WO2009024790A1

WO2009024790A1 - Method and kit for the prognosis of cancer by determining the level of mirna-210

Info

Publication number: WO2009024790A1
Application number: PCT/GB2008/002836
Authority: WO
Inventors: Jonathan Mark Gleadle; Loannis Ragoussis; Adrian Llewellyn Harris
Original assignee: Isis Innovation Limited
Priority date: 2007-08-23
Filing date: 2008-08-22
Publication date: 2009-02-26

Abstract

A method for determining the prognosis of a cancer wherein the method comprises: (a) determining the level of miRNA-210 in a sample from an individual; and (b) using the level of miRNA-210 in the sample to determine the prognosis of the cancer.

Description

METHOD AND KIT FOR THE PROGNOSIS OF CANCER BY DETERMINING THE LEVEL OF MIRNA-210

This invention relates to a method and kit for determining the prognosis of a cancer, in particular the method involves determining the level of an miRNA in a sample from an individual with cancer wherein the level of the miRNA is correlated with the likelihood of recurrence of the cancer and/or survival of the individual.

According to the World Health Organisation cancer kills about 7.6 million people worldwide every year. In particular, cancers of the lung, stomach, liver, colon, kidney and breast are responsible for over half these deaths. Breast cancer is the commonest cancer in women, causing over half a million deaths a year.

In order for treatments of cancer to be effective, individuals require not only an early accurate diagnosis, but also a reliable prognosis of the disease. A reliable prognosis would allow the treatment given to a patient diagnosed with cancer to be adapted to the individual patient, for example, patients having a poorer prognosis may be given the most aggressive treatment.

Whilst methods to diagnose cancer are improving, and it is now possible to diagnose many cancers at very early stages, there are very few accurate and simple methods to determine the prognosis of a cancer. Typically, prognosis is determined on clinical presentation, with histological features, tumour size, invasion and spread to lymph nodes or metastasis being helpful indicators; however, this determination of prognosis is very inaccurate. In particular, it can be difficult to determine the prognosis of a cancer if the cancer is detected at a very early stage when clinical symptoms are not yet evident. Efforts to use molecular signatures to provide a prognosis have been made, for example WO 02/103320 and WO 2006/015312 both describe methods for determining the prognosis of a cancer using a plurality of marker genes. However, these methods are complicated and require many different markers to be analysed and compared.

Therefore, there remains a need for a method of establishing a prognosis that is simple, cost-effective and reliable. The method of this invention has the advantage of being reliable and cost effective and preferably only relies on one prognostic marker, thus making it simple.

According to a first aspect, the invention provides a method for determining the prognosis of a cancer wherein the method comprises: (a) determining the level of miRNA-210 in a sample from an individual; and (b) using the level of miRNA-210 in the sample to determine the prognosis of the cancer.

Preferably, the sample if obtained from an individual who has been diagnosed with cancer.

The sample preferably comprises cancer cells, however this may not be necessary. Preferably the sample comprises cancer cells obtained from a cancer tumour. The sample may be from a biopsy sample or a resection of a sample taken from the site of cancer in an individual. The sample may be a fluid sample removed from a patient, for example, blood or extracts derived from blood, lymph, urine, saliva or any other suitable fluid.

The sample may comprise cancer cells from one or more of cancer of the bone, breast, respiratory tract (e.g. lung) , brain, reproductive organs

(e.g. ovary, cervix) , digestive tract (e.g. gastro-intestinal tract and colorectal tract) , urinary tract, bladder, eye, liver, skin, head, neck, thyroid, parathyroid, kidney, pancreas, blood, ovary, colon, prostate, and metastatic forms thereof.

Preferably the method of the invention is used to determine the prognosis of an individual with breast cancer, renal cancer or a head and neck cancer.

The terms "microRNA" and "miRNA" are used interchangeably herein and refer to non-coding RNA oligonucleotides. Such oligonucleotides have been shown to play a critical role in cell differentiation, proliferation, death, metabolism and more recently in tumourigenesis. miRNAs are first transcribed as primary transcripts of about 70-100 nucleotides in length and are then processed to mature miRNAs of about 19-25 nucleotide RNA molecules.

miRNA-210 as referred to herein preferably has either: (i) the sequence of the primary transcript of SEQ ID No. 1 or a sequence with at least 80%, 85%, 90% or 95% or more identity to the sequence of SEQ ID No. 1; or (ii) the sequence of the mature sequence of SEQ ID No 2 or a sequence with at least 80%, 85%, 90% or 95% or more identity to the sequence of SEQ ID No. 2.

The sequence of the primary transcript of miRNA-210 is ACCCGGCAGUGCCUCCAGGCGCAGGGCAGCCCCUGCCCACCGCA CACUGCGCUGCCCCAGACCCACUGUGCGUGUGACAGCGGCUGAU CUGUGCCUGGGCAGCGCGACCC - SEQ ID No. 1.

The mature sequence of miRNA-210 is CUGUGCGUGUGACAGCGGCUGA - SEQ ID No. 2. The genomic localisation of miRNA-210 can be found at http://www.ensembl.org/Homo_sapiens/contigview?l= 11:556089-560198, and the sequence has been deposited at, for example, http: //microrna. Sanger. ac.uk/cgi-in/sequences/mirna_entry.pl?acc = MI0000286.

The term "identity" in the context of nucleic acid sequences refers to the residues in two sequences which are the same when aligned for maximum correspondence. The length of sequence identity comparison may be over a stretch of at least about nine nucleotides, usually at least about 20 nucleotides. There are a number of different algorithms known in the art which can be used to measure nucleotide sequence identity. For instance, polynucleotide sequences can be compared using FASTA, Gap or Bestfit, which are programs in Wisconsin Package Version 10.0, Genetics Computer Group (GCG) , Madison, Wis. FASTA provides alignments and percent sequence identity of the regions of the best overlap between the query and search sequences (Pearson, Methods Enzymol. 183:63-98 (1990)) . For instance, percent sequence identity between nucleic acid sequences can be determined using FASTA with its default parameters (a word size of 6 and the NOPAM factor for the scoring matrix) or using Gap with its default parameters as provided in GCG Version 6.1. Alternatively, sequences can be compared using the computer program, BLAST (Altschul et al. , J. MoI. Biol. 215:403-410 (1990) ; Gish and States, Nature Genet. 3:266-272 (1993) ; Madden et al. , Meth. Enzymol. 266:131-141 (1996) ; Altschul et al. , Nucleic Acids Res. 25:3389-3402 (1997) ; Zhang and Madden, Genome Res. 7:649-656 (1997)) , especially blastp or tblastn (Altschul et al. , Nucleic Acids Res. 25:3389-3402 (1997)) .

The level of miRNA-210 in a sample obtained from an individual may be determined using any technique suitable for detecting RNA levels in a sample. The level of either the primary transcript of miRNA-210 or the mature sequence of miRNA-210, or the level of both the primary transcript and the mature sequence of miRNA-210, may be determined. Suitable techniques for determining RNA levels, including miRNA levels, in a biological sample are well known to those skilled in the art and include, for example, Northern blot analysis, PCR techniques such as RT- PCR and real-time RT-PCR, in situ hybridisation, microRNA microarrays, RNAase protection assays, immunological assays or other means of detecting low MW RNA.

