The KEAP1-NRF2 pathway: Targets for therapy and role in cancer

Abstract

The KEAP1-NRF2 pathway is the key regulator of cellular defense against both extrinsic and intrinsic oxidative and electrophilic stimuli. Since its discovery in the 1990s, its seminal role in various disease pathologies has become well appreciated, motivating research to elucidate the intricacies of NRF2 signaling and its downstream effects to identify novel targets for therapy. In this graphical review, we present an updated overview of the KEAP1-NRF2 signaling, focusing on the progress made within the past ten years. Specifically, we highlight the advances made in understanding the mechanism of activation of NRF2, resulting in novel discoveries in its therapeutic targeting. Furthermore, we will summarize new findings in the rapidly expanding field of NRF2 in cancer, with important implications for its diagnostics and treatment.

Abbreviations:

15d-PGJ₂

15-deoxy-Δ [12,14]-prostaglandin J₂

17-AAG

17-(Allylamino)-17-demethoxygeldanamycin

2SC

S-(2-succinyl)cysteine

4-HNE

4-hydroxynonenal

AKR1C3

aldo-keto reductase family 1 member C3

AKT

protein kinase B

ARE

antioxidant response element

BACH1

BTB domain and CNC homolog 1

BRAF

B-Raf proto-oncogene

BRG1

Brahma-related gene 1

BTB

Broad complex, Tramtrack, Bric-a-Brac

CDKN2A/B

cyclin-dependent kinase inhibitor 2A/B

CHD6

chromodomain helicase DNA binding protein 6

c-MYC

c-myelocytomatosis oncogene product

CR

cullin repeats

CRBN

cereblon

CREBBP

CREB Binding Protein

CRL

Culling-RING E3 ligase

CSN

COP9 signalosome

CUL1

cullin 1

CUL3

cullin 3

DCN1

defective in Cul neddylation 1 protein

DGR

double-glycine repeat

DON

6-Diazo-5-oxo-l-norleucine

EMSY

EMSY transcriptional repressor, BRCA2 interacting

G6P

glucose 6-phosphate

G6PD

glucose 6-phosphate dehydrogenase

G6PDi-1

glucose-6-phosphate dehydrogenase inhibitor 1

GCV

ganciclovir

GLS

glutaminase

GLUT1

glucose transporter 1

GPNA

L-γ-Glutamyl-p-nitroanilide

GSK-3β

glycogen synthase kinase-3 beta

HSP90

heat shock protein 90

HSV

herpes simplex virus

IFNγ

interferon gamma

IL2

interleukin 2

IVR

intervening region

KEAP1

kelch-like ECH-associated protein 1

KRAS

Kirsten rat sarcoma virus

LUAD

lung adenocarcinoma

MED16

mediator complex subunit 16

NADPH

reduced nicotinamide adenine dinucleotide phosphate

NAE1

NEDD8-activating enzyme E1

NEDD8

neural precursor cell expressed developmentally down-regulated 8

Neh

NRF2-ECH homology

NQO1

NAD(P)H quinone dehydrogenase 1

NRF2

NF-E2–related factor 2

NSCLC

non-small cell lung cancer

OA-NO₂

nitro-oleic acid

p21Cip1/WAF1

cyclin-dependent kinase inhibitor 1

PD-L1

programmed death-ligand 1

PI3K

phosphatidylinositol 3-kinase

PPIs

protein-protein interactions

PPP

pentose phosphate pathway

PROTACs

proteolysis targeting chimeras

PTEN

phosphatase and tensin homolog

R5P

ribose 5-phosphate

RAC3

receptor-associated coactivator 3

RBX1

RING-box protein 1

ROC1

regulator of cullins-1

ROS

reactive oxygen species

RTK

receptor tyrosine kinase

RXRα

retinoid X receptor alpha

SFN

sulforaphane

SLC1A5

solute carrier family 1 member 5

SLC33A1

solute carrier family 33 member 1

sMAF

small Maf (musculoaponeurotic fibrosarcoma)

