WO2009063175A1 - Methods for determining resistance to cancer therapy - Google Patents

Abstract

Methods and kits for assessing breast cancer in an individual are disclosed, in particular for determining the likelihood of the individual responding to endocrine therapy in estrogen receptor α (ERα ) positive breast cancer. This is based on the finding that methylation of cyclin dependent kinase (10) (CDK10), and in particular methylation of the CDK10 promoter, or reduced CDK10 expression is associated with an increased likelihood that the individual will not respond to endocrine therapy.

Description

Methods for Determining Resistance to Cancer Therapy

Field of the Invention

The present invention relates to methods for determining resistance to cancer therapy, and in particular to methods for determining whether a patient is likely to develop resistance to endocrine therapy used to treat breast cancer. These methods may be useful for determining the likelihood of an individual responding to endocrine therapy and therefore for helping to determine appropriate treatments for patients with breast cancer.

Background of the Invention

Approximately 70% of all breast tumours express estrogen receptor α (ERa) and, of these, the majority are dependent on estrogen signalling for their growth (EBCTCG, 1998) . ERa acts as a hormone-dependent nuclear transcription factor,- binding of estrogen to ERa causes dimerization of the receptor and its translocation to the nucleus where the receptor binds Estrogen Responsive Elements (EREs) located in the promoter region of ERa regulated genes. This ER/ERE interaction leads to the recruitment of co-factors that facilitate gene transcription (Mangelsdorf et al, 1995) . Since most ERa positive breast cancers are fully dependent on estrogen signalling for their growth, they can be treated with anti-estrogens or aromatase inhibitors that limit the synthesis of estrogen. Tamoxifen is the most commonly used endocrine therapy and has shown significant patient benefit in the treatment of ERa positive breast cancer (EBCTCG, 1998) . Tamoxifen binds to ERa, blocking the interaction between estrogen and its receptor. This inhibition of ERa leads to suppression of cell cycle progression and arrest at the G₁ cell cycle checkpoint, limiting cellular proliferation and tumour progression (Musgrove et al, 1994; Cariou et al, 2000) . Despite its widespread use, the effectiveness of tamoxifen is limited by the development of drug resistance; all patients with metastatic disease and 40% of early stage breast cancer patients treated with adjuvant tamoxifen, eventually relapse with tamoxifen-resistant disease (Jordan, 1995; Ring and Dowsett, 2004) .

Aromatase inhibitors, which block the synthesis of estrogen, have recently been shown to have superior efficacy and improved tolerability when compared to tamoxifen (Baum et al, 2003) . As such these agents are increasingly used for endocrine treatment of post-menopausal hormone receptor-positive breast cancer patients with both early and advanced disease. Initially it was hoped that resistance to these agents would be less frequent than for tamoxifen but up to 50% of treated patients develop resistance to aromatase inhibitors (Anderson et al, 2007), again limiting effectiveness .

Endocrine therapy resistance has been the subject of intense study. However, the molecular mechanisms that underlie' resistance are not fully understood and as such, effective approaches for preventing and overcoming resistance are not yet available. Two major mechanisms have been proposed by which resistance may occur. First, continued ERa signalling in the presence of an ERa antagonist or the absence of estrogen may occur, an effect termed ligand-independent ERa activation (Shou et al, 2004) . Second, the reliance of tumours upon ERa signalling may be circumvented by the activation of non-ERα growth-promoting pathways (Stoica et al, 2000; Tang et al, 1996; El-Ashry et al, 1997; Oh et al, 2001) . The activity of signal transducing kinases has been implicated in both of these mechanisms (Gee et al, 2001; Perez-Tenorio et al, 2002) . Considerable effort has been expended on the characterisation of individual kinases in endocrine therapy resistance with the notable findings that PAKl and AKT activation can cause resistance to tamoxifen in breast cancer models and tumours (Holm et al, 2006; Rayala et al, 2006; Campbell et al, 2001; Perez- Tenorio et al, 2002) .

In summary, therapies that target estrogen signalling have transformed the treatment of breast cancer, but the effectiveness of anti-endocrine agents is limited by the frequent development of resistance. As the precise mechanisms underlying the response to tamoxifen and tamoxifen resistance remain incompletely understood, it remains a problem in the art to determine whether a patient with breast cancer will respond to endocrine therapy and whether they will develop resistance to endocrine therapy.

Summary of the Invention

Broadly, the present invention relates to methods and kits for assessing breast cancer in an individual, and more particularly for determining the likelihood of the individual responding to endocrine therapy in estrogen receptor α (ERa) positive breast cancer. The present invention is based on the finding that methylation of cyclin dependent kinase 10 (CDKlO) , and in particular methylation of the CDKlO promoter, or reduced CDKlO expression is associated with an increased likelihood that the individual will not respond to endocrine therapy.

The present invention is based on the experiments described herein in which a high throughput siRNA screen was used to identify kinase modifiers of tamoxifen sensitivity. This demonstrated that CDKlO is an important determinant of sensitivity to anti-endocrine agents and showed that CDKlO silencing results in upregulation of the MAPK signalling pathway, circumventing the reliance of breast cancer cells on estrogen signalling for survival and proliferation. The experiments further described a novel signalling axis by which CDKlO silencing increases ETS2 driven transcription of the c-RAF gene resulting in upregulation of the MAPK signalling pathway. The clinical significance of these findings is indicated by the demonstration that patients with ERa positive tumours that express low levels of CDKlO relapse early on tamoxifen. Furthermore, low levels of CDKlO are strongly associated with methylation of the CDKlO promoter suggesting a mechanism by which CDKlO expression is reduced in tumours.

The present invention is based on the results of a study in which simultaneous analysis of all protein kinases was employed to uncover potential novel mechanisms of resistance. Approaches such as high-throughput KNA interference (RNAi) screening now allow such systematic analysis to be performed (Iorns et al, 2007) . RNAi is a naturally occurring mechanism that controls gene expression at the post-transcriptional level. The incorporation of short interfering RNAs (siRNAs) and complementary mRNA transcripts into the RNA-induced silencing complex (RISC) , allows specific mRNAs to be degraded, resulting in silenced gene expression and reduced protein production

(Meister and Tuschl, 2004) . This characteristic of RNAi makes it a valuable laboratory research tool to selectively silence specific proteins in mammalian cells . The development of RNAi libraries, which are composed of reagents that allow the targeting of a wide range of transcripts, has made it possible to conduct high throughput screens that interrogate phenotypes associated with the loss-of-function of many genes (Iorns et al, 2007) .

Accordingly, in a first aspect, the present invention provides a method for determining whether an individual having estrogen receptor oc (ERa) positive breast cancer is likely to show resistance to endocrine therapy, the method comprising: determining cyclin dependent kinase 10 (CDKlO) status in a sample obtained from the individual; and using the CDKlO status to determine whether the individual is likely to respond to endocrine therapy.

In the present invention, the determination of CDKlO status may involve one or more of the following steps:

(a) determining the methylation of the promoter of the CDKlO gene, and correlating methylation of the CDKlO promoter with resistance to endocrine therapy; and/or

(b) determining the expression of the CDKlO gene and correlating reduced expression of the CDKlO gene with resistance to endocrine therapy; and/or

(c) determining the expression of the CDKlO protein and correlating reduced expression of the CDKlO protein with resistance to endocrine therapy.

These assays are discussed in more detail below. By way of reference, the nucleic acid sequence of the CDKlO promoter and gene are provided at Entrez GenelD 8558, NCBI Accession number NC_000016. The CDKlO gene encodes three isoforms NMJD52988 (isoform a), NM_052987 (isoform b) and NM_001098533 (isoform c) . The amino acid sequence of the CDKlO protein isoforms are provided at NCBI Accession numbers NP_443714 (isoform a) ,

NP_443713 (isoform b) and NP_001092003 (isoform c) . Where CDKlO nucleic acid is used in the methods disclosed herein, it preferably comprises or consists of genomic DNA.

Preferably, the methods of present invention further comprise the step of determining whether the breast cancer is ERa positive breast cancer. ERa positivity is routinely assessed using a biopsy taken from the breast tumour by a pathologist and ERa expression may be assessed by immunohistochemistry using an ERa antibody, or using one of the other techniques for determining nucleic acid or protein expression discussed in more detail below with reference to CDKlO expression.

In accordance with normal scientific practise, the methods disclosed herein preferably determine CDKlO status with reference to one or more controls .

As indicated above, the main purpose of the methods of the invention is for determining the type of clinical treatment given to the individual, for example selecting a type of chemotherapy or a chemotherapy regimen for administration to the individual that is likely to have the maximum therapeutic benefit to the individual. Accordingly, when a test on a sample from an individual shows an absence of CDKlO promoter methylation or normal CDKlO gene or protein expression, the individual is treated using endocrine therapy, for example including treatment with tamoxifen, faslodex and/or aromatase inhibitors. Alternatively, when a test on a sample from an individual shows CDKlO promoter methylation or a reduced level of CDKlO gene or protein expression, the individual is unlikely to respond to endocrine therapy and could be treated with signal transduction inhibitors, in particular inhibitors of the MAPK signalling pathway which is activated when CDKlO expression is low.

In aspects of the present invention relating to methylation of CDlO nucleic acid, this may be determined using any suitable method known in the art. Examples of such techniques are discussed in more detail below and include using methylation specific PCR, bisulphite sequencing, hybridisation to a bead array or isolating and sequencing methylated DNA. In one preferred embodiment, the determination of methylation of the CDKlO promoter may performed on genomic nucleic acid extracted from a sample of cells obtained from the breast cancer or from a sample of cancer cells circulating in blood.

Additionally or alternatively, in embodiments of the invention relating to the determination of the expression of the CDKlO gene, the method may comprise extracting RNA from a sample of breast cancer cells and measuring expression by real time PCR or by using a probe capable of hybridising to CDKlO RNA. The probe may conveniently be immobilised as a sequence present in a microarray.

Additionally or alternatively, in embodiments of the invention relating to the determination of CDKlO protein expression, the method may comprise detecting expression in a tumour sample using immunohistochemistry, measuring CDKlO protein levels in a cell lysate by ELISA or measuring CDKlO protein levels using western blotting. The methods for determining CDKlO protein expression may use a binding agent capable of specifically binding to the CDK protein, or a fragment thereof.

All of the methods disclosed herein may comprise an initial step of obtaining a sample from said individual. Examples of samples that may be used include a tumour sample, a blood sample, a tissue sample or a cell sample.