The detection of miRNA levels to determine cancer prognosis has an advantage over alternative prognosis methods that look at conventional RNA and/or DNA molecules as miRNAs display improved stability in the type of samples often available for analysis. For example, tumour specimens are commonly fixed in formalin and preserved in paraffin blocks and this can produce difficulties in conventional RNA and DNA analyses, miRNAs however survive these treatments surprisingly well (Nakajima et al. RNA 2007 Aug 13) .

Preferably, the method of the invention allows a favourable prognosis or an unfavourable prognosis to be given.

The method of the invention preferably provides a method for determining whether a cancer is aggressive. An unfavourable prognosis may be an indication of an aggressive cancer.

The method of the invention may be used to provide a prognosis for a cancer after the cancer has been diagnosed and/or the prognosis may be given after or during treatment for the cancer. The prognosis is preferably determined by comparing the level of miRNA-210 in a sample obtained from an individual with cancer to the level of miRNA-210 in a control sample. Preferably, the control sample is a sample of tissue taken from the same tissue type as used for the cancer patient sample, for example breast tissue or blood, from an individual who does not have cancer in that tissue. Alternatively the control sample may be a sample of tissue taken from near a cancer which is itself not cancerous. Control/normal levels of miRNA-210 may be derived from a pool or set of samples from normal tissue near a cancer or from samples taken from tissue which is not cancerous. Alternatively, control/normal levels of miRNA-210 may be defined from the levels of miRNA observed within a particular set of cancers.

The term 'prognosis' as used herein refers to the likely outcome of a cancer. The prognosis of a cancer may include the likely duration of the cancer, chances of complications of the cancer, probable outcomes of the cancer, the prospects for recovery, the recovery period for the cancer, the likely survival rate, the likely death rate, the likely recurrence of the cancer and other outcome possibilities. Most preferably the method of the invention allows a prognosis to be given indicating the likelihood of cancer recurrence or metastasis in a particular time, and/or the likelihood of death due to the cancer in a particular time.

Preferably a favourable prognosis is that a cancer patient has at least an about 70% or more chance that the cancer will not recur or metastasise within about two, three, four, five, six, seven, eight, nine, ten or more years from taking the sample. Preferably there is a 70% or more chance that the cancer will not recur or metastasise within 10 years. In another embodiment, a favourable prognosis may be that the cancer patient has an 80% or more chance that the cancer will not be fatal to the individual within about two, three, four, five, six, seven, eight, nine, ten or more years from taking the sample, preferably not within 10 years or more years. A favourable prognosis may include reference to recurrence and metastasis as well as, or as an alternative to, reference to the likelihood of patient survival.

Preferably a favourable prognosis will be given when a sample from a cancer patient displays miRNA-210 levels which are below a predetermined level. Preferably this predetermined level is relative to the level of miRNA-210 observed in a control sample. Such a sample may be referred to as having a low level of miRNA-210.

Alternatively, a favourable prognosis may be given when a sample from a cancer patient has less then the median level of miRNA-210, wherein the median level of miRNA-210 is determined from the miRNA-210 levels in at least five, six, seven, eight, nine, ten or more samples taken from different patients with the same cancer. The skilled person will appreciate that the median values typically refer to miRNA-210 relative expression levels normalised by a factor based on the geometric mean of three C/D box snoRNAs (SNORD43, SNORD44 and SNORD48) . For example, in one embodiment, the median level of miRNA-210 is approximately 1, such as from 0.5 to 1.5 (e.g. 1.01) for head and neck cancer and approximately 6, such as from 5 to 7 (e.g. 6.46) for breast cancer. Therefore, references to "low level of miRNA-210" herein described may refer to miRNA-210 values below these median levels. However, it will also be appreciated by the skilled person that other median values may be selected depending upon such factors as the normalisation method used and the control RNAs used.

Preferably an unfavourable prognosis is that a cancer patient has an about 40% or more chance that the cancer will recur or metastasise within about five, six, seven, eight, nine, ten or more years. Preferably, there is an about 50% chance that the cancer will recur or metastasise within about 10 years. In another embodiment, an unfavourable prognosis may be that the cancer patient has an about 30% or more chance that the cancer will be fatal within about five, six, seven, eight, nine, ten or more years. Preferably, in an unfavourable prognosis there is an about 40% or more chance that the cancer will be fatal within about 10 years. Preferably, an unfavourable prognosis may include a combination of the chance of recurrence/metastasis and the chance of survival.

Preferably an unfavourable prognosis will be given when a sample from a cancer patient displays miRNA-210 levels which are greater than a predetermined value. Preferably this predetermined level is relative to the level of miRNA-210 observed in a control sample. Such samples may be referred to as having a high level of miRNA-210.

Alternatively, an unfavourable prognosis may be given when a sample from a cancer patient has greater then the median level of miRNA-210 determined by calculating the miRNA-210 level in at least five, six, seven, eight, nine, ten or more samples taken from different patients with the same cancer. The skilled person will appreciate that the median values typically refer to miRNA-210 relative expression levels normalised by a factor based on the geometric mean of three C/D box snoRNAs (SNORD43, SNORD44 and SNORD48) . For example, in one embodiment, the median level of miRNA-210 is approximately 1 , such as from 0.5 to 1.5 (e.g. 1.01) for head and neck cancer and approximately 6, such as from 5 to 7 (e.g. 6.46) for breast cancer. Therefore, references to "high level of miRNA-210" herein described may refer to miRNA-210 values above these median levels. However, it will also be appreciated by the skilled person that other median values may be selected depending upon such factors as the normalisation method used and the control RNAs used. Preferably reference herein to a "predetermined level" of miRNA-210 refers to a level which is selected from the group comprising 3 or more, 2.9, 2.8, 2.7, 2.6, 2.5, 2.4, 2.3, 2.2, 2.1, 2, 1.5, 1 and less than 1 , fold the level of miRNA-210 observed in a control sample.

Preferably the control sample is a biopsy sample from a cancer free patient, preferably the biopsy is from the same tissue type as the cancer sample. Alternatively, the control sample may be derived from normal tissue near a cancer. Alternatively, the level of miRNA-210 in the cancer sample may be compared to the level of miRNA-210 observed within a particular set of cancers.

The predetermined level may depend on the type of cancer.

Preferably if the cancer is breast cancer, the predetermined level of miRNA-210 is about 2.8 fold the level of miRNA-210 observed in a control sample. Preferably, in this embodiment, a patient who provides a sample which has a level of miRNA-210 which is 2.8 fold or below 2.8 fold the level of miRNA-210 in a control sample will have a favourable prognosis; and a patient who provides a sample which has a level of miRNA-210 which is above 2.8 fold the level of miRNA-210 in a control sample will have an unfavourable prognosis. Preferably the sample from the breast cancer patient is a biopsy sample. Preferably the control sample is a biopsy from the breast tissue of an individual who does not have breast cancer.

Preferably if a cancer sample, such as a breast cancer sample, has less than 1.5 fold more miRNA-210 than a control sample then the chance of the patient surviving about 10 years or more is at least about 90%, and there is an about 70% or more chance that the cancer will not recur or metastasise in the next 5 to 10 years.

Where a patient is given a favourable prognosis a less invasive and/or a less aggressive treatment may be prescribed, particularly if the favourable prognosis indicates that the chance of metastasis is low. This could save a patient from unnecessary distress, costs, and unwanted side effects from treatment such as radiotherapy, hormonal therapies, chemotherapy and further surgical interventions. Where a patient is given an unfavourable prognosis, and the cancer may be considered aggressive, a more invasive and/or a more aggressive treatment may be prescribed. This treatment may include radiotherapy, hormonal therapies, chemotherapy and further surgical interventions. More frequent monitoring of disease progression with medical imaging may also be recommended.

