The mechanisms underlying poor outcome following subarachnoid haemorrhage (SAH) are complex and multifactorial. They include early brain injury, spreading depolarisation, inflammation, oxidative stress, macroscopic cerebral vasospasm, and microcirculatory disturbances. Nrf2 is a global promoter of the antioxidant and anti-inflammatory response and has potential protective effects against all of these mechanisms. It has been shown to be upregulated after SAH, and Nrf2 knockout animals have poorer functional and behavioural outcomes after SAH. There are many agents known to activate the Nrf2 pathway. Of these, the actions of sulforaphane, curcumin, astaxanthin, lycopene,
SAH is a devastating condition and is associated with high levels of morbidity and mortality [
The mechanisms leading to poor outcome after SAH are complex and multifactorial. They include early brain injury [
The purpose of this review is to evaluate the evidence for pharmacological augmentation of the Nrf2 pathway as a treatment for patients with SAH. In order to do this, we have first reviewed the mechanisms underlying poor outcomes after SAH, then reviewed the Nrf2 pathway, before considering how the Nrf2 pathway applies to SAH, and finally reviewed the evidence for specific compounds known to upregulate Nrf2 activity.
The mechanisms leading to poor outcome after SAH can be categorised into primary and secondary injury in a way analogous to head injury, where the immediate damage that occurs directly from the insult is classed as primary and any further subsequent indirect damage as a result of the processes initiated by the insult is secondary. The important clinical distinction is the presumption that secondary injury is potentially treatable, whereas primary injury is not. While this concept is well established in head injury, the terms are not as widespread in SAH where the immediate injury is often classed together with subsequent events in the first 72 hours as early brain injury, and all ensuing events are considered separately. The latter events could also be grouped together as delayed brain injury. While some mechanisms like cerebral vasospasm clearly follow this classification, there are limitations, and others such as spreading depolarisation straddle both time periods.
At the time of SAH, intracranial pressure rises to that of diastolic arterial pressure or higher [
Following lysis of red blood cells in the subarachnoid space, the central nervous system is exposed to high levels of Hb and its degradation products which lead to narrowing of the cerebral vessels and development of delayed cerebral ischaemia (DCI) in 30% of patients [
Following SAH, oxyhaemoglobin (OxyHb) has been shown to induce vasoconstriction in animal models [
However, macrovascular vasospasm does not always correlate directly with the development of DCI. In fact, transcranial Doppler and angiographic studies have only shown a positive predictive value of 57% and 76%, respectively [
It has therefore been proposed that poor outcome is conferred by spasm of the microvasculature rather than macroscopic vasospasm seen on angiography [
An intraoperative study used orthogonal polarisation spectral imaging during aneurysm surgery to visualise the response of the small cortical vessels to hypercapnia. Patients with visible blood clot who underwent early surgery had a more pronounced vasoconstrictive response (39%) compared to the group without visible blood clot who had late surgery (17%) and patients with unruptured aneurysms (7%) [
These findings are consistent with a postmortem study of 53 aneurysmal SAH patients which demonstrated extensive cortical and hypothalamic infarctions with histologic evidence of microangiopathy [
After SAH, free extracellular Hb undergoes oxidation to methaemoglobin (MetHb), which then degrades into haem. Free haem is toxic and acts as a catalyst for formation of ROS causing oxidative stress. Haem toxicity is exerted by its proinflammatory properties as well as damage caused by ROS leading to modification of lipids, carbohydrates, and nucleotides with eventual cell death affecting both neurons [
Following conversion of OxyHb to MetHb, superoxide radicals are released which convert to hydroxyl radicals [
ROS produce vasoactive lipids via reactions with arachidonic acid, resulting in vasoconstriction. Furthermore, free radical oxidation of bilirubin and biliverdin leads to formation of bilirubin oxidation products [
Oxidative stress has been linked to the activation of protein kinase C [
Following SAH, free Hb released in the subarachnoid space stimulates rapid expression of cell adhesion molecules by endothelial cells, attracting neutrophils [
