WO2010151133A1 - Methods for establishing and predicting resistance to endocrine therapy using an mirna profile - Google Patents

Abstract

The present invention relates to a method of predicting or establishing the resistance of a breast cancer tumor to endocrine therapy, comprising the steps of determining the expression profile essentially consisting of mi RNAs hsa-mi R-7, hsa-let-7g, hsa-mi R-17-5p, hsa-mi R-126, hsa-mi R-126#, hsa-mi R- 141, hsa-mi R-145, hsa-mi R-335, hsa-mi R-365, hsa-mi R-489, hsa-mi R-22, hsa- mi R-30a-3p, hsa-mi R-30c, and hsa-mi R-182, preferably hsa-mi R-30a-3p, hsa- mi R-30c, and hsa-mi R-182 in a sample from a breast cancer tumor; comparing said profile with a reference profile, and predicting or establishing the resistance of said breast cancer to said endocrine therapy based on said comparison.

Description

Title: Methods for establishing and predicting resistance to endocrine therapy using an miRNA profile

FIELD OF THE INVENTION

The invention relates to the field of medical diagnostics, more specifically to the field of cancer diagnostics, especially breast cancer.

BACKGROUND OF THE INVENTION

Resistance to anti-estrogens is one of the major challenges in the treatment of breast cancer. For more than 25 years, the golden standard for the endocrine treatment of all stages of estrogen receptor-positive (ER+) breast cancer has been tamoxifen or more recently an aromatase inhibitor. Approximately 30-50% of ER+ breast tumors do not respond to tamoxifen treatment (de novo resistance), and those that do respond often eventually progress to a state in which tumor cell proliferation is no longer inhibited, and may even be stimulated, by tamoxifen treatment (acquired resistance). Also, response rates in patients with ER-α negative primary tumors are very low. Comparable but slightly superior figures are seen for aromatase inhibitors.

The ability to identify tumors that are unlikely to respond to treatment with tamoxifen or other estrogen antagonists is therefore critically needed, as is the development of alternative therapies to treat resistant tumors. Aromatase inhibitors may be more effective than tamoxifen at treating primary breast cancer, and offer a very promising alternative. In addition, many Tam-resistant tumors retain sensitivity to steroidal anti- estrogens such as ICI 182,780 (ICI) (Fulvestrant® or Faslodex®), and this compound is approved as a second line therapy for patients who relapse while undergoing tamoxifen treatment. However, a significant percentage of patients with advanced breast cancer will likely develop resistance to all endocrine therapies, and additional approaches to treat these patients are needed. One promising development is personalized medicine or individualized treatment wherein the right drug is matched to the right patient and in particular to his/her genotype. Therefore biomarkers are needed to identify patients who will not respond to tamoxifen or an aromatase inhibitor and to select patients who are likely to benefit therefrom.

SUMMARY OF THE INVENTION

The present invention provides in a first aspect a method of predicting or establishing the resistance of a breast cancer tumor to endocrine therapy, comprising the steps of: a. determining the expression profile essentially consisting of miRNAs hsa-miR-7, hsa-let-7g, hsa-miR-17-5p, hsa-miR-126, hsa-miR-126#, hsa-miR-141, hsa-miR-145, hsa-miR-335, hsa-miR-365, hsa-miR-489, hsa- miR-22, hsa-miR-30a-3p, hsa-miR-30c, and hsa-miR-182, preferably the expression profile essentially consisting of miRNAs hsa-miR-7, hsa-let-7g, hsa-miR-30a-3p, hsa-miR-30c, and hsa-miR-182, and more preferable the expression profile essentially consisting of miRNAs hsa-miR-30a-3p, hsa- miR-30c, and hsa-miR-182 in a sample from a breast cancer tumor and; b. comparing said profile with a reference profile or a profile of a control sample; c. predicting or establishing the resistance of said breast cancer to said endocrine therapy based on said comparison.

Step a) suitable comprises the step of providing a breast cancer tumor sample. This may be in the form of a tumor biopsie, wherein a small part of the tumor is sampled and taken as a representative sample for the tumor itself. Step b), determining the expression profile, in all aspects of the invention relates to the bonding of specific binding partners to the nucleic acid molecules of the miRNAs to be detected, for instance in the form of sequence-specific hybridization probes or sequence specific amplification primers. Such bonding may involve the liberation and isolation of nucleic acid molecules from the tumor tissue sample and hence disruption of tumor tissue in order to assess the presence and expression level therein of the of the miRNAs. Bonding of the specific binding partners to the miRNAs results in a chemical change by which the miRNAs can be detected. The invention further provides in another aspect a method of predicting or establishing the resistance of a breast cancer tumor to endocrine therapy, comprising the steps of: a. determining the expression level of at least one miRNA, preferably at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, or 15 miRNAs selected from the group consisting of: miRNAs hsa-miR-7, hsa-let-7g, hsa- miR-17-5p, hsa-miR-126, hsa-miR-126#, hsa-miR-141, hsa-miR-145, hsa-miR- 335, hsa-miR-365, hsa-miR-489, hsa-miR-22, hsa-miR-30a-3p, hsa-miR-30c, and hsa-miR-182, preferably from the group consisting of: miRNAs hsa-miR- 7, hsa-let-7g, hsa-miR-30a-3p, hsa-miR-30c, and hsa-miR-182, more preferably from the group consisting of: miRNAs hsa-miR-30a-3p, hsa-miR- 30c, and hsa-miR-182, in a sample from a breast cancer tumor and; b. comparing said expression level with a reference expression level or a expression level of a control sample; c. predicting or establishing the resistance of said breast cancer to said endocrine therapy based on said comparison.

Methods of predicting or establishing the resistance of a breast cancer tumor to endocrine therapy as indicated herein are diagnostic methods. However, emphasis is placed on the analytical part of the diagnosis involving the actual biochemical test. Hence, the present invention in diagnostic aspects refers to methods of analysis of biological samples for the presence of biomarkers that have a predictive, diagnostic and/or prognostic value, wherein said biomarker may be a miRNA profile or a single miRNA expression level.

The preferred miRNAs have high statistical correlation with progression free survival and/ or clinical benefit from tamoxifen therapy

(Tarn. Clinical Benefits) as indicated in Table 4. In a preferred embodiment of methods of the invention, said at least one miRNA is selected from hsa-miR- 30a-3p, hsa-miR-30c and hsa-miR-182.

Further preferred in methods of the invention are embodiments wherein step a) is performed on a sample which is to be exposed to said estrogen antagonist (anti-estrogen) or an aromatase inhibitor prior to determining said level. Alternatively, the existence of resistance during ongoing therapy may be tested.

In a preferred embodiment the difference between the level of said at least one miRNA and said reference value or a said level of a control sample is greater than one-fold. In a highly preferred embodiment said endocrine therapy comprises the administration of an anti-estrogen (most preferably tamoxifen) or an aromatase inhibitor.

In a preferred embodiment of said method according to the invention, said breast cancer tumor is a recurring tumor, preferably an ER+ tumor.

In another aspect, the invention further provides the use of a marker selected from miRNAs miRNAs hsa-miR-7, hsa-let-7g, hsa-miR-17-5p, hsa- miR-126, hsa-miR-126#, hsa-miR-141, hsa-miR-145, hsa-miR-335, hsa-miR- 365, hsa-miR-489, hsa-miR-22, hsa-miR-30a-3p, hsa-miR-30c, and hsa-miR- 182, preferably hsa-miR-30a-3p, hsa-miR-30c, and hsa-miR-182 for diagnostic use in a method for predicting or establishing the resistance of a breast cancer tumor.

