Recent Advances in Understanding ... - Semantic Scholar

CANCER GENOMICS & PROTEOMICS 1: 283-294 (2004)

Review

Recent Advances in Understanding Carcinogenicity of Oral Squamous Cell Carcinoma: From Basic Molecular Biology to Latest Genomic and Proteomic Findings RAVI MEHROTRA1, ENDRE NORMANN VASSTRAND2 and SALAH OSMAN IBRAHIM3 1Department

of Pathology, Moti Lal Nehru Medical College, University of Allahabad, 16/2, Lowther Road, Allahabad-211002, India; 2Dental Faculty-Periodontology, University of Bergen, Årstadveien 17, 5009 Bergen; 3Department of Biomedicine, Section of Biochemistry and Molecular Biology, University of Bergen, Jonas Lies Vei 91, 5009 Bergen, Norway

Abstract. The global increase in incidence and mortality, as well as the poor prognosis of oral squamous cell carcinoma (OSCC), has intensified efforts in the field of prevention and early detection of this disfiguring disease. Prevalence of OSCC is common in areas with high consumption of tobacco products and alcohol. Understanding the carcinogenicity of this cancer, using innovative techniques in genomic and proteomic analysis, is the main focus of current research in OSCC, and the hunt for potential molecular biomarkers is accelerating. Although recent advances in preventive, diagnostic and therapeutic techniques related to OSCC have yielded novel molecular targets, partially uncovered signal pathway dominance and advanced early neoplasia detection, the number of deaths attributed to this disease exceeds that reported for cervical cancer, malignant melanoma and

Abbreviations: OSCC, oral squamous cell carcinoma; TSGs, tumour suppressor genes; CGAP, cancer genome anatomy project; HGP, human genome project; ESTs, expressed sequence tags; LCM, laser capture microdissection; CGH, comparative genomic hybridization; 2-D PAGE, 2-dimensional polyacrylamide gel electrophoresis; MS, mass spectrometry; MALDI-TOF, matrixassisted laser desorption and ionization time-of-flight; SELDITOF, surface-enhanced laser desorption and ionization time-offlight. Correspondence to: Salah Osman Ibrahim, Department of Biomedicine, Section of Biochemistry and Molecular Biology, University of Bergen, Jonas Lies Vei 91, 5009 Bergen, Norway. Tel: +47-55-586423, Fax: +47-55-586360, e-mail: [email protected] Key Words: Oral cancer, epidemiology, carcinogenicity, proteomics, genomics, review.

1109-6535/2004 $2.00+.40

Hodgkin’s disease. Application of advanced molecular biology techniques for classification, profiling of tumour tissues and/or identification of potential markers of OSCCs is on the rise. This review aims at outlining the available knowledge on epidemiology, aetiology, molecular biology, and genomics and proteomics in relation to OSCCs. Epidemiology of oral squamous cell carcinoma (OSCC). OSCC, a malignancy of the lip, mouth or tongue, is predominantly of squamous cell nature with multiple heterogeneous genetic and epigenetic changes that might be involved in the development of this disease (1). The majority of these cancers occur primarily in three anatomical locations: floor of the mouth, soft palate and the ventrolateral aspect of the mobile portion of the tongue (1). The tumour is histologically divided into well-, moderately- and poorly differentiated patterns, with six histological variants recognised in the current WHO classification (1). The disease ranks 11th in frequency and 13th in cancerspecific mortality, being a major cause of cancer morbidity and mortality in the world (2, 3). The overall incidence attributed to OSCC is increasing worldwide, with current estimates of age-standardized rates of 6.6/100,000 and 3.1/100,000 in men and 2.9/100,000 and 1.4/100,000 in women, respectively (2, 3). Over 500,000 new cases of OSCCs are reported annually with three-quarters of these occurring in patients from developing countries, defined as Africa, Central and South America, Caribbean, China, Asia (Outer-Eastern, South-Eastern, SouthCentral and Western), Melanesia and Micronesia/Polynesia (2-5). This geographical variation probably reflects the prevalence of specific environmental influences, rather than genetically-determined or ethnically-related risk factors (3-6).

283

CANCER GENOMICS & PROTEOMICS 1: 283-294 (2004)

Figure 1. Suggested role(s) of tobacco, alcohol, viral infections and/or other DNA -damaging agent(s) in the development of oral squamous cell carcinoma by disturbing cell cycle regulatory genes and/or other tumour suppressor genes.

An estimated 30,000 new cases of this cancer are expected to be diagnosed in the United States in 2003, and approximately 7,200 people will die of the disease (7). In Eastern and Central European countries, a 2 to 3-fold mortality increase has been recorded in the last 3 decades (8). Several factors point to a progressive increase in the incidence and mortality of OSCCs worldwide, with additional factors that might contribute to this increase in the developing countries. Among these are the continuous

284

deterioration in oral health care, lack of diagnostic techniques and methods of early detection, lack of proper coding systems, ambiguity of terms sometimes used as causes of death, inaccuracy of demographic information, lack of autopsy data, urbanization and the changes in lifestyle recently witnessed in many of these countries. In addition, there is very little information on the prevalence of OSCC in many of the developing countries, with other factors leading to lower reporting of OSCCs as compared to

Mehrotra et al: Carcinogenicity of Oral Cancer

the developed world. These include cultural influences, lack and delays associated with the provision of oral health services, low rate of publications in peer-reviewed journals, under reporting of OSCC patients in health centres, deficiency of patient records or OSCC registries, short lifespan, and the fact that poor people (the mostly affected ones) do not have access to general oral health services which are mainly designed for well-off individuals. Aetiology of OSCCs. In the West, epidemiological studies suggest that the social habits of tobacco use and/or alcohol consumption are associated with the increased risk of OSCC (9). In South-East Asia, increased risk for this type of cancer is associated with chewing of tobacco (5, 6, 9). In addition to OSCCs, tobacco use, either smoked or nonsmoked, is a well–known risk factor for multiple neoplasia, including those of the oesophagus, lung and bladder (10). Co-factors in several of these malignancies, in particular OSCCs, include alcohol, dietary factors, immunodeficiency and viral infections (11). Despite the possible role of tobacco/or alcohol exposure in the development of OSCCs, the majority of people who smoke, chew tobacco and/or drink alcohol do not develop the disease, and not all OSCC patients are users of any of these substances (9). These observations point to possible genetic factor(s) that might be involved in the development of OSCCs (9). In head and neck SCCs, epidemiological evidence from case–control studies indicated that a family history of this type of cancer is considered to be a risk factor for the development of the disease with individuals having a positive family history having 3.5–3.71 times the risk of developing the cancer, albeit when tobacco smoking and alcohol consumption are allowed for (12, 13).

