Bay 61-3606 Sensitizes TRAIL-Induced Apoptosis by Downregulating Mcl-1 in Breast Cancer Cells

Breast cancer cells generally develop resistance to TNF-Related Apoptosis-Inducing Ligand (TRAIL) and, therefore, assistance from sensitizers is required. In our study, we have demonstrated that Spleen tyrosine kinase (Syk) inhibitor Bay 61–3606 was identified as a TRAIL sensitizer. Amplification of TRAIL-induced apoptosis by Bay 61–3606 was accompanied by the strong activation of Bak, caspases, and DNA fragmentation. In mechanism of action, Bay 61–3606 sensitized cells to TRAIL via two mechanisms regulating myeloid cell leukemia sequence-1 (Mcl-1). First, Bay 61–3606 triggered ubiquitin-dependent degradation of Mcl-1 by regulating Mcl-1 phosphorylation. Second, Bay 61–3606 downregulates Mcl-1 expression at the transcription level. In this context, Bay 61–3606 acted as an inhibitor of Cyclin-Dependent Kinase (CDK) 9 rather than Syk. In summary, Bay 61–3606 downregulates Mcl-1 expression in breast cancer cells and sensitizes cancer cells to TRAIL-mediated apoptosis.


Introduction
TRAIL/Apo2 ligand selectively kills cancer cells by initiating apoptotic signaling through the engagement of its pro-apoptotic receptors, Death Receptors-4 and -5 [1,2]. TRAIL binding to these receptors results in the formation of 'death-inducing signaling complex (DISC)' inducing caspase-8 activation [3]. After DISC-mediated activation of caspase-8, the intrinsic apoptotic pathway is activated by processing of Bid and its translocation to mitochondria. The joint effort of the extrinsic and intrinsic apoptotic pathways then leads to the activation of downstream caspases (-3 and -7) and apoptotic demise of cells [4].
Although TRAIL shows cancer-selective killing activity, a phase 2 clinical trial failed to demonstrate a clear benefit in a therapeutic window [5]. Parallel to this result, primary tumors were found to be resistant against TRAIL-induced apoptosis. Resistance to TRAIL is partially explained by decoy receptors (DcR1 and DcR2), which have a deleted or truncated death domain [6]. Other defects of cell death pathways, such as dysregulated expression of anti-apoptotic proteins and pro-apoptotic proteins, were identified as mechanisms of resistance [4,7]. However, new biomarkers and molecular targets of TRAIL resistance are still needed for its potential future clinical use.
Myeloid cell leukemia sequence-1 (Mcl-1) is a member of the anti-apoptotic Bcl-2 family proteins that neutralizes pro-apoptotic Bcl-2 proteins such as Bim, Bid, and Bad [8]. The important role of Mcl-1 in TRAIL-mediated cell death has been suggested in a number of published studies. Knockdown of the Mcl-1 gene enhances the apoptotic events induced by TRAIL [9,10]. A recent study of several TRAIL sensitizers revealed that they function via downregulation of Mcl-1 [11][12][13][14].
Cyclin-Dependent Kinases (CDKs) are a group of protein serine/threonine kinases which is activated by specific cyclin co-factors. Multiple CDKs regulate the cell cycle progression and control the cell death [15]. In fact, several CDK inhibitors, i.e. R-roscovitine, CR8, flavopiridol, and CDKI-73 induce Mcl-1 downregulation and thus promote the induction of apoptosis [16][17][18][19]. However, the study that molecular mechanisms and practical approaches downregulate Mcl-1efficiently and safely must still be further clarified. In

Compound Screening and DNA Fragmentation
High throughput, TRAIL-sensitizer screening and DNA fragmentation assay were performed as described previously [12].

Western Blot Analysis
Cell extracts were prepared by scraping cells with lysis buffer containing complete protease inhibitor cocktail (Roche, Basel, Switzerland). After protein concentration measurement using the Bio-Rad DC protein assay (Bio-Rad, Veenendaal, Netherlands), extracts were diluted with 2X Laemmli sample buffer. Cell extract (30 μg) was resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) followed by Western blotting using polyvinylidene difluoride membranes (Millipore, Billerica, MA, USA). The membranes were incubated in TRIS-buffered saline with specific antibodies (1:1000 dilutions). The antibody-protein complex was detected with horseradish peroxidase-conjugated IgG (Bio-Rad) and a chemiluminescent substrate solution (SuperSignal West Pico; Pierce, Rockford, IL). Stripped membranes were then re-probed with antibody to α-tubulin as a loading control.

