(1) The Molecular Mechanisms of NRF2 Regulation by Keap1

The discovery of Keap1 as a negative regulator for NRF2 in 1999 marked the beginning of the NRF2 field. Since embarking on NRF2-related investigations in 2001, I have been privileged to be part of the research unraveling one of natures’ most remarkable defense systems. My group has played a central role in defining this pathway and has elucidated several of the critical steps in controlling homeostatic and induced expression of NRF2. Our pioneering work of NRF2 regulation is summarized in Fig 1. Under homeostatic conditions, NRF2 is bound by Keap1, which acts as a substrate adaptor for the Cullin3 E3 ubiquitin ligase complex. While complexed, NRF2 is continuously ubiquitylated and subsequently degraded by the proteasome. In response to oxidative stress or electrophilic xenobiotics, cysteine residues of Keap1, especially C151, C277 and C288, function as sensors to switch on/off the NRF2 pathway through inhibition of NRF2 ubiquitylation, allowing NRF2 to accumulate, and then to translocate into the nucleus. In the nucleus, NRF2 binds small Maf proteins and turns on antioxidant response element (ARE)-bearing genes. After induction, Keap1 functions as a key postinduction repressor of NRF2. Upon recovery of cellular redox homeostasis, Keap1 translocates into the nucleus to dissociate NRF2 from the ARE. The NRF2-Keap1 complex is then transported out of the nucleus by the Keap1 nuclear export signal. Once in the cytoplasm, the Keap1-NRF2 complex associates with the Cullin3-E3 ubiquitin ligase, resulting in degradation of NRF2 and termination of the NRF2 signaling pathway. These findings set the foundation for the NRF2-Keap1-ARE signaling transduction field.

(2) Physiological (bright side) and pathological roles (dark side) of the NRF2-Keap1-ARE pathway; NRF2 modulators for disease prevention (NRF2 inducers) and treatment (NRF2 inhibitors)

Since the elucidation of NRF2 regulation by Keap1 under physiological conditions, we have discovered many chemopreventive small-molecules that activate this NRF2-mediated adaptive response and demonstrated these NRF2 inducers are effective for disease prevention (cancer and diabetic complications) (Fig. 2A). In 2008, however, we presented a paradigm-shifting concept, the “dark side” of NRF2, demonstrating that uncontrolled NRF2 expression promotes cancer growth, contributes to therapeutic resistance, and poor prognosis. Many genetic (somatic mutations in NRF2, Keap1, and other proteins) or epigenetic alterations that compromise the Keap1-mediated ubiquitylation of NRF2 have been identified in various types of cancer, resulting in constitutively high levels of NRF2 (Fig. 2B). Building on this theme, my team has identified the first compound, brusatol, which inhibits the NRF2 pathway. As predicted by our model, brusatol overcame both intrinsic and acquired chemoresistance. Hence, our programs challenged the classical view of NRF2 by playing a lead role in paving a two-directional therapeutic avenue, using NRF2 inducers to prevent cancer or NRF2 inhibitors to enhance the efficacy of cancer treatment.