The method of the invention may also be used in conjunction with an assessment of clinical characteristics and/or other molecular signatures. Alternatively, the method of the invention may be used alone to provide a prognosis.

Preferably the method is used to determine the prognosis of a mammalian cancer, more preferably a human cancer.

Preferably the method of the invention is carried out in vitro .

The level of miRNA-210 in a sample may be normalised by comparing it to the level of a control gene. The control gene may be RNU43 or any other gene expressed at similar levels in the test and control samples. The ratio of miRNA-210 in cancer and control samples may be normalised to the level of the control in both samples. According to another aspect, the invention provides a method for determining the prognosis of a mammalian cancer wherein the method comprises:

(a) determining the level of miRNA-210 in a sample from a mammalian cancer; and

(b) comparing the level of the miRNA-210 determined in (a) with a control sample;

(c) concluding that a high level of miRNA-210 indicates an unfavourable prognosis of the cancer, or that a low of miRNA-210 indicates a favourable prognosis of the cancer.

A high level of miRNA-210 may be as described previously with reference to the conditions which would lead to an unfavourable prognosis, and a low level of miRNA-210 may be as described previously with reference to the conditions which would lead to a favourable prognosis.

According to yet another aspect the present invention provides a method for determining the therapy to be given to a cancer patient comprising (a) determining the level of miRNA-210 in a sample from the cancer patient;

(b) comparing the level of miRNA-210 in the sample to the level in a control sample;

(c) using the results to decide on the therapy to be given to the patient.

According to a further aspect, the invention preferably provides a kit for performing any method of the invention, comprising one or more probes for detecting the level of miRNA-210 in a sample. The level of miRNA-210 may be detected using PCR, RNAase protection assays, immunological methods, in-situ hybridisation, miRNA microarrays or any other means for detecting low MW RNA.

If PCR is used to determine the miRNA-210 levels in a sample then primers having the sequence of SEQ ID No: 3 and SEQ ID NO: 4 may be used and provided in the kit.

SEQ ID No: 3 - reverse primer (5 '-3') : *CTCAACTGGTGTCGTGGAGTCGGCAATTCAGTTGAGTCAGCCGC*

SEQ ID No: 4 - forward primer (5 '-3 ') : *ACACTCCAGCTGGGCTGTGCGTGTGACAGC*

As a control, miRNA-21 levels in a sample may also be measured. Primers for miRNA-21 may also be included in the kit. Suitable primers for detecting miRNA-21 are identified below as SEQ ID No: 5 and SEQ ID No: 6:

SEQ ID No: 5 - Reverse primer (5 '-3 ') :

**CTCAACTGGTGTCGTGGAGTCGGCAATTCAGTTGAGTCAACATC

*

SEQ ID No: 6 - Forward primer (5 '-3 ') : **ACACTCCAGCTGGGTAGCTTATCAGACTGA*

The kit may also include Taqman probes for miRNA-21 (SEQ ID No: 7) and/or for miRNA-210 (SEQ ID No: 8) :

SEQ ID No: 7 taqman probe (5 '-3 ') : * (6-FAM)TTCAGTTGAGTCAACATC (MGB)*

SEQ ID No: 8 taqman probe (5 '-3') :

* (ό-FAM)TTCAGTTGAGTCAGCCGC (MGB)*

If the level of miRNA-210 is to be determined by an RNAase protection assay an antisense probe to miRNA-210 may also be included in the kit. The probe may be labelled, for example, fluorescently or radioactively. The kit may also include the RNAase enzyme.

The kit may also include reagents for use in performing the method, for example appropriate enzymes and/or buffers.

The kit preferably includes instructions pertaining to the use of the one or more probes. For examples, the instructions may teach the user how to prepare a sample for use with the kit. If the kit comprises primers for use in RT-PCR the instructions may also include the PCR conditions to be used.

The kit may include one or more controls. The kit may include a standard (such as a sample of miRNA-210 of known concentration or a sample of normal breast tissue) against which the results of the method should be compared in order to provide a prognosis.

The skilled man will appreciate that all preferable features referred to with respect to only some aspects of the invention can be applied to all aspects of the invention.

The invention will be further defined by reference to the following examples. It will be apparent to those skilled in the art that many modifications, both to the materials and methods may be practiced without departing from the scope of the invention.

Figure 1 - shows patient demographics including age, nodal status, tumor size, grade, histology and ER status for the 219 patients studied;

Figure 2 - shows hypoxia induced changes in expression of 377 miRNAs in MCF7 cells. Results from analysis comparing miRNAs microarrays data obtained from cells cultured under normoxia

(21% oxygen) and under hypoxia (1% oxygen) for 16 hours. FC = fold change; p = p-value; adjusted p (BH) = p-values adjusted by Benjamini & Hochberg method for multiple test correction. The microRNAs are ordered as up or down regulated according to adjusted p(BH) . Significantly up regulated microRNAs are highlighted by grey shading (multi-test corrected p < 0.05 using Limma) . No significantly down regulated microRNAs were found after multiple test correction of p-values. The prefixes "miR-" and "miRNA" are used interchangeably herein, "hsa" refers to homo sapiens; "mmu" refers to mus niusculus; "rno" refers to ratus norvegicus; "ambi" refers to Ambion-predicted sequences;

Figure 3A - demonstrates disease-specific and recurrence-free survival of breast cancer patients stratified according to miRNA- 210 levels. Expression levels are stratified by median value or in relation to the fold increase relative to miRNA-210 levels in a normal control breast tissue sample. This data was generated from the results obtained from 219 breast tumour samples. Follow-up is limited at 10 years. Log rank test statistics and significance are shown. Figure 3B - demonstrates disease-specific and recurrence- free survival of breast cancer patients stratified according to miRNA-210 levels where the expression levels are stratified by quartiles. The quartiles relate to the miRNA-210 levels compared to median levels of miRNA-210 and miRNA-210 levels in samples from cancer patients compared to miRNA-210 levels in a sample from a normal control patient. Follow-up is limited at 10 years.

Log rank test statistics and significance are shown. In both Figures 3 A and 3B the levels of miRNA-210 are normalised to the levels of the small nuclear RNA RNU43.

Figure 4 - uses Multivariate Cox Analysis of data to show that the risk of dying from breast cancer or of developing a recurrence of breast cancer is highly correlated with the level of expression of miRNA-210. This correlation is much stronger than the correlation of conventional clinical variables such as tumour size, spread to lymph nodes and estrogen receptor status, and is independent of them. The reduced model after Backwards Likelihood elimination (entry p = 0.05, elimination p = 0.10) are shown. Variables included in the analysis were: miRNA-210 as a continuous variable (expression levels are ranked between 0 and 1) , Age (pre- menopausal = <50, post-menopausal= > 50), Tumor Size (small = < 1.5 cm, large = >1.5 cm) , Positive nodes (Neg. = 0, Pos. = > 0) , ER status (Neg = ELISA < 5, Pos = ELISA≥ 5) , Grade (1-3) . Abbreviations used in Figure 4 include "pos" for positive node, "HR" for hazard ratio, and "CI" for confidence interval. In the recurrence free survival data an HR of 1.8 indicates patients are 1.8 fold more likely to have a recurrence of the cancer. In the overall survival data an HR of 11.38 indicates a patient is 11.38 fold more likely to die of cancer. In this figure the levels of has-miR-210 are normalized to the levels of the small nucleolar RNA RNU43. Figures 5A and 5B - show the time course and oxygen sensitivity of miRNA-210 hypoxic induction. Figure 5A shows the results of MCF7 cells exposed to 1% oxygen for 1 , 2, 4, 8, 16, 24 and 48 hours (3 replicates/ time point) . Figure 5B shows the results of MCF7 cells exposed for 16h to different oxygen percentages: 0.1 ,

1, 3 and 5% (3 replicates/ condition) . In both, the expression levels of miRNA-210, miRNA-93 and miRNA-21 were measured by Q-PCR. Results are expressed as mean fold difference (± SD) in miRNA expression between hypoxia and the parallel normoxia control samples using RNU43 as a reference. Significant differences compared to normoxia control were established using ANOVA followed by Scheffe's post hoc test (* p = 0.01, ** p≤O.001) .

