Detection of Oesophageal Cancer Field of the invention

This invention relates to detecting the presence of, or the risk of cancer, in particular oesophageal cancer. Background to the invention

Oesophageal cancer is still uncommon in the West, but has increased in incidence at a rate of 300-500% in the last 30-40 years (Shaheen, N. and Ransohoff, M. D. 2002. Journal of the American Medical Association. 287: 1972-1981), making it the most rapidly increasing cancer in the Western world. Furthermore, it is fatal in the majority of cases. In the US, it is estimated that ~14,520 cases will be diagnosed in 2005, with 13,570 deaths resulting from the disease. This high rate rivals that of pancreatic cancer and is more than 4 times that of rectal cancer (Paz, I. B., et a/ 2005. Esophageal Cancer. In: Cancer Management: A multidisciplinary approach. Eds: Pazdur, R., Hoskins, W. J., Coia, L. R. and Wagman, L. D.), due in part, to the late presentation of symptoms. Dysplasia is generally not noticed until the oesophagus lumen is narrowed to Va - Vs of its normal size, due to its elastic nature, thus symptoms may only become apparent once the tumour is a significant size. Patients frequently have either local invasion or distant metastasis at the time of diagnosis, limiting effective treatment options. The lifetime risk of this cancer is 0.8% for men and 0.3% for women, with risk increasing with age; the mean age at diagnosis is 67 years (Enzinger, P. C. and Mayer, R. J. 2003. New England Journal of Medicine. 349: 2241-2252). Oesophageal cancer is also one of the least studied cancers.

More than 90% of oesophageal cancers are either squamous cell carcinomas (SCC) or adenocarcinomas (AC). Most are located in the lower two-thirds of the oesophagus, with SCC distributed evenly over this region while AC is more commonly found at the lower third, near the gastric-oesophageal junction (Enzinger and Mayer, supra).

Adenocarcinoma (AC) has risen 4-10% per year since 1976 in the US and Europe (Paz, I. B., et a/ 2005. supra) whereas Squamous cell carcinoma (SCC), despite not increasing in incidence during this time, is still the most prevalent histological subtype of oesophageal cancer worldwide. As recently as 1975, ~75% of the cases of oesophageal cancer in the US were SCC, with the remaining 25% being made up mostly of AC. For reasons unknown, squamous cell carcinoma has decreased in both the black and white populations, particularly in men, while the rate of AC has risen by ~450% in white men and 50% among black men. Trends in

smoking, diet and obesity, the use of medication and detergents may account for some of these changes. SCC is 3 times more common in black men than white men, so there are obvious differences in the epidemiology of these 2 types of oesophageal cancer. Currently, there are probably more cases of AC than SCC in the US and Europe (Enzinger and Mayer, supra). Oesophageal cancer is 20-30 times more common in China than in the US, with a recognized belt of occurrence from NE China to the Middle East.

Of the two major types of oesophageal cancer, adenocarcinoma is the best characterized. This type has a very strong association with progressive damage to the lower oesophagus (oesophagitis), followed by gastric acid reflux disease (GERD), Barrett's oesophagus leading to adenocarcinoma. GERD, Barrett's oesophagus and adenocarcinoma are related diseases, with genetic and environmental determinants (Fitzgerald, R. C. 2005. Clinical gastroenterology and hepatology. 3: 529-537). The genetic changes are relevant in that heritable variations in germ-line DNA may influence susceptibility to cancer. There is an accumulation of somatic cell genetic and epigenetic changes within the epithelium during the metaplasia-dysplasia-carcinoma sequence, which may be an important determinant of the likelihood for cancer progression.

Development of oesophagitis represents the failure of many mucosal defenses to counteract refluxed acid or gastro-intestinal contents - the buffering capacity of alkaline saliva, oesophageal mucus and peristaltic clearance should be sufficient. If, however, reflux is persistent orfrequent, tissue damage may result. The regenerating, inflamed epithelium contains immature squamous cells that are sensitive to acid or bile damage. Symptoms of GERD are amongst the most common complaints encountered by the general physician; previously regarded merely as a nuisance, it is now known as having potentially serious complications. GERD is defined as chronic symptoms or mucosal damage produced by the abnormal reflux of gastric contents into the oesophagus (Shaheen and Ransohoff 2002, supra).

