Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Use of allelic loss to predict malignant risk for low-grade oral epithelial lesions Cheng, Xing 2000

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2000-0038.pdf [ 5.96MB ]
Metadata
JSON: 831-1.0089375.json
JSON-LD: 831-1.0089375-ld.json
RDF/XML (Pretty): 831-1.0089375-rdf.xml
RDF/JSON: 831-1.0089375-rdf.json
Turtle: 831-1.0089375-turtle.txt
N-Triples: 831-1.0089375-rdf-ntriples.txt
Original Record: 831-1.0089375-source.json
Full Text
831-1.0089375-fulltext.txt
Citation
831-1.0089375.ris

Full Text

Use of Allelic Loss to Predict Malignant Risk for Low-Grade Oral Epithelial Lesions by Xing Cheng M.Sc.,Xian Medical University, 1986 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (Department of Oral Biological and Medical Sciences, Faculty of Dentistry) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA November 1999 ©Xing Cheng, 1999 In presenting this thesis in partial fulfillment of the requirements for an advanced degree at the University of British Columbia, I agree that the library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of Oral Biological and Medical Sciences The University of British Columbia Vancouver Canada ABSTRACT Oral squamous cell carcinomas (SCC) are believed to develop through progressing stages of oral premalignant lesions (histologically divided into hyperplasia, mild dysplasia, moderate dysplasia, severe dysplasia and carcinoma in situ, CIS) before finally become invasive. Prognosis is poor once invasion occurs and SCC is formed. The key to improve this gloomy prognosis may lie in early diagnosis and proper management of oral premalignant lesions. However, the majority of oral premalignant lesions, particularly low-grade lesions (hyperplasia, mild and moderate dysplasia) do not progress into cancer. Since carcinogenesis is underlain by changes to critical control genes, it is hypothesized that the small percentage of progressing premalignancies differs genetically from the majority of non-progressing lesions. One of the best approaches to identifying genetic changes critical to oral cancer progression is to compare progressing and non-progressing oral premalignant lesions. However, such samples are rare, and little information is available on genetic changes in progressing and non-progressing lesions. This thesis, for the first time, compared genetic changes in 116 cases of progressing and non-progressing low-grade oral premalignant lesions by microsatellite analysis for loss of heterozygosity (LOH) using 19 probes for 7 chromosome arms. The progressing and non-progressing cases showed dramatically different L O H patterns of multiple allelic losses. A n essential step for progression seems to involve L O H at 3p and/or 9p as virtually all progressing cases showed such loss. However, L O H at 3p and/or 9p also occurred in non-progressing cases. Individuals with L O H at 3p and/or 9p but no other arms exhibit only a i i slight increase of 3.8-fold in relative risk for developing cancer. In contrast, individuals with additional losses (on 4q, 8p, l l q , 13q or 17p), which appeared uncommon in non-progressing cases, showed 3 3-fold increases in relative cancer risk. The results suggest that L O H patterns wi l l facilitate the prediction of the malignant potential of low-grade premalignancies. They also demonstrate the predictive value of assaying allelic loss at arms other than 3p and/or 9p. We may more precisely predict the malignant potential of low-grade premalignancies i f we combine molecular analysis with the clinical assessment to premalignent lesions. Clinically, the data also support the belief that the molecular tools such as microsatellite analysis for L O H may be used in differentiating between progressing and non-progressing lesions. The identification of molecular changes that can be used to predict the likely behavior of low-grade lesions would allow the clinician to identify which patients with low-grade lesions should be managed more aggressively and thus, should improve prognosis. i i i TABLE OF CONTENTS A B S T R A C T i i T A B L E O F C O N T E N T iv L I S T O F T A B L E S v i i L I S T O F F I G U R E S v i i i A B B R E V I A T I O N S ix A C K N O W L E D G M E N T S : x i D E D I C A T I O N x i i 1. I N T R O D U C T I O N 1 1.1. Importance of studying oral premalignant lesions 1 1.2.Oral mucosa 2 1.3.Oral premalignant lesions and their relation to oral SCC 2 1.4. Problems with the histological progression model 5 1.5. Molecular biology of carcinogenesis 6 1.5.1. Genetic pathway of carcinogenesis 6 1.5.2. The major genes involved in tumorigenesis 9 1.6. Loss of heterozygosity (LOH) and microsatellite analysis 11 1.7. L O H in oral and head and neck malignant lesions 13 1.8. L O H in oral and head and neck premalignant lesions 18 1.9. A molecular progressing model for oral SCC 21 2. S T A T E M E N T O F T H E P R O B L E M S 23 3. O B J E C T I V E S 25 4. H Y P O T H E S I S 26 iv 5. E X P E R I M E N T A L D E S I G N A N D S T A T I S T I C A L A N A L Y S I S 26 5.1. Experimental design 27 5.2. Statistical analysis 28 6. M A T E R I A L S A N D M E T H O D S 29 6.1. Sample collection 29 6.2. Sample sets 29 6.3. Histological diagnostic criteria for the samples 31 6.4. Slide preparation 32 6.5. Microdissection 33 6.6. Sample digestion and D N A extraction 33 6.7. D N A quantitation 34 6.8. Primer-extension preamplification (PEP) 34 6.9. Coding samples 35 6.10. End-labeling 35 6.11. P C R amplification 35 6.12. Scoring of L O H 36 7. R E S U L T S 38 7.1. Frequency of allelic loss 38 7.2. Pattern of allelic loss 39 7.3. Progression risk 41 7.4. Comparison of L O H pattern in matching premalignant and malignant lesions 44 7.5. Clinical history of the lesions 46 8. D I S C U S S I O N 53 8.1. High frequency of allelic loss characterized dysplastic lesions 53 8.2 Increased frequency of allelic loss in progressing lesions 54 v 8.3. The pattern of allelic loss characterized progressing lesions 55 8.4. L O H as markers to predict cancer risk - further statistic analysis 57 8.5. Most progressed lesions may be derived by clonal outgrowth from the earlier lesions.. 60 8.6. Does evidernce of recurrence serve as a further indicator of cancer risk in premalignant lesions? 61 8.7. Summary 63 9. R E F E R E N C E S 65 10. A P P E N D I X 76 vi LIST OF TABLES Table 1. L O H frequencies (%) in Head and Neck and Oral Cancers 76 Table 2. L O H and oral premalignant lesions 20 Table 3. Samples sets and groups 30 Table 4. Accumulation of allelic loss in progressing and non-progressing lesions 38 Table 5. Patterns of allelic loss in progressing and non-progressing lesions 40 Table 6. The time between initial biopsy and current date /the date when lesion has progressed to S C C 78 Table 7. Probability of lesions not progressing to cancer after 5 years follow-up 43 Table 8. Comparisons of L O H patterns in initial biopsies with that seen in CIS/SCC that later developed at the same anatomical site 44 Table 9. Characteristics of patients with dysplasia 46 Table 10. L O H frequencies and clinical history in progressing lesions rAl Table 11. Allel ic loss in recurrent premalignant lesions without a history of progression 52 v i i LIST OF FIGURES Figure 1. Histological progression model of oral premalignant and malignant lesions 4 Figure 2. Genetic pathway of carcinogenesis 8 Figure 3. Tumor Progression model for head and neck cancers as proposed by Califano et al 1996 22 Figure 4. Probability of having no progression to cancer, according to LOH pattern 42 Figure 5. LOH analysis of 2 patients (a, b) 45 ABBREVIATIONS A P C adenomatous polyposis coli gene B C C A British Columbia Cancer Agency bcl-1,2 B-cell lymphoma C D K Cyclin-dependent kinase CIS Carcinoma in situ D N A Deoxyribonucleic acid docl V H L The gene responsible for von Hippel-Lidadu syndrome erbB Erythroblastosis FHIT Fragile histidine triad H & E Hematoxylin and eosin H H S C C Head and neck SCC L O H Loss of heterozygosity M E N 1 Multiple endocrine neoplasia type 1 N B C C S Nevoid basal cell carcinoma syndrome NF1,2 neurofibromatosis type I and II PEP Primer-extension preamplification PC Phenol-chloroform P C R Polymerase chain reaction p 16/ INK4A/MTS- 1/CDK2A A tumor suppressor gene, encodes a cell cycle protein that halt cell-cycle progression Rat sarcoma ras Rb R F L P Retinoblastoma gene Restriction fragment length polymorphism ix SCC Squamous cell carcinoma SDS Sodium dodecyl sulfate T S G Tumor suppressor gene T0R-II Transforming growth factor type II receptor U B C University of British Columbia W H O World Health Organization x ACKNOWLEDGMENTS I would like to express my deepest thanks to my supervisors, Dr. Lewei Zhang and Dr . Mi r i am P. Rosin, for giving me the opportunity to learn and earn an M . S c . degree under their professional guidance, as well as their support throughout my degree. I am also grateful to Dr. Robert Priddy and Dr. Joel Epstein for their valuable contributions to clinical issues in my thesis. Thanks also to Dr. Douglas Waterfield for chairing my final oral examination. I appreciate very much the assistance of all the people in the laboratory. Special thanks go to my dearest friend, Xiao L e i Zhang for her special support and understanding. xi DEDICATION To my Dear Mom and Dad To my Dearest Daughter 1. INTRODUCTION .1. Importance of studying oral premalignant lesions Head and neck cancers, including those of the mouth and upper air and food passages (oral cavity, oropharynx, nasopharynx, hypopharynx, and larynx), the salivary glands, and thyroid, account for approximately 6% o f all human malignancies in the western world. The incidence is much higher in the Far East and India in particular with up to 40% of malignancies occurring in the head and neck region. The prognosis of oral cancer has not significantly improved during the past two decades: the 5-year-survival rate is still less than 50% and is one of the lowest among the major types of cancers despite recent advances in surgery, radiotherapy, and chemotherapy (Raybaud-Diogene et al 1996; Todd et al 1995). A s oral cancer is believed to progress from oral premalignant lesions, the key to improve the gloomy prognosis of oral cancer may lie in early diagnosis and proper management of oral premalignant lesions. The research of our lab focuses on oral premalignant and malignant lesions. This thesis represents one aspect of the focus the investigation of molecular changes in early oral premalignancies of oral mucosa. 1 1.2. Oral mucosa The oral cavity is lined by oral mucosa, which consists of overlying epithelium and underlying lamina propria. The overlying epithelium of oral mucosa is stratified squamous epithelium. The connective tissue lamina propria underneath the epithelium contains blood and lymphatic vessels, small nerves, fibroblasts, collagen, and elastic fibers. The stratified squamous epithelium could be divided largely into basal and prickle cells in addition to there are other types: merkel, langerhans and melanocyte cells. The one-layered basal cells separate the overlying epithelium from the underlying connective tissue. They are the only cells that have the capacity to divide in the epithelium and subsequently give rise to either more new basal cells or differentiate into prickle cells located above them. The majorities of oral premalignant and malignant lesions arise from this stratified squamous epithelium of oral mucosa and the malignant tumors are called squamous cell carcinoma (SCC) . 1.3. Oral premalignant lesions and their relation to oral SCC Oral S C C is believed to be a result of a multistage carcinogenesis process over a long period of time. This multistage process involves progression from normal to premalignant lesions and finally invasive S C C . A premalignant or precancerous lesion has been defined by the Wor ld Health Organization ( W H O , 1978) as a morphologically altered tissue in which cancer is more l ikely to occur than in its apparently normal counterpart. In the oral cavity, most premalignant lesions present clinically as leukoplakias. Other types o f clinical forms exist, such as erythroplakia and possibly lichen planus ( W H O 1978). Leukoplakia means a "white patch" and occurs on mucous membranes such as the mucosa o f the oropharynx, 2 larynx, esophagus, and genital tract. Not necessarily white, leukoplakias also may appear yellow to light brown, especially in smokers. The World Health Organization ( W H O , 1978) defines leukoplakia in the oral cavity as a white patch or plaque o f oral mucosa that cannot be characterized clinically or pathologically as any other diagnosable disease and is not removed by rubbing. It is only a clinical term. When a biopsy is taken, a leukoplakia w i l l show microscopically hyperkeratosis and/or epithelial hyperplasia (acanthosis) with or without epithelial dysplasia. Dysplastic lesions are further divided into mild, moderate, and severe forms depending upon how much of the tissue is dysplastic. M i l d dysplasia is lesions in which the dysplastic cells are confined to the basal layer and the cells exhibit the smallest degree of the above changes. Wi th moderate and severe dysplasia, the epithelial layers involved and the severity of the cellular changes are progressively increased. In carcinoma in situ, the dysplastic cells occupy the entire thickness of the epithelium (bottom to top changes) although the basement membrane is still intact. Invasion of dysplastic cells through the basement membrane into the underlying stroma and/or the dissemination of these cells to other sites through lymphoid and circulatory systems are events associated with development of invasive S C C . The presence and absence of dysplasia and the degree of dysplasia is believed to have a huge impact on the malignant risk of the premalignant lesions. A l l studies to date have shown that leukoplakia with dysplasia has a higher malignant risk than leukoplakia without dysplasia. A large clinical study by Silverman et al (1984) found that during a mean onset of 7.2 years after presentation, more than 36% of leukoplakia lesions with microscopic epithelial dysplastic features eventually underwent malignant transformation whereas leukoplakia without dysplasias only demonstrated a malignancy rate of 15%. The risk of dysplasia and degree o f dysplasia is further demonstrated by studies from the uterine cervix and other 3 systems and organs including skin and respiratory system. As a result, currently the gold standard for judging the malignant potential of premalignant lesions in these organs and systems, including the oral cavity, is the presence and degree of dysplasia. Using these criteria, a histological progression model has been established for the oral cavity (Fig. 1). Premalignant lesions are classified histologically into categories with progressively increased risk of becoming invasive SCC: epithelial hyperplasia (without dysplasia), mild, moderate and severe dysplasias, and carcinoma in situ (CIS). Figure 1: Histological progression model of oral premalignant and malignant lesions HISTOM-1 JPC Other factors also affect the malignant potential of oral premalignant lesions. These include location and duration of the lesion, gender of the patient, appearance of the lesion (homogenous vs non-homogenous) and presence of C. albicans (Waal, 1997). Most studies on the malignant transformation of oral premalignant lesions have been done on leukoplakias (clinical presentation of oral premalignant lesions), frequently without knowledge of dysplasia 4 for all the study cases. The reported malignant risk for leukoplakia varies from study to study, ranging from as low as 0.13% to as high as 24% depending upon the patient population and follow-up time (Papadimitrakopoulou et al, 1997; Lumerman et al, 1995; Silverman et al, 1984). 