Subscriber access provided by University of Sussex Library
Letter
A small molecule inhibits deregulated NRF2 transcriptional activity in cancer. Michael J. Bollong, Hwayoung Yun, Lance Sherwood, Ashley K. Woods, Luke L. Lairson, and Peter G Schultz ACS Chem. Biol., Just Accepted Manuscript • DOI: 10.1021/acschembio.5b00448 • Publication Date (Web): 13 Aug 2015 Downloaded from http://pubs.acs.org on August 17, 2015
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
ACS Chemical Biology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 17
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Chemical Biology
“A small molecule inhibits deregulated NRF2 transcriptional activity in cancer.” Michael J. Bollong1, Hwayoung Yun1, Lance Sherwood2, Ashley K. Woods2, Luke L. Lairson1, 2*, Peter G. Schultz1, 2* 1
Department of Chemistry, The Scripps Research Institute, 10550, North Torrey Pines La Jolla, CA 92037, USA. 2
California Institute for Biomedical Research, 11119 North Torrey Pines Road, Suite 100, La Jolla, California 92037, United States
Abstract NRF2 serves as the master regulator of oxidative stress resistance in mammalian cells. Although NRF2 activation decreases tumorigenic events in normal cells, accumulating evidence suggests that cancers have broadly selected for NRF2-activating mutations to promote anabolic growth and chemoresistance. Small molecules which inhibit NRF2 activity may therefore offer promise as an alternative anticancer treatment in NRF2 dependent cancers. We have used a high throughput screen to identify small molecules which decrease NRF2 transcriptional activity at antioxidant response element sites. One such molecule, termed AEM1, is capable of broadly decreasing the expression of NRF2 controlled genes, sensitizing A549 cells to various chemotherapeutic agents, and inhibiting the growth of A549 cells in vitro and in vivo. Profiling of multiple cell lines for their responsiveness to AEM1 revealed that AEM1’s activities are restricted to cell lines harboring mutations which render NRF2 constitutively active.
ACS Paragon Plus Environment
ACS Chemical Biology
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
NRF2 (nuclear factor E2-related factor 2), a basic leucine zipper transcription factor, under normal conditions is bound to its repressor protein KEAP1 (Kelch ECH associating protein 1) in the cytoplasm where it is continually sequestered for proteasomal degradation.1 However, in the presence of oxidative stress or electrophilic chemicals, covalent modification of KEAP1 results in the dissociation and nuclear accumulation of NRF2.1 In the nucleus, NRF2 binds to antioxidant response element (ARE) consensus sites to positively effect the expression of a few dozen cytoprotective genes.1 The net effect of NRF2 activation is the clearance of xenobiotics and electrophiles from the cell and cellular resilience in the presence of oxidative stress.1 This beneficial effect on cell survival has been exploited clinically as a host of potent KEAP1-reactive electrophiles have been tested or approved for the treatment of various diseases including multiple sclerosis (dimethyl fumarate), cancer prevention (Oltipraz), and pulmonary arterial hypertension (Bardoxolone methyl), among others.2-4 Despite these beneficial effects, a wealth of new evidence has suggested that many human cancers, including those of the lung, breast, colon, ovaries, and pancreas, have mutations which promote the stability and activity of NRF2.5,6 A number of mechanisms that lead to constitutive NRF2 activity in cancer have been demonstrated, including mutations in KEAP1 and NRF2, mutations in fumarate hydratase leading to KEAP1 succination, KEAP1 promoter hypermethylation, and the expression of oncogenes leading to increased NRF2 transcription.6 Continuous NRF2 activity results in increased proliferation and angiogenesis, as well as resistance to chemotherapy and radiation.7-10 Given the high incidence and dependence of NRF2 expression in cancer, the identification of additional inhibitors of NRF2 is of considerable interest. Herein, we have used a reporter based screen to identify a new class of compounds which functionally inhibit NRF2 activity in a genotype selective manner. To identify small molecule inhibitors of the NRF2 pathway, we first engineered a reporter cell line in which NRF2 transcriptional activity is be constitutively “on.” Previous work by Tuveson and colleagues demonstrated that oncogene expression (K-Ras, B-Raf, and c-Myc) can increase the expression of NRF2 past the suppressive capacity of endogenous KEAP1.9 This results in increased expression of the NRF2 antioxidant transcriptional program, decreased cellular reactive oxygen species, and increased replicative potential.9 To mimic this cellular state and to discourage the identification of compounds which simply inhibit NRF2 transcription, we overexpressed c-Myc in 3T3 mouse embryonic fibroblasts harboring a stably integrated antioxidant response element driving the expression of firefly luciferase (MYC-3T3-ARE-LUC cells) (Supplementary Figure 1A). Consistent with previous results, exogenous expression of c-Myc resulted in a dose dependent increase of transcriptional activity at ARE sites (Supplementary Figure 1B).9 Additionally, MYC-3T3-ARE-LUC cells remained responsive to the commonly-used NRF2 activator tert-butyl hydroquinone (TBHQ), indicating that the observed increase in luciferase signal is due to relevant NRF2-driven transcription (Supplementary Figure 1C)11. We then screened a library of ~30,000 diverse heterocyclic compounds and known biologically active small molecules for ARE suppressive activity. 27 compounds were identified which decreased the ARE-LUC signal > 2 z-scores from plate mean but were not generally cytotoxic or repressive of transcription (Figure 1A). Among these hits, we found the synthetic glucocorticoid dexamethasone to be a potent inhibitor of ARE-LUC activity. Dexamethasone has previously been reported to inhibit both basal and inducible NRF2 transcriptional activity, and, as such, served as a positive control.12 In addition, we found 3 novel chemical scaffolds which dose dependently decreased ARE-LUC signal in MYC-3T3-ARE-LUC cells (Figure 1B,C). Among these, one thienopyrimidine-containing compound, CBR-031-1, was chosen for further study due to its chemical tractability and reproducible ability to broadly inhibit NRF2-controlled gene expression. To confirm the NRF2 suppressive activity of CBR-031-1 in a relevant cancer line, we tested the activity of CBR-031-1 in A549 cells, a lung adenocarcinoma cell line with an inactivating mutation in KEAP1 (G333C) that leads to constitutive NRF2 activation.13 CBR-031-1 dose dependently suppressed luciferase signal in A549 cells with a stably integrated ARE-luciferase construct (A549-ARE-LUC cells) and decreased the total transcript levels of heme oxygenase 1 (HMOX1), a gene previously reported to be under basal transcriptional control of NRF2 in these cells (Figure 2A,B).10 To identify analogs with improved potency but devoid of the potential metabolic liabilities present in CBR-031-1, we undertook a
ACS Paragon Plus Environment
Page 2 of 17
Page 3 of 17
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Chemical Biology
structure activity relationship study around the thienopyrimidine scaffold with ~100 structurally similar commercially available analogs. Testing this series against MYC-3T3-ARE-LUC cells, we identified a number of compounds that displayed half maximal inhibitory activities at sub-micromolar potencies. Among these was a benzodioxole substituted analog which we termed AEM1 for ARE expression modulator 1 (Figure 2C). AEM1 displayed enhanced potency over D341-0389 at inhibiting ARE-LUC signal and suppressing HMOX1 suppression at both transcript and protein levels in A549 cells with a half maximal effective concentration of ~650 nM (Figure 2A,B, Supplementary Figure 2). To confirm that AEM1 broadly suppressed NRF2 controlled transcription, we treated A549 cells with AEM1 and then measured the transcript levels of a panel of genes whose transcription has been previously reported as being either basally or inducibly controlled by NRF2.14 Both 24 and 72 hour treatment with AEM1 resulted in a significant decrease in most transcripts measured, indicating that AEM1 treatment results in the steady state inhibition of NRF2 activity (Figure 2D, Supplementary Figure 3). Additionally, AEM1 treatment resulted in a dose dependent decrease in the content of reduced glutathione (GSH) in A549 cells as measured by GSH-Glo, a result consistent with NRF2’s reported activity in promoting the expression of genes necessary for glutathione synthesis and recycling (Figure 2E).1 Knockdown of NRF2 has previously been reported to