Genetic changes involved in OSCCs Role of tumour suppressor genes (TSGs)/oncogenes. Although the mechanism(s) underlying the development of OSCC remains unclear (14-16), carcinogenesis of this disease has been suggested to be a multi-step process involving interaction(s) between several factors like tobaccoassociated intra-oral carcinogens, alcohol consumption and/or viral infections (reviewed in 14-16). In carcinogenesis, tumour cells proliferate excessively without coordination with the microenvironment and the cells continue to divide as a result of loss of either cell cycle check-points, other cell cycle/growth control regulators, apoptosis, oncogene/TSGs, DNA damage response and/or other cellular regulators (Figure 1). In the literature, a series of reviews have critically summarised findings related to genes/or their protein products that might be involved in the development of OSCCs (reviewed in 17-25). TSGs, representing loss-in-function and defined as gene products

whose loss contribute to the development of cancer, are classified as "gatekeepers" and "caretakers" (26, 27). Gatekeepers, like p53, pRB, ABC and NK1, function to directly suppress cell proliferations and, upon reintroduction of the gene into tumour cells, suppression of tumourigenicity might be achieved (26, 27). Caretakers, however, function to maintain the integrity of the genome through excision repair, mismatch repair and homologous recombinations (26, 27). As a result of the functions of the caretakers, mutations in other genes, including oncogenes or gatekeeper TSGs, might occur (26, 27). TSGs are known to play a crucial role in cell cycle regulation, arrest and apoptosis, and failure of the role(s) played by any/or all of these genes will force the protective system to collapse and enable for a cancer to develop (Figure 1). By now, there are about 2200 genes with known or unknown functions, that have been detected through the Cancer Genome Anatomy Project (CGAP), suggested to be involved in the carcinogenesis of OSCCs (28). Among these genes are known TSGs (like p53, p16ink4a, p15ink4b, p18ink4c, p19ink4d, p14ARF, p21waf1, p27Kip1, pRB) and oncogenes (like hst-1, int-2, EGFR/erbB, c-erbB-2/Her-2, sis, Cyclin D1, ras, raf, stat-3, myc, fos, jun, c-myb, bcl-2, Bax), which were found to be involved in the carcinogenesis of OSCCs, though with conflicting results (reviewed in 14-25). One of the most studied TSGs in cancers, including OSCCs, is the p53 gene, which has been called the "Guardian of the Genome" or "The Molecular Policeman", that has a central role to play in genome integrity and stability, cell cycle progression, cellular differentiation, DNA repair and apoptosis (27, 28). The gene is known as a transcription factor and it exerts its TSG function, in part, by activating target genes after upstream stress signals, such as DNA damage, have been relayed (Figure 1). The p53 gene can be inactivated by several mechanisms including point mutations, deletions and/or binding with other cellular and/or viral protein like the E6 protein of the human papilloma virus types 16/18 (27-29). Estimates suggest that up to half of all human cancers carry some sort of inactivation of the p53 gene via point mutations, overexpressions, deletions and/or binding to viral proteins (27, 29, 30, 31). In oral/head and neck SCCs, the reported incidence varied from 43-96% (reviewed in 32-34) – probably as a result of substantial differences in the detection techniques used (32-34). Mutations of this gene were also found in a minority (range of 36-44%, also subject to the method used) of oral dysplastic lesions, showing that the gene can be mutated before development of the invasive neoplasia (32-36). However, results on p53 gene expression in oral/head and neck SCCs were conflicting regarding the relationship between the expression/or mutations of the gene, clinical as well as the patient’s histopathological parameters (32-34). The role played by mutations of this gene in the progression and vital

285

CANCER GENOMICS & PROTEOMICS 1: 283-294 (2004) prognosis of these cancers has not yet been demonstrated (reviewed in 32-34, 37). For a better understanding of the role played by p53 gene in oral/head and neck SCCs, inactivation of other TSGs like p16ink4a, p15ink4b, p18ink4c, p19ink4d, p14ARF, p21waf1, p27Kip1 and pRB, which are directly or indirectly linked to the transcriptional activity of the gene, have also been examined in these cancers (17-25). Oncogenes, defined as genes with a "gain-in-function", contribute to converting a normal cell into a cancer cell when a normally present proto-oncogene becomes inappropriately activated through either mutation or abnormal expression at high levels (26, 27). These can be represented by growth factors or growth factors receptors (hst-1, int-2, EGFR/erbB, c-erbB-2/Her-2, sis, Cyclin D1), intracellular signal transducers (ras, raf, stat-3), transcription factors (myc, fos, jun, c-myb) and apoptosis (bcl-2, Bax) (26, 27). Activation of oncogenes might affect the gene product itself, leading to direct activation or alteration of its regulatory mechanism, which then leads to increased expression and/or failure to switch off transcription at appropriate times. So far, several studies have examined the expression/or mutations in over a hundred oncogenes in dysplastic epithelium and oral/head and neck SCCs (reviewed in 14-25). In many studies, several known oncogenes were consistently detected in these cancers, some being found to be more strongly associated than others and, interestingly, few being found to be limited to specific cancers associated with tobacco habits (reviewed in 14-25). It has been argued that oncogenes, either directly or indirectly, might be connected to development and/growth disturbances in oral/head and neck SCCs, possibly by acting as cellular ‘switches' at key biochemical points to affect a whole series of phenotypic changes, thus leading to tumour transformation in these cancers (reviewed in 14-25). Therefore, understanding the functional variety of these oncogenes might provide valuable insights into the complex network of events that are involved in the tumourgenesis of OSCCs, and it is hoped that this might lead to both diagnostic and therapeutic benefits for patients (reviewed in 14-25). Most of the studies that have described TSGs/or oncogenes in OSCCs, using either cell culture experiments/or archival formalin-fixed, paraffin-embedded tissues or fresh frozen tissue specimens, have examined either amplifications, deletions/or mutations of a gene at the DNA level, over- or under-expression of the gene product at the RNA/or protein level using traditional methods in molecular biology like PCR, PCR-SSCP, PCR/direct DNA sequencing, RT-PCR, CDGE, blotting, DNA cloning, in situ hybridizations, ELISA, EIA, RIA or immunohistochemistry which generally work on a "one gene in one experiment" basis, indicating that the throughput is substantially limited and the "whole picture" of the gene function(s) is hard to obtain. The fact that multiple TSGs/or oncogenes have been reported in oral dysplastic and malignant epithelium, and the