SOX2

SRY (sex determining region Y)-box transcription factor 2

SQSTM1/p62

sequestosome-1

STING

stimulator of interferon genes

SUCO

SUN domain containing ossification factor

SV-40

simian vacuolating virus 40

TAPT1

transmembrane anterior posterior transformation 1

tBHQ

tert-Butylhydroquinone

TCA

tricarboxylic acid

TK

thymidine kinase

TNFα

tumor necrosis factor alpha

UBE2M

ubiquitin conjugating enzyme E2 M

xCT

cysteine/glutamate antiporter

β-TrCP

β-transducin repeat-containing protein

1. Introduction

The redox-activated transcription factor NF-E2 related factor 2 (NRF2) is the key regulator of cellular defense against both extrinsic and intrinsic oxidative and electrophilic stress. Since its discovery in the 1990s, its seminal role in the protection against environmental insults has become well understood, as is the importance of Kelch-like ECH-associated protein 1 (KEAP1) in the regulation of its activity. Moreover, the protective role of NRF2 in various age-related degenerative diseases has been recognized, providing an opportunity for the development of NRF2-activating drugs for their treatment. In contrast to the transient activation of NRF2 by activating agents, enhanced sustained NRF2 activity in cancer results in resistance to chemo- and radiotherapy and poor prognosis, necessitating the development of approaches to limit NRF2 activity and downstream effects.

In the inaugural issue of Redox Biology in 2013, we provided a graphical overview of the KEAP1-NRF2 system and its role in cancer [1]. Herein, we present an update, focusing on the progress made within the past ten years. Specifically, we highlight the advances in the understanding of the mechanism of NRF2 activation, its upstream and downstream signaling and dysregulation in cancer as well as its role in the tumor microenvironment, in order to illustrate how NRF2 signaling can be targeted for therapy. The other protumorigenic facets of NRF2 such as its role in the initiation, progression, and metastasis have been extensively reviewed elsewhere [[2], [3], [4], [5]].

2. The KEAP1-NRF2 pathway and its activation

The key components of the KEAP1-NRF2 pathway are NRF2, KEAP1 as well as CULLIN3 (CUL3) scaffold protein of the E3 ubiquitin ligase complex (Fig. 1A–C). CUL3 and KEAP1 form dimers, whereas only one NRF2 protein is bound to a dimeric KEAP1 with two binding motifs, DLG and ETGE [6]. KEAP1 interacts with NRF2 via the Kelch domain, which has been structurally characterized in detail by X-ray crystallography [7]. Somatic mutations of NRF2 and KEAP1 within these binding sites are common in many cancers, especially in non-small cell lung cancer (NSCLC) as well as head and neck cancer [8,9]. In NRF2, the mutations are prevalent in the Neh2 domain where the DLG and ETGE motifs are present (Fig. 1A), whereas somatic mutations in KEAP1 occur randomly and throughout the length of the protein [10] (Fig. 1B). Nevertheless, there are certain mutation hotspots implicating that KEAP1 mutations may have cancer-relevant functions beyond NRF2 regulation [11].

Fig. 1.

