Perinatal Hypoxic-Ischemic Encephalopathy and Neuroprotective Peptide Therapies: A Case for Cationic Arginine-Rich Peptides (CARPs)

Perinatal hypoxic-ischemic encephalopathy (HIE) is the leading cause of mortality and morbidity in neonates, with survivors suffering significant neurological sequelae including cerebral palsy, epilepsy, intellectual disability and autism spectrum disorders. While hypothermia is used clinically to reduce neurological injury following HIE, it is only used for term infants (>36 weeks gestation) in tertiary hospitals and improves outcomes in only 30% of patients. For these reasons, a more effective and easily administrable pharmacological therapeutic agent, that can be used in combination with hypothermia or alone when hypothermia cannot be applied, is urgently needed to treat pre-term (≤36 weeks gestation) and term infants suffering HIE. Several recent studies have demonstrated that cationic arginine-rich peptides (CARPs), which include many cell-penetrating peptides [CPPs; e.g., transactivator of transcription (TAT) and poly-arginine-9 (R9; 9-mer of arginine)], possess intrinsic neuroprotective properties. For example, we have demonstrated that poly-arginine-18 (R18; 18-mer of arginine) and its D-enantiomer (R18D) are neuroprotective in vitro following neuronal excitotoxicity, and in vivo following perinatal hypoxia-ischemia (HI). In this paper, we review studies that have used CARPs and other peptides, including putative neuroprotective peptides fused to TAT, in animal models of perinatal HIE. We critically evaluate the evidence that supports our hypothesis that CARP neuroprotection is mediated by peptide arginine content and positive charge and that CARPs represent a novel potential therapeutic for HIE.


Introduction
Perinatal hypoxic-ischemic encephalopathy (HIE; also referred to as birth asphyxia) remains the leading cause of neonatal mortality and morbidity, with an incidence in developed nations of 2-6 and 7 in every 1000 live term (>36 weeks gestation) and pre-term (≤36 weeks gestation) births, respectively [1,2]. The incidence is even higher in developing countries, with rates of up to 30 per 1000 live full-term births [3,4]. Of the individuals affected by HIE, 15-20% will die in the postnatal period, and an additional 25% will develop severe and permanent neurological sequela. Each year, HIE accounts for up to 1.2 million deaths [5] and some form of central nervous system (CNS) dysfunction in an estimated 1.15 million neonates [6] such as cerebral palsy, epilepsy, global developmental delay, intellectual disability, behavioral disorders and autism spectrum disorders [7][8][9][10][11]. While hypothermia is used as a therapeutic intervention to minimize brain injury in neonates suffering HIE, it only provides beneficial outcomes in approximately 30% of patients and has several limitations (reviewed below). Therefore, an effective and easily administrable pharmacological therapeutic agent to improve patient outcomes when used as an adjunct to hypothermia, on its own or when hypothermia cannot be applied is urgently needed.
In recent years, there has been an increased application of putative neuroprotective peptides designed to target cyto-damaging or cyto-protective pathways to reduce injury in acute brain disorders such as HIE. One reason for this interest is the discovery of cell-penetrating peptides (CPPs), allowing for the delivery of neuroprotective peptides and other molecular cargos (e.g., proteins, nucleic acids and nanoparticles) across the blood brain barrier (BBB) and into cells within the CNS. Most studies assessing neuroprotective peptides in animal models of HIE use the cationic arginine-rich CPP, transactivator of transcription (TAT 47-57 YGRKKRRQRRR), as the 'carrier' molecule. However, several years ago, our and other laboratories demonstrated that the TAT peptide has modest intrinsic neuroprotective properties in its own right [12][13][14][15][16][17]. We subsequently demonstrated that other cationic arginine-rich peptides [CARPs; e.g., penetratin, protamine, sodium calcium exchanger inhibitory peptide (XIP), poly-arginine peptides (e.g., R9, R12 and R18; R indicates amino acid arginine and number indicates mer)], possess potent neuroprotective properties in both in vitro and animal models of cerebral ischemia and/or hypoxia [12,[18][19][20][21][22][23][24][25][26][27][28][29], and demonstrated that the arginine content and peptide positive charge are critical determinants of neuroprotection [19,27,28].
Based on the recent application of cationic CPPs for the delivery of putative neuroprotective peptides in animal models of HIE and the recent discovery of the intrinsic neuroprotective properties of CARPs, the objectives of this review are: (i) catalogue studies that have used peptides as neuroprotective agents in HIE, especially those employing a CPP delivery system; and (ii) to evaluate the potential neuroprotective effects of both CPP and cargo molecules in terms of their arginine content and cationic charge, as well as the content of other potentially neuroprotective amino acids such as tryptophan [29]. We will propose that the neuroprotective properties of most, if not all putative neuroprotective peptides fused to cationic CPPs are likely to be mediated in a large part by the actions of the carrier CPP (e.g., TAT) rather than the cargo molecule itself, with potency being further enhanced according to the amino acid content of the cargo (e.g., number of arginine residues) [19]. In addition, we will also review the evidence that CARPs have multiple cellular modes of action that underlie their neuroprotective properties.
First, we provide a brief overview of the major underlying pathophysiological mechanisms associated with HIE and the application of hypothermia as an acute therapy for the disorder.

