Tony Reiman, Room 2229, Cross Cancer Institute, 11560 University Avenue, Edmonton, AB, Canada T6G 1Z2. E-mail: firstname.lastname@example.org
Aurora kinases are potential targets for cancer therapy. Previous studies have validated Aurora kinase A as a therapeutic target in multiple myeloma (MM), and have demonstrated in vitro anti-myeloma effects of small molecule Aurora kinase inhibitors that inhibit both Aurora A and B. This study demonstrated that Aurora B kinase was strongly expressed in myeloma cell lines and primary plasma cells. The selective Aurora B inhibitor AZD1152-induced apoptotic death in myeloma cell lines at nanomolar concentrations, with a cell cycle phenotype consistent with that reported previously for Aurora B inhibition. In some cases, AZD1152 in combination with dexamethasone showed increased anti-myeloma activity compared with the use of either agent alone. AZD1152 was active against sorted CD138+ BM plasma cells from myeloma patients but also, as expected, was toxic to CD138− marrow cells from the same patients. In a murine myeloma xenograft model, AZD1152-inhibited tumour growth at well-tolerated doses and induced cell death in established tumours, with associated mild, transient leucopenia. AZD1152 shows promise in these preclinical studies as a novel treatment for MM.
Aurora kinase B (AURKB) is a key regulator of mitosis as a component of the chromosomal passenger complex (CPC). The CPC has essential functions at the centromere in ensuring correct chromosome alignment and segregation and is required for chromatin-induced microtubule stabilization and spindle assembly as well as orderly cytokinesis. AURKB is a member of a family of three mammalian Aurora kinases (Aurora kinases A, B and C), each with a conserved tyrosine kinase domain but differing in subcellular localization and function. We have reported on the potential of Aurora kinases as therapeutic targets in myeloma (Shi et al, 2007). Initially, our focus was on Aurora kinase A (AURKA) as the prime target for therapy, because it is a component of the centrosome that we have shown to be amplified in myeloma and because of its link to the receptor for hyaluronan-mediated motility (RHAMM), a centrosome-associated protein that confers poor prognosis in myeloma when upregulated or aberrantly spliced (Maxwell et al, 2003, 2004, 2005). The clinical significance of centrosome amplification and centrosome-associated gene expression in myeloma has been validated (Chng et al, 2006). AURKA has been recognized for some time as an important molecule in cancer and this has led to more recent exploration of the role of AURKB (Keen & Taylor, 2004; Girdler et al, 2006). As we report here, in the course of our investigations it has become apparent that AURKB is also a target for therapy in myeloma, even though it is not a component of the centrosome.
AZD1152 is a recently developed specific inhibitor of AURKB that is now involved in clinical trials. AZD1152 inhibits Aurora A, B and C but displays ∼3700-fold lower affinity for Aurora A compared with B or C (Mortlock et al, 2007). In early results from a phase I clinical trial in solid tumour patients, the drug was reported to be well tolerated when administered by IV infusion at doses up to 300 mg with significant disease stabilization in five of eight reported patients, and with transient neutropenia being the only noteworthy toxicity (Schellens et al, 2006). Recent work has also demonstrated the preclinical efficacy of AZD1152 in human acute leukaemia cells (Yang et al, 2007). In the present study, we performed a preclinical evaluation of AZD1152 as a novel drug therapy in the treatment of multiple myeloma (MM).
Materials and methods
Purification of patient bone marrow plasma cells
Bone marrow samples. Bone marrow (BM) samples were obtained from MM patients with their written informed consent and with the approval of our local research ethics board, in accordance with the Declaration of Helsinki. Heparinized and aspirated samples were filtered through a mesh and topped with 1× phosphate-buffered saline (PBS) to a final volume of 50 ml. Samples were divided equally into two tubes and under-laid with 14 ml of Ficoll (GE Healthcare Bio-Sciences Corp., Piscataway, NJ, USA) and spun at 754 g for 25 min. The inter-phase layer was removed and placed into a new 50-ml conical tube and topped to 50 ml with 1× PBS. The samples were spun at 524 g for 10 min and pellets were resuspended in 10 ml of 1× PBS and counted.