Structures of NRF2, KEAP1 and CULLIN3. A) Human NRF2 protein has a complex multi-domain structure consisting of seven distinct functional regions called NRF2-ECH homology (Neh) domains [12]. The Neh2 domain serves as the degron of NRF2 in the KEAP1-Cul3-dependent degradation [13]. It contains the DLG and ETGE motifs which interact with the DGR domains of the KEAP1 homodimer. Between the DLG and ETGE domains, seven lysine residues conjugate with ubiquitin [14]. Neh4 and Neh5 are transactivation domains recruiting transcriptional coactivators and facilitating the interaction between the chromatin remodeling SWI/SNF complex and the mediator complex [15,16]. Neh7 is the retinoid X receptor (RXR) interacting domain that represses NRF2 through physical interaction with RXRα [17]. Neh6 is the KEAP1-independent degron having two serine residues in the DSGIS motif, Ser344 and Ser347, that are phosphorylated by GSK-3β, enabling the interaction with E3 ubiquitin ligase β-TrCP protein and NRF2 ubiquitination via CULLIN1/RBX1. β-TrCP can recognize and ubiquitinate NRF2 also through the DSAPGS motif (382–387) in a GSK-3β-independent manner [18]. The Neh1 domain dimerizes with small-Maf (sMaf) proteins [19,20]. The Neh3 domain is the COOH-terminal domain in NRF2, and it contains chromo-ATPase/helicase DNA-binding protein CHD6 recruiting transactivation domain. B) KEAP1 protein consists of five domain regions [13]. The BTB domain mediates the homodimerization of two KEAP1 molecules and contributes to the interaction with CUL3 [21]. The proximal part of the IVR region called the 3-box motif forms additional interactions with CUL3 [22,23]. The IVR domain also contains reactive cysteines, Cys273 and Cys288, recognized as sensors of oxidative and electrophilic stresses. The C-terminal part of KEAP1 contains six Kelch domains that are required for substrate capture, and which can bind separately the DLG or ETGE motifs of NRF2. C) CUL3 consists of three functional domains. The most N-terminal domain consists of three cullin repeats (CR1, CR2, and CR3), which bind to KEAP1 through the BTB- and IVR-domains of KEAP1 [24]. The Cullin homology domain is important for the binding of RING finger protein ROC1 [25,26]. The C-terminal domain is the Cullin NEDD8 domain that has been found to function as the cullin ubiquitin ligase activation binding site for NEDD8 at lysine 712 and provide a binding site for the zinc-finger motif of RBX1 protein [27,28].

NRF2 activity is largely regulated via its interaction with KEAP1. Under basal conditions, NRF2 is rapidly degraded through the 26S proteasome, KEAP1 serving as a substrate recognition component of the CUL3-dependent E3 ubiquitin ligase complex (Fig. 2A). On exposure to oxidative or electrophilic stress, KEAP1 cysteine residues are modified leading to a conformational change in KEAP1 and inhibition of the proteasome degradation machinery, allowing newly synthesized NRF2 to enter the nucleus and regulate target gene expression (Fig. 2B). KEAP1 is a cysteine-rich protein and studies have shown that three major cysteines, C151, C273, and C288, are modified by electrophiles (Fig. 2B). Interestingly, different electrophiles target distinct KEAP1 cysteines alone or in combination. For example, sulforaphane (SFN) and tert-butylhydroquinone (tBHQ) are categorized as KEAP1 C151-dependent inducers, whereas nitro-oleic acid (OA-NO₂), 4-hydroxynonenal (4-HNE) and 15-deoxy-Δ [29,30]-prostaglandin J₂ (15d-PGJ₂) target C273 and/or C288 [29,31]. In addition, C226, C613, C622, and C624 have been found to be important for sensing hydrogen peroxide [30]. Thus, KEAP1 serves as a redox-sensitive hub of many different stimuli, altering the conformation of the E3 ligase complex and inhibiting NRF2 ubiquitination [5].

Fig. 2.

NRF2 activation. A) In basal conditions, the NRF2-KEAP1 interaction targets NRF2 for proteasomal degradation via ubiquitination in the cytoplasm. B) During classical activation by electrophiles/oxidants, sensor cysteines in KEAP1, including C151, C273, and C288, are modified resulting in a change in KEAP1 conformation, which in turn prevents NRF2 ubiquitination [5,32,33]. C) NRF2 stability is also regulated by GSK-3β through phosphorylation of the serine residues of DSGIS motif within the Neh6 domain of NRF2, allowing the recognition by the β-TrCP-CUL1-based E3 ubiquitin ligase complex and proteasomal degradation [34,35]. Upstream, GSK-3β can be inhibited by PI3K/AKT pathway or activated through PTEN-dependent AKT inhibition. D) Small molecule inhibitors, ETGE motif containing peptides and PROTACs can disrupt the interaction between NRF2 and KEAP1, and lead to NRF2 activation [36]. E) Inhibition of NEDD8-activating enzymes NAE1 or DCN1 by small molecule inhibitors or by increasing deneddylation by CSN, results in the activation of NRF2 [37,38,39]. Activated NRF2 accumulates in the nucleus and interacts with other TFs and cofactors to regulate the transcription of its target genes, encoding various proteins involved e.g., in xenobiotic detoxification, antioxidant pathways, inflammatory and metabolic processes, as well as regulation of cell autophagy and cell death.

