β-TrCP1 degradation is a novel action mechanism of PI3K/mTOR inhibitors in triple-negative breast cancer cells

An F-box protein, β-TrCP recognizes substrate proteins and destabilizes them through ubiquitin-dependent proteolysis. It regulates the stability of diverse proteins and functions as either a tumor suppressor or an oncogene. Although the regulation by β-TrCP has been widely studied, the regulation of β-TrCP itself is not well understood yet. In this study, we found that the level of β-TrCP1 is downregulated by various protein kinase inhibitors in triple-negative breast cancer (TNBC) cells. A PI3K/mTOR inhibitor PI-103 reduced the level of β-TrCP1 in a wide range of TNBC cells in a proteasome-dependent manner. Concomitantly, the levels of c-Myc and cyclin E were also downregulated by PI-103. PI-103 reduced the phosphorylation of β-TrCP1 prior to its degradation. In addition, knockdown of β-TrCP1 inhibited the proliferation of TNBC cells. We further identified that pharmacological inhibition of mTORC2 was sufficient to reduce the β-TrCP1 and c-Myc levels. These results suggest that mTORC2 regulates the stability of β-TrCP1 in TNBC cells and targeting β-TrCP1 is a potential approach to treat human TNBC.


INTRODUCTION
Triple-negative breast cancers (TNBCs), which were first introduced in the medical literature in 2005, are a heterogeneous group of tumors that are immunohistologically defined as the lack of estrogen receptor (ER) and progesterone receptor (PR) expression, as well as human epidermal growth factor receptor 2 expression/amplification. 1 Despite marked increase of studies on TNBCs during the past decade, our knowledge of how TNBCs can be treated is still limited. [2][3][4] Approximately 15 to 20% of all breast cancers are diagnosed as TNBCs. 4 A systemic review demonstrated the highest incidence of TNBCs in women of African ancestry (26.99%) followed by Hispanic (17.5%), Asian (12.19%), Caucasian (11.73%) and other women (8.42%). 5 A recent meta-analysis of large data sets revealed that TNBCs are classified in at least six distinct molecular subtypes that include two basal-like, an immunomodulatory, a mesenchymal, a mesenchymal stem-like and a luminal androgen receptor subtype. 6 However, no successful therapeutic target is currently available to treat TNBC patients. [2][3][4] Beta-transducin repeat containing proteins (β-TrCPs) are members of the F-box/WD repeat-containing protein (FBXW) subfamily of F-box protein families. [7][8][9] As an F-box protein, the β-TrCP is the substrate-recognition subunit of SKP1-cullin 1-F-box protein, E3 ligase complexes and well conserved across species. 8,9 In humans, β-TrCP exists as two homologues, β-TrCP1 (also known as FBXW1) and β-TrCP2 (also known as FBXW11), which are encoded by two distinct genes but share extensive amino acid sequence homology. The differences between these two proteins still remain elusive. 8,9 The function of β-TrCPs in tumorigenesis is either oncogenic or tumor-suppressive in a tissue-specific or cellular contextdependent manner homology. 8,9 Although it has been widely studied that β-TrCP recognizes diverse proteins and regulates their stability, the regulation of β-TrCP itself is not yet understood.
Here, we demonstrated that the expression of β-TrCP1 protein is regulated by mTORC2 and targeting β-TrCP1 is a potential therapeutic approach to treat TNBC cells.

Western blots and antibodies
Western blot analyses were performed as described previously. 10 Antibodies used in this study were as follows: phospho-AKT Immunoprecipitation Immunoprecipitation was performed as described previously 13 with phospho-(Ser/Thr) Phe (#9631) antibody. Then, immune complexes were dissolved on SDS-poly acrylamide gel electrophoresis and western blot analysis was performed with mouse β-TrCP antibody (sc-390629) from Santa Cruz.

