WO2008156827A2 - Molecular grading methods for ductal carcinoma in situ - Google Patents

Abstract

The instant invention provides methods for determining the molecular grade of ductal carcinoma in situ in a subject through molecular profiling. Included in the invention are methods of prognosis and diagnosis of, and methods of determining therapeutic compounds. Also included in the invention are kits.

Description

MOLECULAR GRADING METHODS

The present application claims the benefit of U.S. provisional application no. 60/936,526 filed June 20, 2007, which is incorporated herein by reference in its entirety.

GOVERNMENT SUPPORT

Research supporting this application was carried out by the United States of America as represented by the Secretary, Department of Health and Human Services.

BACKGROUND OF THE INVENTION

Breast cancer represents the most common form of cancer among women. Each year, more than 180,000 and 1 million women in the U.S. and worldwide, respectively, are diagnosed with breast cancer. Breast cancer is the leading cause of death for women between ages 50-55, and is the most common non-preventable malignancy in women in the Western Hemisphere. It is estimated that 2,167,000 women in the United States are currently living with breast cancer (National Cancer Institute, Surveillance Epidemiology and End Results (NCI SEER) program, Cancer Statistics Review (CSR), on the world wide web at seer.ims.nci.nih.gov/Publications/CSR1973 (1998)). Among women in the United States, breast cancer is the second most common form of cancer, after skin cancer, and ranks second only to lung cancer among causes of cancer deaths in women. Nearly 86% of women who are diagnosed with breast cancer are likely to still be alive five years later, though 24% of them will die of breast cancer after 10 years, and nearly half (47%) will die of breast cancer after 20 years. Moreover, based on cancer rates from 1995 through 1997, a report from the National Cancer Institute (NCI) estimates that about 1 in 8 women in the United States (approximately 12.8 percent) will develop breast cancer during her lifetime (NCI's Surveillance, Epidemiology, and End Results Program (SEER) publication SEER Cancer Statistics Review 1973- 1997). In the current era of breast cancer diagnosis, ductal carcinoma confined to the duct and lobular structures of the breast, ductal carcinoma in situ, or DCIS, has become an important clinical entity. The frequency of the diagnosis of DCIS has increased markedly in the United States since the widespread use of screening mammography. In 1998, DCIS accounted for about 18% of all newly diagnosed invasive plus noninvasive breast tumors in the United States. The dramatic increase in incidence has given emphasis to the challenges of managing this important clinical entity. DCIS has become a more common clinical problem and the ability to appropriately diagnose and manage DCIS has become a factor determining the effectiveness of screening programs. Unlike invasive breast cancer, there is no established histopathologic grading system for DCIS in widespread use, nor are there biological markers of prognosis to guide clinical management. The aim of this study was to use molecular profiling to identify robust and clinically applicable indicators of DCIS malignant potential. There remains a need in the field for identification of clinically applicable indicators of DCIS malignant potential. Accordingly, the instant invention provides methods for determining the molecular grade of a cancer, in particular of DCIS.

SUMMARY OF THE INVENTION

The instant invention provides prognostic and diagnostic methods for determining the molecular grade of a cancer.

Accordingly, in one aspect, the invention provides a method for determining the molecular grade of a cancer comprising identifying genes differentially expressed in a cancer sample and a control sample by determining gene expression and gene copy number by molecular profiling, and classifying the cancerous tissue by one or more of gene expression or gene copy number, and thereby determining the molecular grade of the cancer.

In another aspect, the invention provides a method for determining the molecular grade of a cancer in a subject comprising identifying genes differentially expressed in a cancer sample and a control sample by determining gene expression and gene copy number by molecular profiling and classifying the cancerous tissue by one or more of gene expression or gene copy number and thereby determining the molecular grade of a cancer in a subject.

In one embodiment, the above-mentioned methods further comprise obtaining one or more samples of a cancer tissue.

In another embodiment, determining the molecular grade of a cancer further comprises classifying the cancer as low molecular grade or high molecular grade.

In a further embodiment, classifying the cancer by one or more of gene expression or gene copy number comprises calculating a gene expression grade index. In a related embodiment, the gene expression grade index is calculated using the gene expression pattern of the cancer sample versus the gene expression of the control sample.

In a further embodiment, a high gene expression grade index corresponds to a high molecular grade. In still a further embodiment, a low gene expression grade index corresponds to a low molecular grade. In one particular embodiment, the high molecular grade correlates with a low level of survival compared to a control. In another particular embodiment, the low molecular grade correlates with a high level of survival compared to a control.

In one further embodiment, high molecular grade correlates with one or more of a high level of DNA aberrations or a high level of regions of high level amplification compared to a control.

In another further embodiment, low molecular grade correlates with one or more of a low level of DNA aberrations or a low level of regions of high level amplification compared to a control. In one embodiment of the above-mentioned aspects, molecular grade is used to determine the prognosis of a patient. In another embodiment of the above- mentioned aspects, gene expression and gene copy number are determined by molecular analysis.

In another embodiment, the molecular analysis is selected from the group consisting of: Polymerase Chain Reaction (PCR), quantitative real-time reverse trascriptase PCR, targeted differential display, serial analysis of gene expression (SAGE), microarray analysis, multi-analyte profiling beads, and array-based comparative genomic hybridization (CGH).

In one embodiment of the above-mentioned aspects, the samples are selected from invasive cancer, intraductal carcinoma, proliferative disease without atypia, atypical ductal hyperplasia and benign tissue.

In a further embodiment of the above-mentioned aspects, the cancer is breast cancer. In one embodiment, the cancer is an invasive breast cancer. In another embodiment, the breast cancer is ductal carcinoma in situ (DCIS). In another embodiment, the DCIS samples are taken from a subject with invasive breast cancer. In one embodiment, CGH generates a genomic profile for classifying the cancer into a low molecular grade or a high molecular grade. In another embodiment, the molecular analysis generates a gene amplification index (AI) that is correlated with molecular grade. In one further embodiment, the gene amplification index is measured by the number of high level regions of DNA amplification over a threshold. In a further embodiment, greater than two high-level amplifications is correlated with a high molecular grade. In another embodiment, genes differentially expressed in cancer and control samples are selected from one or more of the genes listed in Figures 3A-B.

In one embodiment of the above-mentioned aspects, the differentially expressed genes comprise one or more of genes involved in cell proliferation. In one embodiment, the genes related to cell proliferation are selected from one or more of the genes listed in Figures 3A-B.

In another aspect, the invention features a method of determining the molecular grade of a cancer comprising profiling molecularly to determine one or more of gene expression or gene copy number in a test sample, comparing a test expression profile to a control expression profile, and then determining a molecular grade that corresponds to gene expression and gene copy number, wherein the molecular grade of the cancer is determined.

In one embodiment, the method further comprises the step of providing a test sample of cancerous tissue. In another aspect, the invention features a diagnostic method to determine the molecular grade of cancer using nuclear grade and Ki67 score comprising determining a nuclear grade and a Ki67 score in a carcinoma sample, and determining a molecular grade of cancer using nuclear grade and a Ki67 score, and where the molecular grade of cancer is determined by using nuclear grade and Ki67 score.

In one embodiment, the method further comprises providing a sample of breast cancer tissue from a subject. In a further embodiment, the sample is primary invasive breast cancer tissue.

In a further embodiment of the above-mentioned aspect, the method further comprises the steps of contacting the sample with a candidate compound and determining the cell proliferation profile of the sample.

In another aspect, the invention features a method of identifying a therapeutic compound, the method comprising determining the molecular grade of cancer according to the method as described in the above-mentioned aspect, and then contacting the sample with a candidate compound, and determining the cell proliferation profile of the sample; wherein a therapeutic compound is identified.

In another aspect, the invention features a method of predicting the prognosis of a breast cancer patient comprising determining a nuclear grade and a Ki67 score in a primary invasive breast carcinoma sample and a control sample and comparing the nuclear grade and Ki67 score from the subject sample to the control sample, wherein the molecular grade and the Ki67 score higher in the carcinoma sample than in the control sample is an indication that the prognosis of the test patient is poor.

In one embodiment, the method further comprises providing a sample of primary invasive breast cancer tissue from a subject. In a further embodiment, the breast cancer is intraductal carcinoma. In still a further embodiment, the intraductal carcinoma is DCIS.

In another aspect, the invention features a gene expression profile associated with high molecular grade DCIS comprising one or more of the genes listed in Figures 3A-B. In another aspect, the invention features a single stranded nucleic acid probe comprising: (a) the nucleotide sequence of a tag selected from those listed in Figures 3A-B; or (b) the complement of the nucleotide sequence.

In another aspect, the invention features a microarray comprising a substrate having one or more addresses, wherein each address comprises a capture probe comprising a nucleic acid sequence comprising a tag nucleotide sequence selected from the sequences listed in Figures 3A-B.

In another aspect, the invention features a kit comprising at least 10 probes, each probe comprising a nucleic acid sequence comprising a tag nucleotide sequence selected from those listed in Figures 3A-B.

In another aspect, the invention features a method of identifying the molecular grade of a DCIS sample comprising determining the molecular grade of a sample and comparing the molecular grade of a sample with a reference profile; and thereby identifying the grade of the DCIS sample. In one embodiment, the method further comprises providing a test sample of

DCIS tissue.

In a further embodiment, the method comprises assigning a prognosis to the sample.

DESCRIPTION OF THE DRAWINGS

Figure 1 (A - D) shows 9 panels depicting DCIS histopathology, Ki67 immunohistochemistry and laser capture microdissection. A (i) shows low nuclear grade DCIS associated with grade 1 invasive breast cancer, A (ii) DCIS Ki67 score 9.3%; B (i) shows intermediate nuclear grade DCIS associated with grade 2 invasive cancer, B (ii) DCIS Ki67 score 21.9%; C (i) shows high nuclear grade DCIS with comedo necrosis, C (ii) DCIS Ki67 score 29.5%; D (i-iii) shows laser capture microdissection of DCIS from frozen tissue section.

Figure 2 (A - D) is four graphs that show the correlation between protein measurement and gene expression ratios determined using oligonucleotide microarrays. A. ER status vs ER gene expression (pO.OOOl, n=46, oligonucleotide ID H200000435); B. PR status vs. PR gene expression (pO.OOOl, n=46, oligonucleotide ID H300019820); C. HER2 status vs. HER2 gene expression (pO.OOOl, n=45, oligonucleotide ID H300020181); D. DCIS Ki67 vs. MKi67 gene expression (pO.OOOl, r=0.83, n=47, oligonucleotide ID H200006808).

Figures 3A-B are tables showing oligonucleotide probes that are significantly differentially expressed between DCIS lesions. Table 3 A depicts oligonucleotide probes that are significantly differentially expressed between DCIS lesions associated with grade 1 and grade 3 invasive cancer. Table 3B depicts oligonucleotide probes that are significantly differentially expressed in DCIS lesions according to ER status and HER2 status.

Figure 4 (A & B) depicts the results of clustering of all samples according to expression. A (i) shows hierarchical clustering of 61 samples according to expression of the top 100 differentially expressed oligonucleotide probes (selected by supervised analysis of DCIS lesions associated with Grade 1 and 3 invasive cancer). Columns represent samples; rows represent individual probes. Heatmap depicts high (dark grey) and low (lighter grey) relative levels of gene expression, A (ii) Grade of associated invasive breast cancer, A (iii) DCIS nuclear grade or sample type; B. Gene expression grade index (GGI) is calculated for each sample in the corresponding heatmap column.

