Transcriptional activation of antioxidant gene expression by Nrf2 protects against mitochondrial dysfunction and neuronal death associated with acute and chronic neurodegeneration

Abstract

Mitochondria are both a primary source of reactive oxygen species (ROS) and a sensitive target of oxidative stress; damage to mitochondria can result in bioenergetic dysfunction and both necrotic and apoptotic cell death. These relationships between mitochondria and cell death are particularly strong in both acute and chronic neurodegenerative disorders. ROS levels are affected by both the production of superoxide and its toxic metabolites and by antioxidant defense mechanisms. Mitochondrial antioxidant activities include superoxide dismutase 2, glutathione peroxidase and reductase, and intramitochondrial glutathione. When intracellular conditions disrupt the homeostatic balance between ROS production and detoxification, a net increase in ROS and an oxidized shift in cellular redox state ensues. Cells respond to this imbalance by increasing the expression of genes that code for proteins that protect against oxidative stress and inhibit cytotoxic oxidation of proteins, DNA, and lipids. If, however, the genomic response to mitochondrial oxidative stress is insufficient to maintain homeostasis, mitochondrial bioenergetic dysfunction and release of pro-apoptotic mitochondrial proteins into the cytosol initiate a variety of cell death pathways, ultimately resulting in potentially lethal damage to vital organs, including the brain. Nuclear factor erythroid 2-related factor 2 (Nrf2) is a translational activating protein that enters the nucleus in response to oxidative stress, resulting in increased expression of numerous cytoprotective genes, including genes coding for mitochondrial and non-mitochondrial antioxidant proteins. Many experimental and some FDA-approved drugs promote this process. Since mitochondria are targets of ROS, it follows that protection against mitochondrial oxidative stress by the Nrf2 pathway of gene expression contributes to neuroprotection by these drugs. This document reviews the evidence that Nrf2 activation increases mitochondrial antioxidants, thereby protecting mitochondria from dysfunction and protecting neural cells from damage and death. New experimental results are provided demonstrating that post-ischemic administration of the Nrf2 activator sulforaphane protects against hippocampal neuronal death and neurologic injury in a clinically-relevant animal model of cardiac arrest and resuscitation.

1. Introduction

Following acute brain injury, secondary neurodegeneration often increases the extent and severity of the insult (Schimmel et al., 2017; Visser et al., 2019). Several secondary processes have been implicated in chronic neurodegeneration including altered energy metabolism, oxidative stress, inflammation, and cell death (Hiebert et al., 2015; Narne et al., 2017). Despite their apparent heterogeneity, mitochondria are a common factor that links these processes together. As a result, pharmaceutical interventions that target mitochondria have the potential to impact multiple mechanisms that drive the ongoing injury (Bhatti et al., 2017). This review focuses on nuclear factor erythroid 2-related factor 2 (Nrf2), a transcription factor that interacts with mitochondria to inhibit neurodegenerative processes and provides neuroprotection in numerous animal models of neurologic disorders. New experimental results obtained with an animal model of cardiac arrest are included, demonstrating the neuroprotective efficacy of sulforaphane, a drug that activates the Nrf2 pathway of cytoprotective gene expression and that protects mitochondria from damage caused by elevated Ca²⁺ and oxidized redox state.

2. Acute brain injury

2.1. Traumatic brain injury

Traumatic brain injury (TBI) is a major cause of death and disability worldwide with an estimated 70 million people experiencing a TBI each year (Dewan et al., 2018). The primary injury stems from direct transfer of mechanical energy on the brain, acceleration/deceleration, or pressure waves associated with explosions. Depending on the severity and type of energy injury, the initial insult may cause primarily a focal injury, associated with contusions, lacerations, and/or bleeds, or a disseminated injury, characterized by diffuse axonal injury from stretching/tearing of tissue (Blennow et al., 2012). TBI in patients and lab animals is associated with inhibition of aerobic and stimulation of anaerobic brain energy metabolism within a few hours after the initial injury (DeVience et al., 2017; Robertson et al., 2006b; Xu et al., 2011). Severely damaged neurons and other brain cells may undergo necrotic cell death at this stage, thereby releasing a number of danger-associated molecular patterns (DAMPs), including HMGB1, mitochondrial DNA (mtDNA), ATP, and other molecules that stimulate neuroinflammation (Braun et al., 2017). The excitatory neurotransmitter glutamate is also released which, along with axonal damage, can increase intracellular Ca²⁺ and cause excitotoxicity (Hinzman et al., 2016; Siedler et al., 2014). Elevated intracellular Ca²⁺ can also increase mitochondrial Ca²⁺ levels, which promote reactive oxygen species (ROS) production and opening of the mitochondrial inner membrane permeability transition pore (mPTP), leading to mitochondrial osmotic swelling, energy failure, and cell death (Hurst et al., 2017).

2.2. Ischemic brain injury

According to the World Health Organization, cerebrovascular accidents, including cerebral ischemia, are the second leading cause of death and the third leading cause of disability worldwide (Johnson et al., 2016). Most commonly, ischemia is caused by diminished or interrupted blood supply to tissues from a blockage or cardiac arrest. Due to decreased blood flow, insufficient oxygen supply to the brain disrupts energy metabolism and leads to an ischemic injury, either to just a portion of the brain (focal cerebral ischemia, as occurs with an ischemic stroke), or to the whole brain (global cerebral ischemia, as occurs with cardiac arrest). This lack of sufficient oxygen impairs aerobic energy metabolism and increases glycolysis, resulting in a net decrease in intracellular ATP. Eventually, inadequate energy production promotes several conditions including ionic exchange pump failure, hyperosmolarity, and metabolic lactic acidosis, which can cause necrotic cell death (Sanganalmath et al., 2017; Wu et al., 2018). Necrosis releases glutamate, thereby promoting excitotoxicity and, like TBI, mitochondrial dysfunction through elevated intracellular Ca²⁺ (Andreyev et al., 2018). For a comprehensive review of calcium, ischemia, and excitotoxicity, see Szydlowska and Tymianski (2010). Cerebral ischemia also induces hypoxia inducible factor-1α (HIF-1α)-linked transcription which, in turn, activates NADPH oxidase (NOX) enzymes. NOX are membrane-bound complexes that are a major source of ROS through the generation of superoxide (O₂⁻) (Ma et al., 2017).

