BRCA1/MAD2L1 Deficiency Disrupts the Spindle Assembly Checkpoint to Confer Vinorelbine Resistance in Mesothelioma

Malignant pleural mesothelioma (MPM) is an incurable cancer arising from the parietal pleura. Treatment options for mesothelioma remain limited, and personalized therapy is lacking (1). Antifolate/platinum doublet chemotherapy has been the only licensed treatment since the early 2000s, and there is no FDA- or EMA-approved second-line therapy outside of Japan (2).

The vinca alkaloid spindle poison vinorelbine has demonstrated useful clinical activity in mesothelioma (3–6). We have previously reported that expression of the tumor suppressor breast cancer–associated gene 1 (BRCA1) is an essential regulator of apoptosis induced by vinorelbine in mesothelioma, as evidenced by (i) a positive correlation between BRCA1 expression and in vitro sensitivity; (ii) induction of vinorelbine resistance by BRCA1-targeted RNA interference; (iii) acquired loss of BRCA1 expression in vinorelbine-resistant cells; and (iv) resensitization by ectopic expression of BRCA1 (7). These observations are consistent with other reports (8, 9). Loss of detectable BRCA1 protein has been observed in 38% of mesotheliomas in two independent cohorts of patients, suggesting a potential mechanism of clinical drug resistance (7).

The heterodimeric BRCA1/BARD1 complex modulates mitotic spindle pole assembly, which may contribute to its role in chromosomal stability (10). Loss of BRCA1 has also been reported to inactivate the spindle assembly checkpoint (SAC), causing resistance to the spindle poison paclitaxel (11). The anaphase-promoting complex/cyclosome (APC/C) inhibitor MAD2L1, which is part of the mitotic checkpoint complex, is transcriptionally regulated by BRCA1 via the transcription factor OCT1 (12), implicating a role for BRCA1 in regulating SAC and genome instability.

On this basis, we hypothesized that the BRCA1/MAD2L1 axis would be an essential regulator of vinorelbine response in mesothelioma in vitro, in ex vivo explants and the clinical context, providing a potential biomarker strategy for selecting patients with mesothelioma that are likely to benefit from vinorelbine treatment, in the relapsed setting.

Vinorelbine (cat n. V2264) and docetaxel (cat n. 01885) were obtained by Sigma. Antibodies were used against the following proteins: PARP (cat n. 9542, Cell Signaling Technology), BRCA1 (cat n. OP92, clone MS110, Calbiochem), BUBR1 and BRCA2 (cat n. MAB3612 Clone 8G1, OP95 Clone 2b, Millipore), MAD2L1 (cat n. sc-47747 Clone 17D10, Santa Cruz Biotechnology), and α-tubulin (cat n. ab4074, Abcam). Secondary antibodies were anti-mouse IgG, HRP-linked antibody and anti-rabbit IgG, HRP-linked antibody (cat n. 7076 and 7074, Cell Signaling Technology)

MSTO-211H were purchased from ATCC (cat n. CRL-2081). REN cells obtained were kindly provided by Dr. S.M. Albelda (University of Pennsylvania, Philadelphia, PA). MSTO cells were grown in RPMI Medium 1640 (cat n. 31870, Thermo Fisher Scientific), Glutamax (cat n. 11574466, Thermo Fisher Scientific), and 10% FBS (cat n. F9665 Sigma). REN were grown in Nutrient mixture F12 Ham, Glutamax, and 10% FBS. Vinorelbine-resistant cells (MVR and RVR) were generated by increasing exposure to vinorelbine in a stepwise manner over duration of three months as described previously (7). As apoptosis was induced in the vast majority of cells in the early stage of onset of resistance, it is not possible to determine whether a subclonal expansion accounted for the evolution of the SAC or whether there was a drug-induced alteration in the phenotype (genetic or epigenetic).

A total of 5,000 cells per well were seeded in 96-well plate for the caspase-3 luminescence assay. Vinorelbine in vitro IC₅₀ for REN and MSTO-211H at 48 hours was previously determined (7). Cells were left untreated or incubated with 50 nmol/L (REN) or 100 nmol/L (MSTO-211H) vinorelbine or 20 nmol/L docetaxel. Forty-eight hours following treatment, cells were analyzed using the Caspase-Glo 3/7 Assay (cat n. G8092, Promega).

