Lapatinib–induced NF-kappaB activation sensitizes triple-negative breast cancer cells to proteasome inhibitors

For total cell lysates, cells were washed with ice-cold PBS one time and lysed in RIPA buffer (20 mM Tris–HCl, pH7.4, 150 mM NaCl, 1% NP-40, 1% sodium deoxycholate, 1 mM ethylenediaminetetraacetic acid (EDTA) and 1 mM ethylene glycol tetraacetic acid (EGTA)). For subcellular fractionation, the methods were done as previously described. Protease inhibitors and phosphatase inhibitors cocktails were added in the RIPA buffer. Proteins were separated by SDS-PAGE, transferred to a polyvinylidene fluoride (PVDF) membrane and blotted with indicated antibodies.

Cells were grown on gelatin cover slips and fixed at day 2 with 4% paraformaldehyde in PBS for 15 minutes. For immunofluorescence staining, cells were next treated with 0.5% Triton X-100 in PBS for 15 minutes and blocked with 10% BSA in PBS for 1 hour followed by incubation with anti-p65 antibody at 4°C overnight. After incubation with horseradish peroxidase (HRP)-labeled secondary antibody, cells were further stained with the nucleic acid stain, diamidino-2-phenylindole (DAPI) (Invitrogen, Carlsbad, CA, USA), and mounted with ProLong Gold antifade mounting reagent (Invitrogen).

Total RNA was extracted by Trizol® Reagent (Invitrogen) according to the instruction manual. RNA was quantified at OD260 nm by using a ND-1000 spectrophotometer (Nanodrop Technology, Wilmington, Delaware USA) and qualitated by using a Bioanalyzer 2100 (Agilent Technology, Santa Clara, California USA) with RNA 6000 nano labchip kit (Agilent Technologies). Total RNA (0.5 mg) was amplified by a Quick-Amp Labeling kit (Agilent Technologies) and labeled with Cy3 or Cy5 (CyDye, PerkinElmer, Waltham, Massachusetts USA) during the in vitro transcription process. CyDye-labled cRNA (0.825 mg) was fragmented to an average size of about 50 to 100 nucleotides by incubation with fragmentation buffer at 60°C for 30 minutes. Correspondingly fragmented labeled cRNA was then pooled and hybridized to Agilent Human Whole Genome Oligo 4 × 44 K Microarray (Agilent Technologies) at 60°C for 17 hours. After washing and drying by nitrogen gun blowing, microarrays were scanned with an Agilent microarray scanner (Agilent Technologies) at 535 nm for Cy3 and 625 nm for Cy5. Scanned images were analyzed by Feature extraction 9.5.3 software (Agilent Technologies), an image analysis and normalization software was used to quantify signal and background intensity for each feature, and the data substantially normalized using the rank-consistency-filtering LOWESS method. The data have been deposited in NCBI’s Gene Expression Omnibus and are accessible through GEO Series accession number GSE51889 [23].

NF-κB-targeted gene expressions were overlaid onto a global molecular network developed from information contained in the Ingenuity Pathways Knowledge Base (IPA Ingenuity Systems [24]). The network of these NF-κB-targeted genes was then algorithmically generated based on their connectivity and the molecular relationships between these genes/gene products were presented graphically. The NF-κB-targeted genes or gene products are represented as nodes, and the biological relationship between two nodes is represented as an edge (line). All edges are supported by at least one reference from the literature, from a textbook, or from canonical information stored in the Ingenuity Pathways Knowledge Base. The intensity of the node color indicates the degree of positive (red) or negative (green) correlation. Nodes are displayed using various shapes that represent the functional class of the gene product. Edges are displayed with various labels that describe the nature of the relationship between the nodes.

In vitro cell viability was measured using an MTT colorimetric assay. Cells were trypsinized to seed at a density of 5 × 10³ to 1 × 10⁴ cells/well in 96-well plates. After various treatments and culturing periods, the culture medium was removed and 100 μL of serum-free medium with 5 mg/mL MTT solution (Sigma, St. Louis, MO, USA), 25 μL/well, was added and the cells incubated for three hours. Finally, 100 μL of dimethyl sulfoxide (DMSO) was added to lyse the cells and the O.D.550 wavelength was detected by ELISA reader.

