UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Expression of glutathione S-transferase pi in human salivary gland and salivary gland tumours Zieper, Monica B. 1995

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

Item Metadata

Download

Media
831-ubc_1995-0436.pdf [ 11.1MB ]
Metadata
JSON: 831-1.0086945.json
JSON-LD: 831-1.0086945-ld.json
RDF/XML (Pretty): 831-1.0086945-rdf.xml
RDF/JSON: 831-1.0086945-rdf.json
Turtle: 831-1.0086945-turtle.txt
N-Triples: 831-1.0086945-rdf-ntriples.txt
Original Record: 831-1.0086945-source.json
Full Text
831-1.0086945-fulltext.txt
Citation
831-1.0086945.ris

Full Text

EXPRESSION OF GLUTATHIONE S-TRANSFERASE PI IN HUMAN SALIVARY GLAND AND SALIVARY GLAND TUMOURS by MONICA B. ZIEPER D.M.D. The University of Manitoba, 1986 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE IN DENTAL SCIENCES in THE FACULTY OF GRADUATE STUDIES FACULTYoOF  DENTISTRY  We accept this thesis as conforming to the required standard  THE UNIVERSITY OF BRITISH COLUMBIA March 1995 © Monica B. Zieper, 1995  In presenting degree at the  this thesis  in partial fulfilment  requirements  University of British Columbia, I agree that the  freely available for reference or  by  his  or  for an advanced  Library shall make it  and study. I further agree that permission for extensive  copying of this thesis for scholarly purposes department  of the  her  may be granted  representatives!  It  is  by the  understood  that  head  of my  copying  or  publication of this thesis for financial gain shall not be allowed without my written permission.  O^J.  Department of  -?r\juy(s~<-*-P / flo^t-oJ^^CAJL^^-O  The University of British Columbia Vancouver, Canada  Date  DE-6 (2/88)  ^ ~-  <  UU  JL  7^ *^'^ x  j  ABSTRACT  The studies of tumour markers, such as enzymatic and genetic markers, are of considerable importance in understanding the mechanisms of tumourigenesis, and in the identification, prevention, and possible treatment of tumours. Glutathione S-transferases (GSTs) are a family of multifunctional enzymes which play important roles in the detoxification of xenobiotics, including carcinogens. Placental G S T (GST-JC), an acidic form of G S T , is normally present in epidermis and oral mucosa where there is constant exposure to xenobiotics. G S T - J C is markedly increased during neoplastic processes in a number of tissues and is an excellent tumour marker in these tissues. There is a lack of information on the normal distribution of G S T - 7 t in human salivary glands and salivary gland neoplasms. This study investigated, immunohistochemically, the normal distribution of G S T JC in human major and minor salivary glands and salivary gland neoplasms. The results showed that, in the normal and inflamed salivary glands of all locations (including major and minor salivary glands), the ductal epithelial cells showed moderate to strong GST-TT. staining, myoepithelial cells showed weak staining, and the acinar cells were negative. The staining was mainly cytoplasmic but small areas of nuclear staining were also noted. The staining pattern of the tumour cells was similar to that of their normal counterparts. The tumour cells were generally positive with GST-JC  staining except those tumour cells demonstrating acinic differentiation (serous  cells in acinic cell carcinoma, mucous cells in mucinous cystadenoma and mucoepidermoid carcinomas). Most of the salivary gland tumours, benign or malignant, showed a weak GST-TC staining. Only mucoepidermoid carcinomas demonstrated significantly increased GST-JC reactivity compared to other tumours (p<0.001), which may be a reflection of both malignancy and squamous differentiation.  iii The marked increase in GST-JC activity in mucoepidermoid carcinomas may be useful in serological screening of recurrent and metastatic mucoepidermoid carcinomas. Pleomorphic adenoma (PA) is characterized by relatively high recurrence rate. The most important reason for the recurrence of PAs is believed to be incomplete first removal of the tumours. It is not clear whether biological behaviour of the tumour cells also contributes to recurrence of this tumour. In the second experiment, GST-rc content was compared between non-recurrent PAs and recurrent PAs using a computerized image analysis system. The results showed a significantly higher G S T - J I content in recurrent PAs when compared to nonrecurrent PAs (p<10 ). This study has -8  demonstrated that intrinsic cell behaviour (such as increased GST-rc content) may play a role in the recurrence of PAs.  IV  TABLE OF CONTENTS Page number  Abstract  II  Table of Contents  iv  Acknowledgements  ix  Chapter One Carcinogenesis 1.1  Initiation, promotion, progression and molecular events  1.2  Role of free radicals in carcinogenesis  1.3  Tumour markers  Chapter Two Glutathione S-transferase (GST)  2.8  1 2  Q 1  9  1 1  2.1  Classification and Nomenclature  12  2.2  Structure  -j  4  2.3  Enzymatic mechanisms  1  7  2.4  Metabolic roles  1  9  2.5.  Functions  23  2.51  GST and cellular detoxication  23  2.52  GST function in the cell nucleus  26  2.53  GST and drug resistance  27  2.6  Induction and regulation of GSTs  3  2.7  Tissue distribution of GSTs in normal human tissues  3  4  2.71  Distribution in glandular tissues  4  7  2.72  Distribution in salivary glands  4  7  GST as a marker in neoplasia and preneoplasia  4  8  2.81  4  2  GST and its use in clinical application - serum  1  analysis 2.82  Distribution of GST-TC in human tumours and human tumour cell lines  46  2.9  Sex differences in GST  Chapter Three Human Salivary Gland Tumours  46 50  3.1  Introduction  3.2  Etiology  52  3.3  Classification of salivary gland tumours  55  3.31  3.32  3.4  51  Benign salivary gland tumours  55  3.311 Pleomorphic adenoma  55  3.312 Monomorphic adenoma  57  Malignant salivary gland tumours  58  3.321 Acinic cell carcinoma  58  3.322 Mucoepidermoid carcinoma  60  3.323 Adenoid cystic carcinoma  61  3.324 Malignant mixed tumour  62  Therapy and prognosis  Chapter Four EXPERIMENT #1 The expression of placental  63 68  glutathione S-transferase in normal human salivary gland and salivary gland tumours. Objective and Hypothesis  69  Materials and methods  70  4.21  71  GST-re immunohistochemistry  73  Results 4.31  Normal human salivary gland  73  4.32  Human salivary gland tumours  73  4.321 Monomorphic adenoma  74  4.322 Pleomorphic adenoma  74  4.323 Acinic cell tumour  75  4.324 Adenoid cystic carcinoma  75  vi  4.325 Mucoepidermoid carcinoma  75  Discussion and conclusions  76  Chapter Five EXPERIMENT #2 Quantification of  82  4.4  immunohistochemical reaction of GST-TC in primary and recurrent pleomorphic adenoma. 5.1  Objective and hypothesis  84  5.2  Materials and methods  84  5.21  85  Computer image analysis  5.3  Results  86  5.4  Discussion and conclusions  87  Bibliography Appendix 1. Figures and Tables  92 114  Figures 1  Schematic diagram of cellular protection against  115  oxygen-mediated toxicity. 2  Schematic illustration of multiple derangements of cell  116  metabolism that can be brought about by oxidative stress. 3  Overall pathway of glutathione (GSH) metabolism.  117  4  X-ray crystallographic analysis of pig glutathione S-  118  transferase pi/glutathione sulfonate co-crystal - Stereo drawing of dimeric glutathione S-transferase molecule with the inhibitor glutathione sulfonate included. 5  Ribbon representation of GST crystal structures of pi  119  class human GSTs and pi class dinners. 6  Mean optical density of primary vs. recurrent pleomorphic adenoma.  120  Vll  7  Scatter diagram of mean optical density of primary vs.  121  recurrent pleomorphic adenoma. 8  Skewness of mean optical density of primary vs.  122  recurrent pleomorphic adenoma. 9  Kurtosis of mean optical density of primary vs. recurrent  123  pleomorphic adenoma. 10  Optical density coefficient of variation in primary vs.  124  recurrent pleomorphic adenoma. Tables 1  Species-independant classification of rat, human, and  125  mouse cytosolic glutathione transferase subunits. 2  Nomenclature and classification of glutathione S-  126  transferases. 3  Immunohistochemical detection of GST-re in normal  127  tissues. 4  Histologic classification of salivary gland tumours.'  128  5  GST-re staining intensity in human salivary gland  129  tumours. 6  Comparison of GST-re staining in cell-rich and cell poor  129  pleomorphic adenomas 7  T N M Classification of anatomical extent of disease of  130  salivary gland tumours Appendix 2. Plates  131  Legend to Plate 1.  132  Plate 1.  133  A. B. C. D. E. F. G. H  Legend to Plate 2.  134  Plate 2.  135  A. B. C. D. E. F. G  viii Legend to Plate 3.  136  Plate 3.  137  A. B. C. D. E. F  Appendix 3.  138  Immunohistochemical Method-GST-7C (ABC)  139  Statistical Analyses  143  1. GST-7C in salivary gland tumours  144  Raw Data  144  One Way Analysis of Variance  145  Student-Newmann-Keuls Multiple  146  Comparison Test 2. Sex differences in GST-TC  147  Raw Data  147  Unpaired t-test  148  3. GST-TC in Cell-rich vs. Cell-poor PA  149  4. Image Analysis - Primary vs. Recurrent PA  150  Data Summary - Primary vs. Recurrent PA  150  Data Summary - Primary vs. Recurrent PA ,  151  Mann-Whitney Test  152  Raw Data - Primary PA  153  Raw Data - Recurrent PA  157  ACKNOWLEDGEMENTS  The completion of this thesis would have been very difficult without the assistance of a number of people. I wish to take this opportunity to thank those who have been instrumental in bringing it to fruition: Dr. S.S. Tzang, and Dr. B.Wu, committee members, who provided valuable advice throughout my research, Dr. Y. Xiao, who provided essential technical advice and assistance, Dr. A.K. Nithyanand, for sharing his experience and his computer tips, Dr. N. Poulin, for his patience and perserverence in overseeing the image analyses, Dr. R.W. Priddy, for his continuous support and gentle "nudging", and Dr. L. Zhang, my supervisor, without whose valuable suggestions and research expertise this study could not have been completed. And, of course, thanks Pete, for all of the extra time in dealing with the day to day necessities and my mental health.  Chapter 1. CARCINOGENESIS  1.1  INITIATION, PROMOTION, PROGRESSION AND MOLECULAR  EVENTS  Carcinogenesis is generally believed to be a multi-stage process, consisting of the distinct and sequential stages of initiation, promotion, and progression. Initiation refers to the permanent change in DNA of one or a few cells that have been exposed to a carcinogen at a level that is insufficient to cause a neoplasm (Boutewell, 1989). DNA synthesis is required for the fixation of a mutant gene in daughter cells and provides for the irreversible nature of the initiated cells (Weiner & Cance, 1994). Promotion has been described as reversible clonal expansion of previously initiated cells to grow faster than the surrounding normal cells and to develop into a visible neoplasm (Boutewell, 1989;' Pitot et al., 1989). Both genetic and epigenetic mechanisms are believed to be involved in promotion. These changes result in a monoclonal population of cells which may successfully evolve and evade normal growth and/or developmental control. Progression is believed to be characterized primarily by its karyotypic instability and evolution to malignancy. Some prefer to define progression as a series of qualitative, heritable changes in a subpopulation of initiated cells, resulting in malignancy or an increased potential to progress to malignancy (Sudlinovsky et al., 1991). The development of irreversible, aneuploid malignant neoplasms distinguishes progression from both initiation and promotion. Further selection processes or genetic changes may then result in the development of a more heterogeneous cell population with variable phenotypic properties. They may also provide these cells with the ability to evade normal host immune surveillance and the ability to develop and establish tumour progenitor cells at metastatic sites. The 3-stage process of initiation, promotion, and progression was discovered and defined originally in the mouse skin carcinogenesis model using  chemical carcinogens. Among known carcinogens, the polycyclic aromatic hydrocarbon (PAH) group of carcinogens are most extensively studied as they are the most common environmental carcinogens causing human cancers, including head and neck cancers (Dipple, 1985). In studies using PAH, mouse skin was 'initialized' by topical treatment with a subthreshold dose of a carcingen, an initiator (Periano & Jones, 1989). When the animals received no subsequent treatment, no tumours developed. When the pre-treated mouse skin was subsequently treated with regular, periodic applications of croton oil, a promotor which is not carcinogenic by itself, numerous tumours formed. The number of tumours which formed were dependant upon the dosage of the carcinogen received. The interval period between the applications of a carcinogenic initiator and promoter did not seem to affect the tumour yield, even when the interval was as long as one year (Periano & Jones, 1989). For a carcinogen of any kind to cause cancer in human or animal, it must interact with cellular macromolecules including proteins, lipids and most importantly, with nucleic acids, the genetic material of cells. Some carcinogenic compounds have structures permitting direct reactions with cellular constituents, with no need for metabolic activation, and are called direct acting carcinogens. Many of the direct acting carcinogens lead to alkylation of cellular macromolecules, often at the 7-position of guanine in nucleic acids. Others require metabolic activation to form reactive intermediates called indirect carcinogens, or pre-, or procarcinogens which can then react with cellular constituents. Subsequently these compounds (whether direct reacting or indirect reacting) are metabolized to form further intermediate reactive compounds and go on to bind with cellular receptors or celluar constituents. Alternatively, they may be deactivated by reaction with water or by metabolic conversion. Various enzymes that are involved in metabolic activation and deactivation of carcinogens, therefore, play important roles in  carcinogenesis. Many factors including species, strain, sex, age, diet, immune status, and enzyme inducers and inhibitors can affect the metabolism of these direct or indirect acting carcinogens. For example, a procarcinogen that is carcinogenic in one species may be non-carcinogenic in another species which lacks the necessary enzyme systems to activate the procarcinogen. Individuals lacking the metabolic systems to detoxify carcinogens may be at a greater risk for cancer development. Sex differences have been found to alter the response to xenobiotics, including carcinogens, and in some cases the different responses between genders correlate with differing levels of enzymes in males and females. These factors affecting the carcinogenic potential of compounds are varied and often interrelated (Walkes& Ward, 1994). Preneoplastic and neoplastic cells developed during multistage carcinogenesis are characterized by various genetic and phenotypic changes. However, one common feature shared by all tumour cells is loss of normal growth i  control. The molecular mechanisms of multistage carcinogenesis are still unclear. Cells under normal circumstances evolve numerous pathways whereby cell division, differentiation, adhesion, migration, cell senescence, and death proceed under regulated, processed signals. These processes are the means by which crucial cellular functions are carried out. Inter and intracellular cummunicafion proceeds at multiple levels, regulated by normal active genes, however, should these genes become altered or disrupted, derangement of normal cellular function or growth occurs. It is currently believed that different carcinogens, whether chemical, viral or radiation, cause genetic alterations in a number of critical control genes which are intimately involved in the signal pathways of cell growth and differentiation. Two distinct classes of genes, believed to be associated with cancer  development, are oncogenes and tumour suppressor genes. Activation of oncogenes and deactivation of tumour suppressor genes are believed to be responsible for the loss of normal growth control in tumour cells (Weiner & Cance, 1994). Oncogenes were originally discovered through the study of genetic transduction caused by acutely transforming retroviruses. Transduction is a phenomenon in which DNA is carried from one organism to another by bacteriophage thereby causing genetic change. In the studies of retroviruses, following exposure, transduction lead rapidly to sarcomas in innoculated animals. Oncogenes, and their normal cellular counterparts, proto-oncogenes, have been linked to growth factors, growth factor receptors, protein kinases, regulatory proteins, and transduction pathways (Sudilovsky et al., 1991; Waalkes & Ward, 1994). The cellular proto-oncogenes may be converted to oncogenes by a number of mechanisms including retroviral insertion, mutation, gene amplification, and chromosomal translocation. Chemical carcinogens, in turn, may induce point mutations and gene amplification leading to activation of the cellular protooncogene. Thus proto-oncogenes are normal cellular genes while oncogenes are activated tumour genes which act as positive proliferative signals to neoplastic cells. The earliest identified oncogenes were the H- and K-ras oncogenes (Harvey or Kirsten rat sarcoma) (Barabacid, 1987) and a number of others have subsequently been identified (Waalkes & Ward, 1994). Antioncogenes, or tumour suppressor genes, are genes which under normal circumstances suppress tumour formation. When these genes are inactivated through mutation or are otherwise altered, the protective effects of these genes are lost and tumours may then develop (Sager 1985; Weinberg 1989). The mechanisms through which tumour suppressor genes effect their protection are not entirely clear, but suggested functions include regulation of expression and  function of oncogenes, negative regulation of cell growth through transcriptional and post-transcriptional control of normal gene expression, induction of terminal differentiation and cellular senescence, and production of growth regulatory substances (Sudlinovsky et al., 1991). Much of current research in carcinogenesis emphasizes the understanding of changes occuring during the multi-stage process, including molecular and enzymatic changes, and mechanisms related to these changes.  1.2  ROLE OF FREE RADICALS IN CARCINOGENESIS  A free radical is any species capable of independent existence that contains one or more unpaired electrons. An unpaired electron is one that is alone in an orbital, a region of space within an atom (Halliwell, 1991). Free radicals have been implicated in the pathology of a wide variety of human diseases. Human beings, being aerobic organisms, are susceptible to oxidative stress and damage from free radicals and oxygen species such as as hydrogen peroxide, superoxide anion radicals and hydroxyl radicals and others (Halliwell, 1991). Free radicals and other reactive oxygen species are continuosly formed in the human body both by deliberate synthesis and by chemical side-reactions from sources such as the mitochondrial electron trasport system, cellular peroxidases, monooxygenases, and the autooxidation of flavins, thiols, and lipids. They are usually removed by antioxidant defences (Figure 1). If the balance between antioxidant defences is lost, however, a complex multifactorial nature of oxidative damage which can alter cellular metabolism may result (Figure 2). Not only are reactive oxygen species produced endogenously, but exogenous sources of oxidants are known. Exogenous oxidative damage, or the  7  formation of free radicals, may be produced by xenobiotics such as carcinogens and cytotoxic drugs. Evidence is accumulating which suggests that free radicals are also involved in the multi-step pathway of carcinogenesis and play a role in initiation, promotion and progression. Certain initiatiors, including radiation and chemical carcinogens produce free radicals. Also, the induction of cancer has been found to be associated with oxygen species such as superoxide radicals and hydrogen peroxide . Certain carcinogens can generate free radicals which can cause single 1  or double stranded DNA breaks, altering cellular DNA, and can also eliminate free radical scavengers within a cell (Rice-Evans & Burdon, 1993). Free radicals are known to have a major influence on oncogenes and oncogenesis via a number of mechanisms (Sahu, 1990). They can induce DNA damage by direct oxidation or by indirect DNA-binding of products of lipid peroxidation. Oxygen radicals may induce conformational changes in plasma membranes via lipid peroxidation and protein degradation. These conformational changes, in turn, alter membrane associated cellular acfivites. Oxidative modification of cell lipids also has potential consequences for tumour cell proliferation (Rice-Evans & Burdon, 1993). Free radicals are capable of affecting membrane bound protein kinases, growth factors and their receptors, and consequently may alter signal transduction and cellular communication. Some growth factors, such as platelet-derived growth factor, activate at least two protooncogenes, c-myc and c-fos (Sahu, 1990). Lipid hydroperoxides may influence cellular regulatory proteins, but they may alternatively serve as precursors to the formation of prostaglandins with tumour inhibitory or immunosuppressive properties, and they may cause elevation of oils rich in n-3 polyunsaturated fats which have toxic effects toward tumour cells (Rice-Evans & Burdon, 1993). Although it is felt that the presence of inflammatory cells aid in combatting  tumour development, there is also the theory that inflammation can contribute to tumour promotion and that immune lymphocytes that recognize tumour cells can enhance the colonization of metastatic sites. This is thought to occur due to the release of hydrogen peroxide and superoxide by leukocytes, providing a selective advantage to tumour cells (Rice-Evans & Burdon, 1991). Antioxidants are those substances that, when present in low concentrations compared to the concentration of oxidizable substrates, significantly delay or prevent oxidation of that substrate (Halliwell, 1991). These substances act as antioxidants by scavenging reactive oxygen species, by preventing their formation, or by repairing damage caused by them. Specific mechanisms of antioxidant activity have been theorized including the direct interaction of a carcinogen or its metabolic products with the antioxidant, competition with or blockage of activation of procarcinogens to reactive intermediates, increased efficiency of repair of damaged DNA, and enhanced activities of enzymes which inactivate carcinogens thus limiting cellular transformation (Taylor & Scanlan, 1989). Under normal circumstances, the cellular antioxidant defense system maintains an appropriate balance between necessary oxidative events and those that may be excessive. When this balance is not maintained, an overloading of the cellular redox system occurs and oxygen radicals may then cause cellular damage. A number of naturally occuring antioxidants including vitamin E, vitamin C, beta-carotene, selenium, and synthetic antioxidants, such as 2-(3)-tert-butyl-4-hydroxyanisole (BHA), have been shown to have anticarcinogenic properties (Rice-Evans & Burdon, 1993). The actions of both natural and synthetic antioxidants as modulators of carcinogenesis may be similar. Both endogenous and exogenous reactive oxygen species can be removed by enzymatic and non-enzymatic antioxidant defences. The various mechanisms by which tissues, organs, and cells are equipped  to protect themselves from damages of free radicals include structural integrity of cells, compartmentalization of constituents and functions of cells, usage of antioxidants and the presences of various protective enzymes. The interdependance and interaction of a number of pathways is necessary in providing and maintaining a non toxic, protected balance within and surrounding cells. A number of complex interactions involving enzymatic redox reactions play important roles in maintaining a nontoxic and protected cellular environment. Among these enzymes are an important group of enzymes which function to inactivate pro-oxidants and also metabolize xenobiotics, including carcinogens and cytotoxic drugs, into inactive products. This group of enzymes includes catalase, superoxide dismutase, glutathione peroxidase, DT-diaphorase, gammaglutamyl transpeptidase and the topic of this thesis, glutathione S-transferases (GSTs). As can be seen in Figures 1., these enzymes are highly interdependant. Induction of glutathione transferases and other enzymes in these pathways play a major role in maintaining cellular stability.  1.3  TUMOUR MARKERS  Tumour markers have been defined as specific changes produced in a host suffering neoplasia (Beer & Pitot 1987). These changes consist of specific genetic and phenotypic characteristics, and may include both acquisition of new features or loss of normally existent characteristics (Miller & Miller,1974). Examples of tumour markers include oncogenes, such as c-myc, mutant tumour suppressor genes, such as p53, oncoproteins, and a variety of enzymes and growth factors. The study of tumour markers provides a wealth of information. Markers can provide visual 'labelling' of cells which allows one to study and follow these cells. Identification of tumour markers can provide information on metabolic events, on  the status of cellular differentiation, proliferation, and thus on functional status. Immunohistochemical demonstration of enzymatic or protein tumour markers can be carried out on paraffin embedded tissue sections and can be used in most medical laboratories as routine staining techniques. Understanding of key enzymatic changes in both normal and abnormal cells and tissues may lead to prevention and intervention processes in the treatment of preneoplasia and neoplasia. Early intervention in any form of neoplastic or particularly preneoplastic changes ultimately leads to more positive prognostic value. The study and diagnosis of preneoplastic and neoplastic changes requires recognizable changes indicative of cellular events. In the long process of neoplastic transformation, currently recognizable histomorphologic changes indicative of carcinogen-altered cells occur very late in the process. Presently, histomorphological changes are the primary or the only method used for diagnosis of premalignant and malignant lesions, for assessment of malignant potential of premalignant lesions, and for prediction of prognosis of a tumour. It is well known, however, that morphologically similar lesions may have different clinical behavior, different malignant potential, and vastly different prognoses. Establishment of tumour markers may provide additional tools in identification of carcinogen-altered cells prior to the conventional histomorphological changes. Identification of these morphologically similar but behaviourly disimilar lesions is highly desired. Furthermore, studies of tumour markers are of considerable importance in understanding of the mechanisms of tumourigenesis as they may provide insight into cellular events which occur during neoplasia. The understanding of mechanisms of tumourignesis may in turn provide means in prevention and possible treatment intervention of tumours.  