Ischemic cells that survive the initial injury have a lower concentration of antioxidants (e.g., glutathione) than healthy cells, making them vulnerable to oxidative stress (Wu et al., 2018). As a result, in cases where blood supply is restored to ischemic tissues either through resuscitation or resolution of a blockage, reperfusion can often exacerbate the ischemic injury (Pan et al., 2007). While reperfusion injury occurs due to a variety of mechanisms, excess ROS production is considered a primary driver (Granger and Kvietys, 2015). In the ischemic brain, this increased ROS production comes from a variety of sources including NOX (Walder et al., 1997) and mitochondria (Piantadosi et al., 2008).

2.3. Brain injury treatment window

Given the overlapping mechanisms that drive oxidative stress, inflammation, and energy metabolism dysfunction (for a review, see Witte et al. (2010)), it is not surprising that acute brain injuries can worsen over time due to ongoing pathogenic processes (Balan et al., 2013; Frati et al., 2017; Hazelton et al., 2010; Lee et al., 2018; Proctor et al., 2014; Schimmel et al., 2017). As acute brain injuries generally occur outside of a hospital environment and access to immediate intervention is limited, the hours/days immediately following the injury is a potential target for translational studies aimed to find pharmacologic interventions that prevent functional deterioration and maximize post-injury outcomes. Compounds that activate the Nrf2 pathway of cytoprotective gene expression may be particularly effective due to their ability to modulate multiple mechanisms of cell injury and repair.

3. Nrf2 regulation

Nrf2, the basic leucine zipper transcription factor encoded by the NFE2L2 gene (Alam et al., 1999), is an approximately 100 kilodalton cap’n’collar transcription factor. Nrf2 is a member of a family of conserved proteins that serve important roles in homeostasis and adaptive response to cellular stress in metazoans (Sykiotis and Bohmann, 2010). Nrf2 protein is constitutively expressed and localized in the cytosol [Fig. 1A] (Stewart et al., 2003). Under unstressed conditions, a homodimer of Kelch-like ECH-associated protein (KEAP1) sequesters Nrf2 along the actin cytoskeleton in the cytosol. KEAP1, a Cullin-3 (Cul3)-based E3 ubiquitin ligase adapter, acts to facilitate binding of Nrf2 to the Cul3 ubiquitin ligase complex, thereby promoting the ubiquitination of Nrf2 and subsequent proteasomal degradation (Cullinan et al., 2004; Eggler et al., 2009) [Fig. 1B].

Fig. 1.

Regulation of Nrf2. (A) Nuclear factor erythroid 2-related factor 2 (Nrf2) protein is constitutively generated and, along with the Kelch-like ECH associated protein (KEAP1), is located in the cytoplasm. (B) In the basal state, Nrf2 associates with the Keap1/Cullin3 (KEAP1/Cul3) ubiquitin ligase complex in the cytosol, which ubiquitinates Nrf2 and targets it for proteasomal degradation. (C) Exposure to reactive oxygen/nitrogen species (ROS/RNS), electrophiles, or phosphorylation by certain kinases, such as protein kinase C (PKC), PRK-ER-related kinase (PERK), mitogen activated protein kinase 1 (MAPK1), c-Jun-NH₂-kinase (JNK), or casein kinase 2 (CK2), inhibits the association of the Keap1/Cul3 complex with Nrf2, causing it to build up in the cytoplasm. Cytoplasmic AMP-activated protein kinase (AMPK) phosphorylation of Nrf2 allows it to translocate to the nucleus. Phosphorylated Nrf2 then binds to small Maf (sMaf) proteins and promotes gene expression via binding to antioxidant/electrophile response elements (ARE/EREs) in the promoter regions of genes that affect redox homeostasis, mitochondrial biogenesis, and metabolic up-regulators. Collectively this pattern of gene expression inhibits inflammation. (D) Glycogen synthase kinase-3β (GSK-3β) inhibits Nrf2 expression. This is carried out either through phosphorylation on a different site and recognition by the β-transducin repeats-containing protein/Cullin-1 (β-Trcp/Cul1) ubiquitin ligase complex that targets it for nuclear export and degradation, or by activation and translocation of the nuclear kinase FYN, which may cause dissociation of Nrf2 from ARE/EREs, thereby halting gene expression and leading to proteasomal degradation. GSK-3β can be inhibited by AMPK through activation of peroxisome proliferator-activated receptor-γ coactivator alpha (PGC-1α) and subsequent expression of protein p38, as well as through the phosphatidyl inositol 3-kinase/AKT (PI3K/AKT) pathway.

In the presence of increased ROS, reactive nitrogen species (RNS), electrophiles, or xenobiotics, cysteine thiols within KEAP1 become oxidized (Kansanen et al., 2013) [Fig. 1C], which causes a conformational change and the liberation of Nrf2, allowing it to translocate into the nucleus. Once in the nucleus, Nrf2 forms a heterodimer with small Maf (sMaf) proteins and binds to antioxidant response elements (AREs) or electrophile response elements (EREs). This promotes the expression of proteins involved in the response to oxidative stress and inflammation as well as metabolism and mitochondrial biogenesis (Dinkova-Kostova and Abramov, 2015; Friling et al., 1990; Motohashi and Yamamoto, 2004; Rushmore et al., 1991; Rushmore and Pickett, 1990). There is also evidence that Nrf2-KEAP1 is tethered to the outer mitochondrial membrane, perhaps to react most efficiently to increased mitochondrial ROS generation (Lo and Hannink, 2008; Strom et al., 2016).

While the above pathway is well-defined, the regulation of Nrf2 is complex, with post-translational phosphorylation playing a major role in its release, nuclear translocation, and degradation. Several sites on Nrf2 can be phosphorylated by different kinases to elicit different effects. For instance, it was thought that phosphorylation via protein kinase C (PKC) was necessary for dissociation from the KEAP1 dimer (Huang et al., 2000), but several other kinases have been implicated as potential dissociation agents, including the endoplasmic reticulum-localized PRK-ER-related kinase (PERK) (Cullinan et al., 2003), mitogen-activated protein kinase 1 (MAPK1/ERK), stress response signaler c-Jun-NH₂-kinase (JNK) (Xu et al., 2006), and casein kinase 2 (CK2) (Pi et al., 2007) [Fig. 1C]. Nrf2 translocation into the nucleus may also be stimulated by phosphorylation of Nrf2 via AMP-activated protein kinase (AMPK) (Joo et al., 2016; Wang et al., 2016a). Phosphorylation also plays a critical role in inhibition of Nrf2 activity, especially via glycogen synthase kinase-3β (GSK-3β), a master regulator for several cellular processes including glycogen metabolism, insulin signaling, and apoptosis (Cohen and Frame, 2001). GSK-3β phosphorylation causes sequestration of Nrf2 in the cytosol thereby preventing its influence over gene expression (Salazar et al., 2006); however, phosphorylation of Nrf2 by GSK-3β can also stimulate nuclear translocation, recognition by the nuclear β-transducin repeats containing protein ligase (β-Trcp), a Cullin1 (Cul1)-based E3 ubiquitin ligase adaptor, and degradation (Chowdhry et al., 2013; Rada et al., 2011; Rada et al., 2012). Alternatively, GSK-3β activation of the Src-family tyrosine kinase FYN drives nuclear export of Nrf2 and possibly serves to control gene expression and maintain homeostatic balance (Rizvi et al., 2014; Tebay et al., 2015). GSK-3β can itself be inhibited by the pro-survival phosphatidyl inositol 3-kinase/AKT (PI3K-AKT) signaling pathway (Hayes et al., 2015; Nakaso et al., 2003; Tebay et al., 2015) as well as AMPK (Joo et al., 2016; Wang et al., 2016a), which serve to increase Nrf2 activity [Fig. 1D].