Barrett's oesophagus (BE) occurs as a complication of GERD and is characterized by the replacement (metaplasia) of the distal squamous oesophagus mucosa with a columnar-lined epithelium. A recent inclusion into the description of BE is that this columnar epithelium must contain goblet cells. In addition, the classification is split into long segment (more than 3cm - LSBE) and short segment (less than 3cm - SSBE) columnar epithelium. This is important because of the potential for progression to adenocarcinoma, via the metaplasia - dysplasia - adenocarcinoma

pathway, as described in Bernstein et al (2000. Electronic Journal of Biotechnology. 3: 167-182). Short segment Barrett's (SSBE) is found in 8-20% of adults, making it more prevalent than long segment Barrett's (LSBE) at 1%. Despite this, only 35% of adenocarcinoma will arise out of SSBE, thus LSBE may hold a higher risk for the development of cancer of the oesophagus. Approximately 6-15% of patients with GERD will develop BE and of the patients with Barrett's, 0.5-1 % will develop adenocarcinoma.

The origin of the columnar cells comprising BE is unclear, as these cells differ histologically from those of the gastric cardia, thus upward migration from the stomach does not account for the condition.

To date the most predictive risk factor for progression to adenocarcinoma is the degree of dysplasia or pre-cancerous cellular atypia in Barrett's oesophagus. Although individuals with no dysplasia are unlikely to have adenocarcinoma at subsequent surveillance endoscopies, those with high-grade dysplasia demonstrate a risk of subsequent adenocarcinoma of 25%. As biopsies are taken at random locations along the oesophagus, sampling error in individuals with a high-grade dysplasia is high. For those patients with high-grade dysplasia that undergo oesophageal resection, more than 50% of the specimens demonstrate previously unidentified adenocarcinoma (FaIk et al, 1999. Gastrointestinal endoscopy. 49: 170-176.). The assessment of grade of dysplasia, using endoscopic biopsies is the best marker to identify a high risk, but this is subjective. There are no prospective randomized data demonstrating that periodic endoscopic surveillance prolongs life expectancy or decreases cancer mortality among individuals with Barrett's oesophagus. However, Jankowski et al (1999 American Journal of Pathology. 154: 965-973) stated that cancer patients identified in endoscopic surveillance programmes have a better prognosis (5 year survival of 35-45%) compared to 5-15% for those cancer patients outside the programme. Because the efficacy is unknown, the cost effectiveness is also unknown. In addition, the absolute risk, even to Barrett's patients to cancer is very low. A problem with using reflux symptoms as a marker for increased risk of adenocarcinoma is that many of those developing cancer never experience chronic reflux - 40% in the Largergren et al (1999 New England Journal of Medicine. 340: 825-831) study did not have at least weekly reflux before developing the cancer, therefore screening of the reflux patients alone will miss 40% of those at potential risk. No association was found between reflux symptoms and squamous cell carcinoma risk.

More than 80% of SCC can be attributed to smoking and alcohol, with a synergistic association between the two. Exposure to carcinogens can be linked to SCC if the agents are activated in the oesophagus. It has been shown that this is the case for some carcinogens that are substrates for cytochrome p450, as measured by the antibody for CYP1A2. Patients without glutathione - S - transferase M1 (GSTM1) have a lower ability to detoxify carcinogens, so those patients who are heavy smokers and who are also lacking CYP1A2 and GSTM1 have been shown to have a much higher risk of developing SCC (D'Amico, T. A. and Harpole Jr., D. H. 2000. Chest Surgery Clinics of North America. 10: 451-469). Despite advances, the prognosis for oesophageal patients remains poor. An improved understanding of the molecular biology of this disease may allow improved diagnosis, therapy and prognosis (Jankowski etal, 1999, supra). Numerous molecular markers have been sought as more reliable indicators of cancer risk, but none have gained widespread acceptance. As well as the identification of abnormalities in target genes, it is also possible to determine the global gene expression profile of these diseases and to correlate this profile with clinical outcome. As the prevalence of most markers is low, the use of a panel of markers in detecting the presence of a tumour or dysplasia and determining the type would be a major step forward in the early diagnosis and treatment of oesophageal cancer. Many of the important events that happen in the development of EC first occur in tissues that are neither dysplastic nor malignant.

The identification of early molecular defects in morphologically normal appearing tissue is paramount in identifying subjects at high risk of cancer. It is hoped that these combined strategies will enable patients to be stratified in terms of their cancer risk. At this time, genetics do not influence the clinical management of this devastating form of cancer.