In another aspect, the invention further provides a kit of parts for performing a method of predicting or establishing the resistance of a breast cancer tumor to endocrine therapy comprising RT primers and/or probes specific for the detection of at least one miRNA selected from the group consisting of hsa-miR-7, hsa-let-7g, hsa-miR-17-5p, hsa-miR-30c, hsa-miR- 30a-3p, hsa-miR-126, hsa-miR-126#, hsa-miR-141, hsa-miR-145, hsa-miR- 182, hsa-miR-335, hsa-miR-365, hsa-miR-489, hsa-miR-22, hsa-miR-30a-3p, hsa-miR-30c, and hsa-miR-182, preferably hsa-miR-30a-3p, hsa-miR-30c, and hsa-miR-182 and an instruction for performing the method according to the invention. Preferably, said kit comprises not more than 60, 50, 40, 30, 25, 20, 15, or 10 primers and/or probes.

BRIEF DESCRIPTION OF THE FIGURES

Figure 1: Kaplan-Meier PFS curves of 246 primary breast cancer patients as a function of the levels of hsa-miR-30a-3p (a), hsa-miR-30c (b), and hsa-miR-182 (c). In patients with ER tumors and tamoxifen treated for recurrent breast cancer. Samples are divided in quartiles based on the levels of the respective hsa-miRNAs. Patients at risk at 12-month time intervals are indicated at the bottom of each graph. HR, hazard ratio.

Figure 2: Pathways associated with tamoxifen response and most relevant hsa-miRNAs. Pathway analysis plots of Biocarta pathways most strongly associated with hsa-miR-30c (a,b) and hsa-miR-30a-3p (c,d). A pathway plot [generated by the Global Test program ] shows the genes annotated to the indicated pathway in Biocarta with their association with tumors showing a low versus high expression of the indicated miRNA. Each bar represents a gene in the pathway. The height of the bar indicates the contribution (influence) of each individual gene to the significance of the pathway. Horizontal markers in a bar indicate 1 SD away from the reference point, and two or more horizontal lines in a bar indicate a statistically significant association of the corresponding gene with the subgroups identified by a specific hsa- miRNA. Dark bars indicate positive and light bars indicate negative associations with the tumor subgroup having high versus low expression of a particular hsa-miRNA. DETAILED DESCRIPTION OF THE INVENTION

Definitions

The term "breast cancer" as used herein refers to refers to the erratic growth and proliferation of cells that originate in the breast tissue. A group of rapidly dividing cells may form a lump or mass of extra tissue. These masses are called tumors. Tumors can either be cancerous (malignant) or non-cancerous (benign). Malignant tumors penetrate and destroy healthy body tissues. A group of cells within a tumor may also break away and spread to other parts of the body. Cells that spread from one region of the body into another are called metastases. In preferred embodiments of aspects of the invention the term breast cancer refers to a malignant tumor that has developed from cells in the breast. The teachings of the present invention could be extended to anti-estrogen and aromatase inhibitor-resistant of other tumor types.

The term "predict" as used herein refers to identifying or forecasting whether a breast cancer tumor will exhibit resistance or develop resistance against endocrine therapy.

The term "resistance" as used herein with reference to resistance to treatment or therapy refers the fact that the normal therapeutic efficacy of the drug is not attained and that, for instance, a tumor continues to grow. Preferably, a tumor is classified as being resistant if the patient from whom said tumor was derived is classified as being a non-responder according to, for instance, the criteria used or set forth by the European Organisation for Research and Treatment of Cancer (EORTC) or the Union International Contre Ie Cancer (UICC).

The term "control sample" as used herein refers to a reference sample that is a sample from a tumor which is resistant or that is a sample from a tumor that is not resistant to endocrine therapy or that is a sample from a normal breast tissue, and which is used for comparison purpose with a test sample, that is, in order to classify the test sample — or more specifically the levels of certain miRNAs therein, as being indicative of a resistant of non- resistant tumor. In a preferred embodiment of a method according to the invention, said control sample is from a resistant tumor. The level of a miRNA of a control sample as described herein or an expression profile of miRNAs as described herein is compared to the level or profile in a suspected tumor sample. If there is no significant difference between said levels or profiles, then the suspected tumor sample is determined as having the same indication as that control sample. A control sample may be from a single individual or from multiple individuals.

The term "reference profile" as used herein refers to a collection of expression levels for a multitude of given miRNAs, which expression levels in combination and/or relative to each other provide specific information which can be used for the purpose of comparison with other profiles. The collection in aspects of this invention is indicative for a resistant or non-resistant phenotype of the tumor and can be used as a reference for comparison with the test profile. Preferably, such reference profile is based on samples of one or more non-resistant breast tumors. Alternatively, the reference profile may be based on samples of one or more resistant breast tumors. The skilled person is well aware of methods for comparing data collections such as test and reference profiles of gene expression data, and such methods can suitably be used in aspects of the present invention.

The term "anti-estrogen or an aromatase inhibitor-resistant tumor" as used herein refers to a tumor, including the individual cells therein, that is or becomes refractory to treatment by an anti-estrogen or an aromatase inhibitor. In specific embodiments, the anti-estrogen or aromatase inhibitor- resistant tumor becomes resistant to anti-estrogen or aromatase inhibitor treatment after initiation of the treatment and may occur during the treatment. In further specific embodiments, the resistance to anti-estrogen or aromatase inhibitor manifests at about 2-24 months while the patient is receiving hormonal therapy. In de novo resistance, the patient does not respond to initial therapy. Acquired resistance is where the patient develops metastatic disease during therapy.

Acquired resistance to hormone therapy such as anti-estrogen or aromatase inhibitor is well-known in the art. In particular, breast cancer patients while undergoing treatment with anti-estrogen or aromatase inhibitor have recurrence of the disease. In specific embodiments, the disease metastasizes during therapy with anti-estrogen or aromatase inhibitor, which results in resistant metastases. The present invention predicts the occurrence of potential metastases, provides information concerning present metastases and/or prevents additional mestastases by identifying tumors susceptible to becoming resistant or being resistant.

The term "anti-estrogen or an aromatase inhibitor-sensitive tumor" as used herein refers to a tumor, including the individual cells therein, that is treatable with anti-estrogen or an aromatase inhibitor. In specific embodiments, the anti-estrogen or an aromatase inhibitor-sensitive tumor remains sensitive during the treatment. In further specific embodiments, the anti-estrogen or an aromatase inhibitor- sensitive tumor is still sensitive up to at least about seven to ten years. Preferably, a tumor is classified as being sensitive if the patient from whom said tumor was derived is classified as being a responder as can be classified using EORTC or UICC criteria. Preferably, said responder is classified using EORTC or UICC criteria.

The term "different" as used herein with reference to the comparison of expression profiles or expression levels refers to a degree of difference which is statistically significantly increased or decreased compared to a reference or a control. The term "significantly" or "statistically significant" refers to statistical significance and generally means that values differ two standard deviations (SD). In preferred embodiments, said difference is classified as statistically significant if the expression level is at least a 20 percent increased or decreased compared to expression level of the same expression product in control individuals. Preferably, the increase or decrease is at least 20, 25, 30, 35, 40, 45, 50, 75, 100, 150, 200 or 250 percent. Most preferably, said increase or decrease is at least 100 percent (herein also referred to as "one fold"). The term "level" as used herein refers to the measurable absolute level or a measurable relative level compared to the level of another miRNA. If the expression level is determined of a miRNA from more than one individual (as a control), usually the median or mean expression level of these individuals is used for comparison. The term "miRNA" as used herein refers to MicroRNAs (miRs) which are small RNA molecules encoded in the genomes of plants and animals. They are present within introns of protein-coding genes, in polycistronic transcripts encoding multiple miRNAs and in individual miRNA genes. These highly conserved RNAs with a length of approximately 21-23 nucleotides regulate the expression of genes by binding to the 3'-untranslated regions (3'- UTRs) of specific mRNAs. Each miRNA is thought to regulate multiple genes, and since hundreds of miRNA genes are predicted to be present in higher eukaryotes the potential regulatory circuitry afforded by miRNA is enormous. MiRNAs are believed to act as key regulators of processes as diverse as early development, cell proliferation and cell death, apoptosis and fat metabolism, and cell differentiation. There is speculation that in higher eukaryotes, the role of miRNAs in regulating gene expression could be as important as that of transcription factors. Over 540 human miRNAs have been validated to date; however, computer models suggest there may be thousands more. It is believed that up to 30% of human genes are regulated by miRNAs.

