Genetic and epigenetic regulation of the NRF2-KEAP1 pathway in human lung cancer

Abstract

Electrophilic and oxidative stress is caused when homeostatic mechanisms are disrupted. A major defense mechanism involves the activation of the nuclear factor erythroid 2-related factor 2 (NRF2) transcription factor encoded by the NFE2L2 gene, which can accelerate the detoxification of electrophilic carcinogens and prevent cancer and on the other hand in certain exposure contexts may exacerbate the carcinogenic process. NRF2-target genes activated under these conditions can be used as biomarkers of stress signalling, while activation of NRF2 can also reveal the epigenetic mechanisms that modulate NFE2L2 expression. Epigenetic mechanisms that regulate NFE2L2 and the gene for its adaptor protein KEAP1 include DNA methylation, histone modifications and microRNA. Understanding the activation of the NRF2-KEAP1 signalling pathway in human lung cancer, its epigenetic regulation and its role in oncogenesis is the subject of this review.

Background

Exposure to environmental toxicants, cigarette smoke and air pollution can lead to disease and the exacerbation of the chronic disease. One underlying component of disease pathogenesis is oxidative stress, which is an imbalance between the generation of reactive oxygen species (ROS) and their detoxification by antioxidant systems [1]. ROS excess can result in the damage of DNA, protein and lipids [2]. Consequently, oxidative stress contributes to ageing and age-related diseases such as cancer, cardiovascular disease, chronic inflammation and neurodegenerative diseases [3, 4]. Another underlying component of disease pathogenesis is electrophilic stress, which is the imbalance between electrophile generation and electrophile detoxication, which can lead to DNA and protein adducts [5]. Products of lipid peroxidation are themselves electrophiles, e.g. 4-hydroxy-2-nonenal, and is one example of how oxidative and electrophilic stress are linked. It is perhaps not surprising then that the protective mechanism against both stress signals is mediated via a common pathway involving NRF2-KEAP1 (Fig. 1).

Fig. 1. Regulation of NRF2 Expression Under Different Conditions.

a Regulation of NRF2 under basal conditions; b NRF2 induction by activators; c KEAP1 hypermethylation; d KEAP1 mutation. Under normal conditions, the levels of NRF2 protein are kept low by KEAP1 sequestration and by the CUL3 ubiquitin ligase, which targets NRF2 for proteasomal degradation. Under conditions of electrophilic and oxidative stress, KEAP1 no longer sequesters NRF2, which accumulates in the nucleus where it binds to sMaf, leading to the induction of genes driven by antioxidant-responsive elements (ARE). In lung cancer, hypermethylation of KEAP1 promoter and/or KEAP1 mutation prevents NRF2 from being sequestered so that NRF2 is permanently activated (switched on) leading to oncogenesis. Alternatively, hypermethylation of the NFE2L2 gene promoter can lead to the repression of NRF2 expression.

Oxidative and electrophilic stress is known to disrupt the interaction between NRF2 and KEAP1 [6]. The NRF2 transcription factor is tethered to the Kelch-like-ECH-associated protein (KEAP1), which is attached to the actin cytoskeleton [7]. ROS and electrophiles are known to react with KEAP1 cysteine residues, thereby preventing NRF2 ubiquitination and proteasomal degradation and promoting NRF2 nuclear accumulation [8]. Once NRF2 enters the nucleus, it initiates the transcription of a battery of genes encoding detoxification enzymes by binding to antioxidant response elements (AREs) in their promoter regions. Some of the biological roles of genes directly activated by NRF2 are cellular proliferation, transcription, and regulation of signalling pathways, modulation of redox homeostasis and biomarkers of tumor progression. The focus of this review is the genetic and epigenetic regulation of the NRF2-KEAP1 signalling pathway in the context of lung cancer.

NRF2-KEAP1 system and the ARE

The NRF2-KEAP1 signalling system regulates intracellular homeostasis to ROS and electrophiles. As a result, the NRF2-KEAP1 signalling system has been associated with protection against these insults. KEAP1 acts as an adaptor protein mediating interaction between NRF2 and the Cullin-3-containing ubiquitin–ligase E3 complex and mediates NRF2 degradation [9]. Consequently, the cell constantly synthesises NRF2 to maintain basal antioxidant activity [10]. NRF2 activation by prooxidants or electrophilic compounds leads to disruption of the KEAP1 ubiquitin–ligase activity (Fig. 1).