Chapter 2. GLUTATHIONE S-TRANSFERASE  (GST)  2.1  CLASSIFICATION AND  NOMENCLATURE  The family of GST enzymes has been divided into different classes according to their sequence homology, isoelectric point, and substrate preference (Mannervik, 1985; Mannervik & Danielson 1987; Di llio, 1987) These classifications have also been determined by biochemical, biophysical, immunologic and genetic differences between GSTs. According to their cytological location, the mammalian GSTs are divided into membrane-bound microsomal and cytosolic GSTs. The membrane-bound microsomal group of GSTs shows no sequence homology to the soluble group of the enzymes. They are bound to membrane of endoplasmic reticulum and function as structurally different entities. They form trimers as opposed to dimers formed by the cytosolic group. The cytosolic group of GSTs are classified into alpha (basic), mu (neutral), pi (acidic) and theta forms according to the species indepdendent classification (Ketterer et al., 1992; Mannervik et al., 1985). As a group, the cytosolic enzymes are abundant and together may constitute up to 5% of soluble cellular protein (Waxman, 1990). Recently an enzyme involved in the sythesis of leukotrienes, human leukotriene C4 synthetase, has been identified as belonging to a separate and new class of GSTs (Nicholson et al., 1993). As this thesis investigates GST-JC, emphasis will be given to the pi class in the review of GSTs. Many studies have investigated genes coding individual GST isoenzymes. The rat glutathione S-transferases and their genes have been most completely studied. Five alpha, five mu, one pi, and two theta class of GST subunits have been identified with genes coding these subunits identified. For rat GST-P, one functional gene coding and four pseudogenes have been found (Casenave et al., 1989). Pseudogenes are those DNA sequences that are non-transcribed but bear striking homologies to a structural gene sequence. Psuedogenes may be on  different chromosomes or closely linked to the functional gene and may occur in varying numbers (Shows et al., 1987). The function of these pseudogenes is not clear; however, it is of interest that sequences which are not transcribed have been identified , yet if they were, would provide additional proteins very similar to these groups of enzymes. Similar to rat, multiple forms of human GST isoenzymes have been identified and these isoenzymes are encoded by at least six separate genetic loci. These include A1 and A2 (alpha class), M1, M2, M3, M4, M5 (mu class), and P1 (pi class (Casenave et al., 1989; Morrow et al., 1990; Board et al., 1989; Wilce & Parker, 1994). It is thought that GSTs evolved from a common gene, with the exception of the microsomal GST which shows no obvious sequence relationship to the cytosolic enzymes (Morgenstern et al., 1985). Classification of GSTs has been somewhat confusing. Mannervik, et al. in 1985 demonstrated that the major isozymes of cytosolic glutathione transferase from rat, mouse, and human share structural and catalytic properties which were used as the basis of a species-independant classification system in which GSTs were classified according to n-terminal amino acid sequence, substrate specificity, sensitivity to inhibitors, and immunologic cross reactivity. This classification scheme, over the following few years was further refined, and accepted as outlined in Table 1. The current classification system for human GSTs has been proposed based upon an international system of human gene nomenclature devised in 1987 (Shows et al., 1987). This classification with gene locus and chromosome designation as seen in Table 2, summarizes the human glutathione S-transferases (Table 2, Ketterer et al., 1992; Shows et al., 1987). Other species' enzymes are also being re-classified with a similar type of nomenclature. As these enzymes consist of dimers with two distinct subunits, each subunit  is represented in the enzyme nomenclature. Human cytosolic GST enzymes are all homodimers represented by identical subunits comprising the enzyme. For example, GSTA1 consists of a homodimer and two subunits 1-1. In one of the rat alpha species, however, a heterodimer with disimilar subunits exists as alpha 1-2, orGSTA1-2. The different genes which code for proteins with similar functions are designated by an Arabic numeral, following the characters used for the protein symbol. Hence, GSTA1 distinguishes 1 as the gene locus for this alpha protein while GSTA2 designates a 2 as the gene locus.  2.2  STRUCTURE  The glutathione S-transferases interact with cellular glutathione, a major low molecular weight peptide in cells. Glutathione is also known as gamma-glutamylcysteinylglycine. This molecule contains three amino acids, glutamine, glycine and cysteine, which may all be involved in the metabolism of foreign compounds. It is the nucleophilicity of the sulfur atom of the cysteine residue of glutathione and the antioxidant properties of glutathione that result in the detoxication of xenobiotics and oxidants by this tripeptide (Dekant & Vamavakas, 1993). The biochemical structure of cellular glutathione is as follows:  HS•CH2 • C H • C O • N H • C H 2 • C O O H l NH • C O • CH2 • CH2 • CH • COOH  I  NH2  It is the amino acid cysteine, SHCH2-CH(NH2)COOH, that is involved in  mercapturic acid biosynthesis and has a major role in detoxication, particularly the reactive thiol moeity, that is, the monovalent radical -SH. Until recently, little was known about the structure of GST enzymes. It has been found that the native state of these enzymes is almost always as a dimer, with individual subunits of these dinners functioning independantly. The subunits of these dimers are approximately 24-28 kDa each. They dimerize to form a combination of two like molecules by non-covalent interactions and each subunit of the heterodimer is kinetically independant from the other subunit. The individual subunits have been shown to dimerize only with subunits of a similar class of the enzyme. Full length cDNA encoding of mammalian, cytosolic glutathione Stransferases have been isolated and sequenced for some of the known classes (Manoharan et al., 1992). It has been determined that the open reading frames of the cDNAs encode peptides which range from 208 amino acids among pi class GSTs to 221 amino acids among the alpha class. The first pi class GST to be crystallized was from bovine placenta in 1988 (Schaffer et al., 1988). Subsequently the same group purified and crystallized a pi class enzyme from pig lung (Dirr et al., 1991). It was these crystals that lead to the first determination of the three dimensional structure of a GST in 1991 (Reinemer et al., 1991). The three dimensional structure of a pi class glutathione S-transferase was solved by x-ray crystallographic analysis of a pig glutathione S-transferase pi/glutathione sulfonate co-crystal, shown in Figure 4. The analysis by Reinemer has shown that this pi class GST has a globular shape with molecular dimensions of approximately 55Ax52A x45A. Isolated individual subunits have accessible surface area of 10345A while the accessible surface area for a subunit in a dimer 2  structure is less than the sum of the individual subunits following their joining at 8975A . A prominant feature of the dimeric structure is a large cavity or cleft formed 2  between the two subunits which restricts the accessible surface area upon dimerization. This particular pig form of the pi class is a homodimeric enzyme containing 209 amino acids. The structure of this enzyme and its polypeptide fold is archetypical for the whole family of cytosolic GSTs. Each GST monomer of this pi form contains two domains, a small domain domain I, with 76 amino acids, numbered 1 to 76 and a large domain - domain II, with 126 amino acids, numbered 83 to 209 (Reinemer et al., 1991; Wilce & Parker, 1994). A ribbon representation of this enzyme is seen in Figure 5. The portion of the enzyme which binds to glutathione, the G-site, is contained wholly within the first domain of the protein (Reinemer et al., 1991; Gulick et al, 1992) and is situated within the cleft formed by dimerization (Mannervik et al., 1985). This glutathione binding site is very specific. Unfolding studies have demonstrated the importance of the dimer structure for activity (Wilce & Parker, 1994). A detailed explanation of^* this polypeptide, including folding and sequencing has been summarized by Wilce and Parker, 1994. Monoclonal antibodies have been utilized to study the human glutathione Stransferase pi protein in a number of ways. Antibodies have been used as markers to determine localization of these enzymes in tissues and at the cellular level. They can be generated to bind to specific regions of protein antigens and thus are also useful as probes in determining structural and functional properties of protein molecules. Recently, the ability of a number of monoclonal antibodies to react with GST-TC  in either the native or denatured state has lead to supportive information  regarding overall conformation of the molecule, identification of the binding domain, as well as clarification of the catalytic function of all or part of the glutathione molecules (Gulick et al.,1992). The active site, H-site, to which electrophilic substrates bind, has yet to be identified. Some studies have shown that it may be at the carboxyl terminal region  of the enzyme (Gulick et al., 1992). Reinemer et al. (1991) proposed that it is situated adjacent to the G-site, with three defined regions as possible binding sites: 1. a cavity in domain II, 2. a hydrophobic region in the cleft described above and adjacent to the G-site which would accommodate small molecules, or 3. the cavity or cleft formed between the subunits in a dimeric structure (Reinemer et al., 1991). The substrate binding site is much less specific than the glutathione binding site resulting in the ability of GSTs to react with a wide variety of toxic agents (Wilce & Parker, 1994). The amino acid sequence of five class pi isoenzymes have been established showing substantial structural relationships between the isozymes, with 76% of residues in the pi class fully conserved with positional identifies having approximately 82-85% similarity. The high degree of similarity does not exist between the different classes of glutathione S-transferases, with only approximately 34% of positional identities maintained between alpha and mu classes as compared to pi classes. Thus, this family of enzymes may be functionally related, but appear to differ in their primary structural relationship. Structural differences in G-sites of various isozymes is thought to determine their variable response to differing modifications in the glutathione molecule, while structural differences in the H-sites may impart differing specificity to various electrophilic substrates.  2.3  ENZYMATIC MECHANISM  GSTs catalyze the conjugation of cellular glutathione with physiological and xenobiotic electrophiles to yield products that are less reactive, more hydrophilic and more readily excreted (Manoharan et al., 1992). In mammalian cells, the  sulfhydryl groups of glutathione comprises approximately 10-20% of the nonprotein sulfhydryl groups within a cell. The S atom on glutathione covalently attaches to the electrophilic sites of xenobiotics. A generalized conjugation reaction of glutathione (GSH), catalyzed by glutathione S-transferases, with an electrophilic substrate (R-X), would be represented as follows:  R-X + GSH -> R-SG + H-X  An example of this type of conjugation reaction may be seen with nitrogen mustard as:  R-NR'-CH CH -CI + GSH -> R-NR'-CH CH -SG + HCI 2  2  2  2  In this reaction involving an alkylating agent, the reaction may occur via displacement of the chloride resulting in inactivation of the electrophilic mustard functionality (Waxman, 1990). Glutathione, in nucle'ophilic catalysis, reacts predominantly as an anionic thiolate, with most reactions catalyzed by glutathione S-transferases being nucleophilic substitutions. In these reactions the thiol of bound glutathione is activated by the transferases, although how they activate the thiol is poorly understood. This activation, however, facilitates the attack on electrophiles. Two mechanisms of activation have been proposed, either as a general base-catalysis in which the protein is thought to enhance the nucleophilicity of the thiol by providing a base at the active site of appropriate pKa to allow deprotonating of the thiol group. A second possiblity proposed is that pKa of the thiol moeity of the thiolate anion of bound glutathione is lowered following stabilization by the enzyme's active site (Reinemer et al., 1991; Jakoby, 1978; Mannervik & Danielson,  1988). Clear understanding of these possible mechanisms of activation and detailed understanding of structural and chemical events leading to catalysis are under further investigation. Direct evidence that increased intracellular levels of GST can protect cells from toxins was evidenced in studies by Manoharan et al. (1987). In this study, rat cDNA for GST-1 was transfected into COS or 10T1/2 cells. These transfected cells showed a 260 fold increase GST activity as well a resistance to benzo(a)pyrene anti-diol epoxide as compared to non-transfected cells. The GST enzymes are also able to catalyze a selenium independant glutathione peroxidase activity which leads to the detoxification of lipid and nucleic acid hydroperoxides. In these reactions the reduction of the hydroperoxide (ROOH) consumes two moles of glutathione and proceeds without incorporation of the glutathione into the final product.  R-OOH + 2GSH -> R-OH + GSSG + H 0 2  The hydroperoxides being reduced are formed during redox recycling of many drugs and other xenobiotics, but can also be generated during normal metabolism as by-products of cellular oxygen utilization.  2.4  METABOLIC ROLES  GSTs are homo and heterodimeric proteins which constitute a versatile, multi-functional family of enzymes. These enzymes are present in almost all aerobic cells, and have been found in plants, fish, insects, fungi, yeast, bacteria, animals, and humans (Wilce & Parker, 1994). The substrates or electrophilic  molecules with which these enzymes react are variable, however they are mostly large and hydrophobic. GST-re was initially purified from human term placenta by Guthenberg and Mannervik in 1981, and is thus often referred to as placental GST (Guthenberg & Mannervik, 1981). GSTs are involved in the cellular detoxification of xenobiotics and hydroperoxides and play important roles in mutagenesis and carcinogenesis. It is now clear that GSTs provides several detoxification mechanisms against cytotoxic drugs and carcinogens. The metabolic utilization of glutathione, thought to be the most important cellular antioxidant, follows several pathways including reactions catalyzed by the glutathione S-transferases along the mercapturate pathway. Glutathione is also a substrate of glutathione peroxidases which destroy hydrogen peroxide and organic peroxides. Glutathione also provides the reducing power needed for the conversion of dehydroascorbate to ascorbate as well as for the conversion of ribonucleotides to deoxyribonucleotides and is thus important in the synthesis and repair of DNA and for the folding of newly synthesized proteins. The various metabolic pathways involved in cellular protection and glutathione metabolism can be seen in Figure 3. They are strongly interconnected and the glutathione Stransferases have a prominant role in the mercapturate pathway. The role of glutathione S-transferases in carcinogenesis is related to their ability to conjugate cellular glutathione to different xenobiotics and therefore convert potential carcinogens to less reactive forms. Some of the electrophiles which they bind include alkyl and aryl halides, epoxides, quinones and activated alkenes. The conjugation reaction with compounds such as these proceeds as the first step in the mercapturic acid pathway. The mercapturic acid pathway involves cellular glutathione which is ubiquitous in nature. It ranges in concentration from 0.5 to 10mM and has two characteristic structural features: a sulfhydryl (SH) group and a gamma-glutamyl  linkage (see structure pg 17). These two characteristic features define the biological functions of this tripeptide. The conjugation of glutathione to electrophilic compound by GSTs occurs as the first step in this pathway which then proceeds in the conversion of glutathione S-conjugates to mercapturic acid by three enzymecatalyzed reactions: 1.  the removal of the gamma-glutamyl moeity, catalyzed by membrane bound GGT  2.  the removal of the glycine moeity of the S-conjugate of cys-gly, by a number of aminopeptidases or dipeptidases  3.  N-acetylation of the cysteine conjugate to yield the corresponding mercapturic acid, catalyzed by N-acetyltransferase  Mercapturic acids, the S-conjugates of N-acetylcysteine, may include a number of sulfur substituents including alkyl, aryl, or heterocyclic group with or without halogen, nitro, amino or other groups (Jakoby, 1980). It has become apparent that a single protein with binding properties of lipophilic compounds and the capacity for reversible binding with glutathione provides an extremely versatile means by which an organism may combat potentially toxic substances, either inhaled, ingested, or metabolically produced. Potent carcinogens including aflatoxin B-|, vinyl chloride.and 1,2-dibromoethane (a gasoline additive) form glutathione adducts and are excreted (Degen & Neumann, 1978; Vainio, 1978; Nachtomi, 1970). A number of useful drugs have been tested directly and found to be substrates for GSTs including the following (Jakoby, 1980): Benzo[a]pyrene 4,5-oxide Bromosufophthalein 1-Chloro-2,4-dinitrobenzene 1,2-Dichloro-4-nitrobnezene 1,2-Epoxy-3-(p-nitrophenoxyl)-propane  Ethacrynic acid Idomethane 2-Nitropropane Menaphthyl sulfate p-Nitrobenzyl chloride frans-4-Phenyl-3-buten-2-one Prostaglandin A i Other drugs have been found to be conjugates of glutathione transferases only after the oxidation by the cytochrome P-450 mixed function oxidases which are enzymes also found ubiquitously (Jakoby, 1980). In addition to their catalytic capabilites, transferases serve as carrier proteins for organic substances and endogenous compounds including steroids, prostaglandins and leukotrienes (Mannervik et al., 1985; Bogaard et al., 1992). This ligand binding function facilitates the intracellular transport of numerous compounds, both hydrophobic and amphipathic (of particular configuration), including bile salts, bilirubin, heme and steroids. Binding by these compounds often results in inhibition of the enzyme activity by the bound ligand. Hydrophobic toxins conjugated with glutathione which do not become involved in the detoxification pathways may instead become more water soluble and may therefore be readily excreted from cells. As well, cells may be somewhat protected from toxic substances due to isolation and sequestration of the toxic subsances following high affinity binding with glutathione. Thus, a number of means of protection are provided to cells in association with glutathione Stransferases separate from the conversion of substances to non-toxic or metabolized substrates. GSTs are also thought to be involved in the cellular response to oxidative stress, in particular, the alpha and pi forms of GST which have intrinsic peroxidase  activity. Certain transferases can catalyze a selenium-independant peroxidase activity with lipid and nucleic acid hydroperoxides as substrates. Others serve as true co-enzymes as in the isomerization of delta5-3-ketosteroids (Reinemer et al., 1991) . Thus, they may play a role in protection against oxidative stresses produced by ionizing radiation or other sources (Cholon et al., 1992; Biaglow et al., 1992) . The family of glutathione S-transferases are intimately related to gammaglutamyl transpeptidase (GGT), a group of enzymes involved in many aspects of functional metabolism which can catalyze three types of reactions; hydrolysis, transpeptidation and autotranspeptidation. The metabolic turnover of glutathione is closely related to the activity of GGT as this enzyme is the only known enzyme to catalyze the conversion of cellular glutathione thioethers to mercapturic acids. GGT also functions in the cellular transport of amino acids. A multitude of xenobiotics and constituents of normal food are excreted as mercapturic acids, although there is no evidence that mammals produce significant amounts of mercapturic acids from endogenous compounds (Jakoby, 1980).  2.5  2.51  FUNCTIONS  GST AND CELLULAR DETOXICATION  Human exposure to a vast number and variety of chemicals with immense structural diversity is virtually unavoidable. Potentially harmful and toxic substances may be present in water, air, food or may be administered as drugs. Metabolic transformation of chemicals by the body into polar, readily excretable metabolites facilitates their excretion. Alternative pathways, however, may metabolize chemicals to generate electrophilic compounds which may have high  reactive potential with various cellular macromolecules. Disregulation of and/or the consequences of interactions with cellular macromolecules may result in the phenomenon of toxicity, either cytotoxicity, immunotoxicity, mutagenicity or carcinogenicity. The principle defence mechanism of cells against a variety of chemical agents is the tripeptide glutathione. Glutathione (gamma-glutamylcysteinylglycine) is a major low molecular weight peptide in mammalian cells and participates in a variety of cellular reactions, as previously mentioned (Meisfer, 1994; Dekant &Vamavakas, 1993). It has the ability to neutralize chemical electrophiles and thus impede their reaction with cellular constituents. Should this protein be depleted in tissues, either by poor nutrition or by chemicals, the toxic effect of chemicals may be greatly exacerbated. In the search for anticarcinogenic treatment modalities, selection of therapeutic agents is often based upon an attempt to simulate the detoxication which results in conjugation of chemicals with glutathione. Electrophilic compounds may conjugate spontaneously to a degree with cellular glutathione, however conjugation is usually effected by the family of glutathione S-transferase enzymes. Further roles of glutathione involves its participation in vital defenses against reactive oxygen species. Glutathione, in conjugation with ascorbate and tocopherol, acts as the ultimate antioxidant and can be regenerated by glutathione reductase. Glutathione is also part of the glutathione peroxidase system. Peroxides and hydroperoxides may cause cellular lethality. This system provides protection of cells from peroxides as wellas protects biologic membranes from lipid peroxidation. Some of the glutathione transferases, in particular the alpha and pi forms of GST, have intrinsic peroxidase activity and may also play a role in protection against oxidative stresses produced by ionizing radiation or other sources besides those formed during redox cycling and normal cellular oxygen  25  utilization (Cholon et al., 1992; Biaglow et al., 1992). Radiation produces a number of damaging free radicals, products of free radicals, as well as peroxides, both organic and hydrogen peroxide-H202 (Biaglow et al., 1992). As well, superoxide is produced by radiation. Radiation lethality occurs when peroxide accumulates at a faster rate than the cell can reduce it to less reactive alcohols and water. The major pathway for peroxide reduction is glutathione peroxidase accounting for approximately 70% of inactivation of low levels of peroxide, with catalase accounting for the remaining inactivation. Cellular levels of glutathione in the reduced state is maintained by electrons obtained from the hexose monophosphate shunt and peroxidase and transferase activities are dependant upon glucose oxidation (Biaglow et al., 1992). When this shunt is absent, toxicity from hydroperoxides is increased. In addition to their catalytic functions, GSTs also play a role in detoxication by directly binding and sequestering ligands which are not substrates for conjugation reactions (Cazenave et al., 1989). In early investigations the rat liver isozyme initially purified was referred to as 'ligandin' due to its high affinity binding properties. GSTs reversibly bind and transport compounds including bilirubin, facilitating its transport through the cell, and also bind other bile salts. Spontaneous binding to copper and copper-derivatives also occurs. Other biological mediators, including steroids have shown strong binding affinity for GST as has leukotriene, an important mediator involved in immediate-type hypersensitivity reactions. This occurs via a conjugation reaction, or the direct coupling of glutathione to an endogenous compound leukotriene A4 which yields the paracrine hormone leukotriene C4. GST-mediated ligand binding is often associated with the inhibition of GST activity by the bound ligand and appears to facilitate the intracellular transport of the lipophilic compounds. Although glutathione conjugation has been identified as an important  26 detoxication reaction, a number of glutathione dependant bioactivation reactions resulting in the formation of toxic glutathione S-conjugates have been documented. Different types of compounds, haloalkanes, haloalkenes, hydroquinones, and aminophenol have been found to form toxic glutathione S-conjugates (Dekant & Vamavakas, 1993; Waxman, 1990). As well, the cytotoxicity of the antibiotic neocarzinostatin is greatly enhanced by thiols. Compounds such as those which are activated by glutathione or glutathione/GST and form toxic conjugates could conceivably be directed toward cells and used as cytotoxic or chemotherapeutic agents in cells with elevated levels of GSH or GSH/GST. Conversely, the relatively benign substances might also be transformed to toxic, potentially carcinogenic substances in rare instances, resulting in this enzyme system acting as a contributing factor to the process of carcinogenesis.  2.52  GST FUNCTION IN THE CELL NUCLEUS  The precise function(s) of glutathione S-transferases in the cell nucleus have only been postulated. Given their role when found in the cell cytoplasm, the most obvious extrapolation to their function in the cell nucleus is the biotransformation of electrophilic compounds which may have escaped detoxification in the cytoplasm, thereby preventing their interaction with either DNA or other molecules. In rat liver, GST-P is detectable immunohistochemically in the nuclei of preneoplastic lesions during early stages of chemical carcinogenesis. Tan et al. (1986) reported high GST-P reactivity toward thymine hydroperoxide and it has been suggested that it functions in detoxification or as a carrier protein. In normal rat and human liver, subunits of GST as well as GST-u, and G S T - J C have been detected in free and bound fractions from nuclei and may be of importance in  27  detoxification of DNA peroxides (Ketterer et al., 1988; Campbell et al., 1991). GSTs have been shown to a have possible protective role by inhibiting the formation of covalent bonding between DNA and benzo(a)pyrene in vivo and in vitro (Bennett et al., 1986). Another suggested role of GST is the possibility that they may play a role in regulation of nuclear function. A nonhistone protein has been identified as a glutathione S-transferase, and nuclear localization of subunits of GSTs in or adjacent to regions where transcription or processing of DNA takes place have been found. In particular, they have been found in association with small nuclear RNA's rich in uridylic acid which associate with proteins to form particles proposed to play a critical role in the maturation of hnRNA to mRNA (Bennett et al., 1986). Thus they may play a role in modulating gene expression.  