1.4. Problems with the histological progression model The histological progression model has a better predictive value for high-grade preinvasive lesions (severe dysplasia and CIS) which are believed by many to have a much higher possibility of progression into invasive lesions than low-grade lesions including hyperplasia without dysplasia and those with low-grade dysplasia (mild and moderate dysplasias) i f left untreated. Some even, believe that oral CIS w i l l inevitably become cancer i f left untreated (Regezi et al, 1989). A s a result, high-grade preinvasive lesions are generally treated aggressively, and the histological progression model has served as a good guidance for the aggressive treatment. While the histological progression model has a good or reasonable predictive value for high-grade preinvasive lesions, it is poor at predicting the malignant potential of low-grade lesions. This model is more problematic in guiding the treatment of the low-grade lesions. The majority of these low-grade lesions do not progress into oral cancer, either remaining static or regressing, with only a small percentage progressing. On the other hand, these low-grade lesions constitute the bulk of leukoplakias and account for more than 90% of leukoplakias (hyperplasia without dysplasia, 80.1%; early dysplasia, 12.2%; late dysplasia, 4.5% and S C C , 3.1%) (Waldron Lumerman et al, 1975). This poses a management dilemma for clinicians. Aggressive treatment does not seem to be justified for the majority of these 5 lesions, both in terms of side effects and cost. New methods that could identify that small percentage of progressing low-grade lesions from the majority of non-progressing lesions are highly desired. The significance of establishing these new methods lies in two aspects. First, this w i l l facilitate the understanding of the mechanisms of early carcinogenesis; and second, this w i l l have direct impact on the clinical management of these lesions. If we could understand the critical events occurring during early carcinogenesis, we may not only be able to predict the malignant potential at a very early stage, but can also plan management of this small percentage of progressing lesions accordingly (e.g., aggressive treatment or chemoprevention). Successful treatment of these early lesions and prevention of their progression w i l l decrease the mortality and morbidity of oral S C C drastically. A central dogma of carcinogenesis is that alteration to critical control genes underlies malignant transformation. The investigation of these critical changes in genes has been greatly facilitated recently with the rapid development of molecular biology techniques. This thesis has investigated some of the molecular changes in early oral premalignant lesions. 1.5. Molecular biology of carcinogenesis 1.5.1. Genetic pathway of carcinogenesis In 1976, Nowel proposed that neoplastic transformation occurred in a single cell that had a critical genetic alteration giving it a growth advantage over its neighboring cells. The tissue would then go on to accumulate multiple mutations with these subsequent mutations being random in nature; however, the determination of which mutated cells would expand 6 into genetic clones in the tissue would be dependent on a variety of intracellular and environmental factors, including earlier mutations. Thus, all daughter cells in a tumor would share early initiating events but during tumorigenesis subclones would arise with additional genetic changes (that give them a selective advantage) and lead to heterogeneity. The non-random nature of this mutation selection is responsible for the preferred molecular pathways that are envisaged for many cancers. Ilyas Lumerman et al (1996) summarized that for a normal cell to become malignant, it must acquire the stepwise accumulation of genetic changes and a minimum number of necessary mutations which help it to overcome growth controls. These mutations and the order of development of the mutation profile comprise the "genetic pathway" of carcinogenesis. 7 Figure 2. Genetic pathway of carcinogenesis *Mutation profile: the group of mutations which are essential for tumour development. * * A cell , gets greatest survival advantage over its surrounding cells, w i l l undergo colonial expansion and outgrowth, becomes predominant. ***Not selected, become apoptosis. The understanding of the genetic pathways is essential in identifying those lesions that are following the pathways to cancer from those lesions with less critical genetic changes and with less cancer risk. Since the alteration to critical control genes underlies malignant transformation, it may be logical to assume that progressing premalignant lesions are genetically different from morphologically similar non-progressing lesions. The identification of molecular changes that can be used to predict the l ikely behavior of low-8 grade lesions would allow the clinician to identify which patients with low-grade lesions should be managed more aggressively (either by more frequent screening or by early treatment, using traditional approaches such as surgery, or newer techniques such as chemopreventive regimes), and to better control and prevent the progression of premalignant lesions to cancer. 7.5.2. The major genes involved in tumorigenesis It is now well established that clonal evolution of cancer is due to a progressive accumulation of critical genetic alterations, including at least two large groups: (1) the proto-oncogenes, which can be activated in the tumorigenic process to increase cell proliferation and induce malignant transformation, and (2) the tumor suppressor genes (TSGs), which maintain several growth control checkpoints and control the ability of cells to invade or metastasize. Proto-oncogenes code for proteins that regulate the many functions of the normal cell, including the control of cell division, the production of enzymes that alter cellular activity, and the production of intercellular adhesion molecules and cell surface molecules that are bound by extracellular molecules. They include genes for growth factors, growth factor receptors, protein kinases, signal transducers, nuclear phosphoproteins and transcription factors. Proto-oncogenes act in a dominant fashion to positively regulate cell growth and differentiation. Mutation of these proto-oncogenes to oncogenes can occur in the coding region of gene, resulting in alteration of structure and activity in coded proteins. Many oncogenes have been identified in the literature; however, few of them have been reported to occur in HNSCC. Some of the oncogenes that have been found altered or expressed at 9 abnormal levels in head and neck cancers are ras, cyclin-Dl, myc, erbB, bcl-1, bcl2, int-2 CK8 and CK19 (Staibano et al, 1998; Kannan et al 1996; Bartkova et al 1995; X u et al 1995; Gaffey et al 1995; Michalides et al 1995; Anderson et al 1994; Clark et al 1993; Eversole et al 1993; Wong et al 1993; Riviere et al 1990; Yokota et al 1986). In contrast, TSGs function antagonistically with cellular proto-oncogenes to negatively regulate cell growth and differentiation. The functions of TSGs must be lost in order for tumorigenesis to occur. According to Knudson's hypothesis (1985), both copies of a tumor suppressor gene have to be inactivated for its protective function to be lost in a cell. Experience with known suppressor genes, such as the retinoblastoma gene, suggests that this process involves two separate events, the first quite often involving a point mutation in one allele, followed by loss o f loci containing the wi ld type gene in the remaining allele. Some of the TSGs involved in head and neck cancers include p53, Rb (retinoblastoma), andpl6INK4A (Gallo et al 1999; Jares et al, 1999; Liggett et al, 1999; Papadimitrakopoulou et al, 1999; Sartor et al, 1999; Partridge et al, 1999a and 1998; Pavelic et al, 1997; Reed et al 1996; Gleich et al 1996; Largey et al 1994). Other potential candidates are FHIT (fragile histidine triad), APC (adenomatous polyposis coli), doc-1 VHL (the gene responsible for von Hippel-Lidau syndrome) and TfiR-II (the gene coding for transforming growth factor type II receptor). (Croce et al, 1999; Uzawa et al 1999 and 1994; Mao et al 1998 and 1996; Waber et al 1996;Todd et al 1995). This meticulous balance between growth inducers (coded by proto-oncogenes) and suppressors (coded by tumour suppressor genes) controls the rate of division in normal cells. These genes are altered through a multistep process in which a cell accumulates many genetic changes, breaking the balance of normal cell growth and leading to the malignant phenotype. Recent advancement in the techniques of molecular analysis has rapidly revolutionized our 10 ability to look at these genetic alterations. M y research w i l l focus on loss of tumor suppressor genes (TSGs). Functional loss of TSGs is one of the most common genetic alterations during carcinogenesis (Leis et al, 1996). Therefore, defining chromosomal regions harboring biologically important suppressor genes may have broad practical implications not only on our comprehension of progression of tumors but also on the clinical management of cancers and premalignant lesions. This thesis has studied regions of chromosome loss that contain presumptive TSGs by employing a polymerase chain-based microsatellite analysis for loss of heterozygosity (LOH). 1.6. Loss of heterozygosity (LOH) and microsatellite analysis L O H is defined as a loss of genomic material (as small as a few thousand nucleotides to as large as a whole chromosome) in one of a pair o f chromosomes. The L O H assay is designed to assess polymorphic chromosomal regions that map close to or within putative or known recessive cancer-related genes. The concept of L O H is consistent with Knudson's two-hit hypothesis, which states that inactivation of one of the two alleles by either a germline or somatic mutation w i l l provide a growth advantage to the tumor cell because only one more inactivation of the remaining allele is needed. L O H analysis has been employed as a means of identifying critical loci containing TSGs and has subsequently led to the discovery of several important genes of this class, including the retinoblastoma (Rb) gene and the genes responsible for multiple endocrine neoplasia type 1 {MEN1), the nevoid basal cell carcinoma syndrome (NBCCS), adenomatous polyposis coli (APC), and neurofibromatosis type I and II (NF1 andNFII, respectively) (Fearon et al, 1997 and Ah-See et al, 1994). 11 Two methods have been available for the study of L O H or allelic loss: the more classical approach of restriction frequent length polymorphism (RFLP) analysis, and the newer method of microsatellite analysis. This thesis employed microsatellite analysis for at least two reasons. First, microsatellite repeat markers are highly polymorphic and well-distributed throughout the human genome. They show levels of heterozygosity between 30-80%, significantly above the level observed with the R F L P analysis based on base substitutions at endonuclease recognition sites. Second, this PCR-based approach is much more sensitive than the R F L P analysis and requires only small quantities of D N A (5 nanograms or less per reaction). For these reasons, the microsatellite analysis procedure has become the major tool for the majority of current L O H studies. Microsatellites contain runs of short and tandemly iterated sequences of di, tri, or tetranucleotides, such as G T G T G T . . . or G T A G T A G T A . . . or G T A C G T A C G T A . . . . These short repetitive D N A sequences are called microsatellites. The number of such tandem repeats is found to be highly polymorphic in the population, with each individual typically containing a different number o f copies (generally 4 to 40) o f the repeat at each particular locus. In addition, they are well interspersed throughout the human genome (e.g., estimated every 30-60 K B for C A repeats) and are highly conserved through successive generations (Ah-See et al 1994). Testing of highly polymorphic microsatellite markers from a specific chromosomal region allows rapid assessment of allelic loss by comparing the alleles in tumor D N A to normal D N A . Therefore microsatellites are good way to research the TSGs either close to or within these chromosome spots. Loss of heterozygosity suggests that a putative tumor suppressor gene nearby may be also lost. 12 1.7. L O H in oral and head & neck malignant lesions Recent studies including those from this lab have shown that the loss of specific regions of chromosomes that contain tumor suppressor genes is a common event in oral SCCs . In this thesis, microsatellite markers on chromosome arms 3p, 4q, 8p, 9p, l l q , 13q and 17p were used, since they have been reported to lie within regions most frequently lost in oral SCCs . Each of these regions w i l l be discussed briefly. Chromosome 3: High frequency of L O H at chromosome 3p has been reported i n head and neck cancers (Table 1 in the appendix). The losses appear to center around 3p 13-21.1, 3p21.3-23, and 3p24-25 (Partridge et al, 1999b, 1998; Scully et al, 1996; Partridge et al,\996; Roz et al, 1996; Maestro et al, 1993). The number of regions showing allele loss at 3p (3p 12.1-14.2, 21.3-22.1 and 24-26) is consistent with the progressive accumulation of genetic errors during the development o f oral S C C (Partridge et al, 1996). Each of the three regions is presumed to contain at least one putative T S G . Within the region of 3pl4.2 exists one of the most common fragile site locus, called F R A 3 B , in the human genome. Fragile sites are portions of chromosomes that are extremely weak and break easily. Consequently, these weak areas may be easy targets for carcinogens such as those found in tobacco. The gene, FHIT (Fragile histidine triad), was recently identified at this fragile site and appears to be involved in various cancers such as esophageal, gastric, colonic, breast, cervical, small cell lung, and head and neck carcinomas (Pennisi et al, 1996; Sozzi et al, 1996; Ohta et al, 1996; Wi lke et al, 1996; Mao et al 1996a and W u et al ,1994). It encodes a protein with 69% similarity to a Schizosaccharomyces pombe enzyme, diadenosine 5', 5 " ' - P l , P4-tetraphosphate (Ap4A) asymmetrical hydrolase which cleaves the A P 4 A 13 substrate into 5' - A D P and A M P . Current theories suggest that diadenosine tetraphosphate may accumulate in the cells in the absence of the normal expression of the gene and may eventually lead to D N A synthesis and cell replication (Mao etal, 1996). Several recent studies have shown that FHIT may be significantly involved in oral S C C development (Croce et al, 1999; Mao et al, 1998a and 1998b ) and suggest that alteration to this gene may play an important role in early stage in the development of this cancer (Mao et al, 1996b). It was recently suggested in some tissues and organs, particularly those associated with exposure to environmental carcinogens, alterations in FHIT