KEAP1 has five discrete protein domains as follows: the N-terminal region; the Bric-a-brac/Tramtrack/Broad (BTB) domain, which contains Cys151 [11, 12]; the intervening region (IVR) domain, which contains a group of cysteines important for stress sensing, including Cys273 and Cys288; and a double glycine repeat (DGR) and C-terminal region (CTR) domains collaborate to form a β-propeller structure, which is where KEAP1 interacts with NRF2 and makes up the Kelch domain (KLD) [13] (Fig. 2a). These reactive cysteines under conditions of oxidative and electrophilic stress can undergo oxidation or covalent adduct formation, respectively, leading to the destabilisation of the NRF2/KEAP1 complex.

Fig. 2. Regulation of KEAP1 and NRF2.

a Top: the domain architecture of the KEAP1 protein. KEAP1 consists of 624 amino acids and has five discrete protein domains as follows: the N-terminal region; the Bric-a-brac/Tramtrack/Broad (BTB) domain, which contains Cys151; the intervening region (IVR), which contains a group of cysteines important for stress sensing, including Cys273 and Cys288; and a double glycine repeat (DGR) and C-terminal region (CTR) domains, which collaborate to form a β-propeller structure, where KEAP1 interacts with NRF2 and makes up the Kelch domain (KLD). a Bottom: the methylation analysis of the CpG islands of KEAP1 in lung cancer cells. Panel b shows the domain architecture of the NRF2 protein. NRF2 consists of 605 amino acids and has evolutionarily highly conserved domains, Neh 1–7: Neh 1 is the binding site for ARE; Neh 2 serves as the binding site for KEAP1; Neh 3–5 are transactivation domains for NRF2; Neh 6 is a serine-rich domain that negatively controls the NRF2 stability by β-TrCP interactions; and Neh 7 interacts with RXRα, a nuclear receptor responsible for the suppression of NRF2/ARE signalling pathway.

NRF2 can be divided into seven distinct regions referred to as NRF2–ECH homology (Neh) domains. The individual Neh domains have different functions. For example, the N-terminal Neh 2 domain negatively controls the activity of NRF2 and is crucial for KEAP1-mediated repression of the transcription factor; the Neh 1 domain is a component of the Cap-N-Collar basic leucine zipper domain, which interacts with sMaf its binding partner. Neh 7 is the region of NRF2, which is inhibited by the DNA binding domain of retinoid X receptor α (RXRα) so the expression of NRF2-target genes may be inhibited by the interaction between RXRα and the Neh 7 domain of NRF2 [9], creating a precedent on how NRF2 may be regulated by other orphan nuclear receptors (Fig. 2b).

Upon modification of the KEAP1 Cys residues, NRF2 is activated, translocates to the nucleus and with its binding partner sMaf protein [14] forms a heterodimeric complex on the AREs or electrophilic response elements of responsive genes, where the ARE corresponds to 5′-A/G-TGA-C/G-NNN-GC-A/G-3′ [15]. This sequence distinguishes itself from the palindromic sequence identified by AP-1 transcription factors 5′-TGA-C-TCA-3′ by the presence of the GC dinucleotide at the 3′ end of the consensus sequence.

Activators of the NRF2-KEAP1 system

Activators of the NRF2/KEAP1/ARE pathway include xenobiotics, heavy metals and electrophiles, e.g., diphenols, phenylenediamines and quinones; heavy metal ions (Cd, Co, Cu, Au, Hg and Pb); Michael acceptors; isothiocyanates, thiocarbamates and related sulfur-containing compounds; hydroperoxides; carotenoids and their analogues; and selenium-containing compounds. These will all react with cysteine residues in KEAP1 and hence activate NRF2 expression and induce ARE-regulated genes [16].

Natural products

NRF2 activators can accelerate the detoxification of electrophilic carcinogens and may be useful for cancer chemoprevention [17]. These activators include natural compounds, e.g., R-sulforaphane from broccoli, curcumin from turmeric, carnosic acid from rosemary, resveratrol, quercetin, genistein and andrographolide [18]. R-sulforaphane is an isothiocyanate found in cruciferous vegetables, which is known to decrease NRF2/KEAP1 binding and facilitate NRF2 translocation into the nucleus by enhancing its expression during ageing and oxidative stress [19]. Curcumin present in turmeric is a potent activator of NRF2 with minimal toxicity; its chemoprotective properties have been demonstrated in in vitro studies to protect against As(III) toxicity in lung epithelial cells in an NRF2-dependent manner [20]. Resveratrol has been shown to act as an antioxidant. This stilbene can be activated to an electrophile, which can react with redox reactive cysteines in KEAP1 through covalent modification and disrupt the interaction between NRF2 and KEAP1, allowing NRF2 translocation to the nucleus [21] (Fig. 1).