2.53  G S T AND DRUG RESISTANCE  Chemotherapeutic agents are continually being developed and evaluated with respect to their efficacy in prolonging survival, and curing various human tumours. Several distinctive strategies are pursued when developing chemopreventive agents: 1. identification and validation of markers of premalignant, particularly early premalignant lesions, that can be used as endpoints for measuring chemopreventive activity as opposed to using cancers; 2. identifying and testing potential therapeutic agents on the basis of mechanism of action; 3. evaluating combination therapies to maximize efficacy and minimize toxicity, and; 4. identifying and ranking agents at each stage of development to control choice of the best agents and most effective use of available resources (Kelloff, 1994). Chemotherapeutic agents cure a minority of adult tumours, including Hodgkin's and non-Hodgkin's lymphoma, acute leukemia and teratoma,  and can cure the majority of childhood tumours with early application. However, in the majority of solid tumours there is less than 20% response to chemotherapy and although some tumours are considered curable they may subsequently relapse and show resistance to drug therapies (Hochhauser & Harris, 1991). This resistance may be inherent as a characteristic of particular cells, either in their original form or after having been altered, resulting in a stable change within the cell. Or, resistance may be rapidly induced following exposure to a particular drug. The mechanisms whereby drugs may be ineffective, including the development of drug resistance, must be considered relative to the pharmacokinetics of the drugs as well as the factors which will influence their effectiveness at the molecular level. Efficacy will be affected from the point of administration to the time of achieving cytocidal concentrations at the target site and of great importance, the cellular responses of tissues at the target site. Multidrug resistance (MDR) is the phenomenon whereby exposure to one drug induces cross resistance to a variety of agents of different chemical classes to which the cell has never been exposed (Hochhauser & Harris, 1991). The hallmark of MDR is the expression of p-glycoprotein, although drug resistance is usually mediated through more than one pathway. P-glycoprotein is a 170kD membrane bound glycoprotein believed to function as a drug efflux pump which aids in decreasing the accumulation of toxin in resistant cells. It has six hydrophobic domains and a tandemly duplicated ATP binding domain. Cloned p-glycoprotein nucleotide sequences bare strong resemblances to the well characterized bacterial transport systems which involve not only the membrane bound transport protein, but also periplasmic substratebinding proteins which function in delivering drug or substrate to the membrane bound transport protein. P-glycoprotein acts as a transmembrane exporter of drugs, although normal substrates have not yet been identified. In man, the MDR1  gene (multi drug resistance gene) or p-glycoprotein gene has been located on chromosome 7q21-31. Regulation of this gene or of p-glycoprotein is done by amplification, transcription and translation with relative contributions of each varying throughout the selection process (Hochhauser & Harris, 1991). With the finding of increased levels of GST-TC and the overexpression of pglycoprotein in a multidrug resistant breast cancer cell line, it has been postulated that GSTs may interact with p-glycoprotein as a mammalian drug or toxin transport system (Casenave et al., 1989). The interrelationship of different drug resistance mechanisms can not be denied. Very few studies have been undertaken which have attempted to evaluate more than one resistance mechanism at a time. In a study by Efferth et al. (1992), the expression of drug resistance markers pglycoprotein (P-170), glutathione S-transferase pi (GST TC), and DNA topoisomerase II (Topoll) were analysed in human kidney carcinoma cell lines, human hematologic malignancies, and human breast carcinomas. Co-expression of resistance markers was found as well as evidence that the proliferative activity of tumour cells plays a role in the expression of these markers as well as development of resistance to cytostatic drugs (Efferth et al, 1992). Similarly, a study by Volm et al. (1991) investigating the overexpression of p-glycoprotein and GST-TC  in resistant non-small cell lung carcinomas showed individually significant  levels of the proteins as well as a significant relationship between the overexpression of P-170 and GST-TC. In studies involving co-expression, or multiple pathways of drug resistance, if a dominant enzyme can be implicated in a particular reaction this may provide understanding of catalytic selectiveness of families of enzymes and their relative involvement in different tissue or tumour types. Conversely, in patients evaluated for various enzyme levels, if an enzyme implicated in detoxication of known carcinogens is found to be present in lower  than average levels, patients may be considered at risk for known forms of cancer if exposed to those particular carcinogens. Although it has been stated that the various mechanisms do not likely act totally independantly, contributions made by other enzymes may result in low overall activity, resulting in overall high risk. As well, with respect to drug resistance, the opposite may be of consideration. That is , if an individual enzyme is known to be a predominant catalyst to inactivation of a desired chemotherapeutic agent, the levels of enzyme for a particular individual under treatment may be useful for forcasting for potential inactivation of a chosen therapeutic agent, resulting in possible alteration in or choice of an alternative regimen. Several anticancer drugs may be inactivated by GST catalysis, including melphalan, chlorambucil and cyclophosphamide, with numerous reports citing increased GST levels with the onset of drug resistance (Hochauser & Harris, 1991). has been found to be overexpressed in a number of multidrug resistant cell  GST-JC  lines, particularly those resistant to adriamycin and ethacrynic acid, a substrate of GST-re. Although chemotherapy is not the primary treatment of choice for salivary gland tumours, it has been shown to be effective in treatment of metastatic lesions of individual classes of malignant salivary gland tumours (Regezi & Sciubba, 1993). Others have reported that chemotherapy has no role in the treatment of salivary gland neoplasms (Jones et al., 1993). For a number of tumours of other tissues, chemotherapy is of great use and ongoing investigations as to the role of GSTs as well as other modes of resistance continue. Other mechanisms of drug resistance include topoisomerase II enzymes (Topo II) which play a role in DNA replication, chromosome scaffold formation, and chromosomal segregation. They also have possible roles in recombination and gene transcription. This enzyme acts by producing breaks in single and double  strand DNA and attaching itself to the 5' end of the break, forming covalent enzyme DNA complexes. Some drugs such as doxorubicin interfere with topoisomerase II and stabilize the enzyme-DNA bond. This results in inhibition of ligation of the cleaved DNA strand. Studies on coexpression of resistance markers have found a tendency for co-expression of increased P-170 and G S T - J C in both human kidney carcinoma cell lines and breast carcinomas and increased P-170 with decreased Topo II in kidney carcinoma cell lines (Efferth et al., 1992). The relationship, if any, of GSTs to Topo II is not entirely clear, although the findings of GST in the nucleus, particularly in regions of transcription and maturation of HnRNA is of interest along with a possible interrelationship when co-expressed in certain cells. Multiple mechanisms of DNA repair including excision repair, glycosylases, error-free repair, recombination repair, and mismatch repair, all have roles in drug resistance. In carcinogenesis, drug resistance may be suggested by human repair deficiency diseases where lack of repair may correlate with sensitivity in cell lines. Conversely, repair and other mechanisms may impart more resistance on particular cells. Resistance is usually mediated through more than one pathway, however information with respect to the contribution of individual pathways will aid in understanding their individual roles as related to other modes of resistance, both molecular and kinetic, as well as provide information for planning of drug protocols.  2.6  INDUCTION AND REGULATION OF GLUTATHIONE S-TRANSFERASES  GST-JI  has been found to be present in stomach epithelium of an eighteen  week old fetus, indicating that the expression of this form in human stomach carcinoma may be oncofetal in nature (Sato, 1989). Early studies on the possible  32  activation of oncogenes on the expression of GST's have implicated the human ras oncogene in the regulation of GST expression. Studies by Li et al. (1988) involving transfection of rat liver epithelial cells with human ras expression vector showed a six-fold increase in the level of GST7-7 mRNA expressed. Similar findings were found in N-ras transfected BL8 cells, or by treatment with metabolically activated AfBi by Power et al. (1987). These early studies demonstrate that oncogenic factors alter the expression of GST genes. Induction of glutathione transferase enzymes can occur via a number of different mediators including synthetic food additives. Four such additives [2(3)tert-butyl-4-hydoxyanosole] (BHA), BHT, propyl gallate and ethoxyquin have been shown to modulate experimental carcinogenesis (Taylor & Scanlan, 1989). Ethoxyquin, a very efficient inhibitor of lipid peroxidation and a direct antioxidant, also appears to act via the induction of detoxification enzymes, particularly the glutathione S-transferases. Antioxidants, including BHT, BHA, and oltipraz, along with ethoxyquin are all effective inducers of glutathione reductase, glutathione Stransferase, UDP-glucuronyl transferase and epoxide hydrolase. Studies involving dietary supplements containing 4% ethoxyquin prior to exposure to aflotoxin have shown greater than five-fold induction of multiple forms of glutathione S-transferase in vitro and a 4.5-fold increase in biliary elimination of aflotoxin in vivo as a glutathione conjugate (Taylor & Scanlan, 1989). Organosulfur compounds such as allylic di- and tri-sulfides found in foods such as onions and garlic have been demonstrated to inhibit both initiation and promotion of carcinogenesis. The major means by which these compounds appear to inhibit initiation is by their ability to induce glutathione S-transferase activity (Taylor & Scanlan, 1989). A number of other compounds have shown to induce GSTs including, DDT, indoles, and benzylisothiocyanate. Induction is thought to result from enhanced transcription evoked by these as well as other  33  diverse groups of chemical agents including oxidizable diphenols and phenylenediamines, Michael reaction acceptors, organothiocyanates, electrophiles (alkyl and aryl halides), metal ions (HgCl2 and CdCl2), trivalent arsenic derivatives, vicinal dimercaptans, organic hydroperoxides and hydrogen peroxide, and 1,2dithiole-3-thiones (Prestera et al., 1992). The molecular mechanisms of induction of GST by 28 compounds from the above listed groups have been studied by Prestera et al. (1992). A construct containing the 41-bp enhancer element derived from the 5' upstream region of a mouse liver GST subunit gene ligated to the 5' end of the isolated promoter region of this gene, and inserting it into a plasmid containing human growth hormone receptor gene was prepared. This construct was then transfected into Hep G2 human hepatoma cells. The concentration of the 28 compounds tested required to double growth hormone production and to double quinone reductase specific activities were found to be closely linear. It was shown that all of the chemical inducers tested stimulate the expression of the gene through this 41 base pair enhancer element, regardless of the wide diversity amongst the classes of compounds tested. The sequences of the upstream enhancer elements of this gene that have been found to respond to inducers have been called the electrophile-responsive element (EpRE) and antioxidant-responsive element (ARE), contained within the 41 nucleotide segment. This enhancer element is thought to mediate most, if not all, of the phase II enzyme inducer activity of widely different compounds (Prestera et al., 1992). The gene encoding human G S T - J C has been sequenced including its promotor region (Xia et al., 1993). It has also been demonstrated that the binding site for the concensus activator protein 1, AP1, is essential for its promotor activity. There is also an intronic sequence which upregulates the transcription of GSTP1 and is integral to the promotor (Xia et al., 1993). Modulation of expression of the  human GST-re gene has been shown to be affected by both retinoic acid and insulin, with retinoic acid having a negative effect and insulin a positive effect; Retinoic acid has been shown to be an effective therapeutic agent in treating a number of carcinomas (Xia et al.,1993). The findings of its negative modulation of GSTs have implications in the understanding of mechanisms which lead to the overexpression of GSTs in tumours and the ability of retinoic acid to trigger cell differentiation and to block malignant potential. The findings of positive modulation of GSTs by insulin has provided information on the possible association of c-jun, c-fos, and ras genes. Insulin stimulation causes an increase in c-jun and c-fos mRNA (Xia et al., 1993). The enhanced transcription of GST-re is thus thought to be regulated by Fos and Jun proteins and supports the suggestion by Burgering et al. (1991) that p 2 1  r a s  is an  intermediate of an insulin signal transduction pathway involved in the regulation of genes such as c-jun and c-fos. (Burgering et al., 1991; Xia et al., 1993).  2.7  TISSUE DISTRIBUTION OF GLUTATHIONE S-TRANSFERASES IN NORMAL HUMAN TISSUE  Human alpha, mu and pi class glutathione S-transferases have been localized immunohistologically in a variety of organs. Human tissues vary considerably in overall GST content as well as in expression of GST subunits or isozymes. It has been suggested that quantitative and qualitative differences between tissues is important in determining the susceptibility of a tissue to individual or multiple carcinogens. Variability in expression of different isozymes has also been seen during stages of development in both animal and human studies.  In studies on the expression of glutathione S-transferase during rat liver development, the ontogeny of rat liver GSTs during fetal and postnatal development have shown the expression of the three major classes of GST, alpha, mu and pi, each having differential expression during stages of liver development. The pi form has been found to be essentially a fetal protein regulated at the level of transcription, with alpha and mu classes being 'mature' proteins regulated by a multitude of mechanisms, both transcriptional control during fetal development and post-transcriptional control post-natally (Tee et al., 1992). The significance of the changes in expression of the three enzymes at different levels of development are not entirely clear, however the susceptibility of the fetus and neonate to potentially toxic exposure and the types of enzymes present at these time is of interest. In the rat, GST7-7 is not expressed in the adult liver but is markedly increased during hepatocarcinogenesis. This pattern of expression and its distinct expression during fetal development has also lead to the suggestion that it may be an oncofetal protein (Tee et al., 1992). Studies of GST-TC in human livers have similarly shown the level of expression to be clearly different between adult and fetal livers, determined by levels of mRNA and protein present (Kashiwada, 1991). The isozyme composition of glutathione S-transferase in several human fetal organs has also been studied by Pacific et al.(1986). Human fetal lung, brain, kidney, intestine, liver, and adrenal gland were studied with the findings that lung, brain, kidney and intestine contain only the acidic form of transferase (GST-TC). This suggests that G S J - J C represents a more primitive form present in higher amounts in fetal tissues with a potential multifunctional role and wide substrate specificity (Pacific et al., 1986). Developmental changes in the expression of differing isozymes were noted when levels found in fetal liver, kidney and adrenal gland were compared to the corresponding adult tissues, however no developmental differences were found in fetal and adult lung or brain.  The vast majority of information on different tissue expression of GSTs is from adult human tissues. Terrier et al. (1990), in a study of normal human tissues, found that G S T - J C was expressed predominantly in epithelial cells of urinary, digestive and respiratory tracts, suggesting a possible role in detoxication and elimination of toxic substances. These findings coincide with those found for the expression of p-glycoprotein found in secretory epithelia and provides support for the possibility of their co-regulation in normal and/or malignant cells (Terrier et al., 1990). This has also been supported by studies by Efferth et al.(1992), in which human kidney cell lines, hematologic malignancies and breast carcinomas have shown a tendency for co-expression of P-170 and GST-JT in kidney and breast. No expression was found in hematologic malignancies. (Efferth et al., 1992) The expression of GST-TC in salivary gland was not included in the study by Terrier, however a variety of tissues examined and the main immunoreactive structures in those tissues are summarized in Table 3. Corrigall and Kirsch presented findings on the concentrations of GSTs alpha, mu, and pi in eighteen organs of nine individuals (Corrigall & Kirsch, 1988) and Campbell et al. (1991) have documented human alpha, mu and pi class GSTs, localized immunohistologically, in fifteen human tissues as well as thirty-nine different types of benign and malignant tumours. In these studies, large interorgan variation in alpha transferases was found, with relative constant concentration of pi class transferases. The pi class transferases were present in relative abundance in ductular cells in liver, pancreas, kidney and salivary gland and less so in parenchymal cells. The results of tissue distribution and the presence of GST-rc in bile ducts, pancreatic ducts, epididymis, renal collecting tubules and salivary ducts (Campbell et al., 1991) are important when considering its possible role in detoxication. A striking feature of the results of Terrier et al. (1990) was the strong expression of G S T - J C in the epithelia of urinary, respiratory, and digestive systems,  37  the three major systems involved in the elimination of toxic substances from the J  body, suggesting a role in toxin excretion and metabolism. GST-re isozyme is found most abundantly in kidney and has been found along the nephron, loop of Henle, distal convoluted tubule, collecting duct, ureter and in the urinary bladder. GST-re is also the predominant isozyme found in lung with a possible role in the protection of the respiratory tract from volatile toxins or pollutants.  2.71  DISTRIBUTION OF GST-re IN GLANDULAR TISSUE  The pattern of distribution of enzymes in structures with similar functional or developmental characteristics may, in part, provide information on the role that these enzymes play in vivo. Histologically, breast and pancreas have some similar characteristics as salivary gland and thus the distribution of GST-re in these tissues is of interest when comparing tissues. Strong similarities between some tumours of breast and those of salivary gland exist. Tissue distribution of GST-re in pancreas sharply outlines centroacinar and ductular structures leaving acini and islets unstained (Campbell et al., 1991). This is similar to its expression in salivary gland. Breast, however, revealed staining in both acinar and ductal epithelium, with myoepithelial cells also staining positively (Terrier et al., 1990).  2.72  DISTRIBUTION OF GST-re IN SALIVARY GLAND  The study by Campbell et al. (1991) appears to provide the only previous information available on the expression of GST-re in normal human salivary gland tissue and human salivary gland tumours. Findings for normal glands were based  on only three specimens of unspecified sites and showed that ductular cells stained positively, similar to ducts of the pancreas, while acinar cells stained negatively. Only three benign salivary gland tumours were stained and all were positive. It is known that enzymatic levels in different salivary glands may differ for a given enzyme. It is not known whether GST-TC levels differ in different human salivary glands. There is also a lack of information regarding the expression of GST-JC  2.8  in the variable human salivary gland tumours.  ROLE OF GST AS A MARKER IN NEOPLASIA AND PRENEOPLASIA  GSTs not only have a role in the detoxication of carcinogens, but one isozyme has been reported as a marker of neoplastic transformation in at least one model of chemical carcinogenesis. The Solt-Farber model of carcinogenesis shows that GST7-7, an older nomenclature for a pi class GST in rat, is markedly elevated (Solt et al., 1977). In this model, following exposure to a carcinogen to initiate transformation and a short interval later, rats are subjected to partial hepatectomy to simulate hepatocyte proliferation and are treated with a toxic agent (eg. 2-acetylaminofluorene) which inhibits the normal growth of hepatocytes. A few weeks following exposure to the toxin, rat liver is found to be studded with hyperplastic nodules, the vast majority of which are preneoplastic and regress with time. Over forty different carcinogens have been reported to produce such hyperplastic nodules. Some of the nodules undergo malignant transformation to hepatocellular carcinoma. Several investigators have found increased expression of rat GST7-7 in hyperplastic nodules (Casenave et al., 1989) and the induction of GST 7-7 in premalignant liver cells is predictable, such that it has become a marker  of premalignant changes in rat hepatocytes. Similar findings occur with the acidic form of GST (GST-re) in human hepatocellular carcinomas (Tatsemasu et al., 1985). In the rat studies, the areas in which increased levels of GST were found are not areas with increased rates of DNA replication, implying that overexpression of GST is not associated with proliferation of the cell (Casenave et al., 1989). In a study by Sato (1989), some of the foci induced by the Solt-Farber model stained heterogeneously using an M2 antibody (mu class GST) which, as suggested by these authors, might indicate a variation between lesions generated in different systems. In immunohistochemical studies involving application of initiators, it was found that very small GST-7-7 positive foci or even single cells are detectable before an increase in GST-7-7 content or GGT activity are biochemically evident, in one case appearing 1-2 weeks after a single administration of initiator (Sato, 1989), and in another model within 48 hours of a single dose of diethylnitrosamine (DEN), dimethylnitrosamine (DMN), or aflatoxin Bi (AfBi) (Moore et al., 1987; Sato, 1989). These individual cells and GST7-7 positive foci increased with increasing doses of initiator (DEN) and were not induced by known promotors of liver carcinogenesis including phenobarbitol, methylcholanthrene, polychlorinated biphenyls and isosafrole (Moore et al., 1987). This indicated that GST-7-7 was an accurate marker for very early changes or 'initiated' cells. In addition, nuclear staining was found in earlier and smaller lesions, which did not show positivity with other GST forms, suggesting that GST-7-7 may be present in the nuclei of preneoplastic cells. Moore et al. (1987) demonstrated similar elevation of GST7-7 in pancreatic tissue of Syrian hamsters exposed to carcinogens. Using polynitrosamines, elevated levels of GST7-7 were seen in both preneoplastic and neoplastic pancreatic tissue. GST-7-7 positive foci in rat are particularly useful for practical application in risk assessment and detection of carcinogenic potential of  40 environmental compounds as GST-7-7 positive foci in rat liver are reliable markers for both genotoxic and non-genotoxic carcinogens (Ito et al., 1992). In one class of the GST enzymes, the mu class, significant polymorphism in the expression of the GST-p: gene in humans is seen. Studies by Warholm et al. (1983) have indicated that some individuals express this class of enzyme and that a link exists between the expression of GST-u, and the potential risk of carcinogenicity. Placental GSTs, including GST7-7 (GST-P) and GST-re, have been found to be elevated in preneoplastic and neoplastic lesions in experimental models as mentioned, and also in human tissues (Daly et al., 1991; Efferth et al., 1992; Toffoli et al., 1992; Zhang et al., 1993). In some instances, elevated GGT levels have been noted to occur along with the increased levels of GST. It has been suggested that increased expression of these enzymes (GST and GGT) in premalignant and malignant cells may provide selective growth potential over GST and G G T negative cells and also impart resistance to cytotoxicity. Increased resistance may occur due to the induction of the enzymes GGT and GST. GGT permits cells to utilize extracellular glutathione to preserve their internal glutathione levels, and GST induction allows glutathione utilization for the protection of altered cells in an environment of exposure to xenobiotics. Thus, the combined effects of these enzymes, in a toxic environment, may impart selective growth or enhanced proliferation of cells (Hendrich & Pitot, 1987). Association of both GSTs and GGTs with other detoxication pathways are of importance, specifically in association with glycoprotein P-170 and cytochrome p450. The p450 system named for p450, a hemeprotein distinguished by the wavelength (450nm) of the carbon monoxide derivative of its reduced form, is responsible for many metabolic reactions. One of the initial metabolic reactions for xenobiotics, including carcinogens, involves oxidation to a detoxication product.  41  Metabolic reactions, in turn, have been shown to convert xenobiotics to a form which may be activated or closer to an ultimate carcinogen. In studies on the potential involvement of the detoxication enzymes in some of these reactions, coexpression of p-glycoprotein (P-170) and GST-TC have been found in human kidney carcinoma cell lines, hematological malignancies, and human breast carcinomas (Efferth et al., 1992). A significant relationship has also been found between the two proteins in resistant non-small cell lung carcinomas of smokers (Volm et al., 1991). The significance of this dual expression has not been fully clarified, although it is not unexpected given the nature and interrelationship of these enzymes in processes associated with carcinogenesis. It appears that these enzymes may be involved at many levels of the multistages of carcinogenesis, not only the initial reaction of cells to exposure, but possibly initiation, promotion, and progression of the process of carcinogenesis or tumourigenesis. They may also be under a common regulatory control. Elevated levels of GST may also act in the prevention of lipid peroxidation, a process thought to be involved in tumour progression (Tan, 1986). Studies have revealed that the pi class of glutathione S-transferases appear to be most persistent and strongly expressed in multiple human tumours, with alpha GSTs found in some neoplasms (Campbell et al., 1988; Lewis et al., 1988). In a study of thirty nine different benign and malignant tumours by Campbell et al.(1988), 31 of 39 different types of tumours gave a positive reaction to GST-TC. Numerous other investigations of immunohistochemical detection of GST-TC antibody have been found to be useful for the detection of precancerous states, or dysplasia and differentiated carcinomas in several human organs. Increased levels of GST-TC are found in various premalignant and malignant lesions, including breast, kidney, lung, liver, colon, gallbladder, stomach, esophagus, uterine cervix, as well as oral mucosa (Dillio et al., 1986; 1987, 1988; Nakagawa, 1988; Sato et  42  al., 1987; Kodate et al., 1986; Soma.et al.,1986; Mannervik et al., 1987; Nishihira et al. 1991, 1993; Shea &Henner, 1987; Tsutsumi et al., 1987; Shiratori et al., 1987; Tsuchida, 1989; Zhang et al., 1992, 1993). The over expression of GST-TC in a variety of neoplasms have shown that it may be a marker for these lesions. It has also been found to be associated with oncogenes and thus may be a marker of transformation of protooncogenes or activation of oncogenes. The theta class of GSTs is thought to have possibly been the evolutionary forerunner of the alpha, mu, and pi classes, with the latter classes arising by duplication of the theta gene allowing organisms to adapt to various toxic stresses during the course of evolution (Wilce & Parker, 1994). The pi form of GSTs is the form found prevalently in the fetus and is thought to be a possible primative form and thus may be oncofetal in nature (Kashiwada, 1991; Tee, 1992). Although contradictory to the previous studies, there have been some studies documenting lack of distinction between dysplastic or neoplastic and normal tissue and GST-JC expression, particularly with respect to cervical neoplasia (Maquire et al., 1991; de Camargo & Rodrigues, 1992; Randall et al., 1990). Normal human colon as well as tumourous colonic tissue predominantly express GST-TC, and thus it is not considered a good marker for colonic tumours, however increased levels in tumours may contribute to the relatively high resistance of such tumours to chemotherapeutic agents (Peters et al., 1989).  