Hypermethylation of CD44 promoter in Head and Neck Squamous Cell

Carcinoma

The research resulting in this invention was supported by a grant from the National Institutes of Health, National Cancer Institute Grant number

5RO3CA107828. The U.S. Government has certain rights in the invention. BACKGROUND

1. Field of the Invention

The invention relates to measurement of hypermethylation of CD44 and the use of a panel of biomarkers including methylated CD44 promoter for the diagnosis and prognosis of head and neck squamous cell carcinoma (HNSCC). In particular, methods and kits are provided that measure hypermethylation of CD44 promoter for the diagnosis and prognosis of HNSCC.

2. Background Information Head and neck squamous cell carcinoma (HNSCC) accounts for almost 90% of cancers involving the upper aerodigestive tract (UADT). In the United States in 2005, cancers of the oral cavity, pharynx and larynx are expected to account for nearly 3% of incident cancers and 2% of cancer deaths. There are approximately 500,000 new cases diagnosed world-wide each year. Men are affected over two times more than women. Over half of these cancers involve the oral cavity. The rest are divided equally between larynx and pharynx.

There is no effective early detection program for head and neck squamous cell carcinoma (HNSCC). A recent study from India shows a survival advantage for screening by oral cavity exam. However sensitivity and specificity of this method are only 75%. Screening by physical exam is expensive, skill-dependent, and cannot detect occult disease. Poor detection practices likely contribute to the poor survival noted in black males and patients from lower socioeconomic status. Access to skilled practitioners may pose a challenge to individuals with limited material resources.

Five-year survival rates for HNSCC are low and have not improved in several decades. Moreover, patients with this disease experience severe morbidity including disfigurement, speech, swallowing and breathing problems. Late stage of diagnosis and propensity to recur are challenges that thwart efforts to improve outcomes in these patients. These challenges are more pronounced in black patients compared to

white patients and economically disadvantage*! populations compared to wealthy populations. Effective early detection programs that are targeted to high-risk populations may result in diagnosis of a higher proportion of patients with early stage disease and therefore better outcomes. It has been previously found that elevated levels of soluble CD44 (solCD44) are found in biological samples obtained in patients with HNSCC, and that solCD44 can be used as a diagnostic and prognostic marker for HNSCC (see, e.g. U.S. Pat. Appln. No. 11/090,705, filed March 28, 2005, which is hereby incorporated by reference). In addition, it has been found that hyaluronic acid (HA); hyaluronidase (HAase), and total protein are elevated in biological samples of HNSCC patients (see, e.g. PCT Pat. Appln. No. PCT/US2007/011511, filed July 24, 2007, hereby incorporated by reference). However, a need remains to detect a sub-population of . HNSCC patients who exhibit false-negative results on such tests, e.g. patients with HNSCC who do not show elevated levels of the aforementioned biomarkers. This application claims priority to U.S. provisional application no. 60/845,528, filed September 19, 2006, which is hereby incorporated by reference. All publications and patent documents cited in this application are incorporated by reference in pertinent part for all purposes to the same extent as if each individual publication or patent document were so individually denoted. By their citation of various references in this document, Applicants do not admit any particular reference is "prior art" to their invention.

SUMMARY The invention relates to a method of diagnosing head and neck squamous cell carcinoma (HNSCC) in a subject that comprises measuring hypermethylation of CD44 promoter in a biological sample (e.g. saliva or oral rinse) obtained from the subject.

It has been found that many such patients exhibit hypermethylation of CD44 promoter, and that methylated CD44 promoter will be a useful diagnostic and prognostic marker.

Accordingly, it is one object of the invention to provide a method of diagnosing head and/or neck squamous cell carcinoma (HNSCC) in a subject, comprising measuring hypermethylation of CD44 promoter in the subject. According to the invention, an elevated level of methylation of CD44 promoter in the subject is

indicative of increased likelihood of the presence of HNSCC. Level of methylation of CD44 promoter in the subject can be compared to a baseline level of a control population of subjects, or to prior samples obtained from the same individual. Thus, an at-risk individual may be tested at preselected intervals of time for an elevation over his/her own past levels of methylation. This should also be a useful method for detection of early stage HNSCC and/or gauging the success and predicting the outcome of treatment.

The methods described herein can be used alone or in combination with other methods for diagnosis of HNSCC. Thus, in one embodiment, a method of diagnosing head and/or neck squamous cell carcinoma (HNSCC) in a subject is provided, the method comprising measuring hypermethylation of CD44 promoter in the subject, along with at least one additional biomarker, for example solCD44, wherein elevated levels of these markers are indicative of an increased likelihood of the presence of HNSCC. These measurements are preferably made on the same biological sample, e.g. saliva or oral rinse, but can be made on different samples, or different types of samples (e.g. saliva and blood) if desired. The results of the measurements will be compared with controls, an increase in either solCD44 or methylation of CD44 promoter above a selected baseline being indicative of the increased probability of the presence of HNSCC. Means for establishing such baselines using routine experimentation are well known to those of skill in the art.

In further embodiments, the method may also include the measurement of hyaluronic acid (HA), hyaluronidase, protein and/or interleukin-8 (IL-8) in accordance with the inventors' previous discovery that these substances are also biomarkers for HNSCC. Elevation of any of the tested substances may be indicative of the presence of HNSCC or of tumor stage, with higher levels expected to be correlated with more advanced tumors. Levels may also be indicative of the effectiveness of treatment, decreased levels of the tested substances over time being indicative of treatment progress and/or favorable prognosis.