Environmental exposures

Air pollution

Altered DNA methylation is common in lung cancer, and environmental pollution can cause changes in DNA methylation [22]. Air pollutants contain various compounds in the gaseous phase such as volatile organic compounds, aldehydes, nitrogen oxides and particulate matter (PM), which is impregnated with polycyclic aromatic hydrocarbons (PAHs), nitroarenes and trace metals. An average increase of 10 μg/m³ of PM, <2.5 μm in diameter (PM_2.5), is associated with an approximately 8% increase in the risk of mortality from lung cancer [23]. The increased risk of lung cancer directly correlates with increases in average air pollution levels, and the morbidity and mortality of air pollution-related lung cancer varies significantly across different regions depending on the level of air pollution [24]. PM_2.5 can also be impregnated with lipopolysaccharide (LPS) leading to inflammatory lung disease [25]. Many of the individual constituents found in the gaseous and particulate phases either activate NRF2 directly or through their metabolites as part of the stress response to exposure to these pollutants. The nitroarenes, which are in exhaust emissions, are metabolically activated by nitroreduction [26]. For example, 3-nitrobenzanthrone (3-NBA) is metabolically activated to yield the hydroxylamine derivative, which is further activated by sulfonation to yield DNA adducts via a nitrenium or carbenium ion [27]. The enzymes most implicated in this bioactivation in humans are NQO1 and the aldo-keto reductases (AKRs) [28]. Bioactivation of 3-NBA was eliminated by clustered regularly interspaced short palindromic repeat (CRISPR)-directed gene editing of NFE2L2 showing dependence on this pathway [29]. Thus, in this instance, an unintended consequence of NRF2 activation could be an exacerbation of carcinogenesis by nitroarenes.

Cigarette smoke

Cigarette smoke is a complex mixture containing over 3500 chemicals and includes many carcinogens, e.g. nicotine-derived nitrosamine ketone, PAHs, volatile organic compounds, acrolein, benzene, formaldehyde and heavy metals [30]. Tobacco smoke typically contains electrophilic or redox-active metals such as Cd (65 ng/cig.) and Pb (32 ng/cig.), or Cu (13 ng/cig.) and Fe (16.8 ng/cig.) [31]. Many of these toxicants or their downstream metabolites are electrophiles or produce oxidative stress and will activate NRF2.

The fine particles in cigarette smoke containing N-nitrosamines and PAHs are deposited on bronchoalveolar epithelial cells and can alter gene expression. For example, the PAHs exert direct biological effects by binding to the aryl hydrocarbon receptor (AhR) [32]. The ligand-activated receptor binds to the aromatic hydrocarbon nuclear transporter in the nucleus, where the heterodimer binds to the xenobiotic response element (XRE) on XRE-regulated genes, e.g. CYP1A1 and CYP1B1 [33–35]. Induction of these genes results in PAH being converted to electrophiles and the generation of ROS, which in turn can activate the NRF2 system [36]. For example, P4501A1 and P4501B1 catalyse the conversion of PAH procarcinogens to proximate carcinogens, e.g., PAH trans-dihydrodiols, which are converted to PAH diol-epoxides. These PAH trans-dihydrodiols are also metabolised by AKRs into reactive PAH ο-quinones, which are both electrophilic and redox-active, which can lead to the activation of the NRF2 pathway [37]. Thus, there is important crosstalk between the AhR and NRF2 signalling pathways, which could exacerbate PAH activation.

AKR genes are regulated by oxidative stress so they can be considered a signature of NRF2 activation in humans [38]. Human AKR genes, e.g. AKR1B10, AKR1C1, AKR1C2 and AKR1C3 are among the most upregulated by cigarette smoke and downregulated in ex-smokers [39]. They are part of a smoking gene battery [40]. This is seen in human bronchial epithelial cells, bronchial epithelial brushes and in buccal cells [41]. Importantly, since PAHs are activated by these AKRs to PAH o-quinones, which are both electrophilic and redox active, this represents a feed-forward pathway to exacerbate PAH carcinogenesis in human lung cells.