For clonogenic assay, cells were plated at 5 × 10³ to 1 × 10⁴ cells/well in 6-well plates and pretreated with 1 μM lapatinib or gefitinib for three days followed by addition of MG132 or bortezomib. Colonies were fixed and stained with 30% ethanol containing 1% crystal violet.

Cells were seeded at 2 × 10⁵ cells/well in 12-well plates and transfected with the indicated plasmids in each experiment, including IL-1 β promoter-luciferase plasmids containing NF-κB binding sites. Twenty-four hours after transfection, the luciferase activities in cell lysates were determined by the Luciferase Assay System (Promega, Madison, WI, USA) and normalized to β-galactosiadase activities.

All animal experiments were performed in accordance with a protocol approved by the Institutional Animal Care and Use Committee of China Medical University and Hospital (No. 100-61-N). Female severe combined immunodeficient (SCID) mice at 4 to 6 weeks of age were used in the orthotopic tumor-xenograft model. MDA-MB-231 cells (6 × 10⁶ cells/mouse; re-suspended in a 1:1 mixture of PBS and growth factor–reduced Matrigel (BD Biosciences, San Jose, CA, USA) in a total volume of 50 μL) were injected into the mammary fat pads of SCID mice, and the tumor size was measured regularly. Once the tumor size reached 100 mm³, mice were treated orally with saline, lapatinib (20 mg/kg), bortezomib (0.02 mg/kg), or a combination of lapatinib (20 mg/kg) and bortezomib (0.02 mg/kg) every day. One month later, all mice were sacrificed, and tumors were resected and their size was measured. The tumor growth rates were analyzed by measuring the tumor length (L) and width (W) with calipers and by calculating the volume with the formula LW²/2.

Five-micron thick paraffin-embedded tissue sections were deparaffinized and rehydrated. The tissue sections were incubated for two hours with mouse monoclonal anti-human p65 and anti-Bax antibodies (100 dilution, Santa Cruz, Dallas, Texas USA). After washing to remove unbound primary antibody, sections were treated with a dextran polymer backbone conjugated to secondary antibodies and labeled with HRP according to the manufacturer’s instructions (DAKO Envision system for mouse and rabbit primary antibodies, DAKO Corporation, Carpinteria, CA, USA) for 30 minutes. Tissue sections were incubated in the chromogenic peroxidase substrate, diaminobenzidine, for five minutes. The specificity of labeling by this procedure was verified by negative control reactions using buffer to replace the primary antibody and isotype-specific immunoglobulin G (IgG).

All results are presented as the mean ± SD. A two-tailed Student’s t-test was used to calculate the statistical significance between the groups. The tumor volume was analyzed by a two-sided t-test.

To address whether lapatinib can induce NF-κB activation in both HER2-positive and HER2-negative breast cancer cells, its effect on p65 Ser536 phosphorylation, a critical IKK complex-mediated modification for NF-κB activity [25], was examined. Lapatinib markedly induced p65 Ser536 phosphorylation in HER2-positive SkBr3 and BT474 breast cancer cell lines (Figure 1A) and in HER2-negative MDA-MB-231 and MDA-MB-468 cell lines (Figure 1B). To further examine the activation of NF-κB after long-term treatment with lapatinib acquired resistant clones of SkBr3 and BT474 cell lines, named SkBr3/Lap and BT474/Lap respectively, were established by chronic treatment with increasing concerntrations of lapatinib. p65 Ser536 phosphorylation remained higher in both SkBr3/Lap and BT474/Lap clones compared with their parent cells (Figure 1A) and IKK inhibitors suppressed this phosphorylation in SkBr3/Lap#6 cells (Figure 1C), supporting that lapatinib resistance induced activation of the IKK/NF-κB axis. To examine whether this event occurs in HER2-negative breast cancer cells, lapatinib-selected clones from triple-negative MDA-MB-231, MDA-MB-468, and HBL100 cell lines, named 231/Lap, 468/Lap, and HBL100/Lap respectively, were also established. The induction of p65 phosphorylation by lapatinib was also observed in these lapatinib-selected TNBC clones (Figure 1D) and this phosphorylation in SkBr3/Lap#6, BT474/Lap#3, and 231/Lap#6 cells was reduced once lapatinib was withdrawn from the culture medium (Additional file 1: Figure S1), suggesting that lapatinib activates NF-κB independently of HER2 inhibition. In addition, EGFR specific inhibitors gefitinib (Gef)- or erlotinib (Erl)-selected clones from BT474 and MDA-MB-231 cells were established. However, these two EGFR inhibitors did not increase p65 Ser536 phosphorylation. These results suggest that inhibition of EGFR and HER2 was not involved in the lapatinib-induced NF-κB activation.