Other mechanisms of Nrf2 regulation have been reported. The protein p38 MAPK enhances the interaction between KEAP1 and Nrf2 (Keum et al., 2006), while the autophagy adapter protein p62 interferes with this interaction (Komatsu et al., 2010). The BTB and CNC homology 1 (BACH1) transcription factor competes with Nrf2 for DNA binding sites and is itself exported from the nucleus via antioxidant-induced phosphorylation (Wiel et al., 2019). In addition, the ER-associated E3 ubiquitin ligase synoviolin 1 (SYVN1/Hrd1) has an affinity for Nrf2 and may act independently of the KEAP1/Cul3 and GSK-3β/β-Trcp/Cul1 pathways (Wu et al., 2014).

4. Antioxidant and redox status

The human brain accounts for 20% of the body’s oxygen consumption, despite comprising just 2% of the body weight (Kety, 1957). As a result of its relatively high rate of oxygen consumption and low capacity for ROS detoxification, the brain is particularly vulnerable to oxidative injury (Salim, 2017). The mitochondrial electron transport chain (ETC), a series of complexes involved in the aerobic production of ATP through oxidative phosphorylation, is a primary endogenous source of ROS in the brain. Under basal conditions, low levels of O₂⁻ can be converted from oxygen due to electron leak from the ETC (Handy and Loscalzo, 2012). Despite the constant generation of ROS, cells also possess a robust antioxidant defense capacity that is responsible for ROS detoxification including both enzymatic (e.g. superoxide dismutase 1 (SOD1), catalase (CAT), glutathione peroxidase (GPX)) and non-enzymatic antioxidants (e.g. glutathione (GSH) and coenzyme Q₁₀ (CoQ₁₀)). These phase II antioxidants, a classification for detoxifying enzymes that reduce reactivity of intermediate metabolites, can maintain redox homeostasis under healthy conditions but, in the event of acute injury, the antioxidant defense system may not be sufficient to counteract increased ROS production in the brain.

The Nrf2 pathway is most well-established in its role as a cellular defense mechanism against oxidative and electrophilic stress. Along this pathway, AREs have been identified in the promoter regions of antioxidant genes such as heme oxygenase 1 (HO-1), SOD1, glutamate cysteine ligase catalytic subunit (GCLc), and NAD(P)H quinone dehydrogenase 1 (NQO1), among many others (Hayes et al., 2000; Itoh et al., 1999; Nguyen et al., 2003). In fact, Nrf2 plays a role in the regulation of hundreds of genes, many of which are cytoprotective in nature (Malhotra et al., 2010; Tebay et al., 2015). It follows that ROS levels are usually higher in Nrf2 knockout models (Holmström et al., 2013; Kovac et al., 2015).

One of the most important influences of Nrf2-mediated gene expression is its regulation of redox homeostasis via glutathione regulation. Nrf2 mediates the expression of necessary subunits of GCL, a key enzyme responsible for the biosynthesis of GSH (Wild et al., 1999). Furthermore, Nrf2-controlled expression of glutathione reductase 1 enables the maintenance of glutathione in its reduced state (Harvey et al., 2009; MacLeod et al., 2009). Several enzymes responsible for the generation of NADPH necessary for GSH reduction, including malic enzyme 1 (ME1), isocitrate dehydrogenase 1 (IDH1), 6-phosphogluconate dehydrogenase (PGD), and glucose-6-phosphate dehydrogenase (G6PD), are upregulated by Nrf2 (Lee et al., 2003; Mitsuishi et al., 2012; Thimmulappa et al., 2002; Wu et al., 2011). As shown with regard to antioxidant capacity, Nrf2 deficiency results in lower levels of reduced GSH (Benedict et al., 2012; Higgins and Hayes, 2011; Wakabayashi et al., 2004) as well as a decreased mitochondrial NAD(P) H pool (Holmström et al., 2013).

5. Nrf2 and inflammation

Oxidative stress plays a vital role in many facets of the inflammatory response, including the induction of inflammatory markers such as cytokines, proteins that relay pro-inflammatory signals to other cells, and chemokines, which attract immune cells to the injury site. Activation of these cells can lead to further ROS generation. Thus, ROS removal is the most direct way in which Nrf2 works to mediate inflammatory signaling. For example, the Nrf2 transcription target NQO1 affects the activity of the NLR family pyrin domain containing 3 (NLRP3) inflammasome by preventing the priming necessary for NLRP3 activation by ROS. It follows that NLRP3 is then unable to cleave caspase-1, thereby blocking the production of mature interleukin (IL)-1β, a major pro-inflammatory cytokine (Liu et al., 2017).

Nrf2 also inhibits inflammation by mechanisms other than detoxification of ROS. Carbon monoxide-induced Nrf2 results in an increase in the expression of anti-inflammatory cytokine IL-10 and a decrease in pro-inflammatory tumor necrosis factor-α (TNFα) (MacGarvey et al., 2012). Kobayashi et al. (2016) showed that in mouse macrophages, Nrf2 interferes with the upregulation of the pro-inflammatory genes IL-6 and IL-1β through binding in proximity to their respective gene loci and blocking recruitment of RNA polymerase II. Furthermore, Hu et al. (2018) showed that the mitochondrially-targeted, synthetic antioxidant MitoQ enhances nuclear translocation of Nrf2. The resulting transcriptional activity reduced circulating mtDNA, which acts as a DAMP. They also observed decreased amounts of the cytokines IL-6 and IL-1β. However, this effect may be tissue dependent, as activation of the ARE in proximity to IL-6 causes upregulation of that gene in hepatocytes (Wruck et al., 2011). There is also much evidence that Nrf2 both directly and indirectly suppresses the activity of nuclear factor kappa-light-chain-enhancer of activated B cells (NFκB), a major activator of the inflammatory response (for review, see Ahmed et al. (2017)). For example, Nrf2-induced HO-1 expression inhibits NFκB -linked pro-inflammatory signaling, including the production of cytokines, chemokines, and prostaglandins, in a variety of tissues (Bellezza et al., 2012; Chi et al., 2015; Lee and Jeong, 2014). In Nrf2 knockouts, NFκB activity and proinflammatory cytokine expression is elevated (Jin et al., 2008; Reddy et al., 2011). The protective effects of Nrf2 activation toward inflammation is further demonstrated by improvement in cardiac function and survival in mouse models of sepsis (Liang et al., 2018; MacGarvey et al., 2012).