Forty-eight hours after treatment, cells were lysed in RIPA buffer containing protease inhibitors (cat n. 11836153001, Roche) and whole-cell lysates were clarified by centrifugation. Forty micrograms of total cell lysates was loaded and separated on SDS-PAGE denaturing gels, transferred to nitrocellulose membranes, and blocked in 5% milk-PBS-0.1% tween. Membranes were probed with primary antibodies diluted in 5% milk-PBS-0.1% Tween-20 at 4°C overnight. Signal detection was performed with ECL-plus chemiluminescent system (cat n. GERPN2236, GE Healthcare). Quantifications for Western blot images are provided in Supplementary Table S1.

siRNAs were obtained from Qiagen (Hs_BRCA1_13 FlexiTube siRNA, SI02654575, and Hs_MAD2L1_8 FlexiTube siRNA MAD2L1 SI02653847). On-Target plus MAD2L1 siRNA was obtained from Dharmacon (cat n. J003271–13). siRNA transfections (20 nmol/L) were performed using the RNAiMAX transfection reagent (cat n. 13778075, Invitrogen) according to the manufacturer's instructions.

Total RNA was extracted using TRIzol (cat n. 15596026, Invitrogen) according to the manufacturer's instructions. Reverse transcription was performed with High Capacity RNA-to-cDNA Kit (cat n. 4374966, Applied Biosystems). Real-time PCR was carried out using Power SYBR Green PCR Master Mix (cat n. 4368577, Applied Biosystems) after 24 hours of silencing or 48 hours of treatment. QuantiTect primer assays (Qiagen) were used for BRCA1 (cat n. QT00039305), MAD2L1 (cat n. QT00205709), and actin (cat n. QT00095451).

Cells were treated with vinorelbine 20 nmol/L (this concentration only induced 20% death in both cell lines but was sufficient to induce mitotic arrest and to allow better interpretation of mitotic arrest dynamics). Cells grown on acid-etched glass coverslips were fixed by incubation in 3.7% formaldehyde for 10 minutes followed by permeabilization in 0.5% Triton X-100 for 5 minutes. Cells were blocked in PBS supplemented with 3% BSA before incubation with the appropriate antibody diluted as required in PBS supplemented with 3% BSA.

The primary antibody was mouse BUBR1 (1 μg/mL) and the secondary antibodies was Alexa Fluor 488 goat anti-mouse IgGs (1 μg/mL; cat n. A-11001, Invitrogen).

Imaging was performed on a confocal microscope (TCS SP5; Leica) equipped with an inverted microscope (DMI6000 B; Leica) using a 63× oil objective (NA 1.4). Z-stacks comprising 30 to 50 0.3-μm sections were acquired using LAS-AF software (Leica), and deconvolution of 3D image stacks was performed using Huygens software (Scientific Volume Imaging). To quantify BUBR1 intensity, the mean pixel intensity of individual BUBR1 antibody spots was measured and cytoplasmic background subtracted. A minimum of 15 discrete spots was measured for each cell and the mean of these values calculated. Intensities were scaled so that control intensity was 100%. Time-lapse imaging was performed on a Nikon eclipse Ti microscope equipped with an Andor iXonEM+ EMCCD DU 885 camera using a 10× phase objective. Cells were cultured in multi-well plates and maintained on a stage with a heated incubator at 37°C supplemented with 5% CO₂ using a microscope temperature control system (Life Imaging Services). Images were acquired every 15 minutes for 24 hours using NIS-elements software. Videos were prepared using ImageJ (NIH, Bethesda, MD).

Cells were treated with vinorelbine 20 nmol/L (this concentration only induced 20% death in both cell lines but was sufficient to induce mitotic arrest). Cell-cycle distribution was analyzed by flow cytometry on a FACSCanto II (Becton Dickinson). Cells were collected, washed with PBS, and fixed in 70% ice-cold ethanol for 1 hour or overnight before being stained with propidium iodide (PI; cat n. P4864, Sigma) in PBS containing RNase A (cat n. R6513, Sigma).

