Proteasome inhibition induces IKK-dependent interleukin-8 expression in triple negative breast cancer cells: Opportunity for combination therapy

Triple negative breast cancer (TNBC) cells express increased levels of the pro-inflammatory and pro-angiogenic chemokine interleukin-8 (IL-8, CXCL8), which promotes their proliferation and migration. Because TNBC patients are unresponsive to current targeted therapies, new therapeutic strategies are urgently needed. While proteasome inhibition by bortezomib (BZ) or carfilzomib (CZ) has been effective in treating hematological malignancies, it has been less effective in solid tumors, including TNBC, but the mechanisms are incompletely understood. Here we report that proteasome inhibition significantly increases expression of IL-8, and its receptors CXCR1 and CXCR2, in TNBC cells. Suppression or neutralization of the BZ-induced IL-8 potentiates the BZ cytotoxic and anti-proliferative effect in TNBC cells. The IL-8 expression induced by proteasome inhibition in TNBC cells is mediated by IκB kinase (IKK), increased nuclear accumulation of p65 NFκB, and by IKK-dependent p65 recruitment to IL-8 promoter. Importantly, inhibition of IKK activity significantly decreases proliferation, migration, and invasion of BZ-treated TNBC cells. These data provide the first evidence demonstrating that proteasome inhibition increases the IL-8 signaling in TNBC cells, and suggesting that IKK inhibitors may increase effectiveness of proteasome inhibitors in treating TNBC.

Since there are no effective therapies for TNBC, and the effect of PI on NFκB-dependent transcription in TNBC cells has never been investigated, in this study, we examined the effect of proteasome inhibition on the expression of NFκB-dependent genes in TNBC cells, and tested the hypothesis that proteasome inhibition induces IL-8 expression, resulting in increased proliferation and migration of TNBC cells. Our results are the first to show that proteasome inhibition in TNBC cells specifically upregulates expression of IL-8 and its receptors, CXCR1 and CXCR2. The induced IL-8 expression in TNBC cells is mediated by an increased nuclear accumulation of p65, and IKK-dependent p65 occupancy at the IL-8 promoter. Suppression or neutralization of the induced IL-8, or inhibition of IKK activity, enhances the BZ cytotoxic and anti-proliferative effect in TNBC cells, suggesting that by suppressing the IL-8 expression, IKK inhibitors may increase effectiveness of proteasome inhibitors in TNBC treatment.
Santa Cruz Biotechnology. All other reagents were molecular biology grade and were from Sigma (St Louis, MO).

Transfection with siRNA
Human IL-8 (sc-39631), IKKα (sc-29365), IKKβ (sc-35644), IKKε (sc-39056), and non-silencing (sc-37007) small interfering RNAs (siRNAs) were obtained from Santa Cruz Biotechnology. Prior to transfection, cells were seeded (250,000 cells/ml) into a 6-well plate and incubated in a humidified 5% CO 2 atmosphere at 37˚C in antibiotic-free RPMI medium supplement with 10% FBS for 24 h to about 80% confluence. For each transfection, 80 pmol of either non-silencing siRNA control or specific siRNA were used; cells were transfected 7 h in siRNA transfection medium with siRNA transfection reagent according to manufacturer's instructions (Santa Cruz Biotechnology). After transfection, fresh medium with antibiotics was added, and cells were grown for 24 h before treatment.

Cell proliferation and IL-8 neutralizing assays
Cell proliferation was measured by CellTiter 96 One Solution Cell Proliferation Assay (Promega, Madison, WI, USA). Cells were seeded into 96-well plates at a density of 5000 cells/ 100 μl of medium, and incubated with BZ, CZ, or control DMSO at 37˚C. At indicated time points, 20 μl of CellTiter 96 One Solution Reagent was added to each well, incubated for 4 h at 37˚C, and absorbance at 490 nm was measured.

