Background
The
MYC oncogene is one of the most commonly amplified oncogenes in human breast cancer and contributes to its formation and development [
1‐
3].
MYC gene amplification has been found in approximately 15% of breast tumours, while more than 40% of breast cancers over-express MYC protein, indicating that gene amplification is not the only cause of MYC over-expression [
4,
5]. MYC over-expression results in a number of cellular changes, including transcriptional amplification [
6,
7] and increased protein biosynthesis [
8]. MYC-stimulated cell cycle progression has also been well studied. Cyclin-dependent kinases (CDKs), including three interphase CDKs (CDK2, CDK4 and CDK6) and a mitotic CDK (CDK1), are critical regulators of cell cycle progression in mammalian cells [
9]. Increased cyclin E-CDK2 activity appears to be a principal mechanism contributing to MYC-induced G
1-S phase transition in breast cancer cells [
10,
11], possibly through suppression of the CDK inhibitor p21 [
12,
13] and induction of the CDK phosphatase CDC25A [
14]. Although cyclin D1 and CDK4 are putative MYC target genes, and required for MYC-mediated transformation in keratinocytes [
15,
16], the proliferative effect of MYC in breast cancer cells appears to be independent of cyclin D1/CDK4 activation as evidenced by the absence of cyclin D1 up-regulation and CDK4 activation upon MYC induction [
11].
The key role of MYC activation in the pathogenesis of breast cancer and the high incidence of MYC deregulation make MYC an attractive therapeutic target in breast cancer. However, transcription factors such as MYC are challenging to target directly and clinically-effective pharmaceutical agents targeting MYC are not yet available [
17,
18]. Nevertheless, cancer cells develop dependence on other genes and pathways in order to overcome anti-tumorigenic effects, such as apoptosis and senescence, that result from activation of MYC. These dependencies may provide novel therapeutic options for targeting MYC addiction. Consequently, an alternative approach which has recently received great attention is to identify genes that are synthetically lethal in MYC-dependent cancers. Genome-wide RNAi screens for synthetic lethality in MYC over-expressing cells highlight the potential of targeting cell cycle kinases for MYC-dependent cancers [
19,
20]. Other studies using a candidate approach also identified several cell cycle kinases as MYC-synthetic lethal genes in different types of cancer, including CDK2 [
21], CDK1 [
22] and aurora-B kinase [
23]. Since cellular context and tissue type affect the biological functions of MYC [
24] and thus presumably affect these synthetic lethal interactions, we investigated the therapeutic potential of specific CDK inhibition in MYC-driven breast cancer.
Aberrant CDK activation induces unscheduled proliferation and leads to genomic and chromosomal instability in cancer cells [
25]. Consequently, CDK inhibition has been considered as a potential therapeutic strategy for cancer treatment, and a series of CDK inhibitors have been developed. Disappointingly, CDK inhibitors have yet to demonstrate significant clinical advantages as sole agents [
26]. Accumulating evidence suggests that tumour cells have a selective dependence on specific CDKs, therefore, identification of specific genetic contexts in which tumour cells are the most likely to be responsive to CDK inhibitors, is required to improve effectiveness of CDK inhibitors in clinical trial [
25].
In this study we used an RNAi approach to identify MYC-dependent breast cancer cell lines and then inhibited CDKs including CDK4/6, CDK2 and CDK1 individually by either RNAi or small molecule inhibitors in both MYC-dependent and MYC-independent cells. We found that targeting CDK1 rather than CDK4/6 or CDK2 selectively reduced the viability of MYC- dependent breast cancer cells, suggesting a potential therapeutic value of targeting CDK1 for MYC-driven human breast cancer.
Methods
Cell lines, cell culture and reagents
The cell lines used in this study: AU565, BT20, BT474, BT483, BT549, HCC1143, HCC1500, HCC1569, HCC1937, HCC1954, HCC38, HCC70, Hs578T, MDA-MB-134, MDA-MB-175, MDA-MB-361, MDA-MB-436, MDA-MB-453, MDA-MB-468, SKBR3, and ZR751 were obtained from ATCC, Rockville, MD, USA. MCF-7 cells were obtained from Michigan Cancer Foundation, Detroit, MI, USA. The cell lines HBL100, MDA-MB-157, MDA-MB-231 and T47D were obtained from EG&G Mason Research Institute, Worcester, MA, USA.