The genes that encode miRNAs are transcribed from DNA by RNA polymerase II but not translated into protein. The primary transcripts, which generally have a length of several kilobases, are called pri-miRNAs. Pri- miRNAs are processed in the cell nucleus to shorter, 70-100 nucleotide stem- loop structures known as pre-miRNAs. This processing is performed in animals by the RNase III endonuclease Drosha. Pre-miRNAs are subsequently transported into the cytoplasm, where they are processed to double- stranded miRNAs with a length of 21-23 nucleotides by a second RNase III endonuclease, DICER. One strand of the miRNA duplex is subsequently incorporated into the RNA-induced silencing complex (RISC). As part of the RISC, gene expression of a target gene is counteracted by inhibiting translation and/or by cleaving mRNA. The mature miRNAs are partially complementary to one or more mRNA molecules. If miRNA and mRNA are expressed in the same cell, they can hybridize in a sequence- specific manner, thereby preventing translation of the mRNA into protein, thus importantly regulating specific protein levels. In particular, the term miRNA refers to the single- stranded RNA molecule as listed in Table 5. The term miRNA includes reference to the pri-miRNA, the pre-miRNA, as well as the mature miRNA. The term "aromatase inhibitor", as used herein, relates to a compound which inhibits the estrogen production, i.e., the conversion of the substrates androstenedione and testosterone to estrone and estradiol, respectively. The term includes, but is not limited to, steroids, especially atamestane, exemestane and formestane; and, in particular, non-steroids, especially aminoglutethimide, roglethimide, pyridoglutethimide, trilostane, testolactone, ketokonazole, vorozole, fadrozole, anastrozole and letrozole. Exemestane is marketed as AROMASIN; formestane as LENTARON; fadrozole as AFEMA; anastrozole as ARIMIDEX; letrozole as FEMARA or FEMAR; and aminoglutethimide as ORIMETEN. An aromatase inhibitor is particularly useful for the treatment of hormone receptor positive tumors, e.g., breast tumors.

The term "anti-estrogen", as used herein, relates to a compound which antagonizes the effect of estrogens at the estrogen receptor level. The term includes, but is not limited to, tamoxifen, fulvestrant, raloxifene and raloxifene hydrochloride. Tamoxifen can be administered in the form as it is marketed, e.g., NOLVADEX; and raloxifene hydrochloride as EVISTA. Fulvestrant can be formulated as disclosed in U.S. Pat. No. 4,659,516 and is marketed as FASLODEX. A combination of the invention comprising a chemotherapeutic agent which is an anti-estrogen is particularly useful for the treatment of estrogen receptor positive tumors, e.g., breast tumors.

Tamoxifen breast cancer treatment is widely used as an endocrine therapy in patients with recurrent disease in ER-α and/or PgR positive tumors, and objective response (CR or PR) to tamoxifen approximately is present in 60—70% in women with positive receptors tumors. There is considerable interest in the elucidation of the resistance molecular mechanisms to this important anticancer drug or discover some predictive molecular markers which can predict the tamoxifen response and clinical benefit.

Tissue specific miRNAs expression is commonly deregulated in human cancers. Down regulated hsa-miR-125a and hsa-miR-125b in breast cancer, hsa-miR-let-7 in lung cancer and hsa-miR-143 and hsa-miR-145 in several different cancer types suggest that their regulation plays a significant role in tumor biology.

Recent studies have demonstrated that expression levels of several miRNAs exert an important influence on treatment response. In ovarian cancer, hsa-miR-214 conferred cisplatin resistance by targeting PTEN. Similar studies showed that inhibition of miR-21 and miR-200b sensitized cholangiocytes to gemcitabine.

Thus although altered expression of miRNAs in primary human cancers may be used for tumor diagnosis and prognosis and it's potential involvement in cancer treatment response in general has been implicated, in breast cancer tumors, this has not been suggested in relation to tamoxifen or an aromatase inhibitor therapy and patient behaviour. Here we report on miRNA expression levels of hsa-miR-30a-3p, hsa- miR-30c, and hsa-miR-182 in 246 ER-α positive breast cancer patients treated with tamoxifen for recurrent disease.

It has now been found that ER and PgR niRNA levels are positively correlated with hsa-miR-30a-3p, hsa-miR-30c, and hsa-miR-182. Therefore, the present results show that high expression levels particularly of those three miRNAs are associated with better tamoxifen clinical benefits and better progression free survival. It could be explained through their influence on the most relevant pathways related with tamoxifen or an aromatase inhibitor resistance.

It has now been found that hsa-miR-30a-3p exerts a significant influence on "Ceramide signaling pathway" and modulates the overexpression of MAP2K4 and BCL2 members.

The present results as described below may indicate that high BCL2 may be indicative of an intact ER pathway that is driving tumor growth and should be sensitive to endocrine therapy.

The present results suggest that the mitogen-activated protein kinase kinase (MAPKK) homolog MAP2K4 also called MKK4 is a breast cancer metastasis suppressor gene encoded by human chromosome 17. Hsa-miR-30c has now been found to be correlated to "Rac 1 cell motility signalling pathway" as well as hsa-miR-30a-3p. It suggests that both miRNAs have a collaborative participation on different pathways and conferring better tamoxifen response.

The present invention is based on the discovery of a connection between expression levels of hsa-miR-30a-3p, hsa-miR-30c, and hsa-miR-182 and their association with relevant genes included in pathways that plays an important role on tamoxifen or an aromatase inhibitor treatment response. The determination of the miRNAs expression levels on primary tumors can help clinicians to choose a better treatment for recurrent breast cancer on ER+ tumors and predict the individual clinical benefits for individual patients.

The present invention concerns the prediction or the detection of a tumor of being resistant to cancer therapy, in particular hormone therapy in breast cancer patients, using miRNA expression profiling. Information obtained from the present invention will assist a health care provider in determining whether or not a tumor (including cells therein) will become resistant to the hormone therapy or whether or not a tumor is resistant to the hormone therapy prior to the start of said therapy. In specific embodiments, the present invention will provide direction whether or not to continue with hormone therapy, or whether to start at all. In particular aspects of the invention, tamoxifen or an aromatase inhibitor is the exemplary embodiment described for illustrative purposes only, and a skilled artisan recognizes that the invention can be utilized for other chemotherapeutic drugs also, including other hormone therapy drugs, in particular estrogen inhibitors or estrogen receptor antagonists.

The present invention predicts the occurrence or recurrence of metastases due to therapy resistance, provides information concerning present metastases and/or helps to prevent mestastases by early intervention by identifying tumors susceptible to becoming resistant or being resistant and taking necessary actions based thereon.

The present invention identifies specific miRNAs which are differentially expressed in a tumor that is resistant compared to a tumor that is not resistant. More specifically, tumors with higher expression levels of hsa-miR-30a-3p, hsa-miR-30c, and hsa-miR-182 than a control were shown to be associated with a better response to treatment with tamoxifen or an aromatase inhibitor.

Furthermore, the miRNAs as described herein are useful as biomarkers to avoid hormonal therapies, such as antiestrogens, including tamoxifen or an aromatase inhibitor, raloxifene (Evista), and/or fulvestrant, for example, in case there is a likelihood of developing resistance.

The expression level of an miRNA can be determined by using any suitable method for quantitative detection of RNA. miRNAs can be isolated from the samples by methods well known to those skilled in the art as described, and are commercially available for instance in the form of RNABee ™ kit from Campro, Veenendaal, the Netherlands or the PARIS ™ kit from Applied Biosystems ABI, Nieuwerkerk a/d Ijssel, the Netherlands. Methods for detecting the amount of miRNAs are well known in the art and include, but are not limited to, northern blotting, reverse transcription PCR, real time quantitative PCR and other hybridization methods in particular microarrays or gene-chip arrays. Preferably, miRNA detection is performed using a Taqman® Human MicroRNA Assay.

