Targeting of apoptotic pathways by SMAC or BH3 mimetics distinctly sensitizes paclitaxel-resistant triple negative breast cancer cells

Abstract

Standard chemotherapy is the only systemic treatment for triple-negative breast cancer (TNBC), and despite the good initial response, resistance remains a major therapeutic obstacle. Here, we employed a High-Throughput Screen to identify targeted therapies that overcome chemoresistance in TNBC. We applied short-term paclitaxel treatment and screened 320 small-molecule inhibitors of known targets to identify drugs that preferentially and efficiently target paclitaxel-treated TNBC cells. Among these compounds the SMAC mimetics (BV6, Birinapant) and BH3-mimetics (ABT-737/263) were recognized as potent targeted therapy for multiple paclitaxel-residual TNBC cell lines. However, acquired paclitaxel resistance through repeated paclitaxel pulses result in desensitization to BV6, but not to ABT-263, suggesting that short- and long-term paclitaxel resistance are mediated by distinct mechanisms. Gene expression profiling of paclitaxel-residual, -resistant and naïve MDA-MB-231 cells demonstrated that paclitaxel-residual, as opposed to -resistant cells, were characterized by an apoptotic signature, with downregulation of anti-apoptotic genes (BCL2, BIRC5), induction of apoptosis inducers (IL24, PDCD4), and enrichment of TNFα/NF-κB pathway, including upregulation of TNFSF15, coupled with cell-cycle arrest. BIRC5 and FOXM1 downregulation and IL24 induction was also evident in breast cancer patient datasets following taxane treatment. Exposure of naïve or paclitaxel-resistant cells to supernatants of paclitaxel-residual cells sensitized them to BV6, and treatment with TNFα enhanced BV6 potency, suggesting that sensitization to BV6 is mediated, at least partially, by secreted factor(s). Our results suggest that administration of SMAC or BH3 mimetics following short-term paclitaxel treatment could be an effective therapeutic strategy for TNBC, while only BH3-mimetics could effectively overcome long-term paclitaxel resistance.

Keywords: SMAC mimetics; high-throughput screen; paclitaxel; resistance; triple negative breast cancer.

Figure 1. A High Throughput Screen (HTS) to identify effective compounds against paclitaxel-residual MDA-MB-231 cells

(A) Cartoon of the HTS workflow. Day 0: seeding of 4 × 10³ MDA-MB-231 cells in 384-well white opaque TC plates in 40 μl of growth medium. Day 1: addition of paclitaxel at 5× concentration (final concentration, 8.5 nM) by the GNF instrument, followed by incubation for 96 h. Day 5: recovery in drug-free medium for 96 h utilizing robotic station (Biotek dispenser/Liconic incubator/BRAVO robot). Day 8: seeding of 1.5 × 10³ paclitaxel-naïve cells. Day 9: addition of the library of small molecule compounds by Echo transfer for 72 h in 5 serial dilutions (120 nM-75 μM and 16 nM-10 μM where appropriate) in triplicates. Day 12: Assessment of cell viability by CellTiter Glo luminescent cell viability assay, followed by automatic reading of the luminescent signal (Liconic incubator/BRAVO robot/PheraStar reader). (B) Graphical summary of the results of the HTS. Out of the 208 small molecule inhibitors that were effective against MDA-MB-231 cells, only 6 were selectively potent against paclitaxel-residual cells (green). Among the compounds that were more effective against paclitaxel-naïve cells (N = 198), treatment with paclitaxel caused variable increase of the IC₅₀ as indicated. Notably, for 23 compounds the fold-increase of IC₅₀ was 10–100, whereas for 8 compounds the IC₅₀ was increased above 100-fold. (C–E) Effective small molecule inhibitors against paclitaxel-residual MDA-MB-231 cells. The small molecule inhibitors belong to two main categories, namely SMAC mimetics (C) and BCL-2 family inhibitors (BCL-XL, BCL-2, BCL-w) (D). Decrease in the viability of the paclitaxel-residual compared to parental paclitaxel-naïve cells was also observed following treatment with the SMO/HH pathway antagonist BMS-833923 (XL139) and the CaSR activator AMG-073 HCl (Cinacalcet hydrochloride) (E). PTX: Paclitaxel.

Figure 2. Effects of BV6 and ABT-263 on paclitaxel-residual and –resistant (PTX^R) TNBC cell viability

(A) Fold decrease of the IC₅₀ for BV6 and ABT-263 in the indicated paclitaxel-residual TNBC cell lines compared to parental naïve cells. The results represent the values ± SEM from three independent experiments. (B) IC₅₀ values of Paclitaxel for parental naïve and PTX^R TNBC cell lines that were generated following repeated cycles of drug pulse followed by recovery in drug-free medium. IC₅₀ values are the average of three independent experiments, and where determined following 72 h of treatment. IC₅₀ fold-change represents the ratio of IC₅₀ values of PTX^R compared to naïve cells. (C) Fold-increase of the IC₅₀ for BV6 and ABT-263 in PTX^R cells in the indicated 7 TNBC cell lines compared to paclitaxel-residual cells. The respective fold change is presented over each bar. The results represent the values ± SEM from three independent experiments. (D) IC₅₀ values of BV6 and ABT-263 for parental naïve, paclitaxel-residual and PTX^R TNBC cell lines. PTX: Paclitaxel. PTX-residual: TNBC cells treated with paclitaxel IC₅₀ for 96 h, followed by 96h recovery in drug-free medium. MSL: Mesenchymal stem cell-like; M: Mesenchymal; BL1: Basal-like 1. SEM: standard error of the mean.