The cell lines AU565, BT20, BT474, BT549, HBL100, Hs578T, MCF-7, MDA-MB-134, MDA-MB-157, MDA-MB-175, MDA-MB-231, MDA-MB-361, MDA-MB-436, MDA-MB-453, MDA-MB-468, SKBR3, T47D and ZR571 were cultured in RPMI 1640 medium supplemented with 10% heat-inactivated fetal bovine serum (FBS), 6 mM L-glutamine, 20 mM HEPES and 10 μg/ml human insulin (CSL-Novo, North Rocks, NSW, Australia). The remaining cell lines were cultured in RPMI 1640 medium supplemented with 10% heat-inactivated FBS, 6 mM L-glutamine, 1 mM sodium pyruvate and 20 mM HEPES. The MYC over-expressing MCF7 cells have been previously described [
11,
27] and were cultured in the same conditions as the parental cells.
The CDK4/6 inhibitor PD0332991 was purchased from Selleck Chemicals (Houston, TX, USA), CDK2 inhibitor SNS-032 from Symansis (Auckland, New Zealand) and CDK1 inhibitors, RO-3306 and CGP74514A, from Calbiochem (San Diego, CA, USA).
Cell proliferation and apoptosis analysis
Bromodeoxyuridine (BrdU) incorporation was assayed using the Cell Proliferation ELISA, BrdU (colorimetric) Assay system (Roche, Dee Why, NSW, Australia). Proliferation was also assessed by AlamarBlue (Life technologies, Mulgrave, VIC, Australia). Cell cycle analysis was performed by flow cytometric analysis of propidium iodide-stained, ethanol-fixed cells. The apoptotic cell population was determined by staining methanol-fixed cells with the M30 CytoDEATH antibody (Enzo life Sciences, Farmingdale, NY, USA).
siRNA transfection
The MYC siRNA pool contained equimolar concentrations of siMYC-17 (5′-GGACUAUCCUGCUGCCAAG-3′, Catalogue number D-003282-17-0050) purchased from Dharmacon (Lafayette, Colorado, USA) and MYC silencer (5′-GAGCUAAAACGGAGCUUUU-3′, Catalogue number s9130) purchased from Ambion (Austin, TX, USA). Gene-specific siRNAs including cyclin D1 (L-003210-00), CDK2 (L-003236-00)), CDK1 (J-003224-13) and On-Target Plus Non-Targeting siRNA control (D-001810-10), as well as siTOX transfection control (D-001500-01), were purchased from Dharmacon. siRNA transfection was performed by reverse transfection, where cells were seeded directly onto plates containing transfection reagents and siRNA mixture.
Western blot analysis
Protein lysates were harvested as described previously [
11]. 10 to 30 ug of lysate was separated using NuPage polyacrylamide gels (Life technologies, Mulgrave, VIC, Australia) prior to transfer to polyvinyl difluoride membranes. The membranes were incubated with the following primary antibodies: CDK1 (P34), CDK2 (M2), cyclin A (C-19), cyclin D1 (DCS-6), E2F1 (KH95) and MYC (9E10) from Santa Cruz Biotechnology (Santa Cruz, CA, USA); BIM, cyclin E1, cyclin E2, phospho-CDK2 (Thr160), phospho-pRB (ser795) (Cell Signaling, Danvers, MA, USA); p21
Cip1/Waf1, p27
Kip1 and pRB from BD Pharmingen (San Diego, CA, USA), β-Actin (AC15) from Sigma (St Louis, MO, USA). The secondary antibodies were horseradish peroxidase-conjugated sheep anti-mouse or donkey anti-rabbit antibodies (Amersham, Rydelmere, NSW, Australia), and specific proteins were visualized by chemiluminescence (Perkin-Elmer, Rowville, VIC, Australia). Densitometry was performed using the software ImageJ.
Statistical analysis
All experiments were repeated at least three times. All numerical data are expressed as mean ± SEM. Statistical analyses were done by one-way ANOVA or linear regression using PRISM 6 (GraphPad, San Diego, CA). Error bars on all graphs represent the standard error of the mean between measurements. P < 0.05 was considered significant.