Initiation of the Pathophysiological Cascade in HIE
The etiology of HIE is associated with a variety of maternal, placental and fetal conditions; all of which have variable clinical manifestations. These conditions include, but are not limited to, chronic maternal hypoxia (e.g., chronic utero-placental hypoxia, cardiopulmonary arrest, acute hypotension, pulmonary embolism or vascular disease), pre-eclampsia, nuchal cord, umbilical cord knotting, umbilical cord prolapse, shoulder dystocia and premature placental detachment [30]. Perinatal hypoxia is initiated following an impairment of oxygenated cerebral blood flow (CBF) to the fetus, triggering responses at both the systemic and cellular levels. The fetal brain is particularly sensitive to a HI environment, as there is a requirement of a constant supply of energy in the form of ATP, and when interrupted, excitotoxicity occurs through the uncontrolled release of excitatory neurotransmitters such as glutamate; marking the beginning of the ischemic cascade. The acute and downstream consequences of the ischemic cascade are damaging to neurons and other cells (e.g., glial progenitor cells and astrocytes) at the cytoplasmic and mitochondrial levels, as well as causing disruption to the BBB and the activation of inflammatory responses.

Excitotoxicity
The metabolic rate of the fetal brain is largely determined by the stage of gestational development [30]. To meet metabolic demands, developing neural tissue metabolizes lactate, ketone bodies and glucose. When compared to the developed brain, the fetal brain is more adaptable with respect to energy utilization, thereby increasing its capacity to tolerate hypoxia-ischemia (HI) [31]. However, in the event of a critical depletion of ATP due to prolonged HI, the fetal brain eventually becomes susceptible to injury.
Following prolonged HI, cellular homeostasis is disrupted due to the depletion of ATP and failure to maintain ionic gradients, and as a result, neurons depolarize and release glutamate into the synaptic cleft [32]. High extracellular levels of glutamate cause excitotoxicity in neurons and other cells, namely glial progenitor cells, that express glutamate receptors (NMDA; N-methyl-D-asparate, AMPA; α-amino-3-hydroxy-5-methy-4-isoxazolepropionic acid, kainic acid and metabotropic glutamate receptors), producing a cytotoxic cellular influx of calcium ions [33]. The increased levels of intracellular calcium in neurons and glial progenitor cells in turn results in the activation of calcium-dependent proteases, lipases and deoxyribonuclease (DNase), reactive oxygen species (ROS) production, oxidative stress, cytotoxic edema, mitochondrial dysfunction and the stimulation of pro-cell death pathways [34][35][36][37][38].

Oxidative Stress
Oxidative stress plays an important role in the pathophysiology of HIE, as the developing brain comprises high levels of unsaturated fatty acids susceptible to lipid peroxidation, metals catalyzing free radical reactions and low concentrations of antioxidants [39]. The resulting heightened sensitivity to oxidative stress and ROS production following HI causes significant damage to lipids (lipid oxidation), nucleic acids (DNA degeneration) and proteins (protein oxidation). Sources of ROS following HI include, the mitochondrial electron transport chain (ETC), NADPH oxidases (NOX), xanthine oxidase (XO), arachidonic acid (12/15 lipoxygenase), and nitric oxide synthase (NOS). In particular, NOS has been targeted in HIE, with strategies to ablate or inhibit NOS activity invariably being neuroprotective in animal models of HIE [40].