There is clearly a need for a reliable genetic marker for oesophageal cancer. Summary of the invention

According to a first aspect of the present invention, there is a method for detecting the presence of or the risk of cancer in a patient, comprising the step of:

(i) detecting the expression level of a gene characterised by the polynucleotide sequence identified herein as SEQ ID No. 1 or SEQ ID No. 2, or a similar polynucleotide of at least 15 consecutive nucleotides that hybridises to the complement of SEQ ID NO:1 or SEQ ID NO:2 under stringent hybridising conditions, in a genetic sample isolated from a patient, wherein the expression level of the gene

characterised by SEQ ID No.1 or SEQ ID No.2 or a similar polynucleotide, indicates the presence of or the risk of cancer in the patient from whom the genetic sample was isolated.

According to a second aspect of the invention an isolated polynucleotide comprises the nucleotide sequence identified herein as SEQ ID No;1 or SEQ ID No:2, or a complement thereof, or a polynucleotide of at least 15 consecutive nucleotides that hybridises to SEQ ID No. 1 or SEQ ID No.2 (or their complements) under stringent hybridising conditions.

According to a third aspect of the present invention, an isolated peptide comprises the sequence identified herein as SEQ ID No. 3 or SEQ ID No. 4, or a fragment thereof comprising at least 10 consecutive amino acid residues.

According to a fourth aspect of the present invention, an antibody has an affinity of at least 10- ⁶ M for a peptide as defined above.

According to a fifth aspect of the invention, a polynucleotide that hybridises to or otherwise modifies the expression of an endogenous Rdx-OC-35 (SEQ ID No.1) gene, is used in the manufacture of a medicament for the treatment of cancer, in particular oesophageal cancer, or a precursor of oesophageal cancer.

According to a sixth aspect of the invention, a polynucleotide that hybridises to or otherwise modifies the expression of an endogenous Rdx-OC-40 (SEQ ID No.2) gene, is used in the manufacture of a medicament for the treatment of cancer, in particular oesophageal cancer, or a precursor of oesophageal cancer. Brief Description of the Figures

The invention is described with reference to the following figures, wherein: Figure 1 is a Clustal W alignment of RDX-OC-35 and SAT1 ; Figure 2 is a SPRR2 family alignment;

Figure 3 illustrates the nucleotide alignment of RDX-OC-40 and SPRR2E; and Figure 4 illustrates the alignment of amino acid ORF's from all SPRR2 family members, illustrating the common repeat characteristics of these proline rich proteins. Description of the invention The present invention is based on the identification of genes that are differentially expressed in a patient suffering cancer, in particular, oesophageal cancer and precursors of oesophageal cancer such as GERD and Barret's oesophagus, compared to a patient that does not have cancer or a precursor. Identification of the individual genes (or their gene products) in a sample obtained from a patient indicates the presence of or the risk of cancer in the patient.

The invention further relates to reagents such as polynucleotide and polypeptide sequences, useful for detecting, diagnosing, monitoring, prognosticating, preventing, imaging, treating or determining a pre-disposition to cancer.

Diagnosis can be made on the basis of the presence, absence or level of expression of the gene or gene product in the patient. As used herein, the term "gene product" refers to the mRNA or polypeptide product that results from transcription of the gene. The methods to carry out the diagnosis can involve the synthesis of cDNA from the mRNA in a test sample, amplifying as appropriate portions of the cDNA corresponding to the genes or fragments thereof and detecting each product as an indication of the presence of the disease in that tissue, or detecting translation products of the mRNAs comprising gene sequences as an indication of the presence of the disease.

Useful reagents include polypeptides or fragment(s) thereof which may be useful in diagnostic methods such as RT-PCR, PCR or hybridisation assays of mRNA extracted from biopsied tissue, blood or other test samples; or proteins which are the translation products of such mRNAs; or antibodies directed against these proteins. These assays also include methods for detecting the gene products (proteins) in light of possible post-translation modifications that can occur in the body, including interactions with molecules such as co-factors, inhibitors, activators and other proteins in the formation of sub-unit complexes.

The genes associated with cancer are characterised by the polynucleotides identified herein as SEQ ID No. 1 and SEQ ID No. 2. A gene comprising SEQ ID No.1 , or a homologue, similar molecule or fragment thereof, is referred to herein as an "Rdx-OC-35" gene. A gene comprising SEQ ID No.2, or a homologue, similar molecule or fragment thereof, is referred to herein as an "Rdx-OC-40" gene. Polypeptide products of Rdx-OC-35 and Rdx-OC-40 genes are identified herein as SEQ ID No. 3 (Rdx-OC-35) and SEQ ID No. 4 (Rdx-OC-40).