In a further aspect, the present invention provides the use of a cyclin dependent kinase 10 (CDKlO) status for determining whether an individual having estrogen receptor α (ERa) positive breast cancer is likely to be resistant to endocrine therapy.

In a further aspect, the present invention provides a kit for determining whether an individual having estrogen receptor oc

(ERa) positive breast cancer is likely to respond to endocrine therapy according to a method as disclosed herein.

By way of example, a preferred kit may comprise reagents necessary^' for carrying out the determination of CDKlO status of a sample and instructions for carrying out the test and interpreting the results . Thus , the kit may include one or more of the following reagents :

(a) an antibody capable of recognising CDKlO polypeptides or fragments thereof; and/or

(b) primers based on the nucleic acid sequence of the CDKlO for detecting the presence of CDKlO mRNA; and/or

(c) a probe based on the nucleic acid sequence of the CDKlO gene for detecting CDKlO gene expression; and/or (d) reagents for determining the methylation status of the promoter of the CDKlO gene.

In a further aspect, the present invention provides a demethylating agent for use treating estrogen receptor α (ERa) positive breast cancer, e.g. where the breast cancer is characterised according to its CDKlO status as defined herein. Examples of demethylating agents include 5-azacytidine and decitabine .

in a further aspect, the present invention provides the use of a demethylating agent for the preparation of a medicament for the treatment a patient having breast cancer, wherein the breast cancer is estrogen receptor α (ERa) positive and characterised by a CDKlO status as defined herein (esp. where the CDKlO promoter is methylated) .

In a further aspect, the present invention provides a demethylating agent for use in a method of treating a patient having breast cancer, wherein the breast cancer is estrogen receptor α (ERa) positive and characterised by a positive CDKlO status .

In a further aspect, the present invention provides a method of treating a patient in need having estrogen receptor α (ERa) positive breast cancer, the method comprising administering a therapeutically effective amount of a demethylating agent. In some embodiments, the method may comprise one or more of the steps above to determine whether the breast cancer is likely to be resistant to endocrine therapy.

The present invention includes the combination of the aspects and preferred features described except where such a combination is clearly impermissible or is stated to be expressly avoided. Embodiments of the present invention will now be described by way of example and not limitation with reference to the accompanying figures .

Brief Description of the Figures

Figure 1. A tamoxifen resistance screen with a protein kinase siENA library.

A. High-Throughput Screen (HTS) method. MCF7 cells plated in 96 well plates were transfected with siRNA. Each transfection plate contained 80 experimental siRNAs (SMARTpools of four different siRNA targeting the same gene) supplemented with ten wells of non-targeting siControl. Transfected cells were divided into four replica plates, half treated with ethanol vehicle alone and half with 4OH tamoxifen at 5OnM, the SF80 of MCF7. Cell viability was assessed after seven days of 4OH tamoxifen exposure using CellTiter-Glo Luminescent Cell Viability Assay (Promega) . B. Reproducibility of HTS method. Correlation of the effect of siRNA on cell growth in vehicle treated plates from two replicates of the entire screen. Spearman correlation coefficient, r2=0.71. C. Scatter plot of averaged Z scores from tamoxifen resistance screen carried out in duplicate. Black diamonds - siRNA SMARTpools targeting 779 protein kinase genes, Red diamonds - siControl. D. Validation of resistance hits from the tamoxifen HTS. Tamoxifen sensitivity assay repeated in triplicate with the four different siRNAs originally in each SMARTpool and the SMARTpool, all targeting the same kinase. Surviving fractions following tamoxifen treatment are shown, including those after transfection with siControl (Red) . CRK7 and MAP2K7 revalidate with all 4 siRNA; CDKlO revalidates with 3 siRNA. * - p<0.05 compared to siControl (Student's t-test) . Error bars represent the standard error of the mean (SEM) .

Figure 2. CDKlO silencing decreases sensitivity to tamoxifen and estrogen deprivation .

A. Real time quantitative PCR on cDNA prepared from MCF7 cells transfected with siRNA. Individual siRNAs 1, 3, 4 and the CDKlO SMARTpool significantly reduced CDKlO mRNA levels compared to siControl transfected cells. * - p<0.05 compared to siControl (Student's t-test) . Error bars represent the standard error of the mean (SEM) .

B. Western blot analysis of lysates prepared from MCF7 cells transfected with pReceiver CDKlO HA and siRNA. Antibodies recognising CDKlO were used with β-tubulin as a loading control. Individual siRNAs 1, 3, 4 and the CDKlO SMARTpool significantly reduced CDKlO protein expression compared to siControl transfected cells.

C. Cell viability assay in MCF7 cells transfected with CDKlO SMARTpool, individual CDKlO siRNAs 1 and 4, or siControl and treated with 4OH tamoxifen. CDKlO silenced cells have significantly decreased sensitivity to 4OH tamoxifen. Error bars represent the SEM. D. Cell viability assay in MCF7 cells transfected with CDKlO SMARTpool, individual CDKlO siRNAs 1 and 4, or siControl and deprived of estradiol. CDKlO silenced cells have significantly decreased sensitivity to estrogen deprivation. * - p<0.05 compared to siControl (Student's t-test) . Error bars represent the SEM.

Figure 3. CDKlO silencing reduces tamoxifen induced Gl arrest.

A. Resistance to tamoxifen in CDKlO silenced cells is characterised by a decreased tamoxifen induced Gl arrest. FACS plots showing Propidium Iodide (PI) after treatment with 4OH tamoxifen and either CDKlO SMARTpool or siControl transfection.

B. CDKlO SMARTpool treated cells had significantly less tamoxifen induced Gl arrest and significantly less S phase decrease than siControl treated cells . % Gl arrest was calculated by subtracting the % of cells in Gl before 4OH tamoxifen treatment from the % of cells in Gl after 4OH tamoxifen treatment. % S phase decrease was calculated by subtracting the % of cells in S phase before 4OH tamoxifen treatment from the % of cells in S phase after 4OH tamoxifen treatment. * - p<0.05 compared to siControl (Student's t-test) . Error bars represent the standard error of the mean (SEM) .

C. CDKlO SMARTpool transfected cells have significantly higher cyclin Dl mRNA levels than siControl transfected cells. Real time quantitative PCR on cDNA prepared from MCF7 transfected with siRNA for 48 hours followed by treatment with InM estradiol, or no treatment for 24 hours. Error bars represent the SEM.

D. CDKlO SMARTpool transfected cells have significantly higher expression of cyclin Dl protein than siControl transfected cells. MCF7 cells transfected 48 hours earlier with siRNA were treated with InM estradiol, InM estradiol plus 10OnM 4OH tamoxifen, or no treatment. Lysates were made 24 hours following treatment. Antibodies recognising cyclin Dl were used with β-tubulin as a loading control . E. CDKlO SMARTpool transfected cells have significantly higher levels of phosphorylated Rb than siControl transfected cells. MCF7 cells transfected 48 hours earlier with siRNA were treated with InM estradiol, InM estradiol plus 10OnM 4OH Tamoxifen, or no treatment. Lysates were made 24 hours following treatment. Antibodies recognising phospho Rb and total Rb were used with Ezrin as a loading control .

Figure 4. CDKlO silencing does not cause ligand independent ERa activation.

A. CDKlO SMARTpool transfected cells do not have altered activation or expression of ERa compared with siControl transfected cells.

Left panel: MCF7 cells transfected 48 hours earlier with siRNA were treated with InM estradiol, InM estradiol plus 10OnM 4OH tamoxifen, or no treatment. Lysates were made 24 hours following treatment. Antibodies recognising phosphoserine 118 and total ERa were used with β-tubulin as a loading control .

Right panel : Real time quantitative PCR on cDNA prepared from MCF7 transfected with siRNA for 48 hours followed by treatment with InM estradiol, InM estradiol plus 10OnM 4OH tamoxifen, or no treatment for 24 hours. Error bars represent SEM. B. CDKlO SMARTpool transfected cells do not have increased expression of the ERa regulated gene PR compared with siControl transfected cells.

Left panel: MCF7 cells transfected 48 hours earlier with siRNA were treated with InM estradiol, InM estradiol plus 10OnM 4OH tamoxifen, or no treatment. Lysates were made 24 hours following treatment. An antibody recognising progesterone receptor was used with β-tubulin as a loading control.

Right panel : Real time quantitative PCR on cDNA prepared from MCF7 transfected with siRNA for 48 hours followed by treatment with InM estradiol, InM estradiol plus 10OnM 4OH tamoxifen, or no treatment for 24 hours. Error bars represent SEM. C. CDKlO SMARTpool transfected cells have significantly lower expression of ERa regulated gene TFFl compared with siControl transfected cells. Left panel: MCF7 cells transfected 48 hours earlier with siRNA were treated with InM estradiol, InM estradiol plus 10OnM 4OH tamoxifen, or no treatment. Lysates were made 24 hours following treatment. An antibody recognising TFFl was used with β-tubulin as a loading control .

Right panel : Real time quantitative PCR on cDNA prepared from MCF7 transfected with siRNA for 48 hours followed by treatment with InM estradiol, InM estradiol plus 10OnM 4OH tamoxifen, or no treatment for 24 hours. Error bars represent SEM. D. CDKlO silencing decreases sensitivity to ICI 182780. Cell viability assay in MCF7 cells transfected with CDKlO SMARTpool, individual CDKlO siRNAs 1 and 4, or siControl and treated with ICI 182780. CDKlO silenced cells have significantly decreased sensitivity to ICI 182780. Error bars represent the SEM.

Figure 5. CDKlO silencing activates the p42/p44 MAPK pathway, causing resistance to tamoxifen. A. CDKlO silenced cells have activated p42/p44 MAPK. MCF7 cells transfected 48 hours earlier with siRNA were treated with InM estradiol, InM estradiol plus 10OnM 4OH tamoxifen, or no treatment. Lysates were made 24 hours following treatment. Antibodies recognising phospho and total p42/p44 MAPK were used with β-tubulin as a loading control.

B. CDKlO silenced cells have activated MEKl, 2. MCF7 cells transfected 48 hours earlier with siRNA were treated with InM estradiol, InM estradiol plus 10OnM 4OH tamoxifen, or no treatment. Lysates were made 24 hours following treatment. Antibodies recognising phospho and total MEKl, 2 were used with β- tubulin as a loading control .

C. Silencing of p42/p44 MAPK restores 4OH tamoxifen sensitivity to CDKlO silenced cells.

Left panel: Cell viability assay in MCF7 cells transfected with CDKlO SMARTpool or siControl, alone, or in combination with p42/p44 MAPK SMARTpool siRNAs and treated with 4OH tamoxifen.