Magnetic-activated cell sorting. Cells were washed in 20 ml of cold running buffer, centrifuged at 524 g and resuspended in 90 μl per 1 × 107 cells of cold buffer with 10 μl of CD138 magnetic-activated cell sorting (MACS) beads. Cells were washed, centrifuged and resuspended in 1 ml of cold running buffer. Samples were run on an autoMACS sorter (Miltenyi Biotec Inc., Auburn, CA, USA) using a positive selection sensitive mode. Both CD138+ and CD138− cells were counted and 1 × 105 cells from each sample were used for cytospin slides.
MM cell lines
Seven MM cell lines were used in these experiments: RPMI 8226, U266 and LP-1 were generously supplied by S. Treon (Dana-Farber Cancer Institute, Boston, MA, USA). KMS-11 and KMS-12 were generous gifts from T. Otsuki (Kawasaki Medical School, Kurashiki, Japan). An adherent subclone of KMS12 termed KMS12-A was produced in-house. All cell lines were grown in RPMI media containing 10% fetal bovine serum (FBS) at 37°C with 5% CO2..
Total RNA isolation and cDNA synthesis
Approximately 5 × 106 cells were removed and washed with 1× PBS. The cells were spun at 1500 rpm for 9 min and the supernatant was removed. The cell pellet was resuspended in 1 ml of Trizol® Reagent (Invitrogen Canada Inc., Burlington, ON, USA). The samples were stored at −80°C until RNA was isolated using 200 μl of chloroform per 1 ml of Trizol® sample and the standard protocol. RNA pellets were dried, redissolved in 20 μl of diethylpyrocarbonate-treated water and incubated at 70°C for 10 min. The samples were spun at 5200 g for 2 min and the RNA concentration was determined using a NanoDrop® ND-1000 Spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA). Reverse transcription reactions used 0·5 μg RNA, random hexamer primers and 1 μl of SuperScript II reverse transcriptase (Invitrogen Canada) in a total volume of 20 μl and incubated at 42°C for 50 min followed by enzyme inactivation at 70°C for 15 min.
Real-time polymerase chain reaction
TaqMan® real-time polymerase chain reaction (PCR) was used to determine Aurora kinase B (AURKB) mRNA expression in both the cell lines and patient (BM) samples. TaqMan® reaction mixtures were carried out in a 96-well plate and processed using the ABI 7900HT Fast Real-Time PCR system (Applied Biosystems Canada, Streetsville, ON, USA). Analysis and experimental set up utilized the SDS 2.2.2 software (Applied Biosystems). Briefly, 5 μl of TaqMan® Universal Primer Mix, 0·5 μl of Primer and Probe mix, and 2·5 μl of water for each reaction was combined with diluted cDNA. The plate was spun for 2 min at 325 g before loading onto the ABI 7900HT. All reactions were analysed with ‘relative quantification’ using GAPDH as the endogenous control. All samples were completed in triplicate and all target expression threshold cycle (CT) values were correlated with GAPDH CT values. Analysis was completed using the SDS 2.2.2 software and MS Excel.
Cells were harvested, washed with 1× PBS and lysed in radioimmunoprecipitation assay buffer containing Complete Protease Inhibitor Cocktail (Roche Diagnostics Canada, Laval, QC, Canada). After incubation on ice for 10 min, cellular debris was removed by centrifugation at 12 000 rpm for 10 min in an Eppendorf centrifuge. Protein concentrations were determined using the Bradford assay (Bio-Rad Laboratories Ltd, Mississauga, ON, Canada). Proteins were transferred to polyvinylidene difluoride (PVDF) membrane (Millipore, Billerica, MA, USA) or nitrocellulose (Bio-Rad) and detected with antibodies by Western blotting analysis. The antibodies used for primary staining were rabbit anti-Aurora-B polyclonal antibody [ARK-2 (H-75); 1:1000] purchased from Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA. The PVDF membrane was blocked with 5% non-fat milk/1× PBS/0·1% Tween-20. The primary antibody was incubated with the membrane at room temperature for 1·5 h and then washed 3 × 10 min with PBS/0·1% Tween-20. The primary antibodies were detected using horseradish peroxidase-conjugated anti-mouse or anti-rabbit antibody, as appropriate (1:10 000; Invitrogen – Molecular Probes), and enhanced chemiluminescence detection. The membranes were stripped and re-probed with mouse anti-GAPDH (1:1000; Applied Biological Materials Inc., Vancouver, BC, Canada) to ensure equal loading of cell lysate samples.