The methods may also be used to predict which patients will respond to specific therapies. For example, elevated levels of solCD44 or methylated CD44 promoter are expected to identify patients who will be responsive to therapies that target CD44, e.g. anti-CD44 antibodies, and the like.

In yet another embodiment, a kit for diagnosing/detecting HNSCC or elevated risk thereof in a subject is provided. The kit may also be used for measuring treatment success or predicting recurrence of HNSCC.

In one form, the kit comprises agents and/or means for detecting methylated CD44 promoter, and optionally means for detecting additional biomarkers such as solCD44, hyaluronic acid (HA), hyaluronidase, protein and interleukin-8 (IL-8). Kits of the invention may, for example, contain suitable PCR reagents for detection of methylated CD44 promoter, e.g. specific primers. Means for detection also include antibodies specific for the biomarkers, optionally labeled with fluorescent, colorimetric or radioactive labels to facilitate the detection of antibody-antigen complexes. The kits of the invention may contain other components, such as a substrate or container for collecting a biological sample (e.g saliva, or an oral rinse), printed instructions, specific antibodies, etc.

The kit may contain one or more of: reference standard(s) of biomarker(s) in solution or solid form, one or more antibodies specific for the biomarkers, and directions for carrying out detection assay(s) for the biomarkers with the contents of the kit.

In one specific embodiment, the kit comprises means for detecting methylated CD44, HA, HAase, solCD44 and total protein. In another specific embodiment, the kit comprises: (a) a substrate or container for holding a biological sample isolated from a human subject suspected of having HNSCC, or of being at risk thereof; (b) an agent for detecting methylated CD44; (c) a fluorogenic agent that detects at least one biomarker; (d) a panel of biomarkers; and optionally, (e) printed instructions for reacting the agent with the biological sample or a portion of the biological sample to detect the presence or amount of at least one biomarker in the biological sample. Preferably, the kit comprises a panel of biomarkers of any one or more of: HA and hyaluronidase to be used as standards, along with means of detecting HA and HAase and total protein. The kit may also contain a standard for and/or means for detecting CD44, in particular solCD44. Optionally, the kit comprises antibodies specific for any one or more of biomarkers: HA, HAase, and CD44.

Other aspects are described infra.

Aspects of the invention have been described in Franzmann et al. Cancer Epidemiol. Biomarkers Prev. 2007; 16(7):1348-1354.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 shows that 9/11 HNSCC with low solCD44 were hypermethylated and no normal control subjects showed hypermethylation. PC= positive control, NC= negative control.

DETAILED DESCRIPTION Definitions

The following terms are used as defined below throughout this application, unless otherwise indicated. "Marker" or "biomarker" are used interchangeably herein, and in the context of the present invention refer to a compound or substance (e.g. a polypeptide, protein, nucleic acid) which is differentially present in a sample taken from patients having head and neck squamous cell carcinoma (HNSCC) as compared to a comparable sample taken from control subjects (e.g., a person with a negative diagnosis for HNSCC, a normal or healthy subject). (It will be clear to those of skill in the art that control subjects will in general be matched for characteristics such as smoking and consumption of alcohol, but may be afflicted with conditions other than HNSCC.)

The phrase "differentially present" refers to differences in the quantity and/or the frequency of a marker present in a sample taken from patients having for example, head and neck squamous cell carcinoma (HNSCC) as compared to a control subject, or more preferably a baseline established for control subjects. It is well known in the art how to establish such baselines. For example, a marker can be a polypeptide which is present at an elevated level or at a decreased level in samples of patients with head and neck squamous cell carcinoma (HNSCC) compared to samples of control subjects. Alternatively, a marker can be a polypeptide which is detected at a higher frequency or at a lower frequency in samples of patients compared to samples of control subjects. A marker can be differentially present in terms of quantity, frequency or both.

A marker, compound, composition or substance is differentially present between the two samples if the amount of the marker, compound, composition or substance in one sample is statistically significantly different from the amount of the marker, compound, composition or substance in the other sample. For example, a compound is differentially present between the two samples if it is present at least about 120%, at least about 130%, at least about 150%, at least about 180%, at least

about 200%, at least about 300%, at least about 500%, at least about 700%, at least about 900%, or at least about 1000% the level that is present in the other sample (i. e. increased by 20%, 30%, etc.) or if it is detectable in one sample and not detectable in the other. Alternatively or additionally, a marker, compound, composition or substance is differentially present between the two sets of samples if the frequency of detecting the polypeptide in samples of patients' suffering from head and neck squamous cell carcinoma (HNSCC), is statistically significantly higher or lower than in the control samples. For example, a biomarker is differentially present between the two sets of samples if it is detected at least about 20%, at least about 30%, at least about 50%, at least about 80%, at least about 100%, at least about 200%, at least about 400%, at least about 600%, at least about 800%, or at least about 900% more frequently; or at least about 20%, at least about 30%, at least about 50%, at least about 80%, at least about 90%, at least about 95%, or at least about 99%, less frequently, observed in one set of samples than the other set of samples. These exemplary values notwithstanding, it is expected that a skilled practitioner can determine cut-off points, etc. that represent a statistically significant difference to determine whether the marker is differentially present.

By "methylated CD44 promoter" is meant methylation of the promoter region of the CD44 gene. CD44 methylation may be measured by any suitable means known in the art, include gel based analysis, methylation specific polymerase chain reaction (MS-PCR), etc. (See, e.g., Kagan J, Srivastava S, Barker PE,Belinsky SA, Cairns P. Towards clinical application of methylated DNA sequences as cancer biomarkers: a joint NCI's EDRN and NIST workshop on standards, methods, assays, reagents and tools. Cancer Res 2007;67:4545-9.)

