The NRF2, Thioredoxin, and Glutathione System in Tumorigenesis and Anticancer Therapies

1According to the World Health Organization, cancer is the second leading cause of death, accounting for 9.56 million deaths and an incidence of 18.1 million new cases in 2018, with the GLOBOCAN estimation of 29.5 million new cases and 16.4 million deaths by 2040. Consequently, cancer remains a global burden and an elusive, ever-changing disease with an extremely complex biology. The high inter- and intratumor heterogeneity is governed by cancer’s need to grow and spread in the organism while surviving therapy-induced unfavorable conditions. Genetic alterations inducing survival pathways combined with metabolic reprogramming ensure the action of diverse players contributing to tumor development and progression.

2Reactive oxygen species (ROS) are recognized as essential players contributing to tumorigenesis or suppressing it, with an important role in anticancer therapy as well. Although previously perceived as exogenously and endogenously derived byproducts of cellular metabolism with signaling abilities that affect cellular functions in a concentration-dependent manner, the perception of ROS has been upgraded since. Nowadays, they are considered as cellular tools that are, upon certain stimuli, purposely produced by the cell in an amount that will elicit a specific feedback reaction within the cell or in neighboring ones. The tuning of ROS levels is intertwined with metabolism and antioxidative machinery, as in normal cells and cancer cells, too. However, cancer cells often have more ROS than normal ones, a feature that is exploited by most conventional chemo- and radiotherapy.

3The nuclear factor, erythroid 2 like 2 (NRF2) pathway is the main pathway activated upon ROS production. Its activation induces the expression of over 250 genes, regulating diverse processes from redox homeostasis to detoxification, carbohydrate and lipid metabolism, autophagy, apoptosis, DNA repair, etc.. The antioxidant mechanisms involved in ROS balancing, such as thioredoxin (TRX) and glutathione (GSH) systems, not only protect from carcinogenesis but can support it as well. A recent clinical study on colorectal cancer patients demonstrated that redox biomarkers could have a role in the infiltration of inflammatory cells and tumor budding. The same study also demonstrated that the levels of oxidative stress parameters vary based on the stage of cancer. In addition, GSH and TRX are both depicted as important players in carcinogenesis. While GSH is required for cancer initiation, TRX is a predominant mechanism governing cancer progression by abolishing detrimental ROS levels in already established neoplasm. Noteworthy, the NRF2 pathway is often activated in various types of cancer, emphasizing its dual role in tumorigenesis.

4Only the understanding of all intertwined cancer-related components, particularly the contributors and/or suppressors of cancer growth, will bring us closer to better-targeted therapy. This review will explore the importance of the NRF2 pathway and TRX and GSH systems in carcinogenesis in relation to anticancer treatment strategies.

5ROS are continuously formed in aerobic organisms as byproducts of normal intra/intercellular metabolism or in response to adequate stimuli, as is the case of inflammation. Depending on the ROS type, reactivity, and their diffusion distance, their effect on macromolecules varies. Thus, superoxide anion (O2●−) ROS are generated by the action of complex I and complex III via an electron transport chain in mitochondria, while membrane-bound nicotinamide adenine dinucleotide phosphate (NADPH) oxidase and lipoxygenases also contribute to O2●−. Additionally, huge amounts of ROS, in particular O2●−, are formed during phagocytosis. The dismutation of O2●− occurs spontaneously or enzymatically, with superoxide dismutase (SOD) generating hydrogen peroxide (H2O2), which is a more stable non-radical form of ROS with a greater diffusion distance that may therefore affect targets distant from the initial oxidative insult. At low concentrations, ROS have important signaling roles both intra- and intercellularly; however, at higher concentrations they have toxic effects. Likewise, H2O2 is an important metabolic signal for insulin secretion in β-cells, however excessive H2O2 affects mitochondria and causes β-cell dysfunction. Similarly, ROS generated during the respiratory burst of granulocytes can have dual roles in carcinogenesis. Although an important anticancer effect of granulocytes was demonstrated in vitro, the in vivo results can be contradictory. In experimental animal tumor models, it was shown that ROS can reverse the process of tumor development, leading to spontaneous tumor regression, and can decelerate tumor growth, extending the survival of animals. Opposite to that, it is known that ROS can also promote tumor development, so one also needs to consider the importance of intercellular redox signaling in tumor development. Hence, it was recently shown that also hypochlorous acid intercellular redox signaling is involved in granulocyte-mediated tumor inhibition.