Cell invasion and wound-healing scratch assays
Invasion assay was performed as described by the manufacturer's protocol (Corning, NY, USA). MDA-MB-231 cells were seeded onto the Corning Biocoat Matrigel Chambers (Corning #354480) at a density of 25,000 cells/0.5 ml in serum-free DMEM medium, with bottom wells containing DMEM medium with 10% FBS. Cells in the top chambers were treated with 10 μM Bay-117082 for 12 h, followed by 10 nM BZ for 12 h. After incubation, non-invading cells were scrubbed from the upper surface of the top chambers using cotton swabs. Matrigel invading cells on the bottom membranes were fixed with 100% methanol, stained with 0.5% crystal violet, washed, and air dried. Invading cells were counted in five randomly selected fields under a phase-contrast microscope at 20X magnification, and quantified using ImageJ software.
For wound-healing scratch assay, MDA-MB-231 cells were seeded in 6-well plates at a density of 500,000 cells/ml of medium containing 5% FBS. Once cells reached 90% confluency, they were incubated 12 h with 10 μM Bay-117082 or control DMSO, scratched using a sterile 200 μl pipette tip, and incubated another 24 h with 10 nM BZ. The scratch area was monitored under phase-contrast microscope at 0 and 24 h after BZ treatment. Scratch width was measured in 5 randomly selected areas at 10X magnification using ImageJ software.
Supernatants from cell invasion and wound-healing assays were collected for IL-8 analysis by ELISA.

Western blotting
Whole cell extracts (WCE), and nuclear (NE) and cytoplasmic extracts (CE) were prepared as described previously [42], and separated on 12% SDS gels. To determine equal protein loading, membranes were stripped and re-probed with anti-actin antibody. Contamination of nuclear and cytoplasmic fractions by cytoplasmic and nuclear proteins, respectively, was determined by immunoblotting using lactate dehydrogenase (LDH) and histone H3 as specific markers, as described [42].

Chromatin immunoprecipitation (ChIP)
ChIP analysis was performed as described [43]. The promoter occupancy was calculated by using the human IGX1A negative control primers (SA Biosciences, Frederick, MD), which detect specific genomic ORF-free DNA sequence that does not contain binding site for any known transcription factors. The results were calculated as fold difference in p65 occupancy at the IL-8 promoter in comparison with the IGX1A locus. The IL-8 primers used for real time PCR were the following: forward, 5'-GGGCCATCAGTTGCAAATC-3' and reverse, 5'-GCT TGTGTGCTCTGCTGTCTC-3'.

Statistical analysis
The results represent at least three independent experiments. Numerical results are presented as means ± SE. Data were analyzed by using an InStat software package (GraphPAD, San Diego, CA, USA). Statistical significance was evaluated by using Mann-Whitney U test, and p<0.05 was considered significant. Levels of significance are indicated as Ã p<0.05; ÃÃ p<0.01; and ÃÃÃ p<0.001.

Proteasome inhibition upregulates IL-8 expression in TNBC cells
To test the hypothesis that proteasome inhibition increases IL-8 expression in TNBC cells, we first analyzed expression of IL-8, as well as other NFκB-dependent genes, in bortezomib (BZ)treated MDA-MB-231 cells that are characterized by the lack of ER, PR, and Her2 receptors, high proliferation rates and resistance to hormone therapy, and have been widely used as an in vitro model to study TNBC. Interestingly, while BZ did not have a significant effect on the expression of NFκB-regulated genes cIAP1, cIAP2, Bcl2, Bcl-xL, and PD-L1 in MDA-MB-231 cells, it dramatically increased the IL-8 mRNA levels ( Fig 1A).
To determine whether PI induces the IL-8 expression also in other types of breast cancer cells, we analyzed IL-8 mRNA levels and cytokine release in BZ-treated MDA-MB-468, HCC-1937, and MCF-7 cell lines. Similar to the MDA-MB-231 cells, the MDA-MB-468 and HCC-1937 cell lines are ER, PR, and Her2-negative, and are unresponsive to hormone therapy. In contrast, the MCF-7 cells are ER and PR-positive, and are characterized by hormone sensitivity and lower proliferation rates. As shown in Fig 1B, 24 h incubation with 100 nM BZ, which approximately corresponds to the clinically used concentrations [44], significantly increased IL-8 mRNA levels in TNBC cells MDA-MB-468 and HCC-1937, as well as in ER/PR-positive MCF-7 cells. However, while 100 nM BZ greatly increased the IL-8 release in all three tested TNBC cell lines, the IL-8 release in MCF-7 cells was much lower (Fig 1C). Both IL-8 gene expression ( Fig 1D) and cytokine release ( Fig 1E) were induced in MDA-MB-231 cells also by carfilzomib (CZ), a second-generation proteasome inhibitor [45,46]. Incubation (24 and 48 h) of MDA-MB-231 cells with 100 nM BZ and CZ produced comparable levels of IL-8 release (Fig 1F).