There is also evidence that Nrf2 may interact with a class of receptors that mediate the immune response to both infectious and noninfectious systemic pathologies. Toll-like receptors (TLRs), the most well-studied being TLR4, are pattern recognition receptors that can recognize both pathogen-associated molecular patterns (PAMPs), such as a component of gram negative bacteria (lipopolysaccharide; LPS), as well as DAMPs, such as circulating mtDNA and proteins that are released by stress or injured cells. Ligand binding of these molecules to TLRs induces the production of cytokines, chemokines, and interferons. TLR agonists also activate the Nrf2 pathway (Nadeem et al., 2016) which attenuates TLR-induced inflammatory signaling (Chen et al., 2017; Townsend and Johnson, 2016). For a review, see Mohan and Gupta (2018).

6. Nrf2 and mitochondrial biogenesis

Mitochondrial bioenergetic dysregulation is a hallmark of most neuropathological disorders. Although controlled ROS generation is an important component of homeostatic cell signaling (Shadel and Horvath, 2015), uncontrolled ROS generation leads to oxidative damage, including DNA damage. The contribution of mitochondria to ROS production taken together with the proximity of mitochondrial DNA to these ROS may explain why mitochondrial DNA is relatively sensitive to oxidative damage (Mikhed et al., 2015). Moreover, a disruption in the regular turnover of mitochondria, through both mitochondrial biosynthesis and the removal of damaged mitochondria, a process known as mitophagy, contributes to neurodegeneration (Lemasters, 2005). As a result, interventions that aim to restore healthy mitochondrial stores may be beneficial for the treatment of neurodegeneration in both natural aging (Fivenson et al., 2017; Palikaras and Tavernarakis, 2012; Sun et al., 2016) as well as neurodegenerative disorders such as Alzheimer’s disease (Wilkins and Swerdlow, 2016), Parkinson’s disease (Narendra et al., 2008; Pickrell and Youle, 2015; Ryan et al., 2015), Huntington’s disease (Jodeiri Farshbaf and Ghaedi, 2017; Khalil et al., 2015), and acute brain injury (Hiebert et al., 2015; Niizuma et al., 2009).

The connection between Nrf2 and mitochondrial biogenesis is a more recent discovery (Gureev et al., 2019). Piantadosi et al. (2008) first showed that HO-1, a well-known transcriptional target of Nrf2, enhances the nuclear translocation of nuclear respiratory factor 1 (NRF1), which controls the expression of several ETC components encoded in the nucleus (Evans and Scarpulla, 1990; Li et al., 2017b) and is an ARE target (Picca and Lezza, 2015). NRF1 subsequently promotes the expression of mitochondrial transcription factor A (TFAM), a protein that enters the mitochondrial matrix and binds to the “D-Loop” promoter region of the mitochondrial genome, thereby stimulating mtRNA expression and mtDNA replication (Piantadosi and Suliman, 2006). Upregulation of TFAM and other mitochondrial markers such as citrate synthase and mtDNA copy number led MacGarvey et al. (2012) to conclude that Nrf2-induced mitochondrial biogenesis is a primary driver of increased survival in a rat sepsis model. Nrf2-mediated mitochondrial biogenesis also contributes to normal lipid metabolism. Fatty acid oxidation is decreased in mitochondria isolated from Nrf2 genetic knockout (Nrf2^−/−) mice but is increased in KEAP1 knockout mice (Ludtmann et al., 2014). Furthermore, abnormal lipid metabolism is curbed by the administration of sulforaphane, which increases mitochondrial membrane potential, ETC activity, and utilization of lipids via the promotion of mitochondrial turnover (Lei et al., 2019). The p62-mediated mitophagy inducer (PMI), which acts to disrupt the Nrf2-KEAP1 association and promotes Nrf2-associated activity, induces mitophagy independent of the traditional PINK/Parkin pathway (Bertrand et al., 2015; East et al., 2014). This demonstrates that Nrf2 not only contributes to mitochondrial biogenesis but also to mitophagy, indicating that it is an important regulator of mitochondrial homeostasis.

There is significant crosstalk between mitochondrial biogenesis and the AMP-activated protein kinase (AMPK). AMPK activates peroxisome proliferator-activated receptor-γ (PPAR-γ) coactivator α (PGC-1α), known as the master regulator of mitochondrial biogenesis (Jäger et al., 2007), as a deficiency in AMPK results in decreased mitochondrial complex subunits and mtDNA copy number (Cao et al., 2014). Moreover, p38, which is positively regulated by PGC-1α, inhibits GSK-3β (Choi et al., 2017) [Fig. 1D]. Therefore, while Nrf2’s activity is regulated by PGC-1α activity (and therefore AMPK), they work in tandem to induce mitochondrial biogenesis (Bruni et al., 2010). To further illustrate this point, an Nrf2-sensitive ARE is located in the promoter region for PPAR-γ, the co-activator of PGC-1α (Pi et al., 2010).

Nrf2 may also play a more direct role in mitochondrial metabolic activity and ROS formation. Many transcriptional targets of Nrf2 positively affect mitochondrial membrane potential and the production of ATP (Dinkova-Kostova and Abramov, 2015). A byproduct of increased ETC activity is increased ROS generation, but this effect might be more than compensated by an increase in Nrf2-associated mitochondrial antioxidant gene expression (Holmström et al., 2013). The mitochondrial electron transport chain component NADH dehydrogenase [ubiquinone] 1 α subcomplex subunit 4 (NDUFA4) is upregulated in human breast epithelial cells by treatment with sulforaphane (Agyeman et al., 2012). Similarly, ARE/ARE-like sequences are present in the promoter region of ATP synthase F1 subunit α (atp5F1α) (Abdullah et al., 2012). Overexpression of uncoupling proteins (UCP) can inhibit superoxide formation through increased inner membrane proton conductance (Kim-Han and Dugan, 2005; Mattiasson et al., 2003) and provides neuroprotection in mouse models of Parkinson’s disease (Conti et al., 2005). One of the UCPs, UCP3, is a transcriptional target of Nrf2 (Anedda et al., 2013).

