CROSS REFERENCE TO PRIOR APPLICATION

[0001] The present application claims priority under the Paris Convention to US

Application number 62/099,028, filed on December 31 , 2014. The entire contents of such prior application is incorporated herein by reference.

FIELD OF THE DESCRIPTION

[0002] The present description relates generally to the field of prognosing cancer. More particularly, the description relates to methods for prognosing prostate cancer based on the measurement of Ep-ICD.

STATEMENT REGARDING SEQUENCE LISTING

[0003] A Sequence Listing associated with this application is provided in ASCII format, submitted electronically, and is hereby incorporated by reference into the present specification. The text file containing the Sequence listing is titled "Sequence listing 78602- 71.txt" , was created on December 24, 2015, and is 3 kilobytes in size.

BACKGROUND

[0004] Prostate cancer (PCa) is the second most common cancer in the world, with an estimated 900,000 cases and 258,000 deaths in 2008 (Jemal A., CA Cancer J Clin, 201 1 , 61 :69-90). The United States will have an estimated 239,000 new cases and 29,700 deaths in 2013 alone (Siegel R. et al., CA Cancer J Clin, 2013, 61 :69-90). The ideal treatment for PCa continues to be a challenge for oncologists worldwide. There are curative treatments for PCa however, these are associated with increased patient morbidity. Further, some of prostate cancer patients are over-treated while others are under-treated. The incidence of PCa diagnosis continues to rise with increased use of the screening tool prostate specific antigen (PSA) followed by biopsies. The increase in PCa diagnoses has resulted in an increase in diagnosis of indolent tumors, which are managed by active surveillance wherein patients are biopsied periodically to monitor disease progression (Cooperberg M.R., et al., J Urol, 2007, 178:S14-19). Management of PCa relies heavily on a variety of factors, namely: physical examination; PSA level; Gleason score (GS) at biopsy (Carter et al. J. Clin. Oncol. 2012, 30:4294-4296); clinical stage; and tumor extent, invasion, and imaging. Even with these clinical factors, the prognosis of PCa is difficult to determine. Usually tumor size and microscopic appearance would mandate the patient's treatment. Some patients with good prognosis would receive the same treatment as patients with poor prognosis, leading to under- or over- treatment. Furthermore, some PCa cases have diagnostic uncertainty where the pathology reports state "suspicious for cancer" (Epstein Jl. et al., Mod Pathol, 2004, 17:307-315). The patients with these diagnoses are usually sent for a repeat biopsy, causing additional distress and exposing the patients to the inherent risks of biopsy (e.g., risk of infection).

[0005] Epithelial cell adhesion molecule (EpCAM) has been widely explored as an epithelial cancer antigen (Munz et al. 2009, Cancer Res 69: 5627-5629). It is a glycosylated, 30- to 40-kDa type I membrane protein, expressed in several human epithelial tissues, and overexpressed in cancers as well as in progenitors and stem cells (Munz et al. 2009, Mukherjee et al. 2009; Am J Pathol 175: 2277-2287; Carpenter & Red Brewer 2009, Cancer Cell 15: 165-166; Schnell et al. 2013, Biochim Biophys Acta 1828: 1989-2001 ; Ni et al. 2012, Cancer Metastasis Rev 31 : 779-791). EpCAM is comprised of an extracellular domain (EpEX) with epidermal growth factor (EGF) and thyroglobulin repeat-like domains, a single transmembrane domain, and a short 26-amino acid intracellular domain called Ep-ICD. In normal cells, the full length EpCAM protein is sequestered in tight junctions and therefore note easily accessible to antibodies. In cancer cells, EpCAM is homogeneously distributed on the surface of cancer cells. EpCAM has been explored as a surface-binding site for therapeutic antibodies.

[0006] EpCAM is expressed in a majority of human epithelial cancers, including breast, colon, gastric, head and neck, prostate, pancreas, ovarian and lung cancer and is one of the most widely investigated proteins for its diagnostic and therapeutic potential (Spizzo et al. 2004, Breast Cancer Res Treat 86: 207-213; Went et al. 2004, Hum Pathol 35: 122-128; Saadatmand et al. 2013, Br J Surg 100: 252-260; Soysal et al. 2013, Br J Cancer 108: 1480- 1487). Increased EpCAM expression is a marker of poor prognosis in breast and gall bladder cancers (Gastl et al. 2000, Lancet 356: 1981 -1982; Varga et al. 2004, Clin Cancer Res 10: 3131 -3136), while it is associated with favorable prognosis in colorectal and gastric cancers (Songun et al. 2005, Br J Cancer 92: 1767-1772; Went et al. 2006, Br J Cancer 94: 128-135; Ensinger et al. 2006, J Immunother 29: 569-573). This paradoxical association of EpCAM expression with prognosis in different cancers may be explained by functional studies of EpCAM biology using in vitro and in vivo cancer models (van der Gun et al. 2010, Carcinogenesis 31 : 1913-1921). Taken together these studies suggest that the impact of EpCAM expression in human cancers is likely to be context-dependent. An EpCAM expression-based assay is the only FDA-approved test widely used to detect circulating tumor cells in breast cancer (Cristofanilli et al. 2004, N Engl J Med 351 : 781 -791). [0007] EpCAM-targeted molecular therapies are being studied for several cancers including breast, ovarian, gastric and lung cancer (Baeuerle & Gires 2007, Br. J Cancer 96: 417-423; Simon et al. 2013, Expert Opin Drug Deliv 10: 451 -468) EpCAM expression has been used to predict response to anti-EpCAM antibodies in breast cancer patients (Baeuerle & Gires 2007, Schmidt et al. 2005, Annals of Oncology 23: 2306-2313; Schmidt et al. 2010, Annals of Oncology 21 : 275-282). Clinical trials of anti-EpCAM antibodies targeting the EpEX domain have shown limited efficacy in cancer therapy and its prognostic potential for determining survival of cancer patients remains unclear (Riethmuller et al. 1998, J Clin Oncol 16: 1788-1794; Fields et al. 2009, J Clin Oncol 27: 1941 -1947; Gires & Bauerle et al. 2010, J Clin Oncol 28: e239-240; author reply e241 -232; Schmoll & Arnold 2009, J Clin Oncol 27: 1926-1929; Maetzel et al. 2009, Nat Cell Biol 1 1 : 162-171). This might be explained by the recently unraveled mode of activation of EpCAM oncogenic signaling by proteolysis and the potential of Ep-ICD in triggering more aggressive oncogenesis (Maetzel et al. Nat. Cell Biol. 2009, 1 1 : 162-171). Regulated intra-membrane proteolysis of EpCAM results in shedding of EpEX and release of Ep-ICD into the cytoplasm, nuclear translocation and activation of oncogenic signaling (Carpenter s Brewer, Cancer Cell, 2009, 15: 156-166.

[0008] The present inventors earlier reported on Ep-ICD and EpEX expression analysis and potential for use as diagnostic markers in ten epithelial cancers, including prostate cancer (US Patent Publication No. 201 1/0275530). With respect to diagnosis of prostate cancer, nuclear and cytoplasmic Ep-ICD immunopositivity was observed in 40 of 49 prostate tumors; nuclear Ep-ICD was observed in 2 of 9 normal prostate tissues and cytoplasmic Ep- ICD in 1 of 9 normal prostate tissues (US Patent Publication No. 201 1/0275530).

[0009] European Patent Application No. EP 2772758 examined EpEX and Ep-ICD expression in various malignant tissues using microarrays and found that approximately 10% of cancerous prostate tissues exhibited EpEX positive/Ep-ICD negative expression (FIG. 5). Subcellular localization of Ep-ICD (i.e., identification of cytoplasmic and/or nuclear localization) was not disclosed in prostate cancer samples in European Patent Application No. 2,772,758. Further, prostate cancer samples were not identified by stage and/or Gleason score or otherwise categorized to indicate whether the prostate tissue samples included in the microarray comprised aggressive or non-aggressive cancers. Accordingly, the use of EpEX and/or Ep-ICD as markers of prostate cancer prognosis could not be predicted based on the disclosure of EP 2772758.