In addition to KEAP1-dependent regulation, NRF2 is regulated through the phosphatidylinositol 3-kinase (PI3K)/AKT pathway via glycogen synthase kinase-3 beta (GSK-3β) mediated phosphorylation, resulting in a β-transducin repeats-containing protein (β-TrCP)-CUL1-mediated degradation of NRF2 (Fig. 2C). GSK-3β is thought to play a modulatory role in NRF2 signaling, cytosolic KEAP1-dependent degradation being the primary mechanism by which NRF2 is being degraded [5].

The interaction between NRF2 and KEAP1 is mediated via ETGE and DLG motifs at the Neh2 domain of NRF2 [14] (Fig. 2A). Direct small molecule non-electrophilic and non-covalent inhibitors, with various scaffolds, have been discovered in recent years to disrupt protein-protein interactions (PPIs) between KEAP1 and NRF2 (reviewed in Ref. [40]) or KEAP1 and CUL3 [41] (Fig. 2D). In addition to small molecule inhibitors, ETGE motif containing peptides, such as E78P have been explored [42,43]. Also proteolysis targeting chimeras (PROTACs), bifunctional molecules with covalently linked ligand for protein of interest and E3 ligand, targeting KEAP1 for proteasomal degradation, have been developed [44,36] (Fig. 2D).

Neddylation is a ubiquitin-like posttranslational modification, in which neural precursor cell expressed developmentally downregulated protein 8 (NEDD8) is conjugated to target proteins destined for degradation [45,46]. NEDD8 is conjugated onto Cullin protein, which activates the Cullin-RING E3 ligase (CRL) and increases the ligase activity in CUL3-KEAP1 E3 complex, leading to increased ubiquitination of NRF2 [47]. Thus, inhibition of the cullin neddylation pathway by inhibiting a NEDD8-activating enzyme E1 (NAE1) or Defective in cullin neddylation 1 (DCN1) by small molecule inhibitors results in potent activation of NRF2 [37,38]. In addition, defective neddylation due to the mutations of NEDD8, UBE2M, or CUL3 has recently been shown to activate NRF2 in a subset of leiomyomas [48] (Fig. 2E).

3. Dysregulation in cancer

The KEAP1-NRF2 pathway is one of the central oncogenic pathways in cancers with a strong environmental component [8]. Even though NRF2 stabilization alone is not sufficient to induce tumorigenesis [2,49], in combination with other oncogenic lesions it provides the cancer cells with a survival advantage via upregulation of antioxidant and detoxifying enzymes and metabolic rewiring [[50], [51], [52], [53]]. In cancer cells, the KEAP1-NRF2 signaling system is frequently affected by alterations in the NRF2, KEAP1, or CUL3 encoding genes. Apart from somatic mutations (Fig. 3A), common genomic modifications include domain deletions, amplifications, and splice variants of these genes [54] (Fig. 3B and C). Furthermore, NRF2 activity can be affected by KEAP1 promoter hypermethylation, upstream transcriptional induction of the NFE2L2 gene, and KEAP1 thiol modification by oncometabolites [53,[55], [56], [57], [58], [59], [60]] (Fig. 3D–F). The KEAP1-NRF2 protein-protein interaction can also be disrupted by several proteins, many of which affect NRF2 stability [61]. A well-characterized example is SQSTM1/p62, which binds to and sequesters KEAP1 leading to autophagic degradation of KEAP1 and prolonged activation of NRF2 [62,63]. Also, the cyclin-dependent kinase inhibitor, p21^Cip1/WAF1, competes with KEAP1 for NRF2 binding in the DLG domain, allowing NRF2 stabilization [64] (Fig. 3G).

Fig. 3.