7. Sex differences in Nrf2

Sex is an important biological variable in the response to brain injury. For example, microglia, which are the resident macrophages in the brain, present a sexually dimorphic presence across anatomical regions, including the hippocampus and cortex, that are particularly vulnerable to acute brain injury (Schwarz et al., 2012). Sexually dichotomized responses to inflammatory stimuli have also been observed. In several models of acute brain injury, female animals exhibit milder inflammatory responses to injury than males (Acaz-Fonseca et al., 2015; Bodhankar et al., 2015; Caplan et al., 2017). While the mechanism that underlies these sex differences is unknown, clinical studies indicate that pre-menopausal adult females are more resistant to ischemic brain injury than age-matched males; however, post-menopausal females are more vulnerable (Rahimian et al., 2019) (but see Gupte et al. (2019)). As a result, female sex hormones estrogen and progesterone have been investigated as a potential mediator of the protective effect of sex on brain injury (Robertson et al., 2006a) although clinical trials using estrogen therapy have not yielded promising results (Hendrix et al., 2006). Therefore, alternative explanations to sex hormones for female resilience to injury still need to be investigated. There are several recent reviews of current research on sex differences after brain injury (Bushnell et al., 2018; Gupte et al., 2019; Mollayeva et al., 2018; Spychala et al., 2017).

To date, sex differences in the expression, activation, and signaling of Nrf2 in the brain have not been adequately explored. Indeed, most animal studies using Nrf2 activators rely on a single sex, making it impossible to determine if Nrf2 pathway activation is sexually dimorphic. For example, basal Nrf2 mRNA and/or protein are lower in females compared to males in brain (Zawada et al., 2015) and in liver (Rohrer et al., 2014). It is unclear if these basal differences translate into sex-specific responses to injury or drugs. A cerebral ischemia study showed that estrogen increases Nrf2 signaling in ovariectomized female mice and improves short-term outcomes, including inflammatory signaling, edema, and neurologic function (Li et al., 2017a); it is unclear how intact male or female mice would respond to similar treatment. A study examining sulforaphane induction of glutathione S-transferase activity in the liver failed to reveal sex differences in healthy Fischer and Sprague-Dawley rats, though it is unknown how injured animals may respond. Further research should examine the time course of Nrf2 pathway induction in the brain following trauma or ischemia in both sexes and compare responses to Nrf2 inducers. For recent reviews on sex differences and mitochondrial function in health and disease, see Ventura-Clapier et al. (2017) and Demarest and McCarthy (2015).

8. Nrf2 in animal models of acute brain injury

The Nrf2 pathway is activated in response to several forms of acute brain injury. Following TBI, Nrf2 immunoreactivity is increased in pericontusional neurons (but see Shih et al. (2003)), astrocytes, and microglia, with neuronal-selective localization peaking at 1 day post injury (dpi) and glial expression at 7 dpi (Dong et al., 2019). In both TBI and cerebral ischemia models, Nrf2 is decreased in the cytoplasm and/or increased in the nucleus at 1 dpi (Dai et al., 2018a; Ding et al., 2017b; Li et al., 2013; Li et al., 2016b), indicating release of Nrf2 from KEAP1 and nuclear translocation soon after injury. However, studies on the time course of Nrf2 and oxidative damage following acute brain injury suggest that while the insult activates the Nrf2 pathway, its induction is delayed, thus diminishing its potential to provide antioxidant protection during initial injury-induced increases in ROS production (Miller et al., 2014; Srivastava et al., 2013; Tanaka et al., 2011). The authors suggest that earlier activation with pharmacological Nrf2 inducers could help minimize oxidative damage and improve neurologic outcomes.

Several studies indicate that genetic knockout of Nrf2 results in increased injury severity. In a weight-drop TBI model, Nrf2^−/− mice exhibit increased edema, oxidative stress, inflammation, and cell death. Nrf2^−/− mice are also characterized by reduced glutathione and worse neurological severity scores (NSS) after TBI relative to wild-type mice (Ding et al., 2017b; Jin et al., 2008; Jin et al., 2009; Lu et al., 2015). Similar results were obtained using CCI-induced TBI, with Nrf2^−/− mice showing increased oxidative stress and decreased activation of antioxidant pathways (Cheng et al., 2013; Hong et al., 2010). The same effects of Nrf2 knockout are observed in models of cerebral ischemia. Following focal ischemia (i.e., stroke), Nrf2^−/− animals exhibit increased cerebral infarct, decreased antioxidant gene expression pathway induction, reduced angiogenesis, and worse neurological deficit scores relative to wild-type (Chen et al., 2019; Li et al., 2013; Shih et al., 2005). Importantly, differences in infarct volume between Nrf2^−/− and wild-type are not evident at 24 h; it is only after a delay of several days that this difference becomes apparent, suggesting that persistent oxidative stress, inflammation, and/or mitochondrial dysfunction due to Nrf2 deficiency is to blame rather than a difference in the severity of the initial injury (Shih et al., 2005). Nrf2 knockouts also display increased expression of TLR4, NLRP3, and NFκB, key regulators of inflammatory signaling (Hou et al., 2018; Li et al., 2013; Molteni et al., 2016).

9. Pharmacologic activation of Nrf2 pathways in acute brain injury

The activity of many cytoprotective electrophilic compounds depend at least in part on Nrf2-dependent gene expression. These include sulforaphane, dimethyl fumarate (Tecfidera^®), terpenes/terpenoids (including CDDO-Me, carnosic acid, oleanolic acid, and ursolic acid), and polyphenols (including curcumin, resveratrol, quercetin). Other compounds also directly or indirectly activate Nrf2, such as omega 3 fatty acids and valproic acid (Chen et al., 2018; Zhu et al., 2018), but for the sake of this review, we will focus on the major known activators. See Table 1 for a summary of Nrf2 activators in animal models of acute brain injury.

Table 1.

A summary of animal studies employing agents that target Nrf2 in traumatic and ischemic brain injury including: the injury models used, evidence of Nrf2 involvement, and outcome measures with a fraction of the cited studies reporting this outcome in parentheses.