286

evidence that this type of cancer is a multi-step process involving a number of aberrant molecular events affecting a range of genes, pointed to a great for large scale analysis of genes that might be involved in this type of cancer. Nevertheless, it is now possible to identify over/or underexpressed molecular targets in OSCCs using the new approaches of genomic and proteomic research. Genomic instability. Cytogenetic analysis of OSCCs to detect chromosomal alterations either at the short (p) or long (q) arm of a chromosome following evaluation of the human genome, has become quite efficient using techniques such as restriction landmark genomic scanning (RLGS), fluorescence in situ hybridization (FISH), cross-species colour banding (RxFISH), loss of heterozygosity (LOH), restriction fragment length polymorphism (RFLP), comparative genomic hybridization (CGH), microsatellite instability (MI), chromosome painting and trypsin-giemsa banding (G-banding), which are frequently used to detect patterns of chromosomal imbalances and gross chromosomal regions involved in structural rearrangements (39-41). The most valued findings, thus far, in oral/head and neck SCCs are gains of 3q, 8q, 9q, 20q, 7p, 11q13 and 5p and losses of 3p, 9p, 21q, 5q, 13q, 18q and 8p (reviewed in 22). It has, therefore, been suggested that cytogenetic analysis may be useful for analysing clinical differences in tumour behaviour and response to therapy. Numerous studies are now underway to examine the biology of, and genetic changes in, OSCCs that might lead to additional markers for use as rapid, non-invasive screening methods for individuals at high risk for primary or recurrent cancers (reviewed in 22). Immortalization by telomerase activity. The structures at the ends of eukaryotic chromosomes, "termed telomeres", are random arrays of hexamers TTAGGG, bound to proteins and are located on the ends of each chromosome (42). Since telomeres are lost during cell divisions, the chromosomal ends are no longer protected, which leads to the fusion of chromosomes and karyotypic abnormalities that eventually cause cell death (42). The ribonucleoprotein enzyme telomerase extends the telomeric repeat sequences at the chromosomal ends; it is active in about 90% of human neoplasia, but inactive in most of the normal cells (42). In a study by Mao et al. (43), head and neck SCC cell lines, tumour specimens and adjacent normal and dysplastic mucosa were analysed for telomerase activity. No telomerase activity was found in any of the normal tissues, but it was found in 100% of the cell lines, 90% of the invasive neoplasia and 100% of the dysplastic lesions (43). However, recent developments suggest that telomere maintenance might not be an obligatory requirement for initial tumour formation in some settings and that telomerase activation contributes to tumourigenesis


Figure 2. Schematic illustration of the cDNA and protein microarrays, 2-D PAGE and MALDI/SALDI-TOF-MS techniques in oral squamous cell carcinomas and their pair-wised normal controls.

independently of its role in maintaining telomere length (42). However, the understanding of telomere biology remains incomplete and implicates additional complexity in the relationships among telomere, telomerase and oral/head and neck SCCs, as suggested by others (43-46). Angiogenesis. As normal cells progress toward malignancy, they must switch to angiogenic phenotype to attract the nourishing vasculature that they depend on for their growth, in a process known as angiogenesis. Angiogenesis, defined as the growth of new blood vessels (neovascularization) from pre-existing ones, is a multi-step process, which appears to be regulated by both stimulatory and inhibitory factors and is vital for the continued growth and survival of solid neoplasms (47, 48). The vast majority of highlyvascularized tumours showed a poor prognosis and the

influence of tumour angiogenesis proved to be independent of conventional prognostic indicators. The angiogenesisindicating signals are exemplified by vascular endothelial growth factor (VEGF), some matrix metalloproteinases (MMPs), interleukin-8 (IL-8) and acidic as well as basic fibroblast growth factors (FGF 1/2) (49). Expression of VEGF family members, i.e. VEGF-A and VEGF-C has been reported in head and neck SCCs and has been found to be associated with metastasis (50). Although the quantification of microvessels in tissue sections is mainly used to evaluate angiogenesis, it appears to be subject to biases, as reflected in contradicting research findings where, in particular, the choice of the staining methods seems crucial. However, since studying tumour angiogenesis offers a uniquely attractive therapeutic target, the next decade may produce better methods that can be used to produce a

287

CANCER GENOMICS & PROTEOMICS 1: 283-294 (2004) catalogue of the angiogenic regulatory molecules and factors that may be used in staging of disease for different types of malignancies, whereas the data available for oral/head and neck SCCs are limited.

Dissecting the genome by advanced new technologies Completion of the Human Genome Project (HGP) more than three years ago led to a surge in use of genomic and proteomic technologies for identification of markers for early detection and molecular-targeted therapy in neoplasia (51, 52). The number of defined human genes and expressed sequence tags (ESTs) continues to grow rapidly and new tools are being developed for interrogation of these databases (51, 52). High-throughput innovative screening techniques in genomics and proteomics are now widely available, allowing researchers to rapidly screen and confirm new genes, mRNA transcripts and/or proteins. In particular, differential expression of these molecules between malignant and pair-wised normal tissues allows researchers to identify genes and critical pathways that are involved in cancer development. Laser capture microdissection (LCM). The tumour microenvironment of a carcinoma consists not only of the malignant epithelial component, but also the surrounding stroma and normal tissue. These distinct microcompartments use receptors, cell junctions, and inter- and intracellular signalling molecules to allow tumour cells to communicate with their surroundings, playing an active role in their control/or progression (53). Removing a subpopulation of these cells for growth in an in vitro system interrupts potentially important cell–cell and cell–matrix interactions that might affect the neoplastic behaviour, therefore giving scientists a false impression of the in vivo tumour composition and physiology. The advent of LCM has enhanced our ability to specifically dissect and remove target subpopulations of cells from frozen/or ethanol-fixed tissues, as well as cell culture populations under direct microscopy, which can further be stained or unstained and used for any kind of study that involves changes in DNA, RNA/or proteins (54). Highthroughput analysis of microdissected specimens allows for clean discrimination of events occurring in- and between each of these tissue micro-compartments and, accordingly, researchers are now able to profile low molecular weight proteins in patient serum samples and query them with new, powerful bioinformatics tools to cluster unaffected and cancer patients into their respective groups. To take a more realistic "snapshot" of a tumour’s in vivo inherent properties using LCM, Alevizos et al. (55) carried out large-scale gene expression profiling on tumour and normal oral epithelial cells and identified about 600 genes to be associated with OSCCs. These OSCC-associated genes included oncogenes,