[0010] Fong et al. (J. Clinical Pathol. 2014, 67:408-414) examined EpEX and Ep-ICD expression in various malignant tissues using multi-tissue microarray (TMA) series and found that approximately 51 % of prostate cancer tissue samples (n=45) analyzed exhibited strong or moderate EpEX expression and approximately 60% of those prostate cancer tissue samples exhibited weak or negative Ep-ICD expression (Table 1). Unlike US Patent Publication No. 201 1/0275530, Fong et al. were unable to detect nuclear Ep-ICD staining using immunohistochemistry (IHC) in any of the tissues examined, including the prostate cancer tissues. Any detection of Ep-ICD by Fong et al. was restricted to Ep-ICD identified in the cytoplasm of epithelial cells. Detection of cytoplasmic Ep-ICD in Fong et al. was achieved using mouse monoclonal antibody clone 9-2 (HVD List Science GmbH). The 45 prostate cancer samples examined by Fong et al. were not identified by stage and/or Gleason score or otherwise categorized to indicate whether the prostate tissue samples included in the microarray comprised aggressive or non-aggressive cancers. Accordingly, the use of EpEX and/or Ep-ICD as markers of prostate cancer prognosis could not be predicted based on the disclosure of Fong et al.

[0011] Methods for use in prostate cancer prognosis are desirable, particularly in those patients having Gleason Score (GS) of less than 7.

SUMMARY OF THE DESCRIPTION

[0012] In one aspect, there is provided a method for prognosing prostate cancer in a subject, the method comprising:

(a) measuring an amount of nuclear Ep-ICD polypeptide in a biological sample from the subject; and,

(b) identifying a poor prognosis of prostate cancer in the subject if:

(i) the measured amount is lower than the amount of nuclear Ep-ICD polypeptide in a non-aggressive prostate cancer sample; or,

(ii) the measured amount is equal to or lower than the amount of nuclear Ep-ICD polypeptide in an aggressive prostate cancer sample.

[0013] In another aspect, there is provided a method for detecting abnormal subcellular localization of Ep-ICD in a prostate tissue sample obtained from a subject, the prostate tissue sample comprising cells with each cell having a nucleus and cytoplasm, the method comprising:

(A) determining a nuclear Ep-ICD score and a cytoplasmic Ep-ICD score in the sample comprising:

(i) contacting the sample with a binding agent that specifically binds to Ep-ICD or a portion thereof and a detectable label for detecting binding of the binding agent to Ep-ICD, wherein the detectable label emits a detectable signal upon binding of the binding agent to Ep-ICD;

(ii) measuring:

(a) a first percentage, comprising the percentage of cells in the sample having Ep-ICD in the nucleus bound to the binding agent, and assigning a first quantitative score to the first percentage according to a first scale;

(b) a second percentage, comprising the percentage of cells in the sample having Ep-ICD in the cytoplasm bound to the binding agent, and assigning a second quantitative score to the second percentage according to the first scale;

(iii) measuring:

(a) a first intensity, comprising the intensity of the signal emitted in the nucleus by the label, and assigning a first qualitative score to the first intensity according to a second scale;

(b) a second intensity, comprising the intensity of the signal emitted in the cytoplasm by the label and assigning a second qualitative score to the second intensity according to the second scale;

(iv) calculating the nuclear Ep-ICD score and the cytoplasmic Ep-ICD score by:

(a) adding the first quantitative and qualitative scores to generate the nuclear Ep-ICD score;

(b) adding the second quantitative and qualitative scores to generate the cytoplasmic Ep-ICD score;

(B) calculating an Ep-ICD Subcellular Localization Index (ESLI) value for the sample, the ESLI value being a combination of the nuclear Ep-ICD score and the cytoplasmic Ep-ICD score;

(C) comparing the calculated ESLI value to a reference value, wherein the reference value is:

(i) an ESLI value indicative of a non-aggressive cancerous prostate tissue; or

(ii) an ESLI value indicative of an aggressive prostate cancer; and

(D) detecting abnormal subcellular localization of Ep-ICD in the prostate tissue sample when the calculated ESLI value of the sample is less than the reference value of (C)(i) or is less than or equal to the reference value of (C)(ii). [0014] In one aspect, the biological sample is obtained from a tumor tissue.

[0015] In one aspect, the binding agent is an antibody.

[0016] In one aspect, one or more of the first and second percentages and the first and second intensities are obtained using immunohistochemical (IHC) analysis.

[0017] In one aspect, one or more of the aforementioned steps, or the other steps described herein, can be performed manually or in an automated manner using hardware and/or software elements as described.

DESCRIPTION OF THE DRAWINGS

[0018] The features of certain embodiments will become more apparent in the following detailed description in which reference is made to the appended figures wherein:

[0019] Figure 1 depicts a schematic of the study design used in Example 1.

[0020] Figure 2 depicts representative photomicrographs showing immunohistochemical analysis of Ep-ICD and EpEX in prostate tissues. Panel (a) shows nuclear/cytoplasmic Ep- ICD immunostaining in: (I) normal, (II) benign prostatic hyperplasia (BPH) and prostate adenocarcinoma, (III) Gleason score (GS) < 7, (IV) GS = 7, and (V) GS > 7. Panel (b) shows membrane EpEX immunostaining in: (I) normal (II) BPH, and prostate adenocarcinoma, (III) GS < 7, (IV) GS = 7, and (V) Gleason score > 7. Arrows show nuclear (N) Ep-ICD staining and membrane (M) EpEX staining. (Original Magnification x 400).

[0021] Figure 3 depicts scatter plots showing the nuclear Ep-ICD (left), cytoplasmic Ep- ICD (center) and membrane EpEX (right) of nuclear Ep-ICD (left), cytoplasmic Ep-ICD (center) and membrane EpEX (right) in normal prostate, benign prostatic hyperplasia, prostate intra-epithelial neoplasia and prostate cancer tissues.

[0022] Figure 4 depicts box plot analysis of nuclear Ep-ICD (left), cytoplasmic Ep-ICD (center) and membrane EpEX (right) nuclear Ep-ICD (left), cytoplasmic Ep-ICD (center) and membrane EpEX (right) in prostate cancers with respect to GS.

[0023] Figures 5A-5C depict Kaplan Meier survival analysis based on manual measurement of (i) nuclear Ep-ICD and (ii) membrane EpEX in prostate cancer patients. Kaplan-Maier analysis was performed dividing the patients as biomarker (Ep-ICD nuclear, Membrane EpEX) positive (green) and negative (blue). These were then further stratified by GS, wherein FIG. 5A depicts all PCa cases; FIG. 5B depicts PCa cases comprising a GS of <7; and FIG. 5C depicts PCa cases comprising a GS = to 7. [0024] Figures 6A-6D depict disease free survival in prostate cancer patients with respect to Ep-ICD Subcellular Localization Index (ESLI) scores and Gleason scores (GS). FIG. 6A shows disease free survival in prostate cancer patients with respect to ESLI positivity (> 3) and negativity (<3). FIG. 6B shows disease free survival in prostate cancer patients with respect to GS positivity (< 7) and negativity (> 7). FIG. 6C shows disease free survival in prostate cancer patients with respect to ESLI and GS positivity and negativity. FIG. 6D shows disease free survival in prostate cancer patients having a GS of < 7 with respect to ESLI positivity and negativity.

[0025] Figure 7 depicts representative images of stained prostate cancer epithelial tissues obtained using a NanoZoomer™ scanner and Visiopharm™ software.

[0026] Figures 8A-8C depict Kaplan Meier survival analysis in prostate cancer patients based on automated measurement of subcellular localization of Ep-ICD. FIG. 8A shows survival probability in prostate cancer patients based on measurement of nuclear Ep-ICD (intensity and % positivity). FIG. 8B shows survival probability in prostate cancer patients based on measurement of cytoplasmic Ep-ICD (intensity). FIG. 8C shows survival probability in prostate cancer patients based on calculation of ESLI.

DETAILED DESCRIPTION

[0027] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0028] Definitions

[0029] The term "EpCAM" as used herein, refers to the epithelial cell adhesion molecule having the amino acid sequence set forth in SEQ ID NO: 1 (SEQ ID NO: 1 corresponds to Genbank Accession No. NP_002345). EpCAM comprises an extracellular domain, referred to herein as "EpEX", that is 265 amino acids in length (amino acids 1-265 in SEQ ID NO: 1), a single transmembrane domain that is 23 amino acids in length (amino acids 266-288 in SEQ ID NO. 1), and an intracellular domain, referred to herein as Έρ-ICD", that is 26 amino acids in length (amino acids 289-314 in SEQ ID NO. 1).