By "hypermethylation" or "elevated level of methylation" is meant an increase in methylation of CD44 promoter that is considered statistically significant over levels of a control population. Persons of skill in the art will be able to determine these testing parameters by routine experimentation. Furthermore, "hypermethylation" or "elevated level of methylation" may refer to increased levels seen in the tested individual over time.

DNA hypermethylation occurs by enzymatic addition of a methyl group to the carbon-5 position of cytosine. The majority of methylated cytosines occur as 5'-CpG- 3' dinucleotides that are distributed in islands associated with housekeeping genes and

genes with tissue specific patterns of expression. Normally these islands are unmethylated. When hypermethylated, gene suppression occurs.

"Diagnostic" means identifying the presence or nature of a pathologic condition and includes identifying patients who are at risk of developing HSCC. Diagnostic methods differ in their sensitivity and specificity. The "sensitivity" of a diagnostic assay is the percentage of diseased individuals who test positive (percent of "true positives"). Diseased individuals not detected by the assay are "false negatives." Subjects who are not diseased and who test negative in the assay, are termed "true negatives." The "specificity" of a diagnostic assay is 1 minus the false positive rate, where the "false positive" rate is defined as the proportion of those without the disease who test positive. While a particular diagnostic method may not provide a definitive diagnosis of a condition, it suffices if the method provides a positive indication that aids in diagnosis.

The terms "detection", "detecting" and the like, may be used in the context of detecting biomarkers, or of detecting HNSCC (e.g. when positive assay results are obtained). In the latter context, "detecting" and "diagnosing" are considered synonymous.

A "test amount" of a marker refers to an amount of a marker present in a sample being tested. A test amount can be either in absolute amount (e.g., μg/ml) or a relative amount (e.g., relative intensity of signals).

A "diagnostic amount" of a marker refers to an amount of a marker in a subject's sample that is consistent with a diagnosis of head and neck squamous cell carcinoma (HNSCC). A diagnostic amount can be either in absolute amount (e.g., μg/ml) or a relative amount (e.g., relative intensity of signals). A "control amount" of a marker can be any amount or a range of amount which is to be compared against a test amount of a marker. For example, a control amount of a marker can be the amount of a marker in a person without head and neck squamous cell carcinoma (HNSCC). A control amount can be either in absolute amount (e.g., μg/ml) or a relative amount (e.g., relative intensity of signals). The terms "polypeptide," "peptide" and "protein" are used interchangeably herein to refer to a polymer of α-amino acid residues, in particular, of naturally- occuring α-amino acids. The terms apply to amino acid polymers in which one or more amino acid residue is an analog or mimetic of a corresponding naturally- occurring amino acid, as well as to naturally-occurring amino acid polymers.

Polypeptides can be modified, e.g., by the addition of carbohydrate residues to form glycoproteins. The terms "polypeptide," "peptide" and "protein" include glycoproteins, as well as non-glycoproteins.

"Detectable moiety" or a "label" refers to a composition detectable by spectroscopic, photochemical, biochemical, immunochemical, or chemical means. For example, useful labels include ³² P, ³⁵ S, fluorescent dyes, electron-dense reagents, enzymes (e.g., as commonly used in an ELISA), biotin-streptavidin, dioxigenin, haptens and proteins for which antisera or monoclonal antibodies are available, or nucleic acid molecules with a sequence complementary to a target. The detectable moiety often generates a measurable signal, such as a radioactive, chromogenic, or fluorescent signal, that can be used to quantify the amount of bound detectable moiety in a sample. Quantitation of the signal is achieved by, e.g., scintillation counting, densitometry, or flow cytometry.

"Antibody" refers to a polypeptide ligand substantially encoded by an immunoglobulin gene or immunoglobulin genes, or fragments thereof, which specifically binds and recognizes an epitope (e.g., an antigen). The recognized immunoglobulin genes include the kappa and lambda light chain constant region genes, the alpha, gamma, delta, epsilon and mu heavy chain constant region genes, and the myriad immunoglobulin variable region genes. Antibodies exist, e.g., as intact immunoglobulins or as a number of well characterized fragments produced by digestion with various peptidases. This includes, e.g., Fab' and F(ab)' ₂ fragments. The term "antibody," as used herein, also includes antibody fragments either produced by the modification of whole antibodies or those synthesized de novo using recombinant DNA methodologies. It also includes polyclonal antibodies, monoclonal antibodies, chimeric antibodies, humanized antibodies, or single chain antibodies. "Fc" portion of an antibody refers to that portion of an immunoglobulin heavy chain that comprises one or more heavy chain constant region domains, CHi, CH ₂ and CH ₃ , but does not include the heavy chain variable region.

By "binding assay" is meant a biochemical assay wherein at least one biomarker is detected by binding to an agent, such as an antibody, through which the detection process is carried out. The detection process may involve radioactive or fluorescent labels, and the like. The assay may involve immobilization of the biomarker, or may take place in solution.

"Immunoassay" is an assay that uses an antibody to specifically bind an antigen (e.g., a marker). The immunoassay is characterized by the use of specific binding properties of a particular antibody to isolate, target, and/or quantify the antigen. The phrase "specifically (or selectively) binds" to an antibody or "specifically

(or selectively) immunoreactive with," when referring to a protein or peptide, refers to a binding reaction that is determinative of the presence of the protein in a heterogeneous population of proteins and other biologies. Thus, under designated immunoassay conditions, the specified antibodies bind to a particular protein at least two times the background and do not substantially bind in a significant amount to other proteins present in the sample. Specific binding to an antibody under such conditions may require an antibody that is selected for its specificity for a particular protein. A variety of immunoassay formats may be used to select antibodies specifically immunoreactive with a particular protein. For example, solid-phase ELISA immunoassays are routinely used to select antibodies specifically immunoreactive with a protein (see, e.g., Harlow & Lane, Antibodies, A Laboratory Manual (1988), for a description of immunoassay formats and conditions that can be used to determine specific immunoreactivity).