Mitochondrial Dysfunction
Mitochondria have diverse functions, including ATP generation, intracellular calcium regulation, ROS production, biosynthesis of amino acids, lipids and nucleotides and pro-cell death signaling. The role of mitochondria in HIE has been extensively reviewed [41][42][43]. Here we focus on the role of mitochondria in potentiating secondary pathophysiological cascades contributing to HI. Mitochondria contribute to secondary brain injury following HI via the regulation of cell death pathways, namely apoptosis. Cellular apoptosis following HI is activated via both intrinsic (e.g., in response to mitochondrial dysfunction and damaged DNA) and extrinsic [e.g., in response to cell death receptors such as Fas and tissue necrosis factor-α (TNF-α)] pathways and involve the release of pro-cell death proteins from mitochondria.
In cells affected by HI, several processes such as excitotoxicity, ROS production, activation of p53, DNA damage and altered mitochondrial function can influence the activity of pro-apoptotic Bcl-2 proteins (e.g., Bid and Bax), resulting in Bax-dependent mitochondrial membrane permeability transition (MPT) [44,45]. In addition, excessive mitochondrial ROS generation, due to leakage from the ETC, can oxidize the inner mitochondrial membrane phospholipid, cardiolipin, promoting mitochondrial membrane permeability transition pore (MPTP) opening and subsequent release of cytochrome-c. Cytosolic cytochrome-c, along with deoxyadenosine triphosphate (dATP), interact with apoptotic protease activating factor-1 (Apaf-1) to form the apoptosome, which cleaves and activates pro-caspase-9 [46]. Activated caspase-9 in turn activates caspase-3, which is a major proteolytic enzyme responsible for the ultimate dismantling of cells during apoptotic cell death.
Further alterations to mitochondria, such as mitochondrial outer membrane permeabilization (MOMP), results in the release other pro-apoptotic factors into the cytosol, including apoptosis-inducing factor (AIF), endonuclease G (Endo-G), and Smac/DIABLO [47,48]. Translocation of AIF and Endo-G to the nucleus mediates chromatin fragmentation and Smac/DIABLO interacts with inhibitors of apoptosis to reduce their actions on activated caspases.
TNF cell death receptor signaling mainly stimulates the extrinsic pathway by activating caspase-8, which activates other caspases, including caspase-3. While beyond the scope of this review, it is also important to note that other cell death pathways associated with necrosis, necroptosis, ferroptosis and autophagy may also be activated in response to HI. These pathways can be activated in response to one or more cellular disturbances such as calcium influx, mitochondrial dysfunction, cellular energy depletion, ROS production, TNF receptors cell signaling and other cellular cell signaling pathways (e.g., adenosine monophosphate-activated protein kinase).

Inflammation
The inflammatory response following HIE has been extensively reviewed by [49,50] and is therefore briefly discussed. In the immature brain following HI, an innate immune response occurs within minutes of the injury [51]. Initially, microglia become activated within the cerebral parenchyma, followed by a systemic inflammatory response, mediated by the infiltration of circulating monocytes, neutrophils and T-cells into the brain. Activated microglia develop macrophage-like attributes, such as phagocytic properties, cytokine production, antigen presentation and matrix metalloproteinase (MMP) release. The release and activation of MMPs by microglia can compromise BBB integrity, further promoting the invasion of peripheral leukocytes into the cerebral parenchyma; exacerbating ensuing cerebral injury. A hallmark of HIE is the aggregation of amoeboid microglia within the periventricular white matter, resulting in the production of inflammatory cytokines (TNF-α and IL1-β), NO and other ROS [52]; collectively potentiating the toxic effects of the ischemic cascade in neurons, glial progenitor cells and the cerebral vasculature.
While astrocytes attempt to provide neuroprotective support to neurons and glia via the release of glutathione, superoxide dismutase (SOD), extra-synaptic uptake of glutamate and the maintenance of ion channel gradients, these protective mechanisms can be quickly overrun. As with microglia, astrocytes become hyper-activated due to the effects of pro-inflammatory cytokines, ROS, and dying neurons and glia. Activated astrocytes also release pro-inflammatory cytokines (IL-6, TNF-α, IL-1α/β and INF-γ), further exacerbating HI-induced cell death pathways (e.g., apoptosis). In addition, astrocytes secrete chemokines that attracts systemic immune cells to infiltrate affected tissue, further aggravating tissue injury.