occur quite early in the development of human cancer (Croce et al, 1999). Croce concluded that FHIT loss in bronchial tissue indicates the occurrence of genetic alterations associated with the early steps of carcinogenesis. L O H at 3p l4 has been shown to be involved in oral premalignant lesions (Mao et al, 1996a). Unt i l now there is sufficient evidence for only one gene, FHIT, to be responsible for the L O H at the region 3pl4.3, although the evidence in support of it being a T S G is still considered to be controversial (Mao et al, 1998b). TSGs that are responsible for L O H at the other two regions (3p24-pter, and 3p21.3) are still not clear. For example, the region o f 3p24-25 contains the V H L gene, which is thought to be a member o f a novel class of glycan-anchored membrane proteins that function in signal transduction and cell adhesion (Waber et al, 1996), and its alteration has been reported especially in VHL-associated cancers (van den Berg, et al, 1997; K o k et al, 1997 and Decker et al, 1997). Uzawa also mentioned the possibility that the V H L gene may in involved in oral S C C development (1998). However, mutations o f the V H L gene could not be identified and the V L H gene was not inactivated by hypermethylation in H N S C C . Hypermethylation is an alternation method of inactivity of a gene that does not require direct mutation to the gene. It is possible that allelic loss of 14 chromosome arm 3p in H N S C C involves regions surrounding the V H L locus but does not include the V H L gene. Another T S G in H N S C C may exist in the regions surrounding D3S 1110 at 3p 25 (Uzawa et al, 1998; Waber et al, 1996). Chromosome 9: L O H on 9p is by far the most commonly reported chromosomal defect in head and neck cancers, with L O H reported in 72% of malignant lesions. The most commonly affected region is chromosome 9p21-22. In addition, L O H at 9p22-q23.3 is also common (> 70% of head and neck cancers) (Scully et al 1996 and Nawroz et al, 1994). The putative TSGs are near the interferon locus and are not clearly identified. A t 9p21, the prime T S G candidate involved in the head and neck cancers is pi6 (also know as MTS-1 for major tumor-suppressor 1, INK4a for inhibitor of cyclin-dependent kinase4a, and CDKN2A for cyclin-dependent kinase inhibitor 2A) . The T S G pi6 (INK4A/MTS-1/CDKN2A) encodes a cell cycle protein that inhibits cyclin-dependent kinases ( C D K ) 4 and 6, preventing phosphorylation of Rb protein and consequently inhibiting the cell cycle transition of the G l - S phase (Reed et al, 1996). The major biochemical effect of p i 6 is to halt cell-cycle progression at the G l / S boundary. The loss of p i 6 function may lead to cancer progression by allowing unregulated cellular proliferation (Wil l iam et al, 1998). Although mutations of this gene are not apparently frequent for oral cancer, this might suggest that either this gene is inactivated by an alternative mechanism such as homozygous deletion or by methylation of the 5 'CpG-r ich region, which results in a complete blook of gene transcription (Papadimitrakopoulou et al, 1997; Rawnsley et al, 1997; Matsuda et al, 1996 and Merlo et al, 1995). Reed and Papadimitrakopoulou found that - 8 0 % of the head and neck cancers and premalignant lesions were p i 6 inactivated at the protein and/or D N A level and suggest that inactivation of p i 6 may play an important role in early head and neck cancer development (Papadimitrakopoulou et al, 1997 and Reed et al, 1996). Alternatively, another 15 tumor suppressor gene may exist in this region (Waber et al, 1997; Dawson et al, 1996; Reed etal, 1996). Chromosome 17: L O H on 17p has been reported in 50% of head and neck cancers, most frequently involving 17pl3 and 17p 11.1-12 (Scully et al, 1996; Field et al, 1996; Adamson et al, 1994; Nawroz et al, 1994). The region 17pl3 harbors the gene p53 (17pl3.1), which has been reported to have the highest frequency (~50%) of mutations in human cancers. Mutation at p53 is also one of the most common events in H N S C C ( Lazarys et al, 1995). Its protein functions as mediator in several activities, including transcription activation, D N A repair, apoptosis, senescence, and G I and G2 cell cycle inhibition. In addition, increasing evidence also shows that a region, defined by the cholinergenic receptor B l ( C H R N B 1 ) locus at 17pl 1.1-12, that is tightly linked to thep53 regions may contain a novel T S G . Chromosome 4: L O H on chromosome 4 has been studied in cancers of many systems and organs including hepatocellular, bladder, ovarian and cervical cancers. The putative tumor suppressor locus was localized to a region near the epidermal growth factor gene on 4q25 and 4q24-26. Loss at 4q25 occurs in 70% of head and neck cancers (Pershouse et al, 1997) and loss at 4q26-28 occurs in 47% (Califano et al, 1996; Bockmihl et al, 1996). The combination of allelic deletions and chromosomal transfer studies strongly suggests the presence of a T S G within 4q24-26. This region was involved in >80% of the tumors examined, suggesting that a putative chromosome 4q T S G may play an important role in the evolution of H N S C C (Pershouse et al, 1997). 16 Chromosome 8: Investigation of 8p regions in head and neck squamous carcinoma has shown a relatively high incidence of alterations (31%-67%) (Wu et al, 1997; Califano et al, 1996; Bockmuhl et al, 1996; Scholnick et al, 1996; EI-Naggar et al, 1995; Field et al, 1995; Ah-aee et al, 1994; L i et al, 1994). Deletion mapping of oral and oropharyngeal S C C defines three discrete areas on chromosome arm 8p: 8p23, 8p22, and 8pl2-p21 (Wu et al, 1997 and EI-Naggar et al, 1995). Several studies have linked allelic loss at 8p to a higher stage (Wu et al, 1997) and poor prognosis (Scholnick etal, 1996 and L i et al, 1994). Chromosome 11: L O H on human chromosome 11 has also been commonly reported in a variety of human cancers, including H N S C C (39%-61%) (Lazar et al, 1998; Venugopalam et al, 1998; Bockmuhl et al, 1996; Califano et al, 1996; Uzawa et al, 1996; El-Nagger et al, 1995; Nawroz et al, 1994). The common region of loss at this chromosome seems to be near the INT-2 locus at 1 l q l 3 (Nawroz et al, 1994). It is possible that some of this region's allelic imbalance may be due to amplification rather than L O H (Nawroz et al, 1994). Amplification of this region associated with poor prognosis was also reported (Papadimitrakopoulou et al, 1997). Chromosome 13: More than half of H N S C C s shows L O H of 13q in regions close to the R B (retinoblastoma) locus, but not R B gene (52-67%) (Ogawara et al, 1998; Ishwad et al, 1996; Maestro et al, 1996; Bockmuhl et al, 1996; Califano et al, 1996; Nawroz et al, 1994). A hot spot of D13sl33 at 13ql4.3, which lies just telomeric to the R B gene, was reported (Yoo et al, 1994). A recent study done by Ogawara showed L O H on 13ql4.3 correlated with lymph node metastasis of oral cancer (p<0.0024). His results also suggest that L O H on 13q is a common event in oncogensis and/or progression of oral S C C and the existence of a new 17 suppressor gene near D13S273-D13S176 loci which may play a role in these events since no significant variation in R B protein expression was detected (Ogawara et al, 1998). The study of Harada et al, (1999) confirmed that L O H in chromosome 13 showed a significant correlation with lymph node metastasis in esophageal squamous cell carcinoma, included in H N S C C . They reported that unidentified TSG(s) in region 13ql2-13 might be involved. 1.8. L O H in oral and head & neck premalignant lesions Since tumorigenesis is a sequential accumulation of genetic alterations, analysis of early and late stage lesions may define the genetic changes associated with the development and progression of H N S C C . Few studies (Table 2) have investigated the premalignant stages of the lesions while there are many studies of L O H on oral S C C . The main difficulties lie in the fact that: 1) premalignant lesions are small and therefore it is extremely hard to obtain sufficient amount of D N A for molecular analysis, 2) big hospitals or research centers typically have better access to cancers rather than premalignant lesions, and 3) it is much harder to microdissect premalignant lesions compared to carcinomas. The limited number of studies on premalignant lesions either used only a small number of cases or primers, or did not correlate L O H with degree o f dysplasia and mostly form high-grade dysplasias. Nonetheless, results from these studies clearly show that L O H is a frequent event in premalignant lesions (Califano et al, 1996; Mao et al, 1996a; Emi l ion et al, 1996; Roz et al, 1996; El-Naggar et al, 1995). For example, a similar frequency of L O H at 9p was reported in preinvasive lesions (71%) as in S C C (72%) (Van der Riet et al, 1994). This suggests that loss of 9p is an early event in the progression of oral cancer (Papadimitrakopoulou et al, 1997 and V a n der Riet et al, 1994). Similarly L O H at 3p have been found to occur very early during oral carcinogenesis and in a significant number of oral 18 mild dysplasia or even hyperplasias (Zhang et al, 1997). On the other hand, data from this lab showed that L O H at 17p was not found in reactive hyperplastic lesions and mild dysplasia of oral mucosa, indicating loss at 17p occurs later than L O H at 3p and 9p (Zhang et al, 1997). EI-Naggar and his colleges (1998) recently found L O H at 8p in 27% of the dysplastic lesions and in 67% of the invasive oral and laryngeal SCCs . The highest frequency of allele losses in dysplasia and cancer were detected in the same loci: 8p21 and 8p22. In addition, allelic losses in both dysplastic and corresponding invasive specimens were noted at the same loci , suggesting their emergence from a common preneoplastic clone. The studies suggested that inactivation of TSG(s) within these loci may constitute an early event in the evolution of oral and laryngeal SCCs . Moreover, a study by Mao and co-workers (Mao et al, 1996a) showed that L O H in oral premalignant lesions could be used to predict risk o f cancer progression of these premalignant lesions. They reported that the presence of L O H at 9p21 &/or 3p l4 in oral leukoplakia was associated with a greater probability of progression of this premalignant lesion into S C C : 7 of 19 (37%) cases with such L O H progressed to S C C in their study, as compared to only 1 of 18 (6%) cases without L O H . 19 Table 2. L O H and oral premalignant lesions Authors Chromosome arm studied Degree of oral dysplasias L O H and risk of malignant transformation VanderRiet, 1994 9p21-22 Severe dys/ CIS lesions Not evaluated EI-Naggar, 1995 3p, 5q,8p, 9, 11 q, 17 Severe dysplasia adjacent to cancer LOH at 8p loci has higher risk of tumor's aggressive feature Califano, 1996 3p, 4q, 6p, 8, 9p, 1 lq, 13q, 14q, 17p Dysplasia with no indication of severity Not evaluated Mao, 1996a 3pl4, 9p21 Reported as leukoplakias -some of which were dysplastic Patients with LOH at 9p and/or 3p have higher risk of cancer development Mao, 1996 5q (APC) Dysplasia: 5 cases CIS: 3 cases Not evaluated Emilion, 1996 3p, 17p(p53), 18q(DCC) Primary dysplasias: Mild: 8 cases Moderate: 4 cases Severe: 5 cases Dysplasia adjacent to cancer: Mild: 4 cases Moderate: 6 cases Severe: 3 cases LOH was not associated with the degree of dysplasia, but the number of allelic loss increased while tumor developed. Roz, 1996 3p Severe/CIS lesions only Not evaluated Zhang, 1997 3p, 9p, 17p Reactive lesions: 29 cases Mild dysplasias: 10 cases Moderate dysplasias: 11 cases Severe dys/CIS: 16 cases Cancer: 22 cases Loss of more than one chromosome arm is associated with degree of *"""• 20 Eight paired severe Both lesions El-Naggar, 1998 8p dysplasia and emergent from a corresponding invasive common lesions preneoplastic clone 3p21 31 Dysplasia with no The probability of Partridge, 1998 8p21-23 indication of severity progressing to 9p21 SCC was much 13ql4.2 (Rb) greater for cases 17p 11.2 (tp53) showing AI at two 18q21.1 (DCC) or more relevant loci 1.9. A molecular progressing model for oral S C C In the late 80s, Fearon and Volgelstein, among the first people to describe molecular progression, suggested that a) tumors progress via the activation of oncogenes and the inactivation of TSGs, each generating a growth advantage for a clonal population of cells; b) specific genetic events generally occur in a distinct order of progression; but c) the order of progression is not necessarily the same for each individual tumor, and therefore it is the accumulation of genetic events that determines tumor progression. It has been estimated that at least 6-10 independent genetic events are required in order for head and neck cancers to occur (Emilion et al 1996). It is now accepted that the histologic progression of oral cancer (from hyperplasia mild dysplasia -> moderate dysplasia -> severe dysplasia -> CIS -> SCC) is underlain by the accumulation of such changes to critical genes. In a landmark study by Califano and his colleagues (1996), LOH was investigated in a whole spectrum of oral premalignant lesions including hyperplasia, dysplasia, CIS and SCC. The study proposed a 21 genetic progression model for oral carcinogenesis (Fig. 3). The model proposes that LOH at 9p is the earliest event associated with transition from normal to benign hyperplasia; LOH at 3p and 17p is associated with dysplasia, whereas CIS and SCC were characterized by additional deletions on 4q, 6p, 8, 1 lq, 13q, and 14q (Fig 3). Figure 3 Tumor Progression Model for Head and Neck cancers as proposed by Califano etal 1996 Alternate Precursor Lesion Normal Mucosa 9p loss _ J — \ 3p,17p \ loss i Dysplasia 1 lq, 13q 14q loss N Carcinoma in situ 6p, 8, 4q loss Invasive cancer Benign Squamous Hyperplasia est 22 2. S T A T E M E N T O F T H E P R O B L E M S Oral S C C is believed to progress through stages o f oral premalignant lesions (hyperplasia, mild dysplasia, moderate dysplasia, and severe dysplasia, CIS) before finally become invasive. While the prognosis of oral premalignant lesions are excellent, once invasion occurs and oral S C C is formed the prognosis is poor and about half of the patients die within 5 years of diagnosis despite recent advancement in treatment. Those who survive still face severe cosmetic and functional morbidity. The understanding and intervention of oral premalignant lesions w i l l be critical in the reduction of the mortality and morbidity of oral S C C . However, the majority of oral premalignant lesions do not progress into oral S C C and aggressive treatment of these lesions is not justified. A t present, we can not identify those progressing oral low-grade premalignant lesions from the majority o f non-progressing low-grade oral premalignant lesions. Since cancer is underlain by changes to the critical control genes, the understanding of molecular changes during early oral cancer development may be critical in the establishment of molecular markers for the identification of high-risk oral premalignant lesions. It is l ikely that the progressing low-grade oral premalignant lesions are molecularly different from the morphologically similar non-progressing low-grade premalignant lesions. Wi th the rapid development of molecular biology techniques, there have been numerous studies on the molecular changes in human cancers, including oral SCCs . However, few studies are done in oral premalignant lesions because these lesions are not readily available and are small in size. A number of studies done on premalignant lesions are all limited either in the number of cases used or the number of probes used. Furthermore, most of these studies were either done on high-grade oral premalignant lesions or the degree of dysplasia was not mentioned. There is a marked lack of information on the molecular 23 changes of low-grade oral premalignant lesions, which are the majority of oral premalignant lesions, and are the hardest to predict in terms of their malignant risk. A molecular model has been proposed by Califano et al, (1995) based on studies on oral premalignant and malignant lesions. However, the study merged all dysplasias together. It is well known that the prognosis of low-grade oral premalignant lesions (e.g. mi ld dysplasia) differs drastically from that of high-grade oral premalignant lesions (e.g., severe dysplasia). Furthermore, there is no data available on the characteristics of molecular changes of those oral premalignant lesions that have known to have progress into cancer. The main problem is that such lesions are hard • to find. This thesis w i l l investigate the molecular changes of low-grade oral premalignant lesions, and compare the characteristics of genetic changes of those premalignant lesions that have progressed into CIS or S C C with those with no known progressing history. 