[0030] The term "aggressive" as used herein, refers to a type of cancer that forms, grows and/or spreads more quickly than a "non-aggressive" cancer. For example, a subject having an aggressive prostate cancer may have an expected disease free survival (DFS) time that is less than a subject having a non-aggressive prostate cancer. The DFS is the time period until disease recurrence, metastasis and/or death. [0031] The term "score" as used herein, refers to a rating or grade provided to a result, wherein the rating or grade is measured on a scale that comprises minimum and maximum possible scores for a result.

[0032] The terms "algorithm" and "ESLI algorithm" as used herein, refer to mathematical formulae for numerically characterizing Ep-ICD sub-cellular expression by determining a value (e.g., an Ep-ICD Subcellular Localization Index "ESLI" value). The algorithms are defined further herein.

[0033] "Prognosis", as used herein, refers to a prediction of the probable course and/or outcome of a disease. For example, a poor prognosis may predict a reduced DFS in a patient relative to a patient having a good prognosis. For example, a poor prognosis would predict a DFS of less than about five years and a favourable or good prognosis would predict a DFS of more than about five years.

[0034] As described herein, the inventors have found that prostate cancer patients having a poor prognosis have prostate tissue comprising a decreased amount of Ep-ICD, in particular decreased nuclear Ep-ICD, relative to prostate cancer patients having a favorable prognosis. Methods for prognosing prostate cancer comprising one or more of detecting, measuring, scoring and evaluating subcellular localization of Ep-ICD are discussed further below. In one aspect, the invention provides a numerical scoring method to quantify prognosis, such scoring method is referred to herein as the Ep-ICD Subcellular Localization Index (ESLI). The use of an ESLI value in prognosing prostate cancer is discussed further below. Patient Gleason score status is also considered in view of Ep-ICD subcellular localization measurements, scores and index values. The inventors have also found that subcellular localization of Ep-ICD, in particular nuclear Ep-ICD, and optionally

measurements of membranous EpEX can be used to differentiate normal prostate tissue from cancerous prostate cancer tissue.

[0035] Methods for the Prognosis of Prostate Cancer

[0036] The present disclosure is generally directed to a method for prognosing cancer, in particular prostate cancer, in a subject. The subject, also referred to herein as a patient, may be a mammal that is afflicted with, suspected of having, at risk for pre-disposal to, or being screened for prostate cancer. In a preferred embodiment, the subject is a human.

[0037] In one embodiment, an amount of nuclear and optionally cytoplasmic Ep-ICD is measured in a biological sample from the subject. The biological sample comprises prostate epithelial cells. In a preferred embodiment, the biological sample comprises prostate tissue. In a particularly preferred embodiment, the biological sample comprises prostate cancer tumor cells.

[0038] Measurement of Ep-ICD may be quantitative and/or qualitative. In one embodiment, measurement may be achieved by contacting the biological sample with a first binding agent and measuring in one or more nuclei and optionally cytoplasms of the biological sample the amount of the first binding agent bound to Ep-ICD. In one embodiment, an amount of membranous EpEX is measured in a biological sample from the subject.

Measurement of EpEX may be achieved by contacting the biological sample with a second binding agent and measuring in one or more membranes of the biological sample the amount of the second binding agent bound to EpEX. A binding agent refers to a substance that specifically binds to a specific polypeptide. A binding agent may be, for example, an antibody, a ribosome, RNA, DNA, a polypeptide or an aptamer. For example, an antibody specifically reactive with Ep-ICD may be used to detect Ep-ICD in the biological sample and may be used to determine the subcellular localization of Ep-ICD (i.e., nuclear or

cytoplasmic). General techniques for in vitro detection of antigens in samples are well known in the art. In a preferred embodiment, an Ep-ICD-specific antibody is used to detect Ep-ICD. In a preferred embodiment, an Ep-ICD-specific antibody is used to detect Ep-ICD. The detection of nuclear proteins in general and Ep-ICD in particular can be challenging because antigen retrieval conditions need to be carefully optimized to expose the antigenic epitopes of nuclear proteins such that the antibody can bind to the protein localized in the nucleus. In the context of Ep-ICD this is particularly challenging because the intracellular domain is a short 26 amino acid sequence (amino acids 289-314 in SEQ ID NO. 1).

Optimization of conditions for detection of nuclear Ep-ICD is discussed in Example 1 below. In a preferred embodiment, an EpEX-specific antibody is used to detect EpEX.

[0039] Binding agents specific for Ep-ICD or EpEX may be labelled with a detectable substance which facilitates identification in biological samples based upon the presence of the detectable substance. Examples of detectable substances include, but are not limited to, the following: radioisotopes, fluorescent labels, luminescent labels, bioluminescent labels, enzymatic labels, biotinyl groups, and predetermined polypeptide epitopes recognized by a secondary reporter. Binding agents may also be coupled to electron dense substances, such as ferritin or colloidal gold, which are readily visualized by electron microscopy.

[0040] Indirect methods may also be employed in which a primary antigen-antibody reaction is amplified by the introduction of a second antibody, having specificity for the antibody reactive against an epitope of the target polypeptide. For example, if the antibody having specificity against an Ep-ICD polypeptide is a rabbit IgG antibody, the second antibody may be goat anti-rabbit IgG, Fc fragment specific antibody labelled with a detectable substance, as described herein.

[0041] Methods for conjugating or labelling the antibodies discussed above may be readily accomplished by one of ordinary skill in the art.

[0042] Quantitative and/or qualitative measurement of Ep-ICD and/or EpEX may be automated or it may be done manually.

[0043] One example of manual quantitative and qualitative measurement of Ep-ICD, wherein scores are assigned to nuclear and cytoplasmic Ep-ICD quantitative and qualitative measurements, is described further below.

[0044] One example of automated quantitative and qualitative measurement of Ep-ICD using Visiopharm™ software is described further below.

[0045] In one embodiment, once an amount of nuclear Ep-ICD is measured in a biological sample from the subject, the measured amount is compared to a control and a poor or favorable prognosis is made based on results of the comparison.

[0046] In one embodiment, the control is an amount of nuclear Ep-ICD in a non- aggressive cancerous biological sample, for example, a non-aggressive cancerous prostate tissue or a sample comprising non-aggressive cancerous prostate epithelial cells. In this case, a lower measured amount of nuclear Ep-ICD in the biological sample relative to the control indicates a poor prognosis of prostate cancer and an equal or higher measured amount of nuclear Ep-ICD in the biological sample relative to the control indicates a favorable prognosis.

[0047] In one embodiment, the control is an amount of nuclear Ep-ICD in an aggressive cancerous biological sample, for example, an aggressive prostate tumor or a sample comprising aggressive cancerous prostate epithelial cells. In this case, an equal or lower measured amount of nuclear Ep-ICD in the biological sample relative to the control indicates a poor prognosis of prostate cancer.

[0048] In one embodiment, the method for prognosing prostate cancer may be applied to a subject having a Gleason score (GS) of less than 7. The Gleason Grading System is known in the art and is used to help evaluate the prognosis of men with prostate cancer. A GS may be assigned to a prostate cancer sample based upon microscopic characterization of Gleason patterns (i.e., patterns 1 -5) known in the art. Cancers with higher Gleason scores are thought to be more aggressive and have a worse prognosis than those with lower Gleason scores.

[0049] In one embodiment, a method for prognosing prostate cancer in a subject is provided, wherein the method comprises determining quantitative and qualitative scores corresponding to the amounts of nuclear Ep-ICD and cytoplasmic Ep-ICD. In this method, the quantitative and qualitative nuclear and cytoplasmic Ep-ICD scores are calculated and compared to control values for determining the poor prognosis of prostate cancer in a subject.

[0050] In an aspect of the embodiment, the method may further comprise a step of calculating an Ep-ICD Subcellular Localization Index (ESLI) value for a sample obtained from the subject. The ESLI value, as discussed further below, offers a unique quantitative means of prognosing prostate cancer in a subject.