Drug	Injury model	Evidence of Nrf2 involvement	Outcome measures	References
Sulforaphane	Cerebral Ischemia (middle cerebral artery occlusion; MCAO, carotid artery occlusion)	↑Nrf2 downstream factors ↓Nuclear Nrf2	↑Neurobehavioral function (2/5) ↓Cell death factors (1/5) ↓Lesion volume (3/5) ↓Inflammatory markers (3/5) ↓Oxidative stress markers (1/5)	(Ma et al., 2015; Mao et al., 2019; Ping et al., 2010; Yu et al., 2017; Zhao et al., 2006)
	TBI (controlled cortical impact; CCI)	↑Nrf2 downstream factors ↓Drug effect in Nrf2−/− ↓Nuclear Nrf2	↑Neurobehavioral function (2/3) ↓Cell death factors (1/3) ↓Oxidative stress markers (2/3) ↓Lesion volume (1/3) ↓Edema (1/3)	(Dash et al., 2009; Hong et al., 2010; Zhao et al., 2005)
Dimethyl or Monomethyl Fumarate	Cerebral Ischemia (middle cerebral artery occlusion; MCAO)	↑Nrf2 downstream factors ↓Drug effect in Nrf2−/− ↑Antioxidant measures	↑Neurobehavioral function (3/4) ↓Cell death factors (1/4) ↓Lesion volume (2/4) ↓Inflammatory markers (3/4) ↓Oxidative stress markers (3/4)	(Clausen et al., 2017; Kunze et al., 2015; Singh et al., 2019a; Yao et al., 2016)
	TBI (controlled cortical impact; CCI)	↑Nrf2 downstream factors ↑Antioxidant measures	↓Edema (3/4) ↑Neurobehavioral function (2/2) ↓Lesion volume (2/2) ↓Cell death factors (1/2) ↓Inflammatory markers (1/2) ↓Oxidative stress markers (2/2)	(Casili et al., 2018; Kramer et al., 2017)
Terpenes/Terpenoids (ursolic acid, oleanolic acid, carnosic acid, bardoxolone)	Cerebral Ischemia (carotid artery occlusion, middle cerebral artery occlusion; MCAO)	↑Nrf2 downstream factors ↓Drug effect in Nrf2−/− ↓Nuclear Nrf2	↑Neurobehavioral function (4/5) ↓Cell death factors (3/5) ↓Lesion volume (5/5) ↓Inflammatory markers (3/5) ↓Oxidative stress markers (2/5)	(Li et al., 2013; Wang et al., 2016b; Wang et al., 2018; Yamauchi et al., 2016; Zhang et al., 2012)
	TBI (weight drop, controlled cortical impact; CCI)	↑Nrf2 downstream factors ↓Nuclear Nrf2 ↓Drug effect in Nrf2−/− ↑Antioxidant measures	↑Neurobehavioral function (2/3) ↓Cell death factors (1/3) ↓Oxidative stress markers (2/3)	(Ding et al., 2017a; Maynard et al., 2019; Miller et al., 2015)
Polyphenols (curcumin, resveratrol, quercetin, epicatechin, tBHQ)	Cerebral Ischemia (middle cerebral artery occlusion; MCAO)	↑Nrf2 downstream factors ↓Drug effect in Nrf2−/− ↓Nuclear Nrf2 ↑Antioxidant measures	↑Neurobehavioral function (6/7) ↓Cell death factors (3/7) ↓Lesion volume (6/7) ↓Inflammatory markers (3/7) ↓Oxidative stress markers (2/7)	(Chen et al., 2019; Dai et al., 2018b; Leonardo et al., 2013; Li et al., 2016a; Narayanan et al., 2015; Wicha et al., 2017; Yang et al., 2009)
	TBI (weight drop, controlled cortical impact; CCI)	↑Nrf2 downstream factors ↓Drug effect in Nrf2−/− ↓Nuclear Nrf2 ↑ Antioxidant measures	↑Neurobehavioral function (3/5) ↓Cell death factors (4/5) ↓Lesion volume (2/5) ↓Inflammatory markers (2/5) ↓Oxidative stress markers (5/5) ↓Edema (3/5)	(Cheng et al., 2016; Dai et al., 2018a; Ding et al., 2017b; Dong et al., 2018; Li et al., 2016b)

9.1. Sulforaphane

Sulforaphane is an isothiocyanate derived from cruciferous vegetables, such as broccoli sprouts, and is among the most potent inducers of the Nrf2 expression pathway, including both upregulation of Nrf2-linked antioxidants as well as mitochondrial biogenesis (Lei et al., 2019; Posner et al., 1994; Prestera and Talalay, 1995; Zhu et al., 2008). Sulforaphane has been shown to decrease MPTP sensitivity to Ca²⁺, likely due to increased antioxidant levels in mitochondria following Nrf2-induced gene expression (Greco and Fiskum, 2010; Greco et al., 2011). A synthetic version, SFX-01, is currently in Phase II clinical trials for the prevention of vasospasm and delayed cerebral ischemia following subarachnoid hemorrhage (National Clinical Trial (NCT02614742).

9.2. Monomethyl and dimethyl fumarate

Of all the Nrf2 activators, only one compound is approved for the treatment of a neurodegenerative disease in humans in the USA and Europe. Tecfidera^® (BG-12 or dimethyl fumarate; DMF) is prescribed for relapsing-remitting multiple sclerosis (MS) (Gold et al., 2012). DMF and its metabolite monomethyl fumarate (MMF) have also been investigated in animal models of acute neurological injury. As a pro-drug, DMF is converted into MMF by GSH (Schmidt et al., 2007). DMF and MMF activate the Nrf2 pathway. However, DMF also exhibits immunomodulatory properties, even in Nrf2^−/− mice, confirming that it acts through multiple mechanisms, including mitochondrial biogenesis and NFκB inhibition (Gillard et al., 2015; Hayashi et al., 2017; Linker et al., 2011; Schulze-Topphoff et al., 2016). DMF has also been shown to be protective in an animal model of subarachnoid hemorrhage (Iniaghe et al., 2015), another form of acute brain injury associated with significant oxidative stress linked to neuronal death and mitochondrial dysfunction (Chen et al., 2015; Han et al., 2017).

9.3. Terpenes/terpenoids

Terpenes/Terpenoids are compounds most commonly derived from natural sources such as rosemary, olive oil, garlic, and apples. A related synthetic compound has also been developed, CDDO methyl (CDDO-Me; also called RTA 402 or bardoxolone methyl), that made it to phase 3 clinical trials (BEACON) for chronic kidney disease and type 2 diabetes. While the trial was terminated early due to an increased rate of heart failure, retrospective analyses suggest that, in the absence of certain predictors of heart failure, the risk of cardiac events in drug- and placebo-treated patients was the same (Chin et al., 2014; de Zeeuw et al., 2013), suggesting that the drug may be tested again for other conditions.