(3) The crosstalk between this system and other critical cellular pathways

One of the major efforts undertaken by my laboratory has been to elucidate the interplay of the NRF2 pathway and other cellular signaling networks. These efforts have led to many influential publications reporting the functional significance of crosstalk between the NRF2 pathway and other important pathways/proteins, including p53-p21, p62-autophagy, and the deubiquitylating enzyme USP15. p53-p21: We reported that a direct interaction between p21 and NRF2 prevents Keap1 from binding NRF2, blocking proteasomal degradation and signifying that the previously reported antioxidant function of p53 relies on the p21-mediated stabilization of NRF2. Tumor suppressor p53 regulates many cellular processes such as cell cycle arrest, senescence, survival, cell differentiation, and apoptosis by controlling the expression of a specific set of genes. Our finding, placing NRF2 downstream of the p53-p21 pathway, not only provides the molecular basis of the previously reported antioxidant function of p53 and p21, but also links the p53 prosurvival function to the NRF2-mediated cellular protective response and ROS singling. p62-autophagy: Another important finding of my laboratory is that the autophagy cargo receptor p62 (sequestosome-1) binds directly to Keap1. Accumulation of autophagosomes results in p62-mediated sequestration of Keap1 into the autophagosomes, which inhibits ubiquitylation of NRF2 and activates NRF2 and its downstream genes (Fig. 2C). This represents a novel, noncanonical mechanism of NRF2 activation (p62-dependent), which is distinct from the canonical mechanism of NRF2 regulation (Keap1 C-151-dependent) (Fig. 2C vs 2A). Furthermore, we have provided compelling evidence that canonical versus noncanonical mechanisms of NRF2 activation may dictate the beneficial (disease-preventive, “bright side”) versus harmful (disease-causative, “dark side”) induction of NRF2, respectively. The significance of this finding is illustrated by “Mechanisms of arsenic-mediated carcinogenicity” described in the next paragraph. USP15: Inspired by the existence of the “yin and yang” forces of nature and clues from our early work, we have been searching for the deubiquitylating enzymes (DUBs) that control the NRF2 pathway. We found that Keap1 is a substrate of the DUB USP15. Deubiquitylation of Keap1 by USP15 enhances the activity of the Keap1-Cul3 E3 ubiquitin ligase. Thus, USP15 negatively controls the protein level of NRF2 through increased ubiquitylation of NRF2. These findings further our understanding of how the NRF2-Keap1 pathway is regulated, which is imperative to targeting this pathway for chemoprevention or chemotherapy. For instance, our data suggest that agents that antagonize USP15 could be therapeutically useful to prevent disease onset and progression through activation of NRF2. Detailed mechanistic understanding of NRF2 regulation by endogenous proteins has offered new paradigms and therapeutic opportunities as well as explaining treatment failures and resistance mechanisms.

(4) Canonical vs. non-canonical mechanism of NRF2 regulation; mechanism of arsenic carcinogenicity

Approximately 200 million people worldwide are exposed to arsenic through air, drinking water, or food. Chronic exposure to arsenic can lead to a number of diseases including cancer, metabolic diseases, vascular diseases, or skin diseases. However, the mechanistic underpinnings of arsenic-mediated pathologies remain vague. In contrast to the canonical NRF2 inducers, such as the chemopreventive compound sulforaphane that induces NRF2 through a Keap1-cysteine 151 dependent manner (Fig 2A), my team found that exposure to environmentally relevant low doses of arsenic leads to accumulation of autophagosomes and constitutive activation of NRF2 through the non-canonical p62-dependent mechanism (Fig. 2C), which represents a previously unrecognized connection between prolonged NRF2 activation and arsenic carcinogenicity in humans. Furthermore, my group has leveraged these data to show that NRF2 activation by chemopreventive small molecules (canonical NRF2 inducers) is a strategy to alleviate arsenic-mediated damage. This work should have a broad impact on diseases caused by arsenic exposure that still affect millions of people worldwide.


(5) NRF2 regulation beyond Keap1: a novel E3 ubiquitin ligase Hrd1

In addition to the Keap1-Cullin3 E3 ubiquitin ligase (Fig. 3A), very recently we reported that Hrd1 is another E3 ubiquitin ligase responsible for compromised NRF2 response during liver cirrhosis (Fig. 3B). This critical observation explained a pathological conundrum, as it has long been known that cirrhotic livers have high levels of ROS and, yet, the levels of NRF2 were low. According to the canonical model, ROS should activate NRF2 through inactivation of the Keap1-Cullin3 E3 ubiquitin ligase. Therefore, our findings demonstrating Hrd1 as a novel E3 ubiquitin ligase for NRF2 during activation of the XBP1-Hrd1 arm of ER stress offered an important explanation for the Keap1 independent destruction of NRF2. This work not only revealed the convergence of ER and oxidative stress response pathways, but also the pathological importance of this crosstalk in liver cirrhosis. Furthermore, we demonstrated the importance of therapeutic targeting of Hrd1 and not Keap1 to activate NRF2-mediated protection to alleviate the progression of liver cirrhosis.