NRF2 activity is regulated by genomic or epigenomic alterations in tumors. A robust signature of this pathway is the expression of AKR1B10, AKR1C1, AKR1C2 and AKR1C3 genes in human lung cancer. High levels of expression of AKR1B10, AKR1C1/2 and AKR1C3 have been observed in non-small cell lung cancer cells carrying mutant KEAP1 or NFE2L2. Strong Spearman’s correlations were observed between messenger RNA (mRNA) levels of AKR genes (AKR1B10, AKR1C1, AKR1C2 and AKR1C3) with other NRF2-regulated genes, such as GPX2, GSR, PGD, G6PD, GCLM, GCLC, ME1, SRXN1 and NQO1. Moreover, AKR1B10 and AKR1C3 were expressed at basal levels in cell lines with non-mutated KEAP1/NFE2L2. Elevated AKR levels potentially constitute a biomarker signature of KEAP1 and NFE2L2 mutations compared to normal controls [39].

A novel long non-coding RNA (lncRNA) induced by cigarette smoke has been described in lung cancer cells. The expression of the smoke and lung cancer-associated lncRNA-1 (SCAL1) was regulated transcriptionally by NRF2 as determined by the small interfering RNA (siRNA) knockdown of NFE2L2 and KEAP1. Increased NRF2 binding to its putative binding site on the SCAL1 promoter was identified. The siRNA knockdown of SCAL1 showed a significant potentiation of cytotoxicity induced by cigarette smoke extract in vitro [42].

Cancer chemotherapeutics and somatic mutations in NRF2-KEAP1 signalling

Many of the genes regulated by the NRF2 system are involved in cancer chemotherapeutic drug resistance. This is not surprising since these agents can cause oxidative and electrophilic stress. AKR-mediated drug resistance is observed with the anthracyclines, e.g. daunorubicin, as well as mitomycin, cisplatin, methotrexate, oracin and vincristine [43]. One of the early studies performed by Simpkins et al., showed that ovarian cancer cells grown in the presence of cisplatin became resistant to this therapeutic. Drug resistance was accompanied by the overexpression of AKR1C1. When parental cells were stably transfected with the complementary DNA for this enzyme, the drug resistance phenotype was established [44]. These data implicate NRF2 signalling in chemotherapeutic drug resistance mediated by AKRs [45].

Increased NRF2 expression is one of the major factors of chemoresistance observed in cancer cells [46–49]. CRISPR-directed gene editing to surmount chemoresistance mediated by NRF2 in non-small cell lung carcinoma has been examined [50]. Functional knockout of the NFE2L2 gene in chemoresistant lung cancer cells significantly increased the anticancer activity of cisplatin, carboplatin and vinorelbine [51].

Somatic mutations reduce the association between KEAP1 and NRF2. These acquired mutations in KEAP1 may lead to upregulation of NRF2-regulated genes in cells carrying the mutation, which may lead to either chemoprevention or carcinogen activation e.g. PAH and nitroarene activation. Somatic mutations of KEAP1 in lung cancer patients are seen in the IVR domain, e.g. H200P and R272C, and in the DGR domain, which interacts with NRF2, e.g. A427V, G430C, R415G, G364C and G476R. Detection of KEAP1 mutations in the early stages of lung tumor development may suggest that this is an early event in lung cancer pathogenesis [52].

Somatic mutations of NRF2 in the coding region have been observed in human lung cancer. These mutations occur in amino acids in the DLG or ETGE motifs, resulting in aberrant nuclear cell accumulation of NRF2. Cells expressing NRF2 mutants display constitutive induction of cytoprotective enzymes, which are insensitive to KEAP1-mediated regulation. Somatic mutations of KEAP1 and NFE2L2 occur in almost 30% of lung cancer cases (Fig. 3). TRACERx studies investigate at what stage of tumor development these KEAP1 and NFE2L2 mutations may occur. In this regard, mutations in the candidate driver gene KEAP1 have been found in smokers with stage III adenocarcinoma and squamous cell carcinoma [53]. There are 174 KEAP1 somatic mutations in human lung cancer [54]. Most of them are located within the Kelch domain (KLD) of KEAP1. However, mutations in other regions of KEAP1 have also been found [46, 52, 55]. The mutant KEAP1 protein will exhibit reduced affinity to NRF2, and, consequently, NRF2 will be constitutively activated leading to cells resistant to various anticancer drugs.