Cortical spreading depolarisation (CSD) refers to slow waves of near-total neural depolarisation with resultant cellular swelling due to the influx of cations across the cell membrane. This exceeds the ATP-dependent Na+ and Ca2+ pump activity which leads to shrinkage of the extracellular space due to water influx [
The normal response to a short episode of CSD is hyperaemia. However, in SAH following a single wave of CSD associated with OxyHb [
As well as in animal studies [
Nuclear factor-erythroid 2- (NF-E2-) related factor 2 (Nrf2) is a redox-sensitive transcription factor belonging to the cap’n’collar (CNC) subclass of the basic leucine zipper region containing the protein family. It binds to a specific DNA site, the antioxidant response element (ARE), regulating transcription of an array of detoxifying or antioxidant enzymes. These include gamma-glutamylcysteine synthetase, superoxide dismutase, catalase, glutathione reductase, thioredoxin reductase, peroxiredoxins, and glutathione S-transferase (GST-
Hp has received particular attention following SAH. It is the fourth most abundant plasma protein and is synthesised in the liver and reticuloendothelial system [
Although no suitably powered human study of CSF Hp levels has been done to assess its relationship with outcome, the Hp phenotype has been shown to be important in determining outcome after SAH. Human Hp is composed of two peptide chains:
During normal physiological conditions, Nrf2 is bound to the Kelch-like ECH-associated protein 1 (KEAP1) in the cytoplasm [
Nrf2 regulation. Nrf2 is a redox-sensitive transcription factor that is bound to KEAP1 under physiological condition. KEAP1 is an intracellular redox sensor and targets Nrf2 for ubiquitination. Following oxidative stress, four different mechanisms result in dissociation of KEAP1 from Nrf2. These four mechanisms are as displayed in order: (1) oxidation of cysteine residues by lower molecular weight reactive oxygen species, (2) covalent modification of cysteine residues by electrophiles such as NF-
While KEAP1, as an intracellular redox sensor, regulates the transcriptional response to oxidative stress through Nrf2, this is balanced by the activity of other transcription factors such as nuclear factor kappa-light-chain-enhancer of activated B cells (NF-
Nrf2 is expressed in the central nervous system (CNS), and it is upregulated in response to inflammation and cerebral insults [
A number of studies of ischaemic stroke models [
The Nrf2 pathway is also activated following intracerebral haemorrhage (ICH) in mice. HO-1 was shown to be upregulated after 24 hours, peaking at five days with a return to baseline on day eight [
Nrf2 is also protective against hemin toxicity. Rat astrocyte cultures were pretreated with Hb or vehicle, and then exposed to hemin, the degradation product of Hb. Pretreated astrocytes showed resistance to toxicity induced by hemin. Pretreatment with Hb was shown to induce Nrf2 and HO-1, and the latter led to haem catabolism, in keeping with the protective effect of Hb pretreatment. In support of this mechanism, the protective effect of Hb pretreatment was lost in Nrf2 knockdown cells [
Experimental data have shown that Nrf2 expression is upregulated in the basilar artery of rats after SAH [
Chen and colleagues using a rat SAH model demonstrated that Nrf2 expression is also increased in the cortex at 12 hours, 24 hours, and 48 hours postinjection of blood compared to controls, with a peak at 24 hours postinjection [
Deletion of Nrf2 has been shown to be associated
The pathophysiology of SAH involves oxidative stress and inflammation. The redox state can be assessed using malondialdehyde (MDA) levels and the GSH/GSSG ratio. MDA is a lipid peroxidation product and is elevated after oxidative stress [
There are a large number of known activators of the Nrf2 system. All act by binding KEAP1 releasing Nrf2, which translocates to the nucleus leading to increased transcription. Nrf2 activators are broadly classified as electrophilic cysteine-reactive compounds and nonelectrophilic Keap1-Nrf2 protein-protein interaction inhibitors. Most well-established compounds fall into the former category. However, many are pleiotropic and their primary mechanism of action remains controversial. There are also efforts being made to develop new more selective compounds of the latter category. These have the potential to be more potent inducers with less cross-activation of other pathways [
SAH represents an ideal condition for these treatments. With the wide range of proteins upregulated by Nrf2, it can be postulated that Nrf2 activation may have beneficial effects on any of the described secondary mechanisms underlying poor outcome. The magnitude of this effect is likely to be dictated by the timing of administration of the Nrf2 activator.