The present invention employs miRNA expression profiling to compare biological phenotypes of tumors and evaluating the outcome thereof in the light of the chance that the tumor is resistant. For this, the test sample can be compared to a positive (resistant) control. Alternatively, the test sample can be compared to a negative/susceptible control. Certain miRNA's can be used as internal controls for normalizing expression levels. In one embodiment, miRNA levels in tumors sensitive to endocrine therapy are compared to miRNA level in tumors resistant to endocrine therapy to establish significance levels for the diagnostic test.

Preferably, said tumor comprises a representative part of said tumor. In a preferred embodiment of a method according to the invention, said tumor is classified as ER- or PgR- positive, preferably using E. O. R. T. C. or UICC criteria. Preferably, the cut-point which is used to classify a tumor as ER- or PgR- positive is 10 fmol/mg cytosolic protein. Preferably, said ER or PgR- positive tumor comprises at least 30% nuclei evidently of epithelial tumor origin. More preferably, said at least 30% of the nuclei are distributed uniformly over at least 70% of the section area. Preferably, said tumor sensitive to endocrine therapy is a tumor which has not progressed after about 2 years following initial diagnosis. Preferably, said resistant tumor is a tumor that has pregressed within less than 6 months, preferably during ongoing endocrine therapy treatment. The invention exploits the biological characteristic of primary, circulating or metastatic tumor cells of endocrine therapy sensitive patients being molecularly different from the primary or metastatic, circulating or metastatic tumor cells of endocrine therapy refractory patient.

In a preferred embodiment, the difference between said level of said at least one miRNA and said reference value or said level of a control sample is at least greater than 1, 1.5, 2, 2.5, 3 or 3.5.

In a preferred embodiment said endocrine therapy comprises an estrogen antagonist. In a highly preferred embodiment, said estrogen antagonist comprises tamoxifen or an aromatase inhibitor, most preferably tamoxifen.

EXAMPLES Patients

The Medical Ethical Committee of the Erasmus Medical Center Rotterdam, the Netherlands, approved our study design (MEC 02.953). This retrospective study used 246 blind coded, ER protein-positive (> 10 fmol/mg protein) primary tumor tissues from patients diagnosed between 1981 and 1995, in accordance with the Code of Conduct of the Federation of Medical Scientific Societies in the Netherlands (http:/www.fmvv.nl). These patients were treated either with breast-conserving surgery (34%) or with modified mastectomy (66%). An axillary dissection was performed in 93% of the patients (n=228). None of the patients had received neoadjuvant therapy or was exposed to hormonal adjuvant treatment (hormone-naive). Twenty-eight patients received CMF (cyclophosphamide, methotrexate, 5-fluorouracil) and 18 patients received anthracyclin-containing chemotherapy. Relevant clinicopathologic characteristics of the patients and their primary tumor are given in Table 1. Thirty patients presented with distant metastasis at diagnosis or developed distant metastasis (including supraclavicular lymph node metastasis) within 1 month after primary surgery (Ml-patients). These 30 patients and the 216 patients who developed a recurrence during follow-up [26 patients with local-regional relapse (LRR), 209 patients with distant metastasis (DM)] were treated with first-line tamoxifen (40 mg daily). Of the 216 MO-patients, the median time between primary surgery and start of therapy was 29 months (range, 2-134 months). At the time of surgical removal of the primary tumor the median age of the patients was 60 years (range, 26-89 years) and at start of tamoxifen therapy for recurrent disease 62 years (range, 29-90 years). Response to tamoxifen therapy was defined by standard Union International Contre Ie Cancer (UICC) criteria criteria (1). Objective response was observed in 45 patients (12 complete remission, CR, and 33 partial remission, PR), and 78 patients had a tumor progression of 25% or more, or showed new tumor lesions within 3 months (progressive disease, PD). The 123 patients with no evident tumor reduction of 50% or more (PR) or a tumor-progression, were considered as patients with no change (NC). These patients with NC were divided in 112 patients who had a NC > 6 months (defined as stable disease, SD) and 11 patients with a NC < 6 months. The median progression free survival (PFS) was for CR: 41 months, PR: 14 months, SD: 14 months, NC < 6 months: 6 months, and for PD: 3 months. Because the patients with SD had a PFS similar to patients with PR, we classified these patients as responders to tamoxifen as advised by the E.O.R.T.C. Therefore, clinical benefit was defined in our study as CR + PR + SD. Only patients with measurable disease were evaluated. Hundred-fifty- seven patients (64%) showed clinical benefit of tamoxifen therapy. The median follow-up of patients alive after surgery was 86 months (range, 10- 165 months), and 25 months (range, 1-95 months) after start of tamoxifen therapy. At the end of the follow-up period, 229 patients had developed tumor progression and 181 patients had died.

Methods Tissue Processing

After primary surgery, a representative part of the tumor was selected by the pathologist, frozen in liquid nitrogen, and sent to our laboratory for routine determination of ER and progesterone receptor (PgR) by ligand binding assay or enzyme immunoassay. Tumor cytosols were prepared and processed as recommended by the European Organization for Research and Treatment of Cancer. The cut-point used to classify tumors as ER- or PgR-positive was 10 fmol/mg cytosolic protein. The remainder of the tumor tissue was stored in our liquid nitrogen tumor bank at the Erasmus MC. For RNA isolation, 20 to 60 cryostat sections of 30 μm, corresponding to 30 to 100 mg, were cut from these tissues. Before, during, and after cutting the sections for RNA isolation, 5 μm sections were cut for H&E staining to assess the amount of tumor cells relative to the amount of surrounding stromal cells. In this study, only ER-positive specimens with at least 30% of the nuclei evidently of epithelial tumor cell origin and distributed uniformly over at least 70% of the section area were included.

RNA isolation

Total RNA was extracted with RNABee (Campro, Veenendaal, the Netherlands) according to the manufacturer and stored aliquoted in RNase/DNase-free water at -80⁰C.

miRNA selection

To identify miRNAs possibly related to tamoxifen response, we performed a pre- screening on 15 ER-positive primary tumors from hormone- naive breast cancer patients that were treated first-line with tamoxifen : 10 responders (CR + PR + SD>6 months) versus 5 non responders (PD + SD<6months). For this, the TaqMan Human MicroRNA Assay Set (release date March 2006) from Applied Biosystems (ABI, Nieuwerkerk a/d IJssel, the Netherlands), consisting of 259 unique assays to quantify 247 miRNAs and 12 controls (hsa-miR-132, hsa-miR-374, Z30 and nine different

SNORs/RNUs), was used as described before (Murphy et al. Br J Cancer 2002;87:1411-6) to screen this set of 15 ER+ breast cancer specimens for differentially expressed miRNAs. The use of specific primers with a hairpin structure during cDNA synthesis and mature miRNA specific detection probes precluded the detection of precursor miRNAs. A Mann- Whitney test (2-tailed P<0.05), showed that five out of these 247 miRNA's were differentially expressed between tamoxifen responders and non-responders. Expression levels of those five, hsa-miR-422a (P=O.004), hsa-miR-30a-3p (P=0.004), hsa-miR-187 (P=0.005), hsa-miR-30c (P=0.014) and hsa-miR-182 (P=O.027), were further analyzed real-time by quantitative reverse transcriptase polymerase chain reaction (qRT-PCR) in the remaining 246 tamoxifen treated patients included in this study.

Multiplex cDNA synthesis and miRNA quantification To analyse the expression of the thus selected miRNAs, the corresponding individual TaqMan human MicroRNA Assay kits from ABI were used. In summary, up to 30 different RT primers (250 nM each) were pooled, concentrated for 1 hour in a speed vacuum centrifuge at 50°C and resuspended in nuclease-free ddH2O to a final concentration of 40 nM each. Fifty ng of total RNA sample aliquots were reverse-transcribed in a final volume of 12 μl with a final concentration of 10 nM for each RT primer using the TaqMan MicroRNA Reverse Transcription Kit (ABI) according to the manufacturer's instructions.