Current Clinical Treatments: Hypothermia
Hypothermia is considered one of the most effective neuroprotective treatment interventions in both experimental and clinical settings. While the exact beneficial effects of hypothermia are yet to be elucidated in HIE, it is known to suppress many of the neuro-damaging events associated with HI including excitotoxicity, apoptosis, inflammation, oxidative stress and BBB disruption [53].
In HIE, moderate hypothermia (33.5 • C for 72 h), applied within 6 h of birth, improves the survival and neurodevelopmental outcomes in infants suffering from moderate and severe HIE. While hypothermia is now considered standard care in the treatment of HIE and is well tolerated in term infants [54][55][56][57], it does not provide complete protection. Of the infants who have moderate or severe encephalopathy, therapeutic hypothermia was shown to decrease mortality from 40% to 28%, while in surviving infants neurological morbidity was reduced from 31% to 19% of infants receiving therapeutic hypothermia [58].
Importantly, while hypothermia is safe in pre-term infants with necrotising enterocolitis [59], it has not been adequately evaluated for use in pre-term infants suffering HIE. Information regarding the safety and efficacy of hypothermia to treat pre-term infants suffering HIE is lacking, although information may be available in the near future (NCT1793129). In one study, hypothermia in pre-term infants was associated with higher mortality and increased clinical complications when compared to term infants [60], although the study did not include a normothermic pre-term control group. The inadequate application of hypothermia in the pre-term neonate highlights the need for the development of safer therapeutic interventions in these patients following HI.
In addition, the induction of hypothermia requires specialized equipment, intensive care monitoring and trained staff, limiting its use to centers that have the facilities and medical personnel to undertake the procedure. This develops uncertainty regarding the timing, safety and efficacy of initiating hypothermia in infants before transfer to a centre equipped to deliver hypothermia. For example, the transfer of infants suffering HIE to a medical centre to provide hypothermia can exceed 6 h from the time of birth and that attempts of passive cooling during transport can result in excessive cooling [61]. In light of this, there is an urgent need for the application of an acute neuroprotective therapy for infants who suffer HIE when hypothermia cannot be applied immediately, or is not available. In addition, as the incidences of mortality and morbidity are still relatively high following hypothermia, there is a requirement for additional therapeutic agents, adjuvant to hypothermia, to further improve outcomes in infants suffering HIE.

Peptide Therapeutics
Peptides represent an important and increasingly popular class of therapeutic agent. To date, over 60 general peptide drugs have been approved in the United States, Europe, and Japan, approximately 140 are in active clinical development and more than 500 peptides are being examined in preclinical therapeutic studies [62]. Peptides have several beneficial properties over traditional drugs such as high biological and structural diversity, less toxicity and accumulation in tissue and reduced propensity to cause drug-drug interactions. The main disadvantage of most peptides is their inability to cross the BBB and enter cells, which is an important issue if the peptide has an intracellular target and/or is targeting a CNS disorder. However, as outlined above, the delivery of peptides into the CNS and into cells can be achieved by using CPPs. For this reason, CPPs have been widely utilized as a carrier molecule for the delivery of neuroactive compounds into the CNS, including putative neuroprotective peptides. Below we review experimental studies that have used neuroprotective peptides in HIE models, including putative neuroprotective peptides fused to a CPP. First, we summarize findings from our own and other laboratories highlighting the intrinsic neuroprotective properties of CARPs.
Our laboratory has also demonstrated that CARPs and in particular the poly-arginine peptide R18 are neuroprotective in in vivo experimental brain ischemic and/or hypoxic injury models [18,[20][21][22]24,26,27,29], and that CARPs can reduce neuronal cell surface glutamate receptor levels and excitotoxic calcium influx [19,22,23,25,27,28]. As highlighted previously [19], based on our findings, we have proposed that the neuroprotective properties of most, if not all, putative neuroprotective peptides fused to cationic CPPs (e.g., TAT, R9 and penetratin), as well as other CARPs, are determined by the arginine content and positive charge of the peptide, with potency also influenced by other amino acid residues (e.g., lysine, tryptophan and phenylalanine) [19,29]. Importantly, an arginine and cationic charge mediated neuroprotective mechanism of action of CARPs, including CPP-fused to putative neuroprotective peptides (e.g., TAT-NR2B9c and TAT-NR2Bct), is supported in studies by Marshall et al. [65] and more recently McQueen et al. [66]. Based on studies from our own and other laboratories, we believe that the mechanism of action of putative neuroprotective peptides fused to a CPP need to be critically re-evaluated. In light of the potentially confounding effects of CPPs and/or peptide arginine content and positive charge, we will review studies using neuroprotective peptides in HIE models. First, we provide a summary of the known and potential neuroprotective actions of CARPs.
Interestingly, the neuroprotective potency of CARPs appears to correlate with the same characteristic that improve the ability of cationic CPPs to target and transverse cellular membranes [19,29].
This suggests that CARP-mediated neuroprotection is associated with cellular membrane interactions (e.g., plasma and mitochondrial), anionic structures (phospholipids/cardiolipin, heparan sulfate proteoglycans, chondroitin sulfate proteoglycans, sialic acid resides, anionic regions of surface receptors) and/or increased levels of cellular uptake, which enable the peptides to interact with membrane (e.g., ion channel receptors) and intracellular targets (e.g., proteins, ROS and mitochondria).