Fig. 3. Percentage of KEAP1, NFE2L2 and CUL3 somatic mutations across the different lung cancer subtypes.

KEAP1: lung adenosquamous carcinoma (38.1%), lung adenocarcinoma (14.72%), non-small cell lung cancer (14.7%), lung squamous cell carcinoma (9.74%), and small cell lung cancer (1.57%) (top panel); NFE2L2: lung squamous cell carcinoma (15.87%), non-small cell lung cancer (5.73%), lung adenocarcinoma (2.67%), and small cell lung cancer (1.18%) (middle panel); CUL3: lung squamous cell carcinoma (5.86%), lung adenocarcinoma (2.1%), small cell lung cancer (0.79%) and non-small cell lung cancer (0.72%) (bottom panel) [97, 98].

NFE2L2 mutations are associated with specific lung cancer subtypes, e.g. more mutations are observed in lung squamous cell carcinoma (LSCC) and lung adenosquamous cell carcinoma than adenocarcinoma [56, 57] (Fig. 3). LSCC exhibits copy number alterations and mutations of NFE2L2 and KEAP1 in 34% of cases [58]. Moreover, as demonstrated by Singh’s group, loss of heterozygosity for KEAP1 is observed in 41% of the cases, and heterozygous mutations can induce NRF2 hyperactivation [46]. There is mounting evidence in the literature which supports that KRAS-mutant lung adenocarcinoma carries loss-of-function mutations in KEAP1 [59], also co-mutation signatures were found between PI3K/AKT and NRF2 pathways, e.g., NFE2L2 amplification with loss of STK11 [60].

NRF2 U-shaped dose–response curve

The shape of the dose–response curve for NRF2 activation has been proposed to be U-shaped and similar to that seen for a hormesis model observed with essential nutrients [36]. The use of a nutritional agent is beneficial if there is a deficiency, whereas an excess of supplementation is detrimental. Similarly, for the NRF2-KEAP1 pathway, the absence of NRF2 would be detrimental since there would be no protection against oxidative or electrophilic stress, optimal activation of the pathway lies between the biologically effective dose to activate the pathway to eliminate the stress and a maximal-tolerated dose as a limiting dose before reaching toxicity levels. One result of over-stimulation of the pathway involves increasing the NADPH/NADP⁺ ratio, which would be anabolic in transformed cells. The two different arms of the U-shaped dose–response curve could also be influenced by single-nucleotide polymorphisms (SNPs). SNPs in the NFE2L2 promoter could decrease the induction of this pathway and reduce the expression of detoxication enzymes [61]. By contrast, SNPs in the KEAP1 promoter could increase the silencing of KEAP1 to sustain hyperactivation of NRF2 or do the reverse. For example, somatic mutations in the Kelch region decrease its activity and induce greater nuclear accumulation of NRF2 [46, 62].

Whereas genetic disruption of the pathway through somatic mutations in KEAP1, e.g. glycine-to-serine mutations, leucine-to-arginine mutations and single-nucleotide deletions, lead to missense mutations [46]; NFE2L2 mutations clustered around codons for amino acids 24–34 and 75–82 lead to sustained activation of the pathway [63]. When the magnitude and duration of responses achieved by pharmacological intervention and genetic disruption are compared, the relative activation of the pathway by pharmacodynamic responses may be substantially smaller than that seen by genetic alteration [64].

The U-shaped dose–response curve also indicates the need for both NRF2 activators and NRF2 inhibitors. NRF2 activators can protect against electrophilic and oxidative stress where ROS can initiate tumorigenesis through the oxidation of DNA and subsequent mutation. In these instances, NRF2 activators are sought for the chemoprevention of cancer; by contrast, a prooxidant state can promote carcinogenesis. Increased NRF2 signalling can direct cancer cells toward metabolic reprogramming, i.e. increase the availability of NADPH as a cellular strategy to increase proliferation [65]. In cases like this, NRF2 inhibitors are sought to prevent proliferation. Activation of NRF2 can also lead to chemotherapeutic drug resistance and radioresistance since cytotoxic species are eliminated and, in these instances, NRF2 inhibitors are also sought. Elevated NRF2 activity in cancer cells occurs through two distinct mechanisms: diminished NRF2 turnover and augmented NRF2 mRNA levels [66].