Suppression of BZ-induced IL-8 potentiates BZ cytotoxic and antiproliferative effect in TNBC cells
IL-8 mediates its functions through binding to its receptors, CXCR1 and CXCR2; both receptors are expressed in TNBC cells and are also regulated by NFκB [47,48]. To determine whether proteasome inhibition might regulate the expression of CXCR1 and CXCR2 in TNBC cells, we analyzed their gene and protein expression in BZ (24 h)-treated MDA-MB-231 cells. Interestingly, 100 nM BZ significantly increased both mRNA (Fig 2A) and protein (Fig 2B) levels of both receptors, indicating that proteasome inhibition increases IL-8 signaling in TNBC cells.
To determine whether suppression of the induced IL-8 expression might augment the BZ cytotoxic and anti-proliferative effect, we analyzed cell viability and proliferation in MDA-MB-231 cells transfected with IL-8 specific siRNA or control siRNA, and incubated 24 h with increasing BZ concentrations. IL-8 mRNA levels ( Fig 2C) and cytokine release ( Fig 2D) were significantly decreased in cells transfected with IL-8 specific siRNA compared to control siRNA. As previously observed [49,50], BZ decreased viability and proliferation of TNBC cells (Fig 2E and 2F). Importantly, IL-8 suppression significantly potentiated the BZ-mediated cytotoxic effect (Fig 2E). Furthermore, IL-8 suppression also significantly decreased cell proliferation in both untreated and BZ-treated MDA-MB-231 cells (Fig 2F). To validate the above results and determine whether they can be duplicated by neutralization of the produced IL-8,

BZ-induced IL-8 expression in TNBC cells is mediated by IKK
To understand the mechanism of how proteasome inhibition increases IL-8 expression in TNBC cells, we first investigated IKK involvement in the BZ-induced IL-8 expression, and analyzed IL-8 mRNA levels in BZ-treated MDA-MB-231 cells pre-incubated with a broadspectrum inhibitor of IKKs, Bay-117082 [51], or with a specific inhibitor of IKKβ, SC-514 [52]. Both inhibitors significantly suppressed the BZ-induced IL-8 mRNA expression in MDA-MB-231 cells (Fig 3A), indicating that the IL-8 expression induced by proteasome inhibition in TNBC cells is mediated by IKKβ. To confirm these data, we analyzed IL-8 mRNA expression and cytokine release in BZ-treated MDA-MB-231 cells transfected with IKKα, IKKβ, IKKε, or non-specific control siRNAs (Fig 3B). While transfection with IKKε did not have any significant effect on IL-8 mRNA levels (Fig 3C) or cytokine release (Fig 3D), transfection with IKKα, and particularly IKKβ, significantly suppressed the BZ-induced IL-8 expression (Fig 3C and  3D).

BZ increases nuclear accumulation of p65 NFκB, and IKK-dependent p65 recruitment to IL-8 promoter in TNBC cells
Because IL-8 transcription is regulated predominantly by p65 homodimers [19,34,35], we wanted to determine whether the increased IL-8 expression induced by proteasome inhibition in TNBC cells is associated with increased nuclear levels of p65 NFκB. Analysis of cytoplasmic and nuclear extracts by western blotting showed that p65 was localized predominantly in the nucleus of MDA-MB-231 cells ( Fig 4A); this is consistent with the high constitutive NFκB activity in these cells [21][22][23][24]. Cell incubation with BZ further increased the nuclear accumulation of p65 (Fig 4A and 4B); this is likely caused by the BZ-mediated inhibition of p65 proteasomal degradation [37]. In addition, BZ induced nuclear translocation and accumulation of IκBα in MDA-MB-231 cells (Fig 4A and 4B), as was previously observed in other cancer cells [38][39][40]. However, since p65 homodimers exhibit a low affinity for IκBα [34,35], the nuclear accumulation of IκBα was not associated with the inhibition of IL-8 transcription in TNBC cells.
To determine whether proteasome inhibition increases IL-8 promoter occupancy by p65, we analyzed the kinetics of p65 promoter recruitment by chromatin immunoprecipitation (ChIP) in MDA-MB-231 cells treated with 100 nM BZ or control DMSO. As shown in Fig 4C,  6 h and 24 h incubation with BZ significantly increased p65 promoter occupancy, indicating that proteasome inhibition induces IL-8 transcription by increasing p65 promoter recruitment. To find out whether the BZ-induced p65 recruitment requires the kinase activity of IKK, MDA-MB-231 cells were pre-incubated with Bay-117082 or SC-514 before 24 h treatment with 100 nM BZ. Inhibition of IKK activity by Bay-117082 and SC-514 significantly attenuated the BZ-induced p65 occupancy at the IL-8 promoter (Fig 4D), indicating that IKK activity is required for the BZ-induced p65 recruitment to IL-8 promoter in TNBC cells.