Figures 6A, 6B and 6C - Figure 6A shows miRNA-210 and miRNA-21 expression in breast cancer samples with associated hypoxia score. The expression fold change for miRNA-210 (circles) and miRNA-21 (crosses) is shown as a function of the hypoxia signature score for 73 of the 219 samples. The hypoxia score was calculated as described in the materials and methods.

The expression of miRNA-210 was significantly correlated to the hypoxia signature score (Spearman p = 0.54, p₂.,_ailed < 0.001) , whilst the expression of miRNA-21 did not show significant correlation (Spearman p = 0.14, p₂-i_aii_ed = 0.24) . The distribution of the miRNA-210 (Figure 6B) and miRNA-21 (Figure 6C) fold change levels for samples with negative and positive hypoxia score is illustrated in Figures 6B and 6C. The hypoxia score was calculated using Affymetrix microarray expression data which were available for 73 cases from two previous studies Sotiriou C et al. J Natl Cancer Inst 2006; 98 (4) : 262-72 and Loi S et al J Clin Oncol

2007;25(10) :1239-46. Probe sets in these arrays which were present in a previously derived hypoxic signature were selected and a hypoxia score was calculated considering their median expression value as described therein (Winter SC et al. Cancer Res 2007;67(7) .-3441-9) . In these figures, the levels of miRNA-210 and miRNA-21 are normalised to the levels of the small nucleolar RNA

RNU43.

Figures 7A and 7B - demonstrate correlation among hsa-miR-210 relative expression data sets normalised to the levels of each C/D box snoRNA individually in breast cancer samples (Figure 7A) and

H&N cancer samples (Figure 7B) . Data was transformed on Iog2 scale and correlations were assessed using Spearman's rank tests.

Figures 8A and 8B - Figure 8A shows the overall survival and Figure 8B shows the disease-free survival for patients with breast cancer stratified according to hsa-miR-210 relative expression levels. Relative expression levels were stratified by median value and follow-up was limited at 10 years (log-rank test statistics and significance are given; df = degrees of freedom; sign = significance) . The lines marked with "b" and "a" represent the group of samples with hsa-miR-210 relative expression levels below and above the median value respectively. In these figures, hsa-miR-210 relative expression levels were normalised by a factor based on the geometric mean of three C/D box snoRNAs (SNORD43, SNORD44 and SNORD48) .

Figures 9 A, 9B and 9C compare the levels of three miRNAs, miRNA-210, miRNA-21 , miRNA-lOb, and the hypoxic score of tumour samples. Figures 1OA, 1OB, 1OC and 1OD show the relationship between tumour differentiation (poor, moderate and well) and the level of three microRNAs, miRNA-210, miR-21NA and miRNA-10b.

Figures HA, HB, HC and HD show the relationship between tumour size and microRNA expression.

Figures 12A, 12B, 12C and 12D show the relationship between the presence or absence of lymph node metastasis from tumour, and the level of microRNA in the primary tumour sample.

Figures 13A and 13B show the correlation between microRNA levels and CA9 membrane score.

Figures 14A, 14B, 14C and 14D show the relationship between immunohistochemistry for HIF-I and miRNA levels.

Figures 15 A and 15B show the correlation between microRNA levels and alcohol and smoking.

Materials and Methods

Breast cancer population 219 patients with early first primary breast cancer, treated in Oxford between 1989 and 1992 were studied. Ethical approval for analysis of samples and notes was obtained from the local Research Ethics Committee. Patients received surgery followed by adjuvant chemotherapy, adjuvant hormone therapy, both of these therapies, or no adjuvant treatment. Tamoxifen was used as endocrine therapy for pre- and postmenopausal patients. In patients with invasive tumours who were less than 50 years of age, adjuvant cyclophosphamide, methotrexate and 5- fluorouracil (CMF) was administered if the tumours were lymph node positive, or ER negative and/or more than 3cm in diameter. Patients more than 50 years of age with ER negative, lymph node positive tumours also received CMF. The dataset was complete for age, nodal status, definitive surgery, relapse and survival. The patient demographics are provided in Figure 1, including patient's age, tumor size, grade, histology, nodal status, oestrogen-receptor (ER) status and were collected from clinical and pathological records. Of those still alive, 3 patients that were lost to follow-up have survival data of 6.3, 9.1 and 9.3 years; all other patients alive have more than 10 years.

Breast tumour total RNA samples were analysed for miRNA-210 expression and the expression of the control small nucleolar RNA, RNU43, by Q-PCR and compared with the expression levels in 10 control samples taken from pools of normal breast tissue.

Methods for selection of samples based on hypoxia score

A hypoxic signature has been previously derived in head and neck tumours and assessed for its clinical relevance in a large breast cancer dataset (Loi S et al. , J Clin Oncol 2007;25(10) :1239-46; Winter, SC et al

Cancer Res. 2007 Apr l;67(7):3441-9) . Here, this signature was used to calculate an hypoxia score in 73 of the 219 breast cancer samples where gene expression data were available. This score was correlated to the expression of miRNA-210 and miRNA-21.

In brief, Affymetrix U133a and b gene expression microarrays have been previously performed on 73 of the 219 samples and the details of the processing and normalization have been described elsewhere (Sotiriou C et al. , J Natl Cancer Inst 2006; 98 (4) : 262-72; Loi S et al. , J Clin Oncol 2007;25(10) :1239-46; Winter, SC et al Cancer Res. 2007 Apr l ;67(7) :3441-9) . Expression values were logged (Iog2) and U133a and b arrays were mapped to the Affymetrix U133plus2 arrays where the hypoxic signature was originally derived. The hypoxia score was calculated by selecting the probe sets present in the hypoxic signature and by considering their median expression value as previously described (Loi S et al. , J Clin Oncol 2007;25(10) : 1239-46) .

The hypoxia score was obtained by analysis of genes previously shown to be regulated by hypoxia. The score was obtained by analysing the expression of these genes in the breast cancers and head and neck cancers and patients were ranked according to expression of these genes. The hypoxia score has previously been shown to be a significant prognostic factor for overall survival independent of clinicopathologic risk factors (Winter, SC et al Cancer Res. 2007 Apr l;67(7) :3441-9) .

A high positive hypoxia score reflects a higher than average of genes in the hypoxic signature and indicates a sample is hypoxic. A low negative hypoxia score reflects a lower than average of genes in the hypoxic signature and indicates a sample is normoxic.