2.81  GST AND ITS USE IN CLINICAL APPLICATION - SERUM ANALYSIS  Serum analysis for levels GST-TC have been undertaken for a variety of clinical analyses. Following intravenous pyelography, utilizing the injection of  highly iodinated contrast materials, elevated levels of GST-TC can be found in urine. The presence of these enzymes has been used in the selection process of kidneys for transplantation, with elevated levels considered a sign of cell death, thus negating the use of a particular organ (Feinfeld et al., 1978). Elevated serum levels of GST-TC have also been found in cases of hepatitis, cirrhosis, as well as following liver damage by carbon tetrachloride (Jakoby, 1980). In studies regarding drug toxicity and enzyme levels associated with fascioliasis - an infection with a genus of trematode worms, blood glutathione (GSH), erythrocyte glutathione S-transferase (GST) and serum gamma-glutamyl transferase (GGT) were evaluated before and after treatment with biothionol. It has been found that levels of GSH, GST and GGT were elevated in patients with fascioliasis and that following treatment with biothionol, levels returned to those corresponding to normal control values, confirming the toxic features resulting from infection by Fasciola and suggesting no adverse effects from biothionol treatment (Abdel-Rahman, 1990). Thus blood and serum levels have been useful in early detection of therapeutic response in facioliasis. A number of investigations on the levels of GST-TC in sera of patients with tumours have been undertaken. GST-TC levels in patients with gastric carcinoma and esophogeal carcinoma, have been shown to be elevated. As well, the elevated levels have been found to decrease to within a normal range following surgical removal of tumours in patients with esophogeal carcinoma (Tsuchida et al., 1989). It has been suggested that follow-up monitoring of serum GST-TC levels may be of use in monitoring patients with cancers during the course of treatment (Sato, 1989). Studies evaluating the levels of GST in patients having cancer of the digestive tract have also been documented. Severini (1993) has shown that patients with gastric, liver, and colorectal cancer have significantly elevated mean  levels of GST. Ninety-five patients having neoplasms of the digestive tract were studied, with 90% of those with gastric cancer, 100% of those with liver cancer, and 89% of those with colorectal cancer having levels of GST above those of patients with non-neoplastic disease. In a number of patients (15/95), post-operative levels of GST were found to initially rise and then subsequently decline, suggesting that GST levels in plasma may be useful as a tumour marker in gastric, liver, and colorectal cancer and possibly increased sensitivity when evaluated in combination with other markers such as carcinoembryonic antigen (Severini, 1993). A number of recent studies investigating serum levels of various enzymes in patients before, during, and following cancer therapy have shown that GST-TC levels may be of significance as parameters for predicting therapeutic response to combination chemotherapy regimens. Resistance to chemotherapeutic agents is a major problem in non-small cell lung cancer treatment. The role of GST in drug resistance, and particularly resistance to alkylating agents that include platinum compound, is thought to have an important effect on chemotherapy. In a study by Hida et al. (1994), serum levels of a number of different enzymes including GST-TC, carcinoembryonic antigen (CEA), neuron specific enolase (NSE), and squamous cell carcinoma (SSC) antigen were investigated in patients who received combination chemotherapy with platinum compound. It was found that 50 of 121 (41.3%) patients with non-small cell lung carcinoma had elevated GST-TC serum levels as compared to healthy control subjects. Pre-treatment levels of GST-TC between patients who showed partial versus no response to chemotherapy varied. Patients who showed no response had significantly higher GST-TC pre-treatment levels than those who showed partial response. None of the other enzymes studied showed any relationship to treatment response. Thus, it was concluded that pre-treatment GST-TC levels may be a useful parameter for predicting response  to chemotherapy regimens that include platinum compound (Hida et al., 1994). Detoxication enzymes, referred to as phase II enzymes, include glutathione S-transferases, quinone reductase and UDP-glucuronosyltransferases. These enzymes are considered to be anticarcinogenic due to their detoxication potential. Serum levels of GST and quinone reductase, have been studied with respect to their possible correlation to tissue levels in an investigation of a possible chemoprotected state (Prochaska & Fernandes, 1993). Inducers of these enzymes may be consumed as food additives, medications or constituents of foods and have been shown to produced elevated tissue levels. Dietary BHA [2(3)-tert-butyl-4hydroxyanisole] or dimethyl fumarate is one such inducer and was shown to produce both an increased level of hepatic specific activity as well as increased serum activity of both quinone reductase as well as GST in mice. Serum phase II enzyme activities are dependant on tissue levels, and are determined by measuring quinone reductase and glutathione S-transferase activities in serum. It is suggested that induction of these enzymes by inducers found in fruits and vegetables, may have significance in prevention, and that serum enzyme assays of phase II enzymes may be useful in chemoprevention trials (Prochaska & Fernandes, 1993). Further, studies of serum lipid peroxides, antioxidant vitamins C and E, serum selenium, serum ceruloplasmin, erythrocyte and its membrane lipid peroxidation and antioxidant enzymes including catalase, superoxide dismutase, glutathione peroxidase and glutathione S-transferase levels have been analyzed in post-menopausal untreated women with benign and malignant breast tumours as compared to age matched women with no known neoplasia (Hida et al., 1994). Severe impairment of antioxidant potential was found in women with breast cancer distinguished by increase circulating lipid peroxides, ceruloplasmin and significant decrease in antioxidant vitamins and selenium. Erythrocyte and its membrane lipid  46  peroxidation, including GST, was significantly increased. In this study, various interdependant systems were evaluated and support the concept that anticarcinogenic or antioxidant potential relies upon numerous mechanisms, each contributing to prevention of lipid peroxidation inferred to be elevated in cancer patients (Kumar et al., 1991).  2.82  DISTRIBUTION TUMOUR  CELL  OF  GST-TC  IN  HUMAN  TUMOURS  AND  HUMAN  LINES  A study of GST activity in human tumours and tumour cell lines has been carried out, attempting to assess whether GST activity and isozyme composition in human tumour cell lines reflected those found in vivo. GST activity levels, using 1choloro-2,4-dinitrobenzene (CDNB) as a substrate, were measured in both solid tumour and corresponding tumour cell lines (Lewis, 1989). In most cases the mean activity was higher in solid tumour than in cell lines. As these tumour cell lines were found to express levels of this enzyme differing from those found in corresponding solid tumours, this may have implications in research using cell lines carried out extensively for studies in mechanism of action of anticancer drugs in lieu of tissue samples.  2.9  SEX  DIFFERENCES  IN  GST  The regulation of concentration of GSTs occurs at several levels. Sex differences in the expression of GST in a number of animal and human tissues, both normal and neoplastic have been found. In some animal studies, male rats  were found to generally have higher levels of GST activity based on assay with several substrates (Jakoby, 1978). Certain isozymes (transferase B) have been found, however, to be higher in female animals (Hales & Neims, 1976). As well, hepatic GSTM II, found in large amounts in adult male mouse liver and significant amounts in adult female mouse liver, was inducible to higher levels in females by the injection of testosterone. Levels in male mouse were reducible by castration which resulted in levels similar to those found in females. Isozymes of this form of GST were not affected by administration of testosterone or castration (Mantle et al., 1987). Sexual differences have been found in rat liver, both in metabolism of steroid hormones, but also in the metabolism of xenobiotics. The sex differences in metabolism has been attributed to neonatal 'imprinting' by testicular androgens during the neonatal period (Blanck et al., 1984). Early studies found that sex differences occurred in the development of GST-7-7 positive foci induced by some carcinogens. Areas of GST-7-7 positive lesions in the Solt-Farber model were far larger in male rats than female rats (Sato, 1991). This same effect was found for GGT-positive liver foci in rats . Marked sex differences were found in the area ratio of enzyme altered liver foci in sexually mature male and female Wistar rats and differences in the development of hepatic carcinoma show males developing cancer earlier than females with a tendancy to a longer latency period in males than females (Blanck et al., 1984,1986). Hepatic liver foci, which have been observed to undergo malignant transformation, have been altered by controlling the influence of the hypothalamic-pituitary axis on hormonal controls (Blanck et al., 1986). These studies show a hypothalamic-pituitary influence on at least some of the enzymatic or metabolic pathways of various hepatocarcinogens in rat. In studies of human colon tissue, both male and female, all three classes of GST are expressed with consistent 2-fold higher quantity of alpha GST in males,  the presence of 2 mu class GSTs in females and only one in males, and a higher thermostability of the pi class GST in female colon as compared to male (Singhal et al., 1992). Although only a single gene for the pi isozyme has been suggested, studies on amino acid sequence isolated from a single placenta have shown single amino acid differences in primary structure. Very fine differences such as this have been suggested as possible explanations for differences in thermostability and thus extremely subtle changes in structure may explain differences between enzymes in male and female. Differences in primary structure, discernable only by strict analysis, may be present and may be gender related. As well, given that GSTs bind a number of hydrophobic ligands, inlcuding steroids, differential binding of GSTs to sex-hormones or their metabolites may explain potential variability in concentration and kinetic capabilities between genders. Studies comparing activity, quantitation and kinetic properties of GSTs in human skin showed to be consistent with previous studies on colon (Singhal et al., 1992). Three GST isozymes were expressed in both male and female skin with the major form being GST-re. As in colonic tissue, higher specific activities of GST-re of female skin as compared to male were found. Whether differential expression of GST-re is the sole reason for these findings or whether post-translational modification causes differences in kinetic properties found in the sexes is not known. Differences in the localization and expression of estradiol, progesterone and progesterone receptor in human normal salivary gland and salivary gland tumours have been studied. In tumours of the adenoid cystic type, both sex steroids as well as the receptor for progesterone were found. These findings indicate that the human salivary gland is one of the target tissues for estrogen and some suggest that tumours which express progesterone receptors will respond to endocrine therapy (Ozono et al., 1992). The relationship of glutathione S-transferase  expression and hormonal influence on cellular levels and kinetics in salivary gland is not clear, although there appears to be increasing information that gender differences occur both in the expression of GST in similar tissues in males and females, as found in skin and colon, and that salivary gland is a hormonally sensitive tissue with tumours of the salivary gland showing variable sensitivity. One would thus expect that differences in isozyme levels may be present between genders. It is important to consider whether both qualitative and quantitative differences in expression of particular enzymes translate into the ability to detoxify various xenobiotic, including carcinogens. It has been suggested that observed differences in incidence of some forms of cancers between males and females, particularly of the head and neck, may be related to factors related to gender susceptibility as well as exposure to xenobiotics (Silverburg & Lubera, 1987). The physiologic significance of the gender related differences is not specifically known, yet since GST isozymes are implicated in defence mechanisms against chemical carcinogens and their expression is known to be affected by exposure to xenobiotics, differential expression between genders is a distinct possibility. Differential induction of GSTs may also explain some of the differences found in men and women, with the suggestion that exposure to xenobiotics as a consequence of working environment, use of cosmetics, etc. may result in the varying levels as opposed to there being true gender related differences.  Chapter 3.  HUMAN SALIVARY GLAND TUMOURS  3.1  INTRODUCTION  There exist two types of salivary glands, major and minor. The major glands consist of paired parotid, submandibular, and sublingual glands. Minor salivary glands, although smaller in size, are widely distributed in the palate, tongue, lips, buccal mucosa and floor of mouth. The microscopic structure of secretory units show slight differences and correspondingly the composition of saliva in the glands differs, with sublingual glands producing mainly mucous, the parotid saliva having the consistency of serum, and the submandibular glands producing an intermediate form. Despite these differences, the same tumour types occur in all , salivary glands, both major and minor. Knowledge of tumours of human salivary gland has advanced considerably in recent years. Histologic typing and differences in prognosis and treatment have become better understood by the availability of immunohistochemistry, cytophotometry, hybridization techniques, tissue culture and chromosomal analysis. Specifically, immunocytochemical or histochemical techniques provide information useful for continued classification, functional differentiation and in providing prognostic indicators for therapeutic planning. A number of tumour markers are utilized to distinguish the three main types of epithelial cells in salivary gland - acinic cells, ductal cells and myoepithelial cells (Seifert & Caselitz, 1989; Seifert & Sobin, 1991). Specific markers for acinic cells include amylase, cytokeratin, epithelial membrane antigen (EMA), lactoferrin, lysozyme, secretory component, carcinoembryonic antigen (CEA), and several blood group antigens. Those for ductal cells include cytokeratin, EMA, tissue peptide antigen (TPA), lactoferrin, lysozyme, secretory component, immunoglobulin A, lectin receptors, CEA, and several blood group antigens. Myoepithelial cells can be distinguished by the absence of secretory products, the presence of actin, myosin and S-100  proteins, and the double expression of cytokeratin and vimentin. Leukocyte common antigen and cytokeratin are used for the distinction between undifferentiated carcinomas and malignant lymphomas or sarcomas. Carcinoembryonic antigen and thyroglobulin are useful for the differential diagnosis of primary salivary gland tumours (adenocarcinoma) and metastases of thyroid carcinoma. The expression of detoxication enzymes in human salivary gland or salivary gland tumours, until this time, have had limited investigation.  3.2  ETIOLOGY  Salivary gland cancers differ from other cancers of head and neck in that they have a relative infrequent occurrence, they have a diversity of histologic subtypes with a wide range of biologic behaviours, and they have not been found to be directly associated with use of tobacco products. There has been no associated risk found with respect to exposure to substances as a consequence of any specific occupational or leisure-time activity (Spitz et al., 1990). There are, however, epidemiologic similarities between skin and salivary gland neoplasms possibly attributable to exposure to ultraviolet radiation which have prompted studies on risk of developing skin cancer subsequent to salivary gland cancer and the risk of developing salivary gland cancer subsequent to skin or other cancers. These have shown significant risk of lip cancer in men and melanoma in women following prior salivary gland cancer (Spitz et al., 1990). As ultraviolet radiation is a known risk factor for skin cancer, and a significant association has been found for subsequent salivary gland cancer, it has been suggested that etiology may partially be attributed to exposure to ultraviolet radiation. Some literature suggests that the only established risk factors for salivary gland cancer are radiation exposure and a history of prior cancer (Horn-  Ross et al., 1991). These independant factors have differing effects on the development of salivary gland cancer. A prior cancer likely reflects shared risk factors for a carcinogenic process not solely identified for a secondary tumour, and the occurence of salivary gland tumours subsequent to other carcinomas may not be entirely due to radiation treatment for the primary carcinoma (Horn-Ross et al., 1991). Radon, a naturally occurring element present in trace amounts throughout the crust of the earth produced by radium decay, has been found to have a causal relation to lung cancer in uranium miners. A study undertaken to determine if there is a correlation between radon levels and the incidence of salivary gland tumours in an area of high radon exposure in Pennsylvania found no correlation between incidence of salivary gland tumours and radon levels in two different distant geographic locations (Georgia and Southern California) and Pennsylvania (Miller, 1993). Polyoma virus has been shown to induce tumours in salivary gland of mice, and alteration in frequency, morphology and spectrum of tumour type as well as biologic behaviour have been shown to occur in tumours induced by a polyoma virus mutant with a specifically altered oncogene (Freund et al., 1992). A number of wild type polyoma viruses possess vastly different tumour-inducing abilities in mice, with high tumour strains rapidly inducing tumours of both mesenchymal and epithelial origin and low tumour strains inducing a much lower incidence of tumours of only mesenchymal origin (Dawe et al, 1987; Freund et al., 1992). Specific determinants, including the middle T protein, the major viral capsid protein VP1, the large T antigen, and non-coding regulatory sequences upstream from the early viral promoter, have been found to affect virus replication and tumour induction in target cells of epithelial origin and they have also been found to influence the incidence of tumours in specific subsets of target tissues, including  54  salivary gland.(Freund et al., 1992). Although there is abundant literature on the role of polyoma virus in the etiology of murine salivary gland tumours, there is little information on its role in human salivary gland tumourigenesis. In one study, chromosomal patterns in experimentally induced tumours in mice as compared to those in human tumours have been found to be quite different. The most striking feature found in polyomainduced tumours was a pronounced karyotypic instability, characterized by a multitude of chromosomal deviations; human tumours, instead, were characterized by recurrent translocations and deletion involving specific chromosomal regions (Sandros et al, 1990). The reasons for these differences are not known, but may be related to viral characteristics or inherent characteristics of human or mouse salivary gland tissue. A number of reports are available which discuss the role of oncogenes in the development of salivary gland neoplasms. Elevated levels of H-ras, K-ras and cmyc transcripts having been found in pleomorphic adenomas when compared to normal salivary gland (Spandidos, 1985). More recently, the proto-oncogene cerbB-2, also designated neu and HER-2, has received considerable interest. This oncogene encodes for a glycoprotein of the tyrosine kinase family which has 50% homology with the epidermal growth factor receptor. Amplification or overexpression of this oncogene has been identified in various human neoplasms, predominantly adenocarcinoma arising in kidney, colon, stomach, ovary, and breast (Gullick, 1990). The role of c-erbB-2 has also been investigated, with amplification found in adenocarcinoma (Semba et al., 1985), pleomorphic adenoma (Kahn et al., 1992; Riviere et al., 1991), and adenoid cystic carcinoma (Riviere et al., 1991). Some immunohistochemical studies of c-erbB-2 expression in salivary gland tumours have shown this oncoprotein to be expressed only in malignant salivary gland tumours (Kernohan, 1991; Sugano, 1992). A recent study  by Jordan et al. (1993) has demonstrated overexpression of c-erbB-2 in a number of salivary gland tumours, both benign and malignant as compared to normal gland determined by in situ hybridization, with acinic cell carcinoma and mucoepidermoid carcinoma exhibiting significantly higher expression. Another study has demonstrated that c-erbB-2 oncogene amplification is relatively rare in salivary gland neoplasms (Birek, 1994). Given the differing results of these studies, it has been concluded that overexpression of this oncogene in all salivary gland tumour types cannot be accounted for by gene amplification alone (Jordon et al., 1994). Other molecular events are therefore thought to be involved and remain to be investigated.  3.3  CLASSIFICATION OF SALIVARY GLAND TUMOURS  Diverse histologic subtypes of salivary gland tumours exist. The most recent histologic classification of salivary gland tumours by the World Health Organization is presented in Table 4. Only a number of the more common and important tumours were included in this study and the following is a discussion of those tumours.  3.31  3.311  BENIGN SALIVARY GLAND TUMOURS  PLEOMORPHIC ADENOMA  The benign mixed tumour, or pleomorphic adenoma (PA), is one of the most important salivary gland tumours as it is the most common tumour of both major  \  and minor salivary glands. It also has a relatively high recurrence rate and has the potential for malignant transformation. This tumour is named for its architectural and cellular pleomorphism as it consists of epithelial and modified myoepithelial elements intermingled with connective tissue having mucoid, myxoid, chondroid or hyaline appearance. The ducts of pleomorphic adenoma, like those of normal salivary gland, are composed of two main cell types; inner duct-lining cells and outer myoepithelial cells. In normal glands, the inner ductal cells are prominant, while outer myoepithelial cells are less discernible. In pleomorphic adenoma, the opposite usually occurs, with myoepithelial cells being prominant (Thackray, 1974). These cells play an important role in determining the overall appearance of this tumour. Attempts have been made to classify pleomorphic adenoma on the basis of differentiation of the epithelial cells as well as proportion and differentiation of the stroma. They are often defined as cell-rich and cell-poor types. Cell-rich PAs contain an abundance of epithelial cells, while in some cell-poor pleomorphic adenomas, the myxoid and chondroid areas, which are so characteristic of the this tumour, may be extensive enough to form the greatest part of a neoplasm. Pleomorphic adenomas are slow growing neoplasms. They grow by expansion and by local outbreaks pushing through the shell of fibrous tissue into adjacent gland. Capsular in-growth has been found to occur in microscopic foci, thought to gradually increase in size, resulting in the frequent multilobulation of tumours (Naeim, 1976). Although benign, they have a propensity for recurrence and can also undergo malignant transformation. Various explanations have been proposed for the recurrence of pleomorphic adenomas, with recurrence rate varying from 17 to 50% (Krolls, 1971). Generally, recurrence is associated with incomplete excision (Patey & Thackray, 1958; Krolls, 1971), however variations in recurrence rates may differ significantly depending  upon the surgical procedures used in initial therapy (Naeim, 1976). A higher recurrence rate of pleomorphic adenomas in the parotid region as compared to those of the submandibular glands is likely related to more conservative surgical excision in this area in attempt to preserve and avoid damage to the facial nerve. Projections formed by tumour expansion and/or capsular ingrowth may result in small regions of tumour left during enucleation. As well, in myxoid and chondroid areas, clefts may develop within the tumour, and if parallel to what is considered the deep resection margin, they may create a false plane of resection leaving tumour behind attached to fibrous capsule (Thackray & Lucas, 1973). The most common reason for recurrence, during excision, however, is thought to be rupture of the thin fibrous capsule during removal of the tumour resulting in spillage of tumour fragments into the wound, with multiple subsequent tumour nodules forming in the field of operation (Thackray & Lucas, 1973). It has been suggested that the cell-poor variants of pleomorphic adenoma have a higher risk of recurrence and the cell-rich variants have a higher risk of malignant transformation. However, reports on the relationship between cellularity and recurrence are conflicting (Krolls & Boyer, 1972; Naeim, 1976; Foote & Frazell, 1953; McFarland, 1942; Eneroth, 1965; Evans & Cruikshank, 1970). It is generally accepted, however, that recurrent tumours are predominantly of the cell-poor type. There is also recent evidence that intrinsic cell behaviour may be partly responsible for recurrence of PAs (Takahashi et al., 1992; Gallo et al., 1992).  3.312  MONOMORPHIC ADENOMA  The term monomorphic adenoma was established in1967 to denote specific clinical pathologic entities distinct from the mixed tumour or pleomorphic adenoma.  Monomorphic adenomas are benign tumours in which epithelium forms a regular, usually glandular pattern. In these tumours there is no evidence of mesenchymallike tissue that is a characteristic component of pleomorphic adenoma (Thackray & Lucas, 1973). The absence of myxochondroid tissue, however, is not sufficient reason to label it monomorphic. These tumours consist of monophasic histology composed predominantly of cells of one type along with the absence of connective tissue changes. The epithelium is arranged throughout the tumour in one or another characteristic way, defining individual tumours as belonging to one of a number of subcategories of monomorphic adenoma. A variety of cellular types and architectures can be seen in this group of tumours which include basal cell adenoma, canalicular adenoma, sebaceous adenoma, sebaceous lymphadenoma, oncocytoma, papillary cystadehoma lymphomatosum (Warthin's tumour), and sialadenoma papilliferum or inverted duct papilloma (see Table 4), each having a characteristic configuration of epithelium. Generally, the group of tumours comprising the different monomorphic adenomas are treated by conservative surgical excision including a margin of normal uninvolved tissue, or superficial parotidectomy. They have a low rate of recurrence.  3.32  3.321  MALIGNANT SALIVARY GLAND TUMOURS  ACINIC C E L L CARCINOMA  This tumour of salivary gland is distinctive, although it has a wide spectrum of histopathologic features. It is a malignant epithelial neoplasm that demonstrates some cytological differentiation toward acinar cells. It is thought to originate from  intercalated duct cells or neoplastic transformation of acinar cells. Individual cell characteristics may vary from those of well differentiated acinar cells to intercalated duct cells. They may be vacuolated, clear or non-specific glandular cells. All cell types may be present, but acinar cells, identified by their characteristic appearance of round or polyhedral cells arranged in sheets or acini having basophilic granular cytoplasm, in association with other cells, enables the definition of the broad range of acinic cell carcinoma. Varying growth patterns exist including a solid pattern, microcystic pattern, papillary cystic pattern, and follicular growth pattern. Microscopic examination reveals the majority to be infiltrative. This tumour may be found in all age groups, including children, with a mean occurence in the fifth decade and a female prevalence. They usually present as slow growing lesions, less than 3 cm in diameter, with pain a frequent presenting symptom. They often give a clinical impression of a benign lesion. They present as a range of 1.4% to 20% of malignant parotid tumours with 81% of these tumours presenting in the parotid gland, with an unusual feature of bilateral involvement in approximately 3% of cases (Regezi & Sciubba, 1993). Acinic cell carcinomas, although having a very low metastatic potential, do have a significant tendency to recur, with some patients requiring re-excision of tumour nodules over a period of many years. Regional lymph node metastasis occurs in about 10% of cases, with distant metastasis in about 15% of cases. Poor prognostic features include pain or fixation to surrounding tissue, tumour invasion and microscopic features of desmoplasia, cellular atypia and increased mitotic activity. The morphologic patterns and cellular composition of this tumour do not dictate prognosis. Surgery is the preferred management (WHO, 1991).  3.322  MUCOEPIDERMOID CARCINOMA  These tumours, the most common malignant tumour of salivary gland origin are, as their name implies, epithelial lined mucin producing tumours. They are characterized by the presence of squamous cells, mucin cells, and cells of intermediate type. They are somewhat controversial with respect to presentation and biologic behaviour. Although evidence has been gathered indicating that all mucoepidermoid lesions are carcinomas and that they should be considered able to metastasize, regardless of their macroscopic or microscopic appearance, the most recent classification, although not absolute, grades them as either low or high grade indicating their probable prognosis with respect to local recurrence and metastatic potential. Low grade mucoepidermoid carcinomas typically are well differentiated, usually circumscribed but not encapsulated, predominantly cystic, locally invasive and relatively non-aggressive, while high grade mucoepidermoid tumours are poorly differentiated, macroscopically ill-defined, tend to be solid, and may show focal areas of haemorrhage or necrosis, and have greater metastatic potential. Clinical manifestation of mucoepidermoid carcinoma depends upon the grade of the malignancy. Those of low grade may present as benign tumours do, with an extended period of painless expansion or enlargement. In the oral cavity, they may resemble a mucocele and may be fluctuant due to cystic formation. High grade mucoepidermoid carcinomas grow rapidly and may have associated pain and mucosal ulceration. When present in the major salivary glands, high grade malignancies may present with signs of obstruction or facial nerve involvement. In mucoepidermoid carcinoma, prognostic significance is ascribed to tumour grade. Those of low grade may follow a benign clinical course while those of high grade may present with local and distant metastases in up to 60% of cases. Five  61 year survivial rates for low grade lesions are usually 95% or better. However, survival rates for high grade lesions are approximately 40% at five years, dropping to approximately 25% at fifteen years. Recurrence is highly dependant upon completeness of excision of the primary lesion and detection prior to metastasis.  