Before performing a duplicate real-time PCR for each of the miRNAs separately, RT samples were diluted in nuclease-free ddH2O and amplified in a final volume of 20 μl containing cDNA synthesized from 0.5 ng of total RNA, 6 μl of TaqMan universal PCR master mix without UNG (ABI), 0.6 μl of the individual TaqMan MicroRNA primer and probe assays and 8.4 μl of nuclease-free dH2O. Real-time PCR was performed using the Stratagene Mx3000P QPCR System (Agilent Technologies, Waldbronn, Germany), with cycling conditions according to the manufacturer's instructions. A standard curve prepared from a pool of RNA obtained from a randomly selected set of 33 human breast tissues included in this study was included in each cDNA synthesis and PCR run, and the resulting data were used to normalize for in- between experimental variations. In all cDNA synthesis runs a minus RT reaction was incorporated, which proved to be negative for all assays. PCR efficiency, linearity, and the upper and lower detection limits of each of the individual miRNA assays were validated with the standard curve constructed from the simultaneously run serially diluted pool of RNA. In addition, to verify that the multiplex RT approach had not affected the quantification of specific miRNAs, the performance of all assays were validated with this standard curve set of samples in uniplex RT reactions. To check for RNA input, integrity, and yield of amplifiable cDNA, we proceeded as previously reported. In brief, two miRNA's, hsa-miR-132 and hsa-miR-374, were used as reference to normalize and quantify miRNA expression levels (Murphy et al. Br J Cancer 2002;87:1411-6).

Statistics

For statistical computations, STATA statistical package 10.0 (STATA Corp., College Station, TX) was used. Differences in levels were assessed with the Mann- Whitney U test or Kruskal-Wallis test, including a Wilcoxon-type test for trend, when appropriate. In these tests, patient and tumor characteristics were used as grouping variables. The strengths of the associations between continuous variables were tested with the Spearman rank correlation (Rs). To reduce the skewness most variables were log- transformed or Box- Cox transformed. Four equal parts were used to categorize the variable to low, low-intermediate, intermediate-high and high. Survival curves were generated using the method of Kaplan and Meier (1958) and the logrank test was used to test for differences. The Cox proportional hazard model was used to calculate the hazard ratio (HR) and 95% confidence interval (CI) in the analyses of time to progression free survival (PFS), with the starting point of PFS being defined as the start of the first-line of systemic treatment with tamoxifen for recurrence and the endpoint as the first detection of progression of the disease. Logistic regression analysis was used to examine the relation of miRNA levels with clinical benefit of tamoxifen therapy and for the calculation of the Odds Ratio (OR) and its 95% CI. A two-sided P-value of < .05 was considered statistically significant.

Pathways and global testing

Affymetrix microarray gene expression data (HG-U133A chips) previously deposited in the NCBI/GEO database (entries GSE2034 and GSE5327) were combined with gene expression data from HG-U133-plus2 chips. In total, 425 ER-positive samples were available for analysis. Data were preprocessed as previously described (Campbell et al. J

Biol Chem 2001; 276:9817-24); gene expression signals were calculated using Affymetrix GeneChip analysis software MAS 5.0. Global scaling was performed to bring the average signal intensity of a chip to a target of 600 before data analysis. Next, separately for each chip-type, data were normalized before combining the data-sets. Normalization was performed as described previously (Murphy et al. J Steroid Biochem MoI Biol 2006;102:139- 46.); in short, for each probe set intensities were threshold at 30, and were then expressed relative to the geometric mean of that probe set and were 2- log transformed. After combining the data, two analyses were performed to check if a bias exists between data based on the chip-type; 1) the median levels of all probe-sets were calculated for the samples of each chip-type. Comparing these median levels showed an average difference of only 0.019 between the median levels. 2) SAM analysis was performed to check if significant differentially expressed genes were identified between the 133a- data and plus2-data. No significant genes are present. Next, 71 samples were selected of which both miR and gene expression data were available.

Using the median, a sample was labeled 'high' or 'low' for a miRNA expression. Taken in to account the top 20 or bottom 20 of all samples, according to the expression of that particular miRNA. The 20 samples with the highest hsa-mir-30c expression were compared with the 20 samples with the lowest hsa-mir-30c expression, and likewise for each miRNA of interest.

The Global Test program was used (version 4.4.0) to associate Biocarta pathways (http://www.biocarta.com/) to samples expressing high or low amounts of a particular miRNA, irrespective of their association with response to therapy. All P-values were corrected for multiple testing and checked by re-sampling if an equally sized, randomly chosen group of genes is also significant (1,000 samplings). Pathways were considered of interest if the P-value of the Global Test, after correcting for multiple testing and the resampling P-value were either at or below 0.05. Pathway P-values mentioned in the text are two-sided P-values corrected for multiple testing, except where stated otherwise. The contribution of individual genes in a pathway was evaluated using z- scores calculated by the Global Test program. Genes with z-scores>1.96 were considered significant contributors to the pathway. R version 2.4.1 (http://www.cran.r-project.org) was used to run the Global Test package.

Results

Spearman correlations between ER, PgR mRNA levels and miRNAs . We have used our previously published data where we have confirmed that the sections used for RNA isolation were representative of the whole tumor with respect to ER and PgR mRNA expression.

Spearman rank correlation show positive correlations (P<0.001 for n = 246) between ER mRNA levels and hsa-miR-30a-3p (rs = 0.279), hsa-miR- 30c (rs = 0.308), and hsa-miR-182 (rs = 0.275). Furthermore, PgR mRNA levels are positively correlated with hsa-miR-30a-3p (rs = 0.444), hsa-miR-30c (rs = 0.336) and hsa-miR-182, (rs = 0.177). Levels of hsa-miR-422a were negatively associated with ER and PgR mRNA levels (rs = -0.133; and rs = - 0.152 respectievely).

In addition, hsa-miR-30a-3p expression levels were positively associated with: hsa-miR-30c (rs = 0.744) and with hsa-miR-182 (rs = 0.310). Besides, hsa-miR-30c, was also correlated with hsa-miR-182 (rs = 0.409; PO.001).

miRNAs expression levels and clinicopathologic factors

Tumors and characteristics and their association with the levels of the 5 candidate miRNAs are summarized in Table 1.

Note that the patient's age was associated with hsa-miR-30c (P < 0.001) and hsa-miR-182 (P = 0.008); showing that patients older than 70 years old (better responders) had higher miRNA expression levels compared with younger than 50. Besides, the same is true for postmenopausal patients status (P = 0.036; P = 0.030 respectively).

Univariate and Multivariate analysis for response to tamoxifen therapy.

In Cox univariate logistic regression analysis, using log- transformed continuous variables, increasing levels of hsa-miR-30a-3p (OR, 1.51(1.16-1.96); P = 0.002), hsa-miR-30c (OR, 3.87(2.16-6.93); P O.001), and hsa-miR-182 (OR, 1.53(1.09-2.16); P =0.013), were associated with favourable tamoxifen clinical benefits. In multivariate analysis, in addition to the base model, included the traditional predictive factors: menopausal status, dominant site of relapse, disease free interval and ER an PgR tumor levels, only hsa-miR-30c remains associated with better tamoxifen clinical benefits (OR, 3.14(1.61- 6.12); P = 0.001).

Tumour miRNA expression levels were categorized in quartiles to find out its possible effect in patient's response to tamoxifen therapy (Table 2). In univariate logistic regression analysis, we found a better response to treatment in patients with higher expression levels (4^th quartile) of hsa-miR- 30a-3p (OR, 3.57(1.65-7.73); P = 0.001), hsa-miR-30c (OR, 4.62(2.07-10.2); P < 0.001), and hsa-miR-182 (OR, 3.04(1.37-6.70); P < 0.006), compare to lower expression levels (1^st quartile). In multivariate analysis, only hsa-miR-30a-3p (OR, 2.61(1.05-6.52); P = 0.039), and hsa-miR-30c (OR, 3.64(1.46-9.05); P = 0.005) added an additional significant contribution to the basic model designer including the classical predictive factors.