Error bars represent SEM.

Right panel: Western blot analysis of lysates prepared from MCF7 cells transfected with siRNA. An antibody recognising p42/p44 MAPK was used with β-tubulin as a loading control. The p42/p44

MAPK SMARTpools significantly reduced p42/p44 MAPK protein expression compared to siControl transfected cells. D. Silencing of MEKl, 2 restores 4OH tamoxifen sensitivity to CDKlO silenced cells.

Left panel: Cell viability assay in MCF7 cells transfected with CDKlO SMARTpool or siControl, alone, or in combination with MEKl, 2 SMARTpool siRNAs and treated with 4OH tamoxifen. Error bars represent SEM.

Right panel: Western blot analysis of lysates prepared from MCF7 cells transfected with siRNA. An antibody recognising MEKl, 2 was used with β-tubulin as a loading control. The MEKl, 2 SMARTpools significantly reduced MEKl, 2 protein expression compared to siControl transfected cells .

Figure 6. Activation of the p42/p44 MAPK pathway in CDKlO siRNA transfected cells is due to overexpression of c-RAF. A. CDKlO silenced cells have higher levels of phosphorylated and total c-RAF. MCF7 cells transfected 48 hours earlier with siRNA were treated with InM estradiol, InM estradiol plus 10OnM 4OH tamoxifen, or no treatment. Lysates were made 24 hours following treatment. Antibodies recognising phospho and total c-RAF were used with β-tubulin as a loading control.

B. CDKlO silenced cells have higher levels of c-RAF mRNA. Real time quantitative PCR on cDNA prepared from MCF7 transfected with siRNA for 48 hours followed by treatment with InM estradiol, InM estradiol plus 10OnM 4OH tamoxifen, or no treatment for 24 hours. Error bars represent SEM.

C. CDKlO silenced cells do not have higher levels of activated Ras. Ras activation assay on lysates prepared from MCF7 transfected with siRNA for 72 hours .

D. Overexpression of wild type c-RAF increases the levels of phosphorylated c-RAF. MCF7 cells were transfected with pEF wild type c-RAF or pEF empty vector control and lysates were made 24 hours later. Antibodies recognising phospho and total c-RAF were used with β-tubulin as a loading control .

E. Overexpression of wild type c-RAF increases the levels of phosphorylated p42/p44 MAPK. MCF7 cells were transfected with pEF wild type c-RAF or pEF empty vector control and lysates were made 24 hours later. Antibodies recognising phospho and total p42/p44 MAPK were used with β-tubulin as a loading control .

Figure 7. Increased c-RAF expression in CDKlO silenced cells is regulated by ETS2 transcription factor.

A. CDKlO binds to ETS2. MCF7 cells were transfected with pReceiver CDKlO HA or the empty vector control pReceiver empty HA. Lysates were made 24 hours later and CDKlO was immunoprecipitated with HA conjugated beads. ETS2 binding was detected by western blotting of the immunoprecipitated lysate.

B. Diagram illustrating an ETS2 consensus binding sequence identified in the c-RAF promoter using TFMATRIX (Score 96.0) .

C. CDKlO and ETS2 bind to the c-RAF promoter at the ETS2 binding site. MCF7 cells transfected with pReceiver CDKlO HA or empty vector control pReceiver HA for 24 hours were used for ChIP.

Binding of ETS2 and CDKlO tagged with HA were detected at the c- RAF promoter by PCR. No HA was detected by PCR in the empty vector control pReceiver HA lysates . No binding of normal rabbit IgG was detected by PCR indicating a specific CDKlO and ETS2 interaction. No binding was detected by control PCR further demonstrating specificity of the CDKlO, ETS2 interaction with the c-RAF promoter.

D. Expression of c-RAF in CDKlO silenced cells can be partially restored to normal levels by silencing of ETS2. Left panel: Real time quantitative PCR on cDNA prepared from MCF7 transfected with siRNA for 48 hours followed by treatment with InM estradiol, InM estradiol plus 10OnM 4OH tamoxifen, or no treatment for 24 hours. Error bars represent SEM. Right panel : Real time quantitative PCR analysis on cDNA prepared from MCF7 cells transfected with siRNA. ETS2 SMARTpool significantly reduced ETS2 expression compared to siControl transfected cells .

E. Levels of phosphorylated p42/p44 MAPK and c-RAF protein can be partially restored to normal levels by silencing of ETS2. MCF7 cells transfected 48 hours earlier with siRNA were treated with InM estradiol, InM estradiol plus 10OnM 4OH tamoxifen, or no treatment. Lysates were made 24 hours following treatment. Antibodies recognising phospho and total p42/p44 MAPK and total c-RAF were used with β-tubulin as a loading control.

Figure 8. Low CDKlO expression is associated with significantly poorer survival in patients treated with tamoxifen.

A. Kaplan Meier survival curves of breast cancer patients from data set 1, n=87, with tumours that have low CDKlO expression (defined as lowest quartile expression) were at significantly higher risk of distant relapse, hazard ratio 2.447, 95% confidence interval 1.192 to 8.208, p=0.0205 (Logrank test) when treated with adjuvant tamoxifen.

B. Kaplan Meier survival curves of breast cancer patients from data set 1, n=87, with tumours that have low CDKlO expression had a trend of reduced overall survival, hazard ratio 1.992, 95% confidence interval 0.9053 to 6.048, p=0.0793 (Logrank test) when treated with adjuvant tamoxifen. This is not statistically significant but shows a trend to significance.

C. Kaplan Meier survival curves of breast cancer patients from data set 2, n=38, with tumours that have low CDKlO expression (defined as lowest quartile expression) were at significantly higher risk of distant relapse, (hazard ratio not definable) , p<0.0001 (Logrank test) when treated with adjuvant tamoxifen.

D. Kaplan Meier survival curves of breast cancer patients from data set 2, n=38, with tumours that have low CDKlO expression have significantly reduced overall survival, hazard ratio 13.56, 95% confidence interval 11.82 to 355.6, p<0.0001 (Logrank test) when treated with adjuvant tamoxifen.

E. Methylation status of the CDKlO promoter in data set 2 was significantly associated with expression of CDKlO. Median CDKlO expression in methylated samples = 0.13, median CDKlO expression in unmethylated samples = 1.32, p<0.0001 (Mann Whitney test) .

Figure 9. Methylation of the CDKlO promoter is associated with low CDKlO expression and with significantly poorer survival in patients treated with tamoxifen.

Top panel. Methylation status of the CDKlO promoter in data set 2 was significantly associated with expression of CDKlO. Median CDKlO expression in methylated samples = 0.13, median CDKlO expression in unmethylated samples = 1.32, p<0.0001 (Mann Whitney test) .

Bottom left panel . Kaplan Meier survival curves of breast cancer patients from data set 2, n=38, with tumours that have CDKlO promoter methylation were at significantly higher risk of distant relapse, p<0.0001 (Logrank test) when treated with adjuvant tamoxifen.

Bottom right panel . Kaplan Meier survival curves of breast cancer patients from data set 2, n=38, with tumours that have CDKlO promoter methylation have significantly reduced overall survival, p<0.0001 (Logrank test) when treated with adjuvant tamoxifen.

Detailed Description DNA methylation

In some embodiments, the present invention relates to methods for determining whether an individual is likely to develop resistance to endocrine therapy used for treating breast cancer by analysis of DNA methylation. In particular, the results described herein establish that methylation of the CDKlO gene, and in particular the promoter of the CDKlO gene, correlate with low expression of CDKlO and an increased likelihood of developing resistance to endocrine therapy for breast cancer. Analysis of CDKlO promoter methylation therefore allows the probable clinical progression, or prognosis of breast cancer in a patient to be assessed.

Nucleic acid methylation generally occurs in vertebrates at cytosine bases, especially where the cytosine is flanked by a guanine base (CpG) . Methylation occurs at CpG islands in the promoter region. A CDKlO promoter nucleic acid may include all or part of the CDKlO promoter/gene (Entrez GeneID 8558, NCBI Accession number NC_000016 and Genome sequence NT 010542.15), which has the coding sequence of NCBI Accession numbers NP_443714 (isoform a), NP_443713 (isoform b) and NP_001092003 (isoform c) . A CDKlO nucleic acid preferably comprises or consists of genomic DNA. Preferably CDKlO promoter methylation is determined within a part of the CDKlO promoter (a CpG island) within the genomic sequence NT 010542.15.

A range of different techniques for examining whether or not a particular nucleic acid sequence is methylated are known in the art and may be employed for determining methylation of CDKlO nucleic acid according to the present invention.

By way of example, nucleic acid methylation may be determined using methylation specific PCR or bisulphite sequencing. These techniques may be performed on genomic DNA which can be extracted from a tumour sample or from circulating tumour cells present in the blood stream, i.e. from a blood sample.

Methylation Specific PCR (MSP) is a bisulfite conversion based PCR technique for the study of DNA CpG methylation. For a MSP PCR experiment, two pairs of primers are needed, with one pair specific for methylated DNA (M) and the other for unmethylated DNA (U) . To achieve discrimination for methylated and unmethylated DNA, in each primer sequence, one or more CpG sites are included. First, DNA is modified with sodium bisulfite and purified. Then, two PCR reactions are performed using M primer pair and U primer pair. Successful amplification from M pair and U pair indicated methylation and unmethylation respectively. An example of the use of this method to detect CDKlO promoter methylation is provided in the examples below.

Bisulfite Genomic Sequencing is a method that allows analysis of methylation in a certain region by converting all unmethylated cytosines into thymines, while methylated cytosines remain unchanged. This method has the advantage that it requires small amounts of genomic DNA and therefore is useful for the analysis of clinical samples, where the amount of the sample available for analysis is limited.

Alternatively or additionally, methylation of genes may also be determined by use of array technologies, such as hybridisation to a bead array, such as Illumina' s Goldengate arrays, and/or by any method that profiles methylation by isolating and then sequencing methylated DNA.

Other techniques for assessing methylation of nucleic acid may be employed that rely on the chemical modification of the CDKlO nucleic acid to convert specifically one of an unmethylated or a methylated base, for example an unmethylated or a methylated cytosine base, to another base. ^' Thus, the chemical conversion leads to a detectable difference between the methylated and unmethylated nucleic acid sequences .