Studies Using CARPs in Animal Models of HIE
All descriptive details of the studies discussed below are presented in chronological order of publication (Table 1).
The first study examining the neuroprotective efficacy of the NBD peptide in a model of HIE (P12 rat) administered the peptide [(14,820 nmol/kg: intraperitoneally (IP)] immediately, 6 and 12 h after hypoxia and did not demonstrate any neuroprotection; instead NBD appeared to increase brain injury [104]. To improve delivery to the brain, the NBD peptide was fused to TAT (TAT-NBD: YGRKKRRQRRR-TALDWSWLQTE; net charge +6) and assessed in a P7 rat model of HIE [105], where administration (6917 nmol/kg: IP) immediately and 3 h, or as a single injection at 0, 3 or 6 h, after hypoxia was neuroprotective up to 6 weeks post-HI. When TAT-NBD was administered immediately and 3 h after hypoxia it prevented up-regulation of p53 and activation of caspase-3 24 h post-HI. TAT-NBD treatment also prevented the activation of IKK in the brain 3 h post-HI. Despite TAT-NBD inhibiting IKK and NF-κB activation, it had no effect on HI-induced increases in cytokines (e.g., IL-1β, TNF-α, IL-4, IL-10 and IL-1-RA). The study also assessed a TAT-NBD control peptide, where two tryptophan residues were substituted for two alanine residues (YGRKKKRRQRRR-TALDASALQTE; net charge +6) and was shown not to be neuroprotective. Further assessment of TAT-NBD (6817 nmol/kg: IP) involving administration immediately and 3 h, immediately, 3, 6 and 12 h, immediately, 6 and 12 h, as well as 18 and 21 h after hypoxia [106], revealed long-term neuroprotection only in animals treated immediately and 3 h post-HI. Concomitantly, TAT-NBD treatment immediately and 3 h after hypoxia significantly reduced HI-mediated cytokine levels, whereas immediate, 3, 6 and 12-h treatment did not reduce cytokines. In a subsequent study, TAT-NBD treatment (6818 nmol/kg: IP) immediately and 3 h after hypoxia demonstrated significant improvements in behavioral outcomes up to 14 weeks post-HI [107].
Since intrauterine infections can be associated with HIE, TAT-NBD was assessed in a lipopolysaccharide (LPS)-sensitized (IP injection) HIE model in the P7 rat [108]. Animals were administered with TAT-NBD (477 nmol/kg) intranasally (IN) 10 min after hypoxia, using three HI injury conditions; HI only, 4-h LPS-pre-treatment + HI or 72-h LPS-pre-treatment + HI. Treatment with TAT-NBD significantly reduced brain injury only in the LPS-sensitized HI models, when assessed 7 days post-HI. While it was surprising that TAT-NBD did not reduce brain injury following HI only, its neuroprotective actions in LPS-sensitized HI models may be due to the ability of the cationic TAT-NBD peptide to electrostatically bind and neutralize anionic LPS molecules or by reducing the effects of inflammatory cytokines. Neutralizing LPS would have the effect of suppressing the molecules capacity to evoke an immune response and inflammatory response. In support of this mechanism, previous studies using in vitro cell and in vivo tissue injury models have demonstrated that CARPs (e.g., HBD-2, HNP-1 and indolicidin) can bind to LPS and reduce its inflammatory effects [109]. Furthermore, it is possible that the IN delivery of TAT-NBD was enough to reduce the detrimental effects of LPS, but not ischemic injury associated with HI.