[0051] The present inventors developed the ESLI algorithms by: i) examining subcellular localization of Ep-ICD in samples from subjects having healthy prostates and various stages of prostate cancer; ii) determining associations between Ep-ICD subcellular localization and DFS times in prostate cancer patients; iii) determining that both quantitative and qualitative measurement of subcellular localization of Ep-ICD provided useful prognostic information; iv) generating algorithms for using the quantitative and qualitative data to calculate values with prognostic significance; and v) generating scales and equations for use in the algorithms, wherein the scales are appropriate for scoring the quantitative and qualitative data and weighting the quantitative and qualitative data with respect to one another. In a particularly preferred embodiment, the combination of collecting quantitative and qualitative data regarding Ep-ICD subcellular localization in a prostate tissue sample, applying an ESLI algorithm to the collected data to generate an ESLI value for the sample, comparing the ESLI value of the sample to a reference value facilitates prognosis of prognosis of prostate cancer in subjects. In a particularly preferred embodiment, the quantitative and qualitative data are collected from tissue samples prepared for IHC.

[0052] In one aspect, the prognosis method described herein can be used in conjunction with the traditional Gleason scoring system to prognose prostate cancer patients. For example, the prognosis method described herein may be used as a secondary prognosis tool to verify a poor prognosis based on a Gleason score (e.g., a GS score greater than or equal to 7). Alternatively, the prognosis method described herein may be used to identify prostate cancer patients who have a poor prognosis despite having a GS of less than 7. [0053] Details of the ESLI prognosis method and the ESLI algorithms are discussed further below.

[0054] In order to calculate an ESLI value, a nuclear Ep-ICD score and a cytoplasmic Ep-ICD score are determined for a prostate tissue sample obtained from the subject. The prostate tissue sample would comprise cells (e.g., epithelial cells), each of such cells having a nucleus and cytoplasm.

[0055] In one embodiment, determination of quantitative and qualitative nuclear Ep-ICD and cytoplasmic Ep-ICD scores is done manually and comprises the following four steps.

[0056] (i) The sample is contacted with a binding agent that specifically binds to Ep-ICD or part thereof. A detectable label is used to detect binding of the binding agent to Ep-ICD. As discussed above, the detectable label may, for example, emit a detectable signal upon binding of the binding agent to Ep-ICD. In one aspect, the binding agent may be a labelled antibody specific to Ep-ICD. The label may be chosen from, for example, detectable radioisotopes, luminescent compounds, fluorescent compounds, enzymatic labels, biotinyl groups and predetermined polypeptide epitopes recognizable by a secondary reporter.

[0057] (ii) Subcellular localization of Ep-ICD is measured quantitatively and scored based on the percentages of cells in a tissue sample that are positive for Ep-ICD in (a) their nucleus and (b) their cytoplasm. The percentage of cells in a tissue sample that are positive for nuclear Ep-ICD expression is referred to as the "first percentage". The first percentage is then assigned a score according to a scale that correlates percentage ranges with integer values. Such score and scale are referred to as the "first quantitative score" and "first scale". The percentage of cells in a measured tissue sample that are positive for cytoplasmic Ep- ICD expression is referred to as the "second percentage". The second percentage is then assigned a "second quantitative score" according to the first scale.

[0058] In one aspect, the first percentage (i.e., the percentage positive for nuclear Ep- ICD) and the second percentage (i.e., the percentage positive for cytoplasmic Ep-ICD) are scored according to the following first scale: when less than 10% of cells are positive a score of 0 is assigned; when 10 - 30% cells are positive a score of 1 is assigned; when 31 - 50% cells are positive a score of 2 is assigned; when 51 - 70% of cells are positive a score of 3 is assigned; and when more than 70% of cells are positive a score of 4 is assigned.

[0059] In one embodiment, the first and second percentages are obtained using immunohistochemistry. Immunohistochemistry (IHC) is a known method for demonstrating the presence and location of one or more specific proteins in tissue sections. Briefly, IHC comprises fixing and embedding a tissue sample, sectioning the tissue, mounting the tissue section, deparaffinizing and rehydrating the section, antigen retrieval, immunohistochemical staining, optional counterstaining, dehydrating and stabilizing with mounting medium, and viewing the stained section under a microscope.

[0060] In one embodiment, wherein the first and second percentages are obtained using IHC, a cell that is positive for nuclear and/or cytoplasmic Ep-ICD is one that is

immunopositive (i.e., a cell comprising staining or fluorescence that is detectable upon microscopic examination and indicative of the Ep-ICD-specific antibody used in IHC of the sample).

[0061] In a preferred embodiment, the tissue sample prepared for use in IHC analysis comprises serial formalin fixed and paraffin embedded (FFPE) tissue sections.

[0062] (iii) Subcellular localization of Ep-ICD is measured qualitatively and scored based on the intensity of the signals emitted by a detectable label of an Ep-ICD binding agent in (a) the nucleus and (b) the cytoplasm of cells in the tissue sample. The intensity of the signal detected in the nucleus of the cells in the tissue is referred to as the "first intensity". The first intensity is then assigned a score according to a scale that correlates a categorical assessment of signal intensity (e.g., categories ranging from zero detectable signal to a maximum or near maximum detected signal) with integer values. Such score and scale are referred to as the "first qualitative score" and "second scale". The intensity of the signal detected in the cytoplasm of the cells in the tissue is referred to as the "second intensity". The second intensity is then assigned a "second qualitative score" according to the second scale.

[0063] In one aspect, the first intensity (i.e., the categorical assessment of nuclear Ep- ICD binding agent signal emission) and the second intensity (i.e., the categorical assessment of cytoplasmic Ep-ICD binding agent signal emission) are scored according to the following second scale: when no signal is detected a score of 0 is assigned; when a mild signal is detected a score of 1 is assigned; when a moderate signal is detected a score of 2 is assigned; and when an intense signal is detected a score of 3 is assigned.

[0064] In one embodiment, the first and second intensities are obtained using IHC analysis. In one preferred embodiment, the antibody-antigen interaction (i.e., the anti-Ep- ICD-Ep-ICD interaction) is visualized using chromogenic detection, in which an enzyme conjugated to the antibody cleaves a substrate to produce a colored precipitate at the location of the protein. In another preferred embodiment, the antibody-antigen interaction is visualized using fluorescent detection, in which a fluorophore is conjugated to the antibody and the location of the fluorophore can be visualized using fluorescence microscopy. [0065] (iv) A nuclear Ep-ICD score and a cytoplasmic Ep-ICD score are calculated by adding the first quantitative and qualitative scores to generate the nuclear Ep-ICD score and adding the second quantitative and qualitative scores to generate the cytoplasmic Ep-ICD score.

[0066] In one preferred embodiment, after determining nuclear Ep-ICD and cytoplasmic Ep-ICD scores, an Ep-ICD Subcellular Localization Index (ESLI) value for the sample is calculated. In one example, the ESLI value is the sum of the nuclear Ep-ICD score and the cytoplasmic Ep-ICD score. In one preferred embodiment, the ESLI value is the sum of the nuclear Ep-ICD score and the cytoplasmic Ep-ICD score, divided by two (such arithmetic function being for convenience).

[0067] The calculated ESLI value is then compared to a reference value in order to determine a prognosis of prostate cancer in the subject. The reference value is a predetermined cut-off value, wherein values on one side of the cut off value indicate a poor prognosis of prostate cancer and vales on the other side of the cut-off value indicate a favourable prognosis of prostate cancer.

[0068] In one embodiment, the reference value is an ESLI value indicative of a non- aggressive cancerous prostate tissue. In this embodiment, a poor prognosis of prostate cancer in the subject is determined when the calculated ESLI value of the sample is less than the reference value. In this embodiment, a favourable prognosis of prostate cancer is determined when the calculated ESLI value of the sample is greater than or equal to the reference value.

[0069] In one embodiment, the reference value is an ESLI value indicative of an aggressive prostate cancer. For example, the sample may be obtained from an aggressive prostate tumour tissue. In this embodiment, a poor prognosis of prostate cancer in the subject is determined when the calculated ESLI value of the sample is less than or equal to the reference value.

[0070] In a preferred embodiment, the reference value is determined by retrospectively analyzing a plurality of prostate cancer patients' tissue samples and corresponding patient clinical data regarding time of DFS.

[0071] In a particularly preferred embodiment, wherein an ESLI value is calculated using nuclear and cytoplasmic Ep-ICD scores generated according to the aforementioned first and second scales (i.e., 0-4 and 0-3 for percentage and intensity respectively), a finding of an ESLI value of less than 3 is indicative of aggressive prostate cancer and a poor prognosis of prostate cancer. [0072] In one embodiment, determination of quantitative and qualitative nuclear Ep-ICD and cytoplasmic Ep-ICD scores is automated.

