Non-covalent NRF2 Activation Confers Greater Cellular Protection than Covalent Activation

Abstract

The transcription factor NRF2 confers cellular protection by maintaining cellular redox homeostasis and proteostasis. Basal NRF2 levels are normally low due to KEAP1-mediated ubiquitylation and subsequent proteasomal degradation. KEAP1, a substrate adaptor protein of a KEAP1-CUL3-RBX1 E3 ubiquitin ligase complex, contains a critical cysteine (C151) that is modified by electrophiles or oxidants, resulting in inactivation of the E3 ligase and inhibition of NRF2 degradation. Currently, nearly all NRF2 inducers are electrophilic molecules that possess unwanted off-target effects due to their reactive nature. Here, we report a group of NRF2 inducers, ent-kaurane diterpenoid geopyxins, with and without C151 reactive electrophilic moieties. Among 16 geopyxins, geopyxin F, a non-electrophilic NRF2 activator, showed enhanced cellular protection relative to an electrophilic NRF2 activator, geopyxin C. To our knowledge, this is the first detailed structure-activity relationship study of covalent versus non-covalent NRF2 activators, showing the promise of non-covalent NRF2 activators as potential therapeutic compounds.

Keywords: NRF2; cancer; chemoprevention; drug discovery; geopyxin; natural product; non-covalent; structure activity relationship.

Figure 1.

Chemical structures of geopyxins and their derivatives.

Figure 2.

Induction of autophagosome formation by geopyxin analogs. Immortalized Baby Mouse Kidney Epithelial Cells (iBMK) stably expressing GFP-LC3 were treated as indicated. Autophagosome formation was visualized by the formation of fluorescent puncta using fluorescence microscopy. NT: no treatment, As: 1 µM NaAsO₂, 3: compound 3 (2.33 µM), 5: compound 5 (71.5 µM), and 14: compound 14 (54.5 µM). Each row shows three fields of view that were randomly selected.

Figure 3.

NRF2 and its downstream genes are activated by compounds 3 and 5. (A-B) MDA-MB-231 cells were treated with the indicated dose of 3 (left panel) or 5 (right panel) for 16 hours (A) or with 3 (2.33 µM) or 5 (71.5 µM) for the indicated time (B). The protein level of NRF2 and the indicated downstream genes was measured by immunoblot analysis. (C-D) MDA-MB-231 cells were treated with the indicated dose of compound 3 (C) or compound 5 (D) for 16 hours, mRNA expression of NRF2 and its downstream genes was measured by real-time qRT-PCR. AKR1C1 (Aldo-Keto Reductase 1 Member C1), NQO1 (NAD(P)H Dehydrogenase Quinone 1), HO-1 (heme oxygenase-1), and GCLM (Glutamate-Cysteine Ligase Modifier Subunit) are well-defined NRF2 target genes. All experiments were repeated in triplicate. Bar graphs represent the mean of three independent experiments and error bars represent the standard deviation from the mean. *: p<0.05 compared with the control group. Details of statistical analysis can be found in the STAR Methods.

Figure 4.

Compounds 3 and 5 both block ubiquitylation and stabilize NRF2 in a KEAP1-dependent manner, but only 3 depends on Cys151 in KEAP1. (A) KEAP1-WT and KEAP1^−/− MDA-MB-231 cells were treated with compounds 3 (left panel) or 5 (right panel) for four hours. The protein level of NRF2 was measured by immunoblot (upper panels), the level measured by densitometry, and plotted relative to the GAPDH loading control (lower panels). (B) MDA-MB-231 cells, transfected by expression vectors for NRF2 and HA-Ubiquitin, were treated with one dose of compound 3 (2.33 µM) or 5 (71.5 µM) for 4 hours. Ubiquitylation of NRF2 was measured by immunoprecipitation of NRF2, followed by immunoblot analysis with an anti-HA antibody. (C) MDA-MB-231 cells were left untreated, treated with compound 3 (2.33 µM), or compound 5 (71.5 µM) for the indicated times. The protein level of NRF2 was measured by immunoblot analysis, the intensity of the NRF2 band (normalized to GAPDH) vs time was plotted and the half-life of NRF2 was calculated. (D-E) To determine the KEAP1-Cys151 dependence, MDA-MB-231-KEAP1^−/− cells, established using the CRISPR-Cas9 technique, were transfected with either an expression vector for KEAP1 WT or KEAP1 C151 together with an NQO1-ARE-firefly luciferase and TK-Renilla luciferase vector. Twenty-four hours after transfection, cells were treated with the indicated compound for 16 hours. Both firefly luciferase activity relative to Renilla luciferase activity (D) and the protein levels of NRF2 and KEAP1 (E) were measured (left panel) and the intensity of NRF2 bands (normalized to GAPDH) was plotted (right panel). Experiments were repeated in triplicate. Bar graphs represent the mean of three independent experiments and error bars represent the standard deviation from the mean. *: p<0.05 compared with the control group. Details of statistical analysis can be found in the STAR Methods.

Figure 5.

Compounds 3 and 5 confer protection against xenobiotic insult in an NRF2 dependent manner. (A-B) WT MDA-MB-231 cells and KEAP1^−/− MDA-MB-231 cells, were pre-treated with sulforaphane (2 µM), 3 (2.33 µM), or 5 (71.5 µM) for 4 hours to induce NRF2, then the indicated dose of arsenic (A) or cisplatin (B) was added for an additional 48 hours, in the presence of half the original concentration of sulforaphane (0.5 µM), 3 (1.16 µM), or 5 (35.8 µM). Cellular toxicity was measured by MTT assay. (C-D) WT or KEAP1^−/− MDA-MB-231 cells were pre-treated with compound 3 (C) or 5 (D) at different concentrations (0 x EC₅₀, ¼ × EC₅₀, ½ × EC₅₀, EC₅₀, 2 × EC₅₀, and 4 × EC₅₀) for 4 hours, then arsenic (20 µM) or cisplatin (10 µM) was added for an additional 48 hours. Each experiment was repeated in triplicate, points indicate the average of three experiments, and error bars are the standard deviations from the means. * indicates p < 0.05. Details of statistical analysis can be found in the STAR Methods.