Following SAH, DCI occurs no earlier than three days after the event and is not seen beyond 21 days. Even if the mechanisms leading to it are initiated earlier, Nrf2 preconditioning has been shown to be protective in ischaemic stroke. Therefore, a case can be made for administration as late as 72 hours after SAH, which would be consistent with most previous drug studies in SAH [
Prevention may require earlier treatment. Most discussed mechanisms are either initiated or worsened by Hb. Given that it takes days for red cell lysis, and intracellular Hb to be released, CSF Hb levels progressively rise from day one to day six after SAH [
However, other aspects of inflammation and oxidative stress may result directly from early brain injury. These and indeed early brain injury itself would require earlier treatment still. Preconditioning would offer the greatest chance of benefit but is clearly not practical.
Therefore, aiming for treatment in patients at the earliest available opportunity, but accepting treatment up to 72 hours after SAH where patients do not present immediately, would seem a pragmatic approach, although early phase studies may benefit from shorter recruitment windows to increase the chance of observing an effect. Here, we have reviewed all agents that have been tested and have shown therapeutic potential in SAH. We have summarised the characteristics of each Nrf2 activators in Table
A summary of findings from experimental subarachnoid haemorrhage studies testing agents that activate the Nrf2 pathway, with relevant human data for these agents.
Agent | Curcumin | Astaxanthin | Lycopene | Dimethyl fumarate | Melatonin | Erythropoietin | Sulforaphane | |
---|---|---|---|---|---|---|---|---|
Animal SAH model | Rat, mouse | Rat, rabbit | Rat | Rat | Rat | Rat | Rat, rabbit | Rat |
Timing of administration | 0-4 weeks | 30 min-3 h | 2 h | 0-36 h | Twice daily for 2 d | 0-48 h | 0-72 h | 30 min-72 h |
Method of administration | IP | IT & oral | IP | IP & oral | Oral | IP | SC, IV, & IP | IP |
Animal dose | 150-600 mg/kg | 0.01-75 mg/kg | 40 mg/kg | 12.5-50 mg/kg | 15 mg/kg | 15-150 mg/kg | 400-1000 IU/kg | 5 mg/kg |
Time of tissue evaluation | Days 3-7 | 24-72 h | 24 h | 24-48 h | 48 h | 24-48 h | 24-72 h | 12-72 h |
Time of clinical assessments | 6 h—day 7 | 0-72 h | 24 h | Day 0-8 | Days 2-5 | 24-48 h | Days 0-16 | 72 h |
Biochemical effect | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
Clinical effect | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
Reduced vasospasm | Yes | Yes | Not assessed | Not assessed | Not assessed | Yes | Yes | Yes |
Method of administration in humans | Oral | Oral | Oral | Oral | Oral | Oral | IV | Oral |
Half-life | 6-7 h [ |
28-61 h [ |
20-24 h [ |
12 min [ |
1.8-2.1 h [ |
6-9 h [ |
2.4-2.6 h [ | |
BBB permeability | Yes [ |
Yes [ |
Yes [ |
Yes [ |
Yes [ |
Yes [ |
Yes [ |
Yes [ |
Toxicity | None known [ |
None known [ |
None known [ |
None known [ |
Progressive multifocal leukoencephalopathy & painful dermatitis [ |
None known [ |
Polycythaemia & secondary stroke [ |
None known |
Details of the experimental studies are shown in Table
Animal studies published in English, investigating Nrf2 activators in experimental SAH.