[0073] Automated determination of quantitative and qualitative nuclear Ep-ICD and cytoplasmic Ep-ICD scores may comprise one or more of the same steps described above in the manual method for determination of quantitative and qualitative nuclear Ep-ICD and cytoplasmic Ep-ICD scores. For example, contacting the biological sample with a binding agent that specifically binds to Ep-ICD or part thereof and measuring subcellular localization of Ep-ICD quantitatively and quantitatively may be done as described above. However, in an automated method, subcellular localization measurements and scores may be generated in conjunction with image analysis software, such as, for example, Visiopharm™ software. In one embodiment, automated measurements may be possible on a finer scale than manual measurements. Thus, the scales and scoring for automated measurements may differ from those used in manual methods or they may be the same.

[0074] Calculation of an ESLI value for the biological sample may also be automated and the ESLI algorithm may be the same of different from the algorithm used in a manual method. As in the manual method, an ESLI algorithm for use in an automated method comprises an arithmetic function that weighs the quantitative and qualitative data with respect to one another in a manner that provides a value with prognostic significance.

[0075] In one preferred embodiment comprising automated quantitative and qualitative measurements, the nuclear Ep-ICD score is defined by -0.0071 (percent positivity for nuclear Ep-ICD) + 0.0545(nuclear Ep-ICD intensity score) and the cytoplasmic score is defined by the intensity score of cytoplasmic Ep-ICD. In one preferred embodiment, the ESLI value is the sum of 0.8637(nuclear Ep-ICD score) and 0.0233(cytoplasmic score).

[0076] The calculated ESLI value is then compared to a reference value in order to determine a prognosis of prostate cancer in the subject. The reference value is a predetermined cut-off value, wherein values on one side of the cut off value indicate a poor prognosis of prostate cancer and vales on the other side of the cut-off value indicate a favourable prognosis of prostate cancer. In one embodiment, determination of a reference value is automated. In a preferred embodiment, the reference value is determined by automated retrospective analysis of a plurality of prostate cancer patients' tissue samples and corresponding patient clinical data regarding time of DFS.

[0077] In one aspect, the present method for prognosing prostate cancer comprising calculating an ESLI value for a prostate tissue sample obtained from a subject is used in combination with a Gleason score analysis. As discussed above, cancers with higher Gleason scores are thought to be more aggressive and have a worse prognosis than those with lower Gleason scores (i.e., a GS of less than 7). Accordingly, in one aspect, the method provided herein, comprising calculating an ESLI value for a prostate tissue sample, can be used in combination with a Gleason score analysis in order to verify a poor prognosis determined according to the Gleason score analysis or to verify or refute a favorable prognosis determined according to the Gleason score analysis.

[0078] For example, a patient having a GS of less than 7 is considered to have early stage prostate cancer. This early stage patient group may be treated by active surveillance, in which a practitioner monitors the patient for signs that the cancer may be growing or changing. It is possible that early stage prostate cancer will grow so slowly that it may not cause health problems during a man's lifetime. However, a subset of early stage prostate cancers may be aggressive. The present inventors have found that aggressive early phase prostate cancers may be identified by examining the subcellular localization of Ep-ICD in prostate tissue samples, calculating a corresponding ESLI value for such a sample, and comparing the calculated ESLI value to a reference value in order to prognose the prostate cancer. Identification of aggressive early stage prostate cancers using the method disclosed herein may allow practitioners to identify early stage prostate cancer patients who may benefit from therapeutic treatment, such as, for example, surgery radiation, hormone or combination therapy.

[0079] In one embodiment, a method for detecting abnormal subcellular localization of Ep-ICD in a prostate tissue sample obtained from a subject is provided. In an aspect, the method comprises measuring an amount of nuclear and cytoplasmic Ep-ICD in a biological sample from the subject, comparing the amounts measured in the biological sample to respective controls; and detecting abnormal subcellular localization of Ep-ICD in the prostate tissue sample based on the comparison between the measured amount of nuclear and cytoplasmic Ep-ICD and the controls. Measurement may be quantitative and/or qualitative, as described herein. The controls may be non-aggressive or aggressive prostate cancer, as described herein. Detection of an abnormal subcellular localization of Ep-ICD in a prostate tissue sample is found when the measured amount of nuclear Ep-ICD is less than that of the non-aggressive control or less than or equal to that of the aggressive control.

[0080] In another embodiment, the method for detecting abnormal subcellular localization of Ep-ICD in a prostate tissue sample obtained from a subject comprises the steps of (A) measuring nuclear and cytoplasmic Ep-ICD scores for the sample, (B) calculating an ESLI value for the sample and (C) comparing the calculated ESLI value to a reference value. The measuring and calculating steps may be carried out as discussed above with respect to prostate cancer prognosis. In this embodiment, abnormal subcellular localization of Ep-ICD in the prostate tissue sample is detected when the calculated ESLI value of the sample is less than a reference value corresponding to an ESLI value indicative of a non-aggressive cancerous prostate tissue; or when the calculated ESLI value of the sample is less than or equal to a reference value corresponding to an ESLI value indicative of an aggressive prostate cancer.

[0081] Methods for the Diagnosis of Prostate Cancer

[0082] The present disclosure is also generally directed to a method for distinguishing prostate cancer tissue from non-malignant prostate tissue. Diagnosis of prostate cancer is typically made based on histological examination of prostate tissue biopsy samples. In some instances, an additional or alternative method of diagnosing prostate cancer may be desired. For example, histological examination of prostate tissue biopsy samples may be indefinite or a secondary method of diagnosis to confirm or refute a histological diagnosis might be desired. In one embodiment, a method for diagnosing prostate cancer in a subject comprising measuring subcellular localization of Ep-ICD and optionally membranous EpEX in a biological sample obtained from the subject is provided.

[0083] In one embodiment, the method of diagnosing prostate cancer comprises measuring an amount of nuclear Ep-ICD in a biological sample from the subject. The measured amount is compared to a first control and a diagnosis of prostate cancer or the absence of prostate cancer is made based on results of the comparison. In one embodiment, the first control is an amount of nuclear Ep-ICD in a non-cancerous biological sample, for example, a non-cancerous prostate tissue or a sample comprising noncancerous prostate epithelial cells. In this case, a lower measured amount of nuclear Ep-ICD in the biological sample relative to the first control indicates a diagnosis of prostate cancer and an equal or higher measured amount of nuclear Ep-ICD in the biological sample relative to the first control indicates that the subject does not have prostate cancer. In one embodiment, the first control is an amount of nuclear Ep-ICD in a cancerous biological sample, for example, a prostate tumor sample or a sample comprising cancerous prostate epithelial cells. In this case, an equal or lower measured amount of nuclear Ep-ICD in the biological sample relative to the first control indicates a diagnosis of prostate cancer.

[0084] In a preferred embodiment, nuclear Ep-ICD is measured quantitatively and qualitatively using IHC. In one aspect, quantitative and qualitative measurements are obtained using the methods described herein for prognosis of prostate cancer. If quantitative and qualitative measurements are scored using the aforementioned first and second scales, a finding of a nuclear Ep-ICD score of less than 4 is indicative of a cancerous prostate tissue sample.

[0085] In one embodiment, the method of diagnosing prostate cancer further comprises measuring an amount of EpEX in the biological sample. The measured EpEX amount is compared to a second control and a diagnosis of prostate cancer or the absence of prostate cancer is made based on the results of both the Ep-ICD comparison and the EpEX comparison. In one embodiment, the second control is an amount of membrane EpEX in a non-cancerous biological sample. In this case, a lower measured amount of membrane EpEX in the biological sample relative to the second control indicates a diagnosis of prostate cancer and an equal or higher measured amount of membrane EpEX in the biological sample relative to the second control indicates that the subject does not have prostate cancer. In one embodiment, the second control is an amount of membrane EpEX in a cancerous biological sample. In this case, an equal or lower measured amount of membrane EpEX in the biological sample relative to the second control indicates a diagnosis of prostate cancer.

[0086] In a preferred embodiment, membrane EpEX is measured quantitatively and qualitatively using IHC. In one aspect, quantitative and qualitative measurements are obtained using the methods described herein for prognosis of prostate cancer. If quantitative and qualitative measurements are scored using the aforementioned first and second scales, a finding of a nuclear Ep-ICD score of less than 4 in conjunction with a membrane EpEX score of less than 3 is indicative of a cancerous prostate tissue.