Figure 5 (A - E) shows four graphs and a schematic. A. is a graph showing the correlation between GGI and DCIS Ki67 score (pO.OOOl, r=0.79, n=46); B. is a graph showing Ki67 scores for DCIS samples in low and high MG groups (pO.OOOl, n=46); C. is a graph showing metastasis-free (pO.OOOl) and D. Overall survival (pO.OOOl) analysis for invasive breast cancer of low and high MG defined by a GGI of < or >0 in the van de Vijver et al. dataset (n=294). E. is a schematic showing a two- step classification tree model illustrating prediction of DCIS MG by histopathologic and biomarker features. Figure 6 is a graph showing distant-recurrence free survival for invasive breast cancer of low and high MG defined by GGI < or >0 in the gene expression dataset of Sotiriou et al. (p=0.002, n=179).

Figure 7 (A - E) depicts DNA copy number profiles of low and high molecular grade DCIS. Panels A and B illustrate the frequency of DNA copy number gains (dark grey) and losses (light grey) across the genome (plotted from chromosome lpter to 22qter, X and Y) in DCIS associated with A. grade 1 and; B. grade 3 invasive cancer. Average Iog2 ratio of copy number in DCIS compared with normal male reference DNA is shown in blue; C. Representation of random forest algorithm applied to determine the importance measure for each probe in distinguishing DCIS lesions as associated with grade 1 or grade 3 invasive cancer. Points colored in blue highlight regions that have a much higher copy number in grade 3 cases than in grade 1 ; points in orange have a higher copy number in grade 1 than grade 3. Higher copy number is defined as a difference of 0.25 mean Iog2 ratio or larger; D. Correlation between the number of regions of high level DNA amplification and GGI (p<0.0001, r = 0.62, n^O). E. Number of regions of high level DNA amplification in low and high MG subgroups of DCIS (p=0.003, n=46).

Figure 8 is a Table that lists the sample type and the GEO title for individual case numbers.

DETAILED DESCRIPTION OF THE INVENTION

The increase in incidence of ductal carcinoma in situ (DCIS) associated with mammographic screening for breast cancer has underscored the challenges of managing this important clinical entity. Unlike invasive breast cancer, there is no established histopathologic grading system for DCIS in widespread use, nor are there biological markers of prognosis to guide clinical management. The instant invention uses molecular profiling to identify clinically applicable indicators of DCIS malignant potential.

Definitions

The following definitions are provided for specific terms which are used in the following written description.

As used in the specification and claims, the singular form "a", "an" and "the" include plural references unless the context clearly dictates otherwise. For example, the term "a cell" includes a plurality of cells, including mixtures thereof. The term "a nucleic acid molecule" includes a plurality of nucleic acid molecules.

The term "atypical ductal hyperplasia" is meant to refer to a condition that can occur in the lining of the milk ducts in the breast where the duct contains a characteristic proliferation of abnormal cells. In certain cases, a diagnosis of atypical ductal hyperplasia on its own is associated with an increased risk of subsequently developing invasive breast cancer.). In other certain cases, atypical ductal hyperplasia may be regarded as a pre-malignant condition.

The terms "cancer," and "tumor" are used interchangeably and in either the singular or plural form, refer to cells that have undergone a malignant transformation that makes them pathological to the host organism. Primary cancer cells (that is, cells obtained from near the site of malignant transformation) can be distinguished from non-cancerous cells by well-established techniques, particularly histological examination. The definition of a cancer cell, as used herein, includes not only a primary cancer cell, but also any cell derived from a cancer cell ancestor. This includes metastasized cancer cells, and in vitro cultures and cell lines derived from cancer cells. When referring to a type of cancer that normally manifests as a solid tumor, a "clinically detectable" tumor is one that is detectable on the basis of tumor mass; e.g., by procedures such as CAT scan, MR imaging, X-ray, ultrasound or palpation, and/or which is detectable because of the expression of one or more cancer- specific antigens in a sample obtainable from a patient. The phrase "cell proliferation profile" is meant to refer to a measurement of the proportion of proliferating cells in a group of cells. Cell proliferation profile can be determined by methods and assays such as tritiated thymidine 3H uptake, or methods of high throughput screening in microtiter plates, and high-content screening (HCS) using live cell assays to image cell function, metabolism, and signaling at the level of the individual cell has led to an expanded range of assay formats for measuring cell proliferation, such as fluorescent, luminescent, and colorimetric assays that can determine cell counts, detect DNA synthesis, or measure metabolic activity. The term "comparative genomic hybridization (CGH)" is meant to refer to a molecular cytogenetic method of screening a tumor for genetic changes. In one example, the alterations detected in CGH are differences in the number of copies of genes in cancer relative to the copy number in normal cells. Alterations in cancer are classified as regions of DNA gain and loss and characteristic patterns may be identified. In certain examples, regions of DNA copy number gain may be referred to as regions of DNA amplification. In certain embodiments, CGH can allow rapid screening for DNA copy number gains and losses across the entire genome. In other certain embodiments, CGH is used to generate a genomic profile for classifying the cancer into a low molecular grade or a high molecular grade.

The term "DNA aberration" is meant to refer to any genetic (DNA) abnormality in a cell that is detectable by any molecular method or any form of genetic testing. A DNA aberration may be, for example, the loss or gain of a DNA segment from a chromosome, a structural abnormality in DNA, or a gain or loss in whole gene number.

The term "ductal carcinoma in situ (DCIS)" is meant to refer to a noninvasive cancerous condition. Ductal carcinoma is meant to refer to cancer cells arising from the milk ducts of the breast. In situ is meant to refer to "in place" and refers to the fact that the cancer has not moved out of the duct and into any surrounding tissue. The term "gene" is meant to refer to a polynucleotide that encodes a discrete product, whether RNA or proteinaceous in nature. It is appreciated that more than one polynucleotide may be capable of encoding a discrete product. The term includes alleles and polymorphisms of a gene that encodes the same product, or a functionally associated (including gain, loss, or modulation of function) analog thereof, based upon chromosomal location and ability to recombine during normal mitosis.

The term "gene amplification index (AI)" is meant to refer to a metric that is used to describe the number of discrete regions of DNA amplification exceeding a designated threshold. In certain embodiments, the AI of DCIS is correlated with molecular grade and it can be used to classify samples into high-grade and low-grade. The level of amplifications that is predictive of a high molecular grade tumor may be determined empirically. The level of amplifications that is predictive of a high level tumor can be variable.

The phrase "gene copy number" is meant to refer to the number of copies of a particular gene in the genotype of an individual. In certain embodiments, the gene copy number can be higher in a cancer cell than in a normal cell. In other certain embodiments, the gene copy number can be lower in a cancer cell than in a normal cell. .

The phrase "gene expression grade index" is meant to refer to a metric for quantifying the expression of a certain genes in a sample. Determination of gene expression is made by any method known to one of skill in the art. In certain preferred embodiments, a high gene expression grade index corresponds to a high molecular grade. In other preferred embodiments, a low gene expression grade index corresponds to a low molecular grade. A gene expression grade index corresponds to gene expression above or below a designated threshold. For example, in certain embodiments, a gene expression grade index above a designated threshold corresponds to a high molecular grade. In other certain embodiments, a gene expression index below a certain threshold correlated to a low molecular grade.

The term "gene expression "pattern" or "profile" or "signature" refers to the relative expression of one or more genes, for example one or more genes in combination, between two conditions, which is correlated with being able to distinguish between said conditions, for example, between two or more stages of cancer, or two or more grades of cancer or between an untreated and treated condition, or between a disease and normal sample.

The term "intraductal carcinoma" is meant to refer to any form of cancer confined to normal duct or lobular structures for example the milk ducts of the breast. The term is meant to include ductal carcinoma in situ and lobular carcinoma in situ. The term "Ki67 score" is meant to refer to an index of cellular proliferation. The Ki67 protein is expressed by cells in Gl, S, G2, and M phases of the cell cycle but not Go. A Ki67 score is determined by detection of Ki67 in tissue and determining the number of positive cells/total cells, for example, positive cells/ total tumor cells. In certain embodiments, the Ki67 score or index may increase from low to high molecular grade, for example or from mild to moderate to more severe malignancy in cancer.

The term "microarray" is meant to include a collection of nucleic acid molecules or polypeptides from one or more organisms arranged on a solid support (for example, a chip, plate, or bead). For example, a "microarray" is a linear or two- dimensional array of preferably discrete regions, each having a defined area, formed on the surface of a solid support such as, but not limited to, glass, plastic, or synthetic membrane. The density of the discrete regions on a microarray is determined by the total numbers of immobilized polynucleotides to be detected on the surface of a single solid phase support, preferably at least about 50/cm² , more preferably at least about 100/ cm², even more preferably at least about 500/ cm², but preferably below about 1,000/ cm². As used herein, a DNA microarray is an array of oligonucleotides or polynucleotides placed on a chip or other surfaces used to hybridize to amplified or cloned polynucleotides from a sample. Since the position of each particular group of primers in the array is known, the identities of a sample polynucleotides can be determined based on their binding to a particular position in the microarray.

The term "grade" is meant to refer to a feature of cancer that is determined by one skilled in the art according to pre-defined descriptive criteria. In certain embodiments, high grade reflects a greater degree of difference between cancer cells and normal cells than low grade. In other embodiments, high grade cancer has a greater propensity to grow and spread than low grade cancer and is associated with a worse prognosis. In some cancer grading schemes, an 'intermediate' grade category includes cases with features that do not meet criteria for high or low grade.

The term "molecular grade" is meant to refer to a grade classification based on evaluation of the molecular make-up of a cancer specimen. This could include but is not limited to features of RNA, DNA or protein constituents singly or in combination. RNA features include but are not limited to gene expression. DNA features include but are not limited to gene copy number. Gene expression and gene copy number can be determined by any method known to one of skill in the art, for example any type of polymerase chain reaction (PCR) technique, including PCR, real time PCR, touch start PCR, Quantitative-reverse transcription-PCR, targeted differential display, serial analysis of gene expression (S AGE), microarray analysis including competitive and non-competitive hybridization approaches, multi-analyte profiling beads, array-based comparative genomic hybridization (CGH). In certain embodiments, molecular grade can be determined using microarray based gene expression profiling. For example, microarrays can be used to compare the expression pattern of genes between different samples, for example a cancer sample and a normal sample from the same individual, or cancer samples of different stages or grades. From the gene expression analysis, for example the microarray analysis, groups of genes are identified that can be used to categorize a case as high molecular grade or low molecular grade. In one embodiment, high molecular grade can be categorized by the identification of 1, 2, 4, 6, 10, 15, 20, 30, 40, 50, 75, 100, 200, 400, 500 or more distinguishing genes from low molecular grade. In other embodiments, high molecular grade can be categorized by the identification of a 2-fold, 3-fold, 4-fold, 5-fold, 10-fold or more increase or decrease in gene expression between genes identifies in high and low molecular grade. In certain preferred embodiments, one or more molecular profile characteristics that define the boundary, or cut-off, between high molecular grade and low molecular grade are determined empirically.

The term "molecular profiling" is meant to refer to any number of molecular methodologies that can be used to determine differences in the molecular make-up of two or more samples, for example a cancer tissue and a control tissue. This could include but is not limited to differences in constituent DNA, RNA or protein. Molecular profiling techniques include, but are not limited to a any type of polymerase chain reaction (PCR) technique, including PCR, real time PCR, touch start PCR, Quantitative-reverse transcription-PCR, targeted differential display, serial analysis of gene expression (SAGE), microarray analysis including competitive and non-competitive hybridization approaches, multi-analyte profiling beads, array-based comparative genomic hybridization (CGH).

The term "nuclear grade" is meant to refer to a qualitative evaluation of the size, shape and appearance of the nucleus in tumor cells by microscopic examination by one skilled in the art. In certain embodiments, nuclear grade is associated with molecular grade. For example, a high molecular grade is associated with a high nuclear grade, and a low molecular grade is associated with a low nuclear grade.