Epigenetic regulation of NRF2-KEAP1

The major mechanisms in epigenetic regulation are DNA methylation, histone modifications and miRNA. Human genes contain clusters of CpG sites called CpG islands in their promoters; these regions typically span 200 bp, and the expression of these genes can be epigenetically modulated by DNA methylation. CpG methylation can inhibit the binding of transcription factors and the initiation of transcription. DNA methylation is catalysed by DNA methyltransferases (DNMTs) that transfer a methyl group to the 5′ position of the cytosine residue within CpG dinucleotides. The transcriptional expression of specific genes controlled by CpG methylation has been reported in cancer cells and is one of the potential modifications that occur during the development of human cancers. Deamination of 5-methylcytosine will lead to C:T transitions, which is one of the most common mutations observed in cancer.

Histone modifications are markers for both active and inactive regions of chromatin. Whereas the acetylation of histones marks transcriptionally active regions, deacetylation of histones is found in inactive regions of chromatin. These histone modifications may act sequentially or in combination to constitute a ‘histone code’ that can be read by other cellular processes. Histone readers recognise histone codes for gene transcription. The bromodomain and extra terminal (BET) proteins interact with chromatin complexes and thus, activate or inactivate gene transcription. Modification of histones by methylation is another type of chromatin modification that is involved in the regulation of gene expression. It has been demonstrated that the overexpression of H3F3A promotes lung cancer cell migration by activating metastasis-related genes. H3.3, which is encoded by H3F3A, directly activates GPR87 transcription, G-protein-coupled receptor 87 for lysophosphatidic acid, via its deposition at a specific intronic region of GPR87, where it modifies chromatin status. H3F3A and GPR87 expression levels are associated with lung cancer patients’ prognoses [67].

miRNAs are short non-coding RNAs that contain 20–22 nucleotides and represent another epigenetic mechanism that regulates gene expression at the post-transcriptional level. Epigenetic modifications, e.g., gene expression regulated by miRNAs, are the new targets for chemopreventive strategies that target NRF2 [68]. miRNAs can function as either oncogenic miRNAs or tumor suppressor miRNAs. Among others, miR15 and miR16 are examples of miRNAs with a tumor-suppressive role frequently downregulated in lung cancer cells, and miR-146 has been found to be upregulated in lung cancer cells [68].

The NRF2-KEAP1 pathway may be epigenetically regulated by these mechanisms that do not change the genetic code. Unlike somatic mutations, epigenetic gene regulation may be reversible or heritable during cell division. The epigenetic regulation of KEAP1 and NFE2L2 expression by DNA methylation, histone modification and miRNAs may provide new therapeutic targets for chemoprevention or chemoresistance. The regulation of KEAP1 and NFE2L2 expression by DNA methylation is one of the most common epigenetic modifications that has been studied [69–72].

Epigenetic regulation of KEAP1

The methylation pattern of CpG islands in the KEAP1 promoter differs between normal lung and lung cancer tissues and between human lung carcinoma cells and normal bronchial epithelial cells [73]. The treatment of lung cancer cells with the methylation inhibitor 5′-aza-2′-deoxycytidine restores KEAP1 mRNA levels suggesting an epigenetic mechanism of KEAP1 regulation [74]. Furthermore, CpG islands of the KEAP1 promoter (−291 to 337) are highly methylated in lung cancer cells, but not in normal cells, indicating that epigenetic regulation of KEAP1 might contribute to tumorigenesis (Fig. 2a) [72]. To clarify the mechanism by which KEAP1 is repressed in lung cancer cells, the role of the transcription factor stimulating protein-1 (Sp1) in the KEAP1 promoter region was studied. It was found that the transcription factor Sp1 can directly bind to the KEAP1 promoter region in normal bronchial epithelial cells, but not in A549 cells. These results demonstrated an epigenetic gene silencing mechanism that downregulates KEAP1 expression in A549 cells associated with the inhibition of Sp1 binding to the promoter by hypermethylation [75]. A549 cells also harbor a G333C mutation within the first KLD of KEAP1 [49]. Importantly, a heteroallelic mutation in KEAP1 is sufficient to increase NRF2 activity, as mutated KEAP1 functions in a dominant-negative fashion [76] (Fig. 1).