[0087] Kits

[0088] The present disclosure contemplates kits for carrying out the methods disclosed herein. Such kits typically comprise two or more components required for performing a prognostic prostate cancer assay. Components include but are not limited to one or more of compounds, reagents, containers, equipment and instructions for using the kit. Accordingly, the methods described herein may be performed by utilizing pre-packaged prognostic kits provided herein. In one embodiment, the kit comprises one or more of binding agents, standards, stains, fixatives and instructions. In some embodiments, the instructions comprise one or more reference values for use as controls.

[0089] In one embodiment, the kit comprises one or more binding agents as described herein for prognosing prostate cancer. By way of example, the kit may contain antibodies specific for Ep-ICD, antibodies against the Ep-ICD antibodies labelled with an enzyme(s), and a substrate for the enzyme(s). The kit may further contain antibodies specific for EpEX, antibodies against the EpEX antibodies labelled with an enzyme(s), and a substrate for the enzyme(s). The kit may also contain one or more of microtiter plates, reagents (e.g., standards, buffers), adhesive plate covers, and instructions for carrying out a method using the kit.

[0090] In one embodiment, the kit comprises antibodies or antibody fragments which bind specifically to epitopes of Ep-ICD and means for detecting binding of the antibodies to their epitopes associated with prostate cancer cells, either as concentrates (including lyophilized compositions), which may be further diluted prior to testing, or non-concentrated compositions. For example, a kit for prognosing prostate cancer may contain a known amount of a first binding agent that specifically binds to Ep-ICD, wherein the first specific binding agent comprises a detectable substance or has the capacity to bind directly or indirectly to a detectable substance. In one embodiment, the kit further comprises antibodies or antibody fragments which bind specifically to epitopes of EpEX and means for detecting binding of the EpEX-specific antibodies to their epitopes associated with prostate cancer cells, either as concentrates (including lyophilized compositions), which may be further diluted prior to testing, or non-concentrated compositions.

[0091] In one embodiment, the kit comprises one or more binding agents, standards, stains, fixatives and instructions for measuring nuclear Ep-ICD and optionally membrane EpEX. For example, a kit comprising such binding agents, standards, stains fixatives and instructions may be used to practice methods disclosed herein. In a preferred embodiment, the kit can be used to practice a method disclosed herein that comprises IHC.

[0092] In one embodiment, the kit may further comprise tools useful for collecting biological samples (e.g. prostate tissue samples).

[0093] The following non-limiting examples illustrative of the disclosure are provided.

[0094] Example 1 : Nuclear Ep-ICD Expression Can Be Used to Predict Poor Prognosis in Prostate Cancer Patients

[0095] In Example 1 , the prognostic utility of subcellular Ep-ICD expression and membranous EpEX expression in prostate cancer are examined. Scoring and evaluation of subcellular Ep-ICD expression and membranous EpEX expression is done manually.

Correlation of subcellular Ep-ICD and membranous EpEX expression with clinic-pathological parameters and follow up of prostate cancer patients are also examined. [0096] Methods

[0097] This retrospective study using prostate cancer patients' archived tissue blocks and anonymized clinical data was approved by the Mount Sinai Hospital (MSH) Research Ethics Board, Toronto, Canada.

[0098] Patients

[0099] Patients who were considered to have a "normal prostate" had a diagnosis "negative for malignancy" that remained unchanged for the next 5 years with biopsies in between. These patients were required to have a biopsy due to an increase in prostate- specific antigen (PSA).

[00100] A cancer was considered "aggressive" if a patient had a biochemical recurrence (as defined by the National Comprehensive Cancer Network), distal metastasis, or if a patient died due to prostate cancer; these features were also considered to be clinical end points. All tissue samples identified as aggressive were re-reviewed by a pathologist.

Patients in the aggressive cohort had radical prostatectomy (RP, n = 101) or radiation therapy (RT, n=143) and were free of disease after treatment. Patients considered to have a biochemical recurrence were patients that had an increase in PSA by 2 ng/ml after RT. Patients who had radical prostatectomy considered to have a biochemical recurrence were those that had a rise in PSA by 0.2 ng/ml on 2 occasions during follow-up.

[00101] A schematic illustration of the study design is provided in Figure 1. Inclusion criteria used in this study were based on prostate cancer patients having an event or at least 5 years of follow-up without an event. Patients were excluded if there was incomplete follow- up information (i.e., less than 5 years) or if they were on active surveillance. Patients were followed up for up to 12 years.

[00102] Immunohistochemistry (IHC)

[00103] Serial formalin fixed and paraffin embedded (FFPE) tissue sections (4 μηι thickness) were deparaffinized in xylene, hydrated through graded alcohol series, antigen was retrieved by microwave treatment, endogenous peroxidase activity was blocked and nonspecific binding was blocked using normal horse (EpEX) or normal goat (Ep-ICD) serum (10%) as previously described (Tripathi S.C. et al., PloS One, 2011 , 6:e19213). The sections were incubated with the mouse monoclonal antibody against EpEX (MOC-31 , AbD Serotec, Oxford, UK) or rabbit anti-human Ep-ICD monoclonal antibody (Epitomics Inc., Burlingame, CA) for one hour (Ralhan R. et al., BMC Cancer, 2010, 10:331 ; and Chaudry M.A., et al., Br J Cancer, 2007, 96:1013-1019). Slides were incubated with biotinylated secondary antibody for 20 minutes, followed by VECTASTAIN Elite ABC reagent (Vector labs, Burlingame, CA) using diaminobenzidine as the chromogen. Slides were washed with Tris-buffered saline (TBS, 0.1 M, pH = 7.4), 3-5 times after every step. Sections were counterstained with Mayer's hematoxylin. In negative control tissue sections, the primary antibody was replaced by isotype- specific non-immune mouse/rabbit IgG. The sections were evaluated by light microscopic examination. Various different nuclear antigen retrieval conditions were tested, including pressure cooker treatment, microwave treatment (pre-heat the retrieval buffer for 8 minutes, antigen retrieval - 15 min at full power) and different buffers were tested, including sodium citrate buffer pH = 6.0, Tris-HCI buffer pH = 9.0, with or without 0.05% Tween 20, and heating in an oven at 1 15°C for 3min in Tris-HCI buffer pH = 9.0, with 0.05% Tween 20 to optimize retrieval of nuclear Ep-ICD.

[00104] Immunohistochemistry Evaluation and Manual Scoring

[00105] Immunohistochemistry (IHC) manual scoring was performed under supervision of a pathologist by two researchers who were blinded to clinical outcome. Immunopositive staining was evaluated in the most aggressive areas of tissue sections, as previously described (Tripathi S.C. et al., PloS One, 201 1 , 6:e19213). Nuclear and cytoplasmic Ep-ICD, and membranous EpEX expression were evaluated independently in tumor cell nucleus, cytoplasm and membrane based on intensity and percentage of positive staining. Tissue sections were scored as follows: 0, < 10% cells; 1 , 10 - 30% cells; 2, 31 - 50% cells; 3, 51 - 70% cells; and 4, > 70% cells showed immunoreactivity. Sections were also scored semi- quantitatively on the basis of intensity as follows: 0 = none; 1 = mild; 2 = moderate; and 3 = intense. Scoring was done manually (i.e., visual analog scoring). Finally, a total score (ranging from 0 to 7) was obtained by adding percentage positivity and intensity scores for each of nuclear Ep-ICD, cytoplasmic Ep-ICD and membrane EpEX in the cancer and noncancerous tissue sections (Tripathi S.C. et al., PloS One, 201 1 , 6:e19213).

[00106] Statistical Analysis

[00107] Immunohistochemistry data were subjected to statistical analysis using SPSS 21.0 software (SPSS, Chicago, IL) and GraphPad Prism 5.0 software (GraphPad Software, La Jolla, CA). Scatter plots were used to determine the distribution of total score of cytoplasmic or nuclear Ep-ICD and EpEX expression in all tissues examined, p value < 0.05 was considered significant for statistical analysis. Based on receiver operating curve analysis, the cut-offs of IHC score for preserved expression of cytoplasmic Ep-ICD and nuclear Ep-ICD were > 4.0, and for membrane EpEX were > 3.0 for further analysis.