Cell lines and culture conditions

The human cell lines studied were breast adenocarcinoma cell line MCF7, hepatoblastoma cell line Hep3B, uterine cervix adenocarcinoma cell line HeLa and renal cancer cell line RCC4 stably transfected with either an empty vector or VHL. All cell types were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum (Sigma- Aldrich, Dorset, UK) and 2mM L-glutamine. Cell lines were obtained from ECACC (European Collection of Cell Cultures, Porton, Salisbury, Wiltshire, SP4 OJG) or ATCC (American Type Culture Collection P.O. Box 1549, Manassas, VA 20108, USA)

Exposure of cell cultures to hypoxia was undertaken in an In Vivo2 Hypoxia Work Station (Ruskin Technologies, Kent, UK) in parallel with cells maintained in normoxic conditions (21% oxygen) . All experiments were performed in triplicate from independent cell cultures.

RNA extraction

RNA was extracted from cells using the miRVana miRNA Isolation Kit (Ambion, Austin, USA) . RNA was extracted and purified from liquid nitrogen frozen breast tumour samples or normal breast samples using Tri-reagent (Sigma- Aldrich, Dorset, UK) and ethanol precipitation. In both cases, RNA quality and abundance were determined after extraction using an Agilent 2100 Bioanalyser (Agilent Technologies, Santa Clara, CA, USA) and a Nanodrop ND-1000 spectrophotometer (Nanodrop Technologies) , respectively.

miRNA microarrays

Microarrays were generated using the miRVana Probe set version 1 (Ambion, Austin, USA) . Probes were spotted onto amine reactive slide substrates GE CodeLink (Amersham Biosciences, Bucks, UK) . Each microarray contained four spot replicates for each probe. Small RNA ( < 200 nucleotides) was purified from 100 μg of total RNA using FlashPAGE fractionation gels (Ambion, Austin, USA) and FlashPAGE Clean-up kit (Ambion, Austin, USA) was used for further purification and concentration of the product. The synthetic sequence Control 1 (Ambion, Austin, USA) was used for the control of the labelling process. It was spiked for each normoxic and hypoxic sample in a ratio 1 :10 respectively. 500 ng of small RNA from each sample was labelled using NCode miRNA labelling kit (Invitrogen, Paisley, UK) and two-colour hybridizations were performed. Twelve independent samples (6 hypoxic and 6 normoxic) were used in two series of 3 hybridizations including dye-swapping between both series of experiments. Hybridization signals were detected and quantified using a GenePix 4000B scanner and the GenePix Pro 4.0 respectively (both from Axon Instruments, Molecular Devices, Downingtown, PA, USA) . Median values of each spot were background subtracted by using a local background subtraction method (GenePix Pro 4.0) .

Analysis of microarray data

The Bioconductor (http://www.bioconductor.org/) implementation of Limma (Smyth GK. Limma: linear models for microarray data. In: Gentleman R, Dudoit S, Irizarry R, Huber W, editors. Bioinformatics and computational biology solutions using R and Bioconductor. New York: Springer; 2005. p.397-420) was used to analyse the data. Limma is an established general linear regression algorithm that uses empirical Bayes methods to gain power when a large number of predictors (i.e. probes) are present with a relatively small number of cases (i.e. arrays) . Loess normalisation was performed and average values were calculated for within array replicate spots. To estimate fold changes between hypoxic and normoxic conditions, and relative standard errors, a linear model was fit to each probe in the array; dye-swaps were indicated in the design matrix for this fit. Empirical Bayes smoothing was then applied to the standard errors and the Benjamini & Hochberg method was used to correct for multiple testing. In addition, data was filtered based on fold- changes (using 2-fold as a threshold) . Control 1 was detected according to the expected spike concentration. The miRNA nomenclature was then used according to the miRNA Registry (hsa-miRBase http://microrna.sanger.ac.uk/sequences/) at Sanger Institute, except for Ambion predicted miRNA sequences (designed as ambi-hsa-miR-#) . TaqMan miRNA assays

The following assays were used for quantification of specific target sequences by the quantitative-PCR technique:

Hsa-miR-210 assay (Applied Biosystems Cat. N: 4373089) .

Target sequence: CUGUGCGUGUG.AC AGCGGCUGA (SEQ ID No: 9;

Chr.ll : 558112-558133, corresponding to the mature sequence of hsa- miR-210) .

RNU43 (Applied Biosystems Cat. N: 4373375) .

Target sequence: GAACUU AUUGACGGGCGGACAGAAACUGUGUG

CUGAUUGUCACGUUCUGAUU (SEQ ID NO: 10; Chr.22: 38045003-

38045054, corresponding to the small nucleolar RNA C/D box 43 or SNORD43) .

RNU44 (Applied Biosystems Cat. N: 4373384) .

Target sequence: CCUGGAUGAUG AUAGC AAAUGCUGACUG AAC AU GAAGGUCUUAAUUAGCUCUAACUGACU (SEQ ID NO: 11; Chr. l: 172101729-172101789, corresponding to the small nucleolar RNA C/D box 44 or SNORD44) .

RNU48 (Applied Biosystems Cat. N: 4373383) .

Target sequence: GAUGACCCCAGGUAACUCUGAGUGUGUCGCUG AUGCCAUCACCGCAGCGCUCUGACC (SEQ ID NO: 12; Chr.6: 31911019-31911082, corresponding to the small nucleolar RNA C/D box 48 or SNORD48) .

Hsa-miR-21 assay (Applied Biosystems Cat. N: 4373090) . Target sequence: UAGCUUAUCAGACUGAUGUUGA (SEQ ID No: 13; Ch.17: 55273416-55273437, corresponding to the mature sequence of has-miR-21) .

Real-time reverse transcription PCR. miRNA expression was assessed by real-time PCR according to the TaqMan MicroRNA Assay protocol (Applied Biosystems, Foster City, CA, USA) . cDNA was synthesized from 5 ng of total RNA using TaqMan microRNA-specific primers and the TaqMan MicroRNA Reverse Transcription Kit (both from Applied Biosystems, Foster City, CA, USA) . Real-time PCR was performed using iCycler IQ Detection System (BioRad, Hercules, CA, USA) . Each PCR reaction was done in triplicate and contained 1.33/il RT product, 1 x TaqMan Universal PCR master mix No AmpErase UNG and 1 μl of primers and probe mix of the TaqMan MicroRNA Assay (both products from Applied Biosystems, Foster City, CA, USA) . The 20μl reactions were incubated in a 96-well optical plate at 95°C for 10 min, followed by 40 cycles of 95°C for 15 sec and 60°C for 60 sec. Fold changes in miRNA expression between treatments and controls were determined by the 2"^ΔΛct method (Livak and Schmittgen (2001) Methods 25:402-408), normalizing the results to small nuclear RNA RNU43 expression level. Significant differences were established using ANOVA followed by Scheffe's post hoc test (SPSS version 7, Chicago, IL, USA) . Results were considered significant when p<0.05.

Primers of SEQ ID No: 3 and 4, or hsa-miR-210 assay (Applied Biosystems Cat. N: 4373089) , were used in RT-PCR to amplify miRNA-210 and primers of SEQ ID No: 5 and 6, or has-miR-21 assay (Applied Biosystems Cat. N: 4373090) , were used to amplify miRNA-21.