Twenty patients with histologically confirmed epithelioid malignant pleural mesothelioma scheduled to undergo extended pleurectomy decortication were consented to provide research samples under ethics approval 14/LO/1527. Informed consent to provide research samples was obtained from all patients. Patient characteristics are described in Supplementary Table S2. All primary pleural tissue was sectioned into fragments measuring approximately 8 mm³. Tissue explants were cultured in RPMI Medium 1640, 1% Glutamax, 10% FBS, 1% penicillin/streptomycin, and 1% fungizone. Explants were allowed to recover overnight prior to treatment. Live explants were treated with vinorelbine 1 μmol/L for 72 hours. This concentration is about 10 times in vitro average IC₅₀ taking into account intrapatient heterogenicity, tumor microenvironment, and lower sensitivity compared with 2D models. This concentration is clinically relevant and comparable with the maximum plasma concentration of vinorelbine in clinical studies. After fixation and embedding, 5-μm sections were used for immunohistochemistry, as described previously (13).

A total of 48 formalin-fixed paraffin-embedded mesothelioma tissue samples were collected from 3 centres (UCL ethics approval 06/Q0505/12; Mayo Clinic ethics approval 13–005053; NKI ethics approval N12PRO, N14PLU), enabling the assessment of BRCA1 and MAD2L1 expression and correlation with clinical outcome. Patient characteristics are described in Supplementary Table S3.

Primary antibodies were diluted in 1% goat serum/0.1% BSA/PBS (MAD2L1, Santa Cruz Biotechnology, 1:50 cat n. sc-374131, Clone C-10; cleaved PARP, Abcam 1:6,000 cat n. ab4830; BRCA1, Calbiochem, 1:100 cat n. OP92, clone MS110). All antibodies were incubated at 4°C overnight. The Novolink Polymer Detection Kit (cat n. RE7150-CE, Leica) was used according to the manufacturer's instructions. Sections were counterstained with hematoxylin and mounted using Leica Sub-X mounting medium (Leica). Images were taken on a Hamamatsu Nanozoomer Digital slide scanner. For BRCA1, automated staining has been performed on the Leica Bond III platform and the Leica Bond Polymer Refine DAB was used for detection.

The scoring was carried out by a pathologist with orthogonal validation using image process by Qupath (14). BRCA1-positive tumors were defined as those demonstrating >10% of cells with nuclear staining (7). MAD2L1-positive tumors were defined as those demonstrating >10% of cells with either nuclear or cytoplasmic staining (15). IHC staining of cleaved PARP was scored as percentage of cells with nuclear staining.

Dose response curves were fitted using nonlinear regression (GraphPad Prism version 6.0, GraphPad Software). The significance of the data has been assessed with t test (two tailed), Wilcoxon test, Mann—Whitney, or one-way ANOVA (Tukey multiple comparisons test). For Kaplan–Meier curves, the log-rank (Mantel–Cox) test was applied. All bar charts show means ± SD. Results are from at least 3 biological replicates. All P values less than 0.05 were considered statistically significant.

To directly test whether or not BRCA1 expression is required for SAC activation and consequent apoptosis induced by vinorelbine in mesothelioma, we assessed the effect of depletion of BRCA1 by RNA interference in MSTO-211H and REN mesothelioma cells. Time-lapse imaging after treatment with vinorelbine showed that while parental cells progressed rapidly through mitosis in the absence of vinorelbine, cells treated with vinorelbine rounded up in a manner indicative of mitotic entry and remained in this state for an extended period of time. In contrast, BRCA1-depleted cells progressed through mitosis in the presence of vinorelbine indicating a defective SAC (Fig. 1A). Cell-cycle analysis revealed that 24 hours after vinorelbine treatment, at least 70% of control cells (siNT) were present in the G₂–M fraction as compared with approximately 20% in the absence of vinorelbine, whereas there was no significant accumulation of cells at G₂–M in response to vinorelbine following knockdown of BRCA1, consistent with inactivation of the SAC (Fig. 1B).

