I-BET-762 – 2hydroxy 4methylthiobutanoate

A natural compound, aristoyagonine, is identiﬁed as a potent bromodomain inhibitor by mid-throughput screening

Abstract

Bromodomain-containing protein 4 (Brd4) is known to play a key role in tumorigenesis. It binds acet- ylated histones to regulate the expression of numerous genes. Because of the importance of brd4 in tumorigenesis, much research has been undertaken to develop brd4 inhibitors with therapeutic po- tential. As a result, various scaffolds for bromodomain inhibitors have been identiﬁed. To discover new scaffolds, we performed mid-throughput screening using two different enzyme assays, alpha-screen and ELISA. We found a novel bromodomain inhibitor with a unique scaffold, aristoyagonine. This natural compound showed inhibitory activity in vitro and tumor growth inhibition in a Ty82-xenograft mouse model. In addition, we tested Brd4 inhibitors in gastric cancer cell lines, and found that aristoyagonine exerted cytotoxicity not only in I-BET-762-sensitive cancer cells, but also in I-BET-762-resistant cancer cells. This is the ﬁrst paper to describe a natural compound as a Brd4 bromodomain inhibitor.

1. Introduction

At present, 46 proteins have been reported to contain bromo- domains; they can be divided into eight major families based on sequence similarity [1]. Among the bromodomain-containing proteins, BET (bromodomain and extraterminal domain) proteins, which have two bromodomains and a unique extra-terminal domain, including Brd2, Brd3, Brd4, and Brdt, are known to play a key role in chromatin biology. They modulate gene transcription through recruitment of the transcriptional machinery to the histone-chromatin complex via bromodomains, which bind to acetylated lysine residues of histone tails [2]. Brd4 is the most commonly studied member of BET proteins.

There are three isoforms of Brd4 in humans. One isoform has 1362 amino acid residues, and the others have 722 and 796 resi- dues, respectively. Brd4 is recruited to lineage-speciﬁc enhancers and promoters by binding to acetylated histones. It recruits several proteins, such as NSD3/CHD8, and p-TEFb, which promote tran- scriptional activation, to enhancers/promoters through direct physical interaction [3].
Brd4 has been considered a therapeutic target for cancer since the discovery of the fusion protein Brd4-Nut in midline carcinoma [4]. More than 90% of midline carcinomas contain the Brd4-Nut fusion gene. An in vitro analysis revealed that knockdown of Brd4 markedly decreased Brd4-Nut positive cell proliferation [5]. In addition, Zuber et al. reported that Brd4 knockdown down- regulated c-Myc expression to induce cell death in AML cell line [6]. c-Myc is known as a key player in cancer cell proliferation. However, it has been regarded as a non-druggable target due to the absence of enzymatic activity or any deep pockets for small mole- cule inhibitors [7]. Zuber et al. indicated that Brd4 ablation can be a good strategy for cancer treatment because of its ability to down- regulate c-Myc expression. For this reason, there have been many efforts to develop potent Brd4 inhibitors [8]. At present, more than 30 papers relevant to the development of BET bromodomain in- hibitors have been published [9e16], and 16 compounds are un- dergoing clinical trial [17]. No natural product, however, has been reported as a bromodomain inhibitor.

Here, we performed mid-throughput screening to identify novel Brd4 inhibitors with new scaffolds, and identiﬁed a natural product as a bromodomain inhibitor. We also studied the sensitivity of gastric cancer cells to BET bromodomain inhibitors. Several studies have been performed to investigate whether Brd4 inhibition causes cell death in solid tumors, including breast cancer, prostate cancer, lung cancer, colon cancer, and hepatocellular cancer, whereas few studies have focused on gastric cancer [18e20]. In this study, we suggest that Brd4 inhibitors can induce cell death in gastric cancer cells. In addition, aristoyagonine is effective on gastric cancer cells that are resistant to I-BET-762, which is under clinical trial.