Percentages of patients and tamoxifen treatment response, in function of our three miRNA candidate's expression levels, were calculated comparing the number of responders from total of patients included in each quartile using Pearson Chi- Square test. Patients in 4^th quartile, have better tamoxifen clinical benefits versus 1^st quartile: hsa-miR-30a-3p, 77% (47/61) versus 48% (31/64) (P = 0.007), hsa-miR-30c, 80% (49/61) versus 47% (30/64) (P < 0.001) and for hsa-miR-182, 79% (48/61), versus 55% (34/62) (P =0.026).

hsa-miR-30a-3p, hsa-miR-30c, hsa-miR-182 and Progression free survival (24months)

In order to asses the progression free survival, we have used the miRNA expression levels categorized in quartiles as described above. Figure 1 shows progression-free survival as finction of the categorized miRNA level in all 246 patients. In Cox proportional hazard model analysis, patients in the 4thquartile, had higher expression levels of those three miRNA and also had a better progression free survival at 24 months: hsa-miR-30a-3p (HR, 0.51(0.34-0.76); P = 0.001), hsa-miR-30c (HR, 0.47(0.31-0.70); P < 0.001), and hsa-miR-182 (HR, 0.57(0.37-0.86); P = 0.008 respectively). Additionally, this fact was related to the time to progression (TTP). It means that patients in the upper quartile of each miRNA, had better progression free survival (8 months as a median time to progression) compared with patients in 1^st quartile (data not show).

Pathways and global testing Using global testing pathway analysis to the patients with ER+ tumors, treated with tamoxifen as first-line systemic treatment for recurrent disease and available transcriptome information on these same specimens, allowed us to associate biological pathways with miRNAs and tamoxifen treatment response (Table 3). Hsa-miR-30c was most significantly coupled to "HER2, signal transduction and oncology pathway" (comparative P = 0.014 and FDR adjusted P = 0.007 ), in which ERa and HER4 were overexpressed and EGFR was under-expressed in tumors having high versus those having low hsa- miR-30c expression (Figure 2). hsa-miR-30c also was related to "ERK 1. MAPK signalling pathway" and "PTEN dependent cell cycle arrest and apoptosis" (comparative P = 0.053 /FDR adjusted P = 0.014 and comparative P = 0.054 /FDR adjusted P = 0.014 respectively) (Table 3). Either in MAPK or in MAPK pathways, EGFR gene was under expressed.

Additionally, both hsa-miR-30c and hsa-miR-30a-3p were associated to "RAC 1 cell motility signalling pathway" comparative P = 0.025/ FDR adjusted P = 0.014 and comparative P = 0.016/ FDR adjusted P = 0.016 respectievely. Platelet-derived growth factor receptor α (PDGFRα) and Chimerin 1 (CHNl) genes were under expressed, showing a mutual influence on mentioned pathway. Several studies was observed that overexpression of PDGFRα, a member of breast cancer anti estrogens (BCAR) genes, was associated with breast cancer progression. (Figure 2 (b,c))

Hsa-miR-30a-3p was also associated with "Ceramide signalling pathway" (comparative P = 0.019/ FDR adjusted P = 0.012) in which Bcl-2 and MAP2K4 were overexpressed in tumors having high versus those having low hsa-miR-30a-3p expression. Several studies show a positive correlation between Bcl-2 and ER-α expression levels imply that Bcl-2 is a significant favourable prognostic factor for breast cancer treated with chemotherapy and endocrine therapy (35-37). Additionally, Spearman Rank test showed a significant negative correlation between hsa-miR-30c and EGFR, PDGFRA, CHNl, hsa-miR-30a- 3p only with CHNl and has-miR-182 with PDGFRA mRNA levels.

Hsa-miR-182 was associated to "Control of gene expression by Vitamin D Receptor pathway"comparative P = 0.004/ FDR adjusted P = 0.409.

Example 2

In this study, using the methods and the cohort described earlier (Example 1), we measured using real-time PCR levels over various candidate microRNAs (Hsa-miR-7, hsa-miR-210, hsa-miR-373 and hsa-miR-10b, hsa- miR-205, hsa-miR-98, hsa-miR-374, hsa-miR-335, hsa-miR-22, hsa-miR-214, hsa-miR-212, hsa-miR-21, hsa-miR-187, hsa-miR-17-5p, hsa-miR-132, hsa- miR-125b, hsa-let-7g, hsa-miR-365, hsa-miR-518b, hsa-miR-126*, hsa-miR- 182, hsa-miR-34a, hsa-miR-432, hsa-miR-141. hsa-miR-126, hsa-miR-422a, hsa-miR-489, hsa-miR-520c).

Hsa-miR-7, hsa-miR-210, hsa-miR-373 and hsa-miR-lOb, hsa- miR-205 were selected because us and others have shown that these microRNAs are associated with aggressiveness. For this reason these microRNAs for their role in endocrine resistance in breast cancer. The remaining microRNAs (hsa-miR-98, hsa-miR-374, hsa-miR- 335, hsa-miR-22, hsa-miR-214, hsa-miR-212, hsa-miR-21, hsa-miR-187, hsa- miR-17-5p, hsa-miR-132, hsa-miR-125b, hsa-let-7g, hsa-miR-365, hsa-miR- 518b, hsa-miR-126*, hsa-miR-182, hsa-miR-34a, hsa-miR-432, hsa-miR-141. hsa-miR-126, hsa-miR-422a, hsa-miR-489, hsa-miR-520c) were included because they were predicted targets for factors involved in tamoxifen resistence such as ER, PgR, Her-2, ERBB3, ERBB4, EGFR, EZH2, GRB7, SIAH2, BCARl, TNC (and other extracellular matrix genes), AKT3, NCOl, TFFl or other work had suggested a role in therapy resistance (hsa-miR-212). In the cohort described in example 1 using the methods described above, we found using Cox univariate logistic regression analysis, using log- transformed continuous variables, the levels in the primary tumor of hsa- miR-7, hsa-let-7g, hsa-miR-17-5p, hsa-miR-126, hsa-miR-126#, hsa-miR-141, hsa-miR-145, hsa-miR-335, hsa-miR-365, hsa-miR-489, and hsa-miR-22 were significantly associated with clinical benefit of endocrine therapy in patients receiving tamoxifen as a first line treatment for metastatic breast cancer (Table 2).

Tablel. Associations of biological factors with clinico pathological factors

Median miRs levels (interquartile range) after normalization to the housekeeper miR set

Charactri sties Patients hsa-miR-30a-3p hsa-miR-30c hsa-miR-182 hsa-miR-187 hsa-miR-422a

Age (years)¹

<55 92 0.59 (0.8) 0.94 (0.77) 0.80 (0.75) 0.19(0.70) 0.99 (0.71)

56-70 90 0.68 (1.0) 1.01 (0.95) 1.03 (1.18) 0.19(0.61) 0.95 (0.55)

>70 64 0.76 (1.8) 1.34 (1.23) 0.95 (1.12) 0.26 (1.31) 1.08 (0.77)

P=0.095^§ P=0.001^§ P=0.008^§ P=0.064^§ P=0.075^§

Menopausal status²

Premenopausal 59 0.60 (0.80) 0.87 (0.69) 0.78 (0.81) 0.15 (0.67) 0.96 (0.85)

Postmenopausal 187 0.70(1.17) 1.08(1.07) 0.96 (1.08) 0.21 (0.79) 1.01 (0.66)

P=0.091^§ P=0.036^§ P=0.030^§ P=0.158^§ P=0.272^§

Dominant site of relapse

Local regional relapse 32 0.48 (0.73) 0.88 (0.97) 0.81 (0.43) 0.22 (0.67) 1.13 (0.78)