Cell Viability and Immunocytochemistry
For the quantitative assay of cell survival, a cellular, ATP-based, luminescent kit (CellTiter-Glo; Promega, Madison, WI, USA) was used according to the manufacturer's protocol. The raw luminescence value was measured and the relative cell survival was calculated using the GraphPad Prism software (La Jolla, CA, USA). For immunocytochemistry, an antibody raised against active Bak was used to detect apoptosis. Briefly, cells were seeded on a chamber slide. After treatment, cells were fixed with 4% paraformaldehyde, then permeabilized with 0.5% Triton X-100 in PBS, and blocked with 10% goat serum diluted in PBS. The slide was incubated with primary antibody and secondary antibody with washing steps. After mounting, the sample image was visualized under fluorescence microscopy (Olympus LX71; Tokyo, Japan). [20], pcDNA3-CDK9-HA, pcDNA3-CDK9 (DN)-HA, pGL2-basic, and pGL2-Mcl-1-luc promoter [11] plasmids were acquired from Addgene (Cambridge, MA, USA). One day before transfection, cells were plated at a density of 5 × 10 5 cells in 6-well plates. Cells were transfected with expression plasmids using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA). For luciferase assays, MCF-7 cells were co-transfected with plasmids containing luciferase reporter genes (pGL2-Mcl-1-luc and pRL-TK-luc) using Lipofectamine 2000 for 48 h. Cells were pre-incubated with Bay 61-3606 (2.5 μM) for 1 h and were then exposed to TRAIL (50 ng/ml) for an additional 6 h. After treatment, cells were lysed with lysis buffer and aliquots of the lysates were used to analyze the transcription level of Mcl-1 using the Dual-Luciferase Reporter Assay kit (Promega, Madison, WI, USA).

Reverse Transcription Polymerase Chain Reaction (RT-PCR)
Total RNA of test cell was isolated by the TRI reagent (Molecular Research Center, Cincinnati, OH, USA), and the cDNA was prepared using RevertAid First Strand cDNA synthesis kit (Life Technologies, Carlsbad, CA, USA). PCR was performed with specific primers: Mcl-1 (sense)

In Vivo Xenograft Model Assay
Female BALB/c nude mice (5 weeks old) were used (Japan SLC Inc., Shizuoka, Japan) to generate a tumor xenograft model. For each mouse, a β-estradiol 17-valerate pellet (0.5 mg, 60 days release, Innovative Research of America) was subcutaneously implanted into the neck region to support tumor growth following the implantation of 1 × 10 6 MCF-7 cells subcutaneously into the hind limb. Formation of the tumor and condition of animal were monitored every other day. When the tumor volume reached~100 mm 3 , the mice were divided into four treatment groups (n = 5 / group), i.e. untreated, TRAIL (10 mg/kg), Bay 61-3606 (50 mg/kg), and a combination of Bay 61-3606 and TRAIL. TRAIL with or without Bay 61-3606 was administered twice a week for two weeks by intraperitoneal injection. TRAIL was given 2 h after the injection of Bay 61-3606. The largest longitudinal (length) and transverse (width) tumor diameters were measured using a vernier caliper, and the tumor volume was calculated with the formula: tumor volume = 1/2 (length × width 2 ).

Ethics Statement
All animal experiments were performed under the guide protocol approved by the Institutional Animal Care and Use Committee of the ASAN Institute for Life Sciences and performed in accordance with the institution guidelines. All efforts were made to minimize suffering. For the implantation procedure, animals were anaesthetized by intraperitoneal injection of 1.2% Tribromoethanol at 125~250 mg/kg and recovered on heating pad. Animals were euthanized by CO 2 inhalation in a prefilled CO 2 chamber at the end of designated experiment, or to alleviate untreatable pain or distress such as tumor burden exceeding 10% of body weight, weight loss greater than 20% and other severe clinical signs.