Statistical methods Correlation of miRNA-210 with clinical and pathological variables was assessed using Pearson and Spearman's rank tests. Univariate survival analysis was performed by applying the Log rank test to miRNA-210 expression levels stratified by median value and in 4 quartiles. Multivariate Cox survival analyses were also performed where the miRNAs were introduced as a continuous variable together with other relevant clinical factors. Both disease-specific survival and recurrence- free survival were considered as outcomes. In this case, the fractional rank of the expression was considered, that is the patients were ranked using the miRNA expression levels and the ranks were normalised between 0 and 1. Statistical analyses were performed using SPSS and R (http : //www . r-proj ect . org) .

Results

miRNA expression profile in hypoxic MCF7

The breast cancer line MCF7 was selected for study of miRNA regulation by hypoxia because the expression of the components of the HIF system had previously been characterised in this cell line along with an extensive study of the gene expression profile in response to hypoxia, a prolyl hydroxylase inhibitor dimethyloxalylglycine, and HIF-α isoform manipulations (Elvidge GP et al (2006) J Biol Chem; 281:15215-26) . MCF7 breast cancer cells were grown under conditions of either normoxia (21% oxygen) or hypoxia (1% oxygen) for 16 hours. Among the 377 miRNAs included in the microarray, 4 miRNAs were significantly up-regulated between hypoxic and normoxic conditions (multiple-test corrected p < 0.05 using Limma, see Figure 2) . Nevertheless, only three miRNAs (hsa-miR-210, ambi-miR-7105 and mmu-miR-322-3p) showed greater than 2-fold induction in response to hypoxia (see Figure 2) . No significantly down -regulated miRNAs were found in the analysis after multiple test correction of p- values. Confirmation of the hypoxic induction of miRNA-210 in MCF7 cells by Q-PCR.

To demonstrate the hypoxic regulation of miRNA-210 seen in the array studies, eight independent pairs of RNA samples obtained from normoxic and hypoxic MCF7 cells were analysed with Q-PCR. A substantial and significant induction of miRNA-210 expression was observed in hypoxic cells (4.11 fold + 0.92; p < 0.001) . To exclude the possibility that hypoxia produced an artef actual change in miRNA recovery, the expression of miRNA-93 was also studied as a control miRNA that was not regulated by hypoxia in the microarray assays. miRNA-93 did not show significant regulation by hypoxia (0.92 ± 0.22) when assayed by Q-PCR. The hypoxic induction of miRNA-21, a miRNA whose expression has been reported to be enhanced in a wide range of cancers was also studied. No significant regulation of miRNA-21 by hypoxia was observed when assayed either by microarray or by Q-PCR.

In order to explore further the regulation of miRNA-210 by hypoxia, a time course of miRNA-210 induction by hypoxia was studied. MCF7 cells were cultured in hypoxia (1% oxygen) for 1 , 2, 4, 8, 16, 32 and 48 hours. The expression of miRNA-210, miRNA-93 and miRNA-21, were analysed by Q-PCR. Induction of miR-210 by hypoxia was discernable at 8 hours and showed a progressive increase in expression, becoming significant at 16 hours (3.08 + 0.34, p < 0.001) and being maximal at the latest (48 hour) timepoint (4.78 ± 0.32, p < 0.001) , whilst expression of the miRNA-93 and miRNA-21 were again unaffected by exposure to hypoxia (Figure 5A) . To study the oxygen dependence of the regulation of miRNA-210, further experiments were undertaken in cells exposed to a range of differing oxygen tensions (MCF7 cells were cultured for 16h in 0.1%, 1%, 3%, 5% and 21% oxygen) . A significant induction of miRNA-210 was seen at 0.1% oxygen (2.56 + 0.53, p = 0.011) with more modest regulation at 3 and 5% hypoxia, but the greatest induction occurred with exposure to 1% hypoxia (3.08 + 0.33, p = 0.001) (see Figure 5) . Again the expression of miRNA-21 and miRNA-93 were unaffected by hypoxia.

miRNA-210 as a hypoxic marker in breast cancer

Recent reports have suggested enhanced levels of miR-210 expression in some tumours when assayed by microarray (Iorio MV et al. , Cancer Res 2005;65(16) :7065-70; Volinia S et al. , Proc Natl Acad Sci U S A 2006;103(7) :2257-61) . Given the importance of hypoxia and the HIF system in tumour biology, and the increasing recognition of alteration of miRNA expression in tumours, this study sought to confirm that the level of miRNA-210 in breast carcinoma was correlated to hypoxia. To do this Q-PCR assays were used. 219 breast tumour total RNA samples were analysed for miRNA-210 and miRNA-21 expression as well as 10 control RNA samples obtained from pools of normal breast tissue. The extent to which miRNA-210 levels correlated with other hypoxically regulated genes was also examined. A subset of 73 of the breast cancer samples in which each sample had a hypoxia score determined was studied. The hypoxic score was estimated as described in the methods and it was used to rank the samples from low to high hypoxic tumours. The subset examined were comparable to the whole population of 219 cases in terms of demographics (see materials and methods) ; the only significant difference was a higher percentage of node negative patients (Chi-square test p = 0.002) . Importantly, the expression of miRNA-210 was significantly correlated to the hypoxia signature score (Spearman p = 0.54, p_2-lailcd < 0.001) , whilst the expression of miRNA-21 did not show a significant correlation (Spearman p = 0.14, p₂-_wiii-_d = 0.24) with the hypoxia score (Figure 6 which shows that miR-210 expression is highly associated with hypoxia score whilst a different microRNA miR-21 is not) . These results support the hypothesis that miRNA-210 expression is regulated by hypoxia in breast cancers and can be used as a marker of breast cancer.

miRNA-210 is a prognostic marker in breast cancer. Levels of miRNA-210 correlate with tumour aggressiveness and patient recurrence-free and disease-related survival, and thus can be used to provide a prognosis. To demonstrate this, the relationship between the expression of miRNA-210 in tumour specimens from 219 patients was compared to the outcome of the cancer. The outcome was determined by long term follow up of early breast cancer patients (Figure 3 A and 3B) .

miRNA-210 levels showed a highly significant inverse correlation with recurrence-free and overall survival both in univariate (Figure 3) and multivariate analysis (Figure 4) ; both when considered as a continuous variable and as a binary variable divided by median value (Figure 3A) , quartiles (Figure 3B) or by fold change (Figures 3A and 3B) .

Figures 3 A and 3B chart disease-related survival, which is the interval from the date of diagnosis to the date of death from breast cancer or to the last follow-up date and also the proportion of patients that are free of recurrence of breast cancer over the ten years following initial diagnosis divided according to high or low expression of miRNA-210 (above and below median - Figure 3A) and divided into quartiles of level of miRNA-210 expression (Figure 3B) . The data shows that patients with the highest level of expression of miRNA-210 (greater than median or in the top two quartiles) are more likely to develop a recurrence of breast cancer and are more likely to die from breast cancer than those patients with lower levels of expression. Conversely, patients with low miRNA-210 levels are more likely to survive and less likely to develop breast cancer recurrence. The log Rank Chi-squared statistical analysis shows that these differences are highly significant. For example in Figure 3 A patients having miRNA-210 fold changes between 2.80 and 67.81 (the highest obtained in this study) compared to normal/control levels have a ~ 50% overall survival after 10 years of follow-up, whereas patients with miRNA-210 fold-change ranging from 0.07 to 2.73 represent ~ 80% of overall survival.

The correlation of miRNA-210 levels with mortality in breast cancer and head and neck cancers is both surprising and striking.