24 O B J E C T I V E S To characterize the pattern of genetic changes in premalignant lesions by means of L O H analysis using microsatellite markers for the 7 chromosomal regions known to be frequently lost in oral tumors: 3p, 4q, 8p, 9p, 1 l q , 13q and 17p. To determine whether the L O H profile was significantly altered in cases that have progressed to CIS or S C C as compared to those that have not. 25 4. H Y P O T H E S I S 1) Molecular changes as assayed by the L O H analysis occur early during oral carcinogenesis, including low-grade oral premalignant lesions. 2) Progressing early lesions (low-grade dysplastic and hyperplasia without dysplasia) are genetically different from morphologically similar non-progressing lesions. If the data support the first hypothesis, it would suggest that loss of regions of chromosomes that contain presumptive tumor suppressor genes is critical for early oral carcinogenesis. If the data support the second hypothesis, it would suggest that early lesions progressing to cancer, though morphologically indistinguishable from those without a history of progression, are in fact different genetically. It is this genetic difference that underlines the behavior of these lesions. It would also support the belief that the molecular tools such as microsatellite analysis for L O H may be used in differentiating between progressing and non-progressing lesions. The identification of molecular changes that can be used to predict the l ikely behavior of low-grade lesions would allow the clinician to identify which patients with low-grade lesions should be managed more aggressively, either by more frequent screening or by early treatment, using traditional approaches such as surgery, or newer techniques such as chemopreventive regimes. 26 5. E X P E R I M E N T A L D E S I G N A N D S T A T I S T I C A L A N A L Y S I S 5.1. Exper imental design This is a case — control study (Chap, 1997, p2 and K i n g , 1996, p38). The study subjects were those who already had a certain condition (premalignant lesions known to have progressed to CIS or S C C ) : These subjects are the cases and this thesis investigated the characteristics shared by the cases. The comparison group (those lesions without progression history) — the controls — was selected so that it resembled the cases as closely as possible (Brunette, 1996). We tried to avoid the usual problems usually in a case control design: information bias or data collected in the past under uncertain conditions. First, there was no information bias in our study because there were no significant differences in quality or availability of data between the cases and controls. Second, the experimental data ( L O H data) were collected under the same experimental conditions. Although some data were collected in the past under uncertain conditions, like gender, age, site and smoking history, there was no significant difference between cases and controls in terms of gender distribution, age distribution, site distribution, and smoking history ( Table 9). These matched factors can no longer be evaluated as etiologic agents since they w i l l be equalized in the cases and the controls (Brunette, 1996). Moreover, the histological diagnosis of hyperplasia or mi ld or moderate dysplasia was reconfirmed by two pathologists ( L Z and RP) using criteria established by the Wor ld Health Organization ( W H O collaborating Reference centre 1978). This study was a retrospective design. We obtained some clinical information by tracking recurrent lesions as they appeared in our database and by following the case histories of treatment in hospital charts. However, there was no evidence that treatment was less 27 aggressive among progressing cases compared with those without a history of progression (see details in discussion part). 5.2. Statistical analysis Associations between L O H and progression were examined using Fisher's exact test (two-tailed). Clinical differences between progressing and non-progressing groups were examined using either Fisher's exact test (gender distribution and smoking habit) or unpaired t-test (age and follow up time). A l l p values were two-sided. A P value of 0.05 or less was considered significant. Time-to-progression curves were estimated by the Kaplan-Meier method and comparisons were performed using log-rank test (Chap, 1997 and Armitage et al, 1987). Relative risks were determined using Cox regression analysis (Chap, 1997 and Armitage et al, 1987). 28 6. M A T E R I A L S A N D M E T H O D S 6.1. Sample collection. This study used paraffin-embedded archival samples from the provincial Oral Biopsy Service of British Columbia. This centralized Oral Biopsy Service provides service to dentists and E N T surgeons throughout the province, at no cost to the provider or patient, with more than 3,500 biopsies of oral lesions received per year (19 years archived). This provides a large collection of early lesions that can be followed over time. Cases that progressed into cancer were identified by linking the database o f this Service to the Brit ish Columbia Cancer Registry, which receives notification of all histologically confirmed cases of cancer and CIS diagnosed in the Province. 6.2. Sample sets Two sample (archival paraffin blocks) sets were used (Table 3): Set 1: Oral lesions from patients with no subsequent history of head and neck cancer. We refer to these cases as non-progressing. Set 2: Oral lesions from patients that later progressed to CIS or S C C . Both sets of samples included hyperplasia (without dysplasia) group and low-grade dysplasia (mild or moderate dysplasia) group. 29 Table 3. Samples sets and groups Number of Cases Lesion Type Epi the l ia l Hyperplas ia (without dysplasia) M i l d Dysplasia (low-grade dysplasia) Moderate Dysplasia (low-grade dysplasia) Lesions with no history of head and neck cancer Lesions that later progressed to CIS or SCC 33 31 _ 23 14 The criteria for choosing Set 1 samples included: 1) A histological diagnosis of a case confirmed by two pathologists using criteria established by the World Health Organization (WHO collaborating Reference centre 1978). 2) The provision that the sample was large enough to yield sufficient DNA from both the epithelium and from the connective tissue for multiple LOH analyses. 3) Confirmation that these patients had no prior history of head and neck cancer and that they had not developed such cancer so far was obtained both from hospital records and by using a computer linkage with the British Columbia Cancer Registry. Al l but three of these cases had at least 3 years of follow up time. The inclusion criteria for selection of the set 2 lesions in addition to those described for set 1 included: 1) Both the primary hyperplastic or dysplastic lesions and their matching CIS or SCC had to be from the same anatomical site as recorded on pathology reports and patients charts; 30 2) The interval between the primary lesions and later CIS or S C C had to be longer than 6 months. This criterion was used to exclude cases that might be due to inadequate biopsy, with the time interval chosen for exclusion being arbitrary. 6.3. Histological diagnostic criteria for the samples The generally accepted histological diagnostic criteria for dysplasia includes loss of basal cell polarity, more than 1 layer of basaloid cells, increased nuclear to cytoplasmic ratio, drop-shaped rete ridges, irregular stratification, increased and/or abnormal numbers of mitosis in the basal compartment as well as increased mitotic figures in the superficial half of the epithelium, cellular pleomorphism, nuclear hyperchromatism, enlarged nucleoli, reduction of cellular cohesion, and keratinization of single cells or cell groups in the spinous cell layer ( W H O collaborating Reference centre 1978). Based on the above-mentioned criteria as well as the extent of the epithelial tissue that is dysplastic, dysplasias are divided into mild, moderate, and severe forms. Those classified as mild indicate that the dysplastic cells are confined to the basal layer and that they exhibit the smallest degree o f changes. Wi th moderate and severe dysplasias, the epithelial layers involved and the severity of the cellular changes increase progressively. In carcinoma in situ, the dysplastic cells occupy the entire thickness of the epithelium (from bottom to top), although the basement membrane is still intact. Invasion o f dysplastic cells through the basement membrane into the underlying connective tissue and the dissemination of these cells to other sites through the circulatory systems are characteristics of malignancy (i.e., cancer) (Zhangetal, 1997) (Fig 1). 31 The histological diagnoses of the lesions were performed independently by Dr. R. Priddy and Dr. L . Zhang, oral pathologists at University o f Bri t ish Columbia. Only those cases in which the two pathologists agreed on the diagnosis were used for the study. 6.4. Slide preparation A case was chosen after the pathologists agreed on the diagnosis, and after the determination that the block was big enough and contained sufficient epithelial and connective tissues for molecular analysis. The tissue block for the case was removed from the archive, one 5 micron thick slide was cut and stained with H & E (hematoxylin and eosin) and coverslipped for reference. The actual samples for microdissection were 12 microns thick and approximately 15 slides, each with multiple thick sections, were cut for each tissue block. They were also stained with H & E but left uncoverslipped. The H & E procedure for the slides is described below: Baked at 37°C in an oven overnight, then at 60-65C 0 for 1 hour, and left at room temperature to cool; Deparaffmized by two changes o f xylene for 15 min each; Dehydrated in graded ethanol (100%, 95% and 70%), 5 min each; Hydrated by rinsing in tap water, 2 min; Placed in G i l l ' s Hematoxylin for 5 min; Rinsed in tap water, 2 min; Blued with 1.5 % (w/v) sodium bicarbonate, 1 min; Rinsed in water, 2 min; Counterstained with eosin for 10 sec; 32 Dehydrated in graded alchohol (75%, 95% and 100%), 3 min each; Cleared in xylene, 3 min twice for coverslipping (for the H & E slide) or submitted for microdissection. 6.5. Microdissection Microdissection of the specimens was either performed or supervised by Dr . L . Zhang. Areas o f dysplasia were identified microscopically. Epithelial cells in these areas were then meticulously dissected from adjacent non-epithelium tissue under an inverted microscope using a 23-G needle. The non-epithelial underlying stroma dissected out from the same tissue block was used as controls (Zhang et al, 1997). 6.6. Sample digestion and DNA extraction The microdissected tissue was placed in a 1.5 ml eppendorf tube and digested in 300 ul o f 50 m M T r i s - H C L (pH 8.0) containing 1% sodium dodecyl sulfate (SDS) proteinase K (0.5 mg/ml) at 48C° for 72 or more hours. During incubation, samples were spiked with 10 or 20 ul of fresh proteinase K (20 mg/ml) twice daily. The D N A was then extracted 2 times with PC-9 , a phenol-chloroform mixture, precipitated with 70% ethanol in the presence of glycogen, and washed with 70% ethanol. The samples were then re-suspended in L O T E , a low ionic strength Tris buffer, and submitted for D N A quantitation (Zhang et al, 1997). 33 6.7. DNA quantitation The fluorescence analysis using the Picogreen kit (Molecular Probes) was used to quantitate D N A . Absorbance was read from an S L M 4800C spectrofluorometer ( S L M Instruments Inc. Urbana, IL) . The sample D N A was then determined from the standard curves. A series of dilutions were done subsequently to adjust the concentration of D N A to five ng/pl with L O T E buffer (Zhang et al, 1997). 6.8. Primer-extension preamplification (PEP) If the concentration of D N A was less than 100 ng total, a procedure called P E P was first done. P E P involves amplification of multiple sites of the genome using random primers and low stringency conditions and hence increases the amount of total D N A for microsatellite analysis. It was carried out in a 60 pi reaction volume containing 20 ng o f the D N A sample, 900 m M of T r i s - H C L of p H 8.3, 2 m M of dNTP where N is A , C , G and T, 400 n M of random 15-mers (Operon Techologies, SP 182-2, Poly N , 15 mer), and 1 pi of Taq D N A polymerase (GibcoBRL, 5U/pl) . Two drops of mineral oi l were added prior to the reaction. P E P using the automated thermal cycler (Omigene H B T R 3 C M , Hybaid Ltd) involved 1 cycle of pre-heat at 95 C° for 2 min, 50 cycles of 1) denaturation at 92 C° for 1 min, 2) annealing at 37 C° for 2 min, and 3), ramping from 37 C° to 55 C° at 10 sec/degree, polymerization at 55 C° for 4min. 34 6.9. Coding samples A l l samples were coded in such a way that the analysis of L O H would be performed without the knowledge of the sample diagnosis. 6.10. End-Labeling One more step before P C R was end-labeling of one of the primer pair. The reaction contained a 50ul mixture of P C R water 38ul, 10 x buffer for T4 polynucleotide kinase 5ul, 10 x B S A l u l , one o f the primer pair l u l , T4 polynucleotide kinase 3ul, and [y- 3 2P] A T P (20 u C i , Amersham) 2 ul . It was then run for 1 cycle at 37 C° for 60 min using the thermal cycler (Zhang etal, 1997). 6.11. PCR amplification Microsatellite L O H analysis in this study was done on chromosome arms 3p, 4q, 8p, 9p, l l q , 13q and 17p. The pairs of P end-labeled polymorphic probes (primer pairs, chromosome markers, Research Genetics - Huntsville, A L ) that flank the area of tandem repeats in the chromosomal region of interest mapped to the following regions: 3pl4.2 (D3S1234, D3S1228, D3S1300); 4q26 (FABP2); 4q31.1 (D4S243); 8p21.3 (D8S261); 8p23.3 (D8S262); 8p23.3 (D8S264); 9p21-22 (IFNA, D9S171, D9S736, D9S1748, D9S1751); l l q l 3 . 3 (INT2); l lq22 .3 (D11S1778); 13q32 (D13S170); 13ql4.3 (D13S133); 17p l l . 