An increased level of an Rdx-OC-35 gene product is indicative of the presence of, or risk of, cancer. A decreased level of an Rdx-OC-40 gene product is indicative of the presence of, or risk of, cancer. The skilled person will understand that the terms "increased" and "decreased" refer to the amount of a gene product in a sample, compared to a "control" sample, or a known level of expression, that is indicative of a "healthy" patient that does not have, or is not predisposed to, cancer.

In an alternative embodiment, the amount of gene product in a sample is compared to a "control" sample or known level of expression is indicative of a "diseased" patient that is known to have, or be predisposed to, cancer.

Identification of the genes or their expressed products may be carried out by techniques known for the detection or characterisation of polynucleotides or polypeptides. For example, isolated genetic material from a patient can be probed using short oligonucleotides that hybridise specifically to the target gene. The oligonucleotide probes may be detectably labelled, for example with a fluorophore, so that upon hybridisation with the target gene, the probes can be detected. Alternatively, the gene, or parts thereof, may be amplified using the polymerase chain reaction, with the products being identified, again using labelled oligonucleotides.

Diagnostic assays incorporating any of these genes, proteins or antibodies will include, but not be limited to the following techniques that are known in the art:

Polymerase chain reaction (PCR)

Reverse transcription PCR

Real-time PCR

In-situ hybridisation

Southern dot blots Immuno-histochemistry

Ribonuclease protection assay cDNA array techniques

ELISA

Protein, antigen or antibody arrays on solid supports such as glass or ceramics. Small interfering RNA functional assays.

The present invention is also concerned with isolated polynucleotides that comprise the sequences identified as SEQ ID No. 1 and SEQ ID No. 2, or their complements, or fragments of each thereof that comprise at least 15 consecutive nucleotides, preferably 30 nucleotides, more preferably at least 50 nucleotides, for example 100 nucleotides or more. As used herein, the term "complement" refers to the exact complementary base - paired equivalent of a single stranded polynucleotide, with no mismatches.

Polynucleotides that hybridise to a polynucleotide as defined above, are also within the scope of the invention. Hybridisation will usually be carried out under

stringent conditions, known to those in the art and are chosen to reduce the possibility of non-complementary hybridisation and ensure that only polynucleotides that are similar or identical to at least a region of SEQ ID No.1 or SEQ ID No.2 (or their complements) will hybridise. Suitable conditions are disclosed in Nucleic Acid Hybridisation: A Practical Approach (B. D. Hames and S. J. Higgins, editors IRL Press, 1985). An example of stringent hybridisation conditions is overnight incubation at42°C in a solution comprising: 50% formamide, 5x SSC (15OmM NaCI, 15mM trisodium citrate), 5OmM sodium phosphate (pH7.6), 5x Denhardt's solution, 10% dextran sulphate, and 20mg/ml denatured, sheared salmon sperm DNA, followed by washing in O.ix SSC at about 65 ^° C.

Homologues of SEQ ID Nos. 1-4 are within the scope of the invention. The term "homologue" refers to a sequence that is similar but not identical to one of SEQ ID Nos. 1-4. A homologue may perform the same function as SEQ ID Nos. 1-4, i.e. the same biological function. Homology is routinely calculated as a percentage similarity or identity, terms that are well known in the art. Homologues of SEQ ID Nos. 1-4 preferably have 70% or greater similarity or identity at the nucleic acid or amino acid level, more preferably 80% or greater, more preferably 90% or greater, such as 95% or 99% identity or similarity at the nucleic acid or amino acid level. A number of programs are available to calculate homology; preferred programs are the BLASTn, BLASTp and BLASTx programs, run with default parameters, available at www.ncbi.nlm.nih.gov. For example, two nucleotide sequences may be compared using the BLASTn program with default parameters (score = 100, word size = 11 , expectation value= 10, low complexity filtering = on) and two polypeptide sequences may be compared using the BLASTp program with default parameters (expectation value = 10, word size = 3, matrix = BLOSUM62, Gap costs: existence = 11 & extension

=1).

For the avoidance of doubt, the methods and uses described herein can utilise any of the nucleotide or peptide molecules (i.e. SEQ ID Nos:1-4, homologues, similar molecules and fragments thereof) described herein. The identification of the Rdx-OC-35 gene and the Rdx-OC-40 gene also permits therapies to be developed, with each gene being a target for therapeutic molecules.