9.4. Polyphenols

Polyphenols are naturally-occurring chemicals found in a variety of consumables including tea, red wine, spices, fruits and vegetables, etc. Many polyphenol-rich foods form important parts of diets associated with longevity, such the Mediterranean diet, making them attractive compounds in studies focused on enhancing health (Martinez-Huelamo et al., 2017). In addition to Nrf2 activation, polyphenols have been investigated for their ability to control inflammatory signaling as well as diminish protein aggregation associated with neurodegenerative disease in animal models (Almeida et al., 2016; Caruana et al., 2016; Shimizu, 2017).

10. Nrf2 activators in other neurodegenerative disorders

In addition to testing Nrf2 activators for neuroprotection in acute CNS injury models, they are also being tested for treatment of chronic neurological disorders. As previously mentioned, Tecfidera^® is FDA-approved to treat the relapsing-remitting form of MS, an autoimmune demyelinating disease; it is also being investigated in the primary-progressive form of MS (NCT02959658). A new form of the drug, diroximel fumarate (which is also converted into MMF) is currently in phase 3 trials and may have better gastrointestinal tolerability (Palte et al., 2019). A terpenoid compound (omaveloxolone or RTA 408; NCT02255435) and a polyphenol (micronized resveratrol; NCT03933163) are both in phase 2 trials as therapies for Friedrich’s ataxia, a genetic disorder characterized by progressive ataxia, neurodegeneration, mitochondrial dysfunction, and oxidative stress (Lynch et al., 2018; Rezende et al., 2019).

In animal epilepsy models, sulforaphane acts as an anticonvulsant (Carrasco-Pozo et al., 2015) whereas other drugs with Nrf2-activating properties, such as the type-2 diabetes medication sitagliptin and the vasodilator pentoxifylline, are neuroprotective, decreasing neuroinflammation, oxidative stress, and cell death (Ahmad, 2013; Kaur et al., 2014; Pauletti et al., 2019; Singh et al., 2019b). Perhaps most important for long-term health implications, acute treatment with the KEAP1 inhibitor RTA 408 after an epileptic event resulted in greatly reduced frequency of subsequent occurrences for as much as multiple months, as well as increased ATP production and decreased neuronal loss (Shekh-Ahmad et al., 2018). In age-related neurodegenerative diseases like Parkinson’s and Alzheimer’s disease, pharmacological modification of Nrf2 signaling is similarly beneficial (Ahuja et al., 2016; Campolo et al., 2017; Choi and Lim, 2010; Jazwa et al., 2011; Lastres-Becker et al., 2016; Majkutewicz et al., 2018; Morroni et al., 2013; Pu et al., 2018; Rojo et al., 2018; Siebert et al., 2009; Zhang et al., 2014; Zhou et al., 2016).

Natural aging may also be a target for Nrf2 activators since aging is associated with oxidative stress and mitochondrial dysfunction (Fivenson et al., 2017; Sykiotis and Bohmann, 2010). Impaired Nrf2 signaling has been observed in aged mouse cardiac and human skeletal muscle (Gounder et al., 2012; Safdar et al., 2010). Huang et al. (2019) concluded that decreased physical activity and mitochondrial mass in aged mice is directly caused by diminished Nrf2 activity, as evidenced by middle-aged Nrf2^−/− displaying characteristics of frailty and sarcopenia. In the brain, Navarro et al. (2016) showed that Nrf2-mediated biogenesis can decelerate age-related microglial dysfunction and hinder the progression of neurodegenerative disorders.

11. Protection against cardiac arrest-induced neuronal death and neurologic injury by post-ischemic administration of sulforaphane

11.1. Background

As discussed earlier, a variety of drugs and nutraceuticals have been tested for neuroprotection in animal models of acute ischemic and traumatic brain injury. Sulforaphane is one of these agents that is therapeutically effective in many of these models but has not been tested for neuroprotection in a clinically-relevant model. A study was therefore conducted focusing on the ability of post-ischemic sulforaphane administration to improve neurohistologic and neurobehavioral outcomes in a highly clinically-relevant large animal model of cardiac arrest and resuscitation.

11.2. Methods

A detailed description of this model, the stereologic quantification of hippocampal neuronal death, and the neurologic deficit scoring system can be found in Balan et al. (2006). All animal procedures were performed as approved by the University of Maryland Baltimore Institutional Animal Care and Use Committee and in accordance with ARRIVE guidelines. Ventricular fibrillation cardiac arrest was induced in chloralose-anesthetized and intubated adult female (9–12 kg) purebred beagles (Covance) by stopping the ventilator and applying electrical stimulation to the cardiac surface with a Grass stimulator. Ten min later, dogs were ventilated with 100% O₂ and the heart resuscitated using 3 min open-chest CPR and electrical defibrillation. Following defibrillation, animals were maintained on 100% oxygen for one hr and then 28–35% oxygen for an additional 23 hr. All 11 animals received constant critical care with a morphine drip and additional boluses as needed for analgesia. A subgroup of animals (n=4) received an intravenous 5 min infusion of sulforaphane (LKT Laboratories Inc., catalog number S-8044) at 1 mg per kg in saline containing 0.6% dimethyl sulfoxide immediately after defibrillation. Pancuronium was used intermittently, only as needed, to facilitate ventilation and only after verification that the animal was adequately anesthetized. Ventilation was adjusted to maintain arterial carbon dioxide and oxygen in normal ranges.

At hr 20, anesthesia and analgesia were lightened, and animals weaned from controlled ventilation. At hr 23, the effects of morphine were reversed with naloxone and the neurological deficit score (NDS) was measured by two examiners blinded to treatment protocol. The NDS used has been validated in other laboratories (Bircher and Safar, 1985), as well as ours (Rosenthal et al., 1992) and is a multisystem test of 18 parameters in 5 categories (level of consciousness, respiration, cranial nerve, motor, sensory, behavior), yielding a score between 0 (normal) and 100 (brain death). Following the 5 min NDS procedure, the dogs were re-anesthetized with pentobarbital and α-chloralose, intubated, and mechanically ventilated. At 24 hr post-resuscitation, the deeply anesthetized animals were perfusion-fixed with paraformaldehyde. The fixed brains were removed and processed for histopathology.