6One of the key mechanisms by which ROS achieve their biological effects is the well-documented induction of either reversible or irreversible modifications of proteins, altering their structure and cellular functions. For example, the oxidative modification of erythrocyte spectrin alpha chain, band 3 protein, and glyceraldehyde-3-phosphate dehydrogenase can affect the erythrocyte structure and energy metabolism. Similarly, the oxidation of mitochondrial adenosine triphosphate synthase beta subunit can affect the enzyme catalytic activities and potentially have a detrimental role for neurons.

7Both O2●− and H2O2 can serve as precursors for highly reactive hydroxyl radicals (●OH) that can induce direct oxidative damage to macromolecules. Transition metals, such as iron, can promote ROS and support tumorigenesis. O2●− and H2O2 are accessible to iron via the Fenton reaction, converting them to ●OH, and can contribute to the initiation of ferroptosis, which is also considered in anticancer approaches as a potential target for oncotherapy.

8On the other hand, ROS can also damage DNA, causing strand breaks or oxidative DNA lesions. As an example, the ●OH-induced nucleic acid oxidative damage can yield a variety of base lesions, among which 8-oxo-2′deoxyguanosine (8-OHdG) and 8-oxo-guanosine have received the most attention. The accumulation of unrepaired or defectively repaired DNA damage could eventually promote various disorders, among which cancer is the prominent disease.

9Moreover, the bis-allylic site of polyunsaturated fatty acids is highly susceptible to ROS induced damage. The peroxidation of lipids is profoundly triggered by ●OH and hydroperoxyl radical, yielding as final products reactive aldehydes, among which is the most potent “second messenger of ROS”, 4-hydroxynonenal (4-HNE). 4-HNE was first discovered fifty years ago and was thought to be a toxic product of lipid peroxidation. Since then, a huge amount of effort has been put into investigating the roles of 4-HNE, and today we know that, depending on its location and concentration, it can act as a double-edged sword, exhibiting pathological or desirable physiological effects. The high reactivity of 4-HNE is due to its three functional groups: hydroxyl group, carbonyl group, and C=C double bond. 4-HNE can act either directly or by forming adducts with macromolecules, such as proteins or nucleic acids. Consequently, 4-HNE can modulate various cellular functions of non-malignant as well as cancer cells and can play a role in tumorigenesis.

10Excessive ROS generation alters the cellular redox homeostasis built by antioxidant defense systems to cope with the negative effects of ROS. On the other side, endogenous antioxidant defense systems are affected by the supplementation of exogenous antioxidants, which is common in both healthy and ill people.

11The major endogenous antioxidants responsible for cellular ROS detoxification are the TRX system, the GSH system, peroxiredoxin, SOD, and catalase (CAT). The removal of O2●− by SOD elevates cellular H2O2 that can act as a molecular mediator of various signaling pathways, such as insulin, NRF2, and the c-Jun N-terminal kinase (JNK) signaling pathway. Whether H2O2 will promote cell survival or induce apoptosis depends on its concentration. Thus, cells have several mechanisms for H2O2 detoxification. Catalase, peroxiredoxin, and the TRX and GSH systems all promote the conversion of H2O2 to H2O and molecular oxygen. Although these might be considered as major cellular antioxidant defenses, one should bear in mind that other antioxidants—for example, protein deglycase DJ-1—can also determine the fate of the tumor, interact with NRF2, regulate GSH levels, and promote TRX (reviewed in).

12Therefore, alterations in redox homeostasis are of high importance for cancer development. Although ROS have an important role in the initiation and promotion of tumor development, they are also crucial factors that can inhibit tumor viability and growth, so tumor cells modify their own antioxidant network to enable them to escape the anticancer effects of ROS. Among endogenous antioxidants in the current review, special attention is given to the NRF2, TRX, and GSH antioxidant defense mechanisms in tumorigenesis.