Drug treatment of myeloma cells with Aurora kinase inhibitor AZD1152
AZD1152 and its active moiety, AZD1152-HQPA, were kind gifts from AstraZeneca, Alderley Park, Cheshire, UK. AZD1152-HQPA is a specific inhibitor of AURKB with 50% inhibitory concentration (IC50) of 0·37 nmol/l vs. 1368 nmol/l for AURKA. AZD1152 is rapidly hydrolysed in the plasma to AZD1152-HQPA in vivo. AZD1152-HQPA was used for the in vitro experiments, while AZD1152 was used in the murine experiments described below. Myeloma cell lines or primary MACS-sorted BM plasma cells were incubated in the appropriate medium with varying concentrations of AZD1152-HQPA, dexamethasone (DEX), melphalan (Melph), or drug vehicle alone, singly or in combination, for varying periods ranging from 24 to 120 h, and subjected to further assays as described below.
Apoptosis, cell viability, target inhibition and DNA content assays
Annexin V/propidium iodide apoptosis assay. Cell pellets were resuspended in 500 μl of binding buffer and 100 μl was transferred to fluorescence-activated cell sorting tubes (∼1 × 106 cells). 5 μl of Annexin-V-fluorescein isothiocyanate (FITC) and 10 μl of 1:10 diluted propidium iodide (PI) was added to each tube. Media only controls were set-up for flow cytometry compensation and quadrants as unstained cells, cells stained with Annexin-V-FITC only and cells stained with PI only. The cells were gently mixed and analysed immediately by flow cytometry.
Cell viability assay with MTS reagent. Approximately 1 × 105 cells in RPMI media with 10% FBS media were plated per well in a treated 96-well plate at 24-h intervals for up to 120 h (in triplicate for each time-point). For each time-point, 20 μl of MTS reagent [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt, ProMega, Madison, WI, USA] solution was added to each well and incubated at 37°C for 4 h. The plate absorbance was read at 490 nm on a 96-well spectra max 190 Plate Reader (Molecular Devices Inc, Sunnyvale, CA, USA) using Softmax Pro 4.8 Software. MTS solution was added to media-only wells to correct for background. Setting the control cells at 100% viability, the viability of cells treated with various concentrations of siRNA or drug was determined and graphed using MS Excel. Student’s two-sided t-test was used to compare results in drug- or control-treated cells.
DNA content – cell cycle analysis. Approximately 2·5 × 105 cells suspended in 1 ml of RPMI media containing 10% FBS (R10) were plated per well in a 6-well treated tissue culture plate and treated with AZD1152-HQPA (20 and 200 nmol/l) as described above. The cells were incubated for either 24, 48, 72 or 96 h at 37°C with 5% CO2. At each time-point, the cells were harvested as described above. For the adherent and suspended cells the supernatant was aspirated and the cells were resuspended in 0·5 ml of 1× PBS with 70% cold ethanol added drop wise. The cells were kept in fixative for 2 h or greater before staining with PI. Cells were analysed via flow cytometry using an Ultra Hypersort (Beckman Coulter Canada, Mississauga, ON, Canada).
AZD1152 treatment in a murine myeloma xenograft model
Irradiated non-obese diabetic/severe immunocombined deficiency (NOD/SCID) mice (strain: NOD.CB17-Prkdcscid/J; The Jackson Laboratory, Bar Harbor, Maine, ME, USA) were utilized in the study to further investigate the efficacy of AZD1152 treatment of MM. RPMI 8226 myeloma cells (3·0 × 107) were injected subcutaneously into the hind-flank region of the mice. Treatment with AZD1152 was given at a dose of 30 mg/kg intraperitoneally twice daily for 4 d. Control-treated mice received the vehicle alone (0·9% NaCl) at the same schedule. Caliper measurements of the three longest perpendicular tumour diameters were performed every alternate day and the product of the diameters was used to estimate the tumour volume. Tumour volumes in drug-treated animals were compared with controls with Student’s two-sided t-test. Animals were euthanized when tumours became large or erosive, significant clinical illness and/or weight loss ensued, or after follow-up was complete. At necropsy, tumours were resected, weighed, and dissected for further morphological and molecular analysis. In all statistical analyses, a P-value ≤0·05 was considered significant.