The chemical modification may be achieving by treating the CDKlO nucleic acid with a modifying agent that converts specifically one of an unmethylated or a methylated base to another base. The presence or absence of a base change at a site is indicative of methylation at that site. For example, when the modifying agent modifies methylated bases, a base change is indicative of methylation and when the modifying agent modifies unmethylated bases, the absence of base change is indicative of methylation.

In some preferred embodiments, the modifying agent may convert an unmethylated base into a different base, for example an unmethylated cytosine base to a uracil base. Suitable modifying agents include bisulphite salts, such as sodium bisulphite, which converts an unmethylated cytosine base to a uracil base.

The presence of a CpG dinucleotide is indicative of methylation at the cytosine base, which protects the cytosine from the action of the modifying agent.

A base change may be determined using any convenient technique, for example, by one or more of: sequencing, hybridisation analysis and RFLP analysis . The base change is preferably detected on the sense strand of the CDKlO nucleic acid. By way of example RFLP may be used in situations in which a change in the restriction pattern may occur as a result of methylation dependent base changes in the target sequence of a restriction enzyme, or as a result of the direct blocking by methylation of the target sequence of a restriction enzyme, which may prevent cleavage or reduce the rate of cleavage by the restriction enzyme at the target sequence.

Determination of the expression of CDKlO nucleic acid Alternatively or additionally, the present invention relates to methods for determining whether an individual is likely to develop resistance to endocrine therapy used for treating breast cancer by analysis of CDKlO gene expression. In particular, the results provided herein demonstrate that reduced levels of CDKlO expression are correlated with an increased likelihood of developing resistance to endocrine therapy for breast cancer.

The determination of CDKlO gene expression may involve determining the presence or amount of CDKlO mRNA in a sample. Methods for doing this are well known to the skilled person. By way of example, they include determining the presence of CDKlO mRNA (i) using a labelled probe that is capable of hybridising to the CDKlO nucleic acid; and/or (ii) using PCR involving one or more primers based on a CDKlO nucleic acid sequence to determine whether the CDKlO transcript is present in a sample. The probe may also be immobilised as a sequence included in a microarray.

Preferably, detecting CDKlO mRNA is carried out by extracting KNA from a sample of the tumour and measuring CDKlO expression specifically using quantitative real time RT-PCR, as carried out on clinical data set 2 in the examples below. Alternatively or additionally, the expression of CDKlO could be assessed using RNA extracted from a tumour sample using microarray analysis, which measures the levels of mRNA for a group of genes using a plurality of probes immobilised on a substrate to form the array. This is the type of analysis performed on clinical data set 1 in the examples below. Determination of the expression of CDKlO protein Alternatively or additionally, the present invention relates to methods for determining whether an individual is likely to develop resistance to endocrine therapy used for treating breast cancer by analysis of CDKlO protein expression. In particular, the results provided herein demonstrate that reduced levels of CDKlO protein expression are correlated with an increased likelihood of developing resistance to endocrine therapy for breast cancer.

Preferably, the presence or amount of CDKlO protein may be determined using a binding agent capable of specifically binding to the CDKlO protein, or fragments thereof. A preferred type of CDKlO protein binding agent is an antibody capable of specifically binding the CDKlO or fragment thereof. The antibody may be labelled to enable it to be detected or capable of detection following reaction with one or more further species, for example using a secondary antibody that is labelled or capable of producing a detectable result, e.g. in an ELISA type assay. As an alternative a labelled binding agent may be employed in a western blot to detect CDKlO protein.

Alternatively, or additionally, the method for determining the presence of CDKlO protein may be carried out on tumour samples, for example using immunohistochemical (IHC) analysis. IHC analysis can be carried out using paraffin fixed samples or frozen tissue samples, and generally involves staining the samples to highlight the presence and location of CDKlO protein.

Examples

Cell lines, compounds, plasmids and siRNA

MCF7 cells were obtained from ATCC (USA) and maintained in phenol red free RPMI 1640 (Invitrogen) , supplemented with 10% dextran charcoal treated FCS (10% (v/v) ) , InM estradiol, glutamine and antibiotics. 4OH tamoxifen and estradiol were obtained from

Sigma. ICI 182780 was obtained from Tocris Bioscience UK. The pReceiver CDKlO HA plasmid (EX-Q0187-M08) was obtained from Genecopoeia (USA) . The pEF wild type c-RAF plasmid was provided by Richard Marais (ICR, UK) . MCF7 cells were transfected with SMARTpool siRNAs using Dharmafect 3 transfection reagent according to manufacturer's instructions (Dharmacon) . The protein kinase siRNA library (siARRAY - targeting 779 known and putative human protein kinase genes) was obtained in ten 96 well plates from Dharmacon (USA) . Each well in this library contained a SMARTpool of four distinct siRNA species targeting different sequences of the target transcript.

Antibodies

Antibodies targeting the following epitopes were used: CDKlO (AP7516a, Abgent) , cyclin Dl (2926, Cell Signaling, UK), phospho Rb (ser780) (9307, Cell Signaling, UK), Rb (Ab24, Abeam, UK), phospho ERa (serll8) (2511, Cell Signaling, UK) , ERa (6F11, Novacastra, UK), TFFl (H00007031-M02, Abnova) , progesterone receptor (312, Novacastra, UK), phospho p42/p44 MAPK (Thr202/Tyr204) (4377, Cell Signaling, UK) , p42/p44 MAPK (9102, Cell Signaling, UK) , phospho MEKl, 2 (ser217/221) (9124, Cell Signaling, UK) , MEKl, 2 (4694, Cell Signaling, UK), phospho c-RAF (ser338) (9427, Cell Signaling, UK) , c-RAF (sc-227, Santa Cruz, USA), ETS2 (sc-351, Santa Cruz, USA) and β-tubulin (T4026, Sigma, UK) . All secondary antibodies used for western blot analysis were HRP conjugated.

HTS method

MCF7 cells plated in 96 well plates were transfected 24 hours later with siRNA (final concentration 10OnM) , using Dharmafect 3 (Dharmacon, USA) as per manufacturer's instructions. Twenty four hours following transfection cells were trypsinised and divided into four identical replica plates. At 48 hours following transfection, two replica plates were treated with 5OnM 4OH tamoxifen in media and two replica plates with 0.05% ethanol vehicle in media. Media containing 4OH tamoxifen or vehicle was replenished after 48 hours and 96 hours, and cell viability was assessed after seven days 4OH tamoxifen exposure using CellTiter GIo Luminescent Cell Viability Assay (Promega, USA) as per manufacturer's instructions. The luminescence reading for each well on a plate was expressed relative to the median luminescence value of all wells on the plate. The screen was completed in duplicate. For each transfection the effect on cell growth and tamoxifen sensitivity were calculated. Cell growth: The effect of each individual siRNA SMARTpool on cell growth alone was calculated by dividing mean luminescence in the two replica wells treated with ethanol vehicle by mean luminescence of the replica wells transfected with siControl, and expressed as a percentage. Cell growth effect of siRNA (%) = mean (2 replica wells with siRNA)/ mean (20 replica well treated siControl) x 100. Tamoxifen sensitivity: Sensitivity to 4OH tamoxifen for each siRNA SMARTPool was assessed by calculating the surviving fraction following 4OH tamoxifen. Surviving fraction = log₂ mean (2 replica wells with 4OH tamoxifen) - log₂ mean (2 replica wells with ethanol vehicle) . The surviving fractions were centred on the median surviving fraction of all 80 SMARTpools from one 96 well plate transfection, the results from all ten siRNA plates combined and results expressed as a Z score. For the Z score the standard deviation of the screen was estimated from the median absolute deviation of all siControl wells .

Validation of HTS screen

Four distinct siRNA species targeting each gene were used to revalidate hits from the screen. A significance threshold of p<0.05 (Student's t-test) was used for each individual siRNA.

Validation of gene silencing by siRNA

Validation of RNAi gene silencing was measured by real time RT- PCR. MCF7 cells were transfected with siRNA, and RNA extracted 48 hours later with Trizol and phenol /chloroform extraction followed by isopropanol precipitation. cDNA was synthesized using Superscript III First Strand Synthesis System for RT-PCR (Invitrogen) with oligo dT as per manufacturer's instructions. Assay-on-Demand primer/probe sets were purchased from Applied Biosystems (Foster City, CA) . Real-Time qPCR was performed on the 79 ODHT Fast Real-Time PCR System (Applied Biosystems) , with endogenous control GAPDH. Standard curves were calculated for all reactions with serial dilutions of siControl transfected cells to calculate reaction efficiency. Gene expression was calculated relative to expression of GAPDH endogenous control, and adjusted relative to expression in siControl transfected cells .

Cell viability assays to measure drug sensitivity

MCF7 cells were transfected with siRNA using Dharmafect 3 as per manufacturer's instructions, divided 24 hours later into 96 well plates and exposed to various doses of drug from 48 hours post transfection. Cell viability was assessed by CellTiter GIo Luminescent Cell Viability Assay (Promega, USA) at 9 days post transfection and surviving fraction for each dose of drug assessed by dividing the luminescence value of drug treated by the luminescence value of vehicle.

Western blots

MCF7 cells were transfected with siRNA using Dharmafect 3 as per manufacturer's instructions, after 24 hours the media was refreshed with no estradiol, InM estradiol, or InM estradiol with 10OnM 4OH tamoxifen. Protein lysates were prepared after 48 hours using RIPA lysis buffer (5OnM Tris pH 8.0, 15OmM NaCl, 0.1% SDS, 0.1% DOC, 1% TritonX-100, 5OmM NaF, ImM Na₃VO₄ and protease inhibitors) . lOOμg of total cell lysate was loaded onto prefabricated 4-12% Bis-Tris gels (Invitrogen) , with full range rainbow molecular weight marker (GE Healthcare, UK) as a size reference, and resolved by SDS-PAGE electrophoresis. Proteins were transferred to nitrocellulose membrane (Bio-rad, USA) , blocked and probed with primary antibody diluted 1 in 1000 in

IxTBS-T with 5% BSA overnight at 4°C. Secondary antibodies were diluted 1 in 5000 in IxTBS-T with 5% skim milk and incubated for one hour at room temperature. Protein bands were visualised using ECL (GE Healthcare, UK) and MR or XAR film (Kodak) .