288

TSGs, transcription factors, xenobiotic enzymes, metastatic proteins, differentiation markers and genes that have not, so far, been implicated in OSCCs (55). Genomic array technology. Microarray, or DNA chip, is a new powerful tool for studying gene expression profiles of biological samples to determine the molecular basis of interactions on a large scale that is impossible using conventional analysis (56). Since the start of its application on biological samples in 1995, the fundamental concepts behind the array, the technology needed for making and using these chips, and the multitude of statistical tools needed for analysing the data have been extensively reviewed and critically evaluated (56). This technique makes it possible to address fundamental research questions by examining the expression of thousands of genes simultaneously, with a promise to help in the development of rational approaches to therapy, as well as to facilitate disease diagnosis and prognosis, assuring its entry into clinical practice in specialist centres and hospitals within the next few years (56). The technology is now a well-established method for comparing the expression of genes and ESTs in laboratory and clinical samples, where thousands of genes can be examined simultaneously through amplification of RNA with fluorescent labels and applying the labelled transcripts to array slides containing large numbers of oligonucleotides or cDNAs (Figure 2). Advances in array technology provide a high-throughput approach to monitor the c-DNA arrays for mRNA or total RNA expression, genomic single nucleotide polymorphism (SNPs) and CGH (56). This technology is being used to identify patterns of gene expression, SNPs and CGH that may be indicative of certain diseases, including OSCCs, where a detailed comparison of gene expression profile with respect to normal tissue is possible (Figure 2). After completion of the HGP, there has been a dramatic increase in the amount of genomic information now available in databases, with the prediction that more updated information will be available in databases, thus giving rise to changes in disease research (51, 52). With over 30,000 known human genes, the application of innovative techniques is important for understanding the function and structure of genes involved in a disease process (51, 52). The complete sequence of the human genome will also provide a new starting point for understanding our basic genetic makeup and how variations in our genetic instructions result in disease, in particular, OSCCs. To date, comparison of gene expression profiles between oral/head and neck SCCs and normal tissues, that showed altered expression levels of genes involved in the control of cell growth and differentiation, angiogenesis, apoptosis, cell cycle and signalling, has been shown in a few, wellcontrolled, published microarray studies from the West with significant findings (57-63). Schmalbach et al. (59) reported


the expression profile of 101 genes between metastatic and non-metastatic oral/oropharyngeal SCCs, with a subset of 57 genes that showed significant expression differences between metastatic tumours and normal mucosa. The profile found included genes related to the extracellular matrix, adhesion, motility, inflammation and protease inhibition suggesting that the knowledge gained might facilitate early detection of aggressive tumours and targeted therapeutic investigations (59). Belbin et al. (60) identified 375 genes that showed significant expression differences, and divided the patients with head and neck SCCs into clinically distinct subgroups based on the gene expression patterns, suggesting that gene expression profile can be used as a predictor of outcome (60). In another work by Ginos et al. (61) the gene expression profile was studied in 41 cases of head and neck SCCs, and 2890 genes were found to be associated with proliferation, extracellular matrix production, cytokine/chemokine expression, immune response, invasion and metastasis, which showed significant differences in their expression, suggesting evidence for a new gene expression-based biomarker of local treatment failure in these cancers (61). On the other hand, Leethanakul et al. (62) successfully applied LCM to procure cells from representative sets of head and neck SCC and their matching normal tissues to be used in array experiments. A consistent decrease in expression of differentiation markers like cytokeratins, and an increase in the expression of a number of signal transducing and cell cycle regulatory molecules, as well as growth and angiogenic factors and tissue degrading proteases, were found between cancer and normal cells (62). It has been suggested that the LCM approach might facilitate the identification of candidate markers for early detection of premalignant lesions, as well as novel targets that can be applied for pharmacological intervention in this type of cancer (62). In a recent report, an oral cancer-specific human bacterial artificial chromosome (BAC) array, called the oral cancer genomic regional array (OCGR), has been described to detect copy number alterations in OSCC (63). This was achieved by array CGH, which complements traditional methods for detecting genetic alterations in OSCC, such as microsatellite and CGH analysis, by improving the detection of segmental copy number alterations to single BAC clone resolution (63). The study detected both copy number decrease and increase in 14 LCM clinical specimens of OSCC in nine chromosomal regions known to be involved in the progression of OSCC. Although the number of regions selected in this array was relatively small, single BAC clone alterations were observed at 7p11 and 11q13 regions, which are representative of EGFR and Cyclin D1 respectively, highlighting the need for high-resolution detection techniques. Overexpression of Cyclin D1 is welldocumented in OSCC and has been reported to be

associated with disease progression and poor prognosis. Half of the cases analyzed in the previous study showed copy number increase in this region (63). This work represents the first oral cancer-specific CGH array, and the data presented support the use of high-resolution coverage of specific chromosomal regions to facilitate fine mapping of variations and, ultimately, aid in the discovery of novel oncogenes and TSGs. No doubt, supplementing the OCGR array with new clones, this array can be readily expanded to produce a tool for use in early stage of OSCC diagnosis. Although the burden of OSCC is high in developing countries, studies using cDNA arrays and other advances in the field of molecular biology techniques are rare, which might be attributed to the high cost of these techniques that cannot be met by research laboratories in these countries. In a recent report, Ibrahim et al. (64) studied the differential expression of 588 genes by cDNA macroarray analysis in Caucasians and Africans, and found that alterations of 123 genes are common in OSCCs, regardless of ethnic differences, other socio-cultural risk factors and clinicopathological parameters. Among the genes found were several interesting keratin genes, p19ink4d, Bax, CASPs, MMPs, TOP1, p73, DNA-repair protein XRCCI and DNA excision repair protein ERCC5 (64). In summary, these studies have revealed the implication of different signalling and metabolic pathways, highlighting the potential role of some pathways in the development of oral/head and neck SCCs (57-64). The results have also provided new insights into the carcinogenesis of this disease as well as a source of potential prognostic and predictive markers and targets for prevention and therapeutics (57-64). However, finding a gene expression pattern on a microarray analysis is merely a start, and validation will follow to assist the translation of the statistical values into clinical decisions. Although the sample sizes of the tumours that were used in the previous studies were small (57-64), the results might provide important clues to the understanding of the various gene networks implicated in the carcinogenesis of this disease, and may contribute to the selection of target genes for possible molecular diagnosis and therapy in the future. Proteomics technology. The term "proteomics" indicates proteins expressed by a genome, and is the systematic analysis of protein profiles of tissues paralleling that of genomics (65) (Figure 2). The evaluation of protein structure, function and regulation has evolved rapidly over recent decades, and large amounts of information about a particular protein’s activation status and interactions can now be obtained in a matter of minutes, employing the new high-throughput approaches (65) (Figure 2). The mainstay of proteomics–based expression profiling is still the combination of high-resolution 2-D PAGE gel electrophoresis and mass spectrometry (MS) (66). Recent