Dysregulation of NRF2 in cancer. NRF2 is activated in cancer through different mechanisms. A) A common mechanism of NRF2 activation is through gain-of-function (GOF) mutations in NFE2L2 and loss-of-function (LOF) mutations in KEAP1 genes [9,65]. B) NFE2L2 gene amplification increases the copy number of the NFE2L2 gene in the genome and hence protein expression [8,66]. C) Alternative splicing of NFE2L2 mRNA is found in lung squamous carcinoma and head and neck carcinoma. The resulting transcript variants lack the KEAP1-interacting Neh2 domain [67]. D) Epigenetic silencing of KEAP1 promoter can lead to NRF2 stabilization [57]. E) NRF2 expression can be increased by different oncogenes, such as KRAS^G12D, BRAF^V619E and MYC [53,58]. F) Posttranslational modification of KEAP1 by oncometabolites can result in NRF2 stabilization. In hereditary leiomyomatosis and renal cell carcinoma (HLRCC), LOF mutations in the TCA-cycle enzyme fumarate hydratase (FH), lead to the accumulation of the electrophilic metabolite fumarate, which induces the succination of critical cysteines (2SC) within KEAP1 and NRF2 accumulation [59,60]. G) Several protein-protein interactions have been shown to affect NRF2 activation [[68], [69], [70]]. The p21^Cip1/WAF1 protein binds to the DLG domain of NRF2, altering KEAP1 binding and causing NRF2 activation [64]; phosphorylated SQSTM1/p62 binds to KEAP1 escorting it to autophagic degradation [[70], [71], [72], [73]].

4. The effect of NRF2 on tumor-stroma interactions

Recently, the effect of NRF2 activation on tumor immunity has gained increasing attention, both with respect to cancer intrinsic as well as immune cell-related activation. Given the increasing importance of immune therapies such as immune checkpoint inhibitors as standard therapy for many cancers, the factors affecting their efficacy need to be thoroughly understood. The relationship between KEAP1 mutations and the low number of tumor-infiltrating lymphocytes (TILs) has been observed, suggesting an association with NRF2 hyperactivity and immune evasion [74]. Additionally, in squamous cancers, NRF2 hyperactive tumors comprise a molecular phenotype with co-occurring SOX2 amplification and CDKN2A/B loss that is associated with low lymphocyte and macrophage content irrespective of NRF2 activation mechanism [75]. Mechanistically, NRF2 has been shown to transcriptionally regulate the immune checkpoint molecule programmed cell death ligand 1 (PD-L1) in melanoma, and inhibition of NRF2 and PD-1 synergistically inhibited tumor growth [76]. Recent work showed that the loss of KEAP1 decreased tumoral CD8⁺ cells and increased M2 macrophages via stabilization of EMSY, which suppresses the type I interferon response [77]. Type I interferon response was also inhibited directly in vitro by NRF2 via inhibition of STING [78]. Others have suggested that inhibition of proinflammatory cytokines, enhanced antioxidant capacity, and immunomodulatory metabolites contribute to NRF2-associated immunoevasion [[79], [80], [81]] (Fig. 4).

Fig. 4.

NRF2 and tumor-stroma interaction. In cancer cells, NRF2 hyperactivity results in a low number of lymphocytes and macrophages within tumor stroma [74,75]. NRF2 hyperactivity is also shown to be involved in recruiting PD1+ T-cells while suppressing CD8⁺ T-cells in the tumor microenvironment [76]. Also, the expression and secretion of proinflammatory cytokines such as IFNγ, TNFα, and IL2, are inhibited by NRF2, and NRF2 provides protection against oxidative insults [79,81]. KEAP1 LOF is also associated with increased stabilization of the oncogene EMSY and its activity [77]. This leads to polarization of M2-macrophages, lower infiltration of CD8⁺ T-cells, and reduced secretion of proinflammatory cytokines [78]. By secreting immunomodulatory cytokines, cancer cells can activate anti-inflammatory and pro-tumorigenic M2 macrophages [77].