Quantitative Real Time RT-PCR

MCF7 cells were transfected with siRNA using Dharmafect 3 as per manufacturer's instructions, after 24 hours the media was refreshed with no estradiol, InM estradiol, or InM estradiol with 10OnM 4OH tamoxifen. RNA was extracted 48 hours later with Trizol and phenol/chloroform extraction followed by isopropanol precipitation. cDNA was synthesized using Superscript III First Strand Synthesis System for RT-PCR (Invitrogen) with oligo dT as per manufacturer's instructions. Assay-on-Demand primer/probe sets were purchased from Applied Biosystems (Foster City, CA) . Real-Time qPCR was performed on the 790DHT Fast Real-Time PCR System (Applied Biosystems) , with endogenous control GAPDH.

Standard curves were calculated for all reactions with serial dilutions of siControl transfected cells to calculate reaction efficiency. Gene expression was calculated relative to expression of GAPDH endogenous control, and adjusted relative to expression in siControl transfected cells.

Fluorescence Activated Cell Scanning (FACS) analysis

MCF7 cells were transfected with siRNA using Dharmafect 3 as per manufacturer's instructions, after 24 hours the media was refreshed with no estradiol, InM estradiol, or InM estradiol with 10OnM 4OH tamoxifen. After 48 hours the cells were fixed in 70% ice cold ethanol and stained with 4% propidium iodide (PI) and 10% RNase A in PBS. The sample readout was performed on the FACSCalibur (Becton Dickinson, USA) and the data was analysed using CellQuest Pro (Becton Dickinson, USA) .

Ras activation assay

MCF7 cells were transfected with siRNA using Dharmafect 3 as per manufacturer's instructions, after 24 hours the media was refreshed with RPMI 1640 without phenol red. After 48 hours the cells were lysed and the activation of Ras was measured using the Ras Activation Assay Kit according to manufacturer's instructions (Cell Biolabs, CA) .

Co-iimunoprecipitation (Co-IP)

MCF7 cells were transfected with pReceiver CDKlO HA or the empty vector control, pReceiver empty HA, using Fugene 6 as per manufacturer's instructions. A fter 24 hours the cells were lysed in lysis buffer (5OmM HEPES pH 7.5, 15OmM NaCl, 0.25% (v/v) Igepal CA-630, 2mM EDTA, 5OmM WaF, 15mM Na₄P₂O₇, ImM Na₃VO₄, 0.5mM PMSF, 2mM benzamidine and protease inhibitor cocktail (Roche) ) for 30 minutes on ice, followed by brief sonication. Total cell lysate was clarified by centrifugation and quantified. 5mg of total protein for each lysate was pre-cleared with protein G sepharose (lOOμL) for 1 hour at 4°C. The supernatant was then collected and bound to monoclonal anti-HA agarose conjugate (100 μL, clone HA-7, Sigma Aldrich, A2095) overnight at 4°C.

Immunocomplexes were then washed with 8 x 0.5 iriL wash buffer (as for lysis buffer, with 0.1% (v/v) Igepal CA-630) by gravity flow and eluted with 0.2M glycine, pH 2.5 and ImM EDTA (500μL) . Collected fractions were concentrated to 25μL and loaded onto SDS-PAGE, along with total cell lysates (lOOμg) , followed by electroblotting onto nitrocellulose. The membrane was then probed with anti-ETS2 antibody (1:1000) .

Identification of ETS2 binding site The ETS2 binding site in the c-RAF promoter was identified using the TFMATRIX transcription factor binding site database (Wingender, 1996) .

Chromatin IP (ChIP) assay MCF7 cells were transfected with pReceiver CDKlO HA or the empty vector control, pReceiver empty HA, using Fugene 6 as per manufacturer's instructions. After 24 hours the cells were lysed and the ChIP Assay was performed using the Chip Assay Kit according to manufacturer's instructions (17-295, Upstate, UK). Antibodies used for immunoprecipitation were HA (sc-805, Santa

Cruz, USA), ETS2 (sc-351, Santa Cruz, USA) and normal rabbit IgG (sc-2027, Santa Cruz, USA) . PCR primers were designed to flank the putative ETS2 binding site of the c-RAF promoter at position -545.

Patients and clinical tissue samples for MSP and gPCR

Primary breast cancer samples were obtained as paraffin-embedded tissue sections. Tissues were obtained with fully informed consent and local ethical committee approval. In all cases, the diagnosis and adequate representation of tumour cells were confirmed by independent histopathology review. Patients underwent surgery (mastectomy, quadrantectomy or lumpectomy) plus axillary nodal dissection. Expression of the oestrogen receptor (ER) and progesterone receptor (PR) were determined by immunocytochemistry. The hormone receptor positive patients, who comprise the population described in this study, received tamoxifen 20 mg/day for 5 years. Genomic DNA was isolated from 5μm tissue sections by treatment with xylene to remove paraffin wax, followed by extended incubation in lysis buffer containing lOOμg/ml proteinase K/0.5% SDS and extraction with phenol. The resulting DNA solution from these paired biopsies was used directly for bisulfite modification and methylation-specific PCR. Expression of CDKlO was analysed by gPCR. RNA isolation and cDNA synthesis was as described previously (Crighton et al, 2006) . DNA was subjected to modification with sodium bisulphite as described previously (Smith et al, 2007) . Methylation specific PCR was performed with primer pairs encompassing the CpG island located at the 5' end of the CDKlO gene corresponding to nucleotides 1313375-1314181 of the genomic sequence NT 010542.15. CDKlO MSP Primers: Primer set 1 Left M primer ATCGTTGTTAAGGAGAGGAAGTTC Right M primer CGCGAAAAACTCTAAAACTATCG

Left U primer ATTGTTGTTAAGGAGAGGAAGTTTG

Right U primer CACAAAAAACTCTAAAACTATCATT

Primer set 2

Left M primer GGGATTTGGGAAGAGTAAGTTTC

Right M primer CACAAACAAATAACTCGCGAC

Left U primer GGGATTTGGGAAGAGTAAGTTTT Right U primer CCCACAAACAAATAACTCACAAC Primer set 3

Left M primer GTATATTGGTCGAGTTGTTGGC

Right M primer CGACCAAATATTCTCACTAAACGTA

Left U primer ATATTGGTTGAGTTGTTGGTGT

Right U primer ACCAACCAAATATTCTCACTAAACATA

M- methylated primer U- Unmethylated primer Results

Tamoxifen resistance high throughput siRNA screen (HTS) To identify non-redundant determinants of tamoxifen resistance we designed a robust, high throughput RNA interference screen (Figure IA) targeting 779 known and putative kinases. In summary, this screen initially involved transfecting an ERa positive, tamoxifen sensitive, breast cancer cell line (MCF7) with a 96 well plate arrayed library of siRNA duplexes that enable gene-specific silencing. Twenty-four hours after transfection, cells were divided into replica plates and half were treated with 4OH tamoxifen (the active tamoxifen metabolite) and half with vehicle (Figure IA) . Seven days later, cell viability in each plate was measured to assess the effect of each siRNA on cell growth (as assessed in the vehicle-treated plates) and also the effect of exposure to 4OH tamoxifen. A medium-term time course of drug treatment (seven days) was chosen as opposed to short-term treatments (1-2 days) commonly used in chemoresistance screens to increase the sensitivity of the screen. For the screen we used a library of siRNA arrayed as SMARTpools. Each SMARTpool (contained within one well of a 96 well plate) was composed of four distinct siRNA species targeting different sequences of the target transcript.

We validated the performance of MCF7 cells in the high-throughput format in the following ways: (i) transfection of MCF7 cells with siRNA targeting genes essential for cellular viability (such as PLKl) was observed to cause a reduction in viability of more than 90%, compared to transfection with a non-targeting siRNA, siControl, indicating that high-efficiency transfection could be achieved and (ii) transfection of MCF7 with siControl did not reduce cellular viability more than 20%, compared to mock transfected cells, indicating that these cells could be transfected without excessive non-specific toxicity, which would reduce the sensitivity of the screen. The cell numbers plated for transfection and at the division of cells into replica plates after transfection and before 4OH tamoxifen/vehicle treatment were also titrated. This ensured efficient transfection and prevented cells reaching confluency after vehicle treatment, as this would have the potential to mask 4OH tamoxifen sensitivity. We used the CeIlTitre GIo (Promega) method of cell viability measurement, as opposed to other methods (MTT, MTS assay) as this reagent was shown, in our hands, to generate the most reproducible and sensitive measurements of difference in viability and was ideally suited to the HTS method. Following optimisation, the screen was repeated in duplicate and comparison of data from each screen revealed this approach to be highly reproducible (Figure IB) . Data from the duplicate screens was combined in the final analysis (Figure 1C) .

Validation and exclusion of off-target hits

In addition to silencing target transcripts with perfect complementarity to the siRNA sequence, off-target effects can also occur (Iorns et al, 2007) . To assess the possibility that results observed in the HTS were due to off-target effects, the tamoxifen sensitivity HTS assay was repeated, using each of the four different siRNA species that comprise the SMARTpools of siRNA used in the HTS. It is generally considered that observation of a phenotype caused by two distinct siRNA species indicates that it is unlikely to be the result of an off-target effect (Echeverri et al, 2006) .

The four most potent resistance causing hits were re-examined (Figure ID) . Three were judged as likely to be on-target; Cyclin-Dependent Kinase 10 (CDKlO Entrez GeneID 8558), CDC2 related protein kinase 7 (CRK7 Entrez GeneID 51755) and Mitogen- Activated Protein Kinase Kinase 7 {MAP2K7 Entrez GeneID 5609) (Figure IE) . TTKl was not re-examined because of the excessive toxicity associated with silencing of this gene in the absence of 4OH tamoxifen. We further investigated the mechanism of resistance to tamoxifen induced by CDKlO silencing.

CDKlO silencing causes resistance to tamoxifen and estrogen deprivation Confirmation of CDKlO gene silencing by siRNA was established by quantitative real time PCR and western blotting (Figure 2A,B) . The two CDKlO targeting siRNAs that caused the most significant effects on tamoxifen sensitivity were also shown to cause the most significant CDKlO silencing. To confirm the validity of the results from the HTS, dose response curves were performed over a range of 4OH tamoxifen concentrations (Figure 2C) . CDKlO silencing significantly decreased sensitivity to tamoxifen (siControl SF₆₀ = 28nM, CDKlO SMARTpool SF₆₀ = 488nM; a 17 fold reduction in sensitivity) . Further confirmation that CDKlO silencing caused tamoxifen resistance was provided by replicating the dose response effect with an ON-TARGETpIus SMARTpool (Dharmacon) targeting CDKlO, which also significantly decreased tamoxifen sensitivity (data not shown) . The siRNAs that comprise ON-TARGETplus SMARTpools have been chemically modified to minimise off-target effects (Jackson et al, 2006) .