AURKB is expressed in MM cell lines and purified patient plasma cell samples
A method of real-time quantitative reverse-transcription-PCR (QRT-PCR) was utilized to determine AURKB expression in a broad collection of MM cell lines. The analysis showed significant detectable AURKB transcripts in unsynchronized cells using this method (Fig 1). The protein levels of the cell lines assayed generally correlated with their mRNA expression (Fig 1).
Next we analysed patient BM cells that were sorted into CD138+ and CD138− groups using a MACS sorter. The CD138+ fraction was >95% plasma cells by morphology while the CD138− fraction contained <5% plasma cells. When measured by QRT-PCR, all CD138+ and CD138− cells analysed had quantifiable AURKB mRNA transcripts (Fig 2). It is noteworthy that despite the low-proliferative rate generally attributed to myeloma plasma cells, the AURKB mRNA levels in those cells were of the same order of magnitude as that seen in the accompanying CD138−‘normal’ marrow plasma cells although the CD138− cells generally had higher expression of AURKB mRNA than did the corresponding CD138+ plasma cells from the same patient sample.
AZD1152 induces cell death in MM cell lines in a manner consistent with AURKB inhibition
As AURKB is a key regulator of mitosis and essential for proliferation, we examined the effects of AZD1152 on the cell-cycle profile of several MM cell lines by flow cytometry-based assays. In initial experiments with RPMI8226, after 24-h treatment with 10 nmol/l of drug, there was a marked increase in cells with 4N and 8N DNA content (Fig 3A). In follow-up experiments testing three additional myeloma cell lines, AZD1152 caused significant accumulation of cells with 4N/8N DNA content in KMS12 and U266 and induced apoptosis in KMS18 and U266 as indicated by accumulation of cells with sub-G0 DNA content (Fig 3B). The largest induction of polyploidy was in KMS-12 cells while KMS-18 showed the highest level of induced apoptosis. Dual staining with fluorescent Annexin V antibodies and PI followed by flow cytometry also demonstrated the induction of apoptosis by AZD1152, in a manner paralleling that seen in the DNA content analyses (data not shown). While the degree of polyploidy and the degree of apoptosis did not always correlate, at least one of these two parameters was significantly affected by AZD1152 treatment in all cell lines tested.
To assess the effects on cell viability, we examined the survival of five different MM cell lines after AZD1152 treatment alone and in combination with established therapies. Following drug treatment for 72–96 h, cell death occurred in myeloma cell lines when analysed by MTS assays (Fig 4). In the cell lines examined, there was evidence of reduced cell viability following treatment with AZD1152 alone compared with dimethyl sulphoxide (DMSO). However, the extent of cell killing varied and the effect was not statistically significant in KMS-12 cells, which exhibited resistance to all therapies tested apart from a relatively high concentration of Melph. When AZD1152 was used in combination with DEX, the negative effects on cell viability increased significantly in comparison with single agent in RPMI8226, KMS11 and U266 (P < 0·05 in each case), with a non-significant trend in the same direction seen in LP1. This effect was particularly notable in RPMI8226, as it is a cell line that is resistant to DEX alone. When AZD1152 was used in combination with a high dose of Melph (10 μmol/l), the additive effect was minimal, owing to the high degree of cell killing seen with Melph alone.
Inhibition of AURKB activity by AZD1152 induces cell death in purified patient plasma cell samples
To assess the consequences of AURKB inhibition on cell viability, we examined the survival of MACS-sorted CD138+ and CD138− BM cells from two patients after AZD1152 treatment and in combination with current drug therapies. Following treatments for 96–144 h with 500 nmol/l of AZD1152, cell viability was significantly reduced in CD138+ cells when analysed by MTS assays (P < 0·05 in each case; Fig 5A–B). Toxic effects of AZD1152 treatment were also observed in CD138− cells for both patient samples. There was a trend towards a further reduction in cell viability when AZD1152 was used in combination with DEX, but the results were not statistically significant in comparison with DEX alone in these cases. There was no obvious incremental effect when combining AZD1152 with Melph in primary BM plasma cells.