Transfection of small interference RNA and cell proliferation assay
Transfection of small interference RNA (siRNA) was performed with Lipofectamine 2000 (Invitrogen) as described previously. 14

A PI3K/mTOR inhibitor PI-103 reduces the viable cells of TNBC cell lines
During our previous study, 10 we noticed that a PI3K/mTOR inhibitor PI-103 15 reduced the cell viability of TNBC cells in a dose-dependent manner. As TNBC is a heterogeneous disease which can be further subgrouped into six subtypes, 4 we further tested the effect of PI-103 with expanded TNBC cell lines by MTT assay. As shown in Figure 1a, PI-103 significantly reduced the number of viable cells in a dose-dependent manner in all of the cell lines tested. There is no significant correlation between the EC 50 values for PI-103 and the subtypes of these cell lines (Figure 1b). To further compare the level of proteins in the PI3K/mTOR pathway in these cells, we performed western blot analysis ( Figure 1c). All these cells expressed detectable levels of AKT downstream proteins such as phospho-mTOR (Y2448) and p-S6 (S235/S236). As reported previously, 16 BRCA1defective cells exhibited higher level of phospho-AKT (S473) than other TNBC cells. On the contrary, the expression of phospho-AKT (S473) was barely detectable in MDA-MB-231 cells. Interestingly, significant level of β-TrCP1 protein was detected in all these cells. There is no significant correlation between the EC 50 values of PI-103 and the expression levels of these proteins, whereas high ratio of phospho-AKT to β-TrCP1 correlated with high EC 50 values for PI-103 ( Figure 1b).
To determine the effect of PI-103 on the level of these proteins, we performed western blot analysis with cell lysates   (Figure 1d). As expected, PI-103 reduced the phosphorylation level of AKT substrates including mTOR (S2448) 17 and GSK3β (S9) 18 in all cell lines test. PI-103 also reduced the level of mTORC2 substrate, phospho-AKT (S473). 19 Unexpectedly, the level of β-TrCP1 was also reduced by PI-103 in these cells (Figure 1d).

PI-103 reduces the level of β-TrCP, c-Myc and cyclin E in TNBC cells
The effect of PI-103 on the level of β-TrCP1 was further determined by western blot analyses. Three TNBC cell lines were treated with increasing concentrations of PI-103 for 24 h and western blot analyses were performed. Consistently, the level of β-TrCP1 was reduced by PI-103 in a dose-dependent manner in all these cells (Figure 2a). It has been reported that β-TrCP stabilizes c-Myc through antagonizing FBW7-mediated turnover. 20 It has been also demonstrated that activation of Ras/PI3K/ERK pathway induces c-Myc stabilization in melanoma cells, 21 and inhibition of PI3K/AKT pathway by LY294002 in melanoma cells, 21 or by PI-103 in Burkitt's lymphoma cells, 22 and reduces c-Myc expression. Consistently, the levels of c-Myc and its downstream target, cyclin E, were also reduced by PI-103 in a dose-dependent manner (Figure 2b).
To determine whether the reduction of β-TrCP1 is dependent on proteasome-mediated degradation, western blot analysis was conducted as follows: HS578T and MDA-MB-231 cells were treated with PI-103 for 24 h. To inhibit  proteasome-dependent proteolysis, MG132 was added to the cells 4 h before harvest. Western blot analysis revealed that PI-103-mediated reduction of β-TrCP1 was reversed by MG132 in these cells (Figure 3a). In addition, MG132 treatment increased the basal level of β-TrCP1 in the absence of PI-103 (Figure 3a). Although it has been reported that PI-103 induces autophagy in glioma cells, 17 we further accessed the effect of bafilomycin A1, an inhibitor of autophagic vacuole maturation 23 on the degradation of β-TrCP1 by PI-103. MDA-MB-231 cells were treated with PI-103 for 24 h and either bafilomycin A1 or MG132 was treated for 4 h before harvest. As expected, bafilomycin A1 increased the accumulation of LC3B-II, a marker of autophagy, in the absence of PI-103 (Figure 3b). However, bafilomycin A1 did not affect the PI-103-mediated degradation of β-TrCP1. On the contrary, MG132 reversed the effect of PI-103 on the level of β-TrCP1 (Figure 3b, lanes 8 and  12-14) without affecting the accumulation of LC3B-II in the absence of PI-103 (Figure 3b, lanes 7 and 9-11). In these conditions, the level of β-TrCP1 was increased by MG132 even in the absence of PI-103. These implicate the basal level of β-TrCP1 was regulated by MG132 in these cells. All these results suggest that PI-103 reduced the level of β-TrCP1 by a proteasome-mediated degradation. The degradation of β-TrCP1 was observed as early as 4 h after PI-103 treatment (Figure 3c). Although we cannot completely exclude the possibility of mRNA reduction by PI-103 treatment, rapid reduction of the level of β-TrCP1 in the time-course treatment of β-TrCP1 (Figure 3c) suggests that proteasomal degradation is a major mechanism of β-TrCP1 reduction by PI-103.