Statistical Analysis
All the experiments were independently repeated at least two or more times, data were presented as the mean ± SD using a Two-way ANOVA with software by GraphPad Prism 5.0., and P-values < 0.05 were considered statistically significant.

Bay 61-3606 Sensitizes MCF-7 Cells to TRAIL-Induced Apoptosis
To identify new sensitizers of TRAIL, we performed a compound screen using a library of 1,280 functionally characterized small molecules (LOPAC-1280) in MCF-7 cells. Eight candidates listed in S1 Table could sensitize MCF-7 cells to a subtoxic dose of recombinant TRAIL. Of these, Bay 61-3606 sensitized cells to death induced by TRAIL while exposure to TRAIL alone showed minimal reduction (Fig 1A, bars in far right). Combination treatment induced synergistic cell death in a concentration-dependent manner ( Fig 1B). To characterize the type of death, cells were subjected to immunocytochemistry analysis (Fig 1C), caspase activity assay (Fig 1D), and DNA fragmentation assay ( Fig 1E). Upon combination treatment, Bak and caspases-7 & -8 were activated and DNA was potently fragmented. All of the qualitative cell death assays revealed that Bay 61-3606 sensitized MCF-7 cells to TRAIL-induced cell death with apoptotic characteristics. Moreover, treatment with a variety of cell death-inducing agents, Bay 61-3606 is specific for the death receptor-mediated cell death signaling, i.e. TNF-α (S1A

Bay 61-3606 Sensitizes MCF-7 Cells to TRAIL-Induced Apoptosis by Mcl-1 Down-Regulation
We then analyzed the molecular mechanisms of the sensitization. The combination of Bay 61-3606 and TRAIL massively induced caspase-7 & -8 activation and cleavage of Bid, and PARP (Fig 2A). As shown in Fig 2A and 2B, there was no significant alteration in the expression levels of DR4, DR5, p53, and Bcl-2 family proteins, Bad, Bax, Bak, and Bcl-xL in MCF-7 cells. However, we could detect significantly Mcl-1 downregulation by Bay 61-3606 ( Fig 2B) and Mcl-1 downregulation were concentration-and time-dependent ( Fig 2C). Also, FLICE-inhibitory protein (FLIP) and Inhibitors of Apoptosis Proteins (IAP) family proteins, such as cIAP1, X-linked inhibitor of apoptosis protein (XIAP), and Survivin were downregulated by combination of Bay 61-3606 and TRAIL (Fig 2A and 2B). To examine the contribution of Mcl-1 to TRAIL resistance, the loss of function effect of Mcl-1 was tested. As shown in Fig 2D,
The association of Syk with Mcl-1 expression was also tested in other breast cancer cells with different genetic backgrounds. The mRNA and protein level of Syk were detected in both of MCF-7 and T47D cells, and phosphorylation of Syk was reduced by Bay 61-3606 in both cells. But MDA-MB-231 cell was Syk negative in mRNA and protein expression ( Fig 3C). As shown in Fig 3C, we re-noted that Mcl-1 protein was reduced in MCF-7 cells by Bay 61-3606,  (Fig 4C). However, GSK3 phosphorylation was not modulated by Bay 61-3606 (Fig 4D). These results show that Bay 61-3606 induces the degradation of Mcl-1 by phosphorylation at Ser159 independently from GSK3.
In addition, phosphorylation of Thr163 in the Mcl-1 PEST domain, which is an Mcl-1 stabilizing event, is mediated by ERK [29][30][31]. As shown in Fig 4D, (Fig 5C). This result is comparable with knockdown of CDK9 induces Mcl-1 downregulation by RNA pol II inhibition (Fig 5B), and thus implying that Bay 61-3606 has a negative role on CDK9 and RNA pol II activity.
Moreover, two CDK9 specific inhibitors significantly inhibited simultaneous phosphorylation of CDK9 and RNA pol II and they induced Mcl-1 downregulation in a dose-dependent manner as Bay 61-3606 (S4 Fig). We found that downregulation of Mcl-1 by Bay 61-3606 was reversed by wild type CDK9 (WT CDK9) overexpression (Fig 5D left). Also WT CDK9 reversed the decreased phosphorylation of RNA pol II by Bay 61-3606. In contrast to our WT CDK9 and the kinase-inactive, CDK9 dominant-negative mutant (DN CDK9) decreased RNA