To confirm whether the CDKlO effect was specific to tamoxifen or common to other forms of ERa antagonism, we analysed the effect of CDKlO siRNA on estrogen deprivation, a model for aromatase inhibition (Santen et al, 2005) . CDKlO silencing significantly decreased sensitivity to estrogen deprivation (No E₂; Figure 2D) , indicating that suppression of CDKlO causes resistance to inhibition of ERa signalling generally, rather than tamoxifen specifically. A reduction in CDKlO expression circumvents G₁ cell cycle arrest in cells deprived of estrogen signalling

The anti-proliferative effect of anti-estrogens results in the induction of cell cycle arrest at the Gi checkpoint, as characterised by an increase in cells in the Gi phase of the cycle and a consequential decrease in the S phase proportion (Wilcken et al, 1997) . We explored the possibility that a reduction in CDKlO expression circumvents this tamoxifen-induced cell cycle arrest. siControl transfected cells treated with tamoxifen exhibited the expected increase in Gi and decrease in S phase populations (Figure 3A₇B), whereas CDKlO SMARTpool transfected cells showed reduced tamoxifen induced G_x arrest, with a statistically significantly smaller decrease in the number S phase cells (Figure 3A,B) .

Cyclin Dl expression actively drives transit through the Gi checkpoint. Cyclin Dl binds to CDK4/6, which phosphorylate and inactivate the retinoblastoma protein (Rb) , allowing progression through the restriction point within G₁ (Sherr, 1996; Knudsen and Wang, 1997; Lundberg and Weinberg, 1998; Geng et al, 2001) .

Overexpression of cyclin Dl has also been shown to cause entry into S phase of cells previously arrested at the G₁ checkpoint, reversing the growth inhibitory effects of anti-estrogens (Wilcken et al, 1997) . Both the mRNA and protein levels of cyclin Dl were significantly elevated in CDKlO silenced cells

(Figure 3C,D) and, in addition, the levels of phosphorylated Rb were also significantly elevated (Figure 3E) . These results are consistent with the hypothesis that G₁ cell cycle arrest in tamoxifen treated cells is circumvented by a reduction in CDKlO expression, mediated by an increase in cyclin Dl expression and subsequent phosphorylation of Rb.

Tamoxifen resistance is not caused by ligand independent ERa activation in CDKlO silenced cells Transcription of CCNDl is regulated by the estrogen receptor. ERa stimulation causes cellular proliferation by increasing cyclin Dl expression, driving progression through the G₁ checkpoint and allowing cell cycle progression (Altucci et al, 1996; Doisneau-Sixou et al, 2003) . A possible mechanism of tamoxifen resistance is ligand independent ERa-induced transcription, resulting in continued ERa signalling in the presence of tamoxifen (Shou et al, 2004) . We explored the possibility that the increase in CCNDl observed in CDKlO silenced cells was due to ligand independent ERa activation. Initially we examined the levels of phosphorylated and total ERa in CDKlO silenced cells to determine whether the expression of ERa was increased or whether ERa was hyperactivated. There were no significant changes in either the expression or activation of ERa (Figure 4A) . If ligand independent activation of ERa was occurring, increases in the expression of other ERa regulated genes would also be expected. The progesterone receptor (PR) is a downstream target of ERa signalling (Yu et al, 1981) . We assessed the expression of PR following the silencing of CDKlO and demonstrated that PR levels did not alter significantly (Figure 4B) . We also assessed the expression of TFFl (otherwise known as pS2) , one of the best characterised ERa regulated genes. Transcription of TFFl is tightly controlled by the binding of ERa to an ERE within the TFFl promoter (Jakowlew et al, 1984; Stack et al, 1988) . We found that there was no increase in TFFl expression after CDKlO silencing, instead a reduction in TFFl expression was observed (Figure 4C) . These results suggest that ligand independent activation of ERa signalling is unlikely to be the cause of tamoxifen resistance in CDKlO silenced cells. To further confirm that ERa was not involved in resistance to tamoxifen in CDKlO silenced cells, we examined the effect of CDKlO silencing on sensitivity to ICI 182780, which induces ERa degradation (Jones, 2002) . CDKlO silencing significantly decreased sensitivity to ICI 182780 (SF₆₀ = 4nM, CDKlO SMARTpool SF₆₀ = 32nM; an 8 fold reduction in sensitivity) (Figure 4D) further supporting the hypothesis that CDKlO silencing does not cause tamoxifen resistance by promoting ligand independent ERa signalling. CDKlO suppression activates the p42/p44 MAPK pathway The data described above did not support the hypothesis that the increase in cyclin Dl expression and subsequent tamoxifen resistance in CDKlO silenced cells was caused by ligand independent activation of ERa. The observed reduction in expression of TFFl (Figure 4C) supports the hypothesis that tamoxifen resistance occurs through activation of growth factor signalling pathways that circumvent tumour reliance on ERa signalling. Previous studies have shown that transfection of constitutively active MEKl or c-RAF into MCF7 cells, which hyperactivate p42/p44 MAPK, result in the loss of ERa mediated gene expression, characterised by TFFl suppression and acquisition of anti-estrogen resistance (El-Ashry et al, 1997, Oh et al, 2001) . Additionally, activation of p42/p44 MAPK has been shown to cause an increase in cyclin Dl expression (Lavoie et al, 1996) .

Therefore, we examined whether the p42/p44 MAPK pathway was activated in CDKlO silenced cells. p42/p44 MAPK signalling was assessed by measuring the phosphorylation of p42/p44 MAPK and

MEKl ,2 which are phosphorylated when activated. Phosphorylation of these MAPK components was significantly increased following CDKlO silencing (Figure 5A,B) . To determine whether this activation was the cause of tamoxifen resistance in cells with reduced CDKlO expression, p42/p44 MAPK and MEKl, 2 were silenced by siRNA in conjunction with suppression of CDKlO expression. Sensitivity to tamoxifen in this situation was restored (Figure 5C,D) suggesting that a reduction of CDKlO expression causes tamoxifen resistance by activation of the p42/p44 MAPK signalling pathway, circumventing the reliance upon ER signalling.

CDKlO silencing increases c-RAF expression

To determine at the mechanism by which CDKlO silencing modifies p42/p44 MAPK signalling, the activation of upstream pathway components c-RAF and Ras were examined. Phosphorylation of c-RAF was found to be significantly increased in CDKlO silenced cells, and interestingly, levels of total c-RAF were also increased (Figure 6A) . To determine if the increase in c-RAF protein levels was due to increased transcription of the c-RAF gene, the levels of c-RAF mRNA were assessed using quantitative PCR. c-RAF mRNA levels were significantly increased in CDKlO silenced cells (Figure 6B) . The level of activated Ras was also assessed using a Ras activation assay, but was not found to be significantly increased in CDKlO silenced cells (Figure 6C) . Therefore, activation of the MAPK pathway seemed likely to be mediated by c- RAF levels, caused by the increased transcription of c-RAF. To confirm that overexpression of c-RAF was sufficient to increase the levels of activated c-RAF and to activate downstream components of the MAPK pathway, wild type c-RAF was overexpressed in MCF7 cells. Overexpression of wild type c-RAF did indeed result in increased levels of activated/phosphorylated c-RAF (Figure 6D) and caused increased levels of activated/ phosphorylated p42/p44 MAPK (Figure 6E) suggesting that the increased MAPK activity that characterises tamoxifen resistance in CDKlO silenced cells is driven by elevated c-RAF expression.

Increased c-RAF transcription is mediated by the ETS2 transcription factor

ETS2 is a transcription factor that has previously been identified as a CDKlO interacting protein. ETS2 transactivation is repressed by CDKlO binding (Kasten and Giordano, 2001) . We confirmed that CDKlO binds to ETS2 using co-immunoprecipitation (Co-IP) (Figure 7A) and identified a putative ETS2 binding site in the c-RAF promoter (Figure 7B) . We demonstrated using chromatin IP (ChIP) that ETS2 and CDKlO bind to this site (Figure 7C) . To determine whether the increase in c-RAF transcription observed in CDKlO silenced cells was dependent on ETS2, ETS2 was silenced in combination with CDKlO and the levels of c-RAF transcript were measured using quantitative PCR. The levels of c-RAF transcript were decreased when ETS2 was silenced in combination with CDKlO (Figure 7D) . c-RAF protein levels were also decreased when ETS2 was silenced in combination with CDKlO, as were the levels of phosphorylated p42/p44 MAPK (Figure 7E) . These results suggest that CDKlO silencing removes repression of ETS2 transactivation of the c-RAF promoter causing activation of the MAPK pathway and identifies a novel signalling axis whereby c-RAF expression is modified by CDK10/ETS2 interactions with an ETS2 binding site in the c-RAF promoter.

Low CDKlO expression is associated with clinical resistance to tamoxifen

Our in vitro data suggest that CDKlO expression is a key determinant of sensitivity to tamoxifen. To assess the clinical significance of these findings, CDKlO expression levels were investigated using gene expression microarrays analysed in 87 ERa positive breast tumours from patients treated with adjuvant tamoxifen (Loi et al, 2007) . Low CDKlO expression was associated with a statistically significantly shorter time to distant relapse of disease (p = 0.0205, Figure 8A) , and there was a trend to shorter overall survival (p = 0.0793, Figure 8B) suggesting that reduced CDKlO expression is associated with clinical resistance to tamoxifen in patients consistent with our in vitro studies. Low CDKlO expression was not associated with traditional prognostic factors, including age, size, grade or node positivity (Supplementary Table 2) perhaps suggesting that low CDKlO expression was associated specifically with resistance to tamoxifen rather than a poor prognosis in general .

To confirm these findings, CDKlO expression was measured using qPCR in a second independent set of ERa positive breast tumours from patients treated with adjuvant tamoxifen. Consistent with the data presented above, low CDKlO expression was strongly associated with a statistically significantly shorter time to disease progression (p<0.0001, Figure 8C), and significantly shorter overall survival (p<0.0001, Figure 8D) further suggesting that reduced CDKlO expression is associated with clinical resistance to tamoxifen in patients .

Methylation within the promoter of genes is a common mechanism of transcriptional repression. The methylation status of the CpG island in the CDKlO promoter was assessed in genomic DNA extracted from biopsies taken from the second set of breast tumours, using methylation specific PCR (MSP) . 7/38 (18%) of cases had methylation of the CDKlO promoter and CDKlO promoter methylation was strongly associated with low CDKlO expression (Figure 8E) . Methylation of the CDKlO promoter was also associated with a statistically significantly shorter time to disease progression (p<0.0001) and significantly shorter overall survival (p<0.0001) . This suggests a mechanism by which tumours develop low CDKlO expression levels resulting in tamoxifen resistance.