Examination of the promoter region of KEAP1 in lung cancer shows that there are differences in the contribution of the CpG sites at the 5′ end. Thus, the first seven single CpG sites (1–7, P1a region) of the KEAP1 promoter appeared to be significantly more methylated than the last six CpG sites (8–13, P1b region) of the P1 promoter region in lung cancer cells [74]. Figure 2a shows a composite of the P1a and P1b regions. As a consequence, the hypermethylation of the KEAP1 P1 promoter at the fourth and fifth CpG islands leads to a loss of SP-1 and AP-2 binding, thereby inhibiting its expression [75].

The mechanism by which miRNAs target KEAP1 to regulate tumorigenesis has not been well studied. The overexpression of miR-421 is involved in the downregulation of KEAP1 expression by directly targeting the KEAP1 3′-UTR in A549 lung cancer cells, which further promotes lung cancer cell migration and invasion [77].

Epigenetic regulation of NRF2

NRF2 has been referred to as both an oncogene and a tumor suppressor gene and its methylation status may determine its function. Whereas KEAP1 hypermethylation has been reported in lung cancer [74], the NFE2L2 promoter has exhibited increased demethylation in colorectal tumour samples, demonstrating its potential to be epigenetically regulated [78]. The regulation of NRF2 by BET proteins that can interact with acetylated lysine residues in histone and non-histone proteins will lead to activation or repression of NRF2-dependent genes. BET proteins, particularly Brd2 and Brd4, regulate expression of the NRF2-target genes, HO-1, NQO1 and GCLC, in human airway cells. Moreover, the BET protein inhibitor, JQ1, prevented H₂O₂-induced intracellular ROS production. BET proteins may play a key role in the regulation of NRF2-dependent antioxidant genes and may represent an important target for augmenting antioxidant responses in oxidative stress-mediated diseases [79].

NFE2L2 repression has been investigated using the enhancer of zeste homologue 2 (EZH2), which is a key component of the polycomb PRC2 complex and functions as a histone H3Lys27 (H3K27) trimethyltransferase. EZH2 was downregulated in human non-small cell lung cancer and inhibited lung cancer cell proliferation in vitro and in vivo. In the study, EZH2 bound to the promoter of NFE2L2, where it increased H3K27me3 and repressed NFE2L2 expression [80].

A variety of miRNAs have been implicated in the regulation of the NRF2-KEAP1 pathway, i.e. the overexpression of NRF2 can activate the expression of miR-144-3p by binding to the AREs in the miR-144-3p promoter region. Moreover, targeting miR-144-3p could mediate the downregulation of NRF2, thus decreasing the expression of target genes related to drug resistance [81]. miR-421 is overexpressed in non-small cell lung cancer, which can reduce KEAP1 expression, further promoting lung cancer cell migration and invasion, as well as inhibiting cell apoptosis [77]. The contribution of miRNA machinery, downregulation of the miR-27 family and upregulation of miR-200 family modulates NRF2/KEAP1 activity in NSCLC and SCLC tumours These results suggest that miRNA in lung cancer have a role in the regulation of the NRF2/ARE pathway [82].

NRF2-KEAP1 activators and inhibitors

NRF2 activators prevent excessive cellular damage produced by metabolic, xenobiotic and oxidative stress; thus, they may play an important role in cancer chemoprevention. NRF2 activators protect cells from apoptosis and senescence, promote cell growth and resistance to chemo- and radiotherapy [83]. Recognised NRF2 activators such as curcumin, R-sulforaphane and oltipraz have been found not to be target-specific and might increase the risk of ‘off-target’ toxicity due to their capability to interact with the cysteines of other enzymes and proteins [84]. Triterpenoid compounds are thought to be specific, e.g. triperpenoids such as 1[2-Cyano-3,12-dioxooleana-1,9(11)-dien-28-oyl]imidazole (CDDO-lm). These were developed by Sporn and colleagues [85].

On the other hand, NRF2 inhibitors should sensitise tumor cells to anticancer therapies, but the mechanism by which they block transcriptional activity and prevent its nuclear translocation may not be specific. For example, the quassinoid brusatol blocks protein translation affecting the expression of other short-lived proteins as well as NRF2 [16]. The generation of NRF2 activators and inhibitors may lack specificity and therefore interact with other downstream pathways causing undesired effects.

NRF2 inhibitors have not been established yet as targeted anticancer drugs [86], but they may be effective in the treatment of NRF2-addicted cancers where cancer cells with sustained activation of NRF2 reach an ‘NRF2 addiction’ status, in which inhibition of NRF2 activity effectively restricts their proliferation and tumorigenesis. For example, inhibition of IL-11 signalling may be an effective strategy for suppressing tumorigenesis when NRF2 is expressed in a sustained manner [87].

