MUC1-C Stabilizes MCL-1 in the Oxidative Stress Response of Triple-Negative Breast Cancer Cells to BCL-2 Inhibitors

Aberrant expression of myeloid cell leukemia-1 (MCL-1) is a major cause of drug resistance in triple-negative breast cancer (TNBC) cells. Mucin 1 (MUC1) is a heterodimeric oncoprotein that is aberrantly overexpressed in most TNBC. The present studies show that targeting the oncogenic MUC1 C-terminal subunit (MUC1-C) in TNBC cells with silencing or pharmacologic inhibition with GO-203 is associated with downregulation of MCL-1 levels. Targeting MUC1-C suppresses the MEK → ERK and PI3K → AKT pathways, and in turn destabilizes MCL-1. The small molecules ABT-737 and ABT-263 target BCL-2, BCL-XL and BCL-w, but not MCL-1. We show that treatment with ABT-737 increases reactive oxygen species and thereby MUC1-C expression. In this way, MUC1-C is upregulated in TNBC cells resistant to ABT-737 or ABT-263. We also demonstrate that MUC1-C is necessary for the resistance-associated increases in MCL-1 levels. Significantly, combining GO-203 with ABT-737 is synergistic in inhibiting survival of parental and drug resistant TNBC cells. These findings indicate that targeting MUC1-C is a potential strategy for reversing MCL-1-mediated resistance in TNBC.

The present studies demonstrate that targeting MUC1-C in TNBC cells suppresses activation of the AKT and ERK pathways, and downregulates MCL-1 expression. In addition and importantly, we show that (i) resistance to ABT-737 and its orally active analogue ABT-263 is associated with increases in MUC1-C, and (ii) MUC1-C drives the upregulation of MCL-1. In concert with these results, we also show that targeting MUC1-C is synergistic with ABT-737 and reverses ABT-737 resistance by MCL-1 suppression.

MUC1-C regulates MCL-1 through the ERK and AKT pathways.
The MUC1-C cytoplasmic domain has been linked to activation of the PI3K/AKT and ERK pathways 23,24,35 (Fig. 2A). In this context, we found that DOX-induced downregulation of MUC1-C in MDA-MB-468 cells is associated with decreases in p-ERK and p-AKT (Fig. 2B). In addition and in concert with AKT-mediated phosphorylation of GSK3β , targeting MUC1-C was associated with decreases in p-GSK3β (Fig. 2B). Similar results were obtained with DOX-treated BT-20/tet-MUC1shRNA cells (Fig. 2C). Consistent with these results, stable overexpression of MUC1-C in MDA-MB-468 (Fig. 2D) and BT-20 ( Fig. 2E) cells increased p-ERK, p-AKT, and p-GSK3β levels, supporting a role for MUC1-C in activating the AKT and ERK pathways in TNBC cells.

MUC1-C and MCL-1 are upregulated in MDA-MB-468 cells resistant to ABT-737 and ABT-263.
In concert with a role for BCL-2 in maintaining redox balance 37 , targeting BCL-2 in leukemia cells with ABT-737 is associated with increases in reactive oxygen species (ROS) 38 . In this context, MUC1 expression is induced in the cellular response to oxidative stress and thereby protects against the induction of apoptosis 39 . Accordingly, we asked if treatment of MDA-MB-468 cells with ABT-737 is associated with ROS-mediated upregulation of MUC1-C expression. The results demonstrate that ABT-737 increases ROS and that this response is attenuated by the antioxidant N-acetylcysteine (NAC) (Fig. 6A). Moreover, ABT-737 treatment was associated with increases in MUC1-C expression by a ROS-dependent mechanism, as evidenced by inhibition with NAC (Fig. 6B). Based on these results, we asked if upregulation of MUC1-C contributes to ABT-737 and/or ABT-263 resistance by selecting MDA-MB-468 cells for growth in the presence of increasing concentrations of these agents. Treatment of parental MDA-MB-468 cells with ABT-737 demonstrated a dose-dependent inhibition of growth (Supplemental Fig. S4A). By contrast and as expected, MDA-MB-468/ABT-737R cells were less sensitive to ABT-737-induced growth inhibition (Supplemental Fig. S4A). Notably, we also found that MDA-MB-468/ MUC1-C cells are less sensitive to ABT-737, consistent with MUC1-C-induced upregulation of MCL-1 (Supplemental Fig. S4A). In addition, we treated MDA-MB-468/ABT-737R and MDA-MB-468/MUC1-C cells with the MCL-1 inhibitor A-1210477 and found decreases in survival, consistent with dependency on MCL-1 (Supplemental Fig. S4B, left and right). Analysis of the MDA-MB-468/ABT-737R and MDA-MB-468/ABT-263R cells further demonstrated increases in both MUC1-C and MCL-1 expression (Fig. 6C,D). Similar results were obtained with BT-20/ABT-737R and BT-20/ABT-263R cells (Supplemental Fig. S4C,D). In assessing whether the MDA-MB-468/ABT-737R cells are sensitive to targeting MUC1-C, we found that GO-203 treatment is associated with downregulation of MCL-1 (Fig. 6E), loss of survival (Fig. 6F) and the induction of apoptosis (Fig. 6G).
Targeting MUC1-C is synergistic with ABT-737. Our results invoked the possibility that targeting