Bone 128 0.70 (1.27) 1.05 (0.81) 1.01 (0.99) 0.22 (0.81) 1.00 (0.63)

Visceral 86 0.56 (0.85) 1.01 (1.10) 0.91 (1.34) 0.20 (0.63) 0.94 (0.55)

P=0.318t P=0.816t P= 0.13Ot P=0.623t P=0.198t K)

Disease-free interval (month)

<12 64 0.51 (0.68) 0.88 (0.89) 0.91 (0.85) 0.21 (1.37) 1.13(0.72)

13-36 108 0.74(1.13) 1.05 (0.98) 0.94(1.14) 0.18(0.48) 0.99 (0.54)

>36 74 0.86 (1.35) 1.14(0.99) 1.01 (0.91) 0.22 (0.84) 0.93 (0.71)

P= 0.172t P=0.162t P=0.944t P=0.219t P=0.157t

Tumor Size (cm)

<2 65 0.60 (1.41) 0.98 (0.93) 1.01 (0.78) 0.19(0.83) 0.95 (0.50)

2-5 146 0.66 (0.99) 1.02 (0.85) 0.90 (1.08) 0.21 (0.63) 1.04 (0.73)

>5 35 0.66 (0.89) 1.28 (1.05) 0.91 (1.50) 0.09 (0.54) 0.93 (0.67)

P=0.933t P=0.505t P= 0.911t P=0.214t P=0.196t

Grade

Poor 139 0.70 (1.06) 1.02 (0.97) 0.89 (0.89) 0.23 (0.87) 1.00 (0.73)

Unknown 77 0.65 (0.81) 1.05 (0.75) 1.02 (0.95) 0.19(0.62) 1.01 (0.55)

Good/moderate 30 0.49(1.52) 1.14(1.46) 0.90 (0.86) 0.22 (0.71) 0.95(0.48)

P=0.731t P=0.884t P= 0.496t P= 0.72Ot P=0.548t

Nodal status

NO 96 0.74 (1.01) 1.04 (0.93) 1.00 (1.02) 0.19(0.68) 1.00 (0.53)

Nl-3 59 0.70 (1.22) 1.08(1.12) 0.82 (1.03) 0.22 (0.66) 0.93 (0.64)

N>3 76 0.46 (0.90) 0.94 (0.86) 0.84 (0.77) 0.18(0.55) 1.10(0.77)

P=0.287t P=0.671t P=0.293t P=0.521t P=0.148t

At starts Tamoxifen treatment for recurrent disease; ² At primary surgery;^§ P for Mann-Withney U test; t P for Kruskal-Wallis test

Table 2. Univariate and multivariate analysis for response to tamoxifen therapy

Univariate analysis Multivariate analysis

Patients Clinical OR (95% CI) OR (95% CI)

Factor benefit (%) P P

246 64

All patients

Age (years)¹

<55 92 55 1 1

56-70 90 66 1.53 (0.84-2.78) 0.163 0.91 (0.36-2.32) 0.859

>70 64 73 2.22 (1.11-4.43) 0.023 1.59 (0.59-4.29) 0.351

Menopausal status¹

Premenopausal 59 53 1 1

Postmenopausal 187 67 1.86 (1.02-3.38) 0.040 1.33 (0.52-3.38) 0.544

Dominant site of relapse¹

32 72 1 1

Local regional relapse

Bone 128 59 0.57 (0.24-1.33) 0.196 0.54 (0.21-1.37) 0.196

Visceral 86 67 0.81 (0.33-1.97) 0.644 0.64 (0.23-1.76) 0.396

Disease-free interval

(months)¹

<12 64 41 1 1

13-36 108 70 3.47 (1.81-6.63) O.001 3.92 (1.95-7.90) O.001

>36 74 74 4.23 (2.05-8.70) O.001 4.37 (1.99-9.56) <0.001

246 64 1.28 (1.12-1.46) O.001 1.28 (1.09-1.50) 0.001

ER-cf

PgR² 246 64 1.13 (0.98-1.30) 0.082 1.03 (0.87-1.22) 0.703

Additions to the base model hsa-miR-30a-3p

1^st Quartile 64 48 1 1

2^ndQuartile 62 61 1.68 (0.83-3.42) 0.148 2.14 (0.97-4.71) 0.059

3^rd Quartile 59 69 2.42 (1.15-5.08) 0.019 2.24 (0.95-5.26) 0.064

4^th Quartile 61 77 3.57 (1.65-7.73) 0.001 2.61 (1.05-6.52) 0.039 hsa-miR-30c

1^st Quartile 64 47 1

2°^dQuartile 60 52 1.21 (0.59-2.45) 0.594 1.15 (0.53-2.52) 0.710

3^rd Quartile 61 77 3.80 (1.75-8.24) 0.001 3.08 (1.28-7.41) 0.012

4^th Quartile 61 80 4.62 (2.07-10.2) O.001 3.64 (1.46-9.05) 0.005 hsa-miR-182

1^st Quartile 62 55 1

2°^dQuartile 63 57 1.09 (0.54-2.22) 0.795 0.88 (0.39-1.97) 0.768

3^rd Quartile 60 65 1.52 (0.73-3.17) 0.253 1.47 (0.65-3.32) 0.347

4^th Quartile 61 79 3.04 (1.37-6.70) 0.006 2.17 (0.89-5.28) 0.087

Biological factors were separately introduced as log-transformed continuous variable to the base multivariate model that included the factors menopausal status, dominant site of relapse, disease-free interval, and ER-a and PgR mRNA levels as log-transformed continuous variables 2At start of first-line therapy for recurrent disease Table 3. List of the most significant pathways associated with hsa-miR-30a-3p, hsa-miR-30c and hsa-miR-182

hsa-miR-30a-3p All Tested FDR Comparative

Statistic Q Expected Q SD of Q P-value genes genes adjusted p-value

Rac 1 cell motility signaling pathway 19 13 60.876 11.806 7.1726 3.84E-04 1.64E-02 0.016

Ceramide signaling pathway 20 17 27.812 6.9011 3.2678 1.31E-04 1.22E-02 0.019

Angiotensin II mediated activation of JNK Pathway via Pyk2 dependent signaling 25 18 30.071 7.762 4.0069 8.10E-04 1.64E-02 0.051

HIVI Nef. negative effector of Fas and TNF 48 36 19.999 6.5573 2.6118 6.38E-04 1.64E-02 0.062 Telomeres,Telomerase. Cellular aging and immortality 16 12 37.646 9.5416 5.6001 1.45E-03 2.21E-02 0.066 Keratinocyte differentiation 33 26 30.007 8.9835 4.0308 8.80E-04 1.64E-02 0.072 hsa-miR-30c

Role of ERBB2 in signal transduction and oncology 19 13 47.943 9.3745 5.1236 7.69E-05 7.15E-03 0.014 Raclcell motility signaling pathway 19 13 53.846 10.503 6.3765 4.60E-04 1.46E-02 0.025 Erkl, Erk2 Mapk signaling pathway 25 20 37.058 9.1908 4.7839 6.28E-04 1.46E-02 0.053 PTEN dependent cell cycle arrest and apoptosis 18 11 34.24 8.1098 4.5284 5.61E-04 1.46E-02 0.054 Telomeres, Telomerase. Cellular aging and immortality 16 12 45.07 10.733 6.4698 1.16E-03 2.14E-02 0.076 Chaperones modulate interferon signaling pathway 17 15 46.459 11.341 6.9412 2.21E-03 2.57E-02 0.084 hsa-miR-182

Control of gene expression by vitamin D receptor 26 21 13.376 5.3147 2.4301 8.72E-03 4.09E-01 0.004

The 41BBdependent immune response 15 13 14.925 6.4773 3.4226 2.54E-02 4.09E-01 0.036

Role of ERBB2 in signal transduction and oncology 19 13 24.459 10.523 6.2525 3.58E-02 4.09E-01 0.056

Phospholipids as signaling intermediaries 20 15 26.91 11.491 6.7049 3.34E-02 4.09E-01 0.057

Erkl, Erk2 Mapk signaling pathway 25 20 19.789 9.1129 5.0573 4.07E-02 4.09E-01 0.071

Rac 1 cell motility signaling pathway 19 13 22.661 10.151 6.0536 4.26E-02 4.09E-01 0.071

71 samples were selected of which both hsa-miR and gene expression data were available.