Increased levels of miRNA-210 and miRNA-21 have been detected in several solid tumours including breast cancer and expression levels of miRNA-21 has been shown to correlate with tumour stage. This analysis however shows that the levels of expression of miRNA-21 only showed significant correlation with disease specific and recurrence free survival in univariate analysis, whilst miRNA-210 expression levels were an independent predictor of recurrence free and overall survival, and thus a surprising prognostic marker.

Many cancers are characterised by areas of hypoxia and in one aspect the present invention has identified miRNAs regulated by hypoxia in breast cancer cells and investigated induction by the HIF-VHL system. Of these miRNAs regulated by hypoxia only miRNA-210 has been demonstrated to have a prognostic value.

Microarrays were used to determine changes in miRNA expression under hypoxia and validations were performed by quantitative-PCR. Using siRNA against HIF- lα and HIF-2α as well as RCC4 cells transfected with VHL miRNA-210, expression or miRNA-210 has been indicated to be hypoxically induced in a HIF- lα and VHL dependent fashion. This induction (4-fold, p < 0.001) was also detected in a range of other cancer cells. To conclude, this data further demonstrates that miRNA-210 can be used as an independent prognostic predictor in cancer, and in breast cancer in particular.

The data presented also shows that expression analysis of a single miRNA, miRNA-210, is comparable to the analysis of 97 transcripts used to calculate the hypoxic score, in terms of survival prediction. Thus miRNA-210 offers a simple and cost effective prognostic marker for cancer.

Analysis of miRNA-210 levels and clinicopathological variables The correlation between miRNA-210 expression levels and clinicopathological variables was considered. The results presented in Figure 4 show an association between miRNA-210 levels and size of tumour. However, no significant correlation between miRNA-210 levels and Estrogen Receptor status (ER) was observed. Similarly, no significant interaction effect was observed between ER and miRNA-210 levels or between nodal status and miRNA-210 levels for disease-specific or recurrence-free survival. The miRNA-210 levels are essentially providing an independent prediction of prognosis.

More detailed analysis of pathological variables and miRNA levels was undertaken, which demonstrated miRNA-210 to provide an independent prediction of prognosis.

Figure 9A, 9B and 9C compare the levels of three miRNAs, miRNA-210, miRNA-21 , miRNA-lOb, and the hypoxic score of tumour samples, and demonstrate a strong correlation between miR-210 levels and the hypoxia score. The hypoxia score is calculated from the median level of the distribution of the level of the up-regulated genes in the hypoxia metagene (99 genes strongly clustering with known hypoxia genes) . The score is then converted to fractional rank where patients are ranked on their score, and normalised between 0 and 1 (the X-axis) . The Y-axis shows microRNA values as determined by PCR. Correlation of the hypoxic score with two non-hypoxia-responsive microRNAs (rm^'RNA-21 and miRNA-10b) is provided for comparison.

Figures 9B and 9C show results for Spearman's rho (non-parametric measure of correlation) between the microRNA and hypoxia score. These results show that miRNA-210 has a highly significant correlation coefficient with the hypoxia score.

Figures 1OA, 1OB and 1OC show the relationship between tumour differentiation (poor, moderate and well) and the level of three microRNAs, miRNA-210, miR-21NA and miRNA-lOb. (For these and subsequent graphs, the values for microRNA are normalised to 3 small nucleolar RNA and normal tissue.) For all three microRNAs studied there is a positive correlation between increasing differentiation and increasing level of the microRNA, however by Kruskal-Wallis test (a non- parametric version of the ANOVA) (Figure 10D) , this is only statistically significant for miRNA-lOb.

Figures HA, HB and HC show the relationship between tumour size and microRNA expression. There is no significant change across these groups (Figure HD) .

Figures 12A, 12B and 12C show the relationship between the presence or absence of lymph node metastasis from tumour, and the level of microRNA in the primary tumour sample. There is no statistically significant relationship (Figure 12D) , however the trend for miRNA-210 is a positive correlation with lymph node metastasis while for miRNA-21 and miRNA-lOb the reverse applies. Figures 13 A and 13B show the correlation between the microRNA and

CA9 membrane score. CA9 membrane is a protein known to be induced by hypoxia, and this data shows that miRNA-210 expression correlates with a known hypoxia marker. There is a statistically significant correlation between miRNA-210 and both CA9 membrane and the sum of the scores for CA9 membrane and cytoplasm, as well as between miRNA-21 and CA9 membrane score. There is also a statistically significant relationship between several microRNA: between miRNA-21 and miRNA-210; and between miRNA-10b and miRNA-21.

Figures 14A, 14B, 14C and 14D show there is a significant relationship between immunohistochemistry for HIF-I, and both miRNA-210 and miRNA-lOb.

Figures 15A and 15B show the non-parametric correlation between the microRNA and alcohol and smoking. The results show that miRNA-210 levels are not related to smoking or alcohol use.

Normalisation of miRNA-210 relative expression levels by using several control genes.

The aim of this investigation was to increase the accuracy in the measurement of hsa-miR-210 relative expression levels in a given amount of total RNA by normalising according to the expression levels of three control genes and to relate this to prognosis. This would allow samples to be grouped as having expression levels above or below already established median values based on a representative set of 219 breast cancer and a set of 58 head and neck cancer samples.

The levels of hsa-miR-210 were measured by quantitative-PCR (Q-PCR) in two collections of cancer samples (breast and head and neck (H&N)) and their respective controls. In order to normalise these results, the levels of three C/D box small nucleolar RNAs (C/D box snoRNAs) were also tested in all samples.

C/D box snoRNAs (average length ~ 62 bases) are components of the C/D box small nucleolar ribonucleoparticules (C/D box snoRNPs) which catalyse the 2'0-ribose methylation of target RNAs. They are commonly used as normalisation controls in the quantification of miRNA relative expression levels by Q-PCR.

Two cancer sample collections with long term follow-up were studied. The breast cancer collection was composed of 219 tumours and 10 control samples. The H&N cancer collection included 58 tumours and 11 control samples. Ethical approval for analysis of samples and notes was obtained from the local Research Ethics Committee.

cDNA synthesis and cycling conditions were as recommended by the manufacturer (Applied Biosystems) . The cycle threshold (Ct) values were used to calculate fold changes in hsa-miR-210 expression between cancer samples and controls by the 2"^ΛΔci method (Livak KJ and Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2 (-Delta Delta C(T)) Method. Methods 2001 ;25:402-8) , normalising the results to either the levels of each C/D box snoRNA individually or to a factor based on the geometric mean of the levels of all three snoRNAs (according to the method described by Vandesompele J, De Preter K, Pattyn F, Poppe B, Van Roy N, De Paepe A, Speleman F. Accurate normalisation of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome biology 2002) . A gene-stability factor (developed by Vandesompele et al. (supra)) was calculated for each C/D box snoRNA in each cancer collection by using the corresponding Ct values. This factor is helpful to evaluate the expression stability of control genes on the basis of non-normalised expression levels .

Correlation of hsa-miR-210 relative expression levels normalised to each C/D box snoRNA was assessed using Spearman's rank tests after Iog2 transformation of the raw data.

Univariate survival analysis was performed by applying the Log rank test to hsa-miR-210 relative expression levels stratified by median value and in 4 quartiles. Both disease-specific overall survival and disease-free survival were considered as outcomes.

The results obtained demonstrated that the gene-stability factor was similar between all three C/D Box snoRNAs in each of the cancer sample collections. When comparing values for each snoRNA between collections, gene-stability factors were lower in H&N cancer than in breast cancer indicating that their expression is more stable in H&N than in breast cancer (see Table 1) .