Inhibitors of the interaction between phosphorylated p62 and KEAP1 have potential also as therapeutic agents against human carcinomas. Accumulation of p62 inhibits KEAP1 activity by increasing NRF2 regulation. Therefore, a protein–protein interaction inhibitor that blocks the interaction between phosphorylated p62 and KEAP1 could diminish NRF2 levels [88]. Using a different strategy ML385 binds to the Neh 1 domain and prevents heterodimerisation with sMaf preventing the induction of ARE-regulated genes [89].

NRF2-KEAP1 and oncogenesis

Whether NRF2 acts as a tumor suppressor or as an oncogene is still unclear since NRF2 promotes the survival not only of normal cells but also of cancer cells [90]. Thus, NRF2 activation might not be beneficial in all cancer stages [91]. In healthy cells, activation of NRF2 promotes the transcription of anti-inflammatory and antioxidant genes that suppress the development of DNA damage and mutations that can initiate tumor formation. In contrast, NRF2 supports the survival of transformed cells by protecting against oxidative damage to support their progression to more advanced-stage tumors [92]. Also, NRF2 is involved in the metabolic activation of lung carcinogens, e.g. benzo(a)pyrene, which can increase the prooxidant state, cause tumor promotion and overwhelm the protective effect of NRF2 [36–38]. However, NRF2 hyperactivation in tumors may favour the survival of cancer cells by protecting them from excessive oxidative stress or chemotherapeutic agents. In general, patients with high NRF2 levels in their tumor tissues have a higher risk of recurrence and an overall poor prognosis [47, 93]. To this end, emerging data may define the interactions between NRF2-positive and -negative effects in cancer development.

Several mechanisms have been proposed to cause a persistent upregulation of NRF2 in tumor cells and these include: (1) somatic mutations in KEAP1, e.g. Cys151, which is the cysteine residue on the BTB-IVR KEAP1 domain that decreases the binding of CUL3 to KEAP1. Modification of any of the 27 reactive cysteines on all three domains of KEAP1 may prevent its inhibition of NRF2 signalling [46]. (2) Epigenetic silencing of KEAP1 by hypermethylation prevents its expression so that NRF2 can no longer be sequestered. Several lung tumors and lung cells have been shown to express little or no KEAP1 mRNA. The loss of expression was attributed to hypermethylation of CpG sites in the KEAP1 promoter [72]. (3) Accumulation of p21 and p62, which activate the NRF2 pathway by competing with KEAP1 for NRF2 binding and compromise the KEAP1-dependent ubiquitination of NRF2 [93, 94]. (4) Transcriptional upregulation of NRF2 by oncogenes, TP53, CDKN2A, PTEN, PIK3CA, KRAS, BRAF, and C-MYC [66].

The maintenance of proper homeostatic conditions using NRF2 inhibitors/activators could be a vital therapeutic strategy. Understanding the different stages of NRF2 expression in tumorigenesis and how its level changes may enable the use of this knowledge of NRF2 signalling for the prevention and treatment of cancer. A similar consideration must be given to modulators of KEAP1. Therapeutic approaches to inhibit KEAP1 involve modifications of cysteines (Cys151) so that NRF2 is activated, e.g. R-sulforaphane or CDDO-Im [95, 96].

Future directions

The current review examines the different genetic and epigenetic mechanisms that account for the dysregulation of NFE2L2 and KEAP1 gene expression in human lung cancer. These observations provide insights into lung cancer progression and the regulation of the transcriptional activity of NRF2. These insights identify possible therapeutic interventions for the treatment of lung cancer. Some future directions include the role of BET proteins in oxidative stress-mediated diseases; DNMT inhibitors as regulators of NFE2L2 expression; the epigenetic regulation of the NRF2-KEAP1 pathway that may be used for therapeutic benefit; and the mechanisms by which the transcription factor sMaf is regulated in lung cancer.

Author contributions

NC: manuscript preparation, editing and review; TMP: manuscript editing and review.

Data availability

Data sharing is not applicable to this article as no datasets were generated or analysed during the current study.

Competing interests

The authors declare no conflict of interest.

Ethics approval and consent to participate

This article does not contain any studies involving human participants performed by any of the authors.

Consent for publication

Both authors give consent to publish.