To assess the functional effects of BRCA1 depletion on the SAC in response to vinorelbine, we analyzed the localization of the SAC component, BUBR1, which localizes to kinetochores in the context of a functional checkpoint. Upon vinorelbine treatment of BRCA1-positive cells, BUBR1 exhibited strong staining of kinetochores, indicative of an active SAC. However, in BRCA1-silenced cells, the intensity of BUBR1 was significantly reduced at kinetochores (Fig. 1C). Total cellular BUBR1 protein expression was increased following BRCA1 silencing in MSTO-211H, although this difference was not significant in REN cells (Fig. 1D). Vinorelbine induced apoptosis in control cells (siNT) as shown by a significant increase in caspase-3 activity compared with untreated control, whereas BRCA1, but not BRCA2 depletion (Supplementary Fig. S1), caused a significant reduction in vinorelbine-induced apoptosis (Fig. 1E).

MAD2L1 has been identified as a transcriptional target of BRCA1 (12). Accordingly, following BRCA1 silencing by RNA interference, we observed downregulation of MAD2L1 at both the mRNA (Fig. 2A) and protein levels in MSTO-211H and REN cells (Fig. 2B).

To confirm that BRCA1 mediates the response to vinorelbine through regulation of the SAC via MAD2L1, we determined the effects of MAD2L1 depletion on vinorelbine-induced apoptosis. Treatment with vinorelbine in MSTO-211H and REN control cells (siNT) induced significant activation of caspase-3 compared with untreated control. MAD2L1 depletion by two different siRNAs phenocopied BRCA1 silencing and rescued cells from vinorelbine-induced apoptosis, confirming that the BRCA1/MAD2L1 axis is required to mediate the proapoptotic response to vinorelbine (Fig. 2C).

To determine whether cells under selection for resistance to vinorelbine might acquire SAC deficiency through loss of BRCA1 and MAD2L1, we studied MVR and RVR cell lines, which had been selected for resistance to vinorelbine (7). MVR and RVR cells expressed significantly lower constitutive levels of BRCA1 protein compared with the parental cells, with a reduction in MAD2L1 mRNA and protein expression in MVR and RVR compared with parental cells (Fig. 3A).

Consistent with BRCA1 RNA interference experiments, resistant cells showed a significant reduction in caspase-3 activity in response to vinorelbine, compared with parental cell lines (Fig. 3B). Upon treatment with vinorelbine, parental REN cells rounded up in a manner indicative of mitotic arrest, while RVR cells progressed through mitosis consistent with SAC deficiency. A short delay in mitotic progression upon vinorelbine treatment was observed by live cell imaging in MVR cells although, in contrast to MSTO-211H cells, MVR cells rapidly resumed cell-cycle progression following the transient mitotic arrest (Fig. 3C). Resistant cells showed a reduction in the G₂–M population by flow cytometry after treatment with vinorelbine (Fig. 3D).

Localization of BUBR1 to mitotic kinetochores was reduced by approximately 80% in RVR compared with parental REN cells upon vinorelbine treatment. There was a more modest, albeit significant reduction in BUBR1 intensity (∼50%) in MVR versus MSTO-211H parental mitotic cells (Fig. 3E). This supports the hypothesis that the short delay in mitotic progression observed upon vinorelbine treatment of MVR cells is secondary to acquisition of a defective SAC. An increase in protein levels of BUBR1 in MVR compared with MSTO-211H was observed, although there was no change in expression upon treatment with vinorelbine in either cell line. In contrast, basal levels of BUBR1 were not different between REN and RVR. Vinorelbine induced a decrease in BUBR1 expression in both parental and resistant cells (Fig. 3F). To establish whether this acquired SAC defect would confer resistance to other spindle poisons, the microtubule stabilizer, docetaxel was studied. Apoptosis in response to docetaxel was significantly reduced following BRCA1 silencing in parental MSTO-211H and REN cells (Supplementary Fig. S2A) and in vinorelbine-resistant MRV and RVR cells (Supplementary Fig. S2B). This was associated with loss of SAC-mediated mitotic arrest (Supplementary Fig. S2C).

