Repositioning drugs for traumatic brain injury - N-acetyl cysteine and Phenserine

Phytohemagglutinin (PHA) is a lectin present in particular legumes, particularly red kidney bean (Phaseolus vulgaris), and has potent, cell agglutinating and mitogenic activities that cause the immune activation of peripheral blood mononuclear cells (PBMCs), and their subsequent generation of cytokines. PHA is often used as a tool to challenge PBMCs in culture, and as shown in Fig. 4, results in the production and secretion of pro- and anti-inflammatory cytokines, represented by IL-1β and IL-10, respectively. As illustrated in Fig. 4, (−)-Phen (0.1 to 10 uM) substantially mitigated the PHA-induced elevation in pro-inflammatory IL-1β levels without impacting PHA-induced anti-inflammatory IL-10 levels; thereby mitigating inflammation. Recent in vivo studies in experimental TBI demonstrate that the anti-inflammatory actions seen in ex vivo studies of PBMCs translate into animals by mitigating neuroinflammatory markers associated with microglial cell activation. In the light of extensive studies indicating that chronic neuroinflammation is a common characteristic across neurodegenerative disorders (including AD, PD, TBI and stroke) that drives disease progression, its mitigation by well tolerated agents can be considered beneficial [76].

Glutamate is a key excitatory neurotransmitter in mammalian brain, and when intensely activated can be toxic to neurons over a range of acute CNS injury conditions that encompass TBI, stroke, hypoglycemia and status epilepticus. Excess glutamate is likewise implicated in chronic neurodegenerative disorders, particularly AD. Excessive glutamate activates its postsynaptic receptors, N-methyl-D-aspartate (NMDA), α-amino-3-hydroxy-5-methyl-4-isoxazolepropionate (AMPA) and kainate (KA). Such activation of AMPA receptors depolarizes the cell and concurrently unblocks the NMDA channels (releasing the Mg²⁺ block), and thereby allowing Ca²⁺ entry. Such depolarization opens voltage-activated calcium channels, causing Ca²⁺ ion and water influx into the cell down the osmotic gradient and leading the cells to cytotoxicity. Illustrated in Fig. 5, (−)-Phen treatment provides protection against glutamate-induced excitotoxicity in rat primary hippocamcal cultures. Specifically, glutamate significantly reduced cell viability of cultured primary hippocampal cultures by 53.5%, which was mitigated by (−)-Phen as a return to 73.5% of control levels and protection against anecrotic cell death [58]. Similar neuroprotection was found in human immortal neuronal (SH-SY5Y) cells with phenserine analogs following glutamate challenge.

In relation to the in vivo relevance of cellular studies indicating protection against glutamate excitotoxicity, (−)-Phen has been evaluated in rats challenged with a lethal dose of the organophophate soman, where (−)-Phen both increased the survival rate of animals and provided neuroprotection of neuronal cells in the hippocampus, basolateral amygdala and cingulate cortex [77]. In soman-induced toxicity, the sudden substantial loss of AChE leads to abnormal accumulation of ACh within cholinergic synapses and results in the excessive stimulation of muscarinic and nicotinic receptors within the central and peripheral nervous systems. In the brain, such excessive stimulation of cholinergic neurons induces the release of glutamate, leading to the overactivation of the NMDA receptor, and excessive influx of Ca²⁺ resulting in excitotoxic neuronal cell death [77]. These studies, together, support the notion that neuroprotection provided by (−)-Phen in cellular studies is of in vivo relevance, as additionally supported in anoxia (stroke) in vivo studies in the rat.

Fig. 6 shows that (−)-Phen provides protection against H₂O₂-induced oxidative toxicity in human immortal SH-SY5Y cells. Human SH-SY5Y cells were plated and after 24 h, cells were exposed to (−)-Phen (10 or 30 uM) followed by oxidative stress (100 uM H₂O₂). Cell viability was quantified at 24 h (MTS assay). (−)-Phen treatment significantly ameliorated the H₂O₂-mediated neuronal toxicity and provided protection against apoptotic cell death [58, 78].