3.323 ADENOID  CYSTIC CARCINOMA  This malignant tumour of salivary gland is infiltrative and has been considered one of the most biologically deceptive tumours of the head and neck region, if accounts for approximately 23% of all salivary gland carcinomas with approximately 50 to 75% of them occurring in minor salivary glands of the head and neck. It has various histologic features including three growth patterns: cribriform (glandular), tubular, and solid. Tumours consist of two cell types, the duct lining cells and those of the myoepithelial type. Adenoid cystic carcinoma is characteristic in its microscopic appearance, biologic behaviour, and high rate of local recurrence and systemic spread. Perineural and perivascular spread without stromal reaction is very common and although often indicators of poor prognosis in other tumour types, these features along with mitotic activity and pleomorphism have not been correlated with prognosis. All of these tumours, regardless of type are biologically aggressive and may give rise to metastases long after excision of the primary tumour. Solid types appear to have the worst prognosis often having early recurrence, early metastasis, and high mortality rates. Local extension into bone can indicate a poor prognosis. Statistically, glandular and tubular types have a better prognosis. Clinically, the appearance is of a unilobular mass that is firm to palpation, occasionally with pain or tenderness. They tend to have a slow growth rate,  62 although they may also present as a mass which has undergone rapid growth in a previously inactive tumour. The tumour may be freely movable or fixed to adjacent tissue. Since they have a propensity for infiltration of neural spaces, pain is a fairly common symptom and facial nerve weakness or paralysis may occur. Adenoid cystic carcinoma, due to its slow growth and absence of metastases until late stage, is often judged in terms of 15 to 20 year survival rates. Although early survival rates are good, with approximately 70% at five years, long term outlook has been poor. Fifteen year survival rate is approximately 10%. Apart from metastases, recurrent local disease is common. Poor prognostic factors include facial nerve paralysis, with death occurring within five years of onset in most patients with this complication, incomplete initial excision, tumour size greater than 3cm., and tumours revealing a solid growth pattern with areas of necrosis.  3.324  MALIGNANT MIXED TUMOUR  Malignant mixed tumours are uncommon, comprising only 2-10 % of cases in a number of large series (Naeim, 1976). Histologic criteria and features thought to contribute to recurrence of their benign counterpart, the pleomorphic adenoma, are not considered criteria of malignancy. However, the high recurrence rate of the benign lesions and their potential for malignant transformation distinguishes them from other salivary gland tumours. The only features of malignancy include tumours with an infiltrative and destructive pattern of extension, as well as direct evidence of malignant transformation (Naeim, 1976). An exceptionally rare entity, the metastasizing pleomorphic adenoma, has been reported which presents with typically benign histologic features in distant metastatic sites. The metastatic tumour has similar appearance to the primary tumour but usually presents  63 following multiple local recurrences (Thackray & Lucas, 1973; Freeman et al., 1990; Wenig et al., 1992). Thus, the malignant mixed tumour represents three separate entities including carcinoma arising in pleomorphic adenoma (carcinoma ex pleomorphic adenoma), true malignant mixed tumour (carcinosarcoma), and the metastasizing mixed tumour.  3.4  T H E R A P Y AND PROGNOSIS  Treatment of benign salivary gland tumours involves surgical resection with completeness of excision being a factor of recurrence, particularly in the case of benign pleomorphic adenoma. Surgical resection has also been the traditional treatment for malignant salivary gland tumours. Treatment of pleomorphic adenoma of parotid gland was often previously local enucleation or shelling out of the tumour in an attempt at preservation of the facial nerve. However, on long term follow-up, this was often found to lead to appreciable recurrences. Thus some form of partial parotidectomy with careful preservation of the facial nerve is now considered the operation of choice. In British Columbia, therapy for benign pleomorphic adenoma of parotid gland consists of subtotal parotidectomy with dissection of the facial nerve. If surgical and pathological findings suggest complete excision, no further treatment is indicated. If the initial excision is inadequate, that is, if positive surgical margins are found in the surgical specimen histologically, postoperative radiotherapy may be indicated. If recurrence develops following initial surgical treatment, re-excision followed by radiotherapy is advised (BCCA Cancer Treatment Policies Manual, 1992). In submandibular gland, no large nerve is present and the entire gland and tumour is usually removed, thus the problem of recurrence rarely occurs.  A long cellular doubling time is typical in salivary gland carcinoma which provides for high cure rates. However, a risk of local recurrence and a tendancy for local invasion exists. Nodal metastasis occurs late and distant metastasis occur most frequently in the lungs, but may also present in bones, brain, liver and other viscera. Nodal and distant metastasis are most common in adenoid cystic carcinoma. In a large series, the four most common malignant salivary gland tumours cited were mucoepidermoid (34%), adenoid cystic carcinoma (22%), adenocarcinoma (18%), and malignant mixed tumour (13%) (Spiro &Spiro,1989), with recurrence rates reported by site as 39% in parotid gland, 60% in submandibular gland, and 65% in minor salivary glands (Spiro, 1986). Treatment of a primary mucoepidermoid carcinoma is surgical with the quality of surgical excision being an important factor in local recurrence and prognosis. Low grade lesions usually require excision without neck dissection, while high grade lesions are treated with surgery, often requiring radical neck dissection, plus post-operative radiotherapy to the primary site. The overall five year survival rate is approximately 70%, with low grade tumours having over 90% survival rate but a much poorer prognosis for high grade tumours (WHO, 1991). Treatment of adenoid cystic carcinoma, regardless of the site of the primary lesion, is surgical. When occurring in the parotid gland, wide resection including superficial parotidectomy or superficial and deep lobectomy are often indicated. There is debate over sparing of the facial nerve when the parotid gland is involved. When tumour arises in minor glands wide excision, often including underlying bone to provide a wide margin, is the treatment of choice. Post surgical radiotherapy has been used in cases of recurrence. Chemotherapy is considered ineffective, although has been used in the management of widely metastatic disease. Five year survival rate is approximately 70%, however fifteen year survival rate is approximately 10%.  TNM clinical classification of salivary gland tumours incorporates size of primary tumour, extension to regional lymph nodes, and distant metastasis. (Tprimary tumour, N-regional lymph nodes, M-distant metastasis). A summary of this classification is found in Table 7. The efficacy of adjuvant radiotherapy in improving locoregional control of advanced salivary gland tumours has been reported (Tu et al.,1982; Goepfert et al.; 1983; Horiuchi et al., 1990; North et al., 1990; Shingaki et al, 1992). In one study, it was not found to have significant impact on neck and distant metastases (Shingaki et al., 1992). In another study it was found to be effective therapy for patients with positive surgical margins, particularly those with T1 and T2 stage disease with postoperative raidotherapy. Patients with T3 and T4 disease required more aggressive therapy, and those with nodal metastases in the neck at admission tended to have distant metastases and a poor prognosis regardless of post operative radiotherapy (Sakata et al., 1994). Tu et al. (1982) have reported a survival advantage only in patients presenting with recurrent disease, with a significant advantage for patients treated with combined therapy versus surgery alone. A number of univariate analyses have determined important prognostic indicators of carcinoma of salivary glands (Fu et al., 1977; Jackson et al., 1983; Tu et al, 1982; McNaney et al, 1983). In one of the few multivariate analyses, in which several apparently significant prognostic factors were entered into a multivariate model, only a few remained as significant. These included facial paresis, histologically undifferentiated tumours, male sex, skin involvement and postoperative radiotherapy (North et al., 1989). In this study tumour types included high, intermediate, and low grade mucoepidermoid carcinoma, acinic cell carcinoma, squamous carcinoma, undifferentiated carcinoma, carcinoma expleomorphic adenoma and malignant mixed tumour. It was determined that  postoperative radiotherapy improves both local control and patient survival. The development of fast neutron therapy, or more generally high-linear energy transfer (LET) radiation, based on several radiobiologic principles including the role of hypoxic cells in radioresistance, is also under investigation. This therapy, using neutrons as opposed to photons to potentiate control of malignant tumour growth, has shown interesting results. Locally extensive, inoperable salivary gland tumours have shown superior response to fast neutrons compared to conventional low-LET radiation (Catterall & Errington, 1987; Griffin et al., 1979; Battermann & Mijnheer, 1986; Griffin et al., 1988; Tsunemoto et al., 1989; Griffin et al., 1989). It has been suggested that fast neuron therapy alone should be the treatment of choice for advanced stage salivary gland tumours, and that surgery should be limited to those cases where risk of facial nerve damage is small or a high likelihood of obtaining a negative surgical margin exists (NCI, 1991; Wambersie, 1994). Fast neutron therapy, however, has its greatest potential in only limited settings, one being the treatment of salivary gland tumours. Thus, availablility of this form of treatment is not widely accessible. The treatment of recurrent and metastatic head and neck squamous and salivary gland cancers with chemotherapy is often palliative. Very few randomized trials of chemotherapy have been undertaken on salivary gland malignancies. Studies of single and multiple agents including cisplatin, cyclophosphamide, doxorubicin, fluorouracil, and vincristine have been done with the dose of agent(s) and schedules varying. Due to the heterogeneity of tumour subtypes and small numbers of patients for study, results have been difficult to interpret. It has been suggested, though, that doxorubicin/cyclophosphamide combinations have resulted in response in individuals with adenocarcinoma, adenoid cystic carcinoma, and acinic cell carcinoma (Pinto & Jacobs, 1991). Drugs with higher activity in squamous cell cancers have shown favourable results in patients with  malignant mixed tumour, and undifferentiated carcinoma (Pinto & Jacobs, 1991). However, the difference in survival may be difficult to analyse due to the variable predictability of salivary carcinomas, particularly adenoid cystic carcinoma. A prospective phase II study carried out to test Cisplatin (CDDP) as a single agent in salivary gland carcinomas concluded that CDDP was a moderately active drug in salivary gland carcinomas of varying types, and that it should be included in multidrug regimens to be tested in prospective studies (Licitra et al., 1991). However, a recent randomized trial of Epirubicin and Flurouracin versus Cisplatin suggests that chemotherapy has no place in the treatment of salivary gland malignant tumours (Jones et al., 1993). In British Columbia, treatment of malignant salivary gland tumours including mucoepidermoid carcinoma, adenoid cystic carcinoma, adenocarcinoma, acinic cell carcinoma, squamous cell carcinoma, anaplastic carcinoma and malignant mixed tumour, involves, in parotid gland, total parotidectomy with preservation of facial nerve. The nerve is not sacrificed unless it involves the only area in which adequate excision cannot be achieved. One or more branches of nerve may be sacrificed without sacrificing the main trunk. Postoperative radiotherapy is usually advised for malignant parotid tumours, but may not be necessary for acinic cell carcinoma and low-grade mucoepidermoid carcinoma. Neck disection is advised only if positive nodes have been demonstrated (BCCA Treatment Policies Manual, 1992). In submandibular gland, wide surgical excision is the preferred method of treatment which usually involves suprahyoid dissection for submandbular gland tumours. Composite resection is reserved for tumours with bone or floor of mouth involvement (BCCA Treatment Policies Manual, 1992). In the case of minor salivary glands of the upper aerodigestive tract mucosa, wide excision of the tumour is the treatment of choice when feasible. Inoperable lesions are managed by radiotherapy alone (BCCA Treatment Policies Manual, 1992).  Chapter 4. EXPRESSION  OF  GST-TC IN  EXPERIMENT  NORMAL  SALIVARY  HUMAN  GLAND  #1 SALIVARY  TUMOURS  GLAND  AND  The knowledge of normal distribution of GST-TC is important in our understanding of its normal function and may provide insight in its dysregulation during carcinogenesis. The normal histological distribution of GST-TC has been reported in a number of studies (Campbell et al., 1991; Terrier et al., 1990), showing a wide variation of GST-TC distribution in different organs and tissues. Only one study (Campbell et al., 1991) has investigated the expression of GST-TC in three normal salivary glands of unspecified site and in two salivary gland pleomorphic adenomas. It is not clear if the normal distribution of GST-TC varies among the three major salivary glands and if the normal distribution of GST-TC varies between minor and major salivary glands. There is a lack of information on the status of GST-TC in a wide variety of benign and malignant salivary gland tumours.  4.1  O B J E C T I V E AND HYPOTHESIS  1.  To determine the normal distribution of GST-TC enzyme immunohistochemically in both major and minor human salivary glands.  The experiment will test the hypothesis that normal salivary glands show GST-TC  staining in the ductal system but not in the acini units and that the  GST-TC  staining pattern and staining intensity may differ in the three major  salivary glands, parotid, submandibular and sublingual, and may differ between the major and minor salivary glands.  2.  To determine and compare the immunohistochemical staining pattern and level of GST-TC between monomorphic adenomas and the more aggressive  70  pleomorphic adenomas.  The experiment will test the hypothesis that higher GST-re level will to be found in the benign but aggressive pleomorphic adenomas as compared to other benign monomorphic salivary gland neoplasms.  If the specific hypothesis is proven, it would suggest that increased GST-re levels provide a growth advantage in pleomorphic adenomas and play a role in its aggressive behaviour.  3  To determine and compare the immunohistochemical staining pattern and level of GST-TC between common benign salivary gland tumours and malignant salivary gland tumours.  This experiment will test the hypothesis that malignant salivary gland tumours contain higher levels of GST-JC than benign salivary gland tumours.  If the hypothesis is proven, it would suggest that GST-re plays an important role in the development of malignant salivary gland tumours and may provide a growth advantage for these tumours, and it would suggest that GST-re may be used as a tumour marker and an adjuvant tool in the diagnosis of these tumours as well as in serological screening for recurrence and metastasis of these tumours.  4.2  MATERIALS AND METHODS  Archival pathology specimens, fixed in 10% formaldehyde and embedded in  71  paraffin, were selected from Vancouver General Hospital and Shaughnessy Hospital Archives, Vancouver, British Columbia. Seventy four non-neoplastic salivary glands were studied, including 16 normal glands in biopsies taken for non-neoplastic lesions including lymphoepithelial cysts and sialadenitis and 58 apparently normal or inflammatory glands adjacent to salivary gland tumours. The specimens were from both major salivary glands, including 43 parotid glands, 14 submandibular glands, 1 sublingual gland, and 16 minor salivary glands from palate, lip, buccal mucosa, buccal vestibule, retromolar area, and floor of mouth. The following salivary gland tumours were used: Benign Salivary Gland Tumours(42) - 23 pleomorphic adenomas, 19 monomorphic adenomas, including 5 basal cell adenomas, 1 oncocytic cystadenoma, 1 mucinous cystadenoma, 1 myoepithelioma, 1 oncocytoma, 1 oxyphilic adenoma and 9 Warthin's tumours; Malignant Salivary Gland Tumours(29) - 7 acinic cell tumours, 11 adenoid cystic carcinomas, and 11 mucoepidermoid carcinomas  4.21  GST-TC  IMMUNOHISTOCHEMISTRY  Anti-human polyclonal GST-TC antibody was supplied through the generosity of Professor Y.C. Awasthi, Department of Human Biology, Chemistry, and Genetics, University of Texas Medical Branch, Galveston, Texas, U S A . The specificity of the antibody has been established in previous studies. Avidin-biotin-peroxidase (ABC) kit was purchased from Dimension Laboratories, Mississauga, Ontario. Five micron sections were cut and submitted for GST-TC immunohistochemical staining using the avidin-biotin complex method of Hsu et al. (1981) as detailed in Appendix 3.  Briefly, following incubation in a dry oven, slides were deparaffinized in xylene, rehydrated in graded alcohol, and washed. Endogenous peroxidase activity was blocked by incubation with 0.3% hydrogen peroxide in methanol (30min). The sections were rinsed in phosphate-buffered saline (PBS), blocked with Vectastain non-immune goat serum (30min), and incubated overnight at 4 ° C with the primary GST-pi antibody diluted 1:2000 in PBS. All reactions were carried out in a humidified chamber. Localization of the primary antibody was achieved by incubation with a biotinylated goat anti-rabbit secondary antibody (1 hr) followed by a 30 minute incubation with avidin-biotin complex conjugated to horseradish peroxidase. The slides were washed with PBS after each incubation. The site of peroxidase binding was determined by staining the slides with 0.1% diaminobenzidine tetrahydrochloride (DAB). They were then rinsed and counterstained with hematoxylin. Each set of staining reactions included carcinogen-treated rat liver known to strongly express placental G S T positive nodules to serve as a positive control and ensure reproducibility of the staining. As a negative control for the specificity of the primary antibody binding, rat liver in which the primary antibody was omitted was included. A detailed description of the immunohistochemical staining method is outlined in appendix 3. The intensity of GST-TC staining was judged as negative or equivocal (- or +/-), weakly positive (1), moderately positive (2), or strongly positive (3) idependantly by two observers (MZ, LZ). During the initial grading of the slides, four slides that were graded as 0,1, 2, or 3 respectively were chosen to serve as a standard of reference for each grading intensity. Whenever there were doubts in grading of a slide, these standard slides were referred to in order to ensure consistency of the grading.  4.3  RESULTS  4.31  GST-TC IN  NORMAL AND  INFLAMMATORY  SALIVARY  GLAND  In the normal salivary glands of both major and minor types, the glandular acinar cells, whether mucous-secreting or serous-secreting cells, were negative with GST-TC staining. Ducts stained positively and the intensity of the staining varied, in general, from moderate to strong. The larger excretory ducts tended to show a stronger staining while the intercalated ducts tended to be weaker (Plate 1 B.). The staining was mainly cytoplasmic, however, when the staining was strong, nuclear staining was also noted, particularly in the large excretory ducts. Weakly stained flattened cells were noted at the periphery of and between the acinar cells and they were best viewed in un-counterstained slides. These cells were interpreted as myoepithelial cells. Connective tissues stained negatively. Inflamed salivary glands showed similar GST-TC staining patterns to those of normal glands: negative acini and positive ducts. The proliferating ducts associated with salivary gland inflammation stained moderately with GST-TC staining (Plate 1 A). Morphologically normal or inflammatory salivary glands adjacent to tumours showed similar GST-TC staining to those of normal and inflammatory glands from patients without neoplasms. The staining properties of glands from all sites, including both major and minor glands, were similar.  4.32  GST-TC  IN  HUMAN  SALIVARY  GLAND  TUMOURS  4.321  MONOMORPHIC  ADENOMA  As shown in Table 5, the majority of the monomorphic adenomas showed equivocal or weak GST-TC staining. The staining in general was uniform, although stronger staining was noted in some cells With ductal differentiation (Plate 1 O , D.). The only case that showed strong GST-TC staining was a mucinous cystadenoma. This tumour has two types of tumour cells, mucus-secreting luminal cells and basal cells. The mucus-secreting cells, like normal mucus-secreting acini, were negative, while the basal cells stained strongly positive with GST-TC.  4.322  PLEOMORPHIC  ADENOMA  Most of the pleomorphic adenomas (PAs) showed weak to moderate staining (Table 5). Pleomorphic adenomas are characterized by pleomorphic types of tumour cells and stromal features. The GST-TC staining varied between different types of cells. In general, cells with ductal differentiation showed the highest degree of staining, followed by cells with squamous differentiation. Cells which stained to a lesser degree were the myoepithelial cells, which comprised the bulk of PAs, although the plasmacytoid type of myoepithelial cells tended to stain stronger. When pleomorphic adenomas were sub-typed into cell-rich versus cellpoor, the cell-rich type tended to show weaker staining when compared to that of the cell-poor type (Table 6; Plate 1 E . , F . , G . , H . ) , although the difference was not significant (p=0.0789, Mann-Whitney Test). All cases of pleomorphic adenoma which stained strongly were of the cell-poor variety. The ceilularity of the pleomorphic adenomas varied frequently from area to area within a single tumour. In a number of instances, high staining intensity was found in tumour islands and  particularly amongst scattered cells within the myxoid, cell-poor areas.  4.323  ACINIC CELL CARCINOMA  Acinic cell carcinomas, in general, were equivocal or negative in staining (Table 5). In hematoxylin and eosin stained sections, acinic cell tumours are characterized by two types of cells. One type includes large cells with abundant, frequently granular cytoplasm, similar to serous acinar cells. The other type of cells are small, resembling intercalated duct cells (Plate 2 A.,B.). The proportion of the two types of cells in each tumour varies. In the majority of the tumours studied, the large cells predominated while in some tumours the small cells predominated. All the acini-like large tumour cells were negative with GST-TC staining, similar to normal acinar cells. The small ductal cells, in general, showed weak staining (Plate 2B.).  4.324  ADENOID CYSTIC CARCINOMA  The majority of adenoid cystic carcinomas showed equivocal or weak staining (Table 5). These tumours consisted mainly of myoepithelial cells but also contained, in some areas, ductal cells. Similar to other tumours, the ductal cells showed stronger staining than the myoepithelial cells (Plate 2 F . , G . ) .  4.325  MUCOEPIDERMOID CARCINOMA  This was the only type of salivary gland tumours.which showed markedly  •  ....  increased GST-TC staining (Table 5). These tumours consist of three types of cells: mucous cells, intermediate cells and epidermoid cells. Mucous tumour cells were all negative, similar to normal mucous acinar cells (Plate 2 E.). The remainder of the tumour cells showed a general strong staining (Plate 2 C.,D.). The average staining intensity of mucoepidermoid carcinomas was significantly higher than the other salivary gland tumours (p<0.0001 - Student Newman-Keuls Multiple Comparison Test) The degree of tumour differentiation had no apparent effect on intensity of staining.  4.4  DISCUSSION AND CONCLUSIONS  The knowledge of the normal distribution of GST-TC in salivary gland is important in our understanding of its normal function. It may also provide insight into the dysregulation of this enzyme during carcinogenesis. Campbell etal. (1991) have reported the normal immunohistochemical distribution of GST-TC in three salivary glands from unspecified sites. It is not clear from this study by Campbell, however, whether there is different GST-TC distribution among the three major salivary glands and between the major and minor salivary glands. It is known that activity of an individual enzyme may vary in salivary glands of different locations. For example, gamma-glutamyl transpeptidase (GGT), an enzyme closely related to GST in its metabolism, has been found to be strongly positive in sublingual glands, weakly positive in submandibular glands and negative In parotid glands (Shiozawa et al., 1993). This study has investigated the normal distribution of GST in human salivary glands of all locations. The results are consistent with those of Campbell et al. (1991): ductal cells of salivary gland stained with GST-TC antibody, myoepithelial  76  cells stained weakly positive, and acinar cells were negative. The pattern of GST-TC distribution is similar for normaland inflammatory glands in all locations: parotid, submandibular, sublingual, as well as minor salivary glands. The normal distribution of GST-TC in other glandular tissues has been reported in some studies. It is interesting to observe that morphologically similar glands of different organs demonstrate different GST-TC patterns of expression. In pancreas, acinar cells are negative while ductal cells are strongly positive, similar to those of salivary glands (Campbell et al., 1991). In mammary, sebaceous, and sweat glands all cells, including acinar, ductal, and myoepithelial are positive (Terrier et al., 1990). In the gastrointestinal tract, esophogeal submucosal glands are positive while the duodenal submucosal glands are negative (Terrier et al., 1990). The factors underlying these differences in GST-TC distribution are unclear. The results of this study showed that the staining pattern of salivary gland tumour cells was similar to that of cells in normal and inflamed glands. As in normal gland, cells having acinar differentiation, including serous cells in acinic cell tumours, mucous cells in mucous cystadenoma and mucoepidermoid carcinoma, all stained negatively with GST-TC. Other tumour cells, including ductal cells, epidermoid cells and myoepithelial cells were generally positive with GST-TC. As in normal glands, when a tumour consists of more than one type of tumour cells, such as in pleomorphic adenoma and adenoid cystic carcinoma, the highest staining reaction was noted in cells with ductal differentiation while the weakest staining was noted in myoepithelial tumour cells (Plate 1 E..F.; Plate 2 E.,F.). The expression of isozymes of the glutathione S-transferase family in salivary gland, including GST- alpha, mu, and theta have not been documented. Although there is considerable interest in the pi class of enzymes, the understanding of the other forms of these detoxication enzymes and their possible role in carcinogenesis as related to salivary gland should not be excluded. Both  alpha and mu classes have been found in a variety of neoplasms and preneoplastic lesions in other human tissue types; To fully understand the scope of their involvement in carcinogenesis, future studies of different GST isoforms in salivary gland and the variability of expression within the different cells which form these glands is desired. Histologically, pleomorphic adenomas can be divided into cell-rich and cellpoor types (WHO, 1991). Cell-poor PAs show abundant stroma, frequently hyaline, myxoid or chondromyxoid in nature. They grow by expansion and, characteristically, by local "outbreaks", pushing through the shell of their fibrous capsules into the adjacent gland. These surface projections are often seen to develop in myxoid parts of the tumours (Thackray & Lucas, 1974; Naeim et al., 1976). It is interesting to note that higher GST-TC staining was not infrequently observed in the tumour cells within the myxoid regions of these tumours and that cell-poor PAs tended to show a higher staining intensity than the cell-rich type. It is generally accepted that the high recurrence rate of PAs is mainly a result of incomplete first removal. The fact that recurrent PAs tend to be of the cell-poor type, however, suggests that intrinsic cell behaviour may also play a role in the recurrence. The higher GST-TC staining in cell-poor type PAs seems to support this hypothesis. The difference, however, is not statistically significant ( p=0.0789 Mann-Whitney Test), though the insignificance may be attributed to the small number of tumours used or the insensitivity of subjective grading. Given these findings, a second experiment was designed to investigate whether recurrent PAs, which are usually the cell-poor type, demonstrate higher GST-TC content than nonrecurrent PAs using computer image analysis. This first study, however, showed no differences in GST-TC staining reaction between monomorphic adenomas and the more aggressive pleomorphic adenomas, or between benign and malignant salivary gland tumours with the  exception of mucoepidermoid carcinomas. Whether the increase in GST-TC content in mucoepidermoid tumour represents a phenotype associated with malignancy or a phenotype associated with squamous differentiation remains speculative. Squamous cell carcinomas of various organs including oral squamous cell carcinomas have been found to contain increased levels of this enzyme (Campbell etal., 1991; Hanada etal., 1991; Volm etal., 1991; Zhang etal., 1994). It is possible that the increase in the enzyme activity is associated with both malignancy and/or squamous differentiation. The increase in levels of GST-TC in some of these mucoepidermoid tumours, given its role in cellular protection, may render them resistant to drugs or radiation therapy. Although not considered a primary treament modality, radiotherapy, as adjunctive treament is often used in cases of poor further surgical potential. In these cases, GST levels in the initial surgical specimen, when baseline levels are known, may be useful in indicating prediction of the value of such adjunctive therapy. It has been reported that increases in GST-TC activity in cancers, such as esophageal cancers, can be measured accurately through serum sample, and that serum GST-TC levels decline following resection of the cancers (Tsuchida et al., 1989).  