Live mesothelioma explants cultures were generated from a cohort of 20 patients. BRCA1-positive staining was observed in 85% of samples (n = 17). BRCA1 negativity was observed in 15% of samples (n = 3) and was associated with MAD2L1 negativity in 1 of these samples. Following ex vivo treatment with vinorelbine, nuclear cleaved PARP staining, indicative of apoptosis, was observed only in the context of BRCA1 positivity, in 59% of cases (n = 10), while all cases negative for either BRCA1 or MAD2L1 (n = 3) were resistant to vinorelbine with a 0% response rate (Fig. 4A and B). ROC curves for BRCA1 and MAD2L1 showed a promising fit, although the curves were not significant probably due to the sample size [BRCA1 AUC = 0.853, P = 0.057 (95% confidence interval (CI), 0.677–1.0), MAD2L1 AUC = 0.792, P = 0.186 (95% CI, 0.568–1.0); Supplementary Fig. S3].

This data are consistent with BRCA1/MAD2L1-regulated SAC as a mediator of response to vinorelbine.

In a cohort of 48 patients, BRCA1 protein expression was lost in 39.6% of patients (n = 19), of which 68.4% (n = 13) also lacked MAD2L1 expression (Fig. 4C). BRCA1 and MAD2L1 were not prognostic in this cohort (Fig. 4D and E; Supplementary Fig. S4).

To explore the impact of BRCA1/MAD2L1 expression on clinical outcome following vinorelbine treatment in the relapse setting, patients who had received second-line treatment with vinorelbine were stratified by BRCA1 and MAD2L1 protein expression (n = 10). BRCA1/MAD2L1-negative patients (n = 6) had a worse median overall survival of 5.9 months versus 36.7 months for BRCA1/MAD2L1-positive patients (n = 4; P = 0.03, HR = 3.737), consistent with resistance to vinorelbine (Fig. 4F). The effect of MAD2L1 negativity only (n = 3) on survival was not significant (P = 0.27), although this is likely to be due to the very small sample size (Supplementary Fig. S5).

Despite recent advances in our understanding of interpatient genomic heterogeneity in mesothelioma, targeted therapies are lacking, and no treatment has yet demonstrated an improvement in survival in the relapsed setting. The spindle poison vinorelbine mediates microtubule depolymerization and has demonstrated clinically useful activity in mesothelioma with a reported disease control rate of 68% in the relapsed setting (3, 5, 6). There are at present no validated predictive biomarkers in use to facilitate patient stratification for this agent.

Mesotheliomas exhibit a high degree of aneuploidy and genomic instability (16). The fidelity of the SAC is essential for maintaining genome stability (17), and SAC deregulation has been previously reported in mesothelioma (18). Consistent with our results, sensitivity to the spindle poison paclitaxel has been reported to be dependent on an intact SAC as evidenced by functional genetic studies involving suppression of MAD2L1, BUBR1 (19), or BRCA1 (11) in breast cancer, whereby silencing of BRCA1 induced >1,000-fold cross-resistance to multiple spindle poisons in breast cancer cell lines (8).

BRCA1 has been reported to regulate the response to spindle poisons by affecting the dynamics of SAC activation and mitotic arrest (11, 20). The heterodimeric BRCA1/BARD1 complex mediates E3 ubiquitin ligase activity, and has been shown to transactivate the anaphase-promoting complex APC/C inhibitor MAD2L1 via the Oct1 transcription factor (12). Several studies have uncovered an array of BRCA1 targets involved in the maintenance of chromosome integrity during cell cycle (21–23), and a number of factors are likely involved in regulating the spindle assembly checkpoint. However, the specific clinical relevance of these factors in mesothelioma needs to be addressed. Activation of p21 and p27 by BRCA1 has been shown to lead to arrest cells at the G₁–S phase. Transcriptional repression of cyclin B by BRCA1 is involved in mitotic entry and lack of BRCA1 has been shown to prevent a G₂–M arrest in response to ionizing radiation. The protein expression of Cyclin B has been explored in our model, although no significant change in basal expression was observed between parental and resistant cells or after silencing of BRCA1 (Supplementary Fig. S6). Finally, BRCA1-mediated transactivation of GADD45 was shown to be important in the regulation of the mitotic checkpoint, through the regulation of the CyclinB–cdc2 complex (22).