IKK inhibition enhances BZ cytotoxic and anti-proliferative effect in TNBC cells
Since our data indicated that the BZ-induced, IKK-dependent IL-8 release increases survival and proliferation of TNBC cells, we wanted to test whether inhibition of IKK activity would enhance the BZ cytotoxic and anti-proliferative effect. To this end, cell viability and neutralizing monoclonal antibody. The values represent the mean +/− SE of four experiments; asterisks denote a statistically significant change compared to control. proliferation were analyzed in MDA-MB-231 cells pre-incubated with the IKK inhibitor Bay-117082, and treated 24 h with increasing BZ concentrations. Inhibition of IKK activity by Bay-117082 significantly reduced cell viability ( Fig 5A) and proliferation (Fig 5B) in BZ-treated cells. To confirm the above data by a pharmacologically independent approach, we analyzed cell viability and proliferation in BZ-treated MDA-MB-231 cells transfected with IKKβ siRNA or control non-specific siRNA. Compared to transfection with control siRNA, transfection with IKKβ siRNA significantly decreased viability ( Fig 5C) and proliferation (Fig 5D) of BZ (24 h)-treated cells.
To determine whether inhibition of IKK activity enhances the BZ-cytotoxic effect beyond the 24 h time point, we analyzed cell viability and IL-8 release in MDA-MB-231 cells incubated with BZ with and without Bay-117082 for up to 72 hours. Even though the viability of MDA-MB-231 cells treated with BZ for 48 and 72 h was low, it was further decreased by the inhibition of IKK activity by Bay-117082 (Fig 5E), and this was associated with an inhibition of IL-8 release (Fig 5F).

IKK and proteasome inhibitors have synergistic effect in reducing migration and invasion of TNBC cells
To extend these findings, we examined the effect of proteasome and IKK inhibition on migration and invasion of TNBC cells. Combination of 10 nM BZ and 10 μM Bay-117082 significantly decreased migration of MDA-MB-231 cells measured by the scratch wound-healing assay; both compared to control untreated cells, and compared to cells treated with BZ only (Fig 6A and 6B). In addition, combination of BZ and Bay-117082 significantly decreased the IL-8 release measured in the supernatants of scratched cells; both compared to untreated cells, and compared to cells treated with BZ alone (Fig 6C). Interestingly, 10 nM BZ induced Matrigel invasion of MDA-MB-231 cells (Fig 7A and 7B), and this was associated with increased IL-8 release by the cells (Fig 7C). IKK inhibition by Bay-117082 significantly decreased the invasion potential of BZ-treated cells (Fig 7A and 7B), and suppressed the BZ-induced IL-8 release (Fig 7C). Together, these data show that IKK Proteasome inhibition induces IL-8 in triple negative breast cancer cells inhibition significantly enhances the BZ inhibitory effect on viability, proliferation, and migration of TNBC cells, and inhibits the BZ-induced IL-8 expression and invasion.