The terms "subject", "patient" or "individual" generally refer to a human, although the methods of the invention are not necessarily limited to humans, and should be useful in other mammals.

"Sample" is used herein in its broadest sense. A sample comprising polynucleotides, polypeptides, peptides, antibodies fragments and derivatives thereof may comprise a bodily fluid; a soluble fraction of a cell preparation, or media in which cells were grown; a chromosome, an organelle, or membrane isolated or extracted from a cell; genomic DNA, RNA, or cDNA, polypeptides, or peptides in solution or bound to a substrate; a cell; a tissue; a tissue print; a fingerprint, skin or hair; fragments and derivatives thereof.

The subject sample may be selected, for example, from the group consisting of saliva, an oral rinse, blood, blood plasma, serum, urine, tissue, cells, and liver.

Preferably, the sample is saliva or an oral rinse. Saliva can be collected using many methods. One common method is whole saliva collection. Saliva is collected, often over a set period of time, from the anterior oral cavity, where the majority is released under resting conditions. Oral rinses involve use of a set amount of a fluid, often

saline, that is manipulated in the mouth and helps release substances adherent to the lining of the oral cavity, larynx and pharynx. It is theorized that whole saliva may reflect systemic expression of substances while oral rinses are more reflective of local expression of substances. By "at risk of is intended to mean at increased risk of, compared to a normal subject, or compared to a control group, e.g. a patient population. Thus a subject "at risk of developing HNSCC is at increased risk compared to a normal population, and a subject "at risk of a recurrence of HNSCC may be considered at increased risk of having a recurrence as compared to the risk of a recurrence among all treated HNSCC patients

"Increased risk" or "elevated risk" mean any statistically significant increase in the probability, e.g., that the subject will develop HNSCC, or a recurrence thereof. The risk is preferably increased by at least 10%, more preferably at least 20%, and even more preferably at least 50% over the control group with which the comparison is being made.

"CD44 marker" is intended to include soluble CD44 and isoforms thereof. By "not clinically apparent" is meant an HNSCC tumor that has not yet reached the stage of being detectable upon physical examination.

Tumor staging: In another embodiment, detection of at least one biomarker is diagnostic of tumor staging. It has been shown, for example, in bladder cancer that certain biomarkers are indicative of aggressive tumors (Lokeshwar et al., 2000, J. Urol. 163:348-56), and it is expected that the presence/amount of the biomarkers disclosed herein will be indicative of tumor staging in HNSCC.

Saliva as a screening medium: Saliva is becoming a well-accepted screening medium for various disease processes. It has an advantage over blood because it is readily accessible and noninvasive. The average daily production of whole saliva varies between 1 and 1.5 liters. Components of whole saliva include blood and blood derivatives from intraoral bleeding and gingival crevicular fluid, extrinsic substances such as food, epithelial lining cells, microbes, bronchial, nasal, salivary gland secretions. The majority of saliva, in the unstimulated state, originates from submandibular glands (65%) with 20% from the parotid gland and the remainder from sublingual and minor salivary glands located throughout the upper aerodigestive tract (UADT). Ninety-nine percent water, saliva contains a variety of electrolytes, immunoglobulins, proteins, enzymes, mucins, and nitrogenous products and is

hypotonic especially in the unstimulated state. Normal pH ranges from 6-7. The salivary flow rate is influenced by the size of the salivary glands, hydration status, nutritional state, stimulus, and gender. Total protein concentrations of whole saliva in the unstimulated state give an accurate indication of the hydration state of an individual. Saliva is typically assayed as the product of an oral rinse, as described, for example, below. Detection of Biomarkers

The biomarkers can be detected using a protein assay, binding assay, an immunoassay, or any other suitable assay known to those of skill in the art. Exemplary assays are described in detail in the examples which follow, and in U.S. patent application no. 11/090,705 and PCT Pat. Appln. No. PCT/US2007/011511. For a positive diagnosis of HNSCC or increased risk thereof, the biomarkers and/or total protein detected are elevated as compared to a normal healthy control, a group of controls, a baseline established in a patient population not afflicted with HNSCC, etc. SolCD44: CD44 comprises a family of isoforms expressed in many cell types.

These isoforms arise from alternative splicing of a region of variable exons (exons 5- 14) present in CD44 mRNA. They differ in primary amino acid sequence as well as in amount of N- and O- glycosylation. Isoforms are found in normal cells as CD44 standard (CD44s), CD44 epithelial (CD44E) or CD44v8-10, and CD44v3-10 in keratinocytes. Other CD44 variant isoforms (CD44v) are differentially expressed in some tumors. CD44 mediates a direct link between the extracellular matrix and the cytoskeleton via their conserved extracellular HA binding regions and intracellular ankyrin binding regions. CD44 proteins are also released in soluble form (solCD44) via proteases and are detectable in normal circulation and saliva. Detection methods using solCD44 are described, e.g., in U.S. Appl. No. 11/090,705, filed March 28, 2005, which is incorporated herein by reference.