2 (CHRNB1) and 17p 13.1 {tp53 and D17S786). These markers are localized to regions 35 previously shown to be frequently lost in head and neck tumors (Lazar et al, 1998; EI-Naggar et al, 1998; Zhang et al, 1997; Maestro et al, 1996; Uzawa et al, 1996; Califano et al, 1996; Mao et al, 1996a; Emi l ion et al, 1996; Roz et al, 1996; EI-Naggar et al, 1995; Ah-See et al, 1994; Nawroz et al, 1994; Adamson et al, 1994; W u et al, 1994). Some of the markers, like the regions 3p l4 (instead of 3p21.3-23 and 3p24-25) and 9p21 (instead of 9p22-23), are preferentially chosen for analysis, because high frequencies o f L O H at these regions have not only been reported in head and neck cancers but also have been shown to be associated with the risk of malignant transformation of oral premalignant lesions (Mao et al, 1996a). P C R amplification using the thermal cycler was carried out in a 5 pi reaction volume containing 5 ng of genomic D N A , 1 ng of labeled primer, 10 ng of each unlabeled primer, 1.5 m M each of d A T P , dGTP, dCTP, and dTTP, 0.5 units of Taq D N A polymerase ( G I B C O , B R L ) , P C R buffer [16.6 m M ammonium sulfate, 67 m M Tris (pH8.8), 6.7 m M magnesium chloride, 10 m M (mercaptoethanol, 6.7 m M E D T A , and 0.9% dimethyl sulfoxide], and 2 drops of mineral o i l . Amplification involved 1 cycle o f pre-heat at 95C° for 2 min; 40 cycles of 1) denaturation at 95 C° for 30 sec, 2) annealing at 50-60 C° (depending on the primer used) for 60s, and 3) polymerization at 70 C° for 1 min; and 1 cycle of final polymerization at 70 C° for 5 min. The P C R products were then diluted 1:2 in loading buffer and separated on an 7% urea-formamide-polyacrylamide gels, and visualized by autoradiography. The films were then coded and scored for L O H (Zhang et al, 1997). 6.12. Scoring of L O H For each case, two samples, the epithelial cells in the lesion and the connective tissue underneath (which serves as a control), were subjected to amplification at the same time. A n 36 informative case wi l l yield two alleles, one maternal and the other paternal, with electrophoretic migration dependent on the different sizes. Samples were scored by comparison of the intensity of the autoradiographic bands (which represent the P C R product) of the lesion with that of the normal connective tissue control. Al le l i c imbalance detected as loss or marked reduction (>50%) of one of these allelic bands is termed L O H . A case is non-informative i f both paternal and maternal alleles are o f the same size. In this situation, an alternative information probe was replaced. A l l samples showing L O H would be subjected to repeat analysis after a second independent amplification and re-scored whenever the quantity of D N A was sufficient. 37 7. R E S U L T S 7.1. Frequency of allelic loss LOH was present in 68 of 116 (59%) pre-malignant lesions studied, occurring more frequently among dysplastic (55 of 77, 71%) than hyperplastic (13 of 39, 33%) lesions (P < 0.001). LOH frequencies were dramatically elevated in lesions that later progressed to cancer (Table 4). All progressing lesions (both hyperplastic and dysplastic) showed LOH at one or more of the 19 microsatellite loci tested. LOH was detected in only 21% of the non-progressing hyperplasias and 59% of dysplasias. Multiple chromosomal arm loss was characteristic of progressing lesions (50% of hyperplasia and 91% of dysplasia, see Table 4). It was absent in non-progressing hyperplasia and occurred in only 31% of the non-progressing dysplasias. Table 4. Accumula t ion of allelic loss in progressing and non-progressing lesions • Hyperplas ia Low-grade dysplasia Without With Without With progression progression P progression progression P # of lesions 33 6 54 23 # with L O H a 7(21) 6 (100) 0.001 32 (59) 23 (100) 0.0001 >1 a r m lost 0 3 (50) 0.002 17(31) 21(91) <0.0001 >2 arms lost ~0~ Y(50y 0.002 11 (20) 13 (57) 0.003 a A total of 7 chromosomal loci were tested. Values in parentheses are percentages 38 7.2. Pattern of allelic loss The most common losses for both progressing and non-progressing cases were on 3p and 9p with a higher frequency in the progressing cases (Table 5). Among non-progressing cases, 4 of 30 (13%) hyperplasias and 13 of 53 (25%) dysplasias had a loss at 3p. L O H at 9p was rare in non-progressing hyperplasias (3% of cases), but present in 24 o f 52 (46%) dysplasias. In contrast, 67% of the progressing hyperplasias and 64% dysplasias showed L O H at 3p; 50% of the progressing hyperplasias and 83% dysplasias showed L O H at 9p. The frequency o f loss at other arms (4q, 8p, 1 l q , 13q, 17p) was low for non-progressing cases. Only 2 of 33 hyperplasias (6%) had loss on any of these arms (1 at 1 l q , 1 at 13q). Nineteen of 54 (35%) o f non-progressing dysplasias had loss on these arms, most frequently at 17p (20% of cases) and 8p (15%) followed by 1 l q (12%), 4q (8%), and 13q (4%). Further increases in L O H frequencies at 4q, 8p, 1 l q , 13q and 17p occurred in lesions that progressed to tumors. For dysplasias, this increase was significant for 8p, 1 l q and 13q, with the increase for 4q approaching significance (P = 0.057). There was also a doubling in the frequency of L O H on 17p (from 20% to 41% of cases), although this increase was not significant (P = 0.087). For hyperplasias, increases were significant in comparisons of progressing versus non-progressing lesions for 4q, 8p, and 17p, with 1 l q approaching significance (P = 0.062). 39 Table 5 Patterns of allelic loss i n progressing and non-progressing lesions Hyperplas ia Low-grade dysplasia Without With Without With Arm progression progression P progression progression P 3p 4/30 (13)a 4/6 (67) 0.014 b 13/53 (25) 14/22 (64) 0.003 9p 1/32 (3) 3/6(50) -j- j j j j j j— 24/52 (46) 19/23 (83) 0.005 4q 0/31 (0) 2/6 (33) 0.023 4/48 (8) 6/21 (29) 0.057 8p 0/31 (0) 2/6(33) ' " 0.023 8/51 (15) 11/21 (52) 0.003 l l q 1/31 (3) 3/6 (33) 0.062 6/52(12) 9/23 (39) 0.011 13q 1/32 (3) 0/4 (0) 1 2/53 (4) 7/21 (33) 0.002 17P 0/33 (0) 2/6 (33) 0.020 11/54 (20) 9/22 (41) 0.087 3p & / o r 9p 5/29(17) 6/6 (100) 0.004 30/54 (56) 22/23 (96) 0.004 3p & / o r 9p plus any other a r m 0/32 (0) 3/6 (50) 0.002 15/54 (28) 18/23 (78) 0.0001 a Loss/informative cases (% loss) b Bold means p< 0.05, was considered significant. 40 7.3. Progression r isk Specific L O H patterns in premalignant lesions were examined for association with disease progression by using the Kaplan-Meier method (Armitage et al, 1987). Almost all the progressing lesions (28 of 29, 97%) had L O H on 3p &/or 9p. Time-to-progression curves (see table 6 in appendix of time between initial biopsy and current date or the date on which a lesion has progressed to S C C , with time of first biopsy set at 0) were plotted as a function of L O H at 9p (Fig. 4A) , 3p (Fig. 4B) , or a combination of 3p &/or 9p (Fig. 4C). A l l were significant. A further comparison was made o f cases with loss on these 2 arms in the presence and absence of L O H at any of the other 5 chromosomes (4q, 8p, l l q , 13q and 17p; Fig.4D). A significant difference was again observed. Finally, we separately compared time-to-progression for cases in which 3p &/or 9p L O H was restricted to these 2 arms alone with additional losses on each of the chromosome arms (Fig.4E — 41). Significant P values were observed for combinations that included 8p, 1 l q or 13q. Additional analyses included an assessment of relative risk of progression for each of the L O H patterns presented and a determination of the proportion of cases without progression at 5 years follow-up. These results are tabulated in Table 7 with an assessment of their significance respectively. 41 Figure 4 . Probability of having no progression to cancer, according to L O H pattern Is dl § s ' d p a n d 9 p H e t 3 p ' & / o r 9 p I ^ O H o n l y 3 p & / b r 9 p L d H p l u s o t h e r s p _ v a l u e <.Q1 p _ v a l u e = . 0 9 10 12 14 16 s & 8 3 p & / p r 9 p L O H o n l y \ „ 3 p & / o r 9 p L O H ' p l u s l l q p _ y a l u e = . 0 5 « 9 a » j? 0. i 8 B t i i t 3 p H e t ip'LdH1 p v a l u e <.01 Is — . . . . r 3 p and 9 p H e t 3 p &/or 9 p L t i H ' ' p _ v a l u e <.01 10 12 14 16 3 f l & / o r 9 p L O H o n l y 3 p & / o r 9 p L O H p l u s 4 q 6 8 10 12 Years 3 p & / p r 9 p L O H o n l y 3 p & / o r 9 p L O H p l u s 8 p p _ v a l u e = . 0 3 a 10 12 14 Years 10 12 14 16 1 ^ 1 3 p & / o r 9 p L O H . o n l y 3 p & / o r 9 p L O H p l u s 1 3 q p v a l u e <.01 Is 2 § 8 & / o r 9 p L O H o n l y 3 p & / o r 9 p L O H p l u s 1 7 p p _ v a l u e = . 0 8 8 10 12 14 6 8 10 12 14 Years 8 10 12 14 A, progression as a function of LOH at 9p (No LOH = 69; LOH = 47). B, progression as a function of LOH at 3p (No LOH = 80; LOH = 36). C, progression as a function of LOH at 3p &/or 9p (No LOH = 56, LOH = 60). D, progression as a function of LOH at 3p &/or 9p when this loss occurred in the absence or presence of LOH at any other arm (No additional arms lost = 26, LOH on at least 1 of the following arms: 4q, 8p, 1 lq, 13q or 17p = 34). E, F, G, H , and I, progression as a function of LOH at 3p &/or 9p when this loss occurred with no additional arms lost (n = 26) or with LOH at 4q (n = 10), 8p (n = 18), 1 lq (n = 15), 13q (n = 8) or 17p (n = 21) respectively. 42 Table 7. Probability of lesions not progressing to cancer after 5 years follow-up Proportion (%) L O H pattern #of of R R (95% CI) cases non-progressing cases (95% CI) N o L O H 48 100 -9p*': 9p Het 69 93 (87-99) 1.0 9p L O H 47 61 (48-78) 3.97(1.68-9.15) 3p*2: 3p Het 80 85(76-95) 1.0 3p L O H 36 63 (48-83) 3.74(1.76-7.93) 3p &/or 9p* 3: 3p &/or 9p Het a 56 98 (95-100) 1.0 A l l cases with 3p &/or 9p L O H 60 63 (50-77) 24.1 (3.3-176) 3p &/or 9p L O H (but no other arms) 26 74 (57-95) 3.75 (1.32-10.7) 3p &/or 9p L O H (+ L O H at any other arm) 34 53 (38-74) 33.4 (4.48-249) 3p &/or 9p + others*4: 3p &/or 9p L O H (but no other arms) 26 74 (57-95) 1.0 3p &/or 9p plus 4q L O H 10 60 (36-99) 2.3(0.86-6.16) 3p &/or 9p plus 8p L O H 18 52 (32-84) 2.59(1.05-6.37) 3 p & / o r 9 p p l u s l l q L O H 15 50 (29-86) 2.49 (0.98-6.31) 3 p & / o r 9 p p l u s 1 3 q L O H 8 0 b 7.08 (1.93-25.9) 3 p & / o r 9 p p l u s 17p L O H 21 54 (35-83) 2.2 (0.18-5.5) "Includes 8 cases with L O H at other arms bCalculation does not include 2 non-progressing cases that have less than 5 years follow-up *'The RR estimates the risk of 9p L O H group relative to group of 9pHet. *2The RR estimates the risk of 3p L O H group relative to group of 3pHet. *3The RR estimates the risk of each group relative to group of 3p&/or 9p Het *4The RR estimates the risk of each group relative to group of 3p&/or 9p L O H (but no other arm loss). 43 7.4. Compar ison of L O H pattern i n matching premalignant and malignant lesions In the Table 8, a comparison was made of the regions of loss in biopsy pairs of progressing cases. Twenty-five of 29 progressing cases had later CIS/SCC biopsies available for LOH analysis. In 17 cases (68%), allelic losses in the premalignant lesions (upper versus lower allele) were found in the later lesion (see Fig 5. b), although 14 of these cases showed additional losses on other chromosome arms. The remaining 8 cases showed an alteration in the pattern of loss in the tumor for some of the loci that had LOH in the early biopsy. Table 8: Comparisons of L O H patterns i n ini t ial biopsies wi th that seen i n C I S / S C C that later developed at the same anatomical site # of cases Case ID # Cases i n which later biopsy was available for analysis 25 Those with same LOH pattern in both biopsies 3/25(12%) #281 (6 losses), #470 (3 losses); #365 (1 loss) Those with LOH in first biopsy also present in later biopsy, but with additional losses in the later lesions in the following arms: 14/25 (56%) #385, #405, #416, #271, #293, #399, #185, #377, #245, #286, #464, #121,#223, #460 3p 4/14 (29%) #405, #121, #399, #185 9p 1/14 (7%) #416 17p 3/14(21%) #385,#405, #223 4q 5/14 (36%) #385,#271, #293, #245, #286 8/14 (57%) #405, #385, #223, #293,#185,#377,#245, #460 l l q 6/14 (43%) #385, #405, #416, #185, #286, #464 13q 3/14(21%) #245, #377, #121 Different LOH patterns 8/25 (32%) #173, #446, #472, #401, #95, #406, #122, #473 La te r biopsy not available 4 #471, #363, #197, #79 44 Figure 5. LOH analysis of 2 patients (a, b) a Patient 1 D8S264 IFNA D13S170 D17S786 (Bp) (9p) (13p) (17p) • f t f • C D CD CD CD b Patient 2 IFNA D11S1778 tp53 (9p) (11q) (17p) i l l C D T C D T C D T DNA was isolated from control stroma (C), dysplasia (D) or tumor (T) microdissected from lesion biopsies. Microsatellite markers, the chromosomal arm being assayed and patient numbers are indicated above each block, a: A rare mild dysplasia with multiple allelic loss: loss of lower allele at D8S264; upper allele at IFNA; upper allele at D13S170; and lower allele at D17S786. b: Patient with a mild dysplasia (D) that later progressed to a SCC (T). The mild dysplasia shows the same pattern of multiple allelic loss as the tumor: loss of the lower allele at IFNA and D11S1778; and upper allele at tp53 7.5. Clinical history of the lesions. There was no significant difference between the progressing low-grade dysplasias and those without a history of progressing in terms of gender distribution (56% male in progressing cases vs. 57% of those without a history), age distribution (mean age 58 years in progressing cases vs 55 in those without a history), site distribution, and smoking history (of those with known habits, 78% of progressing cases vs. 85% of those without a history were smokers). However, on average non-progressing cases were monitored for over twice the duration (96 versus 37 months) to ensure that progression did not occur. (Table 9). Table 9 Characteristics of patients with dysplasia Features Non- Progressing P progressing Age (mean, years) " 55 ~~58 0T4T6 Sex (% male) 57 56 1 % with smoking 85 - -^g ^ history Follow-up 96 37 0.0001 (mean, months) A history of recurrence was found for 2 of 6 progressive hyperplasias (compared with none of the 33 cases without a history of progression, p = 0.0192) and for 13 of 25 (52%) progressing low-grade dysplasia (compared to 4 of 55, or 7%, of dysplasias without a history of progression, p < 0.0001, Table 10 and 11). Furthermore, despite the fact that low-grade dysplasias in British Columbia are generally treated conservatively, 6 of the 25 progressive dysplasias had a further wide removal of their lesion (Table 10) probably because of their recurrence and progression. 