For example, there are now many known molecules that have been developed for gene therapy, to target and modify the expression of a specific gene. One particular molecule is a small interfering RNA (siRNA), which suppresses the expression of a specific target protein by stimulating the degradation of the target mRNA. Other

synthetic oligonucleotides are also known which can bind to a gene of interest (or its regulatory elements) to modify expression. Peptide nucleic acids (PNAs) in association with DNA (PNA-DNA chimeras) have also been shown to exhibit strong decoy activity, to alter the expression of the gene of interest. The skilled person will realise that Rdx- OC-35 is upregulated in cancer and, therefore, inhibition of Rdx-OC-35 will be of therapeutic benefit. Likewise, Rdx-OC-40 is downregulated in cancer and, therefore, stimulation of expression (i.e. upregulation) will be of therapeutic benefit.

The present invention also includes antibodies raised against a peptide of any of the genes identified in the invention. The antibodies will usually have an affinity for the peptide of at least 10- ⁶ M, more preferably, 10- ⁹ M and most preferably at least 10- ¹¹ M. The antibody may be of any suitable type, including monoclonal or polyclonal. Assay kits for determining the presence of the peptide antigen in a test sample are also included. In one embodiment, the assay kit comprises a container with an antibody, which specifically binds to the antigen, wherein the antigen comprises at least one epitope encoded by the Rdx-OC-35 gene or the Rdx-OC-40 gene. These kits can further comprise containers with useful tools for collecting test samples, such as blood, saliva, urine and stool. Such tools include lancets and absorbent paper or cloth for collecting and stabilising blood, swabs for collecting and stabilising saliva, cups for collecting and stabilising urine and stool samples. The antibody can be attached to a solid phase, such as glass or a ceramic surface.

Detection of antibodies that bind specifically to any of the antigens in a test sample suspected of containing these antibodies may also be carried out. This detection method comprises contacting the test sample with a polypeptide, which contains at least one epitope of the gene. Contacting is performed for a time and under conditions sufficient to allow antigen/antibody complexes to form. The method further entails detecting complexes, which contain any of the polypeptides. The polypeptide complex can be produced recombinantly or synthetically or be purified from natural sources.

In a separate embodiment of the invention, antibodies, or fragments thereof, against any of the antigens (peptides) can be used for the detection or image localisation of the antigen in a patient for the purpose of detecting or diagnosing the disease or condition. Such antibodies can be monoclonal or polyclonal, or made by molecular biology techniques and can be labelled with a variety of detectable agents, including, but not limited to radioisotopes.

In a further embodiment of the invention, antibodies or fragments thereof, whether monoclonal or polyclonal or made by molecular biology techniques, can be used as therapeutics for the disease characterised by the expression of any of the genes of the invention. The antibody may be used without derivatisation, or it may be derivatised with a cytotoxic agent such as radioisotope, enzyme, toxin, drug, pro-drug or the like.

The term "antibody " refers broadly to any immunologic binding agent such as IgG, IgM, IgA, IgD and IgE. "Antibody" is also used to refer to any antibody-like molecule that has an antigen-binding region and includes antibody fragments such as single domain antibodies (DABS), Fv, scFv, aptamers, etc. The techniques for preparing and using various antibody-based constructs and fragments are well known in the art. Means for preparing and characterising antibodies are also well known in the art.

If desired, the cancer screening methods of the present invention may be readily combined with other methods in order to provide an even more reliable indication of diagnosis or prognosis, thus providing a multi-marker test.

The following example illustrates the invention with reference to the accompanying drawings. Example A number of differentially expressed gene fragments were isolated from cDNA populations derived from matched clinical samples of oesophageal cancer patients, using non-isotopic differential display (DDRT-PCR). Two of these fragments, SEQ ID No. 1 (Rdx-OC-35) and SEQ ID No. 2 (Rdx-OC-40), showed a significant change in expression in oesophageal tumour tissue samples from a number of donors, in comparison to their co-excised normal tissue counterparts. The expression profile of these molecular markers, their genes and corresponding protein sequences are detailed herein. Materials and methods.

Differential gene expression was investigated between matched pairs of normal oesophagus and tumour tissue from the same donor. Tissue samples were obtained, with full ethical approval and informed patient consent. Following surgical removal, one sample of tumour tissue was collected, as was a sample from the adjacent, co-excised normal tissue. Messenger RNA was extracted and cDNA subsequently synthesised, using Dynal dT18-tagged Dynabeads and Superscript III reverse transcription protocols, respectively. Differential display reverse transcription PCR (DDRT-PCR)

was employed to observe differences between the gene expression profiles of these matched samples and individual gene transcripts showing up- or down-regulation were isolated and investigated further.