13The NRF2 pathway is the main pathway activated as a response to oxidative stress. In homeostasis, transcription factor NRF2, encoded by the NFE2L2 gene (nuclear factor, erythroid 2 like 2) is mainly under the control of its repressor, Kelch-like ECH-associated protein 1 (KEAP1), and subjected to ubiquitination and proteasomal degradation. Additionally, NRF2 can also be repressed by β-transducin repeat-containing protein (β-TRCP) in a glycogen synthase kinase-3 (GSK-3) -dependent and -independent manner or in an HRD1 (E3 ubiquitin-protein ligase synoviolin, SYVN1) -dependent manner upon endoplasmic reticulum (ER) stress (Figure 1). Aside from the mentioned mechanisms of NRF2 regulation, numerous proteins and molecules interact with either KEAP1 or NRF2 or even compete with NRF2, such as BTB and CNC homolog 1 (BACH1), contributing to the highly complex regulation/activation of NRF2, the feature particularly accentuated in cancer (reviewed in).

14Current research recognizes the dual role of NRF2 in tumorigenesis. NRF2 was first perceived as a tumor suppressor due to its role in reducing damaging ROS and environmental carcinogens, thus protecting cells from potential neoplastic transformation. Mice Nrf2-knockout models have pointed out the importance of the NRF2 pathway activation in the suppression of carcinogen-induced cancer initiation (e.g., in the skin, breast, and forestomach). In humans, an NRF2 single nucleotide polymorphism (SNP) homozygous allele (-617A7A) exhibits a decreased expression of NRF2 and its target cytoprotective products, and is consequently associated with an increased risk of lung cancer in smokers. In addition, the genetic variants in NFE2L2, NAD(P)H quinone dehydrogenase 1 (NQO1), nitric oxide synthase 3 (NOS3), and heme oxygenase 1 (HMOX-1), exhibit lower ROS detoxification capabilities—although they are not associated with a breast cancer risk individually—in postmenopausal women with three or more of such modifications, increasing the risk of breast cancer, especially when combined with a high iron intake.

15Conversely, the discovery of the hyperactivation of NRF2 in quite a number of tumors, conferring an advantage to tumor cells and resulting in growth promotion and therapy resistance, revealed its new role as an oncogene. In this context, the constitutive activation of NRF2 contributes not only to the progression and chemoresistance in the already-established tumor cells but also to the tumor development itself. Indeed, the constitutive activation of NRF2, associated with an increase in cellular antioxidant enzymes and a diminution in the ROS levels, contributes to arsenite-induced human bronchial epithelial cell transformation.

16Genetic alterations (e.g., NFE2L2 gain-of-function mutations, KEAP1 loss-of-function mutations, and co-occurring driver mutations), interactions with different proteins interfering with the NRF2-KEAP1 bond, and epigenetic and posttranslational modifications can all induce NRF2 activation. Emerging data imply the varying involvement of the NRF2 pathway among tumors (some are listed in Table 1). Cancers with NRF2 hyperactivating mutations share a set of upregulated target genes, including key regulators involved in the thioredoxin and glutathione system, metabolic enzymes, transporters, and others, enabling a cancer growth advantage and consequently a worse overall survival. The upregulation of these genes by NRF2 highly depends upon a more assessable chromatin environment of antioxidant responsive elements (AREs).

17In lung cancer, the co-occurring mutations in KEAP1, NFE2L2, or cullin 3 (CUL3) observed in 7% of non-small cell lung cancer (NSCLC) patients with EGFR (Epidermal Growth Factor Receptor) mutations indicate the activation of the NRF2 pathway as an acquired mechanism, leading to the resistance to usually beneficial EGFR-tyrosine kinase inhibitor therapy. Another protective mechanism observed in lung cancers is metabolic reprogramming to glutaminolysis, mostly observed in KRAS (Kirsten rat sarcoma viral oncogene homolog)-mutant lung adenocarcinoma, which exerts a loss of STK11/LKB1 (serine/threonine kinase 11, liver kinase B1) and co-occurring KEAP1 mutation. In addition, the NRF2 molecular signature, including genes important in activating TRX and GSH systems such as glutathione peroxidase (GPX) 3 and thioredoxin-like 1, is suggested as an excellent predictor of cancer remission and overall survival in patients with lung cancer. The growth advantage and resistance to therapy in NSCLC patients with a worse prognosis are associated with the activation of the NRF2 pathway, contributing to a higher expression of multidrug-resistant protein-3.