The marked increase in GST-TC activity found in mucoepidermoid  carcinomas may be useful in serological screening of recurrent and metastatic mucoepidermoid carcinomas. Elevation of serum phase II enzymes, including glutathione S-transferases, quinone reductase, and UDP-glucuronosyltransferases which act in blocking toxic, mutagenic, and neoplastic effects of carcinogens may also be useful in assessing tissue levels and thereby present as markers of a chemoprotective or radioprotective state. This may be of importance when considering alternative treatment methods to surgery for different tumours. Serum levels may be used to clarify the role that these phase II enzymes play in carcinogenesis, and in the evaluation of anticarcinogens in a clinical setting  (Prochaska & Fernandes). This can be done by utilyzing tissue biopsies or lymphocytes to assess the effectiveness of various compounds to induce these enzymes and potentially provide a protective cellular environment. It is also known that the phase II enzymes act in close relation to a number of other protective mechanism, including p-170 and others. Currently a number of studies looking at the relationship between these different systems within tissues are underway. Similar investigations of multiple enzyme expression in neoplasms of salivary gland may also lend information to the role of these pathways in salivary gland neoplasia. Normal salivary gland excretory ducts showed the highest GST-TC staining reaction as compared to the smaller ducts. It is generally presumed that mucoepidermoid carcinomas arise from the large excretory ducts of salivary gland. The finding of a high content of GST-TT in these tumours may simply reflect the origin of the tumour from cells with a normally high content of the enzyme and support the hypothesis that mucoepidermoid carcinomas arise from large excretory ducts. Differences in expression of GSTs between sexes in both normal and neoplastic tissues have been reported in some studies but not in others (Jakoby, 1978, Hales & Neims, 1976, Sato, 1991, Blanck et al., 1984, 1986, Singhal et al., 1992). The gender distribution of tumours in this study, including benign and malignant tumours, presented as 27 male and 44 female. There was no significant difference in expression of GST-TC in salivary gland tumours between men and women (p=0.3109 Unpaired t-test, two tailed value ). Salivary gland tissue has been reported to be estrogen and progesterone sensitive, with receptor positivity found in ducts, the same location that the enzyme GST-TC  is expressed (Ozono, et al, 1992). There have been no reports of hormone  sensitivity or their relationship to any variant of salivary gland tumours, however  this information would be of interest. Comparison with information known regarding hormone sensitivity in tumours of similar tissues such as breast may lend insight into prognosis and behaviour of some tumours. GSTs are also known to bind steroid hormones, and the relationship, if any, to the finding of hormone receptors in ductal cells only may elucidate the potential interrelation of these enzymes (detoxication), hormone sensitivity, and carcinogenesis. Although these enzymes predominantly provide a protective role to cells, there is documentation of glutathione conjugate mediated toxicities (Monks, et al., 1990). Increased levels of the GSTs in cells may promote increased ability for conjugation reactions, dependant upon cellular levels of glutathione, enzyme, and toxic substrate. Thus in one instance increased levels of GSTs may provide selective cellular protection to those with increased levels of enzyme, while in another instance they may provide selective potential for the formation of increased levels of toxic conjugates thus leading to possible destruction of cells. Depending upon which type of reaction occurs and whether it is occuring within a "normal" or "transformed" cell, the result may be the selective growth of either normal or transformed cell, or the reaction of toxic conjugates and potential alteration or destruction of a normal or transformed cell. The normal expression of these enzymes, and their role in conjugation of a variety of substrates must be understood, before being able to define the ongoing cellular processes as well as their application to clinical trials.  Chapter 5.  EXPERIMENT 2  QUANTIFICATION OF IMMUNOHISTOCHEMICAL  REACTION OF GST-  IN NON-RECURRENT AND RECURRENT PLEOMORPHIC ADENOMA  The recurrence rate reported for pleomorphic adenomas (PAs) varies from 0 to more than 50% and is believed to be relatively high as compared to other benign salivary gland tumours (Krolls, 1971: Thackray & Lucas, 1973). It is generally accepted that the liklihood of further recurrence after one recurrence is even higher. The main reason for the recurrence of PAs is believed to be incomplete first surgical removal. A number of studies have shown that cell-poor PAs have a higher recurrence rate than the cell-rich variety, possibly due to their tendency to have local deficiencies of their capsule as well as having a rich myxoid stroma resulting in easy rupture or 'spillage' of tumour during surgery. The consistency and friability of this type of tumour also results in difficulty of removal (VandenBerg et al., 1964; Krolls & Boyers 1972; Naeim et al., 1976; Goudot et al., 1989; WHO, 1991). Whether the biological behaviour of the tumour cells also attributes to the recurrence of PAs remains a question. In experiment 1, it was noted that cell-poor PAs tended to show a higher staining intensity than the cell-rich type. A similar phenomenon was also noted within individual tumours. When there was variation in cellularity from area to area within a tumour, frequently higher GST-re staining was observed in the tumour cells within the myxoid and cell-poor parts of the tumour. Since recurrent pleomorphic adenomas tend to be more cell-poor than primary pleomorphic adenomas, the results from experimant 1 raised the possibility that recurrent pleomorphic adenomas, in general, contain a higher level of GST-re, which may be partly responsible for the more aggresive behaviour of recurrent pleomorphic adenomas. Identification of a marker that would predict the recurrent tendency of pleomorphic adenomas is highly desirable. The grading of GST-re staining intensity in experiment 1 was carried out semiquantitatively by individual examiners through naked-eye analysis using a light microscope. There are obvious limitations in the degree of precision with such  a quantitative method. Semi-automatic computerized image analysis of the GST-TC staining should provide a more precise grading of the immunohistochemical reaction.  5.1  O B J E C T I V E AND HYPOTHESIS  To determine and compare the immunohistochemical staining intensity of GST-TC  between non-recurrent pleomorphic adenomas and recurrent  pleomorphic adenomas using a computerized image analysis system.  This experiment will test the hypothesis that recurrent pleomorphic adenomas contain a higher level of GST-TC enzyme as compared with nonrecurrent pleomorphic adenomas.  If the hypothesis is proven, it would suggest that recurrence of pleomorphic adenomas may be attributed partly to intrinsic cellular traits and that increased GST-TC levels may provide these tumours with growth advantage and a more aggressive (recurrent) behaviour, and would also suggest that GST-TC  staining may be a useful adjunctive tool in the prediction of recurrrent  behaviour of pleomorphic adenomas.  -  5.2 MATERIALS AND METHOD  The sources of specimens were the same as described in Experiment 1. Two groups of tumours were used: Group 1: Non-recurrent pleomorphic adenomas (12 cases). The time required for recurrent pleomorphic adenoma to manifest clinically is extremely variable, ranging from a few months to over thirty  years. In general the time interval is between two to five years between observation of recurrence and presentation for treatment (Batsakis, 1979). For this reason, the non-recurrent group of lesions was chosen from patients whose pleomorphic adenomas had been removed at least five years prior with no history of recurrence. Archival materials dated 1988 or earlier were used and the confirmation of no recurrence was obtained from family physicians and surgeons; Group 2: Recurrent pleomorphic adenomas (13 cases). The staining procedure was similar to that of experiment 1 with the exception that, following DAB staining, no hematoxylin counterstaining was done. The slides were kept un-coverslipped until the time of imaging when the slides were coverslipped with immersion oil.  5.21  COMPUTER IMAGE ANALYSIS  The grading of the staining intensity was done by a computerized high resolution image analysis system. This was carried out at the British Columbia Cancer Agency in collaboration with Xillix Technologies Corp., Vancouver, B.C. Canada under the direction of Dr. N. Poulin. The image analysis system consists of a camera (Xillix Microlmager 1412 high resolution monochrome CCD-charge coupled device) mounted in the primary image plane of a Nikon Optiphot microscope. A Nikon PlanApo 10x objective was used for all measurements. Brightfield illumination was supplied with a 100 Watt quartz-tungsten-halogen lamp, driven by a stabilized DC power supply. The wavelength of brightfield illumination was selected using a 10nm bandpass filter centered at 470 nm, close to the absorbance maximum of DAB used in staining of the specimens. For each specimen, ten random individual fields were chosen by delineating  these areas using the computer mouse. The non-recurrent and recurrent tumour groups were mixed and thus their recurrence status was not known at the time of selection, although tumour morphology, being cellular or not, was visibly evident. The amount of epithelium present in each of the sections varied considerably and thus the total area selected also varied between specimens. Regardless of area selected, the image analysis computes optical density per area. Thus, whether a number of individual cells were selected, or a large epithelial island of tumour selected, the optical density of the specimen is calculated relative to the total area. An attempt was made for each 10 power field to select as much tumour as possible for this calculation. Digitized images were captured and stored for these areas. This allowed for revisiting of a particular specimen and site if so desired. Within each of the ten individual fields, tumour islands or cells were delineated and the intensity of the immunohistochemical staining in these cells was measured by the system as optical density (OD), thus at least ten measurements were used to calculate the mean optical density for each individual tumour specimen. For each field measured, an area of connective tissue or an area of normal salivary gland acini which stained negatively with G S T - T C was also imaged and measured. These areas served as ah intensity threshold to separate stained objects from the background. Mean optical density (staining intensity) for each specimen was calculated by the system using the results of the ten fields. Mean optical density, coefficient of variation, skewness and kurtosis were calculated arid compared for the two tumour groups.  5.3  RESULTS As shown in Figures 6 and 7, the mean optical density for recurrent  pleomorphic adenomas was significantly higher than non-recurrent (primary)  pleomorphic adenomas using analysis of individual data points for all specimens (p<10- Mann-Whitney Test). The p value, when calculated using the mean optical 8  densities of each of the 25 specimens as opposed to all individual data points, including the ten to 15 points measured for each individual specimen, was calculated as p=0.0007 (Mann-Whitney Test - see Appendix 3) The staining intensity of non-recurrent pleomorphic adenomas was generally weak (Plate 3 A.), although there were occasional more intensely staining cells present in myxoid regions (Plate 3 B.) Tumour islands in the ;  recurrent tumours stained stronger, and staining intensity was prominant at the periphery of tumours (Plate 3 C , D., E., F.). The variation in optical density from area to area within a tumour was determined by distribution statistics - coefficient of variation, skewness, and kurtosis. The non-recurrent pleomorphic adenomas demonstrated more uniform or even distribution of GST-re enzyme thoughout a tumour when compared to recurrent pleomorphic adenomas and skew, kurtosis, optical density coefficient of variation between the two tumour types were also statistically significant (p<10 ) -6  (Figures 8, 9, and 10).  5.5  DISCUSSION AND CONCLUSIONS  The results from experiment 1 suggested that cell-poor pleomorphic adenomas tended to contain a higher level of GST-re than those of the cell-rich pleomorphic adenomas. Since recurrent pleomorphic adenomas are mainly of the cell-poor type and the recurrent tumours tend to have much higher recurrence rate than primary pleomorphic adenomas, the results from experiment 1 raised the question whether increased GST-re played a role in the recurrent behaviour of the tumours. The experiment tested this hypothesis by comparing the GST-re staining  intensity between non-recurrent pleomorphic adenomas and recurrent pleomorphic adenomas In order to minimize subjective errors in microscopic grading of the enzyme content by naked eye, this experiment employed a computerized image analysis system. The results confirmed the hypothesis and showed a significantly higher GST-TC  content in recurrent pleomorphic adenomas when compared to non-  recurrent pleomorphic adenomas (p<10- , Figures 6, 7). Quantitative assay of the 8  GST proteins by this method may be of practical use. It has been found that gel filtration cannot distinguish the various transferases due to similarity in molecular size and overlapping binding specificites, and purification is a laborious and perhaps impractical method for the study of these proteins (Hales & Naeim, 1976). Thus, the value of an immunohistochemical approach, particularly with computer image analysis capable of quantifying protein by the detection of the amount of peroxidase reaction product with a computer assisted microscope photometer is thought to be an important consideration. It is generally accepted that failure in complete first removal of pleomorphic adenomas is the main cause for recurrence. It is not known whether intrinsic cellular behaviour also plays a role in the recurrence. An increased number of recurrences has also been considered a factor for malignant transformation, and thus if intrinsic cellular biologic characteristics can be determined at the time of initial surgical therapy, potential for recurrence, malignant transformation, or metastasis may be possible. Cell biology may be of particular significance in the cases of metastasizing pleomorphic adenomas with morphologic features of benign lesions. This is one of a few studies to demonstrate that cell biology of this tumour plays a role in recurrence. Others have investigated the role of basement membrane production and location in the recurrence of pleomorphic adenoma  (Takahashi et al., 1992), and another has investigated the role of a cell line with myoepithelial phenotype and its production of interleukin-6, postulated to have a role in stimulating growth in PAs (Gallo et al., 1992). This, however, appears to be the only study which provides evidence of an enzymatic role in the recurrence of these tumours. The means by which the increased G S T - J C content contributes to recurrence of the tumour is not entirely clear. Increased GST-re content has been reported in a wide variety of premalignant and malignant tumours. In pancreas, an organ morphologically similar to salivary glands, jt has been shown that placental GST is increased in both premalignant and malignant lesions (Satoh et al., 1985; Moore et al., 1985, 1987; Obara et al., 1986). The increase in GST-re content in these premalignant and malignant lesions are believed to provide these cells increased ability to detoxify toxins and hence have growth advantages, though the exact functions of the enzyme remain speculative. It is possible that the increase in GST-re enzyme in pleomorphic adenomas also provide these cells a growth advantage and hence more aggressive behaviour and more likelihood of recurrence. It is interesting to observe that recurrent pleomorphic adenomas not only demonstrated a higher GST-re content but also demonstrated a larger area to area variation in the staining intensity, and higher staining usually presented in peripheral portions of the tumours. Such variation would suggest the possibility of the evolution of focal clones of cells containing higher GST-re content and development of these focal cells would thus pose even higher recurrence risk. This is consistent with the fact" that recurrent tumours have much higher rate of further recurrence than nonrecurrent pleomorphic adenomas. Bioactivation of a number of substrates including haloalkanes and haloakenes, and amino acid S-conjugates forming mutagenic compunds which form DNA adducts has been found to be glutathione dependent with GSTs playing  a role in this reaction (Monks et al., 1990). This may be a mechanism whereby initiation of tumour formation may occur. As well, glutathione conjugation reactions are categorized as substitution and addition reactions, ie.: GSH + RX -> GRS + HX (substitution reaction) GSH + R <-> GSRH  (addition reaction)  In the latter reaction, the conjugation of glutathione is considered to be reversible. If a particular xenpbiotic which may be potentially toxic is initially bound by glutathione, but may also be released again, this type of reaction may actually act in transport or storage of certain substrates to other areas of the body (Monks et al., 1990). Again, the GSTs responsible for catalyzing these reactions likely play a role in this process. Elevated levels of this enzyme in an environment of high levels of cellular glutathione with substrate availablility could potentiate transport or storage in particular cells. It has been found that some compounds can be toxic to cells, both free or in the GSH-bound form. This was found in the case of allyl isothiocyanate toxicity to liver cells in both the bound and unbound form. The conjugation reaction has also been found to produce conjugates capable of alkylating proteins found to be mutagenic. Other substrates, such as aldehydes have been found to conjugate, and potentially fall apart, leaving the aldehyde to react with other GSH molecules, or possibly with nitrogen nucloephiles in DNA (Monks etal., 1990). As well, it is thought that this reversibility of binding may act in ditributing GSH conjugates to various parts of the body, which may also result in cellullar alteration in different tissues. Although the role of glutathione in its conjugation of xenobiotics is principally a protective one in sequestering and aiding in excretion of toxins, the alternative findings raise the possibility that it may also enhance carcinogenesis by increasing toxicity and mutagenicity of carcinogens. The enzymes involved in these conjugation reactions, the transferases, might thus be implicated by association.  The fact that primary recurrent pleomorphic adenomas showed significantly higher  GST-TC  content than the non-recurrent pleomorphic adenomas suggests that  this enzyme may be used as a tumour marker in the prediction of the recurrent behaviour of this tumour. Since there is overlap in the enzyme content between the non-recurrent pleomorphic adenomas and recurrent pleomorphic adenomas, future studies employing larger number of cases are needed to determine the 'baseline levels', above which there is an increased likelihood for recurrence. Computer image analysis may be particularly useful in the study of salivary gland tumours in which histogenesis is still very controversial. Using this form of analysis, individual cell analysis under high magnification may be carried out and quantification of enzyme levels determined. This will not only provide the "baseline levels needed for this tissue, but levels found in different tumour types as well as individual cell types. Further studies in the classfication of salivary gland tumours and possible identification of cells of differing tumour types by determination of enzyme levels and other markers used to label specific cells could be considered utilizing this system. The varied morphologic characteristics of cells such as the myoepithelial cell, and its role in different tumour types might be better understood if further studies of its enzyme expression, including different isoforms, is better elucidated in normal salivary gland and salivary gland tumours. The same can be considered for the cells of the ductal system.  BIBLIOGRAPHY Abdel-Rahman S.Z. Glutathione and related enzymes in fascioliasis before and after treatment with biothionol. J Trop Medand Hyg 93(5): 337-40, 1990 Aceto A., Martini F., Dragani B., Bucciarelli T., Sacchetta P., and Di llio C. Purification and characterization of glutathione transferase from psoriatic skin. Biochem Med and Met Biol 48: 212-218, 1992 Ahmad H., Singh S.V., Medh R.D., Ansari G.A.S., Kurosky A., and Awasthi Y. Differential expression of alpha, mu and pi classes of isoenzymes of glutathione S-transferase in bovine lens, cornea, and retina. Arch of Biochem and Biophys 266(2): 416-26, 1988 Ahmad H., Singhal S.S., Saxena M., and Awasthi Y.C. Characterization of two novel subunits of the alpha-class glutathione S-transferases of human liver. Biochim et Biophysi Acta 1161: 333-6, 1993 Arias I.M. et al. On the structure, regulation, and function of ligandin. In Glutathione: Metabolism and Function I.M. Arias and W.B. Jacoby, (eds.), Raven, New York, pp. 175-88,1976 Awasthi S. and Ahmad F. Reversed-phase chromatographic method for specific determination of glutathione in cultured malignant cells. J of Chromatography 584: 167-73, 1992 Awasthi S., Srivastava S.K., Ahmad F., Ahmad H., and Ansari G.A.S. Interactions of glutathione S-transferase-pi with ethacrynic acid and its glutathione conjugate. Biochim et Biophys Acta 1164; 173-8, 1993 Awasthi S., Singhal S.S., Srivastava S.K., and Awasthi Y.C. Purification and characterization of glutathione S-transferase of murine ovary and testis. Arch of Biochem and Biophys 301(1): 143-50, 1993 Barbacid M. ras genes. Annu Rev Biochem 56: 779-829, 1987 Batsakis J.G. Tumors of the Head and Neck. Clinical and Pathological Considerations. 2nd ed. Williams & Wilkins, Baltimore/London, 1979, p.26 Battermann J.J. and Mijnheer B.J. The Amsterdam fast neutron radiotherapy project-a final report. Int J Radial Oncol Biol PhysM: 2093-9, 1986 Beer D.G.and Pitot H.C. Biological markers characterizing the development of preneoplastic and neoplastic lesions in rodent liver. Arch Toxicol 10(5): 6880, 1987 Bell D.A., Thompson C . L , Taylor J., Miller OR., Perera F., Hsieh L.L., and Lucier G.W. Genetic monitoring of human polymorphic cancer susceptibility genes by  polymerase chain reaction: application to glutathione transferase mu. Env Health Persp 98:113-7, 1992 Bennett C.F., Spector D.L., and Yeoman L.C. Nonhistone protein BA is a glutathione S-transferase localized to interchromatinic regions of the cell nucleus. J Cell Biol 102:600-9, 1986 Benson A.M .et al Elevation of hepatic glutathione S-transferase activites and protection against mutagenic metabolites of benzo[a]pyrene by dietary antioxidants. Cancer Res. 38: 4486, 1978 Biaglow J., Mitchell J.B., and Held K. The Importance of peroxide and superoxide in the x-ray response. Int J Rad Oncol Biol Phys 22: 665-9, 1992 Birek C , Lui E., and Dardick I., C-fos uhderexpression in salivary gland tumours as measured by in situ hybridization. Am J Pathol 142: 917-23, 1993 Birek C , Lui E., Jordan R.C.K., and Dardick I., Analysis of c-erb-2 oncogene amplification as measured by differential polymerase chain reaction. Oral Oncol Eur J Cancer 30B: 47-50, 1994 Bisgaard H.C. and Thorgeirrsson S.S. Evidence for a common cell of origin for primative epithelial cells isolated from rat liver and pancreas. J of Cellular Phys 147(2): 333-43, 1991 Bishop J.M. Molecular themes in oncogenesis. Cell 64: 235-48, 1991 Blacker K., Olson E., Vessey D.A., and Boyer T.D. Characterization of glutathione S-transferase in cultured human keratinocytes. J of Invest Derm 97(3): 442-6, 1991 Blanck A., Hansson T., Eriksson L . C , and Gustafsson J. On mechanisms of sex differences in chemical carcinogenesis: effects of implantation of ectopic pituitary grafts on the early stages of liver carcinogenesis in the rat. Carcinogenesis 5: 1257-62, 1984 Blanck A., Hansson T., Gustafsson J., and Eriksson L.C. Pituitary grafts modify sex differences in liver tumor formation in the rat following initiation with diethylnitrosamine and different promotion regimens. Carcinogenesis 7(6): 981-5, 1986 Board P.G., Webb G.C. and Coggan M. Isolation of a cDNA clone and localization of the human glutathione S-transferase 3 genes to chromosome bands 11q12 and 12q13-14 Ann Hum Genet 53: 205-13, 1989 Bogaards J.J.P., van Ommen B., and Bladeren P.J. Purification and characterization of eight glutathione S-transferase isoenzymes of hamster. Biochem J. 286:383-8,1992  Booth J., Boyland E., and Sims P. An enzyme from rat liver catalysing conjugations and glutathione Biochem J. 79: 516-24, 1961 Boutwell R.K. Some biological effects of skin carcinogenesis. Prog Exp Tumor Res. 4: 207-250, 1964 Boutwell R.K. Model systems for defining initiation, promotion, and progression of skin neoplasms. In: Skin Carcinogenesis: Mechanisms and Human Relevance. Alan R. Liss Inc., pp. 3-15 Boyland E., and Chasseaud L F . The role of glutathione and glutathione Stransferases in mercapturic acid biosynthesis. AdvEnzymol 32 :173-219, 1969 Bravard A., Hoffshir F., Ricoul M., Cassingena T., Estrade S., Luccioni C , and Dutrillaus B. Alterations of the glutathione cycle enzymes during and after SV40-transformation of human fibroblasts. Carcinogenesis 14(1): 21-4,1993 British Columbia Cancer Agency Treatment Policies Manual, British Columbia Cancer Agency, Vancouver British Columbia, Canada, Spring 1992 Burgering B.M., Medema R.H., Maassen J.A., van de Wetering M.L., van der Eb A.J./McCormick F., and Bos J.L. Insulin stimulaiton of gene expression mediated by p21 ras activation. EMBO 10: 1103-1109, 1991 CampbelU.A.H., Corrigall A.V., and Kirsch R.E. Immunohistologic localization of alpha, mu, and pi class glutathione S-tansferases in human tissue. Cancer 67:1608-13, 1991 Canuto R.A, Muzio G., Maggiora M., Biocca M.E., and Diazani M.U. Glutathione S-transferase, alcohol dehydrogenase and aldehyde reductase activities during diethylnitrosamine^carcinogenesis in rat. Cancer Letters 68:177-83, 1993 Catterall M. and Errington R.D. The implications of improved treatment of malignant salivary gland tumors by fast neutron radiotherapy. Int J Radiat Oncol Biol Phys 13: 1313-8, 1987 Cazenave L A . , Moscow J.A., Myers C.E., and Cowan K.H. Glutathione Stransferase and drug resistance. Drug Res in Cancer Therapy 48: 171-187, 1989 Cholon A., Giaccia A.J., Lewis A.D., Hickson I., and Brown J.M. What role do glutathione S-transferases play in the cellular response to ionizing radiation? Int J Radiation Oncology Biol Phys 22: 759-63, 1991 Clapper M.I., Hoffman S., Carp N., Watts P., Seestaller L.M., Weese J . L , and Tew K. Contribution of patient history to the glutathione S-transferase activity of human lung, breast and colon tissue. Carcinogenesis 12(10): 1957-61, 1991.  