In total, 107 pathways, including 10,520 genes, were included in the Global test. All genes of a pathway on the chip (tested genes) were in a multivariate model tested by logistic regression for their relation with the 20 samples that had the highest levels and 20 samples that had the lowest expression of the indicated miRNA. Statistic Q is the calculated statistic for the model; expected Q is the expected statistic of the model based on chance. The SD of (the expected) Q is also given. With those, a P value and the FDR adjusted P value were calculated. Similarly, a comparative P value is calculated based on the Q-statistic of gene sets randomly drawn from the dataset.

Table 4. miR listing as continuous variable or divided in quartiles and statistical significance.

1 1

1.023 0.73 1.44 0.897 1.159 0.594247 2.26227 0.664 0.833 0.59 1.18 0.301 2.477 1.212599 5.061769 0,013 0.42 0.86 0.003 .. :.:^■ 1.187045 4.965975 0,015 continuous

0.587 0.419 0.823 0.002 3.005 1.436 6.286 0.003 variable

1.178 0.82 1.68 0.369 0.675 0.324673 1.40 0.293 1.09 2.19 0.014 0.540 0.262755 1.11 0.094 1.12 2.28 0.010 •^• ;::. 0.228756 0.97 0.042 continuous

0.577 0.365 0.911 0.018 2.080 0.800 5.409 0.133 variable

h-«a-raiR-2 l 1 1

0.769 0.54 1.09 0.14 2.075 1.038688 4.146551 0.039 0.859 0.61 1.22 0.394 1.901 0.95398 3.786622 0.068 0.769 0.54 1.09 0.138 1.153769 4.711135 0.0S8 continuous bsa-«ύB-≤l 0.868 0.565 1.335 0.520 2.257 0.943 5.403 0.068 variable

1.000 1.000

0.47 0.95 0.025 1.505 0.768413 2.946501 0.233 0.44 0.87 0.008 --0.-1 1.983999 8.435257 0,000 0.35 0.69 0.000 '■ -:^■: 2.129216 9.519694 0,000 continuous

0.574 0.446 0.740 0.000 4.217 2.377 7.483 0.000 variable able 4 (continued).

Table 4 (continued).

2 1 .068 0 .75705 1 .50771 0.707 1 .354 0 .68928 2 .66113 0.379

3 0 .899 0 .63852 1 .26701 0.544 1 .475 0 .75009 2 .90068 0.260

4 0 .43023 0 .87368 0.007 ;< 1 .72212 7 .97009 0,001 ilBBI _™≤r^s 0 .824 0.709 0.959 0.012 1 .649 1.187 2.290 0,003

0.967 0.69 1.36 0.848 1.12637 0.57 2.22 0.730

0.858 0.61 1.21 0.386 1.53846 0.77 3.07 0.222 0.50 1.01 0.054 1.02 4.30 0.044

0.810 0.656 1.000 0.050 1.597 1.037 2.461 0.03 i

0.857 0.61 1.21 0.378 1.70249 0.86 3.35 0.124 0.46 0.93 0.0!? 1.33 5.50 0,006 0.53 1.06 0,09^<! 1.28 5.29 0.008

0.798 0.664 0.958 0.0 ! S 1.665 1.147 2.417 0,007

0.728 0.52 1.03 0.070 1.81685 0.91 3.65 0.093 0.45 0.89 0.010 1.55518 0.78 3.10 0.209 0.47 0.93 O,ϋiΛs 1.42415 0.72 2.83 0.313 continuous

0.769 0.650 0.911 0.002 1.140 0.827 1.571 0.425 variable

0.79301 0.56021 1.12255 0.191 2.26531 1.1315 4.53525 O.'i':]

0.72833 0.5166 1.02684 O.o? 3.56786 1.72898 7.36252 I¹OnI

0.8211 0.58126 1.15989 0.263 1.73371 0.87928 3.41844 0.112 continuous bsa-uύB-≤'2 0.875 0.723 1.061 0.174 1.528 1.087 2.149 O.Oiϊ variable Table 5 Tar et sequences of miRNAs as used herein

Claims

Claims

1. A method of predicting or establishing the resistance of a breast cancer tumor to endocrine therapy, comprising the steps of: a) determining the expression profile essentially consisting of miRNAs hsa-miR-7, hsa-let-7g, hsa-miR-17-5p, hsa-miR-126, hsa-miR-126#, hsa-miR- 141, hsa-miR-145, hsa-miR-335, hsa-miR-365, hsa-miR-489, hsa-miR-22, hsa- miR-30a-3p, hsa-miR-30c, and hsa-miR-182, preferably hsa-miR- 30a- 3p, hsa- miR-30c, and hsa-miR-182 in a sample from a breast cancer tumor; b) comparing said profile with a reference profile, and c) predicting or establishing the resistance of said breast cancer to said endocrine therapy based on said comparison.

2. A method of predicting or establishing the resistance of a breast cancer tumor to endocrine therapy, comprising the steps of: a) determining the expression level of at least one miRNA selected from the group consisting of hsa-miR-7, hsa-let-7g, hsa-miR-17-5p, hsa-miR-126, hsa-miR- 126#, hsa-miR- 141, hsa-miR-145, hsa-miR-335, hsa-miR-365, hsa- miR-489, hsa-miR-22, hsa-miR-30a-3p, hsa-miR-30c, and hsa-miR-182, preferably hsa-miR-30a-3p, hsa-miR-30c, and hsa-miR-182 in a sample from a breast cancer tumor; b) comparing said profile with a reference expression level, and c) predicting or establishing the resistance of said breast cancer to said endocrine therapy based on said comparison.

3. Method according to claim 2, wherein said at least one miRNA comprises hsa-miR-30a-3p, hsa-miR-30c and hsa-miR-182.

4. Method according to any one of claims 1-3, wherein step a is performed on a sample which has been exposed to said endocrine therapy prior to determining said level or profile.

5. The method according to any one of the preceding claims, wherein said comparison reveals a difference between the level of at least one miRNA and a non-resistant control level or profile that is greater than one-fold, or wherein said comparison reveals no difference between the level of at least one miRNA and a resistant control level or profile.

6. The method according to any of the preceding claims, wherein said endocrine therapy comprises anti-estrogen or aromatase inhibitor treatment.

7. The method according to any of the preceding claims, wherein said breast cancer tumor is a recurring tumor and/or an ER+ tumor.

8. Use of a marker selected from miRNAs hsa-miR-7, hsa-let-7g, hsa-miR- 17-5p, hsa-miR-126, hsa-miR-126#, hsa-miR-141, hsa-miR-145, hsa-miR-335, hsa-miR-365, hsa-miR-489, hsa-miR-22, hsa-miR-30a-3p, hsa-miR-30c, and hsa-miR-182, preferably hsa-miR-30a-3p, hsa-miR-30c, and hsa-miR-182 for diagnostic use in a method for predicting or establishing the resistance of a breast cancer tumor.

9. Kit of parts adapted for performing a method according to any one of claims 1-8, comprising up to primers and/or hybridization probes specific for the detection of at least one miRNA selected from the group consisting of hsa- miR-7, hsa-let-7g, hsa-miR-17-5p, hsa-miR-126, hsa-miR-126#, hsa-miR-141, hsa-miR-145, hsa-miR-335, hsa-miR-365, hsa-miR-489, hsa-miR-22, hsa-miR- 30a-3p, hsa-miR-30c, and hsa-miR-182, preferably hsa-miR-30a-3p, hsa-miR- 30c, and hsa-miR-182 and an instruction for performing the method according to any one of claims 1-8.