18The other mechanisms of NRF2 activation include diverse proteins that compete with NRF2 in binding with KEAP1. Examples of these are proteins with an ETGE motif, such as dipeptidyl peptidase 3 protein, cell cycle-related kinase (CDK20), as well as others. The overexpression of CDK20 in lung cancer leads to tumor progression and resistance to radio and chemotherapies. The p62 is another NRF2-binding competitor of KEAP1. Its upregulation was suggested as a protection mechanism of hepatocellular carcinoma (HCC)-initiating cells in the oxidative stress hostile environment that promotes HCC carcinogenesis, as well as anticancer drug tolerance in tumor regions positive for the hepatitis C virus. The alkylation of KEAP1 by succinylacetone or the hypermethylation of the KEAP1 promoter region may lead to NRF2 activation. For instance, the tumor-specific hypermethylation of the KEAP1 promoter region was suggested as a specific feature of a clear cell renal carcinoma.

19Crosstalk between the NRF2 pathway and other proteins can contribute to carcinogenesis and therapy resistance. Hence, it was shown that the progression of HCC requires metabolic changes involving TRAP1 and NRF2 as an early event, leading to an expression pattern of glucose-6-phosphate dehydrogenase that correlates with grading, metastasis, and poor prognosis. Another example of this is the overexpression of TRIM25 (Tripartite motif-containing protein 25) and NRF2, associated with the protection of HCC cells upon ER-induced ROS. The association of BACH1 and NRF2 was shown to promote lung cancer metastasis. Mechanistically, the mutations in KEAP1 found in lung adenocarcinoma patients lead to the activation of the NRF2 pathway and an increase in HMOX-1. Since the degradation of BACH1 by F-box only protein 22 (FBOXO22) necessities heme, the NRF2-induced HMOX-1 reduces heme levels, thus stabilizing BACH1 and promoting the transcription of the pro-metastatic genes. The antioxidant supplementation of N-acetyl cysteine and vitamin E mimics the NRF2-HMOX-1 action, leading to the BACH1 stabilization and glycolysis induction in KRAS-mutant lung cancer. Therefore, lung adenocarcinoma patients with a high BACH1 signature are associated with increased metastasis and worsened survival. Conversely, FBXO22 promotes carcinogenesis in colorectal cancer and HCC by degrading PTEN (phosphatidylinositol 3,4,5-trisphosphate 3-phosphatase and dual-specificity protein phosphatase) and p21, respectively, while revealing both roles (protumorigenic and antimetastatic) in breast cancer.

20The thioredoxin system, one of the key regulators of cellular redox homeostasis, comprises TRX, thioredoxin reductase (TRXR), nicotinamide adenine dinucleotide phosphate (NADPH), and thioredoxin interacting protein (TXNIP) (Figure 2). The TRX system requires NADPH to reduce the oxidized form of TRX. Reduced TRX is needed for the recycling of oxidized thioredoxin peroxidase/peroxiredoxin. The family of thioredoxin-domain containing proteins (TXNC) are also redox regulators, and today 17 members of the TXNDC family are known. When an oxidized substrate is reduced by the action of Cys32 and Cys35 of the reduced dithiol form of TRX, oxidized disulfide TRX is generated. Oxidized TRX is converted back to an active reduced TRX by TRXR at the expense of NADPH. Thus, an adequate supply of NADPH and TRXR activity is crucial for this process.

21In the last few decades, numerous studies have investigated the role of the TRX system and some members of the TXNDC protein family in tumorigenesis (Table 2). The overexpression and hyperactivation of cytoplasmic TRXR (TRXR1) have been reported for various cancer types, such as brain cancer, breast cancer, HCC, lung cancer, oral, and tongue squamous cell carcinoma (Table 2). Mitochondrial TRXR (TRXR2) was also found to be upregulated in tumor tissue. Moreover, tumor cells overexpress TRX to cope with excessive ROS, and its expression is closely related to the pathological grade of the tumor. Additionally, the downregulation of TXNIP, a TRX inhibitor, will result in a decrease in TXNIP-TRX complexes, protecting cells from the effects of excessive ROS and resistance to therapy. Indeed, it is common for various types of cancers to downregulate TXNIP. The TXNIP was suggested to be used as a prognostic marker as its expression inversely correlates with the pathological grade of tumor, while its overexpression can indicate a longer survival of cancer patients. Among the TXNDC protein family, TXNDC5 has been studied the most and was found to be overexpressed in tumor tissues, and its high expression correlates with poor survival. Other TXNDCs reported to be altered in tumors are TXNDC9 and TXNDC17.

