival, hence RIP-deficient mice fail to thrive and die within three days after birth with extensive lymphoid apoptosis associated with failure to activate NFκB due to unfavorable conditions to form complex I [32,119]. Cell culture studies failed to show that RIP-1 kinase modulates NFκB activation [19]. However, in vivo, we have shown that blockade of RIP-1 kinase activity using necrostatin following neonatal HI is associated with prevention of early increase in nuclear translocation of NFκB [7]. This effect is Neurology Research International 7 likely indirect but may be of significance given the toxicity associated with early increases in NFκB levels after neonatal HI [69,120]. Additional confirmation of a possible indirect modulatory effect on NFκB is that transcription of FLIP is downregulated following RIP-1 kinase blockade [7]. Because FLIP is under transcriptional control by NFκB, the decline in early FLIP [121] expression following blockade of RIP-1 kinase with necrostatin may be a reporter for changes in NFκB activity.
NFκB is a transcription factor that also mediates important apoptotic and inflammatory pathways which are central to HI-mediated brain injury in the immature brain [69,120,122]. Innate immune responses are dependent on activation of toll-like receptors (TLRs), recruitment of myeloid differentiation primary response gene (MyD)88 and interleukin-1 receptor-associated kinase (IRAK), association of TRAF6 and MAP3K, phosphorylation of I kappa B kinase (IKK) and release and nuclear translocation of the transcriptional factor NFκB (p65/RelA/p50), resulting in change in cytokine expression [122]. Other proinflammatory receptors linked to NFκB include the nucleotide-binding oligomerization domain (NOD) which with the interleukin (IL)-1 converting enzyme protease-activation factor (IPAF) activates caspase 1 (IL-1β converting enzyme) and forms the inflammasome [123][124][125]. Further details about the inflammatory pathways triggered by NFκB activation may be reviewed elsewhere [122]. Current understanding of the "crosstalk" between programmed necrosis and inflammatory pathways is very limited; however certain interactions can be suspected based on current data. Blockade of programmed cell necrosis and cytokine expression in the neonatal HI model following treatment with necrostatin suggest that inhibition of RIP-1 kinase decreases the activation of the inflammasome, as shown by decreased caspase 1 activity and decreased transcription of IL-1β [7]. Furthermore, TNFα and IL-6 are also downregulated in mice treated by necrostatin following neonatal HI, suggesting that RIP-1 kinase modulates neuroinflammation. However, it remains unclear if these anti-inflammatory changes are a direct effect of blockade of programmed necrosis pathway or whether they are secondary to the overall neuroprotection.
Although astrocytes provide support to neurons, they also release cytokines that instigate and perpetuate neuroinflammation [126]. TLR are constitutively expressed in astrocytes and modulation of these receptors following HI has been characterized [127]. Following induction of programmed necrosis, reactive astrocytes release cytokines and express iNOS [128], suggesting that changes in the cytokine profile associated with RIP-1 kinase blockade in HI may be related to an effect on astrocytes. Our preliminary results show that following neonatal HI, necrostatin decreases iNOS and cytokine expression while preserving astrocyte mitochondrial ultrastructure and attenuating glial fibrillary acidic protein (GFAP) expression at later stages. One possible hypothesis explaining the neuroprotective and antiinflammatory effect associated with RIP-1 kinase inhibition is that in vivo astrocytes are a primary therapeutic target of necrostatin and by protecting and preserving astrocyte structure and function, it protects neurons and prevents neuroinflammation.

Gender Differences in Programmed Necrosis.
Gender differences have been reported in neonatal rodent models of HI brain injury [7,26,50]. These differences may result from intrinsic differences in primary injury pathways. We found a more robust neuroprotection in males than females in response to programmed necrosis blockade [7]. Mechanisms explaining these gender differences are unresolved, but may involve an effect of necrostatin on the more significant decline in NAD + following PARP-1 activation [26] and the preferential nuclear translocation of AIF [50] found in male rodents following neonatal HI. Therefore, necrostatin's blockade of RIP-1/RIP-3 interaction, oxidative damage, and inflammation may reflect mechanisms of action upstream and downstream of AIF translocation in male rodents.

Conclusions
Neonatal HI brain injury remains a common cause of developmental disability despite ongoing advances in obstetrical and neonatal care. With the advent of hypothermia for treatment of some infants with HI, morbidity has begun to decrease [129]. However, hypothermia is only partially neuroprotective after neonatal HI and 45% of all treated infants still suffer severe neurodevelopmental disability or death despite treatment [130]. Development of adjuvant therapies for hypothermia treatment has been limited to date. Novel approaches to understanding neurodegeneration after neonatal HI are needed. The conceptualization of the apoptosis-necrosis "continuum" in neonatal brain injury in 1997 predicted important mechanistic interactions between apoptosis and necrosis pathways [131]. Evidence of programmed necrosis in neonatal HI is in complete agreement with this sentinel observation and provides an important new direction for future research [7]. Programmed necrosis has been well studied in cellular cultures with new findings published routinely but the recognition of its importance in neonatal HI is just beginning. Many components of the signaling pathway now known to also regulate programmed necrosis have been studied over the last decade in models of neonatal HI as part of the apoptotic pathways showing the clear overlap of these pathways ( Table 1). As we now begin to understand the contribution of programmed necrosis to neural cell fate following HI injury, we should take a fresh look at previous findings from these earlier studies. However, many questions remain unanswered with respect to programmed necrosis and neonatal HI including (i) direct effect, if any, of RIP-1 (or the necrosome) in disruption of mitochondrial bioenergetics; (ii) role of calpainmediated lysosomal destabilization in the progression of injury; (iii) link between RIP-1 and PARP-1-AIF feedback cycle; (iv) identification of neural cell types most vulnerable programmed necrosis and the role of individual neural cell types in propagation or resistance to programmed necrosis; (v) the cellular mechanisms activated following necrosome formation in the immature brain; (vi) whether specific inhibitors of programmed necrosis will be clinically 8 Neurology Research International useful; (vii) what effect, if any, current therapies have on programmed necrosis following HI. Studies such as these will provide new perspectives on the mechanisms of neuronal cell death in vivo and may lay the foundation for new effective therapies for neonatal HI.