TAT-mGluR1
mGluR1 is a 14-amino acid peptide (mGluR1: VIKPLTKSYQGSGK 929-942 ; net charge +3), corresponding to an intracellular region of the metabotropic glutamate receptor-1α (mGluR1α) protein that is cleaved by calpain in neurons following excitotoxicity [110]. It is proposed that TAT-mGluR1 (YGRKKRRQRRR-VIKPLTKSYQGSGK, net charge +11) blocks the ability of calpain to cleave mGluR1α and thus maintain the receptors ability to stimulate the neuroprotective phosphoinositide 3-kinase/protein kinase B (PI3K-Akt) signaling pathway and protect neurons from the toxic effects of excitotoxic calcium influx [110]. In a neuronal model of kainic acid mediated excitotoxicity, TAT-mGluR1 prevented mGluR1α truncation and reduced neuronal cell death [110]. In a P7 rat model of HIE [111] TAT-mGluR1 (58,590 nmol/kg; IP) administered 1 h before hypoxia significantly reduced cerebral infarction at 24 h post-HI.
The initial study examining the efficacy of D-JNKI-1 (tdqsrpvqpflnlttprkpr-NH 2 ; net charge +4) in a P7 rat model of HIE when administered (76 nmol/kg: IP) 30 min before and 3, 5, 8, 12 and 20 h after hypoxia did not reveal any reduction in cerebral infarction at 24 h [115], although there was evidence for reduced calpain, caspase-3 and autophagic activation.
A subsequent study demonstrated that the TAT-fused peptide JNKI-1-TATL administered (2446 nmol/kg; IP) immediately and 3 h or 3 h after hypoxia, significantly reduced cerebral infarction at 48 h, while administration 6 h after hypoxia was ineffective [118]. When administered immediately and 3 h after hypoxia, functional benefits were observed 14 weeks post-HI. Despite improvements in cerebral infarct and functional outcomes, JNKI-1-TATL failed to prevent caspase-8-mediated cleavage of Bid, which was unexpected, as activated JNK is known to induce caspase-8 cleavage of Bid and promote mitochondrial pro-apoptotic cell death pathways; this suggests JNKI-1-TATL-mediated neuroprotection was occurring via mechanism independent of JNK suppression.
In our laboratory, administration of JNKI-1-TATD (1000 nmol/kg; IP) immediately after hypoxia resulted in a positive trend in reduced total infarct volume (15% reduction) although it did not improve behavioral outcomes 48 h post-HI [22]. It was also demonstrated that in cultured cortical neurons JNKI-1-TATD dose-dependently reduced glutamic acid mediated excitotoxic intracellular calcium influx.
While it was surprising that the Sab KIM1 scrambled control peptide did not display any neuroprotection, it is possible that the amino acid sequence of the cargo peptide may have influenced neuroprotective potency. Importantly, we and others have confirmed that TAT-fused scrambled peptides such as TAT-p53DMs and TAT-NR2cts [29,66] still retain neuroprotective properties, again pointing to an active role of the TAT carrier and peptide arginine content mediated neuroprotective action.

TAT-BH4
BH4 is a 25-amino acid peptide derived from the Bcl-x(L) protein BH4 homology domain region (RTGYDNREIVMKYIHYKLSQRGYEW 6-30 , net charge +2.1). The BH4 domain is required by Bcl-x(L) to block cytochrome-c release from mitochondria and prevent MPTP opening [120], as well as inhibiting endoplasmic reticulum (ER) calcium release into the cytosol by binding to a regulatory domain of the inositol triphosphate-3 (IP3) receptor [121]; highlighting potential mechanisms whereby the BH4 peptide could exert a neuroprotective action after HI.
The TAT-BH4 (YGRKKRRQRRR-RTGYDNREIVMKYIHYKLSQRGYEW, net charge +10.1) peptide was demonstrated to reduce cerebral tissue damage and caspase-3 activation when administered intracerebroventricularlly (ICV; 5 µL/20 ng) 30 min before hypoxia in a P7 rat model of HI [122]. Interestingly, exposure of cultured adult neural stem cells to TAT-BH4 promoted axonal remodeling, suggesting the peptide exerts other effects that may be beneficial in improving recovery following HI.

Osteopontin (OPN) and TAT-Fused OPN Peptide (TAT-OPN)
Osteopontin (OPN) is a secreted acidic glycoprotein containing an arginine-glycine-aspartate (RGD) motif and possesses multiple cellular functions including chemotactic and cytokine-like actions. The OPN protein is expressed in various tissues including bone, placenta, pancreas and the CNS [123]. Following cerebral ischemia, OPN deficient mice demonstrate increased neurodegeneration [124] and when administered ICV, the protein is neuroprotective in several adult rodent cerebral ischemia and hemorrhagic stroke models [125][126][127][128].

