NRF2/KEAP1 and Wnt/β‐catenin in the multistep process of liver carcinogenesis in humans and rats

Potential conflict of interest: Nothing to report. The group is supported by INCa with the ICGC, the PAIR‐CHC project NoFLIC (also funded by ARC) and the Ligue Nationale contre le Cancer. See Article on Page 851. In humans, liver carcinogenesis in cirrhosis is a multistep process with occurrence of sequential lesions: low‐grade dysplastic nodules > high‐grade dysplastic nodules > early hepatocellular carcionoma (HCC) > advanced HCC. Finally, the genome of each HCC is composed of a unique combination of somatic genetic alterations underlying the complexity of the carcinogenesis mechanisms in the liver. However, somatic mutations recurrently alter a subset of 160 cancer driver genes1 targeting key signaling pathways involved in liver carcinogenesis: telomere maintenance; cell‐cycle regulators; WNT/β‐catenin; epigenetic modifiers; RAS/RAF mitogen‐activated protein kinase; phosphoinositide 3‐kinase/protein kinase B/mammalian target of rapamycin; and nuclear factor erythroid 2‐related factor 2/Kelch‐like erythroid cell‐derived protein with CNC homology ECH‐associated protein (NRF2/KEAP1) pathways.1 NRF2 and KEAP1 are the key players of the oxidative stress pathway that trigger an antioxidant response program when cells are exposed to reactive oxygen species (ROS).4 In an inactivate state, NFR2 (coded by NFE2L2) is a transcription factor retained in the cytoplasm through its binding by KEAP1 and cullin 3 (CUL3; Fig. 1). Thus, NRF2 is ubiquitinated by this complex and degraded by the proteasome. When the level of ROS rises in the cell, KEAP1 dissociates from NRF2 and NRF2 translocates to the nucleus to act as a transcription factor. It increases the transcription of several genes (e.g., NAD(P)H dehydrogenase quinone 1 [NQO1], glutamate cysteine ligase catalytic [GCLC], and glutathione S‐transferase alpha 4 [GSTA4]) and triggers an antioxydant response.4 Because chronic oxidative stress is a classical risk factor of tumor development, initial study using mouse models have proposed the NRF2/KEAP1 pathway as tumor protective owing to its antioxidant properties.5 However, genetic exploration of human cancer genome has challenged this point of view. Next‐generation sequencing identified recurrent activating mutations of NFE2L2 (3%‐6%) and inactivating mutations of KEAP1 (2%‐8%) in human HCC.1NFE2L2 mutations were also found in 12% of hepatoblastoma.6 These mutations lead to constitutive activation of the NRF2/KEAP1 pathway, protect from toxicity of ROS, and give a proliferative advantage.7 Moreover, an association with catenin (cadherin‐associated protein), beta 1, 88kDa (CTNNB1; coding for β‐catenin) mutations has been also demonstrated both in HCC and hepatoblastoma, suggesting cooperation between NRF2/KEAP1 and WNT/β‐catenin activation in liver tumorigenesis.1Figure 1: Multistep process of liver carcinogenesis in humans and rats and potential pharmacological targeting of NRF2/KEAP1 pathway‐activated tumors. NQO1, a surrogate marker of NRF2/KEAP1 pathway activation in humans and rats, catalyzes the reduction of benzoquinone ansamycin Hsp90 inhibitors to the hydroquinone that binds Hsp90 more tightly and therefore displays a more potent Hsp90 inhibitor activity than parent benzoquinone ansamycin. Abbreviations: 2‐AAF, 2‐acetylaminofluorene; DENA, diethylnitrosoamine; HBV, hepatitis B virus; HCV, hepatitis C virus; NASH, nonalcoholic steatohepatitis; NFE2L2, nuclear factor (erythroid‐derived 2)‐like 2.Construction of animal models that summarizes the tumorigenesis mechanisms identified in humans is a major goal to develop the most useful preclinical models. In their study, Zavattari et al. analyzed a model of tumor progression in the rat by combining diethylnitrosamine, 2‐acetylaminofluorene, and two‐thirds hepatectomy in order to study the somatic genetic alterations in preneoplastic nodules, early