Overexpression of normally expressed isoforms also promotes oncogenesis. CD44 transfection increases migration and confers metastatic potential to some cell types, while blocking cell surface CD44 binding to HA reduces tumor cell growth and migration. CD44 associates with other molecules to mediate oncogenic signaling. These include members of the ERBB family of receptor tyrosine kinases such as ERBB 1 and ERBB2. CD44 also functions as a platform for growth factors and members of the matrix metalloproteinase (MMP) family of enzymes, further contributing to signaling events. One member of the MMP family, membrane-type 1

MMP (MTl- MMP) cleaves CD44 to its soluble form. This cleavage results in increased cell migration. While MTl-MMP appears to be one of the main proteases involved in CD44 cleavage, there is evidence that others exist.

Hyaluronic Acid: HA is a nonsulfated glycosaminoglycan (GAG), overexpressed in certain cancers. HA is synthesized by hyaluronan synthase on the surface of cells and is comprised of repeating disaccharide units of D-glucuronic acid and N-acetyl-D-glucosamine. It is present in body fluids, tissues, and extracellular matrix. It interacts with cell surface receptors (e.g., CD44, RHAMM, etc.) and, through these interactions, regulates cell adhesion, migration, and proliferation. Depending upon the type of tumor, HA may be synthesized by stromal cells, tumor cells or both. In tumor tissues, HA supports metastasis by promoting tumor cell migration, offering protection against immune surveillance and causing a partial loss of contact-medicated inhibition of cell growth and migration. Small fragments of HA are angiogenic and have been isolated from urine of bladder cancer patients, prostate cancer tissue, and saliva from HNSCC patients. Concentrations of HA are elevated in several cancers, including colon, breast, prostate, bladder and lung. Tissue expression of HA in tumors such as colon and breast, indicates a poor prognosis.

Hyaluronidase: HAase is an endoglycosidase that degrades HA into small angiogenic HA fragments. HA and HA fragments stimulate endothelial cell proliferation, adhesion and migration by activating the focal adhesion kinase and MAP kinase pathways. HAase alters the expression of CD44 isoforms and is associated with increased tumor cell cycling. Of the 6 human HAases encoded by different genes, three are characterized at the protein level.

Methylated CD 44 promoter. Our previous studies indicated that solCD44 was elevated in almost 80% of the HNSCC patients and none of the controls. However, immunohistochemistry studies in HNSCC show that approximately 10-30% of HNSCC underexpress CD44. To be most useful, it is preferred that a screening test have sensitivity and specificity above 90%. A panel of markers that complement solCD44 would increase the odds of achieving this. We examined oral rinses with low solCD44 and found methylated CD44 in 9/11 HNSCC and 0/10 controls with similar risk factors and demographics, supporting a conclusion that methylation may be associated with low solCD44 levels in HNSCC and thus complement the solCD44 test.

Test panel: CD44, methylated CD44, total protein, IL-8, HA and HAase comprise a related group of complimentary markers involved in tumorigenesis that are detectable in oral rinses, and should be particularly useful in combinations for diagnosis of HNSCC.

Kits

The assays of the present invention are ideally suited for the preparation of kits. Such a kit may comprise a carrier means being compartmentalized to receive in close confinement there with one or more container means such as vials, tubes and the like, each of said container means comprising the separate elements of the immunoassay. For example, there may be a container means containing a first antibody immobilized on a solid phase support, and a further container means containing a second detectably labeled antibody in solution. Further container means may contain standard solutions comprising serial dilutions of the HNSCC biomarkers to be detected, or appropriate quantities of the biomarkers in dry or concentrated form to be made up into standard solutions by the end user. The standard solutions of

HNSCC biomarkers may be used to prepare standard curves with the concentration of each HNSCC biomarker plotted on the abscissa and the detection signal on the ordinate. The results obtained from a sample containing any one of the HNSCC biomarkers may be interpolated from such a plot to give the concentration of each detected biomarker.

In one embodiment, a kit for diagnosing HNSCC or elevated risk thereof in a subject comprising means for determining the presence of hypermethylated CD44 in a subject sample is provided, the kit comprising (a) a substrate for holding a biological sample, (b) one or more means for measuring methylated CD44, such as primer sequences specific for modified (methylated) DNA, methylation-sensitive restriction enzyme(s), and optionally, further detection means for the products of these reactions. The kit may also contain, for example, reagents for DNA purification and for the bisulfate conversion reaction.

The kit may also include standards for additional biomarkers to be detected (e.g. CD44, HA, HAase, protein), one or more fluorogenic agents that detect biomarkers; and printed instructions for reacting the agent(s) with the biological sample or a portion of the biological sample to detect the presence or amount of at least one marker in the biological sample. Preferably, the kit comprises a panel of

biomarkers of any one or more of hyaluronic acid (HA); hyaluronidase, and CD44, in addition to reagents for detection of methylated CD44. Optionally, the kit further comprises antibodies specific for any one or more biomarkers: HA, HAase, and CD44, and means for determining total protein. For a positive diagnosis based on the results of using the kit, at least one biomarker in a patient is elevated as compared to a normal healthy control.

The kit can provide both a panel of HNSCC biomarkers, e.g. to be used for standard curves, and antibodies thereto if desired. The kit will detect biomarkers using antibodies or other suitable detection methods. Examples

General Methology

Selection of Subjects: One hundred and two HNSCC patients and 69 controls with benign disease of the UADT (hereafter referred to as benign disease) were enrolled from otolaryngology clinics at the University of Miami Sylvester Comprehensive Cancer Center (UM/SCCC) and Jackson Memorial Hospital (JMH). We used benign disease patients as our main control group to determine whether common benign diseases confound results. An additional 15 control patients were enrolled as normal volunteers. All subjects were enrolled according to the protocol approved by the Institutional Review Board. To ensure that benign disease patients included mainly smokers and drinkers (as is true of the HNSCC population), they were approached if they answered "yes" to tobacco or alcohol use on the clinic intake questionnaire. Control patients were excluded if they had a potentially malignant lesion or if final diagnosis of their condition was unknown. All HNSCC patients had biopsy proven newly diagnosed or recurrent HNSCC. We included all stages and sites except nasopharynx.