In addition to phosphorylation at Ser536, p65 phosphorylations at other residues, such as Ser276 and Ser529, are also critical for the NF-κB activation in response to various stimuli [26, 27]. However, p65 Ser276 phosphorylation (Additional file 1: Figure S2A and S2B) and Ser529 (Additional file 1: Figure S2C and S2D) were suppressed rather than increased in the SkBr3/Lap, BT474/Lap and 231/Lap clones. To further confirm NF-κB activation, the nuclear localization of p65 in these lapatinib-treated cells was examined. Our data showed that the nuclear p65 level was significantly higher in 231/Lap#2 clone than in the MDA-MB-231 cells (Figure 1G), and this event was observed in immunofluorescence staining/confocal microscope analysis (Figure 1H). All these results indicated that lapatinib possesses off-target activity to induce p65 phosphorylation and nuclear translocation in both HER2-positive and TNBC cells.

We next investigated whether the lapatinib-activated NF-κB can mediate its downstream gene expressions. cDNA microarray analysis was employed to profile gene expressions that were upregulated more than two folds in lapatinib-resistant SkBr3 and BT474 cells (Additional file 2: Table S1). Then, this data were analyzed with Ingenuity Pathway Analysis (IPA) to identify which genes are regulated by NF-κB in response to lapatinib treatment. The result showed 24 genes upregulated by NF-κB (Figure 2A), and some well-documented NF-κB-target genes were further validated. The mRNA levels of IL-1β, IL-6 and TRAF1 in SkBr3/Lap#6 cells were higher than those in SkBr3 cells (Figure 2B-D) and were reduced by p65 shRNA (Figure 2F-H). Similarly, IL-1β, IL-6 and COX-2 transcripts were also induced in 231/Lap#2 cells (Figure 2B, C, and E) and were inhibited by p65 shRNA (Figure 2F, G, and I). IL-1β and IL-6 transcripts were also increased by short-term treatment with lapatinib in SkBr3 and MDA-MB-231 cells (Additional file 1: Figure S3A and S3B) and were reduced in response to the lapatinib withdrawal in SkBr3/Lap#6 and 231/Lap#2 cells (Additional file 1: Figure S3C and S3D). IL-6 promoter activity was dramatically higher in 231/Lap#6 cells than in the parental cells. Mutation of the NF-κB-binding site almost abolished this effect (Figure 2J), indicating that lapatinib-activated NF-κB mediated the gene transcription in response to lapatinib treatment. The protein levels of COX-2 and TRAF1 were also increased in both 231/Lap#2 and 231/Lap#6 cells and were suppressed by p65 shRNA (Figure 2K). These results show that lapatinib treatment not only activates NF-κB but also subsequently induces various NF-κB-targeted gene expressions.