D-TAT-GESV
GESV is a 4-amino acid peptide (GESV, net charge −1) derived from the carboxyterminal PDZ motif (PDZ-G(D/E)XV; net charge −1) of nNOS. Proteins containing PDZ domains play a key role in anchoring receptor proteins in the membrane to intracellular cytoskeletal structures (e.g., actin and microtubules) and cytosolic proteins. Following NMDA receptor stimulation, nitric oxide synthase-1 adaptor protein (NOS1AP) via its PDZ binding domain, binds nNOS and recruits the protein to the cell membrane where it generates nitric oxide (NO). The D-TAT-GESV peptide (ygrkkrrqrrr-GESV, net charge +7) inhibits NOS1AP/nNOS binding, and is believed to confer neuroprotection by reducing NO production and activation of the p38 MAPK stress pathway [144]. In a P7 rat model of HIE, D-TAT-GESV (100 ng: ICV), administered before hypoxia reduced cerebral infarction at 7 days post-HI [145]. In comparison, the control peptide D-TAT-GASA (100 ng: ICV) was ineffective [145], which is not surprising since alanine residues can reduce CARP neuroprotective efficacy [19].

TAT-NR2B9c/NA-1
The development of NR2B9c as a neuroprotective agent has recently been discussed in detail [146]. Briefly, NR2B9c is a 9 amino acid peptide (KLSSIESDV 1479-1487 ) derived from the intracellular terminal carboxyl region of the NMDA receptor NR2B subunit protein, a site that binds the adaptor protein postsynaptic density-95 (PSD-95). PSD-95 couples nNOS to the NR2B subunit and following NMDA receptor activation stimulates NO generation. The NR2B9c peptide was developed on the basis that it blocks NR2B subunit and PSD-95 binding and thereby inhibiting NO production following NMDA receptor stimulation.

Other Peptides Examined in Animal Models of HIE
Several other peptides, not fused to a CPP, have also been examined in animal models of HIE. All descriptive details of the studies discussed below are presented in chronological order of publication and in Table 1.

Connexin 43 (Cx43) Derived Peptides
Astrocytic gap junction hemichannels play an essential role in the regulation of electrolytes and metabolites between the cytosol and extracellular space. Under physiological conditions, hemichannels are usually in a closed state. Specifically, connexin 43 (Cx43) hemichannel opening occurs following hypoxia, ischemia and OGD [153][154][155][156], resulting in a disruption to cellular homeostasis and an exacerbation of excitotoxicity [157]. The non-specific blocking of gap junction hemichannels with octanol and carbenoxolone demonstrates neuroprotective effects in in vitro models of hippocampal hypoxic-hypoglycemic injury [158,159]. To overcome the non-specific blocking of gap junction hemichannels, small mimetic peptides that bind to the extracellular loop of Cx43 hemichannels, blocking hemichannel opening have been developed [160].
Several studies have assessed the administration of a Cx43 mimetic peptide (Peptide 5: VDKFLSRPTEKT, net charge +1) in fetal sheep models of HIE. Administration of Peptide 5 continuously for either 1 (50,000 nmol/kg/h; ICV) or 25 h (50,000 nmol/kg over 1 h and then over 24 h; ICV), beginning 90 min after hypoxia, improved electroencephalographic (EEG) recovery, reduced seizure activity and improved the return of sleep state cycling. In addition, the 25-h infusion of Peptide 5 improved oligodendrocyte survival at 7 days post-HI [161]. Another study demonstrated Peptide 5 administration immediately after hypoxia (50,000 nmol/kg over 1 h and then over 24 h; ICV) improved neuronal and oligodendrocyte cell survival, whereas peptide administration beginning during hypoxia (50,000 nmol/kg for 1 h and then for 24 h; ICV) was not neuroprotective [162]. In contrast, a continuous high dose infusion of Peptide 5 (50,000 nmol/kg/h over 25 h: ICV) immediately after hypoxia did not improve neurological outcome and contributed to cerebral edema in the fetal sheep 7 days post-HI [163]. In a pre-term (103 days gestation) fetal sheep model of HI, Peptide 5 administration (50,000 nmol/kg for 1 h and then for 24 h; ICV) commencing 90 min after hypoxia improved EEG recovery, reduced neuronal death and increased oligodendrocyte cell survival 7 days post-HI [164]. A study using term fetal sheep assessed Peptide 5 and hypothermia (32 ± 1 • C; 72 h), as individual treatments and when combined. Peptide 5 administration 3 h after hypoxia (50,000 nmol/kg for 1 h and then for 24 h; ICV) and hypothermia treatments alone, improved EEG power, neuronal and oligodendrocyte cell survival at 7 days post-HI. However, combined Peptide 5 and hypothermia treatment did not result in any additional improvement in EEG power or histological outcomes [165]. This suggests that neuroprotective targets, other than hemichannels, may be more appropriate to improve the therapeutic effects of hypothermia for HIE.
Since Cx43 mimetic peptides Peptide 5, Gap26 and Gap 27 have only one arginine residue and low cationic charge, it would seem unlikely the peptides are operating via a CARP mechanism of action.