22In addition to the important role of the TRX system and TXNDCs as guardians of cellular redox homeostasis, they also modulate various cellular pathways that might affect tumor development (Table 3), and some of the mechanisms are described below. TRX, mainly located in the cytosol and frequently referred to as cytoplasmic TRX (TRX1), interacts with different proteins in order to maintain cellular redox homeostasis, while in the state of oxidative stress it translocates to the nucleus modulating the activity of transcription factors. The TRX1 induces nitric oxide synthase type 3 and the S-nitrosylation of death receptor CD95, modulating apoptosis. Nitric oxide synthase 2 overexpression induces the S-nitrosylation of mitochondrial TRX (TRX2) and caspase 3, altering their activity and promoting tumor growth. TRX1 induces S100P, leading to an additional increase in TRX1 through a positive feedback mechanism via the upregulation of phosphorylated ERK1/2 and the downregulation of TXNIP. TRX1 overexpression decreases tumor suppressor PTEN and causes an increase in phosphorylated AKT (protein kinase B) that can consequently induce S100A4 and promote epithelial to mesenchymal transition (EMT), migration, and invasion of tumor cells. Recently identified mitochondrial TRXR isoform (TRXR3) was found to reduce TRX2 and promote tumor cell survival. On the contrary, the inhibition of TRXR alters the mitochondrial membrane, reduces tumor growth, and induces apoptosis. Several microRNAs (miR) are known to inhibit TRXR, such as miR-125b-5p, miR-124, and miR-17-3p. Similarly to TRX, TRXR can be inactivated by nitrosylation, while its acetylation increases catalytic activity.

23Furthermore, metabolic and oxidative stress as well as hypoxia or hyperglycemia can promote TXNIP expression, which is also considered as a tumor suppressor. The overexpression of TXNIP induces mitochondrial ROS generation, activates MAPK, and promotes apoptosis and cell cycle arrest. Thus, the downregulation of TXNIP is frequent in tumor cells, and some of the mechanisms by which that is accomplished include targeting the TXNIP N-terminus, affecting the TXNIP promoter, or binding to the 3′-untranslated region (3′-UTR) of TXNIP. The inhibition of TXNIP promotes tumor cell proliferation, EMT, and metastasis. However, TXNIP can be upregulated by inhibiting histone deacetylases, bromodomain and extra-terminal domain, phosphatidylinositol-3-kinase (PI3K)/AKT pathway, the downregulation of HER1/2, or focal adhesion kinase. The poor survival of triple-negative breast cancer patients correlates with elevated c-MYC and decreased expression of TXNIP, which is probably due to c-MYC binding to the TXNIP promoter. The secretome of monocyte-derived foam cells contains 4-HNE and was shown to increase the TXNIP expression of endothelial cells, which could be attributed to 4-HNE’s ability to inhibit the expression of c-MYC.

24Among the TXNDCs, TXNDC5 can have a role in both tumor progression and tumor suppression. The inactivation or downregulation of nuclear receptor 4A1 (NR4A1) downregulates TXNDC5, isocitrate dehydrogenase 1, and the mTOR (mammalian target of rapamycin) pathway, further promoting ROS generation, inducing apoptosis, and inhibiting tumor growth. The downregulation of TXNDC5 was also reported to inhibit angiogenesis. On the other hand, hypoxia-induced TXNDC5 via hypoxia inducible factor 1α (HIF1α) promotes tumorigenesis.