Discussion
MYC is a pivotal regulator of cell growth in breast cancer [
27]. In transgenic mouse models with inducible MYC, withdrawal of MYC expression induces breast tumour regression, indicating these tumours are addicted to MYC function for tumour maintenance [
40‐
42]. We herein exploited this oncogenic addiction to assess the dependence of human breast cancer cells on MYC function through an RNAi approach. Depletion of MYC blocked proliferation of 85% of breast cancer cell lines (22/26 cell lines), which were classified as MYC-dependent breast cancer cells while 15% of cell lines (4/26 cell lines) showed resistance to MYC RNAi and were therefore classified as MYC-independent cells. We further identified that MYC-dependent breast cancer cells possessed high MYC protein expression and high MYC phosphorylation level, suggesting an elevated MYC signalling activity in these cells. Through identification of a MYC transcription gene signature, several studies have uncovered an enrichment of MYC-driven transcription programs in basal breast cancer [
43‐
45]. More recently, Horiuchi
et al. reported that triple negative breast cancer exhibited increased activity of the MYC pathway [
46]. However, we did not found a significant difference between ER + and ER- cells in MYC dependence, although ER- cells tend to be more sensitive to MYC depletion than ER + cells. Instead, our data showed that all
HER2-amplified cell lines were dependent on MYC function, raising the possibility of a biological link between MYC and HER2 pathways. As a downstream target of HER2 signalling, MYC mediates HER2-driven proliferative activity in breast cancer cells [
47].
MYC amplification has also been significantly associated with
HER2 amplification in human breast tumours [
48]. Although patients with
MYC/HER2 co-amplified breast tumours have worse outcomes than patients with single gene amplified tumours [
49], the predictive value of
MYC amplification in the response to adjuvant trastuzumab in HER2-positive breast tumours is still unclear [
50]. Nevertheless, our finding implicated a therapeutic potential of MYC inhibition in
HER2-amplified breast cancers.
Since directly targeting MYC remains a challenge in clinical practice, either targeting components of the MYC pathway, or using a synthetic lethal strategy have been suggested as new options for MYC-dependent malignancies [
17]. In this study we assessed the therapeutic potential of specific CDK inhibition in MYC-dependent breast cancer cells.
CDK4 has been identified as a key MYC target gene in mammals [
16].
CDK4-deficient mice were resistant to skin tumour development induced by MYC [
15], whereas mice lacking cyclin D1 expression and consequently lacking CDK4 activation still developed mammary tumours induced by MYC activation [
51], strongly arguing that the requirement for CDK4 activity in MYC-induced tumorigenesis is affected by cellular context and tissue type. In agreement with
in vivo observations, our study identified a distinct response pattern to MYC inhibition compared to cyclin D1-CDK4/6 inhibition in breast cancer cells. Studies here and by others [
32] demonstrated that cells with luminal ER-positive subtype and a functional pRb pathway were more sensitive to cytostatic effects of cyclin D1 and CDK4/6 inhibition. In contrast, the response to MYC depletion was not dependent on pRb status and was not significantly correlated with the molecular subtypes present in the panel of cell lines. Overall, our data suggested that CDK4/6 inhibitors are unlikely to be useful in the treatment of MYC-dependent breast tumours.
Activation of CDK2, another interphase CDK, is involved in MYC regulation of G
1-S phase transition. As a major event downstream of MYC activation, CDK2 activation can also suppress MYC-induced senescence [
52], which raised the possibility of CDK2 as a potential therapeutic target for MYC-dependent cancers. Consistent with our previous results [
11,
27], this study demonstrated that depletion of MYC reduced CDK2 activity in MYC-dependent cells but not in MYC-independent cells, indicative of a role of CDK2 inactivation in MYC inhibition-induced cell cycle arrest. This occurred in the absence of changes of cyclin E1 expression. Despite our previous study showing cyclin E2 was up-regulated through cyclin D1 but not MYC in MCF-7 cells [
53], we noted that cyclin E2 expression was reduced following MYC RNAi in 5 of 9 MYC-dependent cell lines, suggesting cell type and genetic context-dependent regulation of cyclin E2 expression in breast cancer cells. Although CDK2 inactivation has been reported to induce apoptosis in MYCN-amplified neuroblastoma cells [
21], in this study, siRNA-mediated CDK2 depletion failed to induce apoptosis in breast cancer cells. Inhibition of CDK2 by either siRNA or an inhibitor, reduced cell proliferation of both MYC-independent and MYC-dependent cells, and MYC-independent cells possessed relatively high sensitivity to a CDK2 inhibitor, SNS-032. Therefore, our data do not support a synthetic lethal interaction between CDK2 inactivation and MYC activation in breast cancer cells.