2. Materials and methods

2.1. Molecular cloning & protein expression, and puriﬁcation

Brd4 cDNA was provided by Dr. Stefan Knapp from the Univer- sity of Oxford. N-terminal GST-tagged and C-terminal His-tagged BD1 (GST-BD1-His6) was expressed in E. coli and puriﬁed. BD1 spans 47e170 amino acids. The pGEX 6P-1 vector was digested with EcoRI and XhoI restriction enzymes. BD1 PCR was performed with the BD1_Forward primer (50- ATC TAG GAA TTC CCC CCA GAG ACCTCC AAC CC -30) and BD1_Rev primer (50-ATC TAG CTC GAG TTA GTG GTG GTG GTG GTG GTG TTC GAG TGC GGC CGC AAG CTC GGT TTCTTC TGT GGG TA-30). BL21 Star (DE3) was transformed and induced
by 0.1 mM IPTG overnight at 18 ◦C. The cells were lysed with lysozyme (1 mg/mL) and sonicated in lysis buffer (50 mM NaH2PO4, 300 mM NaCl, 10 mM imidazole, and adjusted pH to 8.0 by NaOH) and centrifuged at 8000 rpm for 30 min. The supernatant was incubated with Ni-NTA beads (Qiagen) for 2 h at 4 ◦C and proteins were eluted with elution buffer (50 mM NaH2PO4, 300 mM NaCl,250 mM imidazole, and adjusted pH to 8.0 by NaOH). Puriﬁed His- tag proteins were further puriﬁed by size exclusion chromatog- raphy on a superdex 16/600 Hiload column (GE Healthcare) using buffer (50 mM Tris-HCl pH 7.4, 200 mM NaCl).

2.2. Alpha-screen enzyme assay

The alpha-screen assay was performed in accordance with the manufacturer’s protocol (PerkinElmer, USA), by using a buffer (50 mM HEPES, 100 mM NaCl, 0.1% BSA, pH 7.4 supplemented with 0.05% CHAPS) and OptiPlate™—384 plate (PerkinElmer, USA). Brieﬂy, 2.5 mL of compound solution and 5 mL of peptide solution [SGRGK(Ac)GGK(Ac)GLGK(Ac)GGAK(Ac)RHRK-biotin] were added to 5 mL of glutathione-S-transferase (GST) and His-tagged BD1 or BD2 in OptiPlate™ 384 plate. Streptavidin-coated donor beads and anti-GST alpha-screen acceptor beads were added under low- light condition. Plate was incubated at 25 ◦C for 60 min using a Thermomixer C (Eppendorf, USA), and read using a Fusion-Alpha™
Multilabel Reader (PerkinElmer, USA). The alpha-screen results were conﬁrmed by using alpha-screen TruHit kits (PerkinElmer, USA).

2.3. ELISA assay

Streptavidin coated 384-well plates (Thermo Fisher Scientiﬁc, USA) were rinsed by phosphate buffered saline/Tween-20 (PBST). Biotin-tagged tetra acetylated or non-acetylated peptides were added and incubated in a cold chamber overnight. The plate was washed by PBST and blocking was performed by using a 1% bovine serum albumin solution. 10 mL of compound solution, and 40 mL of glutathione-S-transferase (GST) and His tagged BD1 were added to each well. After 1 h, 25 mL of primary GST-antibody (0.01 ng/ml, Abcam, USA), and 25 mL of 2nd HRP conjugated antibody (Thermo Fisher Scientiﬁc, USA) were added. Between stages, the plate was washed in PBST. Finally, the absorbance of the wells was measured by using an EnVision Multilabel Plate Reader (PerkinElmer, USA).

2.4. Western blot

For immunoblotting, cells were washed in PBS, lysed in 1 sample buffer (50 mmol/L Tris-HCl (pH 6.8), 10% glycerol, 2% SDS, and 3% b-mercaptoethanol), and boiled for 10 min. Lysates were subjected to SDS-PAGE followed by blotting with the indicated antibodies and detection by Western blotting substrate ECL reagent (Thermo Scientiﬁc). Images were produced using a SensiQ-2000 and Image software. The following antibodies were obtained from Cell Signaling Technology: c-Myc (Catalog No. 5605). Tubulin antibody (Catalog No. T6199) was purchased from Sigma-Aldrich. HRP-conjugated anti-mouse (Catalog No. NCI1430KR), and HRP- conjugated anti-rabbit (Catalog No. NCI1460KR) antibodies were obtained from Thermo Scientiﬁc.