Corrigall A.V. and Kirsch R.E..GIutathione S-transferase distribution and concentration in human organs. Biochem Int 16: 443-8, 1988 Coursin D.B., Cihla H.P., Oberlyey T.D., and Oberley L W . Immunolocalization of antioxidant enzymes and isozymes of glutathione S-transferase in normal rat lung. Am J of Physiology 263(6 pt 1): L679-91, 1992 Daly H.M., Tee L.B.G., Oates P.S., Morgan R.G.H., and Yeoh G.C.T. Glutathione S-transferase (mu class) as an early marker of azaserine-induced foci in the rat pancreas. Carcinogenesis 12(7): 1237-40, 1991 Dawe C.J. et al. Variation in polyoma virus genotype in relation to tumour induction in mice. Characterization of wild type strains with widely differing tumor profiles. Am J Pathol 127: 243-61, 1987 de Carhargo J.L. and Rodrigues M.A. Placental form of glutathione Stransferase as a marker for cervical neoplasia. Acta Cytologica 36(5): 850, 1992 Degen G.H .and Neumann H.G. The major metabolite of aflatoxin B1 in the rat is a glutathione conjugate. Chem- Biol Interact 22: 239-55, 1978 Dekant W. and Vamavakas S. Glutathione-dependant bioactivation of xenobiotics. Xenobiotica 23(8); 873-7, 1993 Di llio C , Del Boccio G., Massoud R., and Federici G. Glutathione transferase of human breast is closely related to transferase of human placenta and erythrocytes. Biochem Int 13(2):263-9,1986 Di llio C , Del Boccio G., Aceto A., Casaccia R., Mucilli F., and Federici G. Elevation of glutathione transferase activity in human lung tumour. Carcinogenesis 9(2): 335-40, 1988 Di llio C , Del Boccio G., Aceto A., and Federici G. Alteration of glutathione transferase isoenzyme concentrations in human renal carcinoma. Carcinogenesis 8(6): 861-4, 1987 Dimery I. W., Legha S.S. Shirian M., and Hong W.K. Fluorouracin, doxorubicin, cyclophosphamide, and cisplatin combination chemotherapy in advanced or recurrent salivary gland carcinoma. J Clin Oncol 8(6): 1056-1062, 1990 Dipple A., Michejda C.J., and Weisburger E.K.. Metabolism of chemical carcinogens. Pharm and Ther 27(3): 265-96, 1985 Dirr H.W., Mann K., Huber R., Ladenstein R., and Reinemer P. Class pi glutathione S-transferase from pig lung. Purification, biochemical characterization, primary structure and crystallization. Eur J Biochem 196: 6938, 1991  Dragan Y.R., Rizvi T., Xu Y-H., Hully J.R., Bawa N., Campbell H.A., Maronpot R.R., and Pitot C. An initiation-promotion assay in rat liver as a potential complement to the 2-year carcinogenesis bioassay. Fundamental and Applied Toxicology 16: 525-47, 1991 Dutton G.Jj Developmental aspects of drug conjugation, with special reference to glucuronidation. Annu Rev Pharmacol Toxicol 18: 17-35, 1978 Efferth T., Mattern J., and Volm M. Immunohistochemical detection of P glycoprotein, glutathione S-transferase and DNA topoisomerase II in human tumors. Oncology 49: 368-375, 1992 Eimoto H., Tsutsumi M., Nakajima A., Yamamoto K., Takashima Y., Maruyama H., and Knoishi Y. Expression of the glutathione S-transferase placental form in human lung carcinomas. Carcinogenesis 9(12):2325-2327, 1988 Eneroth C M . Mixed tumors of major salivary glands: Prognostic role of capsular structure. Ann Otol 74: 944-54, 1965 Evans R.W. and Cruickshank A.H. Epithelial tumours of the salivary glands. Philedelphia W.B. Saunders Co, 1970, pp. 165-225 Eveson J.W. Troublesome Tumours 2: Borderline Tumours of Salivary Glands. J.Clin Pathol 45: 369-377, 1992 Fearon E.R. and Vogelstein B. A genetic model for colorectal tumorigenesis. Cell 61: 759-67, 1990 Feinfeld D.A. et al. Ligandin in perfusate from transplanted kidneys: A test for tubular necrosis. Nephron 21: 38-41, 1978 • • n Foote F.W. and Frazell E.L. Tumors of the major salivary glands. Cancer 6: 1065-1133, 1953 t  Freeman S.B., Kennedy K.S., Parker G.S., and Tatum S.A. Metastasizing pleomorphic adenoma of the nasal septum. Arch Otol H & N Surg 116: 13313,1990 Freund R., Dawe C.J., Carroll J.P., and Benjamin T . L Changes in frequency , morphology, and behaviour of tumors induced in mice by a polyoma virus mutant with a specifically altered oncogene. Am J Pathol 1 41(6): 1409-25, 1992 Fu K.K., Leibel S.A., Levine M.L, Friedlander L.M., Boles R., and Phillips T.L. Carcinoma of the major and minor salivary glands: Analysis of treatment results and sites and causes of failures. Cancer 40: 2882-2890, 1977  Gajewska J. and Szczypka M. Role of pi form of glutathione S-transferase (GST-pi) in cancer: a minireview. Materia Medica Polona Fasc. 1(81); 45-9, 1992 Gallo O., Bani D., Toccafondi G., Almerigogna F., and Storchi O.F. Characterization of a novel cell line from pleomorphic adenoma of the parotid with myoepithelial phenotype and producing interleukin-6 as an autocrine growth factor. Cancer 70(3): 559-68, 1992 Goudot P, Auriol M., Chomette G., Vaillant J.M., Guilbert F. Pleomorphic adenoma of the salivary glands. Impact of the myxoid component on the prognosis. Revue De Stomatologie Et De Chirurgie Maxilo-Faciale. 90: 11922, 1989. Griffin T., Blasko J., and Laramore G. Results of fast neutron beam radiotherapy, pilot studies at the University of Washington. In: Barendsen G.W., Broerse J.J., Berue K. (eds) Highe- LET radiation in clinical therapy. EurJ Cancer(Suppl): 23-9, 1979 Griffin B.R., Laramore G.E., Russell K.J., Griffin T.W., and Eenmaa J. Fast neutron radiotherapy for advanced malignant salivary gland tumors. Radiother Oncol 12: 105-11, 1988 GriffinT.W., Pajk T.F., Laramore G.E., et al. Neutron vs photon irradiation of inoperable salivary gland tumors: results of an RTOG_MRC cooperative randomized study. Int J Radiat Oncol Biol Phys 15: 1085-90, 1988 Guengerich F.P., Shimada T., Raney K.D. Yun C-H., Meyer D.J., Ketterer B., Harris T.M., Groopman J.D., and Kadlubar F.F. Elucidation of catalytic specificities of human cytochrome P450 and glutathione S-transferase enzymes and relevance to molecular epidemiology. Env Health Persp 98: 75-80, 1992 Gulick A.M., Lathrop Goihl A., and Fahl W.E. Structural studies on human glutathione S-transferase pi. J of Biological Chemistry 267: 18946-52,Sept 1992 GullickW.J. Inhibitors of growth factor receptors. In: Genes and cancer., Carney D.S. and Sikora K., eds., Chichester: Wiley, p. 264-73, 1990. Guthenberg C and Mannervik B Glutathione S-transferase (transferase-pi) from human placenta is identical or closely related to glutathione S-transferase (transferase rho) from erythrocyte. Biochim, Biophys Acta 661: 244-60, 1981 Hales B.G. and Neims A.H. A sex difference in hepatic glutathione Stransferase B and the effect of hypophysectomy. Biochem J 160: 223-9, 1976 Halliwell B. How to characterize a biological antioxidant.(Review). Radical Res Comm 9(1): 1-32, 1990.  Free  Halliwell B. Drug antioxidant effects. A basis for drug selection? Drugs 42(4): 569-605, 1991 Halliwell B. Reactive oxygen species in living systems: source, biochemistry, and role in human disease. (Review) Am J of Med 91 (3C): 14S-22S, 1991 Halliwell B and Aruoma OI. DNA damage by oxygen-derived species. Its mechanism and measurement in mammalian systems. FEBS Letters (Federation of European Biochemical Societies) 281(1-2): 9-19, 1991 Hanada K., Ishikawa H., Tamai K., Hashimoto I., and Sato K. Expression of glutathione S-transferase pi in malignant skin tumors. J of Derm Sci 2:18-23, 1991 Hendrich S. and Pilot H.C. Enzymes of glutathione metabolism as biochemical markers during hepatocarcinogenesis. Cancer and Metastasis Rev 6: 155-78, 1987 Hida T et al. Serum glutathione S-transferase-pi level as a tumor marker for non-small cell lung cancer. Potential predictive value in chemotherapeutic response. Cancer 73(5): 1377-1382, 1994 Hill K.A., Hunt T.W., Jones R., Hoover R.L, Burk R.F. Metabolism of nitroglycerine by smooth muscle cells. Involvement of glutathione and glutathione S-transferase. Biochem Pharmacol 43(3): 561-6, 1992 Hirata S., Odajima T., Kohama G., Ishigaki S., Niitsu Y. Significance of glutathione S-transferase-pi as a tumour marker in patients with oral cancer. Cancer 70(10): 2381-7 Hochhauser D. and Harris A . L Drug Resistance. Brit Med Bull 47(1): 178-96, 1991 Horiuchi J., Shibuya H., Suzuki S., Takeda M, and Takagi M. The role of radiotherapy in the management of adenoid cystic carcinoma of the head and neck. In J. Radiat Oncol. Biol. 13: 1135-41. 1987 Horn-Ross P.L., West D.W., and Brown S.R. Recent trends in the incidence of salivary gland cancer. Int J of Epidemiology 20(3): 628-33, 1991 Hrubisko M., McGown A T . and Fox B.W. The role of metallothionine, glutathione, glutathione S-transferases and DNA repair in resistance to platinum drugs in a series of L1210 cell lines made resistant to anticancer platinum agents.Biochem Pharmacol 45(1): 253-6, 1993 Hsu S,. Raine L , and Fanger H. A comparative study of the peroxidaseantiperoxidase method and an avidin-biotin complex method for studying polypeptide hormones with radioimmunoassay antibodies. Am J Clin Pathol 75(5): 734-8, 1981  Hsu S. and Fanger H. Use of avidin-biotin peroxidase complex (ABC) in immunoperoxidase techniques: a comparison between ABC and unlabelled antibody (PAP) procedures. J of Histochem and Cytochem 29(4): 577-80, 1981 Hunter T. Cooperation between oncogenes. Cell 64: 249-70, 1991 loannides C. and Parke D.V. Toxicological aspects of glutathione. Xenobiotica Editorial 23(8): 821-2, 1993 Ito N., Tatematsu M., Hasegawa R., and Tsuda H.. Medium term bioassay system for detection of carcinogens and modifiers of hepatocarcinogenesis utilizing the GST-P positive liver cell focus as an endpoint marker. Toxicol Pathol 17(4 - p a r t i ) : 630-41, 1989 Ito N., Hasegawa R., Imaida K., Masui T., Takahashi S., and Shirai T. Pathological markers for non-genotoxic agent-associated carcinogenesis. Toxicology Letters 64/65:613-20,992 Jackson G.L., Luna M.A., and Byers R.M. Results of surgery alone and surgery combined with postoperative radiotherapy in the treatment of cancer of the parotid gland. Am J Surg 146: 497-500, 1983 Jakoby W.B. The glutathione S-transferases: A group of multifunctional detoxification proteins. Adv Enzymol 46: 383-414, 1978 Jakoby W.B. Enzymatic Basis of Detoxicaton Vol II Jakoby W.B.(ed.) Academic Press, New York, 1980 Jernstrom B., Dock L., Hall M., Mannervik B., Tahir M.K., and Grover P.L. Glutathione transferase catalyzed conjugation of Benzo[a]Pyrene 7,8-Diol 9,10Epoxide with a glutathione in human skin. Chem-Biol Interactions 70:173-80, 1989 Jones, A.S., Phillips D.E., Cook J.A. and Helliwell T.R. A randomized phase II trial of epirubicin and 5-fluorouracil versus cisplatin in the palliation of advanced and recurrent malignant umour of the salivary glands. Br J Cancer 67: 112114, 1993 Jordon R.C.K., Dardick I., Lui E., and Birek C. Demonstration of c-erB-2 overexpression in salivary gland neoplasms by in situ hybridisation. J Oral Pathol Med 23: 226-31, 1994 Kahn H.J., et al. Expression of neu oncogene in pleomorphic adenomas of salivary gland. Arch Pathol Lab Med 116:80-3, 1992  Kashiwada M. et al. Purification and characterization of acidic form of glutathine S-transferase in human fetal livers: high similarity to placental form. J of Biochem 1 10(5): 743-7, 1991 Kelloff G.J. et al. Chemopreventive drug development:perspectives and progress (review). Cancer Epidemiology, Biomarkers and Prevention 3(1): 8598, 1994 Kennedy A.R. Is there a critical target gene for the first step in carcinogenesis? Env Health Persp 93: 199-203, 1991 Kernohan N.M., Blessing K., King G., Corbet I.P., Miller I.D. Expression of cerbB-2 oncoprotein in salivary gland tumours: an immunohistochemical study. J Pathol 163: 77-80, 1991 Ketterer B., Tan K.H., Meyer D.J, and Coles B. Glutathione transferases: A possible role in the detoxification and repair of DNA and lipid hydroperoxides. Mutation Res 214: 433-40, 1989 Ketterer B, et al. Glutathione transferases: A possible role in the detoxification and repair of DNA and lipid hydroperoxides. Env Health Persp 98: 149-63, 1992 Ketterer B., Harris J.M., Talaska G., Meyer D.J., Pemble S.E., Taylor J.B., Lang N.P., Kadlubar F.F. The human glutathione S-transferase supergene family, its polymorphism, and its effects on susceptibility to lung cancer. Env Health Persp 98: 87-94, 1992 Kodate C , Fukushi A., Narita T., Kudo H., Soma Y., and Sato K. Human placental form of glutathione S-transferase(GST-pi) as a new immunohistochemical marker for human colonic carcinoma. Japanese J of Cancer Res 77(3): 226-9, 1986, 1987 Konohana A., Konkhana I., Shhroeder W.T., O'Brien W.R., Amagai M., Greer J., Shimizu N., Gammon W.R., Siciliano M.J., and Duvic M. Placental glutathione S-transferase-pi mRNA is abundantly expressed in human skin. J of Invest Derm 95(2): 119-26,1990 Krolls S.O. and Boyers C. Mixed tumours of the salivary glands. Long term follow up. Cancer 30 : 276-81, 1972. Kumar K, Thangaraju M., and Sachdanandam P. Changes in antioxidant system in the blood of postmenopausal women with breast cancer. Biochem Int 25(2): 371-80, 1991 Lewis A.D., Forrester L.M., Hayes J.D, Wareing C.J., Carmichael J., Harris A.L., Mooghen M., arid Wolf O R . Glutathione S-transferase isoenzymes in human tumours and tumour derived cell lines. Br J Cancer 60(3): 327-31, 1989  Lewis A.D., Hayes J.D., and Wolf C.R. Glutathione and glutathione-dependant enzymes in ovarian adenocarcinoma cell lines derived from a patient before and after the onset of drug resistance: Intrinsic differences and cell cycle effects. Carcinogenesis 9(7): 1283-7, 1988 Li Y et al. MTras T24, a metallothionein-ras fusion gene, modulates expression in cultured rat liver cells of two genes associated with in vivo liver cancer. Proc Natl Acad Sci USA 85:344-8, 1988 Licitra L, Marchini S., Spinazze S., Rossi A., Rocca A., Grandi C , and Molinari R. Cisplatin in advanced salivary gland carcinoma. Cancer 68(9): 1874-7, 1991 Liu P.T., loannides C , and Symons A.M. Role of tissue glutathione in prevention of surgical trauma. Xenobiotica 23(8); 899-911, 1993 Lo J-F, Wang H-F, Tarn M.F., and Lee T-C. Glutathione S-transferase pi in an arsenic-resistant Chinese hamster ovary cell line. Biochem J 288:977-81, 1992 Mantle T.J., Pickett C.B., and Hayes J.D. Glutathione S-transferases and Carcinogenesis. Taylor and Francis, London, 1987 Maguire N.G., Drozd K., Luff R.D., and Kantor R.R.S. Immunohistochemical localization of glutathione S-transferase in preneoplastic and neoplastic lesions of the human cervix. Acta Cytologica 35(1): 94-9, 1991 Malkin, A. Tumour Markers. In: The Basic Science of Oncology. I.F. Tannock and R.P. Hills, eds., Pergammon Press, New York, 1987, pp. 192-3. Mannervik B., Alin P., Guthenberg C , Jensson H., Tahir M.K., Warholm M., and Jornvall H. Identification of three classes of cytosolic glutathione transferase common to several mammalian species: Correlation between structural data and enzymatic properties. Proc Natl Acad Sci USA Biochem 82:7202-6, 1985 Mannervik B. and Danielson U.H. Glutathione transferases-structure and catalytic activity (review). Crit Rev Biochem 23(3); 283-37, 1988 Mannervik B., Awasthi Y.C., Board P.G., Hayes J.D., Di llio C , Ketterer B., Listowsky I., Morgenstern R., Muramatsu M., Pearson W.R., Pickett C.B., Sato K., Widersten M., and Wolf C.R. Nomenclature for human glutathione Stransferases. Bioch J 2 82: 305-8, 1992 Manoharan et al Promotor-glutathione-s-transferase Ya cDNA hybrid genes: Expression and conferred resistance to an alkylating molecule in mammalian cells. J Biol Chem 262: 3739-45, 1987  Mahoharan T.H., Gulick A.M., Puchalski R.B., Servais A.L., and Fahl W. Structural studies on human glutathione S-transferase pi. J of Biol Chem 267(26): 18940-5, 1992 Mantle T.J., Pickett C.B., and Hayes J.D. (eds), Glutathione S-transferases and Carcinogenesis, Taylor & Francis, London, 1987 Marnett L.J. Peroxyl free radical: potential mediators of tumor initiation and promotion. Carcinogenesis 8: 1365-73, 1987 McFarland J . The histopathologic prognosis of salivary gland mixed tumors. Am J Med Sci 203: 502-519, 1942 McGown A T . , Hrubisko M., and Fox B.W. The role of metallothionein, glutathione, glutathione S-transferases and DNA repair in resistance to platinum drugs in a series of L1210 cell lines made resistant to anticancer platinum agents. Biochem Pharmacol 45(1): 253-6, 1993 McLellan L.I., Harrison D.J., and Hayes J.D. Modulation of glutathione Stransferases and glutathione peroxidase by the anticarcinogen butylated hydroxyanisole in murine extrahepatic organs. Carcinogenesis 1 3(12): 225561, 1992 McNaney D.M, McNeese M.D., Guillamondegui O., Fletcher G.H., and Oswald M.J. Postoperative irradiation of malignant epithelial tumors of the parotid. Int J Rad Oncol Biol Phys 9: 1289-95, 1983 McQuaid S., McCann S., Daly P., Lawlor E., and Humphries P. Observations on the transcriptional activity of the glutathione S-transferase pi gene in human haematological malignancies and in the peripheral leukocytes of cancer patients under chemotherapy. Biochem Pharmacol 45(1); 253-6, 1993 Meikle I., Hayes J.D., and Walker S.D. Expression of an abundant alpha-class glutathione S-transferase in bovine and human adrenal cortex tissues. J of Endocrinology 132: 83-92, 1992 Meister A. Glutathione, ascorbate, and cellular protection. Cancer Res Suppl 54: 1969s-1975s, 1994 Mekhail-lshak K., Medina D., and Batist G. Biochemical characteristics of mouse mammary tissues, preneoplastic lesions and tumors. Carcinogenesis 10(12): 2363-66,1989 Meyer D.J., Coles B., Pemble S.E., Gilmore K.S., Fraser G.M., and Ketterer B. Theta, a new class of glutathione transferases purified from rat and man. Biochem J 274: 409-14, 1991  Miller E.C.and Miller J.A. Biochemical mechanisms of chemical carcinogenesis. In: The Molecular Biology of Cancer. Busch H (ed), Academic Press, N.Y., pp 377-402, 1974 Miller AS et al. Search for correlation of radon levels and incidence of salivary gland tumours. Oral Surg, Oral Med, Oral Path 7 5(1): 58-63, 1993. Moore M.A., Satoh K., Kitahara A., Sato K., and Ito N. A protein cross-reacting immunohistochemically with rat glutathione S-transferase placental form as a marker for preneoplasia in Syrian hamster pancreatic and hepatocarcinogenesis. Japan J Cancer Research 76(1-4): 1-4, 1985 Moore MA, et al. Immunohistochemically demonstrated increase in glutathione S-transferase species in propylnitrosamine-induced focal proliferative and neoplastic Syrian hamster pancreatic lesions. Virchows Arch [Cell Path] 52: 479-88, 1987 Morgenstern R., De Pierre J.W., and Jornvall H. Microsomal glutathione transferase. Primary structure. Biol Chem 260: 13976-83, 1985 Morrow C.S., Goldsmith M.E .,and Cowan K.H. Regulation of human glutathione S-transferase pi gene transcription: influence of 5'-flanking sequence and trans-activating factors which recognize AP-1 binding sites. Gene 88: 215-25, 1990 Morrow C.S., Chiu J., and Cowan K.H. Post-transcriptional control of glutathione S-transferase pi gene expression in human breast cancer cells. J of Biol Chem 267(15): 10544-50, 1992 Moscow J.A., Townsend A.J., Goldsmith M.E., Whang-Peng J., Vickers P.J., Poisson R., Legault-Poisson S., Myers C.E., and Cowan K. Isolation of the human anionic glutathione S-transferase cDNA and the relation of its gene expression to estrogen-receptor content in primary breast cancer. Proc Nattl Acad Sci USA 85: 6518-22, 1988 Mukhtar H. and Bend J.R. Serum glutathione S-transferases: Perinatal development, sex difference and effect of carbon tetrachloride administration on enzyme activity in the rat. Life Sci 21: 1277-1286, 1977 Murphy D., McGown A., Hall A., Cattan A., Crowther D., and Fox B.W. Glutathione S-transferase activity and isoenzyme distribution in ovarian tumour biopsies taken before or after cytotoxic chemotherapy. Br J of Cancer 66: 93742, 1992 Nachtomi E The metabolism of ethylene dibromide in the rat. The enzymatic reaction with glutathione in vitro. Biochem Pharmacol 19: 2853-60, 1970 Naeim S., forsberg M.I.Jr., Waisman J., Coulson W.F. Mixed tumors of the salivary glands. Arch Pathol'Lab Med 100: 271-75, 1976.  Nagao K. et al. Histopathologic studies on carcinoma in pleomorphic adenoma of the parotid gland. Cancer 48(1): 113-20, 1981 Nakagawa K., Yokota J. Wada M., Sasaki Y., Fujiwara Y., Sakai M., Muramatsu M., Terasaki T., Tsunokawa Y., TeradaM., et al. Levels of glutathione Stransferase pi mRNS in human lunc cancer cell lines correlate with the resistance to cisplatin and carboplatin. Japanese J Cancer Res 79(3): 301-4, 1988 National Cancer Institute (NCI). Fast neutron radiation therapy in the United States: A twenty year NCI sponsored research program (NTCWG annual report, 1991), National Cancer Insitiute, Bethesda, MD 20892, USA, 1991. Nicholson D.W., Ali A., Vaillancourt J.P., Calaycay J.R., Mumford R.A., Zamboni R.J., and Ford-Hutchinson W. Purification to homogeneity and the N-terminal sequence of human leukotriene C4 synthetase: a homodimeric glutathione Stransferase composed of 18-kDa subunits. Proc Natl Acad Sci USA 90: 201519, 1993 Nishihira T. Expression of glutathione S-transferase-pi messanger RNA in human esophogeal cancer. Cancer 67(10): 2560- 4, 1991. Nishihira T. Glutathione S-transferase in human esophogeal carcinoma. Anticancer Res 13(2): 363-8, 1993 Nishinaka T., Kodaka R., Nanjo H., Terada T., Mizoguchi T., and Nishihara T. Purification and characterization of glutathione S-transferase isozymes in dog lens. Int J Biochemistry 24(11): 1737-42, 1992 North C.A., Lee D-J., Piantadusi S., Zaharak M., Jhons M.E. Carcinoma of the , major salivary glands treated by surgery plus post-operative radiotherapy. Int J Radiat Oncol Biol Phys 18: 1319-26, 1990 Obara T., et al. Comparative histochemical investigation of the glutahtione Stransferase placental form and gamma glutamyl transpeptidase during Nnitrosobis (2-hydroxyprophyl) amine-induced pancreatic carcinogenesis in hamsters. Carcinogenesis 7: 801-5, 1986 Oberly T.D., Oberley L.W., Slattery A.F., and Elwell J.H. Immunohistochemical localization of glutathione S-transferase and glutathione peroxidase in adult Syrian hamster tissues and during kidney development. Am J Pathol 139(2): 356-69 Oude Eleferink P..J, Bakker C.T.M., and Jansen L.M. Glutathione-conjugate transport by human colon adenocarcinoma cells(Caco-2 cells). Biochem J 290: 759-64, 1993  Ozono S. et al. Immunohistochemical localization of estradiol, progesterone, and progesterone receptor in human salivary glands and salivary adenoid cystic carcinoma. Cell Structure and Function 17(3): 169-75, 1992 Pacific G.M., Warholm M., Guthenberg C , Mannervik B., and Rane A. Organ distribution of glutathione transferase isoenzymes in the human fetus: Differences between liver and extrahepatic tissues. Biochem Pharmacol 35(9): 1616-9, 1986 Palcic B., Susnik B., Garner D., and Olivotto I. Quantitative evaluation of malignant potential of early breast cancer using high resolution image cytometry. J of Cell Biochem 17G: 107-113, 1993 Parker L.M., LLauren D.J., Hammock B.D., Winder B., and Hinton D.E. Biochemical and histochemical Properties of hepatic tumours of rainbow trout, Oncorhynchus Mykiss Carcinogenesis 14(2): 211-7, 1993 Patey D.H. and Thackray A.C. The treatment Of parotid tumours in the light of a pathological study of parotidectomy material. Br J Surg 45: 477-87, 1958 Perera M.I.R. et al. Free radical injury and liver tumor promotion. Toxicol Pathol 15: 51-59, 1987 Peters W.H., Nagengast F.M., and Wobbes T. Glutathione S-transferases in normal and cancerous colon tissue. Carcinogenesis 10(12): 2371-4, 1989 Peters W.H.M., Nagengst F.M., and Wobbes T. Glutathione S-transferases in normal and cancerous human colon tissue. Carcinogenesis 10(12): 2371-4, 1989 Peters W.H.M., Roelofs H.M.J., Nagengast F.M., and Van Tongeren J.H.M. Human intestinal glutathione S-transferases. Biochem J 257: 471-6, 1989 Periano C. and Jones C A . The multistage concept of carcinogenesis. In: The Pathobiology of Neoplasia. Sirica AE, (ed.) New York: Plenum Press; p 131-48, 1989 Pitot H.C.et al. Regulation of the expression of some genegenes for enzymes of glutahtione metabolism in hepatotoxicity and hepatocarcinogenesis. Toxicol and Applied Pharm 9 7(1): 23-34, 1989 Pinto H.A. and Jacobs C. Chemotherapy for recurrent and metastatic head and neck cancer. Hematol/Oncol CI of NA 5(4): 667-686, 1991 Ploeman J.H.T.M.,et al. Ethacrynic acid and its glutathione conjugate as inhibitors of glutathione S-transferase. Xenobiotica 23(8); 913-23, 1993 Poulin N, Harrison A, and Palcic B. Quantitative precision of an outomated image cytometric system for the measurement of DNA content and distribution  in cellslabelled with fluorescent nucleic acid stains. Cytometry 16: 227-235, 1994 Power CA et al Carcinogen related expression of glutathione transferase-p. In: Glutathione S-Transferases and Carcinogenesis. Mantle T.J., Pickett C.B., Hayes J.D.,(eds). Taylor and Francis, Philadelphia., 1987 Prestera T., Holtzclaw W.D., Zhang Y., and Talalay P. Chemical and molecular regulation of enzymes that detoxify carcinogens. Proc Natl Acad Sci USA 90; 2965-9, 1993 Primanio T. and Novak R.F. Purification and characterization of class mu glutathione S-transferase isoenzymes from rabbit hepatic tissues. Arch Biochem & Biophys 301(2): 404-410, 1993 Prochaska H.J. and Fernandes C L . Elevation of serum phase two enzymes by anticarcinogenic enzyme inducers: markers for a chemoprotected state? Carcinogenesis 14(12): 2441-5, 1993 Randall B.G., Angus B., Akiba R., Hall A., Cattan A.R., Proctor S.J., Jones R.A., and Home C.H.W. Glutathione S-transferase (placental) as a marker of transformation in the human cervix uteri: An immunohistochemical study. BrJof Cancer 62:614-8, 1990 Raza H., Awatsthi Y . C , Zaim T., Eckert R.L, and Mukhtar H. Glutathione Stransferases in human and rodent skin: Multiple forms and species-specific expression. J Invest Derm 96(4):463-7, 1991 Reinemer P., Dirr H.W., Ladenstein R., Schaffer J., Gallay O., and Huber R. The three-dimensional structure of class pi glutathione S-transferase in complex with glutathione sulfonate at 2.3 A resolution. EMBO J 10(8): 1997-2005, 1991 Remmele W. and Schicketanz K.H. Immunohistochemical determination of estrogen and progesterone receptor content in human breast cancer. Computer-assisted image analysis (QCIC Score) vs. subjective grading (IRS). Path Res Pract 189: 862-866, 1993 Rice-Evans C and Burdon R. Free radical -lipid interactions and their pathological consequences. Prog in Lipid Res 32(1): 71-110, 1993 Riviere A., Becker J., and Loning T. Comparative investigations of c-erbB-2/neu expression in head and neck tumors and mammary cancer.Cancer 67(8): 21224-9, 1991 Robertson I.G.C, Guthenberg C , Mannervik B., and Jernstrom B. Differences in stereoselectivity and catalytic efficiency of three human glutathione transferses in the conjugation of glutathione with 7B,8a-dihydroxy-9a,10a-oxy-7,8,9,10tetrahydrobenzo(a)pyrene. Cancer Res46: 2220-4, May1986  Sager R. Tumor suppressor genes: the puzzle and the promise.Sc/'ence 2 2 8 : 596-7, 1985 Sahu SC Oncogenes, oncogenesis, and oxygen radicals. Biomed & Environ Sci 3(2): 183-201, 1990. Sakata K., Aoki Y., Karasawa K., Nakagawa K., Hasezawa K., Muta N., Terahara A., Onogi Y., Sasaki Y., Akanuma A., et al. Radiation therapy for patients of malignant salivary gland tumors with positive surgical margins. Strahlentherapie und Onkologie 170(6): 342-6, 1994 Sandros J , Stenman G, and Mark J. Cytogenetic and molecular observations in human and experimental salivary gland tumors. Cancer Genet Cytogenet 44: 153-167, 1990 Sato K. Glutathione transferases as markers of prenoplasia and neoplasia. Advin Cancer Res 52:205-255,1989 Satoh K., Kitahara A., and Sato K. Identification of heterogeneous and microheteroheneous subunits of glutathione S-transferase in rat liver cytosol. Arch Biochem Biophys 2 4 2 : 104-11, 1985 Satoh K., Hatayama I, Tsuchida S., and Sato K. Biochemical characteristics of a preneoplastic marker enzyme glutathione S-transferase P-form(7-7). Arch of Biochem and Biophys 285(2): 312-6, 1991 Schaffer J, Gallay 0., and Ladenstein R. Glutathione transferase from bovine placenta. Preparation, biochemical characterization, crystallization, and preliminary crystallographic analysis of a neutral class pi enzyme. J Biol Chem 2 6 3 : 17405-11, 1988 Schecter R.L and Aloui-Jamali M.A. Glutathione S-transferase in chemotherapy resistance and in carcinogenesis. Biochem Cell Biol 7 0 : 34953,1992 Scully C. and Burkhardt A. Tissue markers of potentially malignant human oral epithelial lesions. J Oral Pathol Med 22: 246-56,1993 Seifert G and Caselitz J. Epithelial salivary gland tumours: tumour markers. Prog Surg Pathol 9: 157-87, 1989. Seifert F and Sobin L H . Worid Health Organization's Histologic Classification of Salivary Gland Tumors. A commentary on the second edition. Cancer 7 0 ( 2 ) : 379-85, 1992 Semba K., Kamata N., Toyoshima K., and Yamamoto T. A v-erbB-related protooncogene, c-erB-2, is distinct from the c-erbB-1/epidermal growth factor  receptor gene and is amplified in human salivary gland adenocarcinoma. Proc Nat Acad Sci USA 82(19): 6497-501, 1985 Severini G. Glutathione S-transferase activity in patients with cancer of the digestive tract. J of Cancer Res and Clin Oncol 120(1): 112-4, 1993. Shea T.C. and Henner W.D. Glutathione transferases n in human tumors. In: Glutathione S-Transferases and Carcinogenesis, Mantle T.J., Pickett C.B., and Hayes J.D. (eds.) Taylor & Francis, London, p. 227, 1987 Sherman M.,Titmuss S and Kirsch R.E. Glutathione S-transferase in human organs. Biochem Int 6: 109-18, 1983 Shingaki S., Ohtake K., Nomura T., and Nakajima T. The role of radiotherapy in the management of salivary gland carcinomas. J Cran-Max-Fac Surg 20: 220224, 1992 Shiozawa M., Yasuda K., Yamashita S., and Aiso S. Distribution of gammaglutamyl transpeptidase in human salivary glands: immunohistochemical study using a monoclonal antibody. J Histochem & Cytochem 41(2): 205-13, 1993 Shiratori Y., Soma Y., Maruyama H., Sato S., Takano A., and Sato K. Immunohistochemical detection of the placental form of glutathione Stransferase in dysplastic and neoplastic human uterine cervix lesions. Cancer Res 47: 6806-6809, 1987 Shows T.B., McAlpine P.J., Boucheix C , Collins F.S., Conneally P.M., Frezal J., Gershowitz H., Goodfellow P.N., Hall J.G., Issitt P., Jones C.A., Knowles B.B., Lewis M., McKusick V.A., Meisler M., Morton N.E., Rubenstein P., Schanfield M.S., Schmickel R.D., Skolnick M.H., Spence M.A., Sutherland G.R., Traver M., Van cong N., and Willard H.F. Guidelines for human gene nomeclature. An international system for human gene nomenclature. Cytogenet Cell Genet 46: 11-8, 1987 Silverberg E and Lubera J. Cancer statistics 