The term "nucleic acid" is meant to refer to an oligomer or polymer of ribonucleic acid or deoxyribonucleic acid, or analog thereof. This term includes oligomers consisting of naturally occurring bases, sugars, and intersugar (backbone) linkages as well as oligomers having non-naturally occurring portions which function similarly. Such modified or substituted oligonucleotides are often preferred over native forms because of properties such as, for example, enhanced stability in the presence of nucleases. The term "serial analysis of gene expression (S AGE)" is meant to refer to a method for comprehensive analysis of gene expression patterns. Three principles underlie the SAGE methodology (I) A short sequence tag (10-14bp) contains sufficient information to uniquely identify a transcript provided that that the tag is obtained from a unique position within each transcript; (2) sequence tags can be linked together to from long serial molecules that can be cloned and sequenced; and (3) quantitation of the number of times a particular tag is observed provides the expression level of the corresponding transcript. The output of SAGE is a list of short sequence tags and the number of times it is observed. The SAGE technique allows the quantitative and simultaneous analysis of a large number of transcripts. The term "sample" refers to any biological or chemical mixture for use in the method of the invention. The sample can be a biological sample. The biological samples are generally derived from a patient. The sample can be a "cancer sample" that is a sample that is derived from a neoplasia, cancer or tumor. The sample can be a "control sample." By "control" is meant a standard or reference condition. For example, a control sample can also be a cancer or tumor sample that is untreated. The terms "stage" or "stages" (or equivalents thereof) of cancer as used herein refer to the physical extent of disease spread and are readily known to skilled in the art. Non-limiting examples include in situ cancer, locally invasive cancer, cancer that has spread to regional lymph nodes and cancer that has spread to distant sites in the body. The term "subject" is intended to include vertebrates, preferably a mammal.

Mammals include, but are not limited to, humans.

Ductal Carcinoma in situ (DCIS)

Ductal carcinoma in situ (DCIS), sometimes called intraductal carcinoma, is a group of lesions in which cancer cells have grown within the duct. DCIS may co-exist with invasive cancer or exist in the absence of invasive cancer. A diagnosis of 'DCIS' without further qualification implies that invasive cancer is not present; DCIS characteristically does not invade outside the duct or show metastases at presentation. DCIS is a noninvasive condition, but can progress to become invasive cancer. According to the National Cancer Institute, the widespread use of mammography in the United States has lead to a marked increase in the frequency of the diagnosis of DCIS. In 1998, DCIS accounted for about 18% of all newly diagnosed invasive plus noninvasive breast tumors in the United States. The American Cancer Society estimates that 41,000 new cases of ductal carcinoma in situ (also called DCIS or intraductal carcinoma) will be diagnosed in 2000, making DCIS the most common type of non- invasive breast cancer in women. DCIS accounts for nearly 25% of all breast cancer diagnoses. An estimated 185,000 cases of invasive breast cancer are diagnosed each year, and approximately 20% to 30% of breast cancers detected by mammography are carcinoma in situ. Very few cases of DCIS present as a palpable mass; 80% are diagnosed by mammography alone (Fonseca R, Hartmann LC, Petersen IA, et. al.: Ductal carcinoma in situ of the breast. Ann Intern Med 127 (11): 1013-22, 1997). DCIS comprises a heterogeneous group of histopathologic lesions that have been classified into several subtypes based primarily on architectural pattern: micropapillary, papillary, solid, cribiform, and comedo. Comedo-type DCIS consists of cells that appear cytologically malignant, with the presence of high-grade nuclei, pleomorphism, and abundant central luminal necrosis. Comedo-type DCIS appears to be more aggressive, with a higher probability of associated invasive ductal carcinoma (Fisher ER, Dignam J, Tan-Chiu E, et al.: Pathologic findings from the National Surgical Adjuvant Breast Project (NSABP) eight-year update of Protocol B- 17: intraductal carcinoma. Cancer 86 (3): 429-38, 1999). Comedo DCIS is often a grossly palpable lesion, which was probably considered "cancer" in the 19th and early 20th century and progresses to cancer (without definitive therapy) in at least 50% of patients within three years (Ottesen et al, 1992; Page et al, 1982). Most of the molecular alterations which have been reported in invasive breast cancer have been observed in cases of comedo DCIS (Poller et al, 1993; Radford et al, 1993; and, Tsuda et al, 1993).

The women at high risk for DCIS are similar to those who are at high risk for developing invasive cancers. The shared risk factors include: never having had a full- term pregnancy, having a first full-term pregnancy after age 30, menstrual periods starting early, late menopause, having a parent or sibling with breast cancer, more than five years of hormone replacement therapy (HRT), particularly with the therapy that combines estrogen and progestin, and carrying a mutation in a breast cancer susceptibility gene such as BRCAl or BRCA2).

DCIS Grading

Breast cancer biology is remarkably heterogeneous, and the repeated demonstration that biologic indicators such as histopathologic grade, individual biomarkers and more recently gene expression signatures can predict prognosis of invasive cancer demonstrates a pervasive impact of biology on clinical course [5-9]. Although DCIS exists within the ducto-lobular tree, the same broad spectrum of features that are characteristic of breast cancer biology are present. For example, pure DCIS and DCIS associated with invasive cancer have similar grade associated expression of hormone receptors, p53 and HER2, [10-13] as well as cytogenetic features such as loss of 16q in low and intermediate grade lesions [14-16]. In addition, the strong concordance between grade, biomarker, cytogenetic and gene expression profiles of concomitant in situ and invasive disease further allows the paradigm of invasive breast cancer to inform clinical interpretation of biologic features of DCIS [14-20].

A current impediment to inclusion of biology into clinical evaluation of DCIS is absence of a definitive histopathologic grading system. A number of DCIS grading systems have been proposed, variously using architectural patterns, nuclear morphology and necrosis to distinguish low, intermediate and high-grade lesions [21- 23]. However, there is a wide range of appearances of DCIS and combinations of high and low-grade features are not uncommon [24, 25]. Consequently, while classification of lesions with exclusively high or low-grade features poses no difficulty, 'intermediate grade' DCIS can be particularly heterogeneous [26].

Despite a pressing clinical need to identify robust indicators of DCIS biology, studies aiming to identify such markers are remarkably difficult. In particular, there is an absence of accessible clinical end-points to which candidate markers can be compared. Disease recurrence might seem the best indicator of clinically aggressive DCIS, however this is heavily influenced by treatment and is relatively uncommon [27, 28], implying that very large study cohorts may be required to distinguish true biological differences. An alternative approach for these studies is to use a surrogate end-point which is intermediate between the variable of interest (DCIS malignant potential) and the true clinical end-point (breast cancer specific mortality) [29]. The grade of invasive cancer co-existing with DCIS fulfils these criteria on the basis that i) the malignant potential of DCIS and concomitant invasive cancer are closely related since critical molecular features of the two elements are the same and ii) critical molecular features of invasive cancer are reflected by grade which is in turn correlated with survival. Hence invasive cancer grade occupies an intermediate position between DCIS malignant potential and survival and can be a useful reference for the investigation of DCIS. The ability of gene expression profiling to aid discovery of biological subclasses of disease through stringent computational analysis of detailed datasets has repeatedly been demonstrated [30-32]. This approach is of particular value in circumstances where histopathology is unable to resolve clinically important differences. For example Sotiriou et al. recently reported that a gene expression profile associated with high-grade invasive breast cancer could distinguish intermediate grade cancers according to prognosis [9]. Molecular profiling may therefore be an approach well suited to the problem of DCIS grading.

Microarray

Included in the invention are microarrays. Microarray technology is of use in determining the molecular grade of cancer, in particular the molecular grade of cancer in a subject, as described herein.

The term "microarray" is meant to include a collection of nucleic acid molecules or polypeptides from one or more organisms arranged on a solid support (for example, a chip, plate, or bead). For example, a "microarray" is a linear or two- dimensional array of preferably discrete regions, each having a defined area, formed on the surface of a substrate. The substrate can be a solid support made of, for example, glass, plastic, or a synthetic material. The substrate can be a two- dimensional substrate such as a glass slide, a wafer (e.g., silica or plastic), a mass spectroscopy plate, or a three-dimensional substrate such as a gel pad. Addresses in addition to address of the plurality can be disposed on the array. The density of the discrete regions on a microarray is determined by the total numbers of immobilized polynucleotides to be detected on the surface of a single solid phase support, preferably at least about 50/cm² , more preferably at least about 100/ cm², even more preferably at least about 500/ cm², but preferably below about 1,000/ cm².

An array can be generated by any of a variety of methods. Appropriate methods include, e.g., photolithographic methods (see, e.g.; U.S. Pat. Nos. 5,143,854; 5,510,270; and 5,527,681), mechanical methods (e.g., directed- flow methods as described in U.S. Pat. No. 5,384,261), pin-based methods (e.g., as described in U.S. Pat. No. 5,288,514), and bead-based techniques (e.g., as described in PCT

US/93/04145).

Also included in the invention is a single stranded nucleic acid probe comprising: (a) the nucleotide sequence of a tag selected from those listed in Figures 3A-B; or (b) the complement of the nucleotide sequence.

In particular embodiments, the microarray comprises a substrate having at least 10 or more addresses, wherein each address comprises a capture probe comprising a nucleic acid sequence comprising a tag nucleotide sequence. The tag nucleotide sequence can be one that corresponds to a gene encoding a protein selected from the group of sequences listed in Figures 3A-B. The array can contain at least 10 addresses; at least 25 addresses; at least 50 addresses; at least 100 addresses; at least

200 addresses; or at least 500 or more addresses.

In another embodiment, the array can be used to assay gene expression in a tissue to ascertain tissue specificity of genes in the array. If a sufficient number of diverse samples are analyzed, clustering (for example, hierarchical clustering, k- means clustering, Bayesian clustering) can be used to identify other genes which are co-regulated with the gene of interest. For example, the array can be used for the quantitation of the expression of multiple genes. Thus, not only tissue specificity, but also the level of expression of a battery of genes in the tissue is ascertained. Quantitative data can be used to group, or cluster, genes on the basis of their tissue expression and level of expression in that tissue.

For example, array analysis of gene expression can be used to assess the effect of cell-cell interactions on the expression of a gene of interest. A first tissue can be perturbed and nucleic acid from a second tissue that interacts with the first tissue can be analyzed. In this context, the effect of one cell type on another cell type in response to a biological stimulus can be determined, e.g., to monitor the effect of cell-cell interaction at the level of gene expression.

In another embodiment, the array can be used to monitor expression of one or more genes in the array with respect to time. For example, samples obtained from different time points can be probed with the array. Such analysis can identify and/or characterize the development of a gene X-associated disease or disorder (e.g., breast cancer such as invasive breast cancer); and processes, such as a cellular transformation associated with a gene X-associated disease or disorder. The method can also evaluate the treatment and/or progression of a gene X-associated disease or disorder. In another aspect, the invention features an array having a plurality of addresses. Each address of the plurality includes a unique polypeptide. At least one address of the plurality has disposed thereon a protein or fragment thereof. Methods of producing polypeptide arrays are described in the art [e.g., in De Wildt et al. (2000) Nature Biotech. 18:989-994; Lueking et al. (1999) Anal. Biochem. 270:103-111 ; Ge, H. (2000) Nucleic Acids Res. 28 e3:I-VII; MacBeath, G., and Schreiber, S. L. (2000) Science 289:1760-1763; and WO 99/51773Al]. In a preferred embodiment, each addresses of the plurality has disposed thereon a polypeptide at least 60, 70, 80, 85, 90, 95, or 99% identical to protein X or fragment thereof. For example, multiple variants of protein X (e.g., encoded by allelic variants, site-directed mutants, random mutants, or combinatorial mutants) can be disposed at individual addresses of the plurality. Addresses in addition to the address of the plurality can be disposed on the array.