Understanding of the precise mechanism by which BRCA1 regulates cell fate following mitotic arrest induced by vinorelbine has remained elusive. Here, we show that vinorelbine-induced apoptosis occurs as a result of prolonged activation of the SAC with subsequent mitotic catastrophe that is BRCA1 dependent. Furthermore, we have shown that vinorelbine-resistant cells evolve a dysfunctional SAC associated with downregulation of both BRCA1 and MAD2L1.

The basis for loss of BRCA1 expression in mesothelioma is unknown. In contrast to breast and ovarian cancers, mesotheliomas have not been reported to harbor biallelic somatic alteration of BRCA1 or MAD2L1 (24, 25). BRCA1-associated protein 1 (BAP1) is one of the most commonly mutated genes in mesothelioma (24–27). Deletion of BAP1 has been reported to reduce BRCA1 expression, which can be rescued with ectopic expression of BAP1. This implicates possible control of BRCA1 stability and therefore SAC by BAP1 (28). We have observed a positive correlation between BAP1 expression, SAC activation and consequent sensitivity to vinorelbine and anetumab ravtansine. Also, primary cell lines harboring BAP1 mutation showed reduced BRCA1/MAD2 protein expression (29). This putative association between BAP1 and SAC regulation deserves further exploration. Kumar and colleagues recently reported that BAP1 expression correlates with vinorelbine sensitivity in the MS01 trial cohort (22).

From a translational perspective, loss of BRCA1 and/or MAD2L1 expression in mesothelioma may have implications for predicting the clinical response in patients treated with vinorelbine in the relapse setting. Here, we show that BRCA1 is lost in 39.6% of cases in a cohort of 48 patients, of which 68.4% of BRCA1-negative tumors also lacked expression of MAD2L1 (i.e., double negative). In patients treated with vinorelbine second line, BRCA1/MAD2L loss predicted clinical outcome, suggesting possible utility as a biomarker.

These findings will be explored in an independent validation associated with the Vinorelbine in Mesothelioma (VIM) clinical trial (NCT2139904), a large open-label randomized controlled phase II trial (second-line treatment with oral vinorelbine vs. active symptom control). The analysis of the VIM trial will allow testing for both prognostic and predictive effects of BRCA1 and MAD2L1 either alone or with combined loss.

S. Busacca reports grants from The June Hancock Mesothelioma research fund during the conduct of the study. J. Dzialo reports research support from Astex Pharmaceuticals. A.S. Mansfield reports grants from NIH during the conduct of the study and other from AbbVie, AstraZeneca, BMS, Genentech, Roche, and Mesothelioma Applied Research Foundation outside the submitted work. S.M. Janes reports personal fees from Astra Zeneca, grants from Jansen, Grail Inc., and Owlstone, other from Takeda, and personal fees from Bard1 Bioscience outside the submitted work. P. Baas reports grants and other from BMS and grants from Merck outside the submitted work. D.A. Fennell reports grants from Bayer during the conduct of the study; grants, personal fees, and nonfinancial support from BMS; grants from Astex Therapeutics; grants and personal fees from Boehringer Ingelheim; nonfinancial support from Pierre Fabre; personal fees from Inventiva; and other from Atara Therapeutics outside the submitted work. No disclosures were reported by the other authors.

S. Busacca: Conceptualization, data curation, formal analysis, supervision, investigation, methodology, writing–original draft, writing–review and editing. L. O'Regan: Investigation, methodology. A. Singh: Investigation. A.J. Sharkey: Investigation, methodology. A.G. Dawson: Investigation, methodology, writing–original draft. J. Dzialo: Investigation. A. Parsons: Formal analysis. N. Kumar: Investigation. L.M. Schunselaar: Investigation. N. Guppy: Investigation. A. Nakas: Resources, methodology. M. Sheaff: Investigation, visualization. A.S. Mansfield: Resources. S.M. Janes: Resources. P. Baas: Resources. A.M. Fry: Resources, writing–original draft. D.A. Fennell: Conceptualization, resources, data curation, formal analysis, supervision, funding acquisition, investigation, writing–original draft, writing–review and editing.

S. Busacca was funded by The June Hancock Mesothelioma research fund (JH/09/2011).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.