Figure 3. Affymetrix microarray expression profiling of parental naive, paclitaxel-residual and –resistant (PTX^R) MDA-MB-231 cells

(A) Venn diagram of common and differentially expressed genes (p < 0.05, fold change > 2 and < -2) between the three groups: PTX-residual/naïve, PTX^R/naïve and PTX-residual/PTX^R cells. (B) Principal component analysis illustrating the variance between the different gene expression profiles of the biological replicates of naïve, PTX-residual and PTX^R cells. PTX-residual cells display a markedly different expression profile compared to the other two groups. (C) Hierarchical clustering of the 43 commonly affected genes (Venn diagram, panel A) between either paclitaxel-residual or –resistant cells versus the parental naïve controls (Java TreeView [48]).

Figure 4. Gene expression analysis of paclitaxel-residual MDA-MB-231 cells

(A) Most significantly affected biological processes according to Ingenuity Pathway Analysis in PTX-residual MDA-MB-231 compared to the parental naïve cells. (B) List of top 30 significantly down- and up-regulated genes in the paclitaxel-residual group (Partek Genomics Suite). The fold change values are given as linear ratios of the respective gene expression values. Significantly downregulated or upregulated genes implicated in cell cycle-related processes or cell-death/stress response are highlighted in blue or red, respectively. (C) Pathway enrichment analysis (Metascape server) of the IPA apoptosis-related genelist. The most enriched pathway is the Hallmark TNFα signaling via NFκB, whereas the cell cycle-related pathways are negatively enriched. (D–F) Enrichment plots (GSEA) of the Hallmark genesets related with TNFα/NFκB signaling (D), G2M mitotic checkpoint (E) and E2F targets (F). (G) Heatmap of the significantly enriched genes of the HALLMARK TNFa signaling via NFkB geneset (GSEA). Genes that are implicated in the response to SMAC or BH3 mimetics are highlighted with red. Heatmaps of the negatively enriched (H) Hallmark G2M Checkpoint geneset and (I) Hallmark E2F targets (GSEA). Important genes (e.g. BIRC5, E2F1, AURKB, PLK1, various kinesins) are highlighted with blue. (J) qRT-PCR analysis of the indicated genes in the indicated paclitaxel-residual TNBC cell lines and their corresponding parental cells was used to calculate the relative gene expression levels. The results are presented as fold change of expression in paclitaxel-residual cells compare to parental cells. The mean values ± SEM from three independent experiments are shown. Increase in gene expression is highlighted in red and decrease in blue.

Figure 5. Transcriptional profile of paclitaxel-resistant cells compared to paclitaxel-residual cells

(A) Geneset enrichment analysis (GSEA) for the paclitaxel-resistant (PTX^R) dataset compared to the residual counterparts. The results show an inverse enrichment of the significantly affected genesets with negative enrichment of TNFα/NF-κB signaling (blue), and upregulation of cell cycle-related genes (red). (B–F) Relative expression of the indicated genes in PTX-residual and PTX^R TNBC cell lines. qRT-PCR analysis of the indicated genes in parental naïve, PTX-residual and PTX^R TNBC cells was used to calculate the fold changes between gene expression levels in PTX-residual or PTX^R cells and the respective naïve cells. The mean values ± SEM from three independent experiments are shown. The asterisks denote statistical significance *P < 0.05, **P < 0.01, ***P < 0.001. Statistical assessment was based on two-sided t test.

Figure 6. Effect of taxanes on IL24, BIRC5 and FOXM1 expression in breast cancer samples

Gene expression levels of IL24 (A, D), BIRC5 (B, E) and FOXM1 (C, F) in breast cancer patients (BRCA) irrespective of subtype (N = 488) and TNBC (N = 176) patients, from the publically available dataset GSE25066, treated with anthracyclines followed by sequential treatment (N = 185) or not (N = 303) with the taxanes taxol or taxotere. Differences in gene expression were assessed by a two-sided t test on Affymetrix U133A microarray data normalized using the SCAN method. The asterisks denote statistical significance *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 7. Secreted factor(s) by paclitaxel-residual cells sensitize cells to BV6

(A–D) Supernatants of paclitaxel (PTX)-residual cells sensitize cells to BV6. Parental naïve or PTX^R MDA-MB-231 (A, B) and SUM159T (C, D) cells were incubated with supernatants (50% in complete medium) of the corresponding paclitaxel-residual cells or control untreated cells (control medium), either alone or in combination with different concentrations of BV6 as indicated. Cell viability was measured after 72 hr using the Celltiter Blue assay, and the ratio between viability of cells grown in the presence of BV6 in control or PTX-residual derived media to those grown in the absence of BV6 was calculated. (E–H) TNFα sensitizes cells to BV6. Parental naïve or PTX^R MDA-MB-231 (E, F) and SUM159T (G, H) were incubated with the indicated concentrations of TNFα in the presence or absence of BV6 for 72 hr. Cell viability was measured as described above, and the ratios between cell viability in the presence of both BV6 and TNFα to the viability of cells in the presence of TNFα alone at the indicated concentrations are shown. (I–L) IL24 sensitizes cells to BV6. Parental naïve or PTX^R MDA-MB-231 (I, J) and SUM159T (K, L) were incubated with the indicated concentrations of IL24 in the presence or absence of BV6 for 72 hr. Cell viability was measured as described above, and the ratio between cell viability in the presence of both BV6 and IL24 to the viability of cells in the presence of IL24 alone at the indicated concentration was calculated. Statistical assessment of significant differences in cell viability between cells that were exposed to the combined treatment (BV6 and supernatant/TNFα/IL24) compared to cells that were treated with BV6 alone was based on two-sided t test. The results represent the mean values ± SEM of three independent experiments. The asterisks denote statistical significance *P < 0.05, **P < 0.01, ***P < 0.001. Crystal violet staining below each panel demonstrates the differences in cell viability in each condition.