Collection of oral rinse: Samples are collected from patients at the clinic or screening site. For collection, five milliliters of normal saline is placed in the subject's mouths. Patients are asked to swish for five seconds, gargle for five seconds and then spit into a specimen cup. Saliva is placed on ice for transport and stored at - 80 degrees. As recommended, subjects are asked to refrain from oral hygiene procedures, smoking, eating and drinking for at least 1 hour prior to collection. Samples are stored on ice for transport since solCD44 (including methylated), HA and HAase levels are stable on ice for 8 hours prior to freezing at -80°C. The samples may be fractioned to permit multiple investigations without freeze-thaw cycles and

stored at -8O ⁰ C. Although neither fractioning nor successive freeze-thaw cycles have a significant effect on measurement of solCD44, methylated CD44 promoter, HA or HAase levels, freeze-thaw cycles are avoided so that the samples can be used for future analysis of other tumor markers. It is important that saliva samples that are obtained, have had contact with all mucosal surfaces of interest. HNSCC patients with large tumors, significant pain or tracheotomy tubes may have difficulty gargling, which may contribute to an increased false negative rate. To control for this, normal controls and HNSCC patient's gargles may be graded on a scale from 0 to 2, with "2" being an effective gargle. One is a weak gargle, and zero is inability to gargle. These data can also be analyzed to determine if poor gargling is associated with lower levels of markers in tumor patients and normals.

Freeze-thaw cycles and stability: Three samples were aliquotted into five tubes and stored at -80°. SolCD44 levels were tested for each aliquot of a sample and coefficient of variation between aliquots of the same sample was determined. Aliquots were then taken through five freeze-thaw cycles. SolCD44 levels were measured for each sample at one, three and five freeze-thaw cycles and the coefficient of variation between freeze-thaw cycles of the same sample was measured. Coefficients of variation between aliquots and between freeze-thaw cycles were similar and below 20%. To further test stability, two oral rinse samples were aliquotted and stored on ice for 8 hours then placed in the -80° C freezer or placed immediately in the -80° C freezer. Average coefficient of variation between samples stored on ice versus immediately placed in -80° C freezer was 8.6% indicating that storage on ice does not lead to significant degradation. Statistical Analysis: Variance in duplicate measures (analytical variability) is obtained by squaring the difference in duplicates and dividing by 2. The mean value of these duplicate variances (S _A ² ) can be determined. Then, using the formula to determine coefficient of variation, the square root of the mean variance divided by the overall mean marker level multiplied by 100 will give CV _A . To estimate CVi, the following formula can be used, where Si ² and SA ² correspond to mean biologic within subject variance and mean analytical variance, respectively.

Si ² + S _A ² /2= mean measured within subject variance

Mean measured within subject variance is determined from the variance between samples collected two weeks apart. CVi is then determined using the formula for CV as above.

To estimate CVQ, the variance in the true means, S _G ² is calculated using the following formula. Then CVQ is calculated as previously discussed.

SQ = [Measured mean square between subject variance - (S _A + 2Si )]/4

Once the values for CV _A , CVi, and CV _A are determined they can be incorporated into the following formulas to evaluate endpoints.

Evaluation of endpoints: The criteria to set performance standards was developed by Cotlove and is widely accepted (Fraser CG et al. Critical Rev Clin Lab Sci 1989;27:409-437). The criterion states that the maximum analytic variation should be less than or equal to half of the average within subject biologic variation. This is represented in the formula below where CV is coefficient of variation.

CV _A <l/2CVi

Satisfaction of this criterion results in no more than a 12% increase in the measured over the true within subject variation. This will provide as with a measure to determine whether our analytic variability is sufficiently low.

Determining the significance in changes between serial measurements: The significance of changes in sequential results will be determined, using the index of heterogeneity as described by Fraser and Harris (Critical Rev Clin Lab Sci

1989;27:409-437). If the heterogeneity is nonsignificant then the median of observed within subject variances will be used to determine a significant difference ( p<0.05) between samples using the formula below.

2.77(CVA ² + CV, ² ) ^1/2

If the index of heterogeneity is significant then a distribution of true within subject variances can be developed and the upper percentile points used to calculate the critical difference.

For screening purposes, an individual marker level is determined as positive or negative based on a population based reference level. If the within-subject variation is small compared to the between-subject variation, a significant change in an individual marker level may not be perceived as significant when using population-

based cut-off point. The following formula, which is an accepted index to determine whether individual values can be compared usefully with reference values, can be used:

(CV ² ,+ CV ² _A ) ¹⁷ VCV _G O-O Harris has shown that if index levels are less that 0.6 the population-based reference value is usually insensitive to significant fluctuations in an individual subject. In this case the probability that an observed level will fall within the conventional normal range is greater than the specified probability that is derived from the distribution of the population as a whole (Harris E. K. Clin Chem 1974;20:1535-42). When the reference values is insensitive to individual changes, following marker levels over time in an individual is usually more useful than comparing one set of marker levels to a reference level.

Three components of variation can be determined: the homogeneity of 1) variances in the 60 duplicate measures (analytic variability), 2) variances in the 60 repeat collections (within subject variability), and 3) mean marker level (between subject variability) for 30 subjects.