25The glutathione system, another key regulator of cellular redox homeostasis, comprises glutamate-cysteine ligase (GCL), glutathione synthetase (GSS), reduced glutathione (GSH), oxidized glutathione (GSSG, glutathione disulfide), GPX, glutathione reductase (GR), NADPH, and glutathione S-transferase (GST) (Figure 2). GSH is ubiquitously distributed within the cell and the availability and level of GSH depend on its synthesis by GCL and GSS, the recycling of GSSG by GR and NADPH, and its detoxification activity via GST-mediated conjugation to molecules. GST belongs to phase II metabolism, and today seven classes of GST are known. The detoxification role of GSH and GST is crucial to enable cells to cope with various stressors. However, alterations in GSH systems can promote tumorigenesis (Table 4). Decreased blood GSH was seen in cancer patients. Interestingly, the GSH level was found to be increased in head and neck carcinoma while the opposite was reported for colorectal cancer. Moreover, cancer patients frequently have a decreased blood GSH and GPX activity. Remarkably, the expression of GPX in tumor tissue depends on the GPX isoform. In that regard, GPX1, GPX3, and GPX7 expression were found to be decreased in tumor tissues, while the expression of GPX2 and GPX4 was found to be upregulated. In addition, GSTA1, GSTM1, and GSTZ1 are reported to be downregulated in tumor tissue and can correlate with a poor prognosis, while GSTT1, GSTO1, and GSTK1 are mostly reported to be upregulated in tumor tissue compared to the normal surrounding tissue. The expression level of GSTP1 in tumor tissue is controversial, and its involvement in tumorigenesis could at least in part depend on its hypermethylation.

26Tumor cells tend to modulate the GSH system in order to survive (Table 5), and some mechanisms are described below. The upregulation of the NRF2 pathway promotes GCL, yielding elevated GSH and promoting tumorigenesis. In the case that GSH is depleted, tumor cells overcome protein homeostasis by deubiquitinases. Thus, the combined inhibition of deubiquitinases and GSH generation would be needed for malignant destruction. The methylation of GPX1 promotor downregulates GPX1, which can induce the activation of the AKT/GSK-3β/SNAIL pathway, promoting EMT. GPX2 overexpression is also implicated in EMT. However, miR-17-3p can inhibit GPX2, altering mitochondrial respiration and consequently rendering tumor cells susceptible to anticancer therapy. Several other miRs also modulate the GSH system. Thus, miR-196a targets GPX3, affecting GPX3-mediated cell death and promoting tumorigenesis. Additionally, miR-133b downregulates GSTP1 by targeting GSTP1 3′-UTR. GSTP1 is also downregulated by the methylation of the CpG island, while the recruitment of early B cell factor 1 to its promoter upregulates GSTP1. The overexpression of GSTP1 can induce cell cycle arrest, as well as GPX4 deficiency, and can thus be used as potential targets in anticancer therapy. GSTPs catalytic activity varies depending on the target. For example, GSTA4-4 has the highest activity towards 4-HNE and also the lowest rate with respect to 4-HNE adduction compared to others. The conjugation of 4-HNE with GSH is the major route for 4-HNE elimination from cells.

27The perception that antioxidants can protect cells from detrimental levels of ROS has led to several large-scale studies with somewhat disappointing results. Data suggesting the beneficial effects of antioxidant supplementation are limited, while more show no effects or even imply that antioxidants can increase cancer risk. Among these, a Finnish trial on the effect of vitamin E and beta carotene on the incidence of lung cancer and other cancers in male smokers raised a lot of skepticism against the possible beneficial effects of the fat-soluble vitamins, stressing mostly their negative effects, especially an increase in the lung cancer incidence. However, the authors of the trial have overseen the fact that their treatment resulted in a several-fold overload of the supplemented antioxidants in plasma, indicating their even higher potential overload in the lungs. In the case of persistent oxidative stress, as is the one occurring in heavy smokers, these antioxidants, especially beta-carotene, could be oxidized into multiple toxic, mutagenic, and likely carcinogenic products, although at the same time beta-carotene could further maintain its antioxidant capacity, in particular for the lung cells. Similarly, the Alpha-tocopherol, beta-carotene cancer prevention (ATBC) study in 90′ linked the supplementation of beta-carotene with the increased risk of lung cancer among smokers, while the more recent findings attenuated so negative results showing that supplementary alfa-tocopherol and beta-carotene have no late effects on cancer incidence. Thus, after two decades, the authors of the ATBC trial revealed the preventive effect of moderate-dose alpha-tocopherol on prostate cancer that continued for several years after the trial, reducing prostate cancer mortality. However, the issue of the use of antioxidants as adjuvant therapy is still under debate favoring the laconic “no use” approach, while the increased recurrence and mortality of breast cancer in patients who concurrently took dietary supplements with radio- and chemotherapy is alarming.