2.5. c-Myc knockdown

The pLKO.1 vector was digested with AgeI and EcoRI. pLKO.1 The primers for c-Myc shRNA 2 were: CCGG GAT GAG GAA GAA ATC GAT G CTCGAG C ATC GAT TTC TTC CTC ATC TTTTT and AATTAAAAA GAT GAG GAA GAA ATC GAT G CTCGAG C ATC GAT TTC TTC CTC ATC. The
primers for pLKO.1 c-Myc shRNA 4 were: CCGG CCT GAG ACA GAT CAG CAA CAA CTCGAG TTG TTG CTG ATC TGT CTC AGG TTTTT and AATTAAAAA CCT GAG ACA GAT CAG CAA CAA CTCGAG TTG TTG CTGATC TGT CTC AGG. pLKO.1 c-Myc shRNA plasmids were co- transfected with packaging plasmid and envelope plasmid into 293T cells to produce lentiviral particle.

2.6. Cell cytotoxic assay

For the viability experiments, cells were seeded in 96-well plates at 30% conﬂuency and exposed to chemicals the next day. After 72 h, WST-1 reagent was added, and absorbance at 450 nm was measured by using a Spectramax spectrophotometer (Molec- ular Devices, US) in accordance with the manufacturer’s in- structions. The IC50 values were calculated by using GraphPad Prism version 5 for Windows. The curves were ﬁtted using a nonlinear regression model with a log (inhibitor) versus response formula.

2.7. In vivo xenograft

Female athymic BALB/c (nu/nu) mice (6 weeks old) were ob- tained from Charles River of Japan. Animals were maintained under clean room conditions in sterile ﬁlter top cages and housed on high efﬁciency particulate air-ﬁltered ventilated racks. Animals received sterile rodent chow and water ad libitum. All of the procedures were conducted in accordance with guidelines approved by the Laboratory Animal Care and Use Committee of Korea Research Institute of Chemical Technology. Ty82 cells (5 106 in 100 mL) were implanted subcutaneously (s.c.) into the right ﬂank region of each mouse and allowed to grow to the designated size. Once tu- mors reached an average volume of 200 mm3, mice were ran- domized and dosed via oral gavage daily with the indicated doses of compounds for 14 days. Mice were observed daily throughout the treatment period for signs of morbidity/mortality. Tumors were measured twice weekly using calipers, and volume was calculated using the formula: length width2 0.5. Body weight was also assessed twice weekly.

3. Results and discussion

3.1. Mid-throughput screening for bromodomain inhibitor

Currently, several compounds are undergoing clinical trials for the treatment of patients with hematological and solid tumors [21]. To develop a potent bromodomain inhibitor, we performed mid- throughput screening (MTS) to identify a new chemical scaffold. MTS is likely to result in many false-positive hits [22]. These false- positives make research team to exhaust their time and huge ef- forts, thereby impeding research progress. To remove false-positive hits, we used two orthogonal enzyme assays, alpha-screen and ELISA. False-positives are caused mainly by the interaction of compounds with components of the assay system [23]. Because their assay principles of alpha-screen and ELISA are totally different, we therefore deﬁned a criterion that the compound should be selected as “TRUE” hit only when it exhibited inhibition in both assay systems (Fig. 1A). We performed MTS on approxi- mately 8000 compounds, which represent unique scaffolds, out of 340,000 compounds in the Korea Chemical Bank (KCB). For primary screening, we have done alpha-screen assay. From this assay, 72 compounds were selected as potential hit compounds and sub- jected to secondary screening by ELISA. In ELISA, only one com- pound showed inhibition against the Brd4 bromodomain, while remaining 71 compounds did not. The structure of the hit com- pound, HIT-A, is shown in Fig. 1B. This compound has a unique scaffold, benzo [6,7]oxepino[4,3,2-cd]isoindol-2(1H)-one. The IC50 of HIT-A was 3.27 mM in alpha-screen, and 1.03 mM in ELISA (Fig. 1C). To validate this compound as a “TRUE” hit, we assessed whether this compound down-regulated c-Myc expression in NUT midline carcinoma cells, Ty82 and 10e15, which have brd4-nut fusion gene. Fig. 1D shows that this compound down-regulated c- Myc expression in a similar manner to other known bromodomain inhibitors, JQ1 ( ), I-BET-151, and I-BET-762. Overall, we identiﬁed one novel bromodomain inhibitor out of 8000 compounds.