Discussion

Resistance to endocrine therapies is one of the major limiting factors in the successful treatment of breast cancer, and strategies to enhance their utility would be of major clinical benefit. Intracellular signalling pathways and their effects on the response to endocrine therapies have been the subject of considerable study for some time. While much of this work has been highly informative, it has relied on the study of proteins whose function is consistent with the existing understanding of intracellular signalling. As a complement to these candidate based approaches we identified novel determinants of response to an endocrine therapy using an unbiased approach and screened loss-of-function in 779 kinases and kinase-related proteins. Our intention was to identify strong determinants of tamoxifen sensitivity, with a view to establishing the clinical significance of our in vitro screen. The identification of CDKlO fulfilled these criteria as this protein had not previously been implicated in resistance to endocrine agents. Furthermore, delineation of the mechanism by which CDKlO is likely to determine tamoxifen sensitivity, via ETS2 transactivation of the c-RAF gene, suggests a novel role for this protein. Lastly, the correlations between CDKlO expression, methylation and clinical outcome suggest that such unbiased approaches can identify novel determinants that have clear clinical significance. Integrating complementary functional genomic and clinical approaches

One approach to the identification of determinants of endocrine therapy has been to use gene expression profiling, both on an individual gene and a genome-wide basis (Jansen et al, 2005) . This approach, especially when carried out in a high-throughput fashion and with large cohorts, can be powerful in demonstrating correlations between transcript levels and clinical outcome. However, the interpretation of such analysis is limited by the inability to distinguish expression changes that are causative to the clinical phenotype from those that are merely consequential . The combination of functional analysis such as RNAi screening followed by gene expression analysis in clinical samples provides a powerful unbiased approach to the identification of the key genetic causes of phenotypes such as drug resistance.

Novel CDKlO function in the regulation of c-RAF

Our study has delineated a mechanism by which CDKlO can modulate sensitivity to tamoxifen and other ERa antagonists . By integrating our data with previously published work, we suggest a model by which CDKlO modulates intracellular signalling and determines the response to tamoxifen and other anti-estrogens . We propose that CDKlO normally binds and represses the ETS2 transcription factor, in agreement with previously published work (Kasten et al, 2001) . We identified a novel ETS2 binding site in the c-RAF promoter and using ChIP demonstrated that both CDKlO and ETS2 bind to this site. In the absence of CDKlO activity, c- RAF transcription is significantly upregulated due to relief of ETS2 repression. This increase in c-RAF expression leads to activation of downstream components of the MAPK pathway, including MEKl ,2 and p42/p44 MAPK, which increase the expression of cyclin Dl (Lavoie et al, 1996) , resulting in tamoxifen resistance by circumventing the reliance upon estrogen signalling (Wilcken et al, 1997) .

Clinical significance of CDKlO

Having identified CDKlO as a modifier of tamoxifen sensitivity, we examined CDKlO expression in tamoxifen treated breast tumours to validate its clinical significance. Patients with low CDKlO expressing tumours were resistant to tamoxifen, consistent with the functional effect of silencing CDKlO expression identified in our RNAi screen. Furthermore, we identified methylation of the CDKlO promoter in a significant proportion of tumours (18%) suggesting a mechanism for suppression of CDKlO expression. The significant association of clinical outcome with methylation of the CDKlO promoter provides further evidence that suppression of CDKlO is a key driver of resistance to tamoxifen. Given that the mechanism of tamoxifen resistance identified in tumours with low CDKlO suggests loss of reliance on estrogen signalling, low CDKlO expression may also be associated with clinical resistance to other endocrine therapies such as aromatase inhibitors . Our study identifies a subgroup of low CDKlO expressing, ERa positive breast cancer patients that respond poorly to endocrine therapies and these patients may benefit from alternative therapeutic approaches including the use of signal transduction inhibitors (Dancey et al, 2003) .

It is notable that previous studies have demonstrated an association between both p42/p44 MAPK activation and increased cyclin Dl expression and tamoxifen resistance (Gee et al, 2001; Kenny et al, 1999) . Our data identifies a possible mechanism explaining these associations. In addition, elevated ETS2 expression and phosphorylation have been associated with reduced disease free survival in tamoxifen treated patients (Myers et al, 2005,- Svensson et al, 2005) supporting the hypothesis that increased ETS2 activity can cause resistance to tamoxifen.

Summary

This study reports the first functional RNAi screen to systematically identify the causes of tamoxifen resistance. We identify a novel modifier of tamoxifen sensitivity, CDKlO, and establish its mechanism of action, the regulation of the p42/p44 MAPK pathway. Importantly, we demonstrate the clinical significance of these findings. References

The documents disclosed herein are all expressly incorporated by reference in their entirety.

Altucci et al., (1996) 17beta-Estradiol induces cyclin Dl gene transcription, p36Dl-p34cdk4 complex activation and plO5Rb phosphorylation during mitogenic stimulation of G(I) -arrested human breast cancer cells. Oncogene. 12, 2315-2324.

Anderson et al . , (2007) Predictors of response to aromatase inhibitors. J Steroid Biochem MoI Biol. May 24.

Bardou et al., (2003) Progesterone receptor status significantly improves outcome prediction over estrogen receptor status alone for adjuvant endocrine therapy in two large breast cancer databases. J Clin Oncol. 21, 1973-1979.

Baum et al . , The ATAC (Arimidex, Tamoxifen Alone or in Combination) Trialists ' Group. (2003) Anastrozole alone or in combination with tamoxifen versus tamoxifen alone for adjuvant treatment of postmenopausal women with early-stage breast cancer: results of the ATAC (Arimidex, Tamoxifen Alone or in Combination) trial efficacy and safety update analyses. Cancer. 98, 1802-1810.

Campbell et al . , (2001) Phosphatidylinositol 3-kinase/AKT- mediated activation of estrogen receptor alpha: a new model for anti-estrogen resistance. J Biol Chem. 216, 9817-9824.

Cariou et al., (2000) Down-regulation of p21WAFl/CIPl or p27Kipl abrogates antiestrogen-mediated cell cycle arrest in human breast cancer cells. Proc Natl Acad Sci U S A. 97, 9042-9046.

Crighton et al., (2006) Dram, a p53-induced modulator of autophagy, is critical for apoptosis. Cell. 126, 121-134.

Dancey & Sausville, (2003) Issues and progress with protein kinase inhibitors for cancer treatment. Nature Reviews Drug Discovery. 2, 296 - 313. Doisneau-Sixou et al., (2003) Estrogen and antiestrogen regulation of cell cycle progression in breast cancer cells. Endocr Relat Cancer. 10, 179-186.

Early Breast Cancer Trialists ' Collaborative Group. (1998) Tamoxifen for early breast cancer: an overview of the randomised trials. Lancet. 351, 1451-1467.

Echeverri et al . , (2006) Minimizing the risk of reporting false positives in large-scale RNAi screens. Nat Methods. 3, 777-9.

El-Ashry et al . , (1997) Constitutive Raf-1 kinase activity in breast cancer cells induces both estrogen-independent growth and apoptosis. Oncogene. 15, 423-435.

Gee et al., (2001) Phosphorylation of ERKl/2 mitogen-activated protein kinase is associated with poor response to anti-hormonal therapy and decreased patient survival in clinical breast cancer. Int J Cancer. 95, 247-254.

Geng et al., (2001) Deletion of the p27Kipl gene restores normal development in cyclin Dl-deficient mice. Proc Natl Acad Sci U S A. 98, 194-199.

Guadagno & Ferrell, (1998) Requirement for MAPK activation for normal mitotic progression in Xenopus egg extracts. Science. 282, 1312-1315.

Gutierrez et al . , (2005) Molecular changes in tamoxifen-resistant breast cancer: relationship between estrogen receptor, HER-2, and p38 mitogen-activated protein kinase. J Clin Oncol. 23, 2469-76.

Hammaker et al., (2004) Regulation of c-Jun N-terminal kinase by MEKK-2 and mitogen-activated protein kinase kinase kinases in rheumatoid arthritis. J Immunol. 172, 1612-1618. Hofmann et al . , (2003) HIPK2 regulates transforming growth factor-beta-induced c-Jun NH (2) -terminal kinase activation and apoptosis in human hepatoma cells. Cancer Res. 63, 8271-8277.

Holm et al., (2006) Association between Pakl expression and subcellular localization and tamoxifen resistance in breast cancer patients. J Natl Cancer Inst. 98, 671-680.

Hu et al., (1996) Human HPKl, a novel human hematopoietic progenitor kinase that activates the JNK/SAPK kinase cascade. Genes Dev. 10, 2251-2264.

Iorns et al . , (2007) Utilizing RNA interference to enhance cancer drug discovery. Nat Rev Drug Discov. 6, 556-568.

Jackson et al . , (2003) Expression profiling reveals off-target gene regulation by RNAi. Nature Biotechnology. 21, 635-637.

Jackson et al . , (2006) Position-specific chemical modification of siRNAs reduces "off-target" transcript silencing. RNA. 12, 1197- 1205.

Jakowlew et al . , (1984) Sequence of the pS2 mRNA induced by estrogen in the human breast cancer cell line MCF-7. Nucleic Acids Res. 12, 2861-2878.

Jansen et al., (2005) Molecular classification of tamoxifen- resistant breast carcinomas by gene expression profiling. J Clin Oncol. 23, 732-740.

Jones, (2002) A new estrogen receptor antagonist--an overview of available data. Breast Cancer Res Treat. 75, S19-21.

Jordan, (1995) Third annual William L. McGuire Memorial Lecture. "Studies on the estrogen receptor in breast cancer"--20 years as a target for the treatment and prevention of cancer. Breast Cancer Res Treat. 36, 267-285. Juncker-Jensen et al . , (2006) Insulin-like growth, factor binding protein 2 is a marker for antiestrogen resistant human breast cancer cell lines but is not a major growth regulator. Growth Horm IGF Res. 16, 224-39.

Kasten & Giordano, (2001) CdklO, a Cdc2-related kinase, associates with the Ets2 transcription factor and modulates its transactivation activity. Oncogene. 20, 1832-1838.