Agent | Study | Animal SAH model | Time of doses | Method of administration | Animal dose | Time of tissue evaluation | Time of clinical assessment | Biochemical effect | Clinical effect | Other effects | Vasospasm |
---|---|---|---|---|---|---|---|---|---|---|---|
Curcumin | Wakade et al. 2009 [ |
Mouse | 0, 1, 3, & 24 h | IP | 150/300 mg/kg | 72 & 96 h | Days 0, 1, 2, & 3 | Attenuation of COX-2, IL-1, IL-6, iNOS, TNF- |
No effect | Reduced cerebral infraction | Reduced vasospasm |
Kuo et al. 2011 [ |
Rat | 3 h & then once daily for 6 days | IP | 20 mg/kg | Day 7 | 6 h, days 1, 3, 5, & 7 | Lower glutamate & MDA levels, preserved SOD, & catalase level | Reduced mortality & improved functional outcomes | None | Reduced vasospasm | |
Aydin et al. 2017 [ |
Rat | Single dose | IP | 150/300/600 mg/kg | Blood at 1 h, brain extraction unclear | None | Reduced IL-1, TNF- |
Not done | None | Reduced vasospasm | |
Astaxanthin | Zhang et al. 2014 [ |
Rat |
30 min IT, 3 h Oral | IT, PO | IT 0.01-0.1 mmol/l, PO 25/75 mg/kg | 24 & 72 h | 0, 24, 48, & 72 h | SOD & GSH levels reduced, MDA levels elevated | Neurological improvement only at 24 & 48 h | Reduced BBB permeability, cerebral oedema, & apoptosis and reduced caspase-3 expression | Not assessed |
Wu et al. 2014 [ |
Rat | 30 min | IT | T 0.01-0.1 mmol/l | 24 h | 24 h | Increased expression of Nrf2, GST- |
Better performance at 24 h | Reduced BBB permeability, cerebral oedema, & apoptosis | Not assessed | |
Lycopene | Wu et al. 2015 [ |
Rat | 2 h | IP | 40 mg/kg | 24 h | 24 h | Downregulation of TNF- |
Improved neurological function | Lessened oedema, disruption of BBB, & cortical apoptosis | Not assessed |
Tetra-butyl hydroquinone | Wang and Theeuwes 2014 [ |
Rat | 2, 12, 24, & 36 h | PO | 12.5 mg/kg | 48 h | Days 0, 2, 3, 4, 5, 6, 7, & 8 | Increased Keap1, Nrf2, & HO-1 expression; upregulation of GST- |
Improved performance & learning deficits on days 4 & 5 | Reduced BBB permeability, cerebral oedema, & apoptosis | Not assessed |
Li et al. 2015 [ |
Mouse | 0, 8, & 16 h | IP | 50 mg/kg | 24 h | 24 h | Increased expression of Beclin-1 & the LC3-II to LC3-I ratio | Improvement in neurological deficits | BBB permeability, cerebral oedema, & neuronal degeneration were reduced | Not assessed | |
Dimethyl fumarate | Liu et al. 2015 [ |
Rat | Twice daily for 2 days | PO | 15 mg/kg | 48 h | Days 2, 3, 4, & 5 | Decreased IL-1 |
Reduction of learning deficits | Brain oedema, cortical apoptosis & necrosis decreased | Not assessed |
Melatonin | Aydin et al. 2005 [ |
Rabbit | 0, 2, 12, 24, 36, & 48 h | IP | 5 mg/kg | 48 h | None | Reduced endothelial cellular apoptosis | Not assessed | Reduced cellular apoptosis | Reduced vasospasm |
Ayer et al. 2008 [ |
Rat | 2 h | IP | 15/150 mg/kg | 24 h | 24 h | No effect on MDA | Reduced mortality only | Cerebral oedema reduced | Not assessed | |
Ersahin et al. 2009 [ |
Rat | 0, 24, & 48 h | IP | 10 mg/kg | 48 h | 48 h | Myeloperoxidase activity decreased, chemiluminescence values decrease, MDA decreased, & GSH was preserved | Improved neurological score | Cerebral oedema & BBB permeability reduced | Reduced vasospasm | |
Erythropoietin | Alafaci et al. 2000 [ |
Rabbit | 5 min, 8, 16, & 24 h | IP | 1000 IU/kg | 24 h | None | Increased CSF EPO levels | Not assessed | Decreased neuronal damage | Not assessed |