First described by Liang, P. and Pardee, A. B. (1992 Science. 257: 967-971) differential display reverse transcription PCR (DDRT-PCR) uses mRNA from two or more biological samples as templates for representative cDNA synthesis by reverse transcription, with one of 3 possible anchor primers. Each of the 3 sub-populations was PCR amplified using its respective anchor primer coupled with one of 80 arbitrary 13-mer primers. This number of primer combinations has been estimated to facilitate the representation of 96% of expressed genes in an mRNA population (Sturtevant, J. Clin Microbiol Rev. 2000 Jul;13(3):408-27). This population sub-division results in the reduction of the estimated 12,000-15,000 mRNAs expressed in eukaryotic cells to 100-150 transcripts on completion of second strand cDNA synthesis for each primer set. This facilitates the parallel electrophoretic separation and accurate visualization of matched primer sets on a polyacrylamide gel, leading to the identification of gene fragments expressed in one tissue sample but not the other.

Differentially expressed gene fragments identified in this way were excised from the polyacrylamide gel, re-amplified and direct sequenced to determine their novelty and chromosomal location. Database interrogation with these sequences revealed the extent of homology to known genes, whether any homoiogues were known to be involved in tumour development or oesophageal conditions and whether proteins had been described for these markers. Those fragments that remained novel, either by virtue of not being homologous to known genes, or through a potentially novel use, were used for primer design and creation of full-length transcripts. Fragments not matching known genes were regarded as potentially representing novel markers forthe oesophageal cancer from which they were derived.

Primary screening removed false positive results from the pool and involved real-time PCR assays of each fragment against the tissue cDNA sets from its source donor. Once confirmed as differentially expressed in these tissue samples, more extensive real-time PCR screening was conducted against matched cDNA templates from a panel of oesophageal cancer patients.

Full-length transcripts of the candidate gene fragments, including the open reading frame (ORF), were then derived using either 5' RACE (rapid amplification of cDNA ends), or by extrapolating the sequence to include an overlapping homologous known gene, as determined through database interrogation. Primers were designed

to amplify the putative ORFs of these candidate molecular markers and then confirmed as being correct by sequencing.

The tissue specific expression profile of each molecular marker was tested using gene specific primers against cDNA populations derived from a comprehensive panel of 20 human tissue types, as follows:

1. Bone Marrow

2. Brain cerebellum

3. Colon *

4. Foetal Brain 5. Foetal Liver

6. Heart

7. Kidney

8. Liver

9. Lung 10. Placenta

11. Prostate

12. Salivary Gland

13. ^ Skeletal Muscle

14. Small intestine * 15. Spinal Cord

16. Spleen

17. Stomach

18. Testis

19. Trachea 20. Uterus

The majority of these samples were part of the Human Total RNA panel Il (Clontech), but two RNA samples, marked with asterisks, were obtained separately from Clontech. In addition, assays were performed on a range of ethically approved human tumour samples. cDNA representative of tumours from ovary, testis, stomach, liver, lung, bladder, colon and pancreas were tested against both b-actin and the putative markers, by real-time and conventional PCR.

In conjunction with novel marker expression analysis, each matched pair of oesophageal tissues was subjected to molecular signature analysis, using primers specific to a number of pre-published oesophageal cancer molecular markers in

real-time PCR assays. The expression profile of each marker was determined, tabulated for each sample and used as a reference, against which the novel markers could be compared. This is with the aim of sub-classifying the tumour types and enabling the association of novel markers against such sub-types, increasing the power of the marker considerably.

Results and Discussion: Rdx-OC-35

Using differential display, a gene fragment, SEQ ID No. 1 (Rdx-OC-35), derived from cDNA populations of matched tissue from an oesophageal adenocarcinoma donor, was observed to have significant up-regulation in the tumour cDNA population in comparison to the corresponding normal tissue cDNA. This 278-nucleotide product (SEQ ID NO: 1) was confirmed as differentially expressed and database interrogation determined that Rdx-OC-35 was homologous to a region of human spermidine/spermine N1 -acetyl transferase (SAT1), on chromosome X, which has a complete sequence of 1060-nucleotides (SEQ ID NO:5; the open reading frame starts at nucleotide 166 and ends at nucleotide 681). Figure 1 shows the overlapping sequence, as determined by Clustal W analysis (EBI). From the SAT1 sequence, a presumed ORF of 171 amino acids was derived (SEQ ID NO:3).