11.3. Stereologic quantification

Six dorsal hippocampus sections from a 1-in-24 series representing the entire rostral caudal formation were stained with cresyl violet and quantified with unbiased stereology by a microscopist blinded to treatment. Boundaries of the hippocampal CA1 fields were outlined at low magnification (4x), after which StereoInvestigator software placed within each subfield boundary a set of optical counting frames (55×55 μm) in a systematic random fashion. Cresyl violet stained normal and dying neurons (ie, necrotic and apoptotic cells) were then counted in optical dissectors 7 μm in depth at high magnification (100x), according to stereological principles.

11.4. Statistical analysis

All data are expressed as mean ± SE. NDS scores and cell death counts were analyzed using a t-test where significance was set at p <0.05.

11.5. Results

As described by many labs using cardiac arrest with several species, the CA1 region of the hippocampus is one of the more vulnerable brain regions that undergo considerable neuronal death within just 24 hr following arrest and resuscitation (Back et al., 2004). Fig. 2 provides representative images of hippocampal brain slices at an anteroposterior distance of 40 mm from the rostral olfactory bulb tip that were stained with either cresyl violet or Fluoro Jade B. Cresyl violet–stained neurons in a representative non-ischemic surgical Sham animal appeared normal, as expected. At high resolution, dying neurons in a representative sample from a dog that underwent arrest/resuscitation appeared shrunken, with condensed nuclei lacking a nucleolus. Hippocampal neurons in dogs that received sulforaphane immediately after resuscitation were not entirely normal but displayed identifiable nuclei and were not as condensed as neurons in the brains of dogs that did not receive sulforaphane. Brain slices were also stained with Fluoro Jade B, which is selective for dead or dying neurons (Balan et al., 2006; Park et al., 2013). Sections stained with Fluoro Jade B exhibited little staining in Shams and in brains of dogs treated with sulforaphane; however, Fluoro Jade B staining was much greater in animals that underwent cardiac arrest.

Fig. 2.

Hippocampal neuronal death at 24 hr after cardiac arrest and protection by sulforaphane. Canine brain sections were stained with cresyl violet for morphologic alterations and with Fluoro Jade B (FJB) for neuron-specific damage and death. Following Sham cardiac arrest, hippocampal CA1 neurons appeared healthy with homogeneous discrete nuclei and cell as shown with cresyl violet-stained sections. Following cardiac arrest and 24 hr resuscitation, most neurons displayed dark, pyknotic nuclei and were surrounded by cell debris. Sulforaphane administration reduced but did not eliminate neuronal abnormalities. Sections stained with Fluoro Jade B exhibited little staining in Shams and in dogs following cardiac arrest and treatment with sulforaphane. Fluoro Jade B was much greater in animals that underwent cardiac arrest with no sulforaphane treatment.

Fig. 3 provides the results of comparisons between neurologic deficit scores and cresyl violet-stained hippocampal CA1 neuronal death for dogs treated with sulforaphane and those used previously without sulforaphane administration. The NDS for sulforaphane-treated dogs was significantly lower than the NDS reported previously for animals that received hyperoxic resuscitation but no sulforaphane (24.1 ± 3.8% vs 61.2 ± 4.9%, p < 0.001, t-test, n=4 and 7 for sulforaphane treated and untreated, respectively). The fraction of neurons that were dead or dying was significantly lower in the hippocampus after administration of sulforaphane compared to the fraction previously reported for hyperoxic resuscitation without sulforaphane (6.8 ± 2.5% vs 63.7±3.1%, p< 0.001, t-test, n=4 and 7 for sulforaphane-treated and untreated, respectively).

Fig. 3.

Neurologic deficit scores and hippocampal neuronal death at 23–24 hr post-cardiac arrest and resuscitation. Neurologic deficit scores (NDS) range from zero (normal) to 100 (brain dead). Sulforaphane treatment significantly reduced the NDS by more than 50% (*p < 0.001; n=4–7). Sulforaphane also reduced dead or dying CA1 neurons by almost 90% (p < 0.001; n=4–7).

11.6. Discussion

These results provide preliminary evidence strongly suggesting that one dose of intravenously-administered sulforaphane to young, healthy, adult female beagles following 10 min ventricular fibrillation cardiac arrest provides dramatic neuroprotection in a highly clinically translational model of global cerebral ischemia and resuscitation. This conclusion is, however, limited by several factors. First, the dogs that did not receive sulforaphane also did not receive drug vehicle, and were previously used in another study, albeit with identical experimental methods (Balan et al., 2006). Second, the differences in neurologic characteristics and histopathology between treatment groups was limited to the very short outcome time of 23–24 hr, which is restricted by the guidelines for maintaining lab animals that meet criteria for euthanasia. Other results obtained with the canine cardiac arrest model have been reproduced in a rat global cerebral ischemia model at 30 days post-injury (Hazelton et al., 2010), which was possible because the rats did not meet euthanasia criteria. Third, only females were used in this study and therefore similar tests should be conducted in males. Fourth, the therapeutic potential of sulforaphane was tested using a hyperoxic resuscitation paradigm that generates much greater oxidative stress and neurologic injury than when the animals are maintained under normoxic conditions (Liu et al., 1998; Richards et al., 2007; Richards et al., 2006; Vereczki et al., 2006). Therefore, sulforaphane should also be tested in normoxic resuscitation paradigms that reduce oxidative stress.

11.7. Conclusions

Regulation of Nrf2 activity is mediated by the redox-dependent release of Nrf2 from KEAP1, the phosphorylation state of Nrf2, proteasomal degradation of Nrf2, and the distribution of Nrf2 between cytosolic, nuclear, and possibly mitochondrial compartments. When activated, Nrf2 has important effects on mitochondria, including stimulation of mitochondrial biogenesis and mitophagy, increased mitochondrial levels of antioxidant enzymes, and resistance to redox-regulated mitochondrial permeability transition pore opening.

Nrf2-mediated gene expression moderates several pathogenic processes (e.g. inflammation, oxidative stress, and mitochondrial bioenergetic dysfunction) that contribute to neuronal death and neurologic injury following both acute and chronic neurodegeneration. While we show that sulforaphane reduces hippocampal neuronal death and improves neurologic outcomes in a highly translatable large animal model of cardiac arrest and resuscitation, very few of the pharmacologic agents that stimulate Nrf2-mediated cytoprotective gene expression have been rigorously tested for effectiveness and potential toxicity in clinical trials or even clinically-relevant models of acute neurologic disorders and chronic neurodegenerative diseases. Future research on Nrf2-activating agents should emphasize translatability in model and dosing, as well as investigate potential effects of sex and age on Nrf2-mediated gene expression and neuroprotection.