Buemi et al. 2000 [ |
Rabbit | 0 | IP | 1000 IU/kg | 72 h | 24, 48, & 72 h | No significant increase in CSF EPO concentration | Reduced mortality rate | None | Not assessed | |
Grasso et al. 2002 [ |
Rabbit | 5 min | IP | 1000 IU/kg | 72 h | 72 h | Increase in CSF EPO concentration | Improved neurological score | Reduced ischaemic neuronal damage | Reduced vasospasm | |
Springborg et al. 2002 [ |
Rat | 0 | SC | 400 IU/kg | 48 h | None | No biochemical effect assessed | Not assessed | Normalised autoregulation of cerebral blood flow | Not assessed | |
Grasso et al. 2002 [ |
Rabbit | 5 min, 8, 16, 24, 32, 40, 48, 56, 64, & 72 h | IP | 1000 IU/kg | 72 h | 72 h | Lower S-100 protein concentration in CSF | Improved neurological function | Reduced neuronal damage | Not assessed | |
Murphy et al. 2008 [ |
Rabbit | Days 0, 2, 4, & 6 | IV | 500/1500 IU/kg | 24 h | Days 0, 2, 4, 7, 9, & 16 | Increased haematocrit values | Reduced mortality rate | Improved cerebra blood flow, reduced cellular apoptosis | No change | |
Zhang et al. 2010 [ |
Rat | 15 min, 7, 16, 24, 32, 40, & 48 h | IP | 1000 IU/kg | 48 h | Not assessed | Increased Nrf2 & HO-1 expression, and upregulation of GST- |
Not assessed | Reduced impairment of cerebral oedema, cortical apoptosis, & BBB permeability | Not assessed | |
Sulforaphane | Chen et al. 2011 [ |
Rat | 30 min and 12 & 36 h | IP | 5 mg/kg | 12, 24, & 48 h | Not assessed | Increased Nrf2 & HO-1 expression and upregulation of GST- |
Improved function at 48 h | Decreased cerebral oedema, BBB permeability, & cortical apoptosis | Not assessed |
Zhao et al. 2016 [ |
Rat | 30 min and 24, 48, & 72 h | IP | 5 mg/kg | 72 h | 72 h | Increased Nrf2 & HO-1 expression; upregulation of GST- |
Reduced behavioural deficits | None | Reduced vasospasm |
Sulforaphane (SFN), 1-isothiocyanate-(4R)-(methylsulfinyl) butane, is a widely studied isothiocyanate. SFN stabilizes Nrf2 by inhibiting its ubiquitination. Oxidation of critical cysteine residues of KEAP1 by SFN appears to be essential [
The effect of SFN has been assessed in an
The effects of SFN after SAH
More recently, an experimental study [
In addition to reducing oxidative stress in the subarachnoid space and consequently reducing cerebral vasospasm, SFN may have a beneficial effect on the ischaemia which can follow cerebral vasospasm. In a rat middle cerebral artery (MCA) occlusion, stroke model preconditioning with SFN one hour prior to stroke and reperfusion after four, 24, and 72 hours upregulated Nrf2 and HO-1 expression leading to attenuation of BBB disruption, lesion progression as assessed by magnetic resonance imaging between 24 and 72 hours, and neurological dysfunction (based on motility, grasping reflex, and placing reaction) [
As discussed, microvascular spasm may contribute more to poor outcome after SAH than macroscopic spasm of the main arteries. Microvascular spasm is more difficult to study experimentally, and the effect of SFN on this has not been reported, though unlikely to respond differently to SFN, compared to macrovascular spasm. Whichever predominates, both occur in a delayed manner three days to three weeks after SAH, which offers a good therapeutic window for treatment unlike most other types of ischaemic stroke.