PI-103 reduces the phosphorylation of β-TrCP1
The β-TrCP protein has several potential phosphorylation sites (data not shown). To determine whether β-TrCP1 is  phosphorylated or not, we performed western blot analysis after immunoprecipitation of phospho proteins. To avoid β-TrCP1 degradation, MDA-MB-231 cells were treated with 10 μM of PI-103 for 2 h and phospho-proteins were immuneprecipitated by phospho-Ser/Thr Phe antibody. Then the immune complexes were analyzed by western blot with β-TrCP1 antibody. As shown in Figure 4, β-TrCP1 was phosphorylated in MDA-MB-231 cells and brief treatment of PI-103 reduced this phosphorylation. As a control, the level of mTOR protein was also determined. As expected the phospho-mTOR (S2448) was reduced by PI-103 treatment without affecting the level of mTOR.
Knockdown of β-TrCP1 reduces the proliferation of TNBC cells As PI-103 reduced the viable cell numbers and the level of β-TrCP1 in a broad range of TNBC cells, we questioned whether β-TrCP1 has a role in the proliferation of TNBC cells. To address this question, we performed knockdown of β-TrCP1 by siRNA and determined its effect on the proliferation of TNBC cells. HS578T and MDA-MB-231 cells were transfected with specific siRNAs for β-TrCP1 and for β-TrCP2 as a control and western blot analysis was performed. As shown in Figure 5a, β-TrCP1-siRNAs reduced the level of β-TrCP1.
Next, we determined the number of viable cells at various time points after β-TrCP1 knockdown in HS578T and MDA-MB-231 cells. Interestingly, knockdown of β-TrCP1 was enough to reduce the proliferation of these cells (Figure 5b). The anti-proliferative effect was evident at 3 or 4 days after β-TrCP1 knockdown and sustained up to 5 days. Western blot analysis revealed that the level of β-TrCP1 was slightly reversed over the time in HS578T cells, whereas no significant restoration of β-TrCP1 was observed in MDA-MB-231 cells. Consistent with these, the anti-proliferative effect of β-TrCP1 knockdown was more profound in MDA-MB-231 cells.
Pharmacological inhibition of mTORC2 reduces the level of β-TrCP1 protein To identify the kinase responsible for the phosphorylation of β-TrCP1, HS578T and MDA-MB-231 cells were treated with various PI3K/mTOR inhibitors for 24 h. As shown in Figure 6a, most of the PI3K inhibitors reduced the level of β-TrCP1 in these cells. Notably, PI3K inhibitors, which are known to inhibit preferentially PI3Kα, markedly reduced the β-TrCP1 protein, whereas PI3K inhibitors (TGX221, TG100-115 and IC-87114) specific to other isoforms (β, δ and γ) 15,24 exhibited little or no effect on the level of β-TrCP1 protein. More interestingly, all of the PI3K/mTOR dual inhibitors, such as GDC-0941, 25 PI-103, 15 BEZ235 26 and GSK1059615 27 invariantly reduced the level of β-TrCP1 protein (Figure 6a). Again, the level of c-Myc was concomitantly reduced by inhibitors which reduced the level of β-TrCP1.
To further dissect the pathway to β-TrCP1 degradation, MDA-MB-231 cells were treated with different concentrations of more specific inhibitors. As shown in Figure 6b, a PI3K inhibitor PIK-90 and mTOR-kinase inhibitor WYE-354 28 reduced the levels of β-TrCP1 and c-Myc in a dosedependent manner. On the contrary, the level of β-TrCP1 protein was rather induced by other inhibitors including an allosteric AKT inhibitor MK-2206, 29 an allosteric mTORC1 inhibitor rapamycin, 30 and a GSK3α/β inhibitor CHIR-99021. 31 As an mTOR kinase inhibitor, WYE-354 inhibits both mTORC1 and mTORC2, 28 whereas rapamycin more specifically inhibits mTORC1. 30 Taken together, we speculated that mTORC2 is a potential kinase that stabilizes the β-TrCP1 protein in TNBC cells.