In order to determine outliers for the variances in each subject's replicate measures, and within subject variances, Cochran's test can be used to test the ratio of the maximum variance to the sum of the variances. To test for outliers in between subject variability, Reed's criterion can be used (Reed AH, Henry RJ and Mason Va. Clin Chem 1971; 17:275-84). This criterion analyzes the difference between the extreme value and the next highest (or lowest value). The value is rejected if the difference exceeds one-third the range of all values. This criterion assumes that the true distribution of values for a given parameter are normal. Selecting the best biomarker model for HN screening: In order to determine the combination of biomarkers yielding the highest sensitivity and specificity for detecting HNSCC cancer, the methodology of Li, et al. will be followed (Salivary transcriptome diagnostics for oral cancer detection. Clin Cancer Res 2004; 10:8442- 50.). Here we will be evaluating all five markers to determine the best combination for differentiating cancer from non-cancer. Receiver operator characteristic (ROC) curve analysis will be performed for each biomarker to evaluate the predictive power of each. Since the high-risk control group represents the population targeted for screening, this group and our cases to determine optimal combinations of biomarkers. Area under the curve (AUC) will be computed for each biomarker, and the one with

the largest AUC will be selected as the biomarker having the highest predictive power for detecting HNSCC cancer. Next, using logistic regression (Hosmer DW and Lemeshow S. Applied Logistic Regression, 2nd ed. Wiley, New York, 2000), multivariate classification models will be constructed to determine the best combination of biomarkers for cancer prediction. An indicator for promoter hypermethylation will be included in the multivariate classification models. Additionally, logistic regression will allow us to calculate the association of each biomarker on the dependent variable (cancer/non-cancer) singly and in combination while controlling for potential covariates, such as age, gender, and smoking history. Backward stepwise logistic regression will be used to find the best final model. All 150 from each group will be used for logistic modeling and marker selection. In order to determine the estimated future prediction error in each model, 10-fold Cross Validation will be used as described by Feng and Yasui (Statistical considerations in combining biomarkers for disease classification. Dis Markers. 2004;20:45-51). Briefly, the total sample (n=300) will be split into 10 equal parts. Leaving one part out, the remaining 9 parts will used to fit our multivariate logistic model. The parameters from this model will then be applied to the remaining one part to determine the prediction error. This process will be repeated for each of the 10 parts, and the observed predication errors will be averaged, i.e., the Cross-Validation Error. A final ROC curve can then be computed from the best final logistic model, i.e., the one with the smallest Cross-Validation Error, using the fitted probabilities from the model as possible cutpoints for computation of sensitivity and specificity. Based on our findings, cutpoints will be determined to yield the best trade-off between sensitivity and specificity. Then, this model will also be applied to the low-risk, healthy controls to estimate true specificity. To avoid overfitting, when the final model with cut-points is selected, all further manipulations will be suspended. Then to assess the future prediction error we will use the bootstrap method to test the final model as described by Feng and Yasui. Cluster analysis will be used to produce a classification tree for the entire group (cancers and controls) using the biomarkers and any important covariates (i.e., smoking status) as predictors (Rencher A. Methods of Multivariate Analysis, 2nd Ed. 2002. John Wiley & Sons, Inc.).

Example 1 : Detection of CD44 hypermethylation

We performed methylation specific PCR on 36 HNSCC patients and 49 controls. The groups were balanced with respect to tobacco and alcohol exposure, age and gender. Oral rinse samples were collected and stored as described elsewhere herein. We optimized the test for oral rinses by preparing the standards in a synthetic saliva matrix (Salimetrics) diluted 1 :5 in normal saline (since patients swish and gargle with 5 cc saline) and use a sample diluent (Salimetrics) developed for saliva samples. Samples were vortexed, centrifuged at 3,000 G and the pellet saved (e.g. frozen) for future DNA analyses. DNA was isolated from the pellets using a QIAmp DNA mini Kit (Qiagen, Maryland) according to the manufacturer's protocol. Bisulfite treatment of 2μg of DNA was done using previously described methods (Singal et. al. 1997, Proc Natl Acad Sci USA). Briefly, DNA was denatured with 2N NaOH at 37°C for 15 min. This was followed by bisulfate treatment for 4h at 55°C which converts methylated cytosines to uracil. To measure methylated CD44 promoter, we used methylation specific PCR (MS-PCR). This was performed using primer sequences specific for the modified DNA, but not wild-type DNA. Forward: CD44 MS-S'AGTTTTAGTAGAGTACGGGGCS ¹ . Primer: CD44 MS- 5'ACGAACGAAAAA CACACCCAAACA3'. An additional PCR reaction using primers specific for β-actin was used as a control to verify that DNA was present in the samples (Singal R, et al. Methylation of the minimal promoter of an embryonic globin gene. Proc Natl Acad Sci USA. 1997;94: 13724-9). Relative quantitative MSPCR was performed using SYBR green master mix (Bio-Rad, Hercules, CA). A standard curve was developed using DNA from the LnCap cell line and was run with each individual PCR experiment as a calibrator. The CD44 MS- PCR and β-Actin conditions include an initial melting time at 95 ⁰ C for 14 minutes, followed by 50 cycles of denaturation at 95°C for 30 seconds, annealing at 67°C for 1 minute for CD44MS-PCR and 60 ⁰ C for 1 minute for β-Actin. A melting curve was performed at the end of each run, with a discrete peak noted for CD44MS-PCR and for β-Actin at 81.5 ⁰ C. Relative amounts of CD44MS-PCR and β-Actin for each sample were determined by the standard curve. All measurements were performed in triplicate.