289

CANCER GENOMICS & PROTEOMICS 1: 283-294 (2004) breakthroughs in protein microarrays have produced commercially available protein pathway arrays, that are able to test cell lysates for both their phosphorylated and nonphosphorylated forms using specialized protein chips. Protein microarrays also depend on specific antibody ligand interactions, but avoid some of the complications of immunohistochemistry by allowing rapid, quantitative comparison of the proteins of interest using fluorescencebased imaging programs (65, 66). Matrix-assisted laser desorption and ionization time-of-flight (MALDI–TOF) and surface-enhanced laser desorption and ionization (SELDI)-TOF are two of the methods currently being employed for this purpose (65, 66). The general principle of the MALDI-TOF revolves around the rapid photovolatization of a sample embedded in UV-absorbing matrix, followed by TOF-MS analysis (67). MALDI-TOF-MS has become a popular and versatile method to analyse a wide range of macromolecules of biological origin, from cells to tissues (67). Due to its ability to desorb high-molecular weight thermolabile molecules, its high accuracy and sensitivity, combined with its wide mass range (1-300 kDa), MALDI-TOF-MS is becoming a promising method in the clinical chemistry laboratory for identification of biomolecules in complex samples, including peptides, proteins, oligosaccharides and oligonucleotides (67). The other new protein analysis system, SELDI, has recently been applied for the separation, detection and analysis of multiple proteins in very small amount (~10ng) of microdissected cancer tissue (67). This system facilitates protein capture, purification, analysis and processing from complex biological mixtures directly on to protein chip array surfaces, and the detection of the purified proteins is performed by TOF-MS. A major feature in SELDI-TOFMS is its ability to provide a rapid protein expression profile from a variety of biological and clinical samples. The use of these methods for biomarker identification, as well as the study of protein–protein and protein–DNA interaction, makes it suitable for studies involving OSCCs, since it has been used in projects involving identification of potential diagnostic markers for prostate, bladder, breast and ovarian cancers (66). There are very few studies, all from Western countries, which have applied these techniques in OSCCs to search for related tumour proteins that might have potential clinical applications as biomarkers, enabling tumour identification at an early stage in high-risk individuals, prediction of treatment response and detection of residual or recurrent carcinoma (68-71). He et al. (68) examined the protein expression profile in OSCCs/their corresponding normal resection margins by MALDI-TOF-MS, and found a number of tumour-associated proteins including heat-shock protein (HSP) 60, HSP27, alpha B-crystalline, ATP synthase beta, calgranulin B, myosin, tropomyosin and galectin 1 to be significantly altered in their expression levels in cancers,

290

compared with their paired normal tissues (68). However, due to the complex and heterogenous components of the tissue samples, proteins extracted from tissues are often mixed with many other kinds of molecules such as lipid and carbohydrate, etc. which are often from undesired cells and/or components and can hardly be removed completely during the protein sample preparation, which results in poor separation and low resolution of protein spots on 2-D PAGE. To obtain a homogeneous cell population, Knezevic et al. (69) analysed protein expression in tissue derived from OSCCs, using an antibody microarray approach and LCM to procure total protein from specific microscopic cellular populations. Although the cellular quantities obtained by LCM are generally low, which makes the detection of low abundance proteins difficult, and the fact that fixation and staining procedures involved in LCM can possibly cause some artifacts, they have demonstrated that quantitative, and potentially qualitative, differences in expression patterns of multiple proteins within epithelial cells reproducibly correlate with progression of OSCCs, where most of the proteins identified in both cell types were involved in signal transduction pathways (69). In an attempt to map proteins present in saliva by 2-D PAGE, Ghafouri et al. (70) identified 100 proteins representing 20 different identities according to accession numbers. Abundant proteins found expressed in different forms were: alphaamylase, immunoglobulin A, prolactin-inducible protein, zinc-alpha (2)-glycoprotein and cystatins (S, SA, D and SN), while other proteins found were interleukin-1 receptor antagonist, von Ebner's gland protein (lipocalin-1) and calgranulin A and B (S100A8 and A9) (70), apolipoprotein A-I, beta (2)-microglobulin, glutathione S-transferase P and fatty acid-binding protein, showing that human saliva contains a large number of proteins that are involved in inflammatory and immune responses (70). Wu et al. (71) performed 2-D PAGE/SELDI-TOF-MS to identify proteins differentially expressed in two head and neck cell lines (one derived from the primary tumour and the other from a metastatic lymph node) obtained from the same patient. Two membrane-associated proteins, annexin I and annexin II, and glycolytic protein enolase-alpha were found to be upregulated, and calumenin precursor down-regulated, in the metastatic cell line, suggesting that these proteins might be important molecules in head and neck SCC invasion and metastasis (71). In most of the developing countries, tumour data banks exist only in either formalin/or ethanol-fixed paraffinembedded tissues, which produce excellent histomorphology and good preservation of macromolecules. However, and for proteomic research initiatives to take place in these countries, the use of available data banks needs to be tested and evaluated first in Western research laboratories. In a recent study, it has been shown that proteins were


successfully extracted from ethanol-fixed, paraffin-embedded prostate tissue and, when resolved by 2-D PAGE and stained by standard methods, several hundred protein molecules could be detected and were successfully analyzed by MS (72). The protein profiles obtained from ethanol-fixed tissues were found to be highly similar to those observed from frozen tissues, in contrast to the poor protein recovery from formalin-fixed material, and the protein content of specific cells that were microdissected from ethanol-fixed tissue sections using LCM could also be successfully analyzed by 2D PAGE (72). In the light of these results, ethanol-fixed tissues of oral/head and neck SCCs can be utilized for proteomic analyses. Molecular markers in OSCCs. In a recent review of 169 published articles, Schliephake et al. (73) identified 29 suggested molecular markers of relevance in OSCCs although the prognostic relevance of these tumour markers is still not quite clear. Only 12 out of 23 reports on the prognostic relevance of markers for cell cycle acceleration and proliferation indicated a significant association with prognosis, while 20 out of 29 studies on markers for tumour suppression and anti-tumour response showed prognostic relevance (73). Markers of angiogenesis exhibited only minor importance for the prognosis and treatment of OSCC (73). Results on markers of tumour invasion and metastatic potential appeared to be too premature for a statement regarding their prognostic value. In general, the location of markers within the malignant lesion, and not their quantitative assessment, is important. The analysis of the invasive front of the neoplastic lesion with regard to the occurrence of molecular markers is supposed to be of great importance for prognostication, and it is expected that genomics and proteomics research in OSCCs might provide better and sensitive molecular biomarkers that can be used in these cancers. Gene therapy. The approach of gene therapy is important to the treatment of many diseases and, to achieve promising gene therapy strategies for OSCCs, there is a need to understand the molecular mechanism(s) involved in the development of the disease, yet no single molecule characterises all types of OSCCs. In oral/head and neck SCCs, replacement of TSGs, specially the p53 TSG, in cancer cells has been studied and clinical trials have shown evidence of gene transduction and expression, mediation of apoptosis and clinical responses including pathological complete responses (reviewed in 74-78). However, efforts to improve gene delivery systems, design of immunogene and antiangiogenesis gene therapy, as well as adjuvant use of gene therapy with conventional chemotherapy, radiation therapy and surgery in oral/head and neck SCCs are on the increase, and further refinement of these efforts are to achieve clinical success (73-78).