Multiple studies have demonstrated across different laboratories that (−)-Phen lowers levels of APP in neuronal cell cultures [79–81]. This appears to be a non-cholinergically facilitated action, as it is shared by its (+)-enantiomeric form, Posiphen ((+)-Phenserine tartrate) that lacks anticholinesterase activity, and is mediated post-transcriptionally via an iron response element within the 5′-untranslated region (5’UTR) of APP mRNA [79–81]. The EC₅₀ of this APP lowering action appears to be in the order of 0.64 uM and 1.14 uM to lower secreted versus intracellular levels of APP, respectively, in human immortal neuronal (SH-SY5Y) cells [80]. Notably, primary neurons appear to be more sensitive, with (−)-Phen mediated APP lowering actions occurring at far lower drug doses (100 nM) [81]. As Fig. 7a and b document, (−)-Phen inhibits APP synthesis in vivo, and importantly lowers brain tissue levels of Aβ₄₂. Fig. 7b shows the action of (−)-Phen on Aβ levels in the cortex of transgenic (APP_SWE + PS1) AD mice over-expressing human Aβ, in which a daily dose of 2.5 mg/kg substantially (p < 0.05) lowered APP as well as Aβ. Such (-)-Phen induced APP lowering action in brain translated to rats (Fig. 7a). By contrast, neither donepezil nor (−)-physotigmine (a structural analog of (−)-Phen) shared this action. In line with the described APP lowering actions of (−)-Phen, a similar dose in rats lowered nucleus basalis lesion-induced elevations in APP, as evaluated in CSF samples [82]. Figure 8 shows the suppression of Aβ₄₂ after administration of (−)-Phen to humans in a study of healthy volunteers administered the agent twice daily over 35 days in which the dose was elevated to 15 mg BID ([83], Fig. 8). This same (−)-Phen dose provided an efficacy signal in mild to moderate AD patients [73]. A proof of mechanism clinical study of Posiphen likewise has demonstrated APP and Aβ lowering actions, as evaluated in time-dependent CSF samples obtained after 10 day dosing. Notably, (−)-Phen’s APP lowing action appears to be shared by not only its [+]-enantiomer but also by its 3 primary metabolites at concentrations as low as 100 nM [81]. Furthermore, these actions on APP by (−)-Phen and analogs additionally result in significant reductions in α-syn, which similarly appears to have a regulatory element controlling its translational efficiency within its 5’UTR [84–86].

In the light of several epidemiological studies reporting that a history of brain trauma places a patient at greater risk of developing AD and/or PD [7, 8, 65, 66, 67], (−)-Phen mediated reductions in APP, Aβ and α-syn may translate into potential therapeutic value. In this regard, diffuse axonal injury (DAI) is one of the most frequent and key pathologies that occurs in TBI in both humans and animal models (87). APP, in particular, is routinely present in high concentrations in axons and is conveyed through neurons via fast axonal transport. As a consequence, a rapid and substantial accumulation of APP is routinely evident in damaged axons following experimental and human TBI. In fact, immunohistochemical evaluation of APP accumulation associated with axons, particularly in white matter regions, is routinely used to detect DAI in human brain tissue [87]. The accumulation of APP in axons following TBI is considered an early event, and is associated with upregulated APP gene expression [88, 89, 90]. The extensive co-distribution of APP with Aβ accumulations and plaques has been described in swollen axons associated with DAI within days of an experimental TBI [91]. This, likewise, has been reported in human TBI [87], with Aβ plaques evident within the gray matter and notably also in the white matter in close proximity to swollen axons. Together, these observations indicate that damaged axons provide a prime source of Aβ following a TBI. Such TBI-induced actions, whether occurring early or later in life, can readily increase the vulnerability of the brain to other neurodegenerative events as detailed in the ‘two hit’ hypothesis of the ‘Latent Early-life Associated Regulation’ (LEARn) model of Lahiri et al., [92, 93], whereby genetic and environmental risk factors combine in an epigenetic pathway to trigger the etiology of a later neurobiological disorder (such as TBI leading to AD). Interestingly, the direct infusion of an anti-APP antibody into the affected brain region following TBI in rats resulted in a reduction in neuronal loss, less astrocyte activation, a smaller region of brain damage and less behavioral impairment than was evident in vehicle treated TBI animals [94], supporting a reduction in APP as a therapeutic strategy worth investigating.

Similarly, TBI has been reported to alter the distribution of α-Syn, and post-translationally modified it. An abnormal accumulation of α-Syn has been noted in axonal swellings and dystrophic neurites in TBI brains, with the creation of nitrated as well as conformationally modified forms. In rodent TBI models, striatal axons exhibit the most extensive accumulations of α-Syn forms [87]. Albeit the role of this synaptic protein requires greater elucidation, such TBI-induced changes are likely to impair its physiological role, and potentially induce a pathological one. Thus mechanisms to lower α-Syn may be of importance.