Any reproducible Ct value less than 35 was considered detectable. After this point the icylcer results are unreliable. One sample was excluded because there was no detectable actin. The remainder of the samples contained adequate actin and were used for analysis. Levels were normalized to actin and calibrated with the LNCap control. This method resulted in a very wide range of methylated CD44 levels, probably because the amount of contaminating normal cells varies greatly between subjects (Kagan J, et al. Towards clinical application of methylated DNA sequences as cancer biomarkers: a joint NCI's EDRN and NIST workshop on standards, methods, assays, reagents and tools. Cancer Res 2007;67:4545-9). Detectable methylated CD44 in a sample with detectable actin was considered a positive result.

The results are shown in Table 1. Methylated CD44 is significantly associated with cancer compared to benign disease and with low CD44 levels compared to high CD44 levels within the cancer group. Methylated CD44 was also associated with total protein level < 1 ng/ml (p=0.016). Methylated CD44 was not associated with cancer site, stage, node status, smoking status or drinking status within the cancer group.

Table 1. Methylated C D44

^* CD44 low defined as CD44 < 12 ng/mL

Example 2: β-Actin and methylated CD44 were measured in 11 HNSCC patients with low sol CD44 and 10 matched controls. Characteristics of the HNSCC patients were as follows:

Table 2

*τ L ==Larynx, OC= Oral Cavity, OP= Oropharynx

The cases and controls were compared with regard to race, gender, age, ability to gargle, whether they ever smoked, whether they ever used any tobacco and whether they were nondrinkers/ light drinkers versus moderate/heavy drinkers using Fisher's Exact test. There were no significant differences between the two groups.

The results, shown in Figure 1, demonstrated that 9/11 patients and no normal control subjects showed hypermethylation of CD44.

Example 3 : Panel of Biomarkers

While preliminary results showed that many of the samples negative for the solCD44 test were positive for CD44 hypermethylation, some tumors were missed. In addition, while total protein level holds promise as a tumor marker, studying additional markers may increase the chances of developing a screening test with high sensitivity and specificity. For this reason, we also have tested HA and HAase along with solCD44, methylated CD44 and total protein to examine if various combinations are likely to have high sensitivity and specificity for detecting HNSCC. We studied 12 HNSCC oral rinse specimens and 12 matched oral rinse specimens from patients with benign disease. HNSCC samples were taken in consecutive order from a randomly generated database. Subjects were excluded if they had a limited ability to gargle. There were no significant differences between the HNSCC and control groups with regard to gender, age, race, ethnicity, smoking history, alcohol consumption or oral health. Thus controls were high-risk. CD44 promoter hypermethylation was not included for samples where a pellet was not collected.

Performance of the solCD44 ELISA was described in U.S. application no. 1 1/090,705. HA concentration was measured with an ELISA-like assay that uses the competitive binding principle (Fosang AJ, et al., An ELISA plate-based assay for hyaluronan using biotinylated proteoglycan Gl domain (HA-binding region). Matrix. 1990;10: 306-13; Lokeshwar VB, et al. Tumor-associated hyaluronic acid: A new sensitive and specific urine marker for bladder cancer. Cancer Res. 1997; 57: 773-77). Serial duplicate dilutions of oral rinses were incubated in HA-coated microtiter plates with biotinylated HA binding protein. Plates were washed, HA binding protein was quantitated with an avidin-biotin detection system, and HA concentration was determined via standard graph. HAase levels were also measured using an ELISA- like assay (Stern M, et al., An ELISA-like assay for hyaluronidase and hyaluronidase inhibitors. Matrix. 1992; 12: 397-03; Pham HT, et al., Tumor-derived hyaluronidase: a diagnostic urine marker for high-grade bladder cancer. Cancer Res. 1997; 57: 778- 83). Microtiter wells coated with HA were incubated with duplicate serial dilutions of oral rinse for 16 hours in assay buffer. HA remaining on the wells was determined using the same biotinylated HA-binding protein and avidin-biotin detection system as the HA test. HAase concentration was determined via standard graph.

Table 3 shows results of several biomarkers, including methylated CD44, alone and in combination with other biomarkers. Table 3. Combined tumor marker panels

Several marker combinations appear promising. Data on the combination of CD44, HAase and methylated CD44 represents 6 HNSCC patients and 8 controls.

Precision in replicates on same plate: Because the HA and HAase assays are not as accurate for extremely high and low values, samples with very high or low values were excluded in this experiment. Average % CV was < 10 % for each marker.

Freeze-thaw cycles and stability: For each marker, 3-5 samples were aliquotted to determine whether significant changes in marker levels occur with multiple freeze-thaw cycles or after storage for 8 hours on ice. Our results showed

that all three markers are stable after multiple freeze-thaw cycles and storage on ice for 8 hours (% CV <20%).

Variability in marker levels between repeat collections: We collected oral rinses from three HNSCC patients before treatment on two separate occasions. Coefficient of variation between the two collections was averaged for the three patients for each marker. Average CV was 17.8% for HA and 17.3 %for HAase.

The examples herein are offered by way of illustration, not by way of limitation. While specific examples have been provided, the description is illustrative and not restrictive. Any one or more of the features of the described embodiments can be combined in any manner with one or more features of any other embodiments in the present invention. Furthermore, many variations of the invention become apparent to those skilled in the art upon review of the specification. The scope of the invention should, therefore, be determined not with reference to the description, but instead should be determined with reference to the appended claims along with their full scope of equivalents.