Future directions A progressive increase in incidence and mortality of OSCC patients is expected in developing countries. Although the disease is among the few preventable human cancers, the identification, development and validation of markers related to OSCC is challenging oral cancer research laboratories in both developed and developing nations. Despite some promising molecular markers, to find a useful molecule that can be used in early diagnosis of the disease in otherwise asymptomatic individuals consuming excessive amounts of tobacco/or alcohol has proven to be an elusive goal. Most of the current markers in OSCC are directed towards formulating a clinical decision-making process, following an initial suspicion raised by more conventional means. This points to a need for targets that can be of practical value to the healthcare system, where use of such markers in diagnosis, prognosis and therapeutic selection can become more common and the type(s) of samples to be tested can be expanded. Expected information from these technologies might soon exert a dramatic change in the pace of OSCCs research and may have a positive impact on the care of OSCC patients. There is a need for a new age of OSCCs molecular target discovery and combined proteomic and genomic information will be mined for markers useful to the five main known categories of tumour biomarkers: screening, diagnosis, prognosis, therapeutic monitoring and prediction of OSSC recurrence. Biomarkers, as competing technologies for cancer screening, may be judged on their effectiveness in reducing cancer mortality by the early detection of precancerous lesions, which are amenable to surgical removal or treatment by chemotherapy and chemoprevention. Identification of biomarkers is increasingly being used for investigation and is producing an enormous amount of information, and with the rapid advances in genomic and proteomic technology, the discovery of useful biomarkers has arisen to unravel the unique signatures of cancer cells. These signatures are enabling investigators to pose scientific questions and address problems that, until recently, were inconceivable. For instance, the focus on biomarker-based disease detection has now shifted from one biomarker to a panel of biomarkers, in distinguishing normal cells from precancerous or cancerous cells. Statistical methodologies are being developed to analyze multivariate data in reference to disease outcome, with sophisticated bioinformatics tools and computation algorithms developed to extract knowledge from data generated by highthroughput technology in genomics and proteomics. The role of MALDI-TOF-MS/SELDI–TOF-MS in cancer fingerprinting is very promising and the widespread availability of this expertise will go a long way in helping diagnosis and monitoring of several malignancies, including OSCCs. New signal pathway-

291

CANCER GENOMICS & PROTEOMICS 1: 283-294 (2004) targeted pharmacotherapeutics, such as trastuzumab for breast carcinoma, as well as imatinib for gastrointestinal stromal tumours and chronic myelogenous leukaemia, have come from an enhanced understanding of the genetic and protein alterations within a tumour. It is high time that oral cancer becomes the next candidate for targeted therapies, since genomic searches for novel genetic alterations and proteomic profiling of OSCCs are just the beginning of the changes yet to come in diagnosis and treatment of this disease. The coevolution of genomics and proteomics as complementary approaches to neoplasia will allow us to move closer to the goals of earlier detection, improved prevention and institution of a molecular-based, tumour-specific approach to the treatment of OSCC of each individual patient. Following the era of the human genome and the emergence of innovative technologies, there is a good chance for oral cancer researchers to collaborate and conduct comprehensive analysis of the genome, transcriptomes and proteomes in health and disease by sharing knowledge between developed and developing world. In the former scenario, although prevalence of OSCC is generally low, substantial advances have been made in the application of advanced techniques for early detection, effective treatment and molecular understanding of OSCCs. In the developing part of the world, however, where OSCC is a pressing problem, other types of cancers like lung, breast, colon and leukaemia might overshadow the problem, given the limited resources available. Nevertheless, it is predicted that soon the ever-increasing body of translational research will move from the laboratories to the clinics and revolutionise the gloomy scenario currently facing the increasing numbers of patients with OSCCs.

Acknowledgements The financial support to the first author (RM) by the Department of Biotechnology, Government of New Delhi, India is gratefully acknowledged. The Norwegian Research Council, Meltzer Høyskolefond, Richard With-Johnsen of Hustru Fanny’s Fond and Felleslegatet til Fordel for Biologisk Forskning supported this study.

References 1 Pindborg JJ, Reichart PA, Smith CJ and van der Waal I: World Health Organisation. International Histological Classification of Tumours: Histological Typing of Cancer and Precancer of the Oral Mucosa. 2nd ed. Berlin, Heidleberg: Springer, 1-40, 1997. 2 Landis SH, Murray T, Bolden S and Wingo PA: Cancer statistics. Ca Cancer J Clin 49: 8-31, 1999. 3 Parkin DM, Pisani P and Ferlay J: Estimates of the worldwide incidence of 25 major cancers in 1990. Int J Cancer 80: 827-841, 1999. 4 Gupta PC, Pindborg JJ and Mehta FS: Comparison of carcinogenicity of betel quid with and without tobacco: An epidemiologic review. Ecol Disease 1: 213-219, 1982.