Extensive studies have demonstrated that neurogenesis continues to occur throughout life within key areas of the brain that include the subventricular zone (SVZ) of the lateral ventricles and the subgranular zone (SGZ) of the hippocampal dentate gyrus (DG) in rodents, in non-human primates as well as in humans [95–97]. Newly generated neural stem cells (NSCs) can differentiate into functional mature neurons and integrate into neuronal networks, including those involved in cognitive function [98, 99]. Ischemic brain injury as well as TBI stimulates the proliferation of NSCs localized to the SVZ and SGZ of adult brain, and the resultant newborn cells can migrate to damaged brain areas to potentially differentiate into mature neuronal cells [100, 101]. However, the process of neurogenesis is not particularly efficient, and is impaired by numerous factors initiated and amplified by ischemia and TBI, such as the presence of neuroinflammation [102]. Strategies and in particular, drugs that enhance neurogenesis therefore hold the potential to mitigate TBI and other neurodegenerative disorders.

First, Fig. 9 demonstrates, (−)-Phen enhances neural precursor cell viability in cell culture – increasing neurosphere size and augmenting their survival. Second, in cellular and animal studies high levels of APP (which are elevated by TBI as well as in AD) induce the differentiation of NSCs towards a glial phenotype, and away from a neuronal one. This action is reversed by (−)-Phen [103]. Third, (−)-Phen elevates neurotrophic factor levels in brain – as assessed by measuring BDNF, a key regulator of neurogenesis [78]. In both wild type and AD transgenic mice; administration of (−)-Phen analogs has been demonstrated to augment neurogenesis [78, 104] and, notably, enhance the survival of neurospheres as well as neuronal cells in culture [78].

The most consistent postmortem discovery in fatal head injury is the presence of cerebral ischemia [105, 106], which appears to be a key outcome predictor. Whereas there are numerous studies documenting reductions in cerebral blood flow in models of severe TBI where there is significant tissue and microvascular failure consequent to endothelial swelling, perivascular edema, and microthrombosis, particularly contiguous to focal lesions, its impact on mild and moderate TBI has remained more difficult to conclusively define. However, it is becoming increasingly appreciated that tissue hypoxia following a TBI occurs in a widespread manner in the brain, including within regions that appear to be structurally normal. Additionally, cerebral tissue hypoxia seems to arise independent of ischemia, sometimes in areas of no overlap, which suggests a microvascular etiology. The measurement of brain tissue PO₂, particularly in humans, by using oxygen 15-labeled positron emission tomography (¹⁵O PET) has recently provided definitive evidence of cerebral ischemia occurrence after early TBI [107], which may persist for up to a week after injury. Hence diffusion hypoxia in seemingly normal tissue, distinct from macrovascular ischemia in injured tissue, provides potential targets in TBI for neuroprotective strategies.

Evaluation of (−)-Phen in a classical rodent model of ischemic stroke has been undertaken by two separate research groups to appraise its protective actions in conditions of anoxia. In anesthetized male Sprague-Dawley rats with their right middle cerebral artery ligated and the common carotids clamped to induce focal infarction in the cerebral cortex after 60 min of ischemia, treatment with (−)-Phen (1 mg/kg/day) for four days, compared to placebo, reduced the area of infarction as assessed by digital scanner evaluation of brain slices (p = 0.001). This neuroprotective activity, in which anoxia leads to a focal lesion is supportive of beneficial actions in mild to moderate TBI, in which anoxia is considered less severe than in middle cerebral artery occlusion-induced experimental stroke.

One important pathological loss following head injury results from trauma to the basal midbrain and loss of cholinergic cells located in the NBM and/ or loss of axons providing cholinergic input to the cerebral hemispheres, hippocampus, and other critical brain structures. (−)-Phen has demonstrated efficacious benefits in the presence of NBM cell losses, an early feature of AD neuropathology that leads to elevated levels of APP and Aβ [82, 108].

Illustrated in Fig. 10 is the AChE inhibition induced by (−)- Phen and its metabolites achieved after a single acute administration of (−)-Phen to rodents [109]. As discussed above, (−)-Phen is a highly potent inhibitor of AChE (IC₅₀ = 22-36 nM) in plasma and brain, as are its N1-nor and N1,N8-bisnor metabolites [74]. In rats, at a dose of 1 mg/kg, (−)-Phen achieved a maximal inhibition of 73.5% at 5 min, and this only gradually decreased to 43% at termination of the study at 8 h, with an apparent t_1/2 = 8.25 h (Fig. 10).