46 Table 10 L O H frequencies and clinical history in progressing lesions Id# Age/ sex Site Diagnosis Time (mo.)a Chromosomal arms showing L O H b 405 4 3 / M Tongue Hyperplasia 0 9p S C C 7 9p, 3p, 17p, 8p, l lq 385 66/F Tongue Hyperplasia 0 3p S C C 16 3p, 17p,llq, 8p,4q 406 68 /M Retromolar Hyperplasia 0 9p, 17p, l l q Hyperplasia 8 N A S C C 12 9p, 17p, 3p, 4q 416 5 7 / M Retromolar Hyperplasia 0 3p S C C 72 3p, 9p, l l q 471 70 /M Gum Hyperplasia 0 3p, 4q, 8p CIS 105 N A 363 70/F Tongue Hyperplasia 0 3p,9p, 17p,4q,8p, l l q Hyperplasia 8 N A Moderate dysplasia 31 N A Hyperplasia 46 N A S C C 76 N A 173 7 4 / M Tongue M i l d dysplasia 0 3p, 9p, 8p, l l q , 13q M i l d dysplasia N A S C C 16 3p, 9p, 8p, l l q , 17p 122 56/F Tongue M i l d dysplasia 0 3p, 9p, 8p, l lq , 17p M i l d dysplasia 15 N A C Severe dysplasia, removed with wide margin 32 N A S C C 99 3p, 9p, 8p, 4q, 13q 47 223 29/ M Tongue M i l d dysplasia, 0 9p Persistently recurring leukoplakia treated by multiple wide electrocauterization N A N A S C C 51 9p, 17p,8p 271 69/F Floor of mouth M i l d dysplasia 0 3p,9p, 17p, l l q , 13q Hyperplasia 4 N A CIS 10 3p, 9p, 17p, l l q , 13q, 4q 293 47/ M Tongue M i l d dysplasia 0 3p, 9p, 17p S C C 26 3p, 9p, 17p,4q,8p 399 44/F Tongue M i l d dysplasia 0 9p,17p Clinical recurrence and treated with bleomycin 8 N A S C C 10 9p, 17p, 3p 365 807F Gum M i l d dysplasia 0 N A M i l d dysplasia 11 9p M i l d dysplasia 23 9p S C C (probably from V C ) 26 9p 446 47/F Floor of mouth M i l d dysplasia 0 8p, 9p S C C (very inflamed) 60 8p 472 60/ M Tongue M i l d dysplasia 0 3p ,9p ,4q ,8p , l l q Hyperplasia 2 N A S C C 85 3p ,9p ,4q ,8p , l l q 48 185 50/ M Tongue Moderate dysplasia 0 9p, 4q Clinical recurrence o f leukoplakia and all grossly visible lesion removed with margin (no bx submitted) 48 N A S C C 83 9p, 4q, 3p, 8p, l l q 197 39/ M Tongue Moderate dysplasia 0 3p,9p Moderate dysplasia 6 N A S C C 7 N A 377 52/ M Cheek Moderate dysplasia 0 3p, 4q Moderate dysplasia *7 N A Moderate dysplasia followed by wide laser excision 35 N A S C C 70 3p, 4q, 8p, 13q 473 73/ M Gum Moderate dysplasia 0 3p, 8p, l l q Moderate dysplasia removed with wide margin 5 3p, 8p, l l q , 9p S C C 11 8p, l l q , 9p, 3p, 13q 245 70/F Cheek Moderate dysplasia 0 3p, 9p S C C 56 3p, 9p, 8p, 4q, 13q 281 53/ M Tongue Moderate dysplasia 0 3p, 9p, 17p, 4q, 8p, l l q S C C 52 3p, 9p, 17p, 4q, 8p, l l q 49 401 42/ M Tongue Moderate dysplasia 0 9p, 17p, 4q, l l q , 13q, 8p, SCC 15 9p, 17p, 4q, l l q , 13q 286 65/ M Cheek Moderate dysplasia* 0 3p, 17p, 8p, 13q CIS 11 3p, 17p, 8p, 13q, 9p, 4q, l l q 460 46/ M Floor of mouth Moderate dysplasia 0 3p, 9p, 4q, l l q , 13q Recurrence of leukoplakia but no further biopsy or treatment were given NA NA SCC 46 3p,9p,4q, l l q , 13q, 8p 79 60/F Tongue Moderate dysplasia 0 3p,9p,17p,4q CIS 29 NA 464 71/F Tongue Moderate dysplasia 0 3p, 17p, 8p SCC 27 3p, 17p, 8p, l l q 121 80/F Gum Moderate dysplasia 0 9p, 8p Moderate dysplasia 4 NA SCC 29 9p, 8p, 3p, 13q 470 46/ M Lower lip Moderate dysplasia 0 8p, l l q , 13q Moderate dysplasia 5 NA CIS 6 8p, l l q , 13q 95 76/F Tongue Moderate dysplasia 0 9p, 13q CIS 32 9p Time between biopsies, with time of first biopsy set at 0. b Chromosome arms that are bold are those showing LOH in 1 biopsy but not in the other; italicized arms have different alleles lost in consecutive biopsies *Note: This patient had history of cancer in multiple organs prior to buccal mucosa dysplasia: (1) 11 years ago: cervical lymphoma treated with chemotherapy and prednisolone; (2) 6 years ago: Bladder 50 tumor treated with surgery; (3) 3 years ago: lung SCC treated with surgery; and (4) 1 year ago: prostate cancer treated with surgery and estrogen 51 Table 11 Allelic loss in recurrent premalignant lesions without a history of progression Id# Age/ sex Site Diagnosis Time (mo.)* Chromosomal arms showing L O H 26 6 5 / M Left anterior gingiva Moderate dysplasia 0 N o L O H M i l d dysplasia 10 N A 88 68/F Right floor of mouth M i l d dysplasia 0 3p, 9p, 17p, 8p, 13q, 14q, 8q M i l d to moderate dysplasia 9 N A M i l d dysplasia 15 N A 277 47/? Floor of mouth Hyperplasia 0 17p, 4q (non-informative for 3p) M i l d dysplasia 13 189 43/F Soft palate Moderate dysplasia 0 9p, 4q,8p Moderate dysplasia 32 N A Time between biopsies, with time of first biopsy set at 0. 52 8. D I S C U S S I O N Over the last decade, genetic analysis of tumors has led to significant breakthroughs in our understanding of alterations in tissues that underlie tumorigenesis. One of the current research challenges is to begin to apply such technology to the earliest clinical lesions, both to improve our understanding of the genetic alterations that underlie early cancer development and, hopefully, to provide potential indicators of risk for such lesions. 8.1. High frequency of allelic loss characterized dysplastic lesions This study showed that loss of regions of chromosomes containing presumptive tumor suppressor genes occurred early during oral carcinogenesis. A significant percentage of low-grade oral dysplasias showed allelic loss suggesting that loss of these regions may play an important role in the evolution and growth advantage of these premalignant clones. The presence o f histological evidence of dysplasia is presently the gold standard for judging malignant potential of oral premalignant lesions. This study showed that advent of dysplasia was accompanied by significantly increased frequency of L O H when compared to hyperplasias. When the patterns of loss among hyperplastic and low-grade dysplastic lesions were compared (Table 4), L O H was found in only 13 of 39 (33%) cases showing loss on any of the 7 arms studied in hyperplastic lesions. In contrast, such loss occurred in the majority of dysplastic lesions, 55 of 77 (71%) samples (p < 0.001). This increase in allelic loss in dysplastic lesions also was evident when the percentage of cases showing multiple arm loss was determined. While multiple losses among hyperplastic lesions was limited to 3 of 39 (7.7%) cases, 38 of 77 (49%) of low-grade dysplasias showed > 1 arm lost (pO.OOOl). Our 53 results are consistent with Mao 's : Lesions with dysplastic alteration had a higher rate of microsatellite alterations (3pl4 and 9p21), 50% vs 31% (P = 0.08). It is well accepted that presence of dysplasia has been associated with an increased risk of malignant transformation (van der Waal et al, 1997). The fact that the frequency of L O H was associated with the presence of dysplasia would suggest a cancer predictive value for L O H and also suggest that molecular changes may underline the dysplastic phenotypes. 8.2 Increased frequency of allelic loss in progressing lesions Linkage of specific patterns of genetic alteration to disease progression is often limited by the difficulty of obtaining clinical specimens from the same lesion over time. In this study, microdissected early oral premalignant lesions from 116 patients with or without a history of progression into CIS or invasive S C C were analyzed for L O H at on 7 chromosomes in order to identify genetic differences between progressing and non-progressing lesions, and to identify genetic profiles that have predictive value for early premalignant lesions. One of the more striking observations made in this study is markedly increased frequency of allelic loss in progressing lesions as compared to the non-progressing lesions. A l l such progressing lesions, both hyperplasias and dysplasias, showed loss on at least 1 arm. In contrast, hyperplasias and low-grade dysplasias without a history of progression had loss in 21% and 59% of cases respectively (Table 4). Furthermore, multiple arm loss (>1 arms lost) was absent in hyperplasias without a history of progression but present in 3 of 6 progressing hyperplasias. Also , the majority of progressing dysplasias had >1 arms lost (91% of cases compared to 31% of dysplasias without a history of progression) with 57% of cases having loss at >2 arms (compared with 20% o f dysplasias without a history of progression). 54 These data support the hypothesis that progressing early lesions and those without a history of progression can appear morphologically or phenotypically similar, yet differ significantly at the genetic level. Furthermore, the data suggest that assaying hyperplasias and low-grade dysplasias for L O H frequencies might be o f clinical use in differentiating between low-risk and high-risk lesions. 8.3. The pattern of allelic loss characterized progressing lesions The earliest loss was at 3p and occurred before the advent of dysplasia. Even hyperplasia without a history of progression demonstrated 3p loss in 13% o f the cases. This may suggest that TSGs at 3p, such as FHIT, may play an important role in early oral carcinogenesis. The most common loss in low-grade dysplasia was at 9p (46%), supporting the hypothesis that p i 6 gene dysfunction plays important role in early oral carcinogenesis. The fact that the most common losses for both sets of cases were on 3p and 9p was consistent with results from previous studies that have shown that L O H at 3p l4 and 9p21 (the regions studied in this paper) are frequent occurrences in oral premalignant lesions and likely to occur early in oral carcinogenesis (Zhang et al, 1996; Califano et al, 1996; Mao et al, 1996a; Roz et al, 1996). The loss at 3p &/or 9p was not only the earliest and most common event during oral carcinogenesis, it was also significantly higher in the progressing lesions (Table 5), with this trend apparent when 3p and 9p are considered separately, even more apparent when they are considered together. Virtually all progressing lesions (22/23, 96% in dysplasia and 6/6, 100% in hyperplasia) had loss on 3p &/or 9p, compared to 5 of 29 (17%) hyperplasias without a 55 history of progression and 30 of 54 (56%) of low-grade dysplasias without a history of progression. These data suggest that L O H at 3p &/or 9p is not simply one of those random genetic alterations but rather a prerequisite for progression of oral premalignant lesions. From a clinical point of view, this would suggest that analysis of L O H at 3p and 9p may be used as an important initial screen for the malignant risk of leukoplakia. However, because of the high frequency of L O H at 3p and 9p in early lesions without a history o f progression, especially in dysplasia (56%), these markers may not be the best markers for prediction of prognosis, at least not by themselves. L O H on the other 5 chromosome arms (4q, 8p, l l q , 13q and 17p) occurred rarely or at low frequency in non-progressing lesions. In contrast, a significant number of progressing lesions including both hyperplasia and low-grade dysplasia showed L O H at these chromosome arms. L O H in these chromosome arms have also been reported to occur in oral premalignant lesions; however, such losses (especially for chromosome arm 4q, 8p, l l q and 13q) were generally seen in high-grade preinvasive lesions (severe dysplasia/CIS) or S C C (Califano et al, 1996; unpublished data from this lab). This would suggest that L O H s on these chromosome arms are relatively Tate' events in the preinvasive stages and should be rare or absent in hyperplastic or low-grade dysplastic lesions. This was indeed the case for the low-grade dysplasias without a history of progression in this study. In contrast, progressing low-grade dysplasias had increased frequencies of loss for these chromosome arms. Progressing low-grade dysplasias had significant increases for 8p (p = 0.003), 1 l q (p = 0.01) and 13q (p = 0.002) with the increase for 4q approaching significance (p = 0.057). There was also a doubling in the frequency of L O H on 17p (from 20% to 41% of cases); although this increase was not significant or only marginally significant. 56 The high frequency of L O H on these arms in high-risk advanced oral preinvasive lesions and in lesions with proven progression would suggest L O H on these arms indicates a higher cancer risk (than those with only 3p &/or 9p). In fact, when the percentage of cases that showed a combination of 3p &/or 9p loss with additional losses at any of the remaining arms was determined, this occurrence was shown to be present in 0% of hyperplasias without a history of progression compared to 50% of progressing hyperplasias and in 28% of dysplasias without a history of progression and 78% of progressing dysplasias, both are significantly different (Table 5). These results would suggest that additional losses on these chromosome arms indicate cancer risk. 8.4. L O H as markers to predict cancer risk - further statistic analysis The present study showed that progressing and non-progressing lesions had significantly different L O H profiles. H o w could we use the study data for the evaluation of the cancer risk of oral premalignant lesions? Since loss at 3p &/or 9p was the most common and earliest event and seemed to a prerequisite for progression, it may be used as an initial screening for assessing cancer risk of oral premalignancies. If L O H at 3p &/or 9p had been used as an initial screening for our study set, without knowledge of L O H at other arms, those cases with 3p &/or 9p L O H (with or without L O H at other arms) would have had a 24-fold increase in the relative risk of cancer progression as compared to those without L O H at either 3p or 9p (Fig.4C, Table 7). This is consistent with the results of Mao et al, (1996a) who found L O H at 3p &/or 9p predicts cancer risk of oral leukoplakias. However, since there was a high frequency of allelic loss on these arms in non-progressors and the relative cancer risk for those with L O H at 3p &/or 9p but no other arms was only increased by 3.8-fold in the 57 relative risk of cancer progression as compared to those without L O H at either 3p &/or 9p (Fig.4D, Table 7), additional markers are essential for better prediction of prognosis. The study results suggest that loss at any of the other 5 chromosomes (4q, 8p, l l q , 13q, and 17p) in addition to L O H at 3p &/or 9p seem to provide a better predictive value. Those cases with such losses had a 3 3-fold increased risk of progressing to cancer compared to cases that retained both of these arms (Fig 4D, Table 7). Furthermore, time-to-progression curves showed that lesions that had 3p &/or 9p loss with additional loss on at least one of the indicated arms had a significantly shorter progression time than those with 3p &/or 9p loss only (p<0.01,Fig. 4D). To determine which of the additional losses (on 4q, 8p, l l q , 13q or 17p) would most significantly increase progression risk, the study separately compared those cases with 3p &/or 9p loss alone against those cases with 3p &/or 9p loss plus each of the additional losses (Figure 4E to I). A significantly shorter time-to-progression was observed when either 8p (p<0.03, by the logrank test), l l q (p<0.05) or 13q (p<0.01), L O H was present in addition to 3p &/or 9p L O H . Comparisons with 4q (P = 0.09) or 17p (P = 0.08) were not statistically significant, although a trend was observed. For each premalignant L O H pattern, the probability of having no subsequent progression is summarized in Table 7. Forty to sixty percent of individuals with additional losses at 4q, 8p, l l q or 17p developed cancer within 5 years, corresponding to a 2.2-2.6-fold increase in relative risk of cancer progression as compared to those with L O H at 3p&/or 9p only. Moreover, cases with additional 13q loss had a 7-fold increase in risk of progression. Six of the 8 cases with loss on this arm had 5 years of follow-up and all showed progression within this timeframe. 