SNORD43 SNORD44 SNORD48 Breast cancer 1.319 1.228 1.267

H&N cancer 0.757 0.738 0.636

TABLE 1: Gene-stability measure M of SNORD43, SNORD44 and SNORD48 in breast and H&N cancer samples. The gene-stability measure M is defined as the average pairwise variation of a particular gene with all other control genes (Vandesompele et al. (supra)) . Low M values correspond to more stable expressed genes and vice versa. The three sets of hsa-miR-210 relative expression data obtained after normalisation to each individual C/D box snoRNA expression showed a high degree of correlation in both breast and H&N cancer samples collections (see Figures 7A and 7B) .

The hsa-miR-210 relative expression data set normalised by the factor based on the geometric mean of the expression levels of all three controls was used to perform univariate survival analysis. hsa-miR-210 relative expression levels showed a significant inverse correlation with overall survival and disease-free survival when stratified by median value or quartiles in either breast cancer or H&N cancer (see Figure 8A and 8B) .

This normalisation analysis demonstrated that expression of SNORD43, SNORD44 and SNORD48 is comparable in both breast and H&N cancer sample collections. This is based on the similar gene-stability factor of all three snoRNA in each cancer collection and the high correlation between hsa-miR-210 expression levels normalised by the levels of each snoRNA.

This result does not seem to be affected by the difference in gene-stability expression observed for SNORD43, SNORD44 and SNORD48 between breast and H&N cancers.

These results demonstrate that patients with a hsa-miR-210 relative value above the median value (being 6.46 for breast cancer samples and 1.01 for H&N cancer samples) have poorer outcome than patients with an hsa- miR-210 relative value below the median.

Claims

1. A method for determining the prognosis of a cancer wherein the method comprises: (a) determining the level of miRNA-210 in a sample from an individual; and

(b) using the level of miRNA-210 in the sample to determine the prognosis of the cancer.

2. A method according to claim 1 wherein the sample is obtained from an individual who has been diagnosed with cancer.

3. A method according to claim 1 or claim 2 wherein the sample comprises cancer cells.

4. A method according to claim 1 , 2 or 3 wherein the cancer is selected from the group comprising cancer of the bone, breast, respiratory tract, brain, reproductive organs, digestive tract, urinary tract, bladder, eye, liver, skin, head, neck, thyroid, parathyroid, kidney, pancreas, blood, ovary, colon, prostate, and metastatic forms thereof.

5. A method according to any preceding claim wherein the prognosis is determined by comparing the level of miRNA-210 in a sample obtained from an individual with cancer to a control level of miRNA-210.

6. A method according to claim 5 wherein the control level of miRNA-210 is the level of miRNA-210 in one or more control samples wherein a control sample is selected from the group comprising (i) a sample of tissue taken from the same tissue type as used for the cancer patient sample, but from an individual who does not have cancer in that tissue; (ii) a sample from normal tissue near a cancer; and (iii) a sample from a predetermined set of cancers.

7. A method according to any preceding claim wherein a favourable prognosis is that a cancer patient has at least an about 70% or more chance that the cancer will not recur or metastasise within about five years from taking the sample and/or that the cancer patient has an 80% or more chance that the cancer will not be fatal to the individual within about five years from taking the sample.

8. A method according to claim 7 wherein a favourable prognosis can be given when a sample from a cancer patient displays miRNA-210 levels which are below a predetermined level.

9. A method according to claim 7 wherein a favourable prognosis can be given when a sample from a cancer patient has less then the median level of miRNA-210, wherein the median level of miRNA-210 is determined from the miRNA-210 levels in at least five samples taken from different patients with the same cancer.

10. A method according to any of claims 1 to 6 wherein an unfavourable prognosis is that a cancer patient has an about 40% or more chance that the cancer will recur or metastasise within about five years from the taking of the sample and/or that the cancer patient has an about 30% or more chance that the cancer will be fatal within about five years from the taking of the sample.

11. A method according to claim 10 wherein an unfavourable prognosis can be given when a sample from a cancer patient displays miRNA-210 levels which are above a predetermined the level of miRNA-210.

12. A method according to claim 10 wherein an unfavourable prognosis can be given when a sample from a cancer patient has greater then the median level of miRNA-210 determined by calculating the miRNA-210 level in at least five samples taken from different patients with the same cancer.

13. A method according to claims 9 or 12 wherein the median level of miRNA-210 is approximately 0.5 to 1.5 for head and neck cancer and approximately 5 to 7 for breast cancer.

14. A method according to claim 8 or 11 wherein the predetermined level of miRNA-210 refers to a level which is selected from the group comprising 3 or more, 2.9, 2.8, 2.7, 2.6, 2.5, 2.4, 2.3, 2.2, 2.1, 2, 1.5, 1 and less than 1, fold the level of miRNA-210 observed in a control sample.

15. A method according to claim 14 wherein the predetermined level is 2.8 fold the level of miRNA-210 observed in a control sample.

16. A method according to claim 15 wherein the cancer is breast cancer.

17. A method according to claim 16 wherein the sample from the individual with breast cancer is a biopsy of breast tissue, and the control sample is a biopsy of breast tissue from an individual who does not have breast cancer.

18. A method according to any preceding claim used in conjunction with an assessment of clinical characteristics and/or other molecular signatures.

19. A method according to any preceding claim for use in the prognosis of a mammalian cancer.

20. A method according to any preceding claim wherein the level of miRNA-210 in a sample is normalised by comparison with the level of a control gene.

21. A method for determining the prognosis of a mammalian cancer wherein the method comprises: (a) determining the level of miRNA-210 in a sample from a mammalian cancer; and

(b) comparing the level of the miRNA-210 determined in (a) with the level of miRNA-210 in a control sample;

(c) concluding that a high level of miRNA-210 in the sample from a mammalian cancer indicates an unfavourable prognosis for the cancer, or that a low level of miRNA-210 in the sample from a mammalian cancer indicates a favourable prognosis for the cancer.

22. A method according to claim 21 wherein a high level of miRNA-210 is as defined in claim 11 or claim 12.

23. A method according to claim 21 wherein a low level of miRNA-210 is as defined in claim 8 or claim 9.

24. A method for determining the therapy to be given to a cancer patient comprising

(a) determining the level of miRNA-210 in a sample from the cancer patient;

(b) comparing the level of miRNA-210 in the sample to the level miRNA-210 in a control sample;

(c) using the results to decide on the therapy to be given to the patient.

25. A kit for performing the method of any preceding claim, comprising a one or more probes for detecting the level of miRNA-210 in a sample.

26. A kit or method according to any preceding claim wherein the level of miRNA-210 may be detected using PCR, an RNAase protection assay, an immunological method, in-situ hybridisation, an miRNA microarray or any other means for detecting low MW RNA.

27. A kit or method according to claim 26 wherein PCR and primers having the sequence of SEQ ID No: 3 and SEQ ID NO: 4 are used to determine the miRNA-210 level in a sample.

28. A kit according to claim 25, 26 or 27 also for measuring the miRNA-21 levels in a sample.

29. A kit according to any of claims 25 to 28 also comprising reagents for use in detecting the level of miRNA-210 in a sample.

30. A kit according to any of claims 25 to 29 comprising instructions to the use of the kit and the probes to determine the level of miRNA-210 in a sample.

31. A kit according to any of claims 25 to 30 comprising one or more controls and/or one or more standards.