sensitize A549 cells to various chemotherapies as well as inhibit their ability to proliferate in vitro and in vivo.10 Consistent with these reports, AEM1 treatment significantly reduced the ability of A549 cells to grow in anchorage independent growth conditions, but only modestly reduced their proliferation in monolayer culture (Figure 3A,C). Additionally, treatment of A549 cells with a combination of AEM1 and the known chemotherapeutic agents doxorubicin, etoposide, or 5-fluorouracil revealed that AEM1 additively sensitized these cells to the toxic effects of etoposide and 5-fluorouracil (Figure 3C). Interestingly, co-treatment of AEM1 and doxorubicin proved to be synergistically toxic with higher doses of AEM1 as determined by changes in Bliss independence calculations, indicating that AEM1 might preferentially sensitize cells to reactive oxygen species-inducing chemotherapies (Figure 3C,D). Consistent with this hypothesis, 48 hour pretreatment of A549 cells with AEM1 resulted in a significant sensitization to peroxide-induced cell death (Figure 3B). To confirm that the anti-proliferative and drug-sensitizing properties of AEM1 were due to specific NRF2 antagonism, six structurally similar analogs which were inactive at suppressing ARE-LUC signal or HMOX1 expression were investigated (Supplementary Figure 6 A-C). In contrast to AEM1, these analogs did not suppress the growth of A549 cells and did not sensitize A549 cells to doxorubicininduced cell death, indicating that the toxic effects observed with AEM1 treatment are due to NRF2 inhibition and not an off target activity of the scaffold (Supplementary Figure 6 D-G). Constitutive NRF2 expression in A549 cells promotes metabolite shuttling to the pentose phosphate pathway to provide NADPH reducing equivalents and nucleotide metabolites necessary for sustained tumor growth.15 As such, decreasing total NRF2 levels by genetic or chemical methods results in the inability of A549 cells to form tumors in vivo.10,15,16 We therefore sought to determine if AEM1 could inhibit the in vivo growth of A549 cells in a subcutaneous xenograft model. We first measured the pharmacokinetic profile of orally administered AEM1 (2 hour half-life; 176 ng/mL Cmax; 453 hr*ng/mL AUC(0-24); from a 20 mg/kg dose). Although our in vitro results indicated that maximal anti-proliferative efficacy would likely be observed at doses above 5 µM, our pharmacokinetic results indicated that ~1 µM mouse serum concentrations of drug could be achieved with a 50 mg/kg dose. Despite inherent exposure limitations with AEM1, treatment with a twice daily oral 50 mg/kg dose of compound resulted in a modest yet statistically significant reduction in A549 tumor volume without affecting body weight throughout the treatment period (Figure 3E,F). Compounds with improved potency and pharmacokinetic properties may display enhanced anti-tumor efficacy and are currently under investigation. However, these results provide proof of principle that AEM1 is capable of antagonizing the NRF2 transcriptional program necessary for tumor growth in vivo. Among known pathways controlling NRF stability, both the PI3K/AKT pathway and PKC kinases have been shown to phosphorylate NRF2, a modification which promotes NRF2 nuclear entry and stability.15,17 To determine if AEM1 acts by a similar mechanism of action, A549 cells were treated with AEM1 or chemical inhibitors of PI3K (BEZ235) or PKC (Gö 6983) family members and the levels
ACS Paragon Plus Environment
ACS Chemical Biology
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
of NRF2 and HMOX1 protein were analyzed by Western blotting at various time points. Inhibition of PI3K and PKC both resulted in a transient reduction in NRF2 protein levels, as well as decreased levels of HMOX1 protein, consistent with inhibition of the NRF2 transcriptional program (Supplementary Figure 4A). Surprisingly, NRF2 protein levels were unaffected by AEM1 treatment although HMOX1 protein was significantly reduced at 24 hours (Supplementary Figure 4A). We then profiled AEM1 against a panel of ~100 diverse protein and lipid kinases (including isoforms of PI3K, AKT, and PKC) for inhibitory activity. AEM1 did not inhibit any of the kinases at concentrations up to 5 µM, further confirming that AEM1 does not inhibit any known NRF2-phosphorylating kinases (Supplementary Table 1). Another potential mechanism for NRF2 