Although activation of the IKK complex and its downstream IκBα Ser32/36 phosphorylations were found in SkBr3/Lap and BT474/Lap clones, reduction of IκBα was not seen (Figure 3A and B). Similar phenomena were also observed in lapatinib-treated clones of MDA-MB-231 and MDA-MB-468 (Figure 3C and D). Therefore, whether IκBα degradation was involved in lapatinib-induced NF-κB activation was further investigated. Since IκBα is also a downstream target gene of NF-κB, we hypothesized that lapatinib still can induce IκBα degradation but the protein level of IκBα was retained at a steady state in lapatinib-treated clones due to de novo synthesis. To prove this hypothesis, we treated cells with cycloheximide to inhibit protein biosynthesis and monitored the IκBα level by Western blot analysis. In parental SkBr3 (Figure 3E, left) and MDA-MB-231 (Figure 3F, left) cells, IκBα remained unchanged after treatment with cycloheximide; however, in SkBr3/Lap (Figure 3E, right) and 231/Lap (Figure 3F, right) cells, IκBα declined quickly in the presence of cycloheximide. Furthermore, the IκBα level and Ser32/36 phosphorylations in SkBr3/Lap#6 (Figure 3G) and in 231/Lap#2 (Figure 3H) cells were also elevated when cells were treated with proteasome inhibitor MG132. These results suggest that IκBα protein was continuously degraded but was also re-synthesized simultaneously back to the steady state in response to lapatinib treatment. In support to this notion, an increase in IκBα mRNA level was found in both SkBr3/Lap and 231/Lap clones (Figure 3I), and silencing of p65 expression can down-regulate the transcripts (Figure 3J) and protein expression (Figure 3K) of IκBα. These results suggest that lapatinib-induced NF-κB activation still involved IκBα protein degradation, and the unchanged IκBα level may be due to the elevation of de novo gene expression.

In addition to the IκBα phosphorylations at Ser32/36 by the IKK complex, phosphorylation at Tyr42 by Src can also lead to NF-κB activation by reducing the protein interaction between p65 and IκBα [28, 29]. c-Src/Lck is activated in response to stresses, such as hypoxia and X-rays, and causes Tyr42 phosphorylation of IκBα without its protein degradation [28, 29]. To clarify whether SFK also participates in the NF-κB activation in lapatinib-resistant cells, the SFK activity in resistant cells was examined. Our data revealed that tyrosine phosphorylation of SFK is increased in lapatinib-treated clones of SkBr3, BT474, and MDA-MB-231 (231) cells (Figure 4A). The phosphorylation of SFK was increased by lapatinib treatment for a few days in the parental BT474 cells (Additional file 1: Figure S4A), but was reduced in lapatinib-resistant clones of SkBr3 and MDA-MB-231 cells upon removal of lapatinib from the culture medium (Additional file 1: Figure S4B). Short-term treatment with lapatinib can also induce IκBα Tyr42 phosphorylation in both HER2-positive SkBr3 and HER2-negative MDA-MB-231 cells (Figure 4B and Additional file 1: Figure S4C), which was also observed in SkBr3/Lap#6 and 231/Lap clones (Figure 4C). Both p65 Ser536 phosphorylation and IκBα Tyr42 phosphorylation in lapatinib-treated SkBr3, BT474 and MDA-MB-231 cells were suppressed by Src inhibitors, including dasatinib (Dasa), AZD0530 (AZD) and PP2 (Figure 4D), suggesting that the lapatinib-induced NF-κB activation is Src-dependent. Anti-IκBα immunoprecipitates were immunoblotted with anti-p65 antibody to further confirm the involvement of Tyr42 phosphorylation of IκBα in NF-κB activation, and the interaction between IκBα and p65 was reduced in 231/Lap cells compared to parental cells (Figure 4E). These results suggest that Src-dependent Tyr42 phosphorylation of IκBα may render the de novo IκBα unable to feedback repress NF-κB activity.