3.2. A natural compound, aristoyagonine, is identiﬁed as a new bromodomain inhibitor

Among the 340,000 compounds deposited in KCB, we found 11 compounds with a similar structure to the hit compound (Fig. 1E). Of these 11 compounds, 1A, which has methoxy groups at R1, R2, and R3, exhibited the best efﬁcacy both in bromodomain enzyme assay and in the cell cytotoxicity assay. Compound 1A, with the chemical name 5,7,8-trimethoxy-1-methylbenzo [6,7]oxepino [4,3,2-cd]isoindol-2(1H)-one, is known as aristoyagonine. Aris- toyagonine was isolated from Sarcocapnos enneaphylla plants in 1984, and is the only product in this plant to contain a ﬁve- membered lactam ring [24]. It is a member of the diverse aristo- cularine alkaloid family, which was reported to exert cytotoxic ef- fects against several cancer cell lines [25]. Our further study therefore focused on compound 1A, aristoyagonine.
We investigated on the selectivity of aristoyagonine toward bromodomain family members. The data show that aristoyagonine at 10 mM inhibited several bromodomain-containing proteins including all BET proteins (Brd2, Brd3, Brd4, and Brdt), Baz2A, Baz2B, Brd1, Brd7, Brd9, BRPF1, BRPF3, CECR2, FALZ, TAF1 (2),PBRM1 (5), and TRIM24 (Fig. 2A). Although I-BET-762 is highly speciﬁc to BET proteins, aristoyagonine shows a broad inhibitory spectrum against bromodomain-containing proteins. In addition, aristoyagonine exerted no inhibition in the hERG patch clamp assay (Fig. 2B). To determine the pharmaco-kinetics of aristoyagonine, pharmacokinetic (PK) assay was conducted (Fig. 2C). It shows reasonable PK proﬁle that AUC is 0.13 mg*hr/ml on 10 mpk PO. Based on these data, we decided to test this compound in an in vivo mouse model.

Fig. 1. Discovery of a new bromodomain inhibitor through mid-throughput screening. (A) Workﬂow for bromodomain inhibitor screening in this study. (B) The structure of the hit compound, HIT-A, identiﬁed in mid-throughput screening. (C) The inhibitory activity of HIT-A in alpha-screen assay and ELISA. (D) Ty-82 and 10e15 cells were treated with 2 mM of each bromodomain inhibitor for 6 h, and cell lysates were collected for western blot. (E) Structure-activity relationship (SAR) of HIT-A derivatives.

Fig. 2. Investigation of aristoyagonine activity, speciﬁcity, and pharmacokinetics. (A) Bromodomain panel assay with aristoyagonine at 10 mM (B) hERG patch clamp assay with aristoyagonine. (C) Pharmaco-kinetics of aristoyagonine.

3.3. Aristoyagonine shows anti-tumor efﬁcacy in Ty82-xenograft assay

To determine whether aristoyagonine exhibits tumor growth inhibition, we conducted an in vivo xenograft assay. We generated a Ty82 xenograft mouse model as follows; Ty82 cells were implanted into nude mice and allowed to grow to 100 mm3 or 200 mm3 in size. Subsequently, aristoyagonine was administered orally at daily doses of 100 mpk, and tumor volume was measured for 32 days. As shown in Fig. 3, aristoyagonine exhibited tumor growth inhibition. No weight loss was observed in the mice administered the new bromodomain inhibitors. (data not shown) These results suggest that aristoyagonine is a potent bromodomain inhibitor with a unique scaffold that was effective both in vitro and in vivo.