HCC, and advanced HCC occurring on a cirrhotic‐like background.8 Interestingly, they identified highly frequent somatic mutations in Nfe2l2 or Keap1 in 74% of preneoplastic lesions, 78% of early HCC, and 59% of advanced HCC. Nfe2l2 mutations were the most frequent genetic alterations identified (>75%), all missenses affecting the Keap1‐binding site and leading to Nrf2 activation as further shown in cell models. Because Keap1 physiologically inactivates Nfr2, Keap1 mutations were exclusive from Nfe2l2. Nfe2l2 and Keap1 mutations in preneoplastic lesions were associated with an increased expression of three Nrf2 targets: Nqo1, Gclc and Gsta4. It demonstrates the early constitutive activation of the oxidative stress pathway in this model. Interestingly, in preneoplastic lesions, nonmutated nodules did not harbor Nrf2 activation, whereas in early and advanced HCC, both mutated and nonmutated HCC showed an increased Nqo1, Gclc, and Gsta4 expression. It suggests that additional mechanisms could lead to Nrf2/Keap1 pathway activation in nonmutated early and progressed HCC. In addition, using HEK293T cells and cell lines derived from rat preneoplastic lesions, the investigators confirmed that mutations of NFE2L2 activate the NRF2/KEAP1 pathway and that decreased NRF2 expression impaired tumorigenic ability of tumor cells. All these data strongly support the role of the NRF2/KEAP1 pathway in liver carcinogenesis, both in humans and rodents. Consequently, the rat resistant hepatocyte model will be a useful preclinical model to test drugs targeting the NRF2/KEAP1 pathway. We recently showed that human HCC cell lines were highly sensitive to the benzoquinone ansamycin class of heat shock protein 90 (HSP90) inhibitors (17‐N‐allylamino‐17‐demethoxygeldanamycin a.k.a. 17‐AAG and 17‐dimethylaminoethylamino‐17‐demethoxygeldanamycin a.k.a. 17 DMAG) when they overexpress NQO1, a surrogate marker of NRF2/KEAP1 pathway activation both in humans and rats1 (Fig. 1). Consequently, antitumor effect and toxicity of HSP90 inhibitors could be tested in vivo in the rat model developed by the investigators. Zavattari et al. identified CTNNB1 somatic mutations activating ß‐catenin as a second recurrent oncogenic event. However, ß‐catenin activation was less frequent and found only in advanced HCC (18%). Consequently, the timeframe of genetic alterations in rat liver carcinogenesis indicates Nfe2l2 and Keap1 mutations as early events leading to tumor initiation and malignant transformation. In contrast, activating mutations of β‐catenin were late events involved in tumor progression. This situation mirrors partially the events observed in humans. In patients, mutations in the telomerase reverse transcriptase (TERT) promoter are the most frequent genetic alterations observed in HCC and the only recurrent genetic alterations identified in premalignant nodules on cirrhosis.9 However, in the mouse and rat, such genetic events have not been identified thus far, probably owing to the different telomere and telomerase biology in rodents. Moreover, in humans, mutations of NFE2L2 and KEAP1 were only observed in progressed HCC and not in premalignant nodules and early HCC, suggesting that these mutations are late events in liver carcinogenesis.1 In contrast, both the rat model presented in this study and the analysis of human HCC developed on cirrhosis concluded that CTNNB1 mutations are late genetic events involved in tumor progression (Fig. 1). In conclusion, despite some differences in the sequence of accumulation of genetic alterations between rats and humans, Zavattari et al. have provided new insights in our understanding of the multistep process of liver carcinogenesis in the rat resistant hepatocyte model.8 This model will be also a very useful tool to study the role of NRF2/KEAP1 in liver carcinogenesis and test targeted therapy in vivo (Fig. 1). Author names in bold designate shared co‐first authorship.