There are no published clinical studies of SFN in humans after SAH. Indeed, there are no studies of direct SFN administration in humans at all, due to its relatively short half-life making clinical administration problematic. It has therefore been studied in the context of cruciferous vegetables. Cruciferous vegetables of the genus Brassica, including broccoli, cauliflower, Brussels sprouts, kale, collards, kohlrabi, and mustard, are a rich source of precursors of isothiocyanates called glucosinolates [
The bioavailability of SFN has been studied in animals and humans. Due to its lipophilicity and molecular size, SFN is likely to passively diffuse through enterocytes. It is easily absorbed, conjugated to glutathione, and metabolised via the mercapturic acid pathway sequentially producing cysteinylglycine (SFN-CG), cysteine (SFN-Cys), and N-acetyl-cysteine (SFN-NAC) conjugates which are excreted in the urine [
SFN has been demonstrated in the gastrointestinal and genitourinary tracts as well as the liver, pancreas, lung, and heart, albeit in varying concentrations; bioactivity may differ amongst organs [
In many respects, SFN would appear to be an excellent candidate as a new therapeutic for patients after SAH. The lack of a practical formulation for clinical use has prevented trials to date. However, SFX-01 (Evgen Pharma) represents a novel solution to this by complexing SFN with
Curcumin has been tested in several SAH models. In an
In a SAH perforation model, curcumin was administered at the time of injury and one, three, and 24 hours later. It reduced inflammatory cytokines and the rate of vasospasm and DCI. There was no associated improvement in neurological recovery using rotarod and open-field activity assessment up to day three, despite observing a maximum effect of haemorrhagic infarct volume at day six. Perhaps further behavioural testing should have been performed at a later stage to address this. Interestingly, only a single dose at the time of haemorrhage was found to be associated with reduction in cerebral infarction at day six as well as reduced MCA diameter three days after the haemorrhage. Other treatment time points were not associated with these observed benefits [
Similar improvements in the rate of vasospasm were seen in a recent study comparing the actions of nimodipine, nicorandil, and low and high-dose curcumin on cerebral vasospasm. This demonstrated that high-dose curcumin is associated with a lower rate of vasospasm compared to nimodipine and nicorandil [
Animal behaviour was examined in a rat double-haemorrhage model following intraperitoneal injection of curcumin three hours after SAH induction and daily thereafter for six days. Curcumin was shown to increase superoxide dismutase and catalase and reduce MDA levels in the cortex and hippocampus. Basilar artery perimeter and thickness were significantly altered in the treatment group indicating a reduction in vasospasm. There was reduced neuronal degeneration. Importantly, mortality was reduced and blinded neurological scores improved with curcumin. Curcumin-treated rats had a significantly lower mortality rate assessed during and after the induction of SAH compared to the other groups. Neurological scoring was performed at six hours and days one, three, five, and seven after the haemorrhage induction. Curcumin rats displayed better neurological scores up to day seven, but even in the untreated group, the neurological scores showed a positive trend on day seven [
Curcumin has been associated with an improvement in learning and memory impairment measured by the Morris water maze in a rat SAH model. Treatment duration with curcumin lasted for four weeks, and the authors claim that the positive benefit is secondary to downregulation of hippocampal TNF-
Curcumin has also been found to have benefits in ischaemic stroke similar to SFN. Nrf2 and HO-1 gene and protein levels were measured at three, six, 12, 24, 48, and 72 hours after MCA occlusion. An increase was seen at three hours, peaking at 24 hours poststroke. Infarct volume, brain water content, and early behavioural deficits assessed at 24 hours were reduced in the curcumin group [
Astaxantin (ASTX) is a carotenoid found in algae, fungi, complex plants, and seafood. It has been shown to be a powerful antioxidant [
Lycopene is a natural carotenoid found mainly in tomatoes. It has multiple pleiotropic effects including antioxidant and anti-inflammatory actions [
Lycopene has been tested in a rat SAH model. It was given once, two hours after SAH. Brain oedema, BBB disruption, and cortical apoptosis were significantly reduced at 24 hours. Neurology was only assessed at 24 hours, when neurological dysfunction was markedly reduced. The study showed a beneficial effect of lycopene due to reduction in inflammation as shown by downregulation of IL-1
A phase II clinical trial assessing the effect of lycopene on cerebral vasospasm and autoregulation after SAH has been registered (