292

5 Mehrotra R, Singh M, Kumar D, Pandey AN, Gupta RK and Sinha US: Age specific incidence rate and pathological spectrum of oral cancer in Allahabad. Ind J Med Sci 57: 400-4, 2003. 6 Gupta PC and Ray CS: Smokeless tobacco and health in India and South Asia. Respirology 8: 419-431, 2003. 7 American Cancer Society: Cancer Facts and Figures. Atlanta, Ga: American Cancer Society, 2003. 8 Bray I, Brennan P and Boffetta P: Projections of alcohol- and tobacco-related cancer mortality in Central Europe. Int J Cancer 87: 122-128, 2000. 9 Franceschi S, Bidoli E, Herrero R and Munoz N: Comparison of cancers of the oral cavity and pharynx worldwide: etiological clues. Eur J Cancer (Oral Oncol) 36: 106-115, 2000. 10 IARC Monographs programme on the evaluation of the carcinogenic risk of chemicals to humans. Preamble: IARC Monogr Eval Carcinog Risk Chem Hum 39: 13-32, 1986. 11 Scully C: Oral squamous cell carcinoma; from a hypothesis about a virus, to concern about possible sexual transmission. Eur J Cancer (Oral Oncol) 38: 227-234, 2002. 12 Foulkes WD, Brunet JS, Kowalski LP, Narod SA and Franco EL: Family history is a risk factor for squamous cell carcinoma of the head and neck in Brazil: a case-control study. Int J Cancer 63: 769-773, 1995. 13 Foulkes WD, Brunet JS, Sieh W, Black MJ, Shenouda G and Narod SA: Familial risk of squamous cell carcinoma of the head and neck: a retrospective case-control study. Br Med J 313: 716720, 1996. 14 Das BR and Nagpal JK: Understanding the biology of oral cancer. Med Sci Monit 8: 258-262, 2002. 15 Williams HK: Molecular pathogenesis of oral squamous carcinoma. Mol Pathol 53: 165-172, 2000. 16 Prime SS, Eveson JW, Guest PG, Parkinson EK and Paterson IC: Early genetic and functional events in the pathogenesis of oral cancer. Radiat Oncol Investig 5: 93-96, 1997. 17 URL:http://www.tumour-gene.org/Oral/oral.html. 18 Sugerman PB, Joseph BK and Savage NW: Review article: The role of oncogenes, tumour suppressor genes and growth factors in oral squamous cell carcinoma: a case of apoptosis versus proliferation. Oral Dis 1: 172-188, 1995. 19 Scully C, Field JK and Tanzawa H: Genetic aberrations in oral or head and neck squamous cell carcinoma (SCCHN): 1. Carcinogen metabolism, DNA repair and cell cycle control. Eur J Cancer (Oral Oncol) 36: 256-263, 2000. 20 Scully C, Field JK and Tanzawa H: Genetic aberrations in oral or head and neck squamous cell carcinoma 2: chromosomal aberrations. Eur J Cancer (Oral Oncol) 36: 311-327, 2000. 21 Scully C, Field JK and Tanzawa H: Genetic aberrations in oral or head and neck squamous cell carcinoma 3: clinico-pathological applications. Eur J Cancer (Oral Oncol) 36: 404-413, 2000. 11 Scully C: Oncogenes, onco-suppressors, carcinogenesis and oral cancer. Br Dent J 173: 53-59, 1992. 23 Saranath D, Panchal RG and Nair R: Oncogene amplification in squamous cell carcinoma of the oral cavity. Jpn J Cancer Res 80: 430-437, 1989. 24 Pomerantz RG and Grandis JR: Review: The role of epidermal growth factor receptor in head and neck squamous cell carcinoma. Curr Oncol Rep 5: 140-146, 2003. 25 Jefferies S and Foulkes WD: Genetic mechanisms in squamous cell carcinoma of the head and neck. Eur J Cancer (Oral Oncol) 37: 115-126, 2001.


26 Levine AJ: Tumour suppressor genes. Annu Rev Biochem 62: 623-651, 1993. 27 Levine AJ: p53, the cellular gatekeeper for growth and division. Cell 88: 323-331, 1997. 28 Shillitoe EJ, May M, Patel V, Lethanakul C et al: Genome-wide analysis of oral cancer--early results from the Cancer Genome Anatomy Project. Eur J Cancer (Oral Oncol) 36: 8-16, 2000. 29 Werness BA, Levine AJ and Howley PM: Association with human papillomavirus types 16 and 18 E6 Proteins with p53. Science 248: 76-79, 1990. 30 Greenblatt MS, Bennettm WP, Hollstein M and Harris CC: Mutation in the p53 tumour suppressor gene: clues to cancer etiology and molecular pathogenesis. Cancer Res 54: 48554878, 1994. 31 Olivier M, Eeles R, Hollstein M, Khan MA et al: The IARC TP53 database: new online mutation analysis and recommendations to users. Hum Mutat 19: 607-614, 2002. 32 Nylander K, Dabelsteen E and Hall PA: The p53 molecule and its prognostic role in squamous cell carcinomas of the head and neck. J Oral Pathol Med 29: 413-425, 2000. 33 Jones A: A general review of the p53 gene and oral squamous cell carcinoma. Ann R Australas Coll Dent Surg 14: 66-69, 1998. 34 Raybaud-Diogene H, Tetu B, Morency R, Fortin A and Monteil RA: p53 overexpression in head and neck squamous cell carcinoma: review of the literature. Eur J Cancer (Oral Oncol) 32: 143-149, 1996. 35 Kashiwazaki H, Tonoki H, Tada M, Chiba I et al: High frequency of p53 mutations in human oral epithelial dysplasia and primary squamous cell carcinoma detected by yeast functional assay. Oncogene 15: 2667-2674, 1997. 36 Rowley H, Sherrington P, Helliwell TR, Kinsella A and Jones AS: p53 expression and p53 gene mutation in oral cancer and dysplasia. Otolaryngol Head Neck Surg 118: 115-123, 1998. 37 Shin DM, Lee JS, Lippman SM, Lee JJ et al: p53 expressions: predicting recurrence and second primary tumours in head and neck squamous cell carcinoma. J Nat Cancer Inst 88: 519529, 1996. 38 Ruas M and Peters G: The p16INK4a/CDKN2A tumor suppressor and its relatives. BBA 1378: F115-177, 1998. 39 Fenech M: Chromosomal biomarkers of genomic instability relevant to cancer. Drug Discov Today 7: 1128-1137, 2002. 40 Caburet S, Conti C and Bensimon A: Combing the genome for genomic instability. Trends Biotechnol 20: 344-350, 2002. 41 Cheah PY, Eu KW and Seow FC: Update of genetics in colorectal carcinomas: genomic instability and somatic evolution. Ann Acad Med Singapore 29: 331-336, 2000. 42 Reddel RR: Alternative lengthening of telomeres, telomerase, and cancer. Cancer Lett 194: 155-162, 2003. 43 Mao L, Lee JS and Fan YH: Frequent microsatellite alterations at chromosomes 9p21 and 3p14 in oral premalignant lesions and their value in cancer risk assessment. Nature Med 6: 682-685, 1996. 44 Miyoshi Y, Tsukinoki K and Imaizumi T: Telomerase activity in oral cancer. Eur J Cancer (Oral Oncol) 35: 283-289, 1998. 45 Patel MM, Patel DD, Parekh LJ, Raval GN et al: Evaluation of telomerase activation in head and neck cancer. Eur J Cancer (Oral Oncol) 35: 510-515, 1999. 46 Califano J, Ahrendt SA, Meininger G, Westra WH et al: Detection of telomerase activity in oral rinses from head and neck squamous cell carcinoma patients. Cancer Res 56: 5720-5722, 1996.