The response to immune therapies depends not only on cancer cell-intrinsic factors but also on factors affecting tumor resident immune cells and other stromal cells. NRF2 has well-acknowledged anti-inflammatory functions, and for example, T cell activation is under rigorous redox control in which NRF2 plays a major role [82]. In the context of cancer, the role of stromal NRF2 on tumor progression is ambiguous. It has been noted using mouse models that modest activation of NRF2 in tumor stroma suppressed the progression of NRF2 overexpressing malignant tumors [83], and that NRF2 deficiency exacerbates the formation of lung metastases following implantation of the mouse Lewis lung carcinoma cells [84]. On the other hand, it has been suggested that NRF2-activating drugs for diabetes can promote metastatic spread in xenograft mouse models, as does chronic administration of antioxidants via a mechanism involving BACH1/NRF2 interplay [85,86]. These dichotomous results may be explained by different experimental settings and further clarification of putative mechanisms and stromal cells involved using alternative approaches is warranted. This is important in the clinical context, given the development of NRF2 activators for the treatment of chronic diseases [40]. Therefore, any adverse effects such as the possibility of accelerating the spreading of tumors need to be clarified.

5. Inhibition of NRF2 activity

Given the protumorigenic effects of constitutive NRF2 activity, research for potential NRF2 inhibitors has gained more interest in the past years. However, there are no specific NRF2 inhibitors to date. Brusatol and halofuginone effectively decrease NRF2 protein by blocking its translation [[87], [88], [89]]. However, these compounds affect multiple short-lived proteins, increasing the chance for systemic adverse effects [90]. Through a large high throughput drug screening to search for novel NRF2 inhibitors, small molecules ML385 and AEM1 have been identified [91,92]. ML385 was shown to bind to the Neh1 domain of NRF2 thereby inhibiting the interaction between NRF2 and MafG and subsequent transcription of NRF2 target genes [91] (Fig. 5A). However, the efficacy and specificity of these compounds remain to be confirmed.

Fig. 5.

Inhibition of NRF2 activity. A) ML-385 inhibits NRF2-sMAF interaction and NRF2-dependent transcription [91]. B) Receptor tyrosine kinase (RTK) activation by receptor-specific ligand triggers signal transduction that culminates in the activation of PI3K/AKT axis inhibiting GSK-3β. Inhibition of PI3K by LY294002 activates GSK-3β, which can phosphorylate the serine residues within the DSGIS motif in the Neh6 domain of NRF2. This results in β-TrCP-dependent proteasomal degradation of NRF2 [35]. C) NRF2 regulates the uptake of cystine ((Cys)₂) and simultaneous secretion of glutamate (Glu) by promoting the expression of xCT antiporter [93]. Once inside the cells, (Cys)₂ is reduced to cysteine (Cys) which can be used together with glutamate and glycine for GSH synthesis. NRF2-overactivity causes glutamine addiction that can be exploited therapeutically by targeting the glutamine transporter SLC1A5 with GPNA [94], GLS with CB-839 [[94], [95], [96]], or by using the glutamine antagonist DRP-104 to target simultaneously GLS activity and the glutamine-dependent reactions of nucleotide synthesis pathway. NRF2 transcriptionally controls also the oxidative PPP, to provide the cells with NADPH required for the reduction of oxidative species. The novel small-molecule inhibitor of G6PD [97], G6PDi-1, synergizes with CB-839 in inhibiting cancer growth [98]. D) NQO1- and AKR1C3-activated prodrugs and ARE-driven suicide gene therapy can target specifically NRF2-active cancers [99]. NQO1 is involved in the bioactivation of HSP90 inhibitor 17-AAG [100]. AKR1C3 catalyzes oxidoreduction of PR-104 to the respective alcohol with anticancer activity [101]. In suicide gene therapy, a lentiviral vector is used to transfer the ARE-driven suicide gene construct into the cells [102]. TK is produced in cancer cells with high NRF2 activity, leading to the phosphorylation of the ganciclovir (GCV) prodrug. Bioactivated GCV is toxic to the cells and promotes apoptosis.