Since NF-κB is activated in both HER2-positive and HER2-negative breast cancer cells in response to lapatinib treatment, we next examined the effect of p65 shRNA on the cell viability of SkBr3 and MDA-MB-231 cells and their lapatinib-resistant clones in both the presence and absence of lapatinib. Although silencing of p65 reduced cell viability in SkBr3 but not in MDA-MB-231 cells, treatment with lapatinib enhanced their sensitivity to p65 shRNA in both cell lines (Figure 5A and B). The cell viability inhibition by p65 shRNA was more dramatic in both SkBr3/Lap#6 and 231/Lap#2 cells in the presence of lapatinib, but the suppression was attenuated after lapatinib withdrawal (Figure 5A and B). Similarly, lapatinib also enhanced the sensitivity of MDA-MB-231 cells to IKKα inhibitor (Figure 5C). The 231/Lap#2 cells also showed greater sensitivity to IKK inhibitor than their parental cells, and this effect was diminished upon lapatinib withdrawal from the culture medium (Figure 5C). These results suggest a critical role of IKK/NF-κB activation in the lapatinib-resistance of both HER2-positive and TNBC cells. Prevention of IκBα degradation via inhibiting proteasome activity efficiently impairs NF-κB activation. Treatment with proteasome inhibitors including MG-132, Lactacystin and Proteasome inhibitor I (PSI) did not further enhance the anti-cancer activity of lapatinib in SkBr3 cells due to the high sensitivity of this cell line to lapatinib (Figure 5D-F). Nevertheless, these proteasome inhibitors circumvented the lapatinib resistance in SkBr3/Lap#6 cells but this effect was not observed after lapatinib was removed from the culture medium (Figure 5D-F). These results suggest that inhibition of NF-κB activation by proteasome inhibitors may overcome the acquired lapatinib resistance in HER2-positive breast cancer cells. Our data further showed that the cell viability of MDA-MB-231 cells was little affected by proteasome inhibitors, but the inhibition was significantly enhanced by lapatinib (Figure 5G-I). Interestingly, proteasome inhibitors dramatically inhibited the cell viability of 231/Lap#2 cells in the presence of lapatinib, and lapatinib withdrawal diminished the sensitivity (Figure 5G-I). Poly ADP ribose polymerase (PARP) and caspase-3 protein cleavages induced by the proteasome inhibitor bortezomib were also obviously enhanced in 231/Lap clones (Figure 5J). These results not only suggest that lapatinib can enhance the oncogenic addiction of TNBC cells to NF-κB, but also imply that co-treatment of lapatinib may enhance their sensitivity to proteasome inhibitors.

Next, we tested our hypothesis that lapatinib can enhance or switch the oncogenic addiction of TNBC cells to NF-κB and, in turn, enhance their sensitivity to proteasome inhibitors both in vitro and in vivo. Bcl-2 is a direct transcription target of NF-κB and responsible for anti-apoptosis [30]. Our data showed that the anti-apoptotic Bcl-2 expression was induced by treatment with 1 μM lapatinib for three days in MDA-MB-231 cells and this effect was blocked by bortezomib (Figure 6A). Nevertheless, lapatinib and bortezomib in combination, but not individually, can induce pro-apoptotic Bax expression (Figure 6A) and enhance bortezomib-induced PARP and caspase-3 cleavages in MDA-MB-231 and HS-578 T cell lines, further demonstrating the synergistic effect of lapatinib on the proteasome inhibitor-induced apoptosis. In clonogenic assays, lapatinib and gefitinib did not reduce the viability of MDA-MB-231 and HS-578 T TNBC cell lines due to HER2-negative expression and the dispensable role of EGFR in these cells (Figure 6B). However, pretreatment with lapatinib but not gefitinib can enhance the anti-cancer activity of MG-132 and bortezomib (Figure 6B).

To further confirm this synergistic activity in vivo, MDA-MB-231 cells were injected into the mammary fat pad of SCID mice followed by treatment with lapatinib, bortezomib or lapatinib plus bortezomib, respectively. Treatment with bortezomib and lapatinib combined, but not individually, significantly suppressed the tumor growth (Figure 6C). The nuclear localization of p65 was enhanced in the lapatinib-treated tumors, and this effect was inhibited by treatment with bortezomib (Figure 6D). The combined treatment not only inhibited NF-κB but also induced Bax expression in the xenograft tumor tissues (Figure 6D). These results suggest that lapatinib may sensitize TNBC cells to proteasome inhibitors by increasing their oncogene addiction to NF-κB.