3.4. The effect of aristoyagonine against gastric cancer cell lines

We decided to compare aristoyagonine with I-BET-762 in gastric cancer cells. There have been many reports on the effect of the bromodomain inhibitor in hematological cancer and solid tumors including prostate cancer, breast cancer, and lung cancer, whereas few studies have focused on gastric cancer [26,27]. First, we knockdowned c-Myc expression by using c-Myc shRNA (Supplementary Fig. 1). As c-Myc is known to be a key player in cancer cell proliferation, we thought that all the gastric cancer cells would die after c-Myc knockdown. However, among 9 gastric cancer cell lines tested, the growth of only 5 cell lines, MKN1, MKN28, MKN74, SNU638, and SNU719, was suppressed signiﬁ- cantly, while the growth of other 3 cell lines, MKN7, MKN45, and SNU668, was n’t suppressed at all, and the growth of SNU-216 was suppressed only slightly (Supplementary Fig. 2). Next, we investi- gated the c-Myc down-regulation by I-BET-762 (Supplementary Fig. 3). Since c-Myc expression is known to be determined by bromodomain activity, we postulated that I-BET-762 causes c-Myc down-regulation in all the gastric cancer cell lines. However, I-BET- 762 induced down-regulation of c-Myc expression only in 5 cell lines out of 9 tested gastric cancer cell lines. Surprisingly, these 5 cell lines are exactly same as the cell lines in which c-Myc knockdown caused cell death. It means that I-BET-762 results in c- Myc down-regulation in cell lines in which c-Myc plays a key role in cell proliferation. We currently have no information to explain this phenomenon at this moment. Anyway, from these data, we antic- ipated that Brd4 inhibitors might cause cell death only in 5 cell lines, MKN1, MKN28, MKN74, SNU638, and SNU719. To prove this, a cytotoxicity assay in 9 gastric cancer cell lines was performed with I-BET-762, aristoyagonine, and one negative control compound, 11A (Fig. 4A). As expected, 11A was not cytotoxic to any cell lines at all, and I-BET-762 showed cytotoxic effect only in 5 cell lines in which c-Myc expression was down-regulated by I-BET-762. Unexpectedly, aristoyagonine showed cytotoxic effect in all 9 gastric cancer cell lines. Even in the cell lines in which I-BET-762 was not effective and c-Myc knockdown didn’t cause cell death, aristoyagonine showed cytotoxic effects. As 11A, which has no inhibitory activity against bromodomain at all and is similar to aristoyagonine in structure, was not cytotoxic to any gastric cancer cell line at all, the cytotoxic effect of aristoyagonine is thought to be a result of its inhibitory effects on the bromodomain. Western blot showed that aristoya- gonine suppressed c-Myc expression in SNU638 like I-BET-762 (Fig. 4B). However, c-Myc expression was not suppressed by aris- toyagonine in MKN45, which is resistant to I-BET-762, although aristoyagonine inhibited MKN45 cell proliferation, as shown in Fig. 4A. Therefore, aristoyagonine is likely to have exerted cyto- toxicity through a different mechanism other than c-Myc down- regulation in I-BET-762-resistant cells. In the bromodomain panel assay, aristoyagonine showed inhibition against a wider range of bromodomains compared with I-BET-762 (Fig. 2A). We anticipate that this makes aristoyagonine cytotoxic to a wider range of cancer cell lines compared to I-BET-762.

Fig. 3. In vivo xenograft assay with aristoyagonine. Ty82 cells were implanted into mice and allowed to grow to the designated size. Vehicle or aristoyagonine was orally administered daily at doses of 100 mpk. Tumor sizes were measured twice per week throughout the treatment period by using calipers. A signiﬁcant difference, *P < 0.05 (Student's t-test), was observed between vehicle-treated mice and aristoyagonie- treated mice. Fig. 4. Aristoyagonine suppresses the growth of gastric cancer cells. (A) Gastric cancer cell lines were treated with bromodomain inhibitors at the indicated concentrations, and cell proliferation was measured after 3 days. (B) SNU-638 and MKN-45 cell lines were treated with aristoyagonine for 24 h at indicated concentrations and c-Myc level was measured by western blot. 3.5. Summary In this study, for the ﬁrst time, we introduce a natural product as a Brd4 bromodomain inhibitor. Aristoyagonine showed potent inhibitory activities in both in vitro and in vivo system. It is expected that a new bromodomain inhibitor will be developed based on the structural scaffold of aristoyagonine.