Discussion
MCL-1 is one of the most frequently amplified genes in human cancers and is of importance to the development of resistance to anti-cancer agents [40][41][42] . In breast cancers, overexpression of MCL-1 is associated with a poor prognosis 2 , consistent with the dependency of breast cancer cells, including those of the TNBC subtype, on MCL-1 for survival 3,4,9 . Nonetheless, few insights have been available regarding the mechanisms responsible for MCL-1 overexpression in breast cancer. The present results demonstrate that silencing MUC1-C in TNBC cells results in downregulation of MCL-1 expression. In addition and in concert with this observation, we found that enforced expression of MUC1-C is associated with increases in MCL-1. MUC1-C has been linked to the inhibition of BAX by direct binding to the BAX BH3 domain and thereby suppression of the intrinsic apoptotic pathway 43,44 ; however, there has been no known association between MUC1-C and MCL-1. Our findings that MUC1-C increases MCL-1 protein, and not mRNA levels, provided support for a post-transcriptional mechanism. MUC1-C binds directly to PI3K, promotes activation of AKT and thereby AKT-mediated suppression of GSK3β 23,24 . MUC1-C also activates ERK signaling [26][27][28][29][30] . In concert with activation of the AKT and ERK pathways, we found that silencing MUC1-C decreased p-AKT, p-GSK3β and p-ERK levels in TNBC cells. In turn, targeting MUC1-C in TNBC cells was associated with (i) decreases in ERK-mediated phosphorylation of MCL-1 on Thr-163 and (ii) increases in GSK3β -induced MCL-1 phosphorylation on Ser-159, both of which result in downregulation of MCL-1 stability and expression 18,19,36,45 .
ABT-737 and ABT-263 target BCL-2, BCL-X L and BCL-w 46,47 . By contrast, ABT-737 and ABT-263 are ineffective against MCL-1, and resistance to these agents is often associated with upregulation of MCL-1 expression [13][14][15][16] . The present results demonstrate that treatment of TNBC cells with ABT-737 is associated with increases in ROS and thereby induction of MUC1-C expression by a ROS-mediated mechanism (Fig. 7D). Based on these findings, we selected cells for resistance to ABT-737 and ABT-263. Intriguingly in this regard, we found that MUC1-C and MCL-1 expression are both increased in the resistant cells, invoking the possibility that the upregulation of MUC1-C expression is upstream to that for MCL-1 (Fig. 7D). Indeed, inhibiting MUC1-C function with GO-203 in the ABT-737-and ABT-263-resistant cells resulted in suppression of MCL-1 expression. Moreover, GO-203 treatment was associated with induction of cell death, consistent with dependence of the ABT-737-and ABT-263-resistant cells on MCL-1 for survival. In concert with these results, the combination of GO-203 and ABT-737 was synergistic in the treatment of ABT-737-resistant TNBC cells, indicating that targeting MUC1-C, and thereby downregulating MCL-1, reverses ABT-737 resistance. These findings and the demonstration that GO-203 is synergistic with ABT-737 in drug-naïve TNBC cells provide support for the notion that targeting MUC1-C could be effective in both preventing and abrogating MCL-1-mediated resistance to ABT-737 or ABT-263. Of note, ABT-737 resistance has also been linked to upregulation of the anti-apoptotic BFL-1 protein in lymphoma cells 13 . Thus, further investigation will be needed to determine whether targeting MUC1-C can suppress upregulation of BFL-1 expression.
Targeting MUC1-C in breast cancer cells is associated with increases in ROS and the induction of late-apoptosis/necrosis 20,48 . In addition, targeting MUC1-C with silencing or GO-203 in TNBC cells inhibits self-renewal capacity and tumorigenicity 34 . The present results demonstrating that MUC1-C drives MCL-1 expression in (i) drug-naïve and (ii) ABT-737-and ABT-263-resistant TNBC cells can explain, at least in part, why MUC1-C is an effective target for decreasing TNBC cell survival 20 . Efforts have been underway toward the development of small molecule and peptidic MCL-1 inhibitors; however, there are presently no agents that target MCL-1 in the clinic 3,12,[49][50][51] . The findings that MCL-1 confers resistance to anti-tubulin agents, such as taxol, and radiation has emphasized the importance of developing approaches that target MCL-1 for the treatment of TNBC 41,52,53 . Of potential relevance to the present results, previous studies of breast cancer cells had shown that (i) MUC1-C blocks the apoptotic response to cytotoxic chemotherapeutic drugs 54 , and (ii) GO-203 is synergistic with taxol in inducing apoptosis 55 . Accordingly, studies will be undertaken that address whether targeting MUC1-C can also reverse MCL-1-mediated resistance to anti-cancer agents used for the treatment of TNBC 8 . A Phase I trial of GO-203 has been completed in patients with advanced solid tumors, and this drug has been formulated in polymeric nanoparticles for sustained delivery in the treatment of TNBC and other malignancies 56 . The present findings provide support for considering combinations of GO-203 with BCL-2 inhibitors that are limited by the development of MCL-1-mediated resistance.

Establishment of ABT-737-and ABT-263-resistant cells.
ABT-737-and ABT-263-resistant cells were prepared by continuous exposure to increasing concentrations of drug for 2-3 months. Parental cells were initially exposed to a concentration of 1.0 μ g/ml. Cells were selected for growth in final concentrations of 7.5 μ g/ml ABT-737 and 5 μ g/ml ABT-263.

RNA preparation and real-time quantitative reverse-transcription PCR.
Total RNA was isolated using with Trizol reagent (Invitrogen) following the manufacturer's protocol. Complementary DNA was synthesized from 2.0 μ g total RNA using the with High Capacity cDNA Reverse Transcription Kit (Applied Biosystems), as described 59 . The Power SYBR Green PCR Master Mix (Applied Biosystems, Grand Island, NY, USA) was used with 1 μ l of diluted cDNA for each sample. The samples were amplified using the 7300 Realtime PCR System (Applied Biosystems). Primers used for RT-PCR analysis are listed in Supplemental Table S1.