Expression data were analyzed to determine significant correlations between preserved expression of Ep-ICD/ EpEX, clinical parameters and prognosis of prostate cancer patients. Correlation of Ep-ICD/ EpEX expression with patient survival (i.e., disease free survival, DFS) was evaluated using life tables constructed from survival data with Kaplan-Meier plots, as previously described (Tripathi S.C. et al., PloS One, 201 1 , 6:e19213). Multivariate analysis was carried out using Cox regression models to determine the performance of preserved expression of EpEX/Ep-ICD as a marker in comparison with other clinical and pathological prognostic parameters including age, Gleason score (GS), AJCC (American Joint Committee on Cancer) stage, clinical risk (based on PSA), and grade in prostate cancer patients. Bootstrap analysis was done to show an internal validation of data (Tables 1 and 2).

[00108] Table 1 : Internal Validation for risk assessment through bootstrap method.

7 101

>7 40 — — — — —

RISK*

Low 157

Int. 50

High 20 — — — — —

Recurrence

No 159

0.13(0.19- 0.82(0.14-

Yes 90 0.012 0.234 0.170 0.11(0.21-0.62)

0.69) 3.28)

[00109] Table 2: Bootstrap validation for cox multivariate.

[00110] Ep-ICD Subcellular Localization Index (ESLI) Scoring

[00111] Following the evaluation and scoring of the IHC data, a calculation was made of the ESLI. The ESLI was calculated according to the following equation: ESLI = ½ x

(%positivity score of Nuclear Ep-ICD + intensity score of Nuclear Ep-ICD + %positivity score of Cytoplasm Ep-ICD + intensity score of Cytoplasm Ep-ICD). As indicated above, the %positivity score comprises a score on a scale of 0 to 4 and the intensity score comprises a score on a scale of 0 to 3. A cut off of 3 was used for the ESLI score, wherein ESLI scores less than or equal to 3 indicate a poor prognosis and ESLI scores greater than 3 indicate a favorable prognosis.

[00112] Results

[00113] Immunohistochemical analysis of Ep-ICD and EpEX expression in prostate tissues

[00114] Representative photomicrographs depicting the subcellular localization of Ep-ICD and EpEX in normal prostate tissue, benign prostate hyperplasia (BPH) and prostate cancer tissues having different Gleason scores (GSs) are shown in Figure 2. Distribution of IHC scores for nuclear Ep-ICD (left), cytoplasmic Ep-ICD (center), and membranous EpEX (right), in normal, BPH, prostate intraepithelial neoplasia (PIN) and, prostate cancer( adenocarcinoma) tissues is shown in Figure 3 . A box plot analysis of IHC scores for nuclear Ep-ICD (left), cytoplasmic Ep-ICD (center) and membranous EpEX (right) with respect to GS is shown in Figure 4. Among the normal prostate tissues, 99 of 100 (99%), 100 of 100 (100%) and 81 of 100 (81 %) expressed strong nuclear Ep-ICD, cytoplasmic Ep-ICD and EpEX membrane respectively (Table 3), while the prostate cancer showed significant decrease in expression of these proteins (84/249, 33.7%; 240/249, 96.4%; and 142/249, 57.0% respectively; Table 3). Benign prostate hyperplasia tissue samples showed a similar decrease in nuclear Ep-ICD, cytoplasmic Ep-ICD and EpEX expression in comparison with normal prostate tissues, though the decrease was smaller than in prostate cancer samples (Table 3). Among prostate cancer samples, there was a decrease in membranous EpEX with increasing AJCC staging and increasing GS (Table 3). Patients who had recurrence showed reduced nuclear Ep-ICD (p = 0.002) and decreased membranous EpEX expression (p < 0.001) (Table 3). Comparison of prostate cancer tissues with normal tissues showed that nuclear Ep-ICD expression had the highest specificity (99%) and sensitivity (66%) with an area under the curve (AUC) of 0.909 (Table 4). Comparison of BPH tissues with prostate cancer tissues showed that nuclear Ep-ICD expression had a sensitivity of 51%, specificity of 66% and AUC of 0.656.

[00115] Table 3: Analysis of Ep-ICD and EpEX expression in BPH and prostate cancer and correlation with clinic-pathological parameters ( ¹ Normal vs BPH, ² BPH vs

Cancer, ³ Normal vs Cancer *BPH-Benign Prostatic Hyperplasia, **PIN - Prostatic Intra- epthelial Neoplasia, ^# RISK as defined by AJCC, data not available for 22 patients)

[00116] Table 4: Biomarker analysis of Ep-ICD and EpEX in benign prostate hyperplasia and prostate cancer. Ep-ICD (Cytoplasm) > 4 1.00 0.04 0.512 0.720

Membrane EpEX > 3 0.81 0.43 0.630 <0.001

BPH vs Cancer

Ep-ICD (Nuclear) > 4 0.51 0.66 0.656 <0.001

Ep-ICD (Cytoplasm) > 4 0.84 0.04 0.369 <0.001

Membrane EpEX > 3 0.77 0.43 0.616 0.001

[00117] Ep-ICD and EpEX as Prognostic Markers of Prostate Cancer

[00118] In Kaplan Meier analysis for all prostate cancer tissue samples (all GS taken together), patients with preserved expression of nuclear Ep-ICD had a mean DFS of 9.5 years, while those with reduced expression had a mean DFS of 7.8 years (p = 0.001);

patients with increased membranous EpEX had a mean DFS of 9.5 years while those showing loss of membrane EpEX had mean DFS of 7.1 years, p < 0.001 ; Figure 5). For prostate cancer patients with a GS of less than 7, nuclear Ep-ICD and membrane EpEX (increased expression mean DFS = 10.5 years, reduced expression mean DFS = 9.0 years, respectively) were associated with DFS (Figure 5, p = 0.001 , p = 0.042, respectively).

Patients with a GS of 7 showed significant association of nuclear Ep-ICD (increased expression mean DFS = 8.9 years, reduced expression mean DFS = 7.4 years) and membranous EpEX (increased expression mean DFS = 8.7 years, reduced expression mean DFS = 6.5 years) with survival (Figure 5, p = 0.05 and p = 0.027 respectively).

However, none of these proteins were associated with disease prognosis in prostate cancer with GS greater than 7.

[00119] In Cox regression multivariate analysis for all prostate cancer samples where nuclear Ep-ICD, risk, stage, membrane EpEX, and GS were included in the model, membrane EpEX, GS and nuclear Ep-ICD emerged as the most significant factors (p < 0.001 , Hazard ratio (HR) = 0.294; p < 0.001 , HR = 5.158; and p < 0.001 , HR = 0.359 respectively, Table 5). In the same model including age and excluding risk patients with a GS < 7, nuclear Ep-ICD emerged as the most significant factor (p = 0.05, HR = 0.025). For patients with a GS of 7 in the same model, AJCC stage (p = 0.001 , HR = 6.214) and membranous EpEX (p = 0.019, HR = 0.452) emerged as the most significant factors for DFS (Table 5). None of these factors significantly correlated with prognosis for patients with a GS greater than 7 in multivariate analysis.

[00120] Table 5: Univariate and multivariate analysis with respect to Gleason score groups.

[00121] ESLI Results

[00122] A significant association was observed between ESLI values > 3 and increased disease free survival in all prostate cancer cases (p< 0.001); patients having an ESLI value > 3 had a mean survival of 8.94 years and median survival of 1 1.23 years, while patients having an ESLI value < 3 had a mean survival of 5.17 years and median survival of 3.74 years (Figure 6A).

[00123] A significant association was observed between Gleason scores of less than 7 and increased disease free survival (p< 0.001) in patients; GS positive cases (GS <7) had a mean survival of 10 years, while the GS negative cases (GS > 7) had a mean survival of 7.07 years (Figure 6B).

[00124] A significant association was observed between ESLI positivity (i.e., an ESLI value of > 3) and GS positivity (i.e., a GS of <7) with increased disease free survival in all prostate cancer cases (p< 0.001); ESLI and GS positive cases had a mean survival of 10.39 years in comparison with all the other prostate cancer cases that had a mean survival of 6.98 years (Figure 6C).