Anti-Tumor Activity of TRAIL with Bay 61-3606 In Vivo
The in vivo anti-tumor activity of Bay 61-3606 was assessed by treating MCF-7 tumor xenograft-bearing BALB/c nude mice with a combination of Bay 61-3606 and TRAIL (Fig 6A). Interestingly, the TRAIL single agent led to slight reductions in tumor growth and the administration of Bay 61-3606 alone (P <0.05) was more efficacious than that of TRAIL alone. After  (Fig 6B). These results suggest that combination treatment with Bay 61-3606 and TRAIL suppresses tumor growth in vivo.

Discussion
This study examined whether Bay 61-3606 sensitizes cells to TRAIL-induced apoptosis and if so, clarified the detailed molecular mechanisms in human breast cancer cells. Our results provide a hypothetical mechanism regarding how Bay 61-3606 downregulates the Mcl-1 protein level and gene transcription (Fig 6C). First, we have shown that Bay 61-3606 induces Mcl-1 protein degradation by phosphorylating Ser159. The phosphorylated Mcl-1 at Ser159 interacts with ubiquitin-ligase and promotes proteasomal degradation [32][33][34]. We could also detect   (Fig 4) [28,31,32]. This suggests that other upstream kinases regulate Mcl-1 stabilization and degradation by multiple alternative routes.
Secondly, we found that Bay 61-3606 inhibits Mcl-1 gene transcription by inhibiting CDK9-mediated RNA pol II activity. To address the effect of Bay 61-3606 on Mcl-1, we suspected that Syk plays essential role to regulate Mcl-1 in cells because Bay 61-3606 was originally developed for the inhibitor of Syk [21]. However, we have determined that Syk is not involved in the Bay 61-3606-mediated Mcl-1 downregulation (Fig 3). This result is comparable with that of previous reports indicating that Syk is a tumor suppressor in breast cancer cells [35]. Meanwhile, we noticed that Bay 61-3606 interacts with CDK9 at the ATP binding site [23] and inhibits CDK9 kinase activity. In our data, we could show that CDK9 inhibition results in RNA pol II dephosphorylation at Ser2 and the downregulation of Mcl-1. The combination of Bay 61-3606 and TRAIL significantly decreased the phosphorylation of CDK9 and RNA pol II. Moreover, knockdown of CDK9, specific CDK9 inhibitors, and kinase inactivation of CDK9 (Fig 5B-5D and S4 Fig) can strongly suppress RNA pol II-mediated Mcl-1 transcription and can decrease Mcl-1 expression similarly to that of Bay 61-3606. In addition, the direct inhibition of CDK9 with RNA pol II inactivation by Bay 61-3606 has been suggested ( Fig 5C) and CDK9 is the main target of Bay 61-3606 to control Mcl-1. Our results are consistent with those of previous reports regarding roscovitine [17] and flavopiridol, both of which are inhibitors of CDK9 [18]. It has recently been shown that a novel CDK9 inhibitor inhibits RNA pol II phosphorylation and Mcl-1 transcription, thus showing strong anti-cancer therapeutic activity [19]. All this taken together, CDK9 is an important target kinase that Because each of these two phosphorylation events represents opposite results for Mcl-1 (degradation vs stabilization), it can be expected that there is a more delicate mechanism to regulate the balance between stable and fragile Mcl-1 over simple phosphorylation induction. Lastly, as Fig 2B shows, the anti-apoptotic proteins including Flip and IAP family proteins were similarly downregulated by Bay 61-3606 requesting more detail mechanistic investigation.
In conclusion, we suggest that Bay 61-3606 is a promising agent for downregulating Mcl-1 expression in cancer cells. Our results also strongly suggest that Bay 61-3606 reduces Mcl-1 mRNA transcription by inhibition of CDK9-dependent RNA pol II activation and induces rapid protein degradation via the UPS. Based on these results, specific CDK9 inhibitors and identification of a new Bay 61-3606 target for Mcl-1 degradation may be beneficial for the treatment of cancers in which Mcl-1 contributes to the development of resistance to anticancer therapeutics.