Tetra-butyl hydroquinone (tBHQ) has been evaluated in two SAH models [
Dimethyl fumarate (DMF) is an ester of fumaric acid conventionally used in the treatment of psoriasis [
The effects of DMF have been investigated in a rat prechiasmatic cistern injection model utilising autologous blood in rats. DMF was administered orally twice daily for two days, but the exact timing of administration relative to SAH was not specified. Two sets of experiments were performed. In the first group, tissue analysis took place only once 48 hours after surgery. The second experiment involved Morris water maze assessment of trained animals up to five days after the haemorrhage. Activities of KEAP1, Nrf2, and HO-1 were significantly increased within glial cells and neurons of animals treated with DMF. Cortical MDA was decreased, and superoxide dismutase and glutathione peroxidase activities were increased. Levels of proinflammatory cytokines IL-1
Melatonin is a well-known antioxidant with the ability to scavenge free radicals probably acting through multiple mechanisms [
In a rat SAH model, animals were treated with intraperitoneal injection of melatonin 150 mg/kg at two and 24 hours after the induction of SAH. Neurological scores and brain tissues were examined at 48 hours. Nrf2 and HO-1 were upregulated at 48 hours in the SAH group, mainly expressed on neurons. The levels of HO-1, NQO1, and GST-
Erythropoietin (EPO) is a pleiotropic molecule with known effects on Nrf2. In an experimental SAH model, EPO was injected intraperitoneally five minutes after SAH and every eight hours up to 48 hours. HO-1, NQO1, and GST-
EPO is unique amongst the agents identified in having been assessed in human studies. In a small case series of seven patients, EPO was shown to be effective in improving brain tissue oxygen tension if given over three consecutive days. This showed anti-inflammatory properties as well as restoration of cerebral autoregulation [
There is good experimental evidence suggesting that early Nrf2 activation reduces deficits early after SAH although more studies examining their effects on long-term outcome are needed. The reasons underlying the paucity of studies examining long-term functional outcome are unclear. This may be due to poor experimental design, practical reasons, or difficulty in inducing significant late deficits without excessive early mortality in rodent SAH models. Other than EPO, there have been no completed human clinical trials of Nrf2 activation in SAH. Experimental studies suggest biochemical and early functional improvements following treatment, although it is difficult to test for the more subtle neurocognitive deficits most prevalent in patients with SAH. The timing of administration of first dose in animal studies was generally early (often within 2 hours of SAH), with few studies providing data on later use. Although this is a potential concern for human studies, even if data on later administration was available in animal models, extrapolation of the therapeutic window from animals to humans is notoriously difficult if not impossible, and given the generally much slower evolution of SAH in humans compared to rodents, trials administering at the earliest available opportunity, up to 72 hours after ictus when patients start to deteriorate, could be considered. There are a number of potential agents that could be used in this context. There are no head-to-head comparisons in the literature, and they are all reported to penetrate the CNS, have relatively good safety profiles, and with exception of EPO, can be given orally. There is therefore little to guide which may be most suitable.
Outcomes following SAH remain poor despite advances in treatment. The mechanisms underlying recovery from SAH are multifactorial; however, Nrf2 activation appears to play a key protective role. There is overwhelming evidence for the therapeutic potential of several Nrf2 activators, with studies replicated in different SAH models and different laboratories. In the absence of any human data, there is a clear need for clinical studies to examine the safety and efficacy of Nrf2 activation after SAH.
Antioxidant response element
BTB and CNC homology 1
CREB-binding protein
Heme-oxygenase 1
Haptoglobin
Kelch-like ECH-associated protein 1
Musculoaponeurotic fibrosarcoma
Nuclear factor-erythroid 2- (NF-E2-) related factor 2
Phosphate group
Prostaglandin
Sulphide side chain reduced by a group R
Sulfhydryl side chain.
Diederik Bulters is the chief investigator, and Ian Galea and Ardalan Zolnourian are investigators of the SAS (SFX-01 after SAH) trial sponsored by Evgen Pharma.
Ian Galea and Diederik Bulters are joint senior authors.
This work has not been funded by an external body. Ardalan Zolnourian’s research fellowship has been granted by Royal College of Surgeons of England. We thank Freya Davis (BSc Hons, Pharmacology), who has kindly proofread the manuscript.