The polypeptide array can be used to detect a protein -binding compound, e.g., an antibody in a sample from a subject with specificity for a protein of interest or the presence of a protein of interest-binding protein or ligand.

Nucleic Acids

The nucleic acid molecules of the invention include those containing or consisting of the nucleotide sequences (or the complements thereof) isolated from samples, for example microdissected areas of DCIS. The nucleic acid molecules of the invention can be cDNA, genomic DNA, synthetic DNA, or RNA, and can be double-stranded or single-stranded (i.e., either a sense or an antisense strand). Segments of these molecules are also considered within the scope of the invention, and can be produced by, for example, the polymerase chain reaction (PCR) or generated by treatment with one or more restriction endonucleases. A ribonucleic acid (RNA) molecule can be produced by in vitro transcription. Preferably, the nucleic acid molecules encode polypeptides that, regardless of length, are soluble under normal physiological conditions.

The nucleic acid molecules of the invention can contain naturally occurring sequences, or sequences that differ from those that occur naturally, but, due to the degeneracy of the genetic code, encode the same polypeptide. In addition, these nucleic acid molecules are not limited to coding sequences, e.g., they can include some or all of the non-coding sequences that lie upstream or downstream from a coding sequence. They can also contain irrelevant sequences at their 5' and/or 3' ends (e.g., sequences derived from a vector). The nucleic acid molecules of the invention can be synthesized (for example, by phosphoramidite-based synthesis) or obtained from a biological cell, such as the cell of a mammal. The nucleic acids can be those of a human, non-human primate (e.g., monkey), mouse, rat, guinea pig, cow, sheep, horse, pig, rabbit, dog, or cat. Combinations or modifications of the nucleotides within these types of nucleic acids are also encompassed.

In addition, the isolated nucleic acid molecules of the invention encompass segments that are not found as such in the natural state. Thus, the invention encompasses recombinant nucleic acid molecules incorporated into a vector (for example, a plasmid or viral vector) or into the genome of a heterologous cell (or the genome of a homologous cell, at a position other than the natural chromosomal location). Recombinant nucleic acid molecules and uses therefor are discussed further below.

Techniques associated with detection or regulation of genes are well known to skilled artisans. Such techniques can be used to diagnose and/or treat disorders (e.g., DCIS or invasive cancer) associated with aberrant expression of the genes corresponding to those identified using any method of molecular analysis known to one of skill in the art for example, but not limited to, Polymerase Chain Reaction (PCR), targeted differential display, serial analysis of gene expression (S AGE), microarray analysis, array-based comparative genomic hybridization (CGH).

Polypeptides and Polypeptide Fragments Polypeptides of the invention include all those encoded by the nucleic acids described above and functional fragments of these polypeptides. The polypeptides embraced by the invention also include fusion proteins that contain either a full-length polypeptide, or a functional fragment thereof, fused to unrelated amino acid sequence. The unrelated sequences can be additional functional domains or signal peptides. The polypeptides can be any of those described-above but with not more than 50 (e.g., not more than: 50; 40; 30; 25; 20; 15; 12, 10; nine; eight; seven; six; five; four; three; two; or one) conservative substitution(s). Conservative substitutions typically include substitutions within the following groups: glycine and alanine; valine, isoleucine, and leucine; aspartic acid and glutamic acid; asparagine, glutamine, serine and threonine; lysine, histidine and arginine; and phenylalanine and tyrosine. All that is required of a polypeptide with one or more conservative substitutions is that it have at least 5% (e.g., at least: 5%; 10%; 20%; 30%; 40%; 50%; 60%; 70%; 80%; 90%; 95%; 98%; 99%; 100%; or more) of the activity (e.g., ability to inhibit proliferation of breast cancer cells) of the relevant wild-type, mature polypeptide.

Polypeptides of the invention and those useful for the invention can be purified from natural sources (e.g., blood, serum, plasma, tissues or cells such as normal breast or cancerous breast epithelial cells (of the luminal type), myoepithelial cells, leukocytes, or endothelial cells). Smaller peptides (less than 50 amino acids long) can also be conveniently synthesized by standard chemical means. In addition, both polypeptides and peptides can be produced by standard in vitro recombinant DNA techniques and in vivo transgenesis, using nucleotide sequences encoding the appropriate polypeptides or peptides. Methods well-known to those skilled in the art can be used to construct expression vectors containing relevant coding sequences and appropriate transcriptional/translational control signals. See, for example; the techniques described in Sambrook et al., Molecular Cloning: A Laboratory Manual (2nd Ed.) [Cold Spring Harbor Laboratory, N. Y., 1989], and Ausubel et al., Current Protocols in Molecular Biology [Green Publishing Associates and Wiley Interscience, N.Y., 1989]. Polypeptides and fragments of the invention, and those useful for the invention, also include those described above, but modified for in vivo use by the addition, at the amino- and/or carboxyl-terminal ends, of a blocking agent to facilitate survival of the relevant polypeptide in vivo. This can be useful in those situations in which the peptide termini tend to be degraded by proteases prior to cellular uptake. Such blocking agents can include, without limitation, additional related or unrelated peptide sequences that can be attached to the amino and/or carboxyl terminal residues of the peptide to be administered. This can be done either chemically during the synthesis of the peptide or by recombinant DNA technology by methods familiar to artisans of average skill.

Methods of the Invention

Included in the invention are methods for determining the molecular grade of a cancer. The method as described herein comprises identifying genes differentially expressed in a cancer sample and a control sample by determining gene expression and gene copy number by molecular profiling, and then classifying the cancerous tissue by gene expression and gene copy number to thereby determine the molecular grade of the cancer.

Also included are methods for determining the molecular grade of a cancer in a subject. The method comprises identifying genes differentially expressed in a cancer sample and a control sample by determining gene expression and gene copy number by molecular profiling, and then classifying the cancerous tissue by gene expression and gene copy number to then determine the molecular grade of a cancer in a subject.

The method comprises obtaining one or more samples of a cancer tissue. Further included in the invention are methods of identifying the molecular grade of a DCIS sample comprising determining the molecular grade of a sample and then comparing the molecular grade of a sample with a reference or control profile, and thereby identifying the grade of the DCIS sample.

Also included in the invention are methods of determining the molecular grade of a cancer comprising profiling molecularly to determine gene expression and gene copy number in a test sample, and then comparing a test expression profile to a control expression profile, and determining a molecular grade that corresponds to gene expression and gene copy number, where the molecular grade of the cancer is determined.

According to the methods of the invention, determining the molecular grade of a cancer comprises classifying the cancer as low molecular grade or high molecular grade. The term "molecular grade" is meant to refer to a molecular-based classification based on one or more of gene expression and gene copy number patterns. Gene expression and gene copy number can be determined by any method known to one of skill in the art for example any type of polymerase chain reaction (PCR) technique, including PCR, real time PCR, touch start PCR, Quantitative- reverse transcription-PCR, targeted differential display, serial analysis of gene expression (SAGE), microarray analysis including competitive and non-competitive hybridization approaches, multi-analyte profiling beads, array-based comparative genomic hybridization (CGH). In certain embodiments, molecular grade can be determined using microarray based gene expression profiling, as described above. Any method of gene expression known to one of skill in the art can be used to determine a gene expression profile associated with high molecular grade DCIS. In preferred embodiments of the invention, a gene expression profile associated with high molecular grade DCIS comprises one or more of the genes listed in Figures 3A- B. In certain examples, microarrays can be used to compare the expression pattern of genes between different samples, for example a cancer sample and a control sample, or cancer samples at different stages or grades. From the gene expression analysis, for example data obtained from a microarray analysis, groups of genes are identified that can be used to categorize high molecular grade and low molecular grade. In one embodiment, high molecular grade can be categorized by the identification of 1, 2, 4, 6, 8, 10, 15,20, 30, 40, 50, 75, 100, 200, 400, 500 or more genes that distinguish the high molecular grade from a low molecular grade. In other embodiments, high molecular grade can be categorized by the identification of a 2- fold, 3-fold, 4-fold, 5-fold, 10-fold or more increase or decrease in gene expression between genes identifies in high and low molecular grade. Using this method, a profile of genes that are highly correlated with one stage or molecular grade category relative to another may be used to assay a sample from a subject afflicted with, or suspected of having, breast cancer to identify the stage or molecular grade category of breast cancer to which the sample belongs. Such an assay may further be used as part of a method to determine the therapeutic treatment for said subject based upon the stage or the molecular grade of breast cancer identified. In certain embodiments of the method as described, classifying the cancer by gene expression or gene copy number comprises calculating a gene expression grade index. The gene expression grade index is calculated using the gene expression pattern of the cancer sample derived from molecular profiling versus the gene expression associated with the control sample. In certain examples, a high gene expression grade index corresponds to a high molecular grade. Likewise, a low gene expression grade index corresponds to a low molecular grade. In invasive breast cancer high molecular grade correlates with a lower rate of survival compared to a control, and the low molecular grade correlates with a high level of survival compared to a control. In certain embodiments, a high molecular grade correlates with, but is not limited to a high number of regions of high level DNA amplification compared to a control or a low molecular grade. In other certain embodiments, a low molecular grade correlates with, but is not limited to, a low number of regions of high DNA level amplification compared to a control or a high molecular grade. It is advantageous in certain clinical situations to use the molecular grade to determine the prognosis of a patient.

In certain examples, gene expression and gene copy number are determined by molecular analysis. A number of methods of molecular analysis are known to one of skill in the art, and are available for use to determine differential gene expression.

Examples include, but are not limited to, any type of polymerase chain reaction (PCR) technique, including PCR, real time PCR, touch start PCR, Quantitative-reverse transcription-PCR, targeted differential display, serial analysis of gene expression (SAGE), microarray analysis including competitive and non-competitive hybridization approaches, multi-analyte profiling beads, array-based comparative genomic hybridization (CGH). In certain preferred embodiment, CGH generates a genomic profile for classifying the cancer into a low molecular grade or a high molecular grade. CGH is a molecular cytogenetic method of screening a tumor for genetic changes. The alterations are classified as DNA gains and losses. In a typical CGH measurement, total genomic DNA is isolated from test and reference cell populations, differentially labeled and hybridized to metaphase chromosomes or, more recently, DNA microarrays. The relative hybridization intensity of the test and reference signals at a given location is then proportional to the relative copy number of those sequences in the test and reference genomes. If the reference genome is normal, then increases and decreases in the intensity ratio directly indicate DNA copy-number variation in the genome of the test cells. Regions of DNA copy number gain may be referred to as regions of DNA amplification. More than two genomes can be compared simultaneously if distinguishable labels are available (Pinkel D et al. Array comparative genomic hybridization and its applications in cancer. Nature Genetics, 37, S11 - S17 (2005)). In certain examples, an amplification index (AI) is generated from the results of CGH. The amplification index (AI) is a metric used to described the number of regions of DNA showing copy number gain over a specified threshold. The amplification index (AI) is correlated with molecular grade. In certain examples, not meant to be limiting, greater than two high-level amplifications is correlated with a high grade tumor.