Discussion
Triple negative breast cancer comprises 15-20% of newly diagnosed breast cancer cases, and compared with other breast cancer subtypes, exhibits increased aggressiveness and less favorable prognosis. Since TNBC patients are unresponsive to current targeted therapies, new therapeutic strategies are vital. Although previous studies have shown that proteasome inhibition induces apoptosis in TNBC cells [49,50], suggesting a potential therapeutic activity in TNBC, clinical trials using PI as single agents have produced disappointing results. In this study, we demonstrate that proteasome inhibition induces expression of IL-8 and its receptors CXCR1 and CXCR2, resulting in increased survival, proliferation, and migration of TNBC cells. The induced IL-8 expression is mediated by IKK, and associated with an increased nuclear accumulation of p65 NFκB, and IKK-dependent p65 recruitment to IL-8 promoter. Inhibition of IKK activity significantly decreases proliferation, migration, and invasion of BZ-treated TNBC cells. These results are the first to show that proteasome inhibition increases IL-8 signaling in Proteasome inhibition induces IL-8 in triple negative breast cancer cells TNBC cells, indicating that targeting IKK activity might increase PI effectiveness in the treatment of TNBC.
Hundreds of genes are targeted by proteasome, suggesting that proteasome inhibition might not be compatible with normal cellular functions. Interestingly, a number of studies have demonstrated an increased susceptibility of transformed cancer cells, compared to untransformed cells, to proteasome inhibition. While the exact molecular basis for this increased sensitivity remains incompletely understood, several studies have shown that rapidly dividing and proliferating cells are more susceptible to PI-induced apoptosis than quiescent cells [53]. Since NFκB signaling is increased in cancer cells, and promotes their proliferation and migration, it likely represents one of the main targets of PI in cancer cells, compared to normal cells.
Interestingly, however, from the tested NFκB-regulated genes, only expression of IL-8, and its receptors CXCR1 and CXCR2, was induced by proteasome inhibition in TNBC cells (Figs  1A and 2A), demonstrating that the regulation of NFκB-dependent transcription by PI is gene specific. What are the mechanisms regulating specificity of NFκB-dependent transcription in response to proteasome inhibition? Both IκBα and p65 can serve as proteasome substrates [8][9][10]33]. However, while IκBα stabilization by PI downregulates NFκB-dependent transcription, p65 stabilization has the opposite effect, and may increase transcription of p65-regulated genes [54,55], especially genes regulated by p65 homodimers. Thus, the effect of PI on NFκB-dependent transcription likely depends on the subunit composition of NFκB complexes at the corresponding promoters. Unlike other NFκB-dependent genes, the IL-8 transcription is regulated predominantly by NFκB p65 homodimers that have a low affinity for IκBα, resulting in the relative resistance to inhibition by nuclear IκBα [34,35]. Therefore, even though proteasome inhibition stabilizes IκBα and increases its nuclear accumulation, resulting in the inhibition of genes regulated by NFκB p65/50 heterodimers, it concomitantly induces expression of IL-8, and perhaps other NFκB-dependent genes that are regulated by p65 homodimers. In this regard, it will be interesting to determine in the future whether transcription of CXCR1 and CXCR2, which are also increased by BZ in TNBC cells (Fig 2), is also regulated by p65 homodimers.
Importantly, our results demonstrate that the proteasome inhibition-induced IL-8 expression in TNBC cells is dependent on IKK activity, and indicate that it is mediated by IKKα and IKKβ, but not IKKε (Fig 3). Despite the limited effectiveness of PI and IKK inhibitors as single agents in the treatment of TNBC and other solid tumors, accumulating evidence indicates that combining PI and IKK inhibitors increases their cytotoxic effect in solid tumors [23,39,40,56,57]. Combination of BZ and an IKK inhibitory peptide had a synergistic effect on inhibiting proliferation of ER-negative breast cancer MDA-MB453 cells [23]. Inhibition of IKKβ activity also increased the BZ cytotoxic effect in ER-positive breast cancer MCF-7 and T47D cells [56]. Furthermore, IKK inhibition increased effectiveness of BZ in reducing ovarian tumor growth in vivo [57].
Our results indicate that the mechanistic basis of this synergy consists of the induced, IKKdependent expression of IL-8, which increases survival, proliferation, and migration of solid cancer cells. However, it is important to note that our study was done in vitro, using three TNBC cell lines, MDA-MB-231, MDA-MB-468, and HCC-1937. These cell lines are basal-like cells, characterized by increased proliferation and invasiveness compared to the ER/PR-positive MCF-7 cells, and mutated tumor suppressor p53 [58][59][60]. In addition, compared to the ER/PR-positive MCF-7 cells, these TNBC cells are characterized by increased constitutive expression of IL-8, which has been associated with the increased invasive phenotype of these cells [3,4,6]. However, even though these cells have been widely used as an in vitro model to study TNBC, they lack proper tumor microenvironment formed by tumor cells, immune cells, and the surrounding tumor stroma [61]. Thus, future studies should determine whether IKK inhibition increases effectiveness of PI in TNBC animal models, and ultimately in TNBC patients.

Conclusions
Together, our results show that proteasome inhibition upregulates IL-8 signaling in TNBC cells. Suppression or neutralization of the BZ-induced IL-8 potentiates the BZ cytotoxic and anti-proliferative effect. The IL-8 expression induced by proteasome inhibition is mediated by IKK, and inhibition of IKK activity decreases proliferation, migration, and invasion of BZtreated TNBC cells. These data provide a mechanistic rationale for clinical studies using proteasome inhibitors in combination with IKK inhibitors in the treatment of TNBC.