28While considering the possible benefits or risks of using antioxidants in oncology, one should recall the pathophysiology of oxidative stress, which is aware of beneficial roles played by endogenous antioxidants, especially in the case of acute stress response (such as those induced by irradiation, toxic compounds, or ischemia/reperfusion), for which optimal interplay between water-resistant and water-soluble antioxidants is essential. If that is not the case, many antioxidants (in particular, those water-resistant—i.e., lipid-soluble—present in biomembranes) could become free radicals, even more harmful than the initial ROS that started oxidative stress. That is even more relevant for chronic oxidative stress, as occurs in cancer and inflammation, while the long-lasting or excessive use of exogenous antioxidants even bears a risk of uncontrolled interference with endogenous antioxidants, which is the main reason why antioxidants cannot be considered as panacea. Similarly, even in the case of mild and acute oxidative stress (like in exercise) the use “preventive” exogeneous antioxidants can block desirable hormetic feedback effects of exercise, that can cause a rise in the endogenous antioxidant capacities, mostly affecting NRF2, which may eventually increase the resistance of oxidative-stress associated disorders.

29The NRF2 pathway is the main cellular defense mechanism activated upon exposure to oxidative stress, electrophilic stress, and xenobiotics. It regulates the expression of a great variety of cytoprotective genes, enabling cells to withstand unfavorable conditions and restore the homeostatic state. In an effort to reduce cancer incidence, past research has been focused on the activation of what was considered as a cancer-protective mechanism, the NRF2 pathway. Thus, the extensive search for activators of the NRF2 pathways begun. Many candidates emerged, some reaching clinical trials. Sulforaphane is one such candidate abundantly present in broccoli sprouts. Clinical trials with broccoli sprout extracts revealed the attenuation of cancer risk in individuals exposed to aflatoxins and air-borne toxins and insufficient, although safe, anticancer activity in prostate cancer patients, while others are still ongoing (extensively reviewed in). Curcumin is another NRF2 activator with pleiotropic activity. To date, clinical trials investigating the effectiveness of curcumin treatment in diverse cancers revealed the omission of the expected effect. Yet, curcumin was found to be well-tolerable and safe. Other NRF2 activators enrolled in clinical trials include resveratrol, bardoxolone-methyl (CDDO-Me), oltipraz, RTA-408 (omaveloxolone), etc. (reviewed in).

30While the translation of in vitro and in vivo observed NRF2 activator-induced benefits to the clinic are still scarce, caution concerning their usage arises. The negative effect of NRF2 activation was observed in diabetic patients. Standard diabetic drugs, saxagliptin and sitagliptin, were shown to increase the risk of a metastatic spread in patients who already have cancer, although they did not enhance cancer incidence. The underlying mechanism included prolonged NRF2 activation. Therefore, the use of NRF2 activators should be cautiously evaluated, particularly in cancer patients, considering the hyperactivation of the NRF2 pathway observed in the vast majority of cancers.

31The constitutive activation of NRF2, observed in many cancers, causes research to rethink the new possibilities for treating cancer that can amend acquired resistance to conventional therapy. Inhibitors of the NRF2 activation, and its target products such as TRX and GSH, have become a focus of this research. All-trans retinoic acid (ATRA), clobetasol propionate (CP), and apigenin are some of the examples of the NRF2 inhibitors under clinical investigation, while others show promise in cellular and animal models, including ARE expression modulator 1 (AEM1), ML385, 1-(2-cyclohexylethoxy)aniline (IM3829), malabaricone-A (MAL-A), etc.. ATRA forms a complex with retinoic acid receptors (RARs) that bind with NRF2, thus interfering with the binding of NRF2 to ARE sequences and blocking the activation of the NRF2 pathway. Although already in use for treating acute promyelocytic leukemia patients, its applicability in the treatment of solid tumors is still ambiguous and heavily ongoing (reviewed in). CP is another NRF2 inhibitor currently evaluated in phase 2 clinical trials (NCT02368886) in patients with refractory metastatic colorectal cancer. CP prevents nuclear accumulation and promotes the β-TRCP-dependent degradation of NRF2 in a glucocorticoid receptor- and a GSK-3-dependent manner. While apigenin was shown to inhibit NRF2 activation, further research revealed its multiple modes of action. Apigenin did reach the clinical trial (NCT00609310) investigating the prevention of neoplasia recurrence, yet the study has been suspended. Auranofin, an FDA-approved drug for the treatment of rheumatoid arthritis, and buthionine sulfoximine (BSO), an inhibitor of GSH synthesis, have been in consideration as anticancer agents as well. While BSO failed in the clinic, auranofin is still under investigation in several clinical trials. Its mode of action includes the inhibition of the activity of TRXR and thus the disruption of the TRX system.