The correlated genes may be used singly with significant accuracy or in combination to increase the ability to accurately discriminate between various molecular grades of cancer or stages in the development of cancer, for example but not limited to, breast cancer. Thus, the method described herein provides a means for correlating a molecular expression phenotype with a physiological (cellular) stage or state. This correlation provides a way to molecularly diagnose and/or monitor a cell's status in comparison to different cancerous versus non-cancerous phenotypes as disclosed herein. Additional uses of the correlated gene(s) are in the classification of cells and tissues, including non-malignant or pre-malignant conditions, determination of diagnosis and/or prognosis; and determination and/or alteration of therapy. In a further aspect, the gene(s) identified by a model as capable of discriminating between breast cancer stages may be used to identify the cellular state of an unknown sample of cell(s) from the breast. Preferably, the sample is isolated via minimally- invasive means. The expression of said gene(s) in said unknown sample may be determined and compared to the expression of said gene(s) in control data of gene expression patterns from the various stages of breast cancer. Optionally, the comparison to a reference or a control samples may be by comparison to the model(s) constructed based on the reference or the control samples.

In certain embodiments, the genes expressed in cancer samples are selected from one or more of the genes listed in Figures 3A-B.

Using the methods of molecular profiling as described herein it is possible to group the differentially expressed genes into categories of genes based on, for example function or structure. As such, the differentially expressed genes may comprise one or more genes involved in, for example, but not limited to, cell proliferation.

Diagnostic Methods

Also included in the invention are diagnostic methods. Generally, diagnostic methods can be used to determine the molecular grade of a cancer. For example, a diagnostic method can include determining the molecular grade of cancer using nuclear grade and Ki67 score. In one example, the method comprises determining a nuclear grade and a Ki67 score in a primary invasive breast carcinoma sample, wherein the molecular grade of cancer is determined by using nuclear grade and Ki67 score.

Prognostic Methods

Included in the invention are methods of prognosis, for example, a method of predicting the prognosis of a cancer patient. In certain examples, the methods of prognosis are used for a breast cancer patient. In certain examples, the method comprises determining a nuclear grade and a

Ki67 score in a primary invasive breast carcinoma sample and a control sample, and comparing the nuclear grade and Ki67 score from the subject sample to the control sample, where a molecular grade and the Ki67 score higher in the carcinoma sample than in the control sample is an indication that the prognosis of the test patient is poor.

Therapeutics

Included in the invention are methods of identifying therapeutic compounds. The method comprises determining the molecular grade of cancer according to the methods as described herein, contacting the sample with a candidate compound, and then determining the cell proliferation profile or molecular grade of the sample, thus identifying a therapeutic compound.

In certain examples, cells can be contacted with a therapeutic agent. The expression profile of the cells is determined using the array, and the expression profile is compared to the profile of like cells not contacted with the agent. For example, the assay can be used to determine or analyze the molecular basis of an effect of the therapeutic agent, for example the effect of the therapeutic agent on cell proliferation. If an agent is administered to a cell, the invention provides an assay to determine the molecular basis of the effect of the therapeutic. Moreover, undesirable biological effects of a therapeutic agent can be determined at the molecular level using the same methods and thus the effects of an agent on expression of other than the target gene can be ascertained and counteracted.

The methods are also useful for ascertaining the effects of therapeutics on the expression of a gene on the expression of other genes in the same cell or in different cells (e.g., ascertaining the effect of gene X expression on the expression of other genes). This provides, for example, for a selection of alternate molecular targets for therapeutic intervention if the ultimate or downstream target cannot be regulated.

Test Samples

Test samples according to the invention include any tissue that one of skill in the art, for example a clinician, wants to determine a molecular profile, for example, any tissue useful in a method of prognosis, diagnosis, or therapy. The methods described herein comprise steps of providing samples of tissues, for example cancerous tissues or test tissue. Samples to be used are limited only by what is useful to the skilled practitioner; however examples include intraductal carcinoma, atypical ductal hyperplasia and epithelium. Cancerous tissue or test tissue can be any cancerous or non-cancerous tissue, including breast tissue. Test samples can comprise primary invasive breast cancer tissue from a subject. Cancerous tissue or test tissue can be DCIS tissue.

Kits

Included in the invention are kits for performing any of the methods as described herein.

In certain examples, the kits comprise at least 10 probes, each probe comprising a nucleic acid sequence comprising a tag nucleotide sequence selected from those listed in Figures 3A-B.

EXAMPLES

A means to predict the malignant potential of DCIS has important implications for management of this increasingly important condition. In the study of gene expression profiles presented herein, a signature that distinguishes between low and high grade DCIS cases is presented. The ability of histopathologic nuclear grade and Ki67 score to essentially recapitulate this molecular classification suggests a novel DCIS grading scheme suitable for routine use. Such a grading scheme has the advantage of practical application and the support of an objective and fully integrated analysis of biology.

The invention will now be further illustrated with reference to the following methods and examples. It will be appreciated that what follows is by way of example only and that modifications to detail may be made while still falling within the scope of the invention. Example 1: Defining the Cohort

The cohort included 46 cases of invasive breast cancer: 45 with concomitant DCIS and one with LCIS only. Tumor characteristics are summarized in Table 1, Table 2 and Table 3 shown below.

Table 1 : Tumor Characteristics n(%)

Invasive Tumor Size (mm)

(n=46) <5 5(10.9)

5-10 5(10.9)

10-20 16(34.8)

20-50 17(37.0)

>50 3 (6.5)

Lymph Node Status 1

(# positive nodes) 0 13(28.9)

(n=45°) 1-3 23(51.1)

4-9 6(13.3)

>10 3(6.7)

Histological Subtype

(n=46) Ductal NOS 41(89.1)

Lobular, pleomorphic 1 (2.2)

Cribriform 1 (2.2)

Mixed^b 2(4.3)

Too small to type 1 (2.2)

Grade

(n=4?) 1 14(32.6)

2 19(44.2)

3 10(23.3)

Lymphovascular Invasion

(n=46) Absent 10(21.7)

Present 36(78.3)

% Overall Tumour Comprised by in situ Carcinoma

(n=46) <25% 17(37.0)

25-50% 15(32.6)

50-75% 4(8.7)

>75% 10(21.7)

Uniform/Mixed DCIS Morphology

(n=4f) Uniform 32(69.6)

Mixed 14(30.4)

" 1 case lymph node status not available b Mixed subtypes included a mixture of NOS/tubular/cribriform and NOS/mucinous c 3 cases too small to grade ^d excludes one sample from a tumor with LCIS only Abbreviations: NOS, not otherwise specified

Table 2: Features of samples microdissected from invasive breast cancers assessed by gene expression microarray analysis.

Characteristic A. Total n(%)

Sample Type

(n=61) Benign epithelium 7(11.5)

PDWA 1(1.6)

ADH 4(6.6)

LCIS^a 2(3.3)

DCIS 47(77.0)

DCIS Histopathologic Features

DCIS Nuclear Grade

(n=4t) Low 10(21.3)

Intermediate 18(38.3)

High 19(40.4)

Necrosis

(n=4?) Absent/Punctate 11 (23.4)

Comedo 36(76.6)

Architectural Pattern

(n=4t) Polarised 15(31.9)

Mixed 24(51.1)

Non-Polarised 8(17.0)

Cell Polarisation

(n=47^b) Present 39(83.0)

Absent 8(17.0)

Calcification

(n=47^b) Absent 14(29.8)

Present^d 33 (70.2)

Whole Tumour Features

ER

(n=45) Positive 38(84.4)

Negative 7(15.6)

PR

(n=45) Positive 35 (77.8)

Negative 10(22.2)

HER2

(n=44°) Negative (o, +, ++) 35 (79.5)

Positive (+++) 9(20.5) p53

(n=4l) Negative 35(81.4) Positive 8 (18.6)

Table 3: Gene ontology enrichment analysis of differentially expressed genes

List

EASE Score (P- List Hits Total Population Population

System Category value)" (LH)" (LT)^C Hits (PH)^d Total (PT)'

Biological Process mitotic cell cycle 1.36E-18 25 77 266 8700 cell cycle 3.71E-17 31 77 550 8700 mitosis 4.91E-15 16 77 101 8700

M phase of mitotic cell 6.65E-15 cycle

16 77 103 8700 nuclear division 9.11E-15 17 77 128 8700

M phase 1.79E-14 17 77 133 8700 cell proliferation 3.47E- 13 32 77 826 8700 regulation of cell cycle 7.17E-09

17 77 310 8700 cell growth and/or 1.19E-08 maintenance

48 77 2639 8700 cellular physiological 0.00000019 process

50 77 3070 8700

DNA metabolism 0.00000123 16 77 395 8700 regulation of mitosis 0.00000166

6 77 24 8700

DNA replication and 0.0000038 chromosome cycle

10 77 143 8700 cytokinesis 0.00000608 8 77 81 8700 cellular process 0.0000797 59 77 4767 8700

G2/M transition of 0.000468 mitotic cell cycle

5 77 42 8700 cell cycle checkpoint 0.0022 4 77 30 8700 response to DNA 0.00235 damage stimulus

7 77 156 8700 response to endogenous 0.00275 stimulus

7 77 161 8700 chromosome 0.0047 condensation

3 77 12 8700

DNA repair 0.00546 6 77 130 8700

Gl phase of mitotic cell 0.0064 cycle

3 77 14 8700

DNA replication 0.0155 5 77 110 8700

S phase of mitotic cell 0.0159 cycle

5 77 1 1 1 8700 regulation of CDK 0.0392 activity

3 77 36 8700 transcription from Pol II 0.0464 promoter 77 378 8700

Cellular Component microtubule 0.00000289 cytoskeleton 10 67 144 7813 nucleus 0.0000161 37 67 2298 7813 spindle 0.000249 5 67 37 7813 intracellular 0.000273 53 67 4519 7813 chromosome, 0.000592 pericentric region 4 67 20 7813 chromosome 0.00829 6 67 149 7813 cytoskeleton 0.0179 10 67 476 7813 microtubule associated 0.0279 complex

4 67 78 7813

Molecular Function

ATP binding 0.0264 13 68 900 9264

DNA binding 0.0275 18 68 1457 9264 adenyl nucleotide 0.0287 binding

13 68 91 1 9264 motor activity 0.0482 4 68 113 9264 ^a EASE Score (P-value) represents the level of confidence that this term is over-represented in the DCIS discriminative gene list b LH - number of genes with this term in the DCIS discriminative gene list c LT - number of genes in the DCIS discriminative gene list mapped to any term in this ontology

("system") d PH - number of genes with this gene ontology term on the background gene list (ie. the entire oligonucleotide microarray) e PT - number of genes on the entire oligonucleotide microarray mapped to any term in this ontology ('system")

Example 2: Gene expression profiling of DCIS

RNA was extracted for gene expression profiling from microdissected areas of in situ carcinoma and, in a proportion of cases, adjacent atypical ductal hyperplasia, proliferative disease without atypia and benign epithelium (Figure 1). Of the 14 tumors with two distinct morphologic subtypes of DCIS, both of these were present and separately captured from frozen sections in two cases. In the remaining 12 cases, one subtype only was available for molecular analysis. Histopathologic features of samples analyzed are summarized in Table 2, above.

In support of the reliability of the gene expression data, there was a significant correlation between protein levels and relative expression at oligonucleotide probes corresponding to ER (pO.OOOl), PR (pO.OOOl), HER2 (pO.OOOl) and Ki67 (pO.0001 r=0.83), as shown in Figure 2. Example 3: Grade associated gene expression profile of DCIS

To determine a gene expression profile that could distinguish DCIS cases according to malignant potential, supervised analysis of gene expression in DCIS associated with grade 1 invasive cancer (n=14) versus DCIS associated with grade 3 invasive cancer (n=9) was performed. In this analysis, the predominant DCIS type only was included for cases with multiple types separately sampled (n=2) and the single tumor with LCIS only was excluded.