inhibition might involve the up-regulation of KEAP1 protein levels to compensate for the inactivating mutation observed in KEAP1 in A549 cells. Measuring the protein level of KEAP1 by Western blotting over the course of a 24 hour treatment period in A549 cells revealed that AEM1 had no effect on KEAP1 protein abundance (Supplementary Figure 4B). Together, these results suggest that AEM1inhibits NRF2 activity by a novel mechanism of action which does not involve altering the protein levels of NRF2 or KEAP1. To determine if the mechanism of action by which AEM1 inhibits NRF2 activity was generalizable to other cell types, we next profiled a panel of cell lines and primary cells for responsiveness to AEM1 treatment. Treating cells with either 1 or 10 µM AEM1 for 24 hours and then quantifying the relative transcript levels of HMOX1 as a readout of NRF2 activity revealed that responsiveness to AEM1 was highly variable among the cell types tested. While some cell types (HeLa and HEK293T cell lines) displayed a modest suppression of HMOX1 transcript level, other cells types (SH-SY5Y and primary human lung fibroblasts) displayed a slight but reproducible trend in the upregulation of HMOX1 transcript, indicating that AEM1 can actually activate NRF2 in some cell types (Figure 4). Among the most responsive cell lines were H838 and H460, both of which displayed comparable levels of HMOX1 suppression to A549 (Figure 4). Interestingly, these cell lines have previously been characterized by Biswal and colleagues as harboring NRF2-activating mutations in KEAP1 (D236H for H460 and a stop codon at position 443 for H838) in a similar manner to A549.13 These results suggest that AEM1’s NRF2 inhibitory activities are restricted to cell lines in which mutations have induced a constitutively active NRF2 transcriptional program. A number of small molecule inhibitors of NRF2 that display in vivo antitumor activity have been reported to date, including the natural product brusatol and the coffee-derived alkaloid trigonelline.5,16 Both compounds act by destabilizing NRF2 and therefore inhibit NRF2 activity in both normal and cancer cells. In contrast, using a reporter cell line mimicking the deregulated NRF2 system observed in cancer, we identified AEM1 as a functional inhibitor of NRF2 activity. Treating the NRF2 dependent cell line A549 with AEM1 resulted in the broad suppression of NRF2 driven genes, sensitization to chemotherapies and oxidative stress, and the reduction of their growth in vitro and in vivo. Further, AEM1 suppressed NRF2 activity in other cancer cell lines with NRF2 activating mutations in KEAP1 while displaying minimal inhibitory activity in other cell types. The observation that AEM1 only inhibits NRF2 in cells with a constitutively active NRF2 program may suggest a unique dependence on a previously undescribed factor necessary for prolonged NRF2 activation. Efforts to identify the cellular target and associated mechanism of action of AEM1 are currently underway. Understanding AEM1’s mechanism should provide insight into NRF2 cancer dependency and may ultimately inform the design of alternative anticancer treatments.
ACS Paragon Plus Environment
Page 4 of 17
Page 5 of 17
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Chemical Biology
Figure 1. A high throughput screen identifies small molecule inhibitors of NRF2 transcriptional activity. (A) Z-scores of luciferase signal for 27 hits selective for activity in MYC-3T3-ARE-LUC cells when compared to a historical toxicity screen. Relative ARE-LUC signal in MYC-3T3-ARE-LUC cells (B) and chemical structures (C) of 3 compounds identified by high throughput screening.
ACS Paragon Plus Environment
ACS Chemical Biology
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Figure 2. AEM1 broadly decreases NRF2-driven gene expression in A549 cells. (A) Relative luminance values of A549-ARE-LUC cells treated with the indicated doses of AEM1 or CBR-031-1 for 24 hours. (B) Relative expression of HMOX1 mRNA in A549 cells treated with the indicated doses of AEM1 or CBR-031-1 for 24 hours. (C) Structure of AEM1. (D) A549 cells were treated for 72 hours with 5 µM AEM1 and the relative mRNA levels of NRF2 controlled genes measured by qRT-PCR. (E) Relative levels of reduced glutathione content of A549 cells treated for 24 hours with the indicated doses of AEM1. (n=3, mean and standard deviation.*P