DISCUSSION
In the present study, we identified β-TrCP1 as a potential target to treat the TNBC cells. Inhibition of PI3K/mTOR by PI-103 reduced the level of β-TrCP1 in diverse TNBC cell lines in a proteasome-dependent manner. siRNA-based depletion of β-TrCP1 profoundly reduced the proliferation of TNBC cell lines, HS578T and MDA-MB-231. Pharmacological inhibition of mTORC2 but not mTORC1 was sufficient to reduce the level of β-TrCP1 and c-Myc in TNBC cells.
Treatment of a series of specific kinase inhibitors revealed that mTORC2 is a plausible regulator of β-TrCP1 stability in TNBC cells. Treatment of PI-103 reduced the level of phosphoβ-TrCP1. PI3K/mTOR inhibitors reduced the level of β-TrCP1, whereas MK-2206 (an allosteric AKT inhibitor) did not reduce the β-TrCP1 protein in TNBC cells. Because the inhibition of PI3K reduces the activity of AKT that activates mTORC1 activity, these results are somewhat complicated. However, a recent report suggested that the activity of mTORC2 can be directly regulated by PI3K in a TSC-independent manner. 32,33 In addition, WYE-354, an mTOR-kinase inhibitor, reduced the level of β-TrCP1. On the contrary, rapamycin, an mTORC1 inhibitor, did not affect the β-TrCP1 level. All these results suggest that the level of β-TrCP1 protein is regulated by mTORC2 in TNBC cells (Figure 7). Future studies with specific siRNAs for components of PI3K/mTORC2 pathway will further delineate the signaling pathway that is important to regulate β-TrCP1 stability.
The stability of F-box itself is also regulated by posttranslational regulation. As an example, S-phase kinaseassociated protein 2, another oncogenic F-box protein, has been reported to be phosphorylated and stabilized by AKT1 in human cancers. 34 However, the regulation of β-TrCP1 stability through post-translational modification has not been attributed yet. Our present data, which demonstrates that PI-103 reduced the level of phospho-β-TrCP1, suggest that the stability of β-TrCP1 might be also regulated by protein modification. Recently, S-phase kinase-associated protein 2 has been identified as an E3 ligase that is responsible for degradation of β-TrCP. 35 In fact, β-TrCP itself is an E3 ligase for degradation of Emi1, 36,37 an inhibitor of the APC/C E3 ligase which mediates S-phase kinase-associated protein 2 degradation. 38,39 More recently, SIRT1 has been identified as a negative regulator of β-TrCP through protein degradation. 40 Although reported to function as a tissue-specific tumor suppressor, multiple studies have supported that β-TrCP has roles as an oncogene and promotes tumorigenesis when overexpressed and elevated levels of β-TrCP has been reported various human cancers including hepatoblastoma, melanoma, colorectal, pancreatic, prostate, breast and gastric cancers. 8,9 In fact, potential tumor-suppressor function is a common property of many β-TrCP substrates such as inhibitor of nuclear factor-κB (IκB), programmed cell-death protein 4 (PDCD4), DEP domain-containing mTOR-interacting protein (DEPTOR), RE1-silencing transcription factor (REST) and metastasis suppressor 1 (MTSS1). 8,9,41,42 Interestingly, phosphorylation of DEPTOR, a negative regulator of mTORC1/2, generates phosphodegron that binds β-TrCP and leads to degradation of DEPTOR. [43][44][45] In addition, depletion of β-TrCP by shRNA induced accumulation of DEPTOR, reduced mTOR and S6 kinase activity and activated autophagy to reduce cell growth. 44,45 These results suggest that β-TrCP1 is an important candidate of druggable target protein to treat cancers via modulating cell proliferation and/or survival.

CONFLICT OF INTEREST
The authors declare no conflict of interest.  This work is licensed under a Creative Commons Attribution 3.0 Unported License. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http:// creativecommons.org/licenses/by/3.0/