Notably, these levels of AChE inhibition reflect the drug plus metabolite concentrations and result in elevated brain levels of ACh [109]. As noted in Fig. 3, (−)-Phen after oral dosing is first-pass metabolized by hepatocytes into three active (−) compounds: the N-1 Nor-, N-8 Nor-, and N-1,N-8-Bisnor-phenserine derivatives, which can all readily enter brain. Thus any given initial concentration of (−)-Phen will provide longer AChE inhibition effects than other non-cholinergic mediated drug and metabolite effects. In Fig. 10b, (−)-Phen inhibits AChE in a dose-response relationship in humans [72] that has been shown efficacious in improving cognition lost in AD where NBM lesions cause ACh deficiencies and at least partially reversible cognitive losses [73].

In Fig. 11, as evaluated by the novel object recognition (NOR) test at 7 days following a mild concussive TBI (a 30 g free falling weight from 80 cm striking a 30 g mouse on the left side of the head in the area of the parietal cerebral cortex above the hippocampus), (−)-Phen at two clinically relevant doses (2.5 and 5.0 mg/kg BID for 5 days initiated after mTBI) mitigated mTBI-induced cognitive impairment [58]. These TBI conditions (a 30 g weight and 30 g mouse) were created to mirror a human falling on their head from a three foot fall, and are considered mild concussive injuries [111], which certainly instigate neuroinflammation [112]. Notably, the rapid metabolic degradation of (−)-Phen and metabolites ensures that no pharmacological concentrations are evident in brain during cognitive evaluations performed 2 days and later after cessation of (−)-Phen dosing, thereby ensuring that mitigation of TBI cognitive deficits are not symptomatically induced by (−)-Phen’s known cholinergic actions.

Our prior studies highlighted in Figures illustrated above, additionally demonstrated in brain subjected to mild TBI and treated with (−)-Phen that markers of oxidative stress Thiobarbituric Acid Reactive Substances (TBARS) were reduced vs. mTBI alone, when evaluated in hippocampus 5 and 14 days after injury. This reduction in oxidative stress was consequent to a (−)-Phen induced upregulation in the activity/expression of the endogenous antioxidant proteins superoxide dismutase [SOD] 1 and 2, and glutathione peroxidase [GPx] [58].

These studies demonstrate that clinically relevant doses of (−)-Phen can provide a unique and broad range of beneficial pharmacological actions that may favorably impact the programmed cell death that results following a TBI, with such apoptosis being a common feature across many neurodegenerative disorders. (−)-Phen illustrates how a drug that was opportunistically developed to supplement cholinergic activity in AD, and that has proven well-tolerated and provided a consistent evidence of efficacy [73, 113], can – consequent to its more recently discovered important broad spectrum of pharmacological actions – be optimized to not only provide potential efficacy in TBI, but also provide a pharmacological tool to understand how TBI can lead to AD. It has become increasingly clear that multiple processes combine together following an insult (whether an acute TBI or a chronic degenerative disorder such as in AD or PD) to induce the programed cell death of neurons. The modulation of (i) inflammation either directly or via cholinergic mechanisms, (ii) oxidative stress, (iii) neurosphere/NPC apoptosis/survival, (iv) glutamate excitotoxity, and (v) APP/Aβ/α-syn over-expression, as well as ability to augment endogenous trophic factors like BDNF and stimulate other such mechanisms, provides a means to both limit cell death and optimize endogenous regenerative actions. Clinical trials in TBI and AD of experimental drugs that act via a single mechanism only, such as anti-inflammatory or Aβ lowering approaches, have failed to address the full range of pathologies that lead to neuronal loss and cognitive impairment. (−)-Phen’s described activation of multiple pathways, including the augmentation of endogenous antioxidant, neurotrophic, neuroprotective, anti-inflammatory, pro-angiogenesis, APP/Aβ/α-syn-lowering as well as cholinergic and others provide neuroprotection across multiple animal models. The revelation of these multiple activities of (−)-Phen and analogs over many years exemplifies how initial notions of a drug’s mechanism of action may mislead investigators away from its full spectrum of benefits for human health.