[00125] Among the GS positive prostate cancer patients (GS <7), a significant association was observed between ESLI positivity (i.e., an ESLI value of > 3) and increased disease free survival (p< 0.001); patients having an ESLI value > 3 had a mean survival of 10.39 years, while patients having an ESLI value < 3 had a mean survival of 5.47 years (Figure 6D). [00126] Discussion

[00127] Evidence gathered in Example 1 suggests that prostate cancer patients having an increase in nuclear Ep-ICD and membranous EpEX (across all GSs) were found to have better survival. In patients with a GS of less than 7, preserved expression of nuclear Ep-ICD emerged as the most significant marker in multivariate analysis for prolonged disease free survival, wherein patients had no recurrence during the follow up period and wherein patients having reduced nuclear Ep-ICD expression were at high risk of having aggressive disease. Prostate cancer patients with a GS of less than 7 are often thought to have indolent cancer (i.e., non-aggressive). However, the present results indicate that patients having a GS of less than 7 may have an aggressive cancer that is identifiable by reduced nuclear Ep- ICD expression and/or ESLI value. Aggressive prostate cancer is associated with a poor prognosis.

[00128] The inventors' earlier report on Ep-ICD and EpEX expression analysis in ten epithelial cancers described Ep-ICD expression in prostate cancer with respect to diagnosis of prostate cancer, wherein nuclear and cytoplasmic Ep-ICD immunopositivity was observed in 40 of 49 tumors; nuclear Ep-ICD was observed in 2 of 9 and cytoplasmic Ep-ICD in 1 of 9 normal prostate tissues (US Patent Publication No. 201 1/0275530). The present results, which were generated using a larger dataset that is independent from that of US

201 1/0275530, showed nuclear Ep-ICD immunopositivity in the majority of normal prostate tissues, while 33.7% of prostate cancer samples were nuclear Ep-ICD immunopositive. This difference in results may be attributed to the fact that there was a larger cohort of both prostate normal as well as cancers in the present study.

[00129] In Example 1 , nuclear Ep-ICD expression was identified as the most significant marker for distinguishing poor from favorable prognosis of prostate cancer; higher amounts of nuclear Ep-ICD indicating a favorable prognosis and lower amounts of nuclear Ep-ICD indicating a poor prognosis. Accordingly, methods for prognosing prostate cancer that rely on nuclear Ep-ICD as a marker must comprise techniques capable of reliably detecting subcellular localization of Ep-ICD, including localization in the nucleus of epithelial cells.

[00130] The present results indicate the prognostic significance of nuclear Ep-ICD, which may be used to stratify prostate cancer patients with a GS of less than 7 who are at increased risk of disease recurrence. In some cases, patients who have a GS of less than 7 may have a recurrence of prostate cancer even though the traditional system of cancer grading had deemed those patients to be at a low risk for disease recurrence. It is contemplated herein that patients who have a GS of less than 7 and are nuclear Ep-ICD negative would benefit from rigorous follow up and management (i.e., a standard of care prescribed to a patient having an aggressive prostate cancer). It is also contemplated that analysis of nuclear Ep-ICD expression may deter overtreatment of patients who are at low risk of disease recurrence, thereby reducing harmful side effects of therapy as well as reduce the economic burden on health care providers.

[00131] Example 2: Automated evaluation of Nuclear Ep-ICD Expression Can Be Used to Predict Poor Prognosis in Prostate Cancer Patients

[00132] In Example 2, automated methods are used to evaluate subcellular Ep-ICD expression in the prostate cancer samples described in Example 1. An ESLI algorithm and calculation of ESLI values based on the algorithm and Ep-ICD expression data generated using the automated methods are provided.

[00133] Automated Analysis of Immunohistochemistry Data

[00134] Tissue samples subjected to IHC, as described above, were provided on slides and scanned using a NanoZoomer™ scanner at The Toronto Centre for Phenogenomics (TCP, Toronto, Ontario, Canada) at 20x magnification. Scanned images were uploaded to Visiopharm™ software for analysis using applications developed in-house. Analysis of the scanned digital images was performed using the Visiopharm Integrator System™ (VIS, version 4.6.3.857; Visiopharm, Hoersholm, Denmark). Regions-of-interest (ROI) were manually drawn on each digital image. Regions within the ROIs were analyzed by the VIS to measure 3,3'-Diaminobenzidine (DAB) staining in in the nuclei, cytoplasm and/or membrane of epithelial cells and to measure the intensity of staining in the cell nuclei, cytoplasms and/or membranes. Results of this analysis were then used to stratify patients based on their risk for disease reoccurrence.

[00135] Results of Automated Analysis of Immunohistochemistry Data

[00136] Representative images of prostate epithelial tissues obtained using the

NanoZoomer scanner are provided in FIG. 7. Nuclear and cytoplasmic Ep-ICD DAB staining (right) and membrane EpEX DAB staining (left) are shown in prostate cancer tissues comprising a GS of <7 (top) and a GS of > 7 (bottom).

[00137] A Kaplan Meier survival curve was generated based on nuclear Ep-ICD expression measured using Visopharm software (FIG. 8A). The percentage of cells in a sample expressing nuclear Ep-ICD and the intensity of nuclear Ep-ICD staining were assessed and scored for each sample analyzed. Using these data, a "prediction score 1 " was derived using the following algorithm: prediction score 1 = -0.0071 (% positive nuclear Ep-ICD score) + 0.0545(nuclear Ep-ICD intensity score). The % positive nuclear Ep-ICD scores and nuclear Ep-ICD intensity scores are generated by the Visiopharm software used. A prediction score 1 of >_ 3.94 was suitable for identifying patient samples having a high risk for disease recurrence (i.e., those having a poor prognosis of prostate cancer) and a prediction score 1 of < 3.94 was suitable for identifying patient samples having a low risk of disease recurrence (i.e., those having a good prognosis of prostate cancer). The P value of prediction score 1 in the cox regression analysis was 4.14e-05. The P value of the log-rank test for the difference between the identified high and low risk groups was 1.78e-8.

Prediction score 1 is representative of a nuclear Ep-ICD score.

[00138] A Kaplan Meier survival curve was generated based on cytoplasmic Ep-ICD expression measured using Visopharm software (FIG. 8B). Specifically, the intensity of cytoplasmic Ep-ICD staining was assessed and scored for each sample analyzed. Using these data, a "prediction score 2" was derived using the following algorithm: prediction score 2 = cytoplasmic intensity score. Cytoplasmic intensity score is generated by the Visiopharm software and is based on the intensity of cytoplasmic Ep-ICD staining in the region of interest in the stained tissue section. A prediction score 2 of >_161.2 was suitable for identifying patient samples having a high risk for disease recurrence (i.e., those having a poor prognosis of prostate cancer) and a prediction score 2 of < 161.2 was suitable for identifying patient samples having a low risk of disease recurrence (i.e., those having a good prognosis of prostate cancer). The P value of prediction score 1 in the cox regression analysis was 0.0026. The P value of the log-rank test for the difference between the identified high and low risk groups was 3.64e-5. Prediction score 2 is representative of a cytoplasmic Ep-ICD score.

[00139] An ESLI value was calculated based on the % positive nuclear Ep-ICD score, the nuclear Ep-ICD intensity score and the cytoplasmic Ep-ICD intensity score. In this case, ESLI (also referred to as a prediction score in FIG. 8C) is equal to 0.8637(prediction score 1) +0.0233 (prediction score 2). A Kaplan Meier survival curve was generated based on the ESLI value calculated from the automated data (FIG. 8C). An ESLI value of > 7.08 was suitable for identifying patient samples having a high risk for disease recurrence (i.e., those having a poor prognosis of prostate cancer) and an ESLI value of < 7.08 was suitable for identifying patient samples having a low risk of disease recurrence (i.e., those having a good prognosis of prostate cancer). The P value of the ESLI value in the cox regression analysis was 1.04e-05. The P value of the log-rank test for the difference between the identified high and low risk groups was 3.87e-10. [00140] Although the above description includes reference to certain specific

embodiments, various modifications thereof will be apparent to those skilled in the art. Any examples provided herein are included solely for the purpose of illustration and are not intended to be limiting in any way. Any drawings provided herein are solely for the purpose of illustrating various aspects of the description and are not intended to be drawn to scale or to be limiting in any way. The scope of the claims appended hereto should not be limited by the preferred embodiments set forth in the above description, but should be given the broadest interpretation consistent with the present specification as a whole. The disclosures of all prior art recited herein are incorporated herein by reference in their entirety.