Effects of AZD1152 treatment on murine xenograft tumour models
To further investigate the efficacy of AZD1152 treatment of MM, further studies were conducted on in vivo growth of xenografted RPMI 8226 cells. In comparison with controls, mice treated with AZD1152 for 4 d starting 24 h after tumour cell inoculation showed significantly attenuated growth of xenograft tumours (Fig 6A). At 4 d, differences in tumour volume could be measured and by 10 d post-treatment, control mice (n = 5) had tumours ranging from 400 to 600 mm3 in size while four of 10 drug-treated mice were tumour-free and the remaining six mice had tumours ranging from 16 to 80 mm3 in size (Fig 6A and B). The difference in mean tumour volume between drug- and control-treated mice was statistically significant (P < 0·0001). Histological examination of tumour tissue harvested 24 h following drug treatment-revealed extensive tumour cell death. There was no obvious impact of AZD1152 treatment on the general appearance or health of the mice and the drug did not induce anorexia or weight loss. Transient leucopenia was noted during the first 7 d after initiation of therapy. The depth of leucopenia was generally modest (approximately two- to 10-fold reduction in white blood cell count) and brief (complete recovery within 1 week), and no infectious complications were noted.
This study has shown high levels of AURKB mRNA and protein expression in a panel of human myeloma cell lines and primary myeloma cells. We have established that treatment of myeloma cell lines and primary cells with the AURKB inhibitor AZD1152 induces apoptotic cell death, in a manner consistent with the expected cell cycle phenotype of AURKB inhibition. As one might expect, some but not all myeloma cell lines tested were sensitive to AZD1152 monotherapy, but the degree of activity observed was nonetheless encouraging. In some cases, increased cell killing was seen when AZD1152 was used in combination with DEX, an active agent in myeloma with a non-overlapping toxicity profile, providing a compelling rationale for exploring their clinical use in combination. Furthermore, we have demonstrated efficacy and tolerability of AZD1152 using an in vivo murine xenograft model.
Myeloma cell lines retain some features of the primary disease from which they are derived, but they are also more proliferative in culture than is usual for primary myeloma plasma cells. It was interesting to see evidence of efficacy in AZD1152-treated primary myeloma cells. It may be that AURKB is serving an important role in the survival of non-dividing myeloma plasma cells, but further study in this regard is required. Recently reported evidence from myeloma gene expression profiling data highlights that elevation in AURKB transcript levels correlate with poor prognosis, providing further justification for the pursuit of this therapeutic target (Chng & Fonseca, 2006). The observed toxicity of AZD1152 to normal marrow elements in culture is not unexpected given the mechanism of action of the drug and its known tendency to cause leucopenia in early clinical trials. As with other myelosuppressive anti-cancer agents, the effective application of AZD1152 in the clinic will require due consideration of this issue.
It has previously been shown that treatment of myeloma cells with pan-Aurora kinase inhibitors, such as VX-680, induces apoptosis, an effect which is potentiated by RHAMM overexpression and which is attenuated by Aurora A overexpression or by RHAMM knockdown (Shi et al, 2007). RHAMM and Aurora A are both found at the centrosome, and are related by their strong interaction with TPX2. We have also reported in abstract form that Aurora A siRNA knockdown induces myeloma cell apoptosis (Reiman et al, 2006). These findings validate Aurora A as a therapeutic target in myeloma. Interestingly, we have observed that myeloma cell treatment with the pan-Aurora kinase inhibitor VE-465 induces apoptosis and accumulation of cells with 4N/8N DNA content in a manner similar to that seen with the more selective Aurora B inhibitor AZD1152 (Reiman et al, 2006). This phenotype is associated with Aurora B inhibition rather than Aurora A inhibition (Girdler et al, 2006). The anti-myeloma effects of VE-465 have been confirmed (Negri et al, 2006). Taken together, the evidence indicates that both Aurora A and B are important targets in this disease.
In summary, there is consistent evidence from our group and several others of the potential of Aurora kinase inhibition as a therapeutic strategy in myeloma. The data presented here provides a strong preclinical rationale for the evaluation of AZD1152 alone or in combination with DEX as a novel treatment for patients with myeloma whose disease has relapsed or is refractory to standard therapy. The adaptation of some of the methods presented here to the study of BM plasma cells from patients on such a clinical trial could provide insight into the achievement of target inhibition and molecular correlates of response to therapy.
The authors acknowledge grant support from the Alberta Heritage Foundation for Medical Research (Clinical Investigator Award - TR), Alberta Cancer Board/Alberta Cancer Foundation (TR), and the Canadian Institutes of Health Research (LMP, TR, ARB). RPE is the recipient of a Fellowship from the Lymphoma Foundation of Canada.