58 In conclusion, although prospective studies involving large numbers of subjects over time are necessary to fully understand the relation between chromosomal loss and tumorigenesis, the study data do suggest that L O H patterns w i l l facilitate the prediction of the malignant potential of low-grade premalignancies. Can we use this information clinically? In addition to being used for deletion mapping of tumor suppressor genes, the L O H assay has begun to be used to assist pathologists in screening for patients with premalignant lesions and cancer. Mao (1996c) analyzed L O H in urothelial cells obtained from urine sediments of cancer patients and correctly identified 95% of the patients with cancer. There have also been prognostic studies that attempt to correlate the degree of L O H and survival rate (Partridge et al, 1996; Scholnick et al, 1996; Nawroz et al, 1996; Field et al 1995). For example, Nawroz (1996) found that patients with L O H at any of their 12 markers all had advanced disease (stage III or IV) , most had nodal metastases, and half of them died of disease. Recently, Califano (1999) tried to use detection of genetic alterations for molecular identification of the site of origin o f the primary tumor; Harda (1999) and Ogawara (1998) both found L O H on 13q was significantly correlated with lymph node metastasis in oral S C C and H N S C C ; Partridge (1999c) found that patients with allelic imbalance at 3p24-26, 3pl3 and 9p21 have an approximately 25 times increase in their mortality rate relative to patients retaining heterozygosity at these loci . She suggested that it would be possible to develop a molecular staging system which w i l l be a better predictor of outcome than conventional clinicopathological features, as the molecular events represent fundamental biological characteristics of each tumour; Matsuura (1998) found L O H of chromosome 9p21 and 7p31 is correlated with high incidence of recurrent in H N S C C . Finally, Mao (1998) has suggested that clinical and histologic assessments of the response to chemporeventive agents may be 59 insufficient to determine their efficacy and that critical genetic alterations could be used as independent biomarkers to augment the ability to evaluate the efficacy of such agents. A l l these studies positively support our inquiry. Our data suggest that L O H at 3p &/or 9p would place the patient into an at risk category since nearly all progressing cases in our cohort showed such alteration. Their relative risk increases with additional loss on other arms. Although additional losses on any of these arms is an indicator of probable progression, the loss of 13q might signal the need for active intervention with either traditional or novel forms of therapy such as chemoprevention. 8.5. Most progressed lesions may be derived by clonal outgrowth from the earlier lesions. Comparing the L O H pattern o f premalignant and malignant in patients that later progressed to CIS or S C C , we have found that most of L O H patterns in the premalignant lesions were found in their matched later lesions, although 14 of these cases showed additional losses on other chromosome arms and the remaining 8 cases showed an alteration in the pattern of loss in the tumor for some of loci that had L O H in the early biopsy (Fig. 5 b and Table 8). However, in 3 of the latter cases this difference was restricted to a single arm out of 5 or 6 arms lost. For example, in case #173 the early lesion contained a L O H at 13q that was not found in the later lesion; however, the pair showed loss of the same alleles on 3p, 9p, 8p and l l q . Our analysis of lesions with progressed cases suggests that for most progressing lesions, the later cancer was derived by clonal outgrowth from the earlier lesions. In other words, the later cancer progressed from the early lesion. 60 8.6. Does evidernce of recurrence serve as a further indicator of cancer risk in premalignant lesions? We obtained some clinical information for our subjects by tracking recurrent lesions as they appeared in our database and by following the case histories of treatment in hospital charts. Wi th the information available, the data showed a significantly higher number of recurrences among progressing lesions (61% of progressing dysplasias vs. 7% dysplasias without a history of progression) (Table 10 and 11). Although there are some studies that correlated L O H and recurrence in head and neck cancer (Matsuura et al, 1998; Lydiatt et al, 1994 and 1998), no study showed the relationship between recurrence and cancer risk in premalignant lesions. There are at least two possible explanations for our results: 1) the recurrence resulted from failure to totally remove these lesions and this failure was more prevalent among progressing cases; and 2) the recurrence reflected a more aggressive lesion with an elevated rate of genetic change. The study does suggest that a large percentage of progressing lesions were not totally removed. In 17 of 25 (68%) progressing lesions, loci lost in the early lesions showed the same pattern of loss (upper versus lower allele) in the later lesion, suggesting that they were derived by clonal outgrowth from the earlier lesions. Interestingly, many of these lesions were considered to be totally eliminated using histological and clinical criteria. However, there is no evidence that treatment was less aggressive among progressing cases compared with those without a history of progression. In fact, in 6 of the 25 progressing dysplasias the decision was made to treat aggressively again, based on the recurrence and/or progression of the lesions (Table 10: #473, #122, #223, #399, #185 and #377). Despite wide removal of the 61 lesion or wide laser excision or multiple wide electrocauterization or treatment with bleomycin, the same site later still progressed into cancer. Our data support the second probability that these progressing lesions were indeed more aggressive in nature, possibly due to the specific genetic alterations that they had undergone. The recurrence reflected a more aggressive lesion with an elevated rate of genetic change. It is well known that many oral leukoplakias, which histologically consist of hyperplasia &/or dysplasia, w i l l not progress into cancer even when left untreated. This would imply that incomplete removal is only important clinically when the premalignant lesion is a high-risk lesion (genetic profile more important). In a rare, long-term. study, Banoczy (1977) followed 670 patients with oral leukoplakia for up to 30 years (average 10 years). The treatment for many o f these lesions was described as 'removal of local irritants and conservative treatment'. A large percentage of these lesions either regressed (29.7%) or remained static (25.8%). It is possible that we did not see as many recurrent lesions without a history of progression as progressing lesions because the clones of genetically altered cells regressed or remained static in lesions without a history of progression, even i f they were not totally removed during the incisional biopsy. The result that 3 o f the 4 recurrent premalignant lesions without a history of progression had multiple losses (Table 11) supports the hypothesis that the recurrent lesions were biologically more aggressive and that this behavior was associated either directly or indirectly with the amount of genetic damage in the lesion. Moreover, whether or not the failure of these recurrent lesions to progress was a result of their complete removal or insufficient time for follow-up is still not clear although a significantly greater length of follow-up was observed for non-progressing lesions (progressing cases, 37 62 months vs. non-progressing, 96 months, P = 0.0001) (Table 9). We may pay more attention to the recurrent lesions clinically. 8.7. Summary The results from this study suggest that L O H patterns w i l l facilitate the prediction of the malignant potential of low-grade premalignancies. In this study, we have demonstrated the predictive value of allelic loss subsequent to 3p &/or 9p loss, especially with 13q loss which increases the relative risk in progression by 7 fold. We may more precisely predict the malignant potential of low-grade premalignancies i f we combine L O H patterns with the predictive value of dysplasia and clinical attributes such as recurrence after treatment. We are only beginning to understand the processes that control the malignant transformation of oral premalignancies. For example, despite the significant association between multiple L O H and cancer risk, the speed of the malignant transformation was not affected by the number of the L O H in a lesion (Table 10). Some premalignancies had a relatively low number of regions showing L O H and yet progressed into cancer rapidly while others, with numerous losses, seemed to 'sit ' for a long time before S C C occurred. This suggests that other changes, either genetic or epigenetic, must control or influence the malignant transformation. Such changes could include not only allelic losses at other chromosomal regions but also mutations to proto-oncogenes and TSGs or even epigenetic changes such as D N A methylation or histone deacetylation (Bakin et al, 1999). Alternatively, the immune system or homeostatic regulation of premalignant cells by surrounding normal cells such as epithelial-mesenchymal interaction could also play a role in controlling outgrowth of such lesions. 63 Furthermore, the study also suggests that not all cells in the high-risk premalignancies were removed even though the margins appeared clinically and histologically clean. Previous studies have shown that the presence of genetically damaged cells at the margins of surgical resection of oral SCC could account for tumor recurrence, despite gross and histologically free surgical margins (Westra et al, 1998). The same principle might be applied to oral premalignant lesions. If so, the malignant transformation rate of oral premalignant lesions could be reduced significantly in the future if we guide our surgical margin not only by gross and histological criteria, but also by molecular markers, especially for those premalignant lesions designated as high-risk. 64 9. R E F E R E N C E S 1. Adamson R, Jones A S , Field JK: Loss of heterozygosity studies on chromosome 17 in head and neck cancer using microsatellite markers. Oncogene 1994, 9:2077-2082. 2. Ah-See K . , Cooke T., Pickford I., Soutar D . , and Balmain A . A n allelotype of squamous carcinoma o f the head and neck using microsatellite markers. Cancer Research, 54: 1617-21, 1994. 3. Alberts B . , Bray D . , Lewis J., Raff M . , Roberts K . , and Watson J. Molecular Biology of the cell (3 r d Edition). Garland Publishing, Inc. N e w York & London. 1994 4. Armitage P and Berry G Statistical methods in medical research, Second Edition • Blackwel l Scientific Publications, 1987 Chapter 14, Survival Analysis, pp421-439 5. Bakin A V , Curran T: Role of D N A 5-methylcytosine transferase in cell transformation by fos. Science 1999,283:387-390. 6. Banocy J: Follow-up studies in leukoplakia. J Max-fac Surg: 69-75, 1977 7. Boiocchi M : Chromosome 13q deletion mapping in head and neck squamous cell carcinomas: identification of two distinct regions of preferential loss. Cancer Res 1996, 56:1146-1150. 8. Bockmuhl, U . , Schwendel, A . , Dietel, M . , and Petersen, I. Distinct patterns of chromosomal alterations in high- and low-grade head and neck squamous cell carcinomas. Cancer Research, 56: 5325-29, 1996. 9. Bouquot JE: The Pathology and Progression of Oral Premalignancy. Epithelial Dysplasia Symposium. 5 t h International Congress on Oral Cancer. Edited by Bouquot JE. Royal College of Physicians, London, United Kingdom, 1997. 65 10. Brennan J A and Mao L et al Molecular assessment of histopathological staging in squamous-cell carcinoma of the head and neck N Engl J M e d 332:429-435, 1995 11. Brunette D M Critical thinking-understanding and evaluating dental research Edited by Patricia Bereck Weikersheimer. Quintessence Publishing Co, Inc 1996, p l29 . 12. Califano J, van der Riet P, Westra W , Nawroz H , Clayman G , Piatadosi, S, Corio R, Lee D , Greenberg B , K o c h W , Sidransky D : Genetic progression model for head and neck cancer: Implications for field cancerization. Cancer Res 1996, 56:2488-2492 13. Califano, J., Westra W H . , Koch , W . , Meininger G , Reed A . , Y i p L . , Boyle , JO, Lonardo, F and Sidransky D : Unknown primary head and neck S C C : Molecular identification of the site of origin J. Nati.Can Institute 91(7): 599-604,1999 14. Dawson C D , Chang K W , Solt D B . M T S 1 gene mutations in archival oral squamous cell carcinoma. J Oral Pathol Med. , 25: 541-546, 1996. 15. Chap, T L E , Appl ied survival analysis John Wi ley & Sons, Inc. 1997 16. Croce, C M . , Sozzi , G . and Huebner, K Role of F H I T in human cancer J C l i n Oncol 17: 1618-1624, 1999 17. El-Naggar A K , Hurr K , Batsakis J G , Luna M A , Goepfert H , Huff V : Sequential loss of heterozygosity at microsatellite motifs in preinvasive and invasive head and neck squamous carcinoma. Cancer Res 1995, 55(12):2656-2659. 18. El-Naggar A K , Coombes M M , Batsakis J G , Hong W K , Goepfert H , Kagan J: Localization of chromosome 8p regions involved in early tumorigenesis of oral and laryngeal squamous carcinoma. Oncogene 1998, 16:2983-2987. 19. Emi l ion G , Langdon JD, Speight P, Partridge M : Frequent gene deletions in potentially malignant oral lesions. Brit J Cancer 1996, 73(6):809-813 66 20. Fearon, E . Human cancer syndromes: clues to the origin and nature of cancer. Science, 278: 1043-50, 1997. 21. Field, J., Kiaris , H . , Risk, J., Tsiriyotis, C , Adamson, R. , Zoumpourlis, V . , Rowley, H . , Taylor, K . , Whittaker, J., Howard, P., Birne, J., Gosney, J. , Woolgar, J., Vaughan, E . , Spandidos, D . , and Jones, A . Allelotype of squamous cell carcinoma of the head and neck: fractional allele loss correlates with survival. Brit ish Journal of Cancer, 72: 1180-8, 1995. 22. Gallo, O and Sardi., I. Multiple primary of the upper aerodigestive tract: is there a role for constitutional mutations in the p53 gene? Int. J. Cancer: 1999, 82:180-186. 23. Harada, H . and Tanaka, H et al. Lymph node metastasis is associated with allelic loss on chromosome 13ql2-13 in esophageal squamous cell carcinoma. Cancer Research, 59:3724-3729, 1999. 24. H u L F , Eiriksdottir G , Lebedeva T, Kholodniluk I, A l imov A , Chen F, Luo Y , Zabarovsky E R , Ingvarsson S, K l e i n G , Ernberg I. Loss of heterozygosity on chromosome arm 3p in nasopharyngeal carcinoma. Gene Chrom Cancer, 17: 118-126, 1996. 25. Ilyas, M & I .P .M. Tomlinson. Genetic pathways in colorectal cancer Histopathology 28: 389-399, 1996 26. Ishwad, C , Ferrell, R. , Rossie, K . , Appel , B . , Johnson, J., Myers, E . , Law, J., Srivastava, S., and Gol l in , S. Loss of heterozygosity of the short arm o f chromosomes 3 and 9 in oral cancer. International Journal of Cancer (Pred. Oncology), 69: 1-4, 1996. 27. Jares P and Nadal A et al Disregulation of pi6 MTS1/ CDK41 protein and m R N A expression is associated with gene alterations in squamous-cell carcinima of the larynx. Int. J. Cancer 81:705-711, 1999 67 28. Jiang W , Zhang Y - J , Kahn S M , Hollstein M C et al. Altered expression of the cyclin D l and retinoblastoma genes in human esophageal cancer Proc. Natl Acad Sci U S A 90: 9026-30,1993 29. Jordan R C K and Daley T Oral squamous cell carcinoma: new insights. Canadian Dental Association 63 (7) 517-525 30. K i n g , R J B , Cancer biology. Addison Wesley Longman Limited, 1996 31. Lazar A D , Winter M R , Noguerira C P , Larson PS, Finnemore E M , Dolan R W , Fuleihan N , Chakravarti A , Zietman A , Rosenberg C: Loss of heterozygosity at 1 lq23 in squamous cell carcinoma of the head and neck is associated with recurrent disease. C l i n Cancer Res 1998, 4:2787-2793. 32. Leis JF, Livingston D M . The tumor suppressor genes and their mechanisms of action. In Bishop J M , Weinberg R A ed, Molecular Oncology. Scientific A m . , Inc. p p l 11-41, 1996. 33. L i X , Lee N K , Y e Y W , Waber P G , Schweitzer C, Cheng Q C , Nisen P D . Al le l i c loss at chromosomes 3p, 8p, 13q and 17p associated with poor prognosis in head and neck cancer. J Natl Cancer Inst., 86(20): 1524-1529,1994. 