SAT1 is a highly regulated enzyme that catalyzes the acetylation of polyamines as part of the polyamine degradation pathway and is involved in the regulation of polyamine transport out of ceils. The polyamines, e.g. putrescine, spermidine, and spermine, constitute a group of cell components that are important in the regulation of cell proliferation and cell differentiation. There is also evidence suggesting a role for polyamines in programmed cell death, thus their disrupted expression may have a major influence in the progression of cancer. The importance of the polyamines in cell function is reflected in a strict regulatory control of their intracellular levels. Suitable cellular polyamine levels are achieved by a careful balance between biosynthesis, degradation and uptake of amines. Some of the regulatory mechanisms involved in maintaining a balance in the cellular polyamine pools are unique. The polyamine biosynthetic pathway consists of two highly regulated enzymes, ornithine decarboxylase and S-adenosylmethionine decarboxylase, and two constitutively expressed enzymes, spermidine synthase and spermine synthase. The biological half-lives of the two regulatory enzymes ornithine decarboxylase and S-adenosylmethionine decarboxylase (5-60 min) are among the shortest known for mammalian enzymes, allowing the cell to rapidly change the cellular polyamine levels.

It could be speculated that increased expression of SAT1 in a tissue sample may lead to increased degradation of the polyamines, which may lead to an imbalance in this normally finely regulated system of cell proliferation and differentiation. Increased degradation of these regulating polyamines could potentially cause increased proliferation and reduced apoptosis, If the cells' other normal cell cycle control checks have been impaired, through mutations in p53, for example, this may go unchallenged, leading to the formation of a tumour.

A detailed real-time expression profile of this fragment was undertaken using cDNA populations derived from a number of matched normal and tumour oesophageal tissue samples donated by other patients. Of the donor samples screened, many exhibited notable increases in expression, suggesting that Rdx-OC-35 is a molecular marker for the presence of an oesophageal tumour. The table below shows the number of patient samples tested and the expression profile of this putative marker.

No. %

Increased in tumour tissue: 7 58

Increased in normal tissue: 2 16

No difference: 3 25

No expression detected: 0 0

Total matched samples used: 12 100

Rdx-OC-35 was further tested using cDNA populations derived from a panel of 20 human tissue types by real-time PCR analysis. In addition, assays were performed on a range of ethically approved human tumour samples, obtained through Medical Solutions pic, to ascertain whether the marker was oesophageal cancer specific or a less specific marker for the presence of cancer. cDNA, representative of tumours from ovary, testis, colon, stomach, liver, lung, bladder and pancreas were used. Rdx-OC-35 was evident in most of the cDNA samples, so this marker is not tissue specific.

Despite being detected in the majority of samples tested, Rdx-OC-35 is useful as an indicator for the presence of a tumour, through its increased expression in a number of tumour samples, in comparison to co-excised normal tissue from the same donor. This candidate marker has particular utility for RNA expression profiling when using biopsies or other localised material, as direct comparisons are being made between the normal and tumour samples from the same tissue type. In relation to its use as a serum marker, normal levels will need to be established, so that the detection

of an increase in these levels will be indicative of the potential for an oesophageal tumour.

Results and Discussion: Rdx-OC-40

Using differential display, a gene fragment, SEQ ID No.2 (Rdx-OC-40), derived from cDNA populations of matched tissue from an oesophageal adenocarcinoma donor, was observed to have a significant reduction in expression in the tumour cDNA population in comparison to the corresponding normal tissue cDNA. This 333-nucleotide product (SEQ ID No:2) was confirmed as differentially expressed and database interrogation determined that Rdx-OC-40 was homologous to a number of members of the small proline rich protein gene family on chromosome 1 , namely SPRR2A, SPRR2B, SPRR2D, SPRR2E and SPRR2F. Figure 2 shows the homologous sequence of each of these family members, as determined by Clustal W analysis (EBI). A further Clustal analysis revealed that Rdx-OC-40 was most homologous to SPRR2E. These similarities (95% over the Rdx-OC-40 sequence) are highlighted in the cluster shown in Figure 3. The presumed peptide sequence of SPRR2E is included as SEQ ID No:4. In this 72 amino acid residue peptide, the common repeat region of all SPRR2 proteins stretches from residues 21 (proline) to 47 (proline).

Of the homologous SPRR2 gene family members, only 2E has the exact match of the forward and reverse expression primers used to assess the candidate markers potential. Two other SPRR2 family members, however, namely 2B and 2D, had only a single mismatch on the reverse primer, so these genes may also have been represented in any expression profiling performed.