32The recent opinion suggests combinational therapy to be more adequate in cancer treatment because anticancer agents usually affect different pathways, not offering cancer cells to adapt so quickly as in monotherapy and is thus reducing the occurrence of chemoresistance. In addition, the mutual effect lowers the effective dose, which in turn attenuates the unwanted side effects of some drugs. PI3K/AKT inhibitors, such as MK2206 (Merck), have shown some promise in clinical trials. Yet, in vitro and in vivo studies have revealed the non-responsiveness of such monotherapy in NSCLC, suggesting combinational therapy as an advantageous strategy. Thus, a synthetic lethality induced by MK2206 and TXNRD1 inhibitor auranofin, found to be dependent on the genetic status of KEAP1, shows promise in KEAP1 wild-type over mutant.

33A more personalized approach, such as the analysis of patients’ mutational status and putting them in context with the known mechanisms of the anticancer treatments, could improve patients’ outcomes. For example, combining clinical data and in vivo and in vitro approaches revealed the importance of KEAP1 and NFE2L2 mutations in lung adenocarcinoma, correlating them with advanced stages and worse survival. Associating these mutations with KRAS mutations revealed the high dependence on glutaminolysis of KRAS/KEAP1 or KRAS/NFE2L2 mutants, a potentially exploitable feature in future therapy. In addition, KRAS/KEAP1 mutants were found to arise from a bronchiolar cell of origin and keep the pentose phosphate pathway active, another possible exploitable feature in therapy. Moreover, in lung adenocarcinoma tumors co-occurring KEAP1 mutations and STK11/LKB1 loss lead to metabolic reprogramming (glutamine metabolism), activating the pentose phosphate pathway and the tricarboxylic acid (TCA) cycle to maintain redox balance, suggesting a glutaminase inhibitor as a possible treatment strategy. Xu et al. investigated the relationship between NFE2L2/KEAP1 mutations, tumor mutational burden (TMB), and programmed death ligand 1 (PD-L1) expression. They found that NFE2L2/KEAP1 mutations are present in various cancers, with the highest incidence found in lung squamous cell carcinoma. These mutations were linked with higher TMB and PD-L1 expression. Since the NFE2L2/KEAP1 mutations in cancer are often associated with poorer overall survival, a survival analysis of NSCLC patients receiving immunotherapy revealed improved clinical outcomes in comparison to other treatments, suggesting its possible beneficial use for patients with mutations in NFE2L2/KEAP1. In addition, NFE2L2/KEAP1 mutations cause the constitutive activation of the NRF2 pathway and enhanced ARE activity, a feature suggested as being exploitable for a cancer suicide gene therapy. Leinonen et al. used a lentiviral vector expressing herpes simplex virus thymidine kinase (HSV-TK/GCV) under the regulation of ARE. They evaluated this approach in human lung adenocarcinoma cells. They showed this approach to be effective in both in vitro and in vivo and suggested it as a promising treatment in conjunction with traditional therapies. Table 6 contains a summary of antioxidant defense system modulators and strategies in cancer incidence and therapy.

34Cancer biology is a very complex process that includes the multifaceted interplay between antioxidant systems and ROS in determining cancer development, progression, metastasis, and regression. Such complexity is sometimes overseen while evaluating the effects of potential anticancer antioxidants, as in the ATBC trial in the 1990s. In the modern era, realizing that a more personalized and integrative biomedical approach could give more benefits for the prevention and therapy of cancer, we should focus on altered oxidative homeostasis in cancer cells.

35Therefore, the relevance of the NRF2 pathway and TRX and GSH systems in carcinogenesis and in anticancer therapies has been extensively investigated. Targeting only one system can be beneficial, while the combined modulation of multiple antioxidant systems can give better anticancer results. Besides various synthetic agents, miRs should also be considered in anticancer therapies, as they have been shown to have an important role in the modulation of the NRF2 pathway and TRX and GSH antioxidant system efficiencies.