34. Liggett W H and sidransky D Role of the p i 6 tumor suppressor gene in cancer Journal of Clinical Oncology 16(3)1197-1206, 1998 35. Lydiatt W M , Davidson B J , Shah J, Schantz SP, Changanti R S K . The relationship of loss of heterozygosity to tobacco exposure and early recurrence in head and neck squamous cell carcinoma. A m J Surg, 168: 437-440, 1994. 36. Lydiatt, W. , Davidson, B . , Schantz, S., Caruana, S., and Chaganti, S. 9p21 deletion correlates with recurrence in head and neck cancer. Head and Neck, 20(2): 113-8, 1998. 68 37. Maestro, R., Gasparotto, D . , Vukosavljevic, T., Barzan, L . , Sulfaro, S., and Boiocchi , M . Three discrete regions of deletion at 3p in head and neck cancers. Cancer Research, 53: 5775-9,1993. 38. Maestro R, Piccinin S, Doglioni C , Gasparotto D , Vukosavljevic T, Sulfaro S, Barzan L , Boiocchi M . Chromosome 13q deletion mapping in head and neck squamous cell carcinoma: identification of two distinct regions of preferential loss. Cancer Res., 56: 1146-1150, 1996. 39. Matsuura K and Shiga K et al Loss of Heterozygosity o f Chromosome 9p21 and 7p31 is correlated with high incidence of recurrent Tumor in head and neck squamous cell carcinoma Anticancer Research 18:453-8, 1998 40. Matsuda H and Konishi N et al Alterations of p l 9 / C D K N 2 , p53 and ras genes in oral \ squamous cell carcinomas and premalignant lesions. J Oral Pathol M e d 25: 232-238, 1996 41. Merlo A and Herman J G et al 5' C p G island methylation is associated with transcriptional silencing of the rumour suppressor p l 6 / C D K N 2 / M T S - l in human cancers Nature M e d 1:686-692, 1995 42. Mao, E . , Oda, D . , Haigh, W. , and Beckman, A . Loss of the Adenomatous Polyposis C o l i gene and human papillomavirus infection in oral carcinogenesis. Oral Oncology, European Journal of Cancer, 32B: 260-3, 1996. 43. Mao , L . , Lee, J., Fan, Y . , Ro , J., Batsakis, J., Lippman, S., Hittelman, W. , and Hong, W . Frequent microsatellite alternations at chromosomes 9p21 and 3p l4 in oral premalignant lesions and their value in cancer risk assessment. Nature Medicine, 2: 682-5, 1996a. 44. Mao L , Fan Y H , Lotan R, Hong W K . Frequent abnormalities of FHIT , a candidate tumor suppressor gene, in head and neck cancer cell lines. Cancer Res., 56: 5128-5131, 1996b. 69 45. Mao, L and El-Naggar, A K et al Phenotype and Genotype o f advanced premalignant head and neck lesions after chemopreventive therapy J. Nati . Can. Institute 90 (20): 1545-1551, 1998a 46. Mao L . Tumor suppressor genes: Does FHIT fit? J of the national cancer institute 90 (6) 412-414,1998b 47. Nakanishi, H and Wang, X L et al, Localization of a novel tumor suppressor gene loci on chromosome 9p21-22 in oral cancer Anticancer Research 19:29-34, 1999 48. Nawroz H , van der Riet P, Hruban R H , K o c h W , Ruppert J M , Sidransky D : Allelotype of head and neck squamous cell carcinoma. Cancer Res 1994, 54(5):1152-1155. 49. Nowel l P C . The clonal evolution of tumor cell populations. Science, 194: 23-28, 1976. 50. Ogawara K , Miyakawa A , Shiiba M , Uzawa K , Watanabe T, Wang X L , Sato T, Kubosawa H , Kondo Y , Tanzawa H . Al le l i c loss of chromosome 13ql4.3 in human oral cancer: correlation with lymph node metastasis. Int J Cancer: 79: 312-317, 1998. 51.0hta , M . , Inoue, H . , Cotticelli, M . , Kastury, K . , Baffa, R., Palazzo, J., Siprashvili, Z . , M o r i , M . , McCue , P., Druck, T., Croce, C , and Heubner, K . The F H I T gene, spanning the chromosome 3pl4.2 fragile site and renal carcinoma-associated t(3;8) breakpoint, is abnormal in digestive tract cancers. Ce l l , 84: 587-97, 1996. 52. Papadimitrakopoulou, V . , Izzo, J., Lippman, S., Lee, J., Fan, Y . , Clayman, G . , Ro , J., Hittelman, W. , Lotan, R., Hong, W. , and Mao, L . Frequent inactivation of p i 6 in oral premalignant lesions. Oncogene, 14: 1799-803, 1997a. 53. Papadimitrakoulou V , Izzo J, Lippman S M , Lee JS, Fan Y H , Clayman G , Ro J Y , Hittleman W N , Lotan R, Hong, W K , Mao L . Frequent inactivation of p l 6 / I N K 4 a in oral premalignant lesions. B r i J Cancer, 14: 1799-1803, 1997b. 70 54. Partridge, M . , Kiguwa, S., and Langdon, J. Frequent deletion of chromosome 3p in oral squamous cell carcinoma. Oral Oncology, European Journal of Cancer, 30: 248-51, 1994. 55. Partridge M . , Emi l l ion G. , and Langdon J. L O H at 3p correlates with a poor survival in oral squamous cell carcinoma. British Journal of Cancer, 73: 366-71, 1996. 56. Partridge M . and Emil l ion G et al A l l e l i c imbalance at chromosomal loci implicated in the pathogenesis of oral precancer, cumulative loss and its relationship with progression to cancer Oral Oncology 34:77-83, 1998 57. Partridge M . , Emi l l ion G . , Falworth M , Hern R A , Phillips E Pateromichelakis S, and Langdon J Patient-specific mutation databases for oral cancer. Int. J. Cancer (pred. Oncol.): 84:282-292, 1999a 58. Partridge M , Head and neck cancer and precancer: can we use molecular genetics to make better predictions? A n n R C o l l Surg Engl 81:1 -11, 1999b 59. Partridge M . , Emi l l ion G . , Pateromichelakis S, Hern R A , Phillips E arid Langdon J The prognostic significance of allelic imbalance at key chromosomal loci in oral cancer. British Journal of Cancer, 79(11/12): 1821-7, 1999c. 60. Pennisi, E . New gene forges link between fragile site and many cancers. Science, 272: 649, 1996. 61. Perhouse M A , El-Naggar A K , Hurr K , L i n H , Yung W K A , Steck P A . Deletion mapping of chromosome 4 in head and neck squamous cell carcinoma. Oncogene, 14, 369-373, 1997. 62. Raybaud-Diogene, H . , Tetu, B . , Morency, R., Fortin, A . , and Montei l R. p53 overexpression in head and neck squamous cell carcinoma: review o f the literature. Oral Oncology, European Journal of Cancer, 32B(3): 143-9, 1996. 71 63. Rawnsley JD, Srivatsan E S , Chakrabarti R, Bil l ings K R , Wang M B . Deletion analysis of the p l 6 / C D K N 2 gene in head and neck squamous cell carcinoma using quantitative polymerase chain reaction method. Arch Otolaryngol Head Neck Surg., 123: 863-867, 1997. 64. Regezi J A , Sciubba JJ: Oral Pathology. Clinical-Pathologic Correlations. Edited by Regezi J A , Sciubba JJ. Toronto, W B Sanders Company, 1989, pp. 94-99. 65. Rosin M P , Cheng X , Poh C F , Priddy R, Berean K , Nakamur H , Zhang L : Leukoplakias that progressed to cancer have specific patterns of allelic loss. Cancer Res 1999, 40:136, Abstract #905. 66. Roz L , W u C L , Porter S, Scully C , Speight P, Read A , Sloan P, Thakker N : Al le l i c imbalance on chromosome 3p in oral dysplastic lesions: an early event in oral carcinogenesis. Cancer Res 1996, 56(6): 1228-1231. 67. Rowley H and Jones A et al, Definition of a tumor suppressor gene locus on the short are of chromosome 3 in squamous cell carcinoma of head and neck by means o f microsatellite markers. A r c h Otolaryngol Head Neck Surg.: 122:497-501, 1996 68. Sartor M and Steingrimadottir H et al Role of p l 6 / M T S l , cyclin D l and R B in primary oral cancer and oral cancer cell lines British Journal of Cancer 80(1/2)79-86, 1999 69. Sard L and Accornero P et al. The tumor-suppressor gene FHIT is involved in the regilation of apoptosis and in cell cycle control. Proc Natl Acad Sci U S A , 96 (15):8489-92,1999 70. Selwyn, M . R . Principles of experimental design for the life sciences C R C Press, Inc. 1996 72 71. Scholnick S B , Haughey B H , Sunwoo JB , El -Mofty S K , Baty JD, Picciri l lo JF, Zequeira M R . Chromosome 8 allelic loss and the outcome of patients with squamous cell carcinoma of the supraglottic larynx. J Natl Cancer Inst., 88(22): 1676-1682, 1996. 72. Schreiber G , Pitterle D , K i m Y and Bepler G Molecular Genetic analysis of primary lung cancer and cancer metastatic to the lung Anticcancer Research 19: 1109-1116, 1999 73. Scully, C. and Field, J. Genetic aberrations in squamous cell carcinoma of the head and neck ( S C C H N ) , with reference to oral carcinoma (Review). International Journal of Oncology, 10:5-21, 1996. 74. Silverman S, Jr, Gorsky M , Lozada F: Oral leukoplakia and malignant transformation. A follow-up study of 257 patients. Cancer 53:563-568, 1984 75. Somers K D , Cartwright S L , Schechter G L . Amplification of the int-2 gene in human head and neck squamous cell carcinoma. Oncogene, 5: 915-920, 1990. 76. Sozzi, G . , Veronese, M . , Negrini, M . , Baffa, R. , Cotticelli, M . , Inoue, H . , Torniell i , S., Pilotti, S., De Gregorio, L . , Pastorino, U . , Pierotti, M . , Ohta, M . , Heubner, K . , and Croce, C. The FHIT gene at 3pl4.2 is abnormal in lung cancer. Ce l l , 85: 17-26, 1996. 77. Sozzi G and Pastorino U et al. Loss of FHIT function in lung cancer and pervasive bronchial lesions. Cancer Res 58(22): 5032-7, 1998 78. Sunwoo JB and Holt M S et al Evidence for multiple tumor suppressor genes on chromosome arm 8p in supraglottic laryngeal cancer Genes Chromosomes Cancer 16(3): 164-169,1996 79. Thiberville L , Payne p et al. Evidence of cumulative gene losses with progreaaion of premalignant epithelial lesions to carcinoma of the bronchus. Cancer Res:55 (22):5133-9, 1995 80. Todd, R., McBride , J., Tsuji, T., Donoff, R., Nagal, M . , Chou, M . , Chiang, T., and Wong, D . Deleted in oral cancer-1 (doc-1), a novel oral tumor suppressor gene. F A S E B , 9: 1362-70, 1995. 81. Uzawa K . , Suzuki H , Komiya A , Nakanishi H , Ogawara K , Tanzawa H , Sato K : Evidence for two distinct tumor-suppressor gene loci on the long arm of chromosome 11 in human oral cancer. Int J Cancer 1996, 67:510-514. 82. Uzawa N , Yoshida M A , Hosoe S, Oshimura M Amagasa T and Ikeuchi T: Functional Evidence for involvement of multiple putative tumor suppressor gene on the short arm of chromosome 3 in human oral squamous cell carcinogenesis. Cancer Genet Cytogenet 107:125-131, 1998 83. V a n der Riet P, Nawroz H , Hruban R H , Corio R, Tokino K , K o c h W , Sidransky D . Frequent loss of chromosome 9p21-22 early in head and neck cancer progression. Cancer Res., 54: 1156-1158, 1994. 84. V a n der Waal I, Schepman K . P., van der Meij H . E.and Smeele L . E . . : Oral leukoplakia: a clinicopathological review. Oral Oncology, 33 (5) 291-301,1997 85. Venugopalan M , Wood T F , Wilczunski SP, Sen S, Peters J, M a G C , Evans G A , Srivatsan ES . Loss of heterozygosity in squamous cell carcinomas of the head and neck defines a tumor suppressor gene region on 1 l q l 3 . Cancer Genet Cytogenet, 104: 124-132, 1998. 86. Waber, P., Lee N . , and Nisen P. Frequent allelic loss at chromosome arm 3p is distinct from genetic alterations of the Von-Hippel Lindau tumor suppressor gene in head and neck cancer. Oncogene, 12: 365-9, 1996. 87. Waber P, Dlugosz S, Cheng Q C , Truelson J, Nisen P D . Genetic alterations of chromosome band 9p21-22 in head and neck cancer are not restrected to p l 6 / I N K 4 a . B r i J Cancer, J U N : 1699-1704, 1997. 74 88. Waldron C A and Shafer W G Leukoplakia Revisi ted—A clinicopathologic study 3256 oral Leukoplakias, Cancer 36:1386-1392,1975 89. Westra W H , Sidransky D : Phenotypic and genotypic disparity in premalignant lesions: of calm water and crocodiles. JNCI 1998, 90:1500-1501. 90. W H O Collaborating Reference Centre for Oral Precancerous Lesions: Definition of leukoplakia and related lesions: an aid to studies on oral precancer. Oral Surg 1978, 46:517-39. 91. Wi lke , C , Ha l l , B . , Hoge, A . , Pardee, W. , Smith, D . , and Glover, T. F R A 3 B extends over a broad region and contains a spontaneous H P V 1 6 integration site: direct evidence for the coincidence of viral integration sites and fragile sites. Human Molecular Genetics, 5: 187-95, 1996. 92. W u C L , Sloan P, Read A P, Harris R, Thakker N . Deletion mapping on the short arm of chromosome 3 in squamous cell carcinoma of the oral cavity. Cancer Res 1994, 54: 6484-6488. 93. W u C L , Roz L , Sloan P, Read A P , Holland S, Porter S., Scully C, Speight P M , Thakker N . Deletion mapping defined three discrete areas of allelic imbalance on chromosome arm 8p in oral and oropharyngeal squamous cell carcinoma. Gene Chrom Cancer, 20: 347-353, 1997. 94. Y o o G H , X u H J , Brennan J A , Westra W , Hruban R H , K o c h W , Benedict W F , Sidransky D . Infrequent inactivation of the retinoblastoma gene despite frequent loss of chtomosome 13q in head and neck squamous cell carcinoma. Cancer Res., 54: 4603-4606, 1994. 95. Zhang L , Michelsen C , Cheng X , Zeng T, Priddy R, Rosin M P : Molecular analysis of oral lichen planus: A premalignant lesion? A m J Pathol 1997, 8:323-327. 75 10. A P P E N D I X Table 1. L O H Frequencies (%) in Head and Neck and Oral Cancers Chromo Oral Cancer Head and Neck Cancers some arm Wu Uzawa Roz Ishwad Ah- Nawroz Adamso Field El-1994 1996 1996 1996 See 1994 n 1994 1995 Maestr Naggar 1994 oe 1995 1996 IP * 14 30 27 lq 0 23 28 2p 0 19 28 2q 8 15 13 3P 52 50 58 44 67 52 47 3q 0 50 13 4p 12 38 8 4q 11 37 13 5p 11 19 17 5q 43 25 29 6p 7 0 38 21 6q 5 23 25 7p 8 23 8 7q 11 0 29 7 8 P 10 40 35 53 8q 7 38 21 9p 48 24 72 62 72 9q 35 13 20 35 76 iop 0 23 9 lOq 0 21 13 l i p 5 17 13 l l q 56 6 45 61 23 33 12p 7 18 14 12q 12 25 15 13p 13q 0 54 27 67 14p 14q 5 39 11 15p 15q 6 5 12 16p 0 10 13 16q 0 10 13 17p 31 52 50 50 17q 19 31 30 18p 0 27 16 18q 0 23 49 19p 0 32 0 19q 0 40 29 20p 0 30 0 20q 7 9 9 21p 21q 5 26 8 22p 22q 16 29 0 *Blanks indicate information not available. 77 Table 6 The time between initial biopsy and current date /the date on which a lesion has progressed to SCC Groups Case# Arms with L O H Arms with Nl # arms lost Time between initial biopsy and current date/SCC (mo.) Without progression set Mild dysplasia 108 4q 0 65 180 0 122 184 0 161 186 3p,9p,17p,llq 4 164 187 9p,4q 2 183 188 9p 1 123 191 3p,9p,llq 3 155 201 9p 1 145 206 0 130 207 0 112 209 9p 1 159 233 0 103 239 0 28 240 9p 1 145 241 9p 4q 1 127 242 l l q 9p 1 122 246 3p,Hq mmmm 135 248 0 112 256 0 112 258 3p,9p,17p 3 103 270 8p 1 110 78 277 17p,4q 3p,llq 2 147 12 1 65 21 0 66 26 0 65 51 0 52 55 3p 1 52 60 8p 0 54 64 0 41 88 9p,17p,8p,13q 3p 4 42 92 9p 8p,13q 1 40 Moderate 143 3p,9p,17p,8p,13q 5 37 dysplasia 163 3p,9p,17p,llq 4 37 167 9p 1 141 168 3p,9p,17p,8p l l q 4 62 169 3p,9p 2 142 176 0 164 177 3p,9p,17p,8p 4 32 189 9p,4q,8p 3 170 190 3p,9p,17p,8p 4q 4 145 214 3p,9p 2 105 232 9p 1 147 234 H H H H H 1 109 . . . ................. 243 0 104 257 8p 0 110 274 0 Z o 22 9p,17p 2 65 25 4q 1 64 79 32 3p,9p,17p,llq 4q 4 44 34 8p 1 47 66 0 47 71 4q,9p 0 59 76 0 66 Hyperplasia 101 0 41 102 4q 0 41 103 0 41 111 8p,Hq 0 41 127 3p 0 35 129 3p 1 37 130 3p 1 37 131 13q 1 36 132 0 36 133 0 37 134 3p 0 35 135 0 38 136 0 35 137 0 36 138 3p 41 139 0 41 140 0 37 141 0 36 142 0 42 144 0 36 146 9p 1 37 148 l l q 1 36 80 149 13q 0 36 150 4q 0 37 151 mmm 37 152 0 36 171 3p,8p,llq 0 34 172 9p 0 34 179 0 34 183 0 34 11 0 43 23 0 42 28 0 43 With progression set Mild and 173 3p, 9p, 8p,llq,13q 5 16 moderate dysplasia 122 3p,9p,17p,8p,llq 5 99 223 1 51 271 3p,9p,17p,llq,13q 8p 5 10 293 3p,9p,17p 3 26 399 9p,17p 2 10 365 9p 1 26 446 9p,8p 2 60 472 3p,9p, 4q, 8p, l l q 5 85 185 9p,4q 2 83 197 3p,9p 4q,13q 2 7 377 3p,9p 2 70 473 3p,8p,llq 3 11 81 245 3p,9p 17p 2 56 281 3p,9p,17p,4q,8p,llq 6 52 401 9p,17p,4q,8p,llq,13q 3p 6 15 286 3p,17p,8p,13q 4 11 460 3p,9p,4q,llq,13q 5 46 79 3p,9p,17p,4q 13q 4 29 464 3p,17p,8p 3 27 121 9p,8p o L 29 470 8p,llq,13q 3 6 95 9p,13q 4q 2 32 Hyperplasia 405 9p 13q 1 7 385 1 16 406 9p,17p,llq 3 12 416 HHHHHi 13q 1 72 471 3p,4q,8p 3 105 363 3p,9p,17p,4q,8p,llq 6 76 82 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0089375/manifest

Comment

Related Items