There was a marked difference between the two groups at a gene expression level with significant differential expression detected at 173 oligonucleotide probes, corresponding to 146 individual genes and 13 expressed sequence tags (ESTs) (FDR<0.05). This is shown in the tables shown in Figures 3 A-B. Moreover, differential expression at many oligonucleotides was highly significant (FDR for the top 50 probes 0.0002 - 0.0078).

Example 4: Determining the 'molecular grade' of DCIS

Clustering of all samples according to expression at the top 100 grade associated probes demonstrated two principal clusters (Figure 4A). In addition, a 'gene expression grade index' (GGI) was calculated for each sample as a standardized representation of combined expression at all 173 grade associated probes (Figure 4B). A GGI cut-off of 0 delineated cluster membership of all samples with the exception of a single PDWA case, and was therefore set to define low and high 'molecular grade' (MG) subgroups.

The low MG subgroup included 21 samples of DCIS and the single LCIS only case in the cohort (Table 2B). In addition, all samples of ADH (n=4), PDWA (n=l) and benign epithelium (n=7) were low MG although corresponding samples of DCIS from all but 2 ADH cases were high MG. Twenty six DCIS samples were designated high MG, as was one sample of LCIS co-existing with DCIS and ADH of low MG.

Low and high DCIS nuclear grade corresponded to MG although intermediate nuclear grade DCIS was divided between low and high MG subgroups. Moreover, despite a statistical correlation between high MG and recognized high-grade features of breast cancer, such as ER and PR negativity (p=0.02, p=0.03) and HER2 and p53 positivity (p=0.006, p=0.01), there was considerable variation within the high MG subgroup with respect to these features.

Example 5: Molecular grade, proliferation, and clinical outcome

Proliferation

Ranking of over-represented functional gene categories amongst the 173 grade associated probes indicated a dominant influence of cell cycle and proliferation. For example, the top three over-represented gene ontology categories were mitotic cell cycle (EASE Score 17 p=1.36 x 10-18), cell cycle (p=3.71 x 10-17) and mitosis (p=4.91 x 10-15) (full list in Figures 3 A-B). In keeping with this observation, there was a strong correlation between GGI and DCIS Ki67 scores (pO.OOOl, r=0.79, Figure 5A) and Ki67 scores were significantly different between DCIS in low and high MG subgroups (pO.OOOl, Figure 5B). Clinical Outcome

Molecular grade is associated with clinical outcome of invasive breast cancer. To verify that the MG classification of DCIS was based on a clinically significant biological difference, MG was compared to clinical outcome in two independent invasive breast cancer cohorts. Gene expression data from van de Vijver et al. 18 (n=294) and Sotiriou et al. 14 (n=179) were used for these analyses and MG was designated according to a GGI calculated from the 172 and 114 available unique grade associated features (representing 170 and 99 unique genes) in each dataset respectively. In both cohorts, metastasis free and overall survival were significantly worse in the high MG subgroup (DFS and OS p<0.00001 van de Vijver et al. Figure 5C&D; DFS p=0.002 Sotiriou et al., Figure 6).

Example 6: Predicting molecular grade

Histopathologic and biomarker features of DCIS were examined to determine whether routinely accessible information could be used to predict MG. DCIS nuclear grade, necrosis, cell polarization, ER, PR, HER2, Ki67, and p53 were considered for inclusion in a classification tree model 19. The tree based on DCIS nuclear grade and Ki67 score (Figure 5E) was an accurate predictor of MG with 44/46 cases (95.7%) correctly assigned; the 10-fold cross-validated error rate for this predictor was 6.52%.

Example 7: DNA copy number profiles of low and high molecular grade DCIS A genomic profile correlate of the grade associated gene expression profile was sought by comparing array-CGH profiles of DCIS associated with grade 1 (n=15) and grade 3 (n=l 1) invasive breast cancer (Figures 7A&B). This showed some striking differences with distinct regions of the genome differentially altered between the two groups. A random forest algorithm used to determine the importance of each CGH probe to the distinction between grade 1 and grade 3 associated DCIS showed particular influence of regions on chromosomes 8, 11 and 17 with smaller contributions from other chromosomes (Figure 7C).

Genomic alteration was directly associated with the grade associated gene expression as evidenced by a positive correlation between GGI and the number of high level DNA amplifications (pO.OOOl , r=0.62) (Figure 7D). Consistent with this, the number of high-level amplifications was significantly different between low and high MG DCIS subgroups (p=0.003, Figure 7E).

In the data presented herein, gene expression profiling was used to determine genes differentially expressed between the intraductal component of grade 1 and grade 3 invasive breast cancer. A binary low/high molecular grade (MG) classification based on expression at grade associated oligonucleotide probes classified benign epithelium and ADH as low grade and importantly divided intermediate nuclear grade DCIS between low and high MG sub-groups. A strong correlation between this classification and survival in two independent invasive breast cancer cohorts provided further verification that DCIS MG was a meaningful indicator of malignant potential.

Discrimination of DCIS into low and high MG clearly demonstrates the feasibility of an informative biological classification of DCIS and omission of the 'intermediate grade' category is a major improvement on other proposed DCIS grading schemes 20, 21. Moreover, the difficulty of arriving at such a classifier by histopathologic assessment is apparent from the diversity of individual pathologic features in each MG subgroup. For example presence of comedo type necrosis, which has been an influential indicator of high grade in many proposed histopathologic DCIS grading schemes 22, was present in 61.9% of cases in the low MG group. Area to area morphologic heterogeneity is a further characteristic of DCIS that has frustrated attempts to devise a robust histopathologic classification 23, 24. This was evident in the current study with two distinct forms of DCIS identified in 14/46 tumors (30.4%). Delineation of individual DCIS sub-types and their separate sampling by laser capture microdissection greatly increased the precision of molecular analysis which proved a major advantage in determination of the MG DCIS classifier. Relatively few gene expression profiling studies of DCIS have been published to date and most have focused on identification of progression-associated genes by comparison of in situ and invasive disease 25-27. In a study by Ma et al., the existence of a grade associated gene expression profile was clearly demonstrated from microdissected samples of ADH, DCIS and invasive cancer 25. There was little overlap between the list of grade associated genes identified by these investigators and our own. However, as Ma et al. adjusted gene expression values according to patient matched normal breast epithelium to emphasize aspects of disease progression, this may not be surprising. In contrast, our list of 173 grade associated probes shows striking concordance with 128 probe sets showing differential expression between grade 1 and grade 3 invasive cancer reported by Sotiriou et al. 14. In this instance there were 50 common elements including 18 of the 35 (51.4%) most significant differentially expressed from our study. Analogous to the findings presented herein in DCIS, Sotiriou et al. reported that a grade associated gene expression profile provided a clinically meaningful resolution of intermediate grade invasive cancer; a finding that has been replicated in a more recent report (28).

Over-expression of genes involved in cellular proliferation in high-grade cases was a dominant feature of the grade associated gene list presented herein and is furthermore consistent with findings from studies that have compared high and low grade cancers from a variety of tissues (29). This feature enabled a simple approach to estimating MG sub-grouping since a classification tree model showed 95.7% of DCIS samples could be accurately assigned to low or high MG subgroups on the basis of nuclear grade and Ki67 score. This combination of routinely accessible markers is similar to invasive breast cancer histopathologic grading and offers a practical clinical reporting alternative to gene expression profiling. However, it does not avoid the principal criticism of histopathologic grading which is its subjective nature and vulnerability to inter-observer variability. Moreover, a commitment to rigorous Ki67 scoring using a cell counting approach is required, although this does offer the additional benefit of buffering variations in nuclear grading under the proposed scheme. Clearly the cut-off Ki67 score for distinguishing high and low MG cases presented in this report was derived from model-fitting to the dataset and determination of an appropriate cut-off for clinical reporting will rely on further studies of larger sample cohorts.

Array-based CGH analysis revealed distinct differences in the character and degree of genomic aberration between DCIS associated with grade 1 or grade 3 invasive breast cancer. In combination with the positive correlation between GGI and high level DNA amplification, this provides further verification that the gene- expression based DCIS classifier reflects true differences in malignant phenotype. It is also consistent with the recent report from Chin et al. showing an association between regions of DNA amplification and both grade and Ki67 score in invasive cancer (30).

Materials and Methods of the Invention

The results reported herein were obtained using the following Materials and Methods:

Patient samples

The study cohort consisted of 46 cases identified from a collection of frozen tumor samples taken from therapeutic excisions of breast cancer performed at Westmead Hospital Australia between 1989 and 1998. The principal inclusion criterion was DCIS identified in frozen tissue sections by morphologic assessment. Cancers with lobular carcinoma in situ (LCIS) only were not included; however, one case judged initially as DCIS but reassigned LCIS following detailed review was retained. All patient information and materials were de-identified and the study was conducted with institutional Human Research Ethics Committee approval. Histopathology review

Histopathologic features of each case were documented by review of archival diagnostic tissue sections using a dual observer protocol. A standardized reporting format was used including invasive cancer grade according to the criteria of Elston and Ellis 12 and individual features of DCIS following reference to criteria listed in the 1997 Consensus Conference Committee report 13. In 14/46 tumors (30.4%), two distinct morphological subtypes of DCIS were identified and separately described.

Determination of tumor ER, PR, HER2 andp53 expression

Results from clinical enzyme immunoassay or immunoperoxidase staining assessment of tumor estrogen (ER) and progesterone receptor (PR) content were used. Tumor HER2 and p53 expression were determined by immunohistochemical staining.

DCIS Ki67 scoring

Immunohistochemical staining for Ki67 was performed on both frozen and paraffin embedded tissue sections using a rabbit polyclonal antibody (Novocastra, Newcastle on Tyne, UK), at 1 : 1000 dilution. The Ki67 score (percentage positive cells) was determined by a single observer (LW) by manual counting of positive and negative in situ carcinoma cell nuclei using the manual tag function in the ImagePro Plus 4.0 software (Image Processing Solutions, MA, USA). In tumors with two morphologic subtypes of DCIS present, a separate Ki67 score was determined for each subtype. There was a high level of concordance between Ki67 scores in frozen and paraffin sections (r=0.70), but to maintain consistency with areas sampled for molecular analysis as far as possible, scores from frozen sections were used except in 3 cases where DCIS had cut out of the frozen tissue block.

Laser capture microdissection and microarray analysis

DCIS foci were isolated from 10mm serial frozen tissue sections by laser capture microdissection (PALM Microlaser Technologies AG, Bernried, Germany). In addition, LCIS and co-existing areas of atypical ductal hyperplasia (ADH), proliferative disease without atypia (PDWA) and benign epithelium were sampled from a proportion of cases. For in situ lesions care was taken to capture pure intraduct cell populations. Benign epithelium samples were lobular tissue collected with intralobular stroma.

Microarray analysis

Oligonucleotide microarrays used for both gene expression and comparative genomic hybridisation (CGH) experiments were the Array-Ready Oligo Set for the Human Genome Version 3.0 (Qiagen Inc., CA, USA) printed onto glass slides. This consisted of 34,580 60-mer probes representing 24,650 genes and 37,123 gene transcripts.

Details of RNA and DNA extraction, amplification, labeling and array hybridization and analysis methods are described below. The complete microarray raw data are available through the Gene Expression Omnibus (GEO) data repository, GEO accession number GSE7882.

Gene expression grade index

Expression at oligonucleotide probes showing differential expression with a false discovery rate (FDR) <0.05 between DCIS associated with grade 1 and grade 3 invasive cancer (n=173) were used to calculate a gene expression grade index (GGI), according to the formula of Sotiriou et al. 14. Note that the GGI is standardized such that the grade 1 associated DCIS cases had a mean GGI of -1 and grade 3 associated DCIS a mean score of 1. Each sample was left out of the standardization process to determine its own GGI.