Unlike CDK4, CDK6 and CDK2 which are redundant for the mammalian cell cycle, CDK1 is essential for cell division and sufficient for driving the cell cycle in all cell types [
25,
54]. CDK1 regulates chromosome condensation and microtubule dynamics to facilitate the transition from G
2 to M phase. Goga
et al. reported that CDK1 inhibition resulted in a synthetic lethality in mouse lymphoma and hepatoblastoma with MYC hyper-activation [
22]. We showed here that breast cancer cells were also selectively sensitive to CDK1 inhibition. The different sensitivity did not appear to be related to CDK1 expression, since CDK1 expression did not vary markedly between the cell lines used here, consistent with previous data showing that there are not large variations in CDK1 expression in breast cancer cell lines [
55]. Instead, the sensitivity to CDK1 inhibition appeared to reflect a synthetic lethal interaction between MYC and CDK1 [
22,
46]. One potential mechanism for this synthetic lethal interaction is that loss of CDK1 leads to substantial mitotic catastrophe [
56], which possibly increases MYC-induced replication stress, and subsequently activates checkpoint signalling, resulting in cell death. Thus cells harbouring MYC hyper-activation might be more vulnerable to mitotic disruption. Indeed, several MYC synthetic lethal genes identified in recent studies, including aurora-B kinase, CHK1/2 and SUMO-activating enzyme, are all involved in maintaining mitotic fidelity [
20,
23,
57]. Moreover, high-throughput screens display enrichment of the components of the mitotic spindle among MYC synthetic lethal candidates [
19,
20]. Therefore, specific targeting of CDK1 might be effective for breast tumours dependent on MYC activation and this synthetic lethal strategy may overcome some problems with side effects induced by CDK1 inhibition.
Previous studies showed that up-regulation of BIM expression was required for MYC overexpression-induced apoptosis [
58] and contributed to cell death induced by the CDK inhibitor purvalanol A in breast cancer cells [
46]. Consistent with these studies, we found that CDK1 inhibitors induced BIM expression and that elevated BIM expression was associated with increased sensitivity to CDK1 inhibitors in cells with high MYC expression. p53, however, appears to be dispensable for increased cell apoptosis induced by CDK1 inhibitors, although loss of p53 has been reported to reduce cell apoptosis associated with MYC overexpression [
59]. p53-independent apoptosis was also observed in MYC-overexpressing mouse embryo fibroblast cells treated with purvalanol A [
22]. Therefore, specific apoptotic pathways appear to be involved in CDK1 inhibitor-induced MYC-dependent cell death, providing a mechanistic insight into MYC-CDK1 synthetic lethality in breast cancer cells.
Acknowledgements
This research was supported by the National Health and Medical Research Council of Australia (535903, 427601), the Cancer Institute NSW (11/CDF/3-26, 09/RIG/1-18), the Australian Cancer Research Foundation (ACRF Unit for the Molecular Genetics of Cancer), the Petre Foundation, and the RT Hall Trust. RLS was a NHMRC Senior Principal Research Fellow. EAM was a Cancer Institute NSW Fellow.
We thank Gillian Lehrbach and Christine Lee for their help with tissue culture and Drs Kim Moran-Jones and Radhika Nair for helpful discussions and comments on the manuscript. We would also like to mark the contribution of Robert L. Sutherland, who sadly passed away in October 2012.
Competing interests
The authors declared that they have no competing interests.
Authors’ contributions
JK designed the study, carried out the experiments, analysed the data, drafted the manuscript and read and approved the final manuscript. CMS carried out the experiments and read and approved the final manuscript. RLS conceived of the study, and participated in its design and coordination. EAM conceived of the study, participated in its design and coordination, helped to draft the manuscript and read and approved the final manuscript.