In addition to direct targeting of NRF2, efforts have focused on identifying vulnerabilities upstream or downstream of NRF2. Oncogenic KRAS [53,58] and PI3K/AKT pathway [50] have been shown to support NRF2 transcription and activity in cancer. Therefore, PI3K inhibition is an effective strategy to inhibit NRF2 activity and increase chemotherapeutic efficacy in NSCLC and breast cancer [18,103] (Fig. 5B). Downstream of its activation, NRF2 is an important regulator of metabolic pathways [4]. It controls glutathione (GSH) biosynthesis, cystine uptake via xCT antiporter as well as NADPH synthesis via the pentose phosphate pathway (PPP), necessary to regenerate GSH. By increasing the use of intracellular amino acids for GSH synthesis, NRF2 creates metabolic vulnerabilities that can be exploited to target NRF2 active cancers. NRF2-dependent upregulation of xCT antiporter [93] increases cystine uptake at the expense of intracellular glutamate, rendering the cells exquisitely reliant on exogenous glutamine for glutamate generation [94,95] (Fig. 5C). Inhibition of glutaminase (GLS), the enzyme responsible for the conversion of glutamine to glutamate, with CB-839, has been relatively successful in NRF2-active NSCLC in vitro and in vivo [[94], [95], [96]]. Unfortunately, a clinical trial using CB-839 in combination with standard first-line therapy in NSCLC patients with somatic mutations in KEAP1 or NRF2 has been terminated prematurely due to a lack of clinical benefit (NCT04265534). The newly developed prodrug of the glutamine antagonist 6-Diazo-5-oxo-l-norleucine (DON), DRP-104, allows the targeting of all glutamine-dependent reactions selectively in cancer, due to the high enrichment of esterases and peptidases in the tumor microenvironment [104]. Glutamine antagonism can also potentiate the effector T-cell activation against cancer cells [104,105]. A clinical trial using DRP-104 in NSCLC patients with KEAP1 or NFE2L2 somatic mutations is ongoing (NCT04471415). Inhibition of glucose-6-phosphate dehydrogenase (G6PD), the rate-limiting enzyme in the PPP, synergizes with CB-839 in NRF2-active NSCLC [98] (Fig. 5C). A recent study, utilizing CRISPR-Cas9 screening, identified endoplasmic reticulum-resident proteins, i.e., SLC33A1, SUCO, and TAPT1, to be synthetic lethal in KEAP1-mutant lung adenocarcinoma (LUAD) cells [106]. This study highlights the power of functional genomics approaches for the identification of novel therapeutic targets. Prodrugs activated through NRF2 target genes represent a promising category of novel anticancer treatments (Fig. 5D). Several NAD(P)H:quinone oxidoreductase 1 (NQO1) and aldo-ketoreductase 1C 3 (AKR1C3) bioactivatable compounds have been described [[100], [101], [107], [108], [109]]. Both NQO1 and AKR1C3 are bona fide NRF2 target genes and are highly expressed in many cancers. Constitutive ARE activity has also been exploited in suicide gene therapy using ARE-driven lentiviral vectors expressing the thymidine kinase (TK) gene of herpes simplex virus (HSV) [102] (Fig. 5D).

6. Conclusions and future perspectives

In this graphical review, we outlined novel discoveries in NRF2 signaling with the perspective to highlight aspects relevant for therapeutic targeting. With respect to drugs augmenting NRF2 activity, electrophilic NRF2 activators covalently modifying KEAP1 thiols have entered clinical trials, and dimethyl fumarate [40] and omaveloxolone [110] have already an FDA approval for multiple sclerosis and Friedrich's ataxia, respectively. However, given their ability to react with cysteines on other proteins, covalent NRF2 activators have potential adverse off-target effects and therefore preclinical studies are focusing on the development of PPIs and PROTACs for specificity. Regarding inhibition of NRF2 activity in cancer, direct inhibition of NRF2 has proven to be challenging and therefore indirect approaches to target upstream signaling pathways affecting NRF2 activation or downstream processes e.g., metabolic reprogramming are more promising, especially in combination with other therapies. Finally, given the advent of immune therapies in cancer treatment, we need a better mechanistic understanding of the effect of NRF2 on immune functions in both cancer cells as well as stromal immune cells. This is important in order to fully appreciate the role of NRF2 in anti-tumor immune responses and identification of targets to overcome immunotherapy resistance.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

No data was used for the research described in the article.