CGH analysis

Data were segmented 15 to provide a list of discrete regions across the genome each with an associated copy number estimate. The randomForest package 16 in the R statistical programming language was applied to the segmented data to calculate the random forest classification and importance measures. 'High level amplification' was defined as a discrete amplification exceeding a threshold corresponding to the 95% of the distribution of segment means (for all tumors, all segments) from the segmentation algorithm.

Determination of tumour ER, PR, HER2 andp53 expression Results from clinical assessment of tumour estrogen (ER) and progesterone receptor (PR) content were used. For 44 cases these were determined by enzyme immunoassay of tumour cytosol preparations according to a previously described method [I]. For both ER and PR, measures of <10fmol/mg protein were regarded as negative. In the remaining 2 cases ER and PR had been determined by immunoperoxidase staining of frozen tissue or fine needle aspiration biopsy material. Tumor HER2 and p53 expression were determined by immunohistochemical staining of 4um formalin-fixed paraffin embedded tumour sections that had been stored at 4°C. For HER2, mouse monoclonal anti-human cerbB2 clone CBl 1 was used at a concentration of 1 :40 (Novocastra, Newcastle on Tyne UK), and for p53 mouse monoclonal anti-human p53 BP53.12 at 1 : 100 (Zymed, CA, USA). For each stain, antigen retrieval was performed by autoclaving at 121 °C 15 psi in 0.0 IM sodium citrate pH 6. Following treatment with H2O2 (3% v/v) to quench endogenous peroxidase and blocking with normal goat serum (50% v/v in phosphate buffered saline), sections were incubated for one hour in a humidification chamber with the primary antibody in 0.5% triton-X 100 in phosphate buffered saline. Sections were subsequently incubated with biotinylated goat-anti-mouse immunoglobulins (Dako, Glostrup Denmark) and avidin-biotin complex (Zymed, CA, USA). Positive results were revealed by treatment with diaminobenzadine (1 mg/ml (w/v), 0.02% (v/v) H2O2). A positive control section was included in each staining run. For HER2 staining, a negative control section with the primary antibody omitted was included for each case and minor non-specific staining on these sections was taken into account in interpretation of test sections.

HER2 expression was evaluated according to the Dako HerceptTest scoring protocol (accessed at www.dakousa.com) and +++ staining was designated positive. For p53 the proportion of tumour cells positive was estimated and the pattern of staining was noted (extensive or scattered). Cases were designated p53 positive if >50% of cell nuclei were positively stained in an extensive pattern.

RNA and DNA extraction and amplification RNA was extracted from microdissected tissue using the Absolutely RNA

Microprep Kit (Stratagene, CA, USA) and quantitated using the RiboGreen RNA Quantitation Reagent (Molecular Probes, OR, USA). RNA quality was assessed by running RNA isolated from each tissue sample on the Agilent 2100 Bioanalyser (Agilent Technologies, Victoria, Australia). This showed a degree of RNA degradation in all samples but 18S and 28S ribosomal bands were clearly distinguished in 39/46 (84.8%) cases. Depending on tissue sample size, lOOng, 50ng or <20ng of RNA was subject to two rounds of amplification using the RiboAmp RNA Amplification Kit (Arcturus, CA, USA). Two-round amplified Stratagene Universal Human Reference RNA (Stratagene, CA, USA) was the reference in gene expression microarray experiments.

DNA was extracted from laser microdissected material using the QIAamp DNA Micro Kit (Qiagen Inc., CA, USA) and quantitated using the PicoGreen dsDNA Quantitation Reagent (Molecular Probes) both according to the manufacturers instructions. Approximately IOng of DNA was amplified using the GenomiPhi DNA Amplification Kit (Amersham Biosciences, NJ, USA) according to the manufacturer's instructions. Amplified Promega Normal Male DNA (Promega, WI, USA) was used as the reference for array CGH.

RNA labeling Due to the antisense orientation of the amplified RNA samples, a labelling protocol based on a method published by Schlingemann et al [2] was used to generate fluorescent labelled antisense cDNA for hybridisation to sense-oriented oligonucleotide microarrays.

One microgram of amplified RNA was incubated for 5 minutes at room temperature in the presence of 0.5μg of pdN6 (Amersham Biosciences) in a total volume of 5.5μL then chilled on ice. Reverse transcription mixture was added, yielding final concentrations of IX First Strand Buffer (Invitrogen, CA, USA), 1OmM DTT (Invitrogen), 500μM each of dATP, dCTP, dGTP, dTTP (Amersham Biosciences), 2U/μL RNasin ribonuclease inhibitor (Promega) and 1 OU/μL Superscript II reverse transcriptase (Invitrogen) in a total volume of lOμL. The following temperature profile was employed for reverse transcription: 37oC 20 minutes, 42oC 20 minutes, 50oC 10 minutes, 55⁰C 10 minutes, 65oC 15 minutes; reactions were then held at 37oC. RNase H (Promega) digestion (IU per reaction) was carried out for 30 minutes at 37oC, followed by 2 minutes at 95⁰C to degrade enzymes. Samples were chilled to 4oC. cDNA labelling by Klenow fragment and amino-allyl incorporation was performed using the Bioprime Kit with a modified protocol. The lOμL cDNA sample (unpurified) was mixed with 9OuL of Klenow mixture to yield a reaction mixture that contained IX random primer solution (Invitrogen), 650μM each of dATP, dCTP, dGTP, 400μM dTTP (Amersham Biosciences), 260μM amino-allyl dUTP (Sigma, MO, USA) and 1. OU/μL Klenow fragment (Invitrogen). DNA polymerisation was carried out at 37⁰C for 16 hours. cDNA was purified using the QIAquick PCR purification kit (Qiagen Inc., CA, USA). Samples were eluted in 30μL DEPC- water (warmed to 50⁰C) and dried in a speed vac on medium heat. Samples were re-suspended in DEPC-H2O and coupled for 1 hour in the dark to the appropriate NHS ester Cy dye (Cy3 or Cy5, Amersham Biosciences) that had been prepared in DMSO and bicarbonate buffer (pH 8.5). Uncoupled dye was removed by cleaning the samples using the QIAquick PCR purification kit. Cleaned probes were eluted in lOOμL elution buffer (Qiagen). Individual labelling reactions (Cy3 and Cy5) were analysed using spectrophotometry by measuring absorbance at 260, 550 (Cy3) and 65OnM (Cy5). An arbitrary wavelength (75OnM) was also measured to normalize the baselines between samples. The labelled Cy3 and Cy5 probes pairs were combined in an equimolar ratio and concentrated in a speedvac to 18μL. Purified, labelled cDNA was mixed with IX hybridisation buffer (50% formamide, 1 OXSSC, 0.4% SDS) and 50μg Cotl DNA and lOOμg yeast tRNA were added to block repetitive sequence elements. Probes were denatured at 95⁰C for 5 minutes and pre-blocked for 60 minutes at 42⁰C. DNA labeling

Three micrograms of amplified test and reference DNA were directly labelled with either Cy3-dUTP or Cy5-dUTP. Bioprime random primer mix (Invitrogen) was added (IX final concentration) to 3μg of diluted DNA; the samples were vortexed, denatured at 98oC for 5 minutes and cooled on ice. Labelling master mix was added containing 120μM each of dATP, dCTP, dGTP, 60μM dTTP (Amersham Biosciences), 2.5nmol Cy3-dUTP or Cy5-dUTP (Amersham Biosciences), 0.8U/μL Klenow Fragment (Invitrogen). Samples were vortexed and incubated in a thermocycler at 37⁰C for 2 hours. DNA was purified with the QIAquick PCR purification kit and eluted in 30μL Buffer EB (warmed to 50⁰C) (Qiagen). The labelled Cy3 and Cy5 probes pairs were combined and concentrated in a speedvac to 18μL. A hybridisation mix (warmed to 42⁰C) containing 50μg Cotl DNA, lOOμg yeast tRNA, 50% formamide, 5X SSC and 0.2% SDS was added to each sample. The probe was denatured at 95oC for 5 minutes and pre-blocked by incubation at 42oC for 60 minutes in a water bath.

Oligonucleotide microarray hybridisation and analysis

Microarrays were pre-blocked in 5X SSC, 1% BSA and 0.2% SDS at 42oC for 60 minutes. Slides were washed in 2 changes of dH2O and 100% isopropanol for 1 minute each prior to drying by centrifugation. Probes were applied to the oligonucleotide microarrays mounted in a Bio-Micro Maui Hybstation and hybridised for 20 hours at 45oC. Following hybridisation slides were washed in 0.5XSSC, 0.05% SDS; 2 changes of 0.5XSSC and 0. IXSSC for 5 minutes each and then dried by centrifugation. Microarray slides were scanned on an Agilent DNA Microarray Scanner.

Image analyses were using DeArray software (Scanalytics, VA, USA) [3]. The open source statistical environment R (www.r-project.org) was used for statistical analysis. Gene expression microarrays were loess normalized using the limma package [4]. Supervised analysis was performed using the limma package from Bioconductor. Except where otherwise noted, gene lists were defined using a false discovery rate (FDR) of 0.05. Microarray based CGH analysis

Scanning and image analysis of CGH microarrays were as for the gene expression arrays. No normalization was performed. Segmentation provides a means of "smoothing" the data while maintaining its underlying genomic structure. As part of the segmentation process the "center" of the data, representing the modal copy number, was determined by performing multiple segmentation runs using different offsets for the raw data; the offset that produced the smallest number of probes counted as gained or lost was chosen as the "center" of the data. Doing so corresponds to the hypothesis that the most parsimonious "center" of the data is that which produces the fewest number of aberrant probes. Note that we cannot determine the absolute copy number easily from the CGH data themselves and then aberrations are then measured relative to the centre of the data as defined above.

Statistical analysis of pathologic feature and subgroup covariates

Statistical analyses were performed using SPSS for Windows Version 12.0 (SPSS Science, IL, USA). A Mann- Whitney or Student's t-test was used to test for association between categorical and continuous variables. Spearman rank correlation (r) was used to quantify the degree of association between ordered categorical or continuous variables. Odds ratios were calculated using exact logistic regression analysis (LogXact 4, Cytel Software Corporation, MA, USA). A regression tree model [5] was used to study the relationship between gene expression groupings and DCIS histopathological and biological predictor variables; and a 10-fold cross- validation was used to obtain an estimated error rate. The regression tree model was fitted using S-Plus Version 6.2 (Insightful Corporation, WA₅ USA). Kaplan-Meier survival curves were used to illustrate the survival distributions by molecular grade group and log-rank tests were used to test for significant differences in survival between the groups.

Incorporation by Reference The contents of all references, patents, pending patent applications and published patents, cited throughout this application are hereby expressly incorporated by reference.

Equivalents

Variations, modifications, and other implementations of what is described herein will occur to those of ordinary skill in the art without departing from the spirit and scope of the invention and claims herein. All patents, patent publications, international applications, and references are incorporated by reference herein in their entireties.

REFERENCES:

1. Li CI, Daling JR, Malone KE. Age-specific incidence rates of in situ breast carcinomas by histologic type, 1980 to 2001. Cancer Epidemiology, Biomarkers & Prevention 2005;14(4): 1008-11.

2. Ernster VL, Ballard-Barbash R, Barlow WE, et al. Detection of ductal carcinoma in situ in women undergoing screening mammography. Journal of the National Cancer Institute 2002;94(20): 1546-54.

3. Emster VL, Barclay J, Kerlikowske K, Wilkie H, Ballard-Barbash R. Mortality among women with ductal carcinoma in situ of the breast in the population-based surveillance, epidemiology and end results program. Archives of Internal Medicine 2000;160(7):953-8.