47 Bergers G and Benjamin LE: Tumorigenesis and the angiogenic switch. Nat Rev Cancer 3: 401-410, 2003. 48 Bouma-ter Steege JC, Mayo KH and Griffioen AW: Angiostatic proteins and peptides. Crit Rev Eukaryot Gene Expr 11: 319334, 2001. 49 Kerbel RS: Tumour angiogenesis: past future and present. Carcinogenesis 21: 505-515, 2000. 50 O-charoenrat P, Rhys-Evans P and Eccles SA: Expression of vascular endothelial growth factor family members in head and neck squamous cell carcinoma correlates with lymph node metastasis. Cancer 92: 556-568, 2001. 51 Venter JC, Adams MD, Myers EW, Li PW et al: The sequence of the human genome. Science 291: 1304-1351, 2001. 52 Lander ES, Linton LM, Birren B, Nusbaum C et al: Initial sequencing and analysis of the human genome. Nature 409: 860-921, 2001. 53 Patel V, Leethanakul C and Gutkind JS: New approaches to the understanding of the molecular basis of oral cancer. Crit Rev Oral Biol Med 12: 55-63, 2001. 54 Fend F, Emmert-Buck MR, Chuaqui R, Cole K et al: ImmunoLCM: laser capture microdissection of immunostained frozen sections for mRNA analysis. Am J Pathol 154: 61-66, 1999. 55 Alevizos I, Mahadevappa M, Zhang X, Ohyama H et al: Oral cancer in vivo gene expression profiling assisted by laser capture microdissection and microarray analysis. Oncogene 20: 61966204, 2001. 56 DeRisi J, Penland L, Brown PO, Bittner ML et al: Use of a cDNA microarray to analyse gene expression patterns in human cancer. Nat Genet 14: 457-460, 1996. 57 Warner GC, Reis PP, Makitie AA, Sukhai MA et al: Current applications of microarrays in head and neck cancer research. Laryngoscope 114: 241-248, 2004. 58 Kim DS and Watkinson JC: Gene chip expression analysis in head and neck cancer. Clin Otolaryngol 27: 296-303, 2002. 59 Schmalbach CE, Chepeha DB, Giordano TJ, Rubin MA et al: Molecular profiling and the identification of genes associated with metastatic oral cavity/pharynx squamous cell carcinoma. Arch Otolaryngol Head Neck Surg 130: 295-302, 2004. 60 Belbin TJ, Singh B, Barber I, Socci N et al: Molecular classification of head and neck squamous cell carcinoma using cDNA microarrays. Cancer Res 62: 1184-1190, 2002. 61 Ginos MA, Page GP, Michalowicz BS, Patel KJ et al: Identification of a gene expression signature associated with recurrent disease in squamous cell carcinoma of the head and neck. Cancer Res 64: 55-63, 2004. 62 Leethanakul C, Patel V, Gillespie J, Pallente M et al: Distinct pattern of expression of differentiation and growth-related genes in squamous cell carcinomas of the head and neck revealed by the use of laser capture microdissection and cDNA arrays. Oncogene 19: 3220-3224, 2000. 63 Garnis C, Campbell J, Zhang L. Rosin MP and Lam WL: OCGR array: an oral cancer genomic regional array for comparative genomic hybridization analysis. Eur J Cancer (Oral Oncol) 40: 511-519, 2004. 64 Ibrahim SO, Aarsaether N, Holsve MK, Kross KW et al: Gene expression profile in oral squamous cell carcinomas and matching normal oral mucosal tissues from black Africans and white Caucasians: the case of the Sudan vs. Norway. Eur J Cancer (Oral Oncol) 39: 37-48, 2003.

293

CANCER GENOMICS & PROTEOMICS 1: 283-294 (2004) 65 Pandey A and Mann M: Proteomics to study genes and genomes. Nature 405: 837-846, 2000. 66 Schramm A, Apostolov O, Sitek B, Pfeiffer K et al: Proteomics: techniques and applications in cancer research. Klin Padiatr 215: 293-297, 2003. 67 Marvin LF, Roberts MA and Fay LB: Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry in clinical chemistry. Clin Chim Acta 337: 11-21, 2003. 68 He Q-Y, Chen J, Kung H-F, Yuen A P-W and Chiu J-F: Identification of tumor-associated proteins in oral tongue squamous cell carcinoma by proteomics. Proteomics 4: 271278, 2004. 69 Knezevic V, Leethanakul C, Bichsel VE, Worth JM et al: Proteomic profiling of the cancer microenvironment by antibody arrays. Proteomics 1: 1271-1278, 2001. 70 Ghafouri B, Tagesson C and Lindahl M: Mapping of proteins in human saliva using two-dimensional gel electrophoresis and peptide mass fingerprinting. Proteomics 3: 1003-1015, 2003. 71 Wu W, Tang X, Hu W, Lotan R et al: Identification and validation of metastasis-associated proteins in head and neck cancer cell lines by two-dimensional electrophoresis and mass spectrometry. Clin Exp Metastasis 19: 319-326, 2002. 72 Ahram M, Flaig MJ, Gillespie JW, Duray PH et al: Evaluation of ethanol-fixed, paraffin-embedded tissues for proteomic applications. Proteomics 3: 413-421, 2003.

294

73 Schliephake H: Prognostic relevance of molecular markers of oral cancer–a review. Int J Oral Maxillofac Surg 32: 233245, 2003. 74 Moon C, Oh Y and Roth JA: Current status of gene therapy for lung cancer and head and neck cancer. Clin Cancer Res 9: 5055-5067, 2003. 75 Xi S and Grandis JR: Gene therapy for the treatment of oral squamous cell carcinoma. J Dent Res 82: 11-16, 2003. 76 Frank DK: Gene therapy for head and neck cancer. Surg Oncol Clin N Am 11: 589-606, 2002. 77 Nemunaitis J and O'Brien J: Head and neck cancer: gene therapy approaches. Part 1: adenoviral vectors. Expert Opin Biol Ther 2: 177-185, 2002. 78 Nemunaitis J and O'Brien J: Head and neck cancer: gene therapy approaches. Part II: genes delivered. Expert Opin Biol Ther 2: 311-324, 2002.

Received June 6, 2004 Accepted August 23, 2004