Do All CARPs Including TAT-Fused Peptides Share a Common Neuroprotective Mechanism of Action?
Based on previous studies in our and other laboratories, examining the neuroprotective properties of various CARPs, we view it highly likely that the neuroprotection demonstrated by the CARPs described above (i.e., TAT-NBD, TAT-mGluR1, JNKI-1-TATL, JNKI-1-TATD, Sab KIM1 , TAT-BH4, P5-TAT, D-TAT-GESV, TAT-NR2B9c, R18, R18D, COG133, Apelin-36 and IDR-1018) is largely mediated by peptide arginine content and positive charge [19,27,29,65,66]. There are several lines of evidence that support this hypothesis, which we provide below. In addition, we previously provided an explanation why we believe a CARP-mediated neuroprotective mechanism of action is operating for the TAT-NR2B9c and JNKI-1-TATD peptides, rather than the mechanism of action in which they were originally developed [27].
It has been demonstrated that CARPs inhibit neuronal excitotoxic calcium [19,22,23,25,27,28] and potassium [64,75] influx and in doing so inhibit many of the pathophysiological processes downstream from the ischemic excitotoxic cascade (e.g., calpain activation, mitochondrial dysfunction and ROS production, calcium induced cell signaling and lipase and DNase activation). Therefore, given that the vast majority of the proposed mechanisms of action of developed TAT-fused and other neuroprotective CARPs are molecular targets downstream from excitotoxicity (e.g., NF-κB, JNK and Cdk5), it is not surprising that they are effective and appear to be inhibiting their intended targets.
Interestingly, in addition to the inhibition of excitotoxic cationic ion influx, CARPs also have the capacity to modulate other cellular processes that are neuroprotective following HI such as proteolytic enzyme activation (e.g., proteasome and MMPs), ERK-1/2 and Akt phosphorylation and the modulation of immune responses. For example, in the case of TAT-NBD, which was developed to block NF-κB activation, the ability of CARPs to inhibit the proteasome [84,85,176,177] would have the effect of preventing NF-κB activation by preventing proteasomal degradation of the NF-κB inhibitor protein i-κB.
Future studies examining the efficacy of putative neuroprotective peptides fused to TAT would benefit with respect to confirming neuroprotective mechanisms of action by fusing the peptide to a non-cationic CPP (e.g., MTS; AAVALLPAVLLALLP) and/or replacing any negatively charged aspartate (D) and glutamate (E) residues within the peptide sequence with arginine residues; we predict these two manipulations will decrease and increase neuroprotective efficacy, respectively. It is also advisable to confirm if the TAT-fused peptide has the capacity to reduce excitotoxic calcium influx.

Conclusions
There is now a large body of data from various sources demonstrating that CARPs comprise a class of peptide with significant neuroprotective potential for development as an adjuvant to hypothermia, or when hypothermia cannot be applied, for the treatment of HIE. While further studies are required to elucidate neuroprotective mechanisms, available evidence indicates that CARP neuroprotection is determined by the peptide's arginine content and cationic charge and more specifically by the guanidine head group in the arginine residues. Interestingly, CARP neuroprotective potency appears to correlate with the same characteristics that improves the ability of the cationic CPPs to target and traverse cellular membranes [19,29]. Taken together it is clear that CARPs possess many properties enabling them to target both intracellular and extracellular damaging and protective pathways that are likely to be beneficial following HIE, not to mention their ability to cross the BBB and enter cells.
In conclusion, CARPs, including poly-arginine peptides, offer an exciting therapeutic approach for HIE with the potential to reduce the severity of brain injury via potentially several neuroprotective mechanisms of action.

Future Directions
There are several further considerations to be addressed before translation of CARPs as a therapeutic for HIE in the clinic. It will be important to firstly identify a CARP that has high efficacy and the greatest potential to be effective in HIE, in both pre-term and term infants. In addition, preclinical studies will need to examine CARP treatment in combination with hypothermia to ensure safety, and to identify any synergistic effects in terms of improving patient outcomes. It is also important that beneficial results obtained in rodent models of HI, especially with combining CARPs and hypothermia, are confirmed in a large animal model (e.g., piglet or lamb), allowing for assessment in a gyrencephalic brain with greater comparability to the human and that the efficacy of IV administration should be established prior to clinical trials.