4. Duffy SW, Agbaje O, Tabar L, et al. Overdiagnosis and overtreatment of breast cancer: estimates of overdiagnosis from two trials of mammographic screening for breast cancer. Breast Cancer Research 2005;7(6):258-65.

5. Evans AJ, Blanks RG. Should breast screening programmes limit their detection of ductal carcinoma in situ? Clinical Radiology 2002;57(l 2): 1086-9.

6. Mokbel K, Cutuli B. Heterogeneity of ductal carcinoma in situ and its effects on management. Lancet Oncology 2006;7(9):756-65.

7. Weng EY, Juillard GJ, Parker RG, Chang HR, Gornbein JA. Outcomes and factors impacting local recurrence of ductal carcinoma in situ. Cancer 2000;88(7):1643- 9.

8. Silverstein MJ, Lagios MD, Groshen S, et al. The influence of margin width on local control of ductal carcinoma in situ of the breast. New England Journal of Medicine 1999;340(19):1455-61. 9. Schatzkin A, Gail M. The promise and peril of surrogate end points in cancer research. Nature Reviews Cancer 2002;2(l): 19-27.

10. Leong AS, Sormunen RT, Vinyuvat S, Hamdani RW, Suthipintawong C. Biologic markers in ductal carcinoma in situ and concurrent infiltrating carcinoma. A comparison of eight contemporary grading systems. American Journal of Clinical Pathology 2001;115(5):709-18.

11. Webster LR, Bilous AM, Willis L, et al. Histopathologic indicators of breast cancer biology: insights from population mammographic screening. British Journal of Cancer 2005;92(8): 1366-71.

12. Elston CW, Ellis IO. Pathological prognostic factors in breast cancer. I. The value of histological grade in breast cancer: experience from a large study with long- term follow-up. Histopathology 1991;19(5):403-10.

13. Consensus Conference Committee. Consensus Conference on the classification of ductal carcinoma in situ. The Consensus Conference Committee. Cancer 1997;80(9): 1798-802.

14. Sotiriou C, Wirapati P, Loi S, et al. Gene expression profiling in breast cancer: understanding the molecular basis of histologic grade to improve prognosis. Journal of the National Cancer Institute 2006;98(4):262-72.

15. Lipson D, Aumann Y, Ben-Dor A, Linial N, Yakhini Z. Efficient calculation of interval scores for DNA copy number data analysis. Journal of Computational Biology 2006;13(2):215-28.

16. Liaw A, Wiener M. Classification and regression by randomForest. R News 2002;2(3): 18-22. 17. Hosack DA, Dennis G, Sherman BT, Lane HC, Lempicki RA. Identifying biological themes within lists of genes with EASE. Genome Biology 2003;4(10):R70.

18. van de Vijver MJ, He YD, van't Veer LJ, et al. A gene-expression signature as a predictor of survival in breast cancer. New England Journal of Medicine 2002;347(25): 1999-2009.

19. Brieman L, Friedman JH, Olshen RA, Stone CJ. Classification and regression trees. Belmont, CA: Wadsworth; 1984.

20. Holland R, Peterse JL, Millis RR, et al. Ductal carcinoma in situ: a proposal for a new classification. Seminars in Diagnostic Pathology 1994;11(3):167-80.

21. Silverstein MJ, Poller DN, Waisman JR, et al. Prognostic classification of breast ductal carcinoma-in-situ. Lancet 1995;345(8958): 1154-7.

22. Shoker BS, Sloane JP. DCIS grading schemes and clinical implications. Histopathology 1999;35(5):393-400.

23. Quinn CM, Ostrowski JL. Cytological and architectural heterogeneity in ductal carcinoma in situ of the breast. Journal of Clinical Pathology 1997;50(7):596-9.

24. Harrison M, Coyne JD, Gorey T, Dervan PA. Comparison of cytomorphological and architectural heterogeneity in mammographically-detected ductal carcinoma in situ.

Histopathology 1996;28(5):445-50.

25. Ma XJ, Salunga R, Tuggle JT, et al. Gene expression profiles of human breast cancer progression. Proceedings of the National Academy of Sciences of the United States of America 2003; 100(10):5974-9. 26. Nishidate T, Katagiri T, Lin ML, et al. Genome-wide gene-expression profiles of breast-cancer cells purified with laser microbeam microdissection: identification of genes associated with progression and metastasis. International Journal of Oncology 2004;25(4):797-819.

27_; Schuetz CS, Bonin M, Clare SE, et al. Progression-specific genes identified by expression profiling of matched ductal carcinomas in situ and invasive breast tumors, combining laser capture microdissection and oligonucleotide microarray analysis. Cancer Research 2006;66(10):5278-86.

28. Ivshina AV, George J, Senko O, et al. Genetic reclassification of histologic grade delineates new clinical subtypes of breast cancer. Cancer Research 2006;66(21): 10292-301.

29. Rhodes DR, Yu J, Shanker K, et al. Large-scale meta-analysis of cancer microarray data identifies common transcriptional profiles of neoplastic transformation and progression. Proceedings of the National Academy of Sciences of the United States of America 2004;101(25):9309-14.

30. Chin K, DeVries S, Fridlyand J, et al. Genomic and transcriptional aberrations linked to breast cancer pathophysiologies. Cancer Cell 2006;10(6):529-41.

Claims

What is claimed is

1. A method for determining the molecular grade of a cancer comprising: identifying genes differentially expressed in a cancer sample and a control sample by determining gene expression and gene copy number by molecular profiling; and classifying the cancerous tissue by one or more of gene expression or gene copy number; thereby determining the molecular grade of the cancer.

2. A method for determining the molecular grade of a cancer in a subject comprising: identifying genes differentially expressed in a cancer sample and a control sample by determining gene expression and gene copy number by molecular profiling; and classifying the cancerous tissue by one or more of gene expression or gene copy number; thereby determining the molecular grade of a cancer in a subject.

3. The method of claim 1 or 2, further comprising obtaining one or more samples of a cancer tissue.

4. The method of claim 1 or 2, wherein determining the molecular grade of a cancer further comprises classifying the cancer as low molecular grade or high molecular grade.

5. The method of claim 1 or 2, wherein classifying the cancer by one or more of gene expression or gene copy number comprises calculating a gene expression grade index.

6. The method of claim 5, wherein the gene expression grade index is calculated using the gene expression pattern of the cancer sample versus the gene expression of the control sample.

7. The method of claim 6, wherein a high gene expression grade index corresponds to a high molecular grade.

8. The method of claim 6, wherein a low gene expression grade index corresponds to a low molecular grade.

9. The method of claim 7, wherein the high molecular grade correlates with a low level of survival compared to a control.

10. The method of claim 8, wherein the low molecular grade correlates with a high level of survival compared to a control.

11. The method of claim 7, wherein high molecular grade correlates with one or more of a high level of DNA aberrations or a high level of regions of high level amplification compared to a control.

12. The method of claim 8, wherein low molecular grade correlates with one or more of a low level of DNA aberrations or a low level of regions of high level amplification compared to a control.

13. The method of claim 1 or 2, wherein the molecular grade is used to determine the prognosis of a patient.

14. The method of claim 1 or 2, wherein gene expression and gene copy number are determined by molecular analysis.

15. The method of claim 14, wherein the molecular analysis is selected from the group consisting of: Polymerase Chain Reaction (PCR), quantitative real-time reverse trascriptase PCR, targeted differential display, serial analysis of gene expression (SAGE), microarray analysis, multi-analyte profiling beads, and array-based comparative genomic hybridization (CGH).

16. The method of claim 1 or 2, wherein the cancer samples are selected from invasive cancer, intraductal carcinoma, proliferative disease without atypia, atypical ductal hyperplasia and benign tissue.

17. The method of claim 16, wherein the cancer samples are from breast cancer.

18. The method of claim 17, wherein the breast cancer is ductal carcinoma in situ (DCIS).

19. The method of claim 15, wherein CGH generates a genomic profile.

20. The method of claim 19, wherein the genomic profile is used to classify the cancer into a low molecular grade or a high molecular grade.

21. The method of claim 15, wherein the molecular analysis generates a gene amplification index (AI).

22. The method of claim 21, wherein the AI is correlated with molecular grade.

23. The method of claim 22, wherein the AI is measured by the number of high level regions of DNA amplification over a threshold.

24. The method of claim 23, wherein the number of high level regions of DNA amplification over a threshold greater than two is correlated with a high molecular grade.

25. The method of claim 1 or 2, wherein genes differentially expressed in cancer and control samples are selected from one or more of the genes listed in Figures 3A-B.

26. The method of claim 1 or 2, wherein the genes that are differentially expressed comprise one or more genes involved in cell proliferation.

27. The method of claim 26, wherein the genes related to cell proliferation are selected from one or more of the genes listed in Figures 3A-B.

28. A method of determining the molecular grade of a cancer comprising: profiling molecularly to determine one or more of gene expression or gene copy number in a test sample; comparing a test expression profile to a control expression profile; and determining a molecular grade that corresponds to gene expression and gene copy number; wherein the molecular grade of the cancer is determined.

29. The method of claim 28, further comprising the step of providing a test sample of cancerous tissue.

30. A diagnostic method to determine the molecular grade of cancer using nuclear grade and Ki67 score comprising: determining a nuclear grade and a Ki67 score in a carcinoma sample; and determining a molecular grade of cancer using nuclear grade and a Ki67 score; and wherein the molecular grade of cancer is determined by using nuclear grade and Ki67 score.

31. The method of claim 30, further comprising providing a sample of breast cancer tissue from a subject.

32. The method of claim 30, wherein the sample is primary invasive breast cancer tissue.

33. The method of determining the molecular grade of a cancer according to claim 28, further comprising the steps of contacting the sample with a candidate compound and determining the cell proliferation profile of the sample.

34. A method of identifying a therapeutic compound, the method comprising: determining the molecular grade of cancer according to the method of claim 28; contacting the sample with a candidate compound; and determining the cell proliferation profile of the sample; wherein a therapeutic compound is identified.

35. A method of predicting the prognosis of a breast cancer patient comprising: determining a nuclear grade and a Ki67 score in a primary invasive breast carcinoma sample and a control sample; and comparing the nuclear grade and Ki67 score from the subject sample to the control sample; wherein the molecular grade and the Ki67 score higher in the carcinoma sample than in the control sample is an indication that the prognosis of the test patient is poor.

36. The method of claim 35, further comprising providing a sample of primary invasive breast cancer tissue from a subject.

37. The method of claim 36, wherein the breast cancer is intraductal carcinoma.

38. The method of claim 37, wherein the intraductal carcinoma is DCIS.

39. A gene expression profile associated with high molecular grade DCIS comprising one or more of the genes listed in Figures 3A-B.

40. A single stranded nucleic acid probe comprising: (a) the nucleotide sequence of a tag selected from those listed in Figures 3A-B; or (b) the complement of the nucleotide sequence.

41. A microarray comprising a substrate having one or more addresses, wherein each address comprises a capture probe comprising a nucleic acid sequence comprising a tag nucleotide sequence selected from the sequences listed in Figures 3A-B.

42. A kit comprising at least 10 probes, each probe comprising a nucleic acid sequence comprising a tag nucleotide sequence selected from those listed in Figures 3A-B.

43. A method of identifying the molecular grade of a DCIS sample comprising: determining the molecular grade of a sample; and comparing the molecular grade of a sample with a reference profile; and thereby identifying the grade of the DCIS sample.

44. The method of claim 43, further comprising providing a test sample of DCIS tissue.

45. The method of claim 43, further comprising assigning a prognosis to the sample.