Kenny et al., (1999) Overexpression of cyclin Dl messenger RNA predicts for poor prognosis in estrogen receptor-positive breast cancer. Clin Cancer Res. 5, 2069-2076. Kim et al., (2007) Tamoxifen-induced activation of p2lWafl/Cipl gene transcription is mediated by Early Growth Response-1 protein. Cellular Signalling. 19, 1290-1300.

Knudsen & Wang, (1997) Dual mechanisms for the inhibition of E2F binding to RB by cyclin-dependent kinase-mediated RB phosphorylation. MoI Cell Biol. 17, 5771-5783.

Lavoie et al., (1996) Cyclin Dl expression is regulated positively by the p42/p44MAPK and negatively by the p38/HOGMAPK pathway. J Biol Chem. 271, 20608-16.

Lee et al., (2006) JNK pathway mediates apoptotic cell death induced by tumor suppressor LKBl in Drosophila. Cell Death Differ. 13, 1110-1122.

Loi et al . , (2007) Definition of clinically distinct molecular subtypes in estrogen receptor-positive breast carcinomas through genomic grade. J Clin Oncol. 25, 1239-46.

Lundberg & Weinberg, (1998) Functional inactivation of the retinoblastoma protein requires sequential modification by at least two distinct cyclin-cdk complexes, MoI Cell Biol. 18, 753- 761. MacKeigan et al . , (2005) Sensitized RNAi screen of human kinases and phosphatases identifies new regulators of apoptosis and chemoresistance. Nat Cell Biol. 7, 591-600.

Mandlekar et al., (2000) Activation of caspase-3 and c-Jun NH2- terminal kinase-1 signaling pathways in tamoxifen-induced apoptosis of human breast cancer cells. Cancer Res. 60, 5995- 6000.

Mangelsdorf et al . , (1995) The nuclear receptor superfamily: the second decade. Cell. 83, 835-839.

Meister et al., (2004) Human Argonaute2 mediates RWA cleavage targeted by miRNAs and siRNAs. MoI. Cell. 15, 185-197.

Meister & Tuschl, (2004) Mechanisms of gene silencing by double- stranded RNA. Nature 431, 343-349.

Musgrove & Sutherland, (1994) Cell cycle control by steroid hormones. Semin Cancer Biol. 5, 381-389.

Myers et al., (2005) Associations and interactions between Ets-1 and Ets-2 and coregulatory proteins, SRC-I, AIBl, and NCoR in breast cancer. Clin Cancer Res. 11, 2111-2122.

Oh et al . , (2001) Hyperactivation of MAPK induces loss of ERalpha expression in breast cancer cells. MoI Endocrinol. 15, 1344-1359.

Perez-Tenorio et al . , (2002) Activation of AKT/PKB in breast cancer predicts a worse outcome among endocrine treated patients . Br J Cancer. 86, 540-545.

Rayala (2006) P21-activated kinase 1 regulation of estrogen receptor-alpha activation involves serine 305 activation linked with serine 118 phosphorylation. Cancer Res. 66, 1694-1701. Ring & Dowsett, (2004) Mechanisms of tamoxifen resistance. Endocr Relat Cancer. 11, 643-658.

Sherr, (1996) Cancer cell cycles. Science. 274, 1672-1677.

Shou et al . , (2004) Mechanisms of tamoxifen resistance: increased estrogen receptor-HER2/neu cross-talk in ER/HER2-positive breast cancer. J Natl Cancer Inst. 96, 926-935.

Simpson et al . , (2005) Columnar cell lesions of the breast: the missing link in breast cancer progression? A morphological and molecular analysis. Am J Surg Pathol. 29, 734-46.

Smith et al . , (2007) Epigenetic inactivation implies independent tumour suppressor functions for IGFBP-rPl and the related IGFBPLl in breast cancer. Clin. Cancer Res. 15, 4061-4068.

Stack et al . , (1988) Structure and function of the pS2 gene and estrogen receptor in human breast cancer cells. Cancer Treat Res. 40, 185-206.

Stoica et al . , (2000) Regulation of estrogen receptor-alpha gene expression by epidermal growth factor. J Endocrinol. 165, 371- 378.

Stoica et al . , (2000) Role of insulin-like growth factor-I in regulating estrogen receptor-alpha gene expression. J Cell Biochem. 76, 605-614.

Svensson et al . , ERK phosphorylation is linked to VEGFR2 expression and Ets-2 phosphorylation in breast cancer and is associated with tamoxifen treatment resistance and small tumours with good prognosis. Oncogene. 24, 4370-4379.

Swanton et al., (2007) Regulators of mitotic arrest and ceramide metabolism are determinants of sensitivity to paclitaxel and other chemotherapeutic drugs. Cancer Cell. 11, 498-512. Tang et al., (1996) Involvement of heregulin-beta2 in the acquisition of the hormone-independent phenotype of breast cancer cells. Cancer Res. 56, 3350-3358.

Tournier et al . , (1997) Mitogen-activated protein kinase kinase 7 is an activator of the c-Jun NH2-terminal kinase. Proc Natl Acad Sci U S A. 94, 7337-7342.

Whitehurst et al., (2007) Synthetic lethal screen identification of chemosensitizer loci in cancer cells. Nature. 446, 815-819.

Wilcken et al . , (1997) Different points of action of retinoids and anti-estrogens in Gl phase identified in synchronized T-47D breast cancer cells. Int J Cancer. 70, 291-296.

Wilcken et al . , (1997) Inducible overexpression of cyclin Dl in breast cancer cells reverses the growth-inhibitory effects of antiestrogens . Clin Cancer Res. 3, 849-854.

Wingender et al., (1996) TRANSFAC: a database on transcription factors and their DNA binding sites. Nucleic Acids Res. 24, 238- 241.

Yasuda et al . , (1999) The JIP group of mitogen-activated protein kinase scaffold proteins. MoI Cell Biol. 19, 7245-7254.

Yu et al . , (1981) Effects of 17 beta-estradiol on progesterone receptors and the uptake of thymidine in human breast cancer cell line CAMA-I. Cancer Res. 41, 5004-5009.

Claims

Claims :

1. A method for determining whether an individual having estrogen receptor α (ERa) positive breast cancer is likely to be resistant to endocrine therapy, the method comprising: determining a cyclin dependent kinase 10 (CDKlO) status in a sample obtained from the individual; and using the CDKlO status to determine whether the individual is likely to be resistant to endocrine therapy.

2. The method of claim 1, wherein determining the CDKlO status comprises determining the methylation of the promoter of the CDKlO gene and correlating methylation of the CDKlO promoter with resistance to endocrine therapy.

3. The method of claim 1 or claim 2, wherein determining CDKlO status comprises determining the expression of the CDKlO gene and correlating reduced expression of the CDKlO gene with resistance to endocrine therapy .

4. The method of any one of claims 1 to 3 , wherein determining CDKlO status comprises determining the expression of the CDKlO protein and correlating reduced expression of the CDKlO protein with resistance to endocrine therapy.

5. The method of any one of the preceding claims, wherein the method further comprises the step of determining whether the breast cancer is ERa positive breast cancer.

6. The method of any one of the preceding claims, wherein CDKlO status is determined with reference to one or more controls.

7. The method of any one of the preceding claims, wherein the method is used for determining clinical treatment given to the individual .

8. The method of claim 7, wherein determining the clinical treatment comprises selecting a type of chemotherapy or a chemotherapy regimen for administration to the individual.

9. The method of claim 7 or claim 8, wherein when a test on a sample from an individual shows an absence of CDKlO promoter methylation or normal CDKlO gene or protein expression, the individual is treated using endocrine therapy.

10. The method of claim 9, wherein endocrine therapy is treatment with tamoxifen, faslodex and/or aromatase inhibitors.

11. The method of claim 7 or claim 8, wherein when a test on a sample from an individual shows CDKlO promoter methylation or a reduced level of CDKlO gene or protein expression, the individual is treated using signal transduction inhibitors .

12. The method of any one of the preceding claims , wherein the methylation of the CDKlO promoter is determined using methylation specific PCR, bisulphite sequencing, hybridisation to a bead array or by isolating and sequencing methylated DNA.

13. The method of claim 12, wherein the determination of methylation of the CDKlO promoter is performed on genomic nucleic acid extracted from a sample of cells obtained from the breast cancer or from a sample of cancer cells circulating in blood.

14. The method of any one of the preceding claims , wherein determining the expression of the CDKlO gene comprises extracting RNA from a sample of breast cancer cells and measuring expression by real time PCR or by using a probe capable of hybridising to CDKlO RNA.

15. The method of claim 14, wherein the probe is immobilised in a microarray.

16. The method of any one of the preceding claims, wherein^' determining CDKlO protein expression comprise detecting expression in a tumour sample using immunohistochemistry.

17. The method of any one of the preceding claims, wherein the determining CDKlO protein expression comprises measuring CDKlO protein levels in a cell lysate by ELISA or western blotting.

18. The method of any one of the preceding claims, wherein determining CDKlO protein expression comprises using a binding agent capable of specifically binding to the CDK protein, or a fragment thereof.

19. The method of any one of the preceding claims, wherein the method comprises the initial step of obtaining a sample from said individual .

20. The method of any one of the preceding claims, wherein the sample is a tumour sample, a blood sample, a tissue sample or a cell sample.

21. Use of a cyclin dependent kinase 10 (CDKlO) status for determining whether an individual having estrogen receptor α

(ERa) positive breast cancer is likely to be resistant to endocrine therapy.

22. A kit for determining whether an individual having estrogen receptor α (ERa) positive breast cancer is likely to develop resistance to endocrine therapy according to the method of any one of claims 1 to 20.

23. The kit of claim 22, wherein the kit comprises reagents necessary for carrying out the determination of CDKlO status of a sample and instructions for carrying out the test and interpreting the results .

24. The kit of claim 22 or claim 23, wherein the kit comprises one or more of the following reagents:

(a) an antibody capable of recognising CDKlO polypeptides or fragments thereof; and/or (b) primers based on the nucleic acid, sequence of the CDKlO for detecting the presence of CDKlO mKNA; and/or

(c) a probe based on the nucleic acid sequence of the CDKlO gene for detecting CDKlO gene expression,- and/or (d) reagents for determining the methylation status of the promoter of the CDKlO gene.

25. Use of a demethylating agent for the preparation of a medicament for the treatment a patient having breast cancer, wherein the breast cancer is estrogen receptor α (ERa) positive and characterised by a positive CDKlO status.

26. A demethylating agent for use in a method of treating a patient having breast cancer, wherein the breast cancer is estrogen receptor oc (ERa) positive and characterised by a positive CDKlO status.