Primary screening confirmed this fragment to be differentially expressed in matched tissue from the source donor, so a more detailed expression profile was undertaken using cDNA populations derived from a number of matched normal and tumour oesophageal tissue samples donated by other patients. Of the donor samples screened, many exhibited notable decreases in tumour expression, suggesting that Rdx-OC-40 is a putative molecular marker for the presence of an oesophageal tumour. The table below shows the number of patient samples tested and the expression profile of this putative marker.

No. %

Increased in tumour tissue: 2 16.6

Increased in normal tissue: 8 66.6 No difference: 2 16.6 No expression detected: O O Total matched samples used: 12 100

Rdx-OC-40 was further tested using cDNA populations derived from a panel of 20 human tissue types and 8 non-matched tumour samples, by real-time PCR analysis. From this analysis, the candidate marker was weakly expressed in a number of the normal tissue samples screened, namely, bone marrow, foetal brain, lung, placenta, stomach, trachea and uterus, so is not very tissue specific. In addition, expression was also noted in tumour samples from bladder, liver, ovary, pancreas, stomach and testis, with bladder and ovarian tumours having the highest levels of this marker.

All of the SPRR genes are located in a region of chromosome1q21 , which spans 2MB and also contains S100 and other structurally, functionarily and evolutionarily co-related genes. Collectively, this region is known as the epidermal differentiation complex (Marenholz et a/., 1996. Genomics. 37: 295-302), with at least 43 genes that are expressed during keratinocyte differentiation (Elder, J. T. and Zhao, X. 2002. Dermatology. 11 : 406-412). There are 3 main gene families, encoding proteins relating to the cell envelope (CE), intermediate filament-associated proteins and calcium binding proteins (Marenholz, I et a/. 2001. Genome Research. 11 : 341-355). The small proline rich proteins (SPRR's), with loricrin and involucrin, are the major precursors of the CE, a highly insoluble and rigid structure that is essential for the barrier function of the skin.

The assembly of the CE starts with the formation of a scaffold made up of involucrin and envoplakin. Reinforcing proteins, such as cystatin and the SPRR's are added to form the complete CE structure, which serves as a platform for lipid attachment (Cabral, A. etal, 2001. Journal of Biological Chemistry. 26: 19231-19237). The characteristics relating to rigidity or flexibility and strength appear to be dictated by the SPRR proteins. The genetic characterisation of the multigene family determines 2 distinct groups. Group 1 has 3 classes (SPRR1 , SPRR3 and SPRR4), with different protein structures, whereas group 2 (SPRR2) has 7 genes with a high degree of homology, being distinguished more by their regulation and expression than their structure. It is this regulation of the highly homologous gene family that underscores their importance in CE formation and flexibility to respond to external stimuli (Cabral et

al., 2001 supra). The SPRR2 genes all contain 3 clusters of homologous residue repeats, each consisting of the consensus PKCPEPCPP, with one or 2 residues of variance (Figure 4). In many cases, the difference between the family members is only 1 or 2 amino acids over the open reading frame, but sequence divergence is more pronounced in the promoter regions. Thus the expression differences of these SPRR's are key to the differentiation of the CE, dependant on external challenges.

Cytokeratin expression is a reliable biochemical indicator of the epithelial differentiation process. Keratins 5 and 14 represent the basal layer, whereas 1 and 10 are expressed as the keratinocyte cells become larger and differentiate. In squamous cell carcinoma, CK4, CK13 and CK15 have been shown to be down-regulated, whereas CK16 and CK17 were up-regulated (Luo, A., 2004. Oncogene. 23: 1291-1299). Reduced expression is also a characteristic of many of the calcium binding proteins, such as the S100's and the SPRR's, which are abundant in the oesophagus. This was the case with Rdx-OC-40, which displayed reduced tumour expression in a number of the matched tissue samples tested, providing extra evidence that this differential display product belongs to the small proline rich protein family. In particular, the highest homology over the full sequence of this candidate marker is with SPRR2E, which has not previously been associated with oesophageal cancer, or indeed, been dealt with in any detail. Epidemiological data have shown that in areas of China that show increased risk of ESCC, the calcium levels of wheat and maize are lower than in the same cereals of lower risk areas (Chang, S. et al, 1999. Wei Sheng Yan Jiu. 28 (6): 364-366.) and may result in deficiencies of calcium in the diets of the population in such regions. The results of Luo et al. (supra), suggest that calcium, or lack of it, may play a significant role in the progression of ESCC, and perhaps of oesophageal cancer in general, thus the expression of calcium-dependant proteins, particularly SPRR2E, may act as excellent markers for the diagnosis, prognosis and monitoring of this disease.