1987. Cancer J.Clin 37: 2-19, 1987 Sliverberg E.and Lubera J . Cancer statistics 1988. Cancer J Clin 38(1): 5-22, 1988 Singh S.V., Leal T., Ansari G.A.S., Awasthi Y.C. Purification and characterization of glutathione S-transferases of human kidney. Biochem J 246: 179-86, 1987 Singhal S.S., Saxena M., Awasthi S., Mukhtar H., Zaidi S.I.A., Ahmad H., and Awasthi Y.C. Glutathione S-transferases of human skin: Qualitative and quantitative differences im men and women. Biochim et Biophys Acta 1163: 266-72, 1993  Singhal S.S., Saxena M., Awasthi S., Ahmad H., Sharma R., and Awasthi Y.C. Gender related differences in the expression and characteristics of glutathione S-transferases of human colon. Biochim et Biophys Acta 1171: 19-26, 1992 Singhal S.S., Saxena M., Ahmad H., Awasthi S., Haque A.K., and Awasthi Y.C. Glutathione S-transferases of human lung: Characterization and evaluation of the protective role of the alpha-class isozymes against lipid peroxidation. Arch Biochem & Biophys 299(2): 232-41, 1992 Singhal S.S.,Gupta S., Ahmad H., Sharma R., and Awasthi Y.C. Characterization of a novel alpha-class anionic glutathione S-transferase isozyme from human liver. Arch Biochem & Biophys 279(1): 45-53, 1990 Solt D.B., Medline A. and Farber E. Rapid emergence of carcinogen-induced hyperplastic lesions: A new model for the sequential analysis of liver carcinogenesis. Am J of Pathol 8 8 : 595-610, 1977 Soma Y., Satoh K., and Sato K. Purification and subunit-structural and immunological characterization of five glutathione S-transferases in human liver, and the acidic form as a hepatic tumor marker. Biochimic et Biophysica Acta 869: 247-58, 1986 Spandidos D.A,. Lamothe A., and Field J.K. Multiple transcripiton activation of cellular oncogenes in human head and neck solid tumours. Anticancer Res 5: 221-6, 1985 Spiro R.H. Salivary neoplasms: overview of a 35 year experience with 2807 patients. Head and Neck Surg. 8: 177-84, 1986 Spiro R.H. and Spiro J. D. Cancer of the salivary glands. In Cancer of the Head and Neck. Myers E.N. and Suen J.Y. (eds). Churchill Livingstone: New York, 2nd edition, pp. 645-8, 1989 Spitz M.R., Sider J.G., and Newell G.R. Salivary gland cancer and risk of subsequent skin cancer.Head & Neck 12(3): 254-6, 1990 Spitz M.R., Fueger J.J., Goepfert H., and Newell G.R. Salivary gland cancer. A case control investigation of risk factors. Arch Otolaryngol Head & Neck Surg 116: 1163-6, 1990 Sreebny L.M. The Salivary System C R C Press Inc., 1987 Stenius U. Different inhibition of DNA synthesis by transforming growth factor B and phenobarbitol on GST-P-positive and GST-negative hepatocytes. Carcinogenesis 14(1): 159-61, 1993 Stehnman G., Sandros J., Nordkvist A., Mark J., and Sahlin P. Expression of th ERBB2 protein in benign and malignant salivary gland tumots. Genes, Chromosomes & Cancer 3: 128-135, 1991  110  Strange R.C., Fryer A.A., Matharoo B., Zhao L , Broome J., Campbell D.A., Jones P., Pator I.C., and Singh R.V.P. The human glutathione S-transferases: comparision of isoenzyme expression in normal and astrocytoma brain. Biochimica et Biophysica Acta 1139:222-88,1992 Strohmeyer T., Klone A., Wagner G., Hartmann M., and Sies H. Glutathione Stransferases in human testicular germ cell tumors: Changes of expression and activity. J Urology 147: 1424-8, 1992 Sudilovsky O,. Pitot H O , and Liotta L.A. Boundries Between Promotion and Progression During Carcinogenesis. Basic Life Sciences Vol 57, Plenum Press, New York, 1991 Sugano S., Mukai K., Tsuda H., et al. Immunohistochemical study of c-erbB-2 oncoprotein overexpression in human major salivary gland carcinoma: an indicator of aggressiveness. Laryngoscope 102: 923-7, 1992 Tahir M.K. and Mannervik B. Simple inhibition studies for distinction between homodimeric and heterodimeric isoenzymes of glutathione transferase. J Biol Chem 261(3): 1048-51, 1986 Takahashi M., Hokunan K., and Unno T. Immunohistochemical study of basement membrane in pleomorphic adenomas of the parotid gland: comparison between primary treated tumor and recurrent tumor. Nippon Jibinkoka Gakkai Kaiho (Journal of the Oto-Rhino-Laryngological society of Japan) 95(11): 1759-64, 1992 Tan K.H., Meyer D.J., Coles B., and Ketterer B. FEBS Letter 207: 231-3, 1986 Tatematsu M., Mera Y., Ito N., Satoh K., and Sato K. Relative merits of immunohistochemical demonstrations of placental, A, B and C forms of glutathione S-transferase and histochemical demonstration of gamma-glutamyl transferase as markers of altered foci during liver carcinogenesis in rats. Carcinogenesis 6(11): 1621-1626, 1985 Taylor S.G. Head and neck cancer. Cancer Chemotherapy and Biol Resp Mod 14: 427-441, 1993 Taylor S.L. and Scanlan R.A. Food toxicology. A perspective on the relative risks. Basic Symposium Series Institute of Food Technologists. Chicago, III. Marcel Dekker, Inc Newyork and Basel, 1989 pp. 104, 107, 108, 116, 127, 170, 178,179,189,355,358,360 Tee L.B.G., Gilmore K.S., Meyer D.J., Ketterer B., Vandenberghe Y., and Yeoh G.C.T. Expression of glutathione S-transferase during rat liver development. Biochem J 28: 209-18, 1992  Terrier P., Townsend A.J., Coindre J.M., Triche T.J., and Cowan K.H. An immunohistochemical study of pi class glutathione S-transferase expression in normal human tissues. Am J Pathology 137(4): 845-53, 1990 Thackray A.C., Lucas R.B., (eds.) Tumors of the Major Salivary Glands. Armed Forces Institute of Pathology. Washington, 1974, p. 16-39, p. 107-118. Tidefelt U., Elmhorn-Rosenborg A., Paul C , Hao X-Y.m Mannervik B., and Eriksson L. Expression of glutathione transferase pi as a predictor for treatment results at different stages of acute nonlymphoblastic leukemia. Cancer Res 52: 3281-5, 1992 Toffoli G., Viel A., Tumiotto L., Giannini F., Volpe R., Quaia M., and Boiocchi M. Expression of glutathione S-transferase-pi in human tumours. Eur J Cancer  28A(8/9): 1441-6, 1992  Tsuchida S., Sekine Y., Shineha R., Nishihira t., and Sato K. Elevation of the placental glutathione S-transferase form (GST-pi) in tumor tissues and the levels in sera of patients with cancer. Cancer Res 49: 5225-9, 1989 Tsuda H., Ozaki K., Uwagawa S., Yamaguchi S., Hakoi K., Aoki t., Kato T., Sato K., and Ito N. Effects of modifying agents on conformity of enzyme phenotype and proliferative potential in focal preneoplastic and neoplastic liver cell lesions in rats. Jpn J Cancer Res 83: 1154-65, 1992 Tsunemoto H., Morita S., Satoh S., lino Y., Yul Yoo. Present status of fast neutron therapy in Asian countries. Strahlenther Onkol 165: 330-6, 1989 Tu G-Y., Hu Y-H., Jiang P-J., and Qin D-X. The superiority of combined therapy (surgery and postoperative irradiation) in parotid cancer. Arch. Otolaryngol. 108: 710-3, 1982 Vainio H. Vinyl chloride and vinyl benzene (styrene) - metabolism, mutagenicity, and carcinogenicity. Chem-Biol Interact 22: 117-24, 1978 VandenBerg H.J.Jr., Kambouris A., Pryzybylski T., Rachmaninoff N. Salivary Tumors. Clinicopathologic review of 190 patients. Am J Surg 108: 480-84, 1964 Vander Jagt D.L, Hunsaker LA., Garcia K.B., and Royer R.E. Isolation and characterization of the multiple glutathione S-transferases from human liver. Evidence for unique heme-binding sites. J Biol Chem 260(21): 11603-10, 1985 van der Zee A.G.J., van Ommen B., Meijer C , Hollema H., van Bladeren P.J., and de Vries E.G.E. Glutathione S-transferase activity and isoenzyme composition in benign ovarian tumours, untreated malignant ovarian tumours, and malignant ovarian tumours after platinum/cyclophosphamide chemotherapy. Br J of Cancer 66: 930-6, 1992  Volm M., Mattern J., and Samsel B. Relationship of inherent resistance to doxorubicin, proliferative activity and expression of P-glycoprotein 170, and glutathione S-transferase-pi in human lung tumors. Cancer 70(4): 764-9,1992 Volm M., Mattern J., and Samsel B. Overexpression of P-glycoprotein and glutathione S-transferase-pi in resistant non-small cell lung carcinomas of smokers. BrJ of Cancer 64: 700-4, 1991 Vuillaume M. Reduced oxygen species, mutation, induction and cancer initiation. Mut Res 186: 43-72, 1987 Waalkes M.P. and Ward J.M. Carcinogenesis. Target Organ Toxicology Series. Raven Press, New York, 1994 Wambersie A., Richard F., and Breteau N. Developemennt of fast neutron therapy worldwide. Radiobiological, clinical and technical aspects. Acta Oncologica 33(3): 261-274, 1994 Warholm M et al Identification of a new glutathione S-transferase in human liver. Acta Chem Scand 34: 607-10, 1980 Waxman D.J. Glutathine S-transferases: Role in alkylating agent resistance and possible target for modulation chemotherapy - a review. Cancer Res. 50: 6449-6454, 1990 Weinberg R.A. Oncogenes, antioncogenes, and the molecular basis of multistage carcinogenesis. Cancer Res 49: 3713-21, 1989 Weiner T. and Cance W.G. Molecular mechanisms involved in tumorigenesis and their surgical implications. Am J of Surg 167: 428-34, 1994 Wenig B.M, Hitchcock C.L., Ellis G . L , and Gnepp D.R. Metastasizing mixed tumor of salivary glands. A clinicopathologic and flow cytometric analysis. Am J Surg Path 16(9): 845-58, 1992 WHO International histologic classification of tumours. Histologic typing of salivary gland tumours. G. Seifert, (ed) in collaboration with pathologists in 6 countries. 2nd edition. Springer-Verlag, 1991. Wilce M.C.J. & Parker M.W. Structure and function of glutathione Stransferases. Bioch et Biophys Acta 1205: 1-18, 1994 Xia C. Taylor J.B., Spencer S.R. and Ketterer B. The human glutathione Stransferase P1-1 Gene: Modulation of expression by retinoic acid and insulin. Biochem J 292:845-50, 1993 Yokoyama Y., Tsuchida S., Hatayama I., and Sato K. Lack of peroxisomal enzyme inducibility in rat hepatic preneoplastic lesions induced by mutagenic  carcinogens: Contrasted expression of glutathione S-transferase P form and enoyl CoA hydratase. Carcinogenesis 14(3): 393-8, 1993 Zhang L, Mock D., and Cameron R. Development of glutathione S-transferase placental form (GST-P) stained foci during hamster buccal pouch mucosa carcinogenesis. Cancer Letters 64:241-7,1992 Zhang L, Xiao Y and Priddy R Increase in placental glutathione S-transferase in human oral dysplastic and neoplastic lesions. J of Oral Pathol and Med. 23(2): 75-9, 1994 Zhang L The value of placental glutathione S-transferase and y-glutamyl transpeptidase as markers of altered foci during hamster pouch carcinogenesis. Carcinogenesis 15(1): 105-9, 1994 Zhong S., Howie A.F., Ketterer B., Taylor J., Hayes J.D., Beckett G.J., Wathen C.G.,Wolf R . C , and Spurr N.K. Glutathione S-transferase mu locus: Use of genotyping and phenotyping assays to assess association with lung cancer susceptibility. Carcinogenesis 12(9): 1533-7, 1991  114  APPENDIX 1  115  Figure 1. Schematic diagram of cellular protection against oxygenmediated toxicity: UQH-semiquinone of ubiquinone. D,DH, AH2, BH2 various cellular substrates involved in enzymatic redox reactions. (From Taylor & Scanlan, 1989 p 174)  1  116  Oxidative stress  r  1  DNA damage  Rlsos In  Increased  intracellular  lipid  tree Iron  peroxidation  Poly A D P ribose synthetase activation  NAD (H) depletion  GSH  3  depletion  Inhibition of  Direct damage  A T P synthesis  to proteins  r  GSH depletion  r Rises In  Cyloskeletal  Intracellular  damage  free C a -  Membrane  Membrane  Increased  blebbing  peroxidation  damage to DNA,  and destruction  proteins, lipids  Metal Ion release Into surrounding tissues. Injury to adjacent cells  Figure 2. Schematic illustration of multiple derangements of cell metabolism that can be brought about by oxidative stress. (From Halliwell, 1991)  \  \  OXIDATION-REDUCTION PATHWAYS GSSQ  Figure 3. Overall pathway of glutathione (GSH) metabolism: 1gamma glutamylcysteine synthetase, 2-glutathione synthetase, 3-gammaglutamyl transpeptidase, 4-cysteinyl glycine hydrolases, 5-gamma-glutamyl cyclotransferase, 6-5-oxoprolinase, 7-glutathione S-transferase, 8-transport and reduction of gamma-Glu-0(Cys)2, 9-oxidation reduction pathways involving cleavage of ATP to ADP and Pj. (From Meister, 1994)  Figure 4. X-ray crystallographic analysis of pig glutathione Stransferase pi/glutathione sulfonate co-crystal - Stereo drawing of dimeric glutathione S-tranferase molecule (thin line) with the inhibitor glutathione sulfonate included (thick line). etal., 1991)  (From Reinemer  119  Figure 5. Ribbon representation of GST-pi class dimers, two orthogonal views. (From Wilce & Parker, 1994)  120  121  122  123  124  Table 1 Species-independant classification of rat, human, and mouse cytosolic glutathione transferase subunits (From Sato, 1989)  SPECIES-INDEPENDENT CLASSIFICATION OF RAT, HUMAN, AND MOUSE CYTOSOLIC GLUTATHIONE TRANSFERASE SUBUNITS"  Class  Rat  alpha  1 (Ya) (25) 2 (Yc) (28) 8 (Yk or Ya') (24.5) 3 (Ybi) (26.5) 4 (Yb ) (26.5) 6 (Yn, Yn,, or Yb ) (26) ? (Yn ) (26) 7 (Yp or Yf) (24)  mu  Human  fo  2  (YO, B (Y ) (26) Subunit 2 (28.5) c  Bi  2  — — Y (Y ) (27) e  M  3  — —  3  2  Pi  Mouse  Y„..(Y ) (24.5) 5  4  m, (25.5)  m  — —  (27) (GT-9.3, GT-8.7)' (/  3  m (24) 2  Table 2 Nomenclature and classification of glutathione S-transferases  CLASS  NOMENCLATURE  LOCUS  CHROMOSOME  DESIGNATION alpha  GSTA1-1  GSTA1  6  alpha  GSTA2-2  GSTA2  6  mu  GSTM1a-1a  GSTM1  1  mu  GSTM1b-1b  GSTM1  1  mu  GSTM2-2  GSTM2  _  mu  GSTM3-3  GSTM3  pi  GSTP1-1  GSTP1  11  GST12  12  theta microsomal  Table 3 Immunohistochemical detection of GST-re in normal tissues (From Terrier et al., 1990) Tissue Urinary System Kidney Ureter, urinary bladder Digestive Tract Esophagus Stomach  Main immunoreactive structures Henle loop, distal convoluted tubule, collecting ducts, Epithelium, muscularis, Epithelium, submucosal glands, Surface epithelium and fundic glands, Epithelium, crypts of Lieberkuhn,  Duodenum, jejunum, colon Digestive Glands Liver Biliary ducts, Gall bladder Epithelium, muscularis, Pancreas Ducts, Respiratory System Trachea, bronchi Epithelium, glands, cartilage, Lung Alveolar epithelium, Reproductive system Uterus Cervical epithelium, Endometrial epithelium and glands, myometrium, Tube epithelium, smooth muscle, Ovary Follicles, cortex cells, Breast Acinar and ductal epithelium, Placenta Cytotrophoblast, Testis Seminiferous tubules, epididymis, smooth muscle, Prostate Glandular epithelium, smooth muscle, Endocrine System Thyroid Follicle cells, Adrenal Cortex, medulla, Nervous System Brain Neurones, glial cells, Peripheral nerves Schwann cells, Vascular System Heart Myocardium, Blood vessels Endothelium, smooth muscle, Other Skin Epidermis, sebaceous and sweat glands, Lymph nodes Germinal centers.  Table 4 Histologic Classification of Salivary Gland Tumours Adenomas (Benign Salivary Gland Tumours) Pleomorphic adenoma Myoepithelioma(Myoepithe|ial adenoma) Basal cell adenoma Warthin tumour (Adenolymphoma) Oncocytoma (Oncocytic adenoma) Canalicular adenoma Sebaceous adenoma Ductal papilloma Inverted ductal papilloma Intraductal papilloma Sialadenoma papilliferum Cystadenoma Papillary cystadenoma Mucinous cystadenoma Carcinomas (Malignant Salivary Gland Tumours) Acinic cell carcinoma Mucoepidermoid carcinoma Adenoid cystic carcinoma Polymorphous low grade adenocarcinoma (Terminal duct adenocarcinoma) Epithelial-myoepithelial carcinoma Basal cell adenocarcinoma Sebaceous carcinoma Papillary cystadenocarcinoma Mucinous adenocarcinoma Oncocytic carcinoma Salivary duct carcinoma Adenocarcinoma Malignant myoepithelioma (Myoepithelial carcinoma) Carcinoma in pleomorphic adenoma (Malignant mixed tumour) Squamous cell carcinoma Small cell carcinoma Undifferentiated carcinoma Other carcinomas Nonepithelial tumours Malignant lymphomas Secondary tumours Unclassified tumours Tumour-like lesions Sialadenosis Oncocytosis Necrotizing sialometaplasia (Salivary gland infarction) Benign lymphoepithelial lesion Salivary gland cysts Chronic sclerosing sialadenitis of submandibular gland (Kuttner tumour) Cystic lymphoid hyperplasia in AIDS  129  Table 5 GST-TC staining intensity in human salivary gland tumours Diagnosis  Total # of cases 19  Staining intensity (%) -/±  + 10 (53)  ++ 3(16)  Monomorphic 5(26) adenomas 23 Pleomorphic 2(9) 14 (61) 4(17) adenomas 7 4 (57) Acinic cell 2 (29) 1 (14) carcinomas Adenoid cystic 11 4(36) 5 (46) 2(18) carcinomas Mucoepidermoid 11 0 0 2(18) carcinomas *A mucinous cystadenoma All are the cell-poor type of pleomorphic adenomas  1 (5)*  Average staining intensity +  3 (13)**  +  0  •±  0  +  9 (82)  +++  +++  Table 6 Comparison of GST-TC staining in cell-rich and cell-poor pleomorphic adenomas  GST-pi staining intensil v (%) Histological Total # -or± + ++ +++ diagnosis of cases Cell-rich 9 0 7 (78) 1 OD 1 (1D Cell-poor 3 (30) 10 3(30) 3 (30) 1 (10) * Of the 23 pleomorphic adenomas studied, 4 were classified as intermediate cellularity  I  Table 7 T N M Classification of anatomical extent of disease of salivary gland tumours Tx  primary tumour cannot be assessed  To  no evidence of primary tumour  T1  tumour < or = 2cm  T2  tumour >2cm , but no greater than 4 cm  T3  tumour >4cm, but no greater than 6 cm  T4  tumour >6cm (All categories divided as a.-no local extension and b.- with local extension)  NX  regional lymph nodes can not be assessed  NO  no regional metastasis  N1  single ipsilateral node < or = 3cm  N2  single ipsilateral node >3cm, no greater than 6cm multiple ipslilateral nodes < or = 6cm bilateral or contralateral nodes < or = 6cm N2a  metastasis single ipsilateral node 3cm, no greater than 6cm  N2b  metastasis multiple ipsilateral nodes, none greater than 6cm  N2c N3  >6cm  metastasis in bilateral or contralateral nodes, none >6cm  131  APPENDIX 2  A.  INFLAMED SALIVARY GLAND  Portions of this lobule of salivary gland show inflammatory changes including atrophy of acinar units, proliferation of ducts, and periductal inflammation. The GST-TC staining in these regions is similar to that found in normal gland: ducts stain positively with larger excretory ducts tending to have stronger staining and intercalated ducts tending to stain weaker. Acini are negative. Orig. Mag.X24  B.  NORMAL SALIVARY GLAND  staining of normal salivary gland showing acinar cells staining negatively, with ducts staining positively. There is slight variation in intensity of ductal staining. Orig. Mag.X62.5  GST-TC  C.  MONOMORPHIC ADENOMA  The GST-TC staining in this monomorphic adenoma is uniformly weak. Orig. Mag.X62.5  D.  MONOMORPHIC ADENOMA - WARTHIN'S TUMOUR  The majority of this tumour stains weakly with GST-TC with some cells of ductal differentiation staining slightly stronger. Orig. Mag.X24  E.  PLEOMORPHIC ADENOMA - CELL-RICH TYPE  This solid area of pleomorphic adenoma shows polygonal and stellate epithelial cells, with no ductal differentiation The bulk of this tumour stains weakly with GST-TC. Orig. Mag.X62.5  F.  PLEOMORPHIC ADENOMA - CELL-RICH TYPE  This richly cellular tumour consists of solid regions of epithelial cells interspersed with hyalin or myxoid acellular areas. Some ductal differentiation is evident. The tumour stains weakly with GST-TC. Orig. Mag.X24  G.  PLEOMORPHIC ADENOMA - CELL-POOR TYPE  The majority of this portion of tumour consists of loose myxoid stroma. Islands of epithelial cells, some with strong ductal differentiation are present and stain strongly with GST-TC. Orig. Mag.X24  H.  PLEOMORPHIC ADENOMA - CELL-POOR TYPE  High power of the epithelial formation within tumour seen in G . showing strong GST-TC staining of epithelial cells, including prominant nuclear staining. Orig. Mag.X163  Legend to Plate 1.  133  A. ACINIC C E L L CARCINOMA (HEMOTOXYLIN & EOSIN STAIN) This well differentiated acinic cell tumour consists of cells with abundant granular cytoplasm arranged in acinar groups. Between these cells are lightly staining cells resembling those of intercalated duct. Orig. Mag.X62.5 B.  ACINIC C E L L CARCINOMA GST-TC staining of the tumour seen in A . shows equivocal to weak staining, with acinar cells being negative and the remaining epithelial cells faintly positive. Orig. Mag.X62.5 C. MUCOEPIDERMOID CARCINOMA This tumour consists of epithelial tumour islands composed of mucus-secreting cells, epidermoid cells and intermediate cells. GST-TC staining intensity is generally strong. Orig. Mag.X24 D. MUCOEPIDERMOID CARCINOMA This higher power view of the tumour island in C . shows strong GST-TC staining with areas of nuclear staining evident. Orig. Mag.X62.5 E. MUCOEPIDERMOID CARCINOMA This high power view of portions of tumour seen in C . and D. show mucus tumour cells staining negatively with GST-TC and the remainder of tumour cells staining strongly, with nuclear staining evident. Orig. Mag.X163 F. ADENOID CYSTIC CARCINOMA This tumour is composed predominantly of myoepithelial cells and contains, in some areas, ductal cells. GST-TC staining is generally equivocal or weak, interspersed with stronger staining of some cells. Orig. Mag.X62.5 G. ADENOID CYSTIC CARCINOMA This higher power view of the tumour seen in F. shows weak staining of the majority of cells with small numbers of cells staining stronger. Orig. Mag.X163  Legend to Plate 2.  135  PLATE 2  A. NON-RECURRENT PLEOMORPHIC ADENOMA This is a cellular tumour showing very weak GST-TC staining, except in the areas of ductal differentiation which stain moderately. Orig. Mag.X62.5 B. NON-RECURRENT PELOMORPHIC ADENOMA In some non-recurrent tumours, cell-poor areas are present. Epithelial cells in these areas, on occasion, stained stronger. Orig. Mag.X62.5 C. RECURRENT PLEOMORPHIC ADENOMA This low power view shows portions of recurrent tumour along with adjacent normal appearing salivary gland. Orig. Mag.X24 D. RECURRENT PLEOMORPHIC ADENOMA Epithelial tumour islands as well as individual cells, within the myxoid stroma, stain strongly with GST-TC. Cells at the periphery of the tumour, adjacent to capsule stain more intensely. Orig. Mag.X62.5 E. RECURRENT PLEOMORPHIC ADENOMA Individual tumour islands and cells from peripheral regions of tumour seen in D. show strong GST-TC staining, with nuclear staining. Orig Mag.X163 F. RECURRENT PLEOMORPHIC ADENOMA Stellate epithelial or myoepithelial cells in central regions of tumour seen in D. show strong GST-TC staining, with nuclear staining. Orig. Mag.X163  Legend to Plate 3.  137  138  APPENDIX 3.  139  GST-P IMMUNOHISTOCHEMICAL STAINING (ABC METHOD) General Points The buffer used is a 10mM sodium phosphate pH 7.5, 0.9% saline (PBS): 25.5g NaCI 3.21 g Na2HP04 1.17g NaH2P04. 2H20 3 litres distilled water prepared as a x10 stock solution Tris buffer 0.1 M, pH 7.2 MWofTris: 120.14 To make 10 ml of Tris, dissolve g of Tris in distilled water and adjust pH to 7.2. ABC kit should be kept in the fridge To prevent sections from detaching from the glass, slides can be coated with gelatin or acid cleaned with 10ml concentrated HCI in 1 litre of 70% ethanol. Do not use egg albumin coated slides. Traces of egg white avidin may affect staining. Hand lotions can cause sections to detach from slides or may prevent adequate penetration of reagents. Avoid touching rinse baths with oily hands. One drop is approximately 50ul. To avoid adsorption of the antobody to plastic or glass container in which the final dilution is made, all dilutions of primary antibody, link antibody, or ABC reagents should be made in a buffer containing 0.1% crystalline grade bovine serum albumin or dilute block serum (2 drops concentrated stock in 10ml of buffer) included in the ABC kit. Only crystalline grade serum should be used, as other preparations can contain undesired impurities (see Vector instructions). When dropping reagents, especially antibody, onto sections, avoid air bubbles. The mixing bottles in the ABC kit should be washed 20 times with tap water, 20 times with distilled water, then dried after each use. Preparation of stock DAB at 50 mg/ml, ie. 5% (10x the final concentration of 5 mg/ml, ie. 0.5%) To make 20 mis of 5% DAB (50mg/ml): mix 1g of 3,3'Diaminobenzidine tetrahydro-chloride (chromogen) in 20 mis of 0.1 M Tris buffer, pH 7.2, and make suitable aliquots stored frozen at -20°C for extended periods of time.  PART 1 1. Draw circles around sections to be stained with a diamond pen (not too big, otherwise they will consume too much reagent). This prevents the spread of liquid from the sections. Label (+) and (-) controls respectively. 2. Incubate in the dry oven (37°C) for 30 mins to allow for condensation (from being in the fridge) to evaporate, ensuring no moisture exists. 3. Deparaffinize in xylene (2x 1hr, room temp, or overnight) at room temp in hood. 4. Hydrate in graded alcohol: absolute alocohol (2x3min), 95% (3min), 70% (3min). 5. Place in distilled,water for 5 min, dipping constantly. Prepare 0.3% hydrogen peroxide (methanol peroxide) (1ml of 30% hydrogen peroxide/100ml methanol): For 1 slide tray or 1 staining jar, about 200 ml is needed. Hence, add 2 ml of 30% hydrogen peroxide in 198 ml of methanol. 6. Incubate in 0.3% hydrogen peroxide in methanol (w/v, or 1% v.v) for 30 min at room temperature in a staining jar to block endogenous peroxidase activity in hood. 7. Wash in cold, running tap water for 20 min. 8. Place in distilled water for 10 min. Shake frequently. Wet moist chamber with paper towel which should not touch spacer bars. If moist chamber unavailable, petri dishes can be used. Prepare diluted normal serum: Add two drops of stock goat serum (small yellow-labelled bottle) to 6.66 ml of buffer in mixing bottle (yellow label). This is enough to do approximately 24 slides with up to three sections per slide. Usually 3 drops of goat serum is added to 10ml of PBS. 9. Drain and wipe slides except within the circled area (one slide at a time). Add diluted normal serum to each section so as to form a dome of liquid over it (approx 2 drops, or 0.1-0.2ml) to block non-specific binding. Incubate in a moist chamber, room temperature, 1/2 hour. Note: It is very important to wipe all areas of the slides dry, except for the sections, before adding subsequent reagents. This helps prevent the spread of reagents across the slides. Never let sections dry out!!  To avoid slides drying out, wipe one slide at a time and immediately add reagent. Normal serum is serum from the species in which the secondary antiserum is made. Prepare primary antibody: 0.015 ml to 4.5 ml PBS in a blue cap tube (1:300), and completely wash the antibody vial with same PBS win which the antibody has been added (the pipette volume should be >0.0015 ml). 10. Drain off excess serum and wipe around (do not wash). For positive staining slides, add primary antibody (anti-rat GST-P) to each section (about 2 drops). Incubate in a moist chamber overnight in a fridge, (approx 17 hrs, ie. 4p.m. to 9 a.m.) Add PBS on (-) labelled liver control or other negative controls. Note: 15uJ of GST-P (batch supplied)is enough to do approximately 40 slides with 1-3 sections per slide. The anti-rat GST-P antiserum was raised in rabbit. The concentration supplied was 1:300. The optimal concentration of primary antibody was determined by titration on positive control sections. Anytime this protocol is used, a positive and negative control must be used. Positive control: liver. Negative control: ommision of the primary antibody on liver sections. PART  2  1. Wash with PBS; 3x5min at room temp. Note: control slides washed separately to avoid (-) labelled control being contaminated. Prepare moist chamberPrepare secondary antibody: Add 25uJ of secondary antibody (small blue labelled bottle) to 5ml of buffer in mixing bottle (blue label).  2. Dry around sections, one slide at a time, and add secondary antibody (Biotin-labelled), about 2 drops per section. Incubate in a moist chamber at room temp for 1 hour (do a T hour timing and a 50 min timing - see below). Ten minutes before the incubation is finished, prepare ABC reagent: add exactly 1 drop of reagent A (small orange-labelled bottle) to 5 ml of buffer in the ABC reagent mixing bottle (brown label) and mix immediately. Then add exactly 1 drop of reagent B (small brown-labelled bottle) to the same ABC reagent mixing bottle and mix immediately. This mixture is to stand for approximately 30 minutes before use. 3. Drop IgG from slides onto paper towel, transfer slides in slide holder and wash 3x5min in PBS. 4. Dry one slide at a time and add the ABC reagent. Incubate in a moist chamber for 30 min at room temp. 5. Wash 3x5min in PBS, dipping constantly. Prepare the two components of the substrate solution in two white cap tubes: A - 6ml of 0.02% H2O2: add 0.004ml of 30% H2O2 to 6ml of dH20. B - 2.5 ml of 0.1M Tris buffer (pH 7.2): add 0.25 ml of 10x stock (1M) to 2.25 ml of dH20. Note: For ease of handling, DAB may be made up at 5mg/ml or higher in Tris buffer and suitable aliquiots stored at -20oC for extended periods of time. 6. After drying 7 slides, incubate with substrate solution for a maximum of 7 min. Blot off excess substrate and stop enzymatic reaction by washing once in PBS (a dip in PBS is sufficient). Note: do 7 slides at a time as the duration of this incubaiton is critical! 7. Place under running water for a minute or two. 8. Counterstain with hematoxylin for 15-40 seconds (such time as to obtain light staining). Note: for computer image analysis, omit counterstaining and mounting steps. 9. Wash in cold running tap water until hematoxylin colour disappears from the water. 10. Let air dry in hood (may leave overnight). 11. Once dry, mount by dipping in xylene and using a permanent mount (entellan).  STATISTICAL  ANALYSES  GST-TC  Raw  IN  SALIVARY  GLAND  TUMOURS  Data  Title Column ID Column Title Raw or Mean 1 2 3 4 5 • 6 7 8 9 10 11 12 13 14 1 5 16 17 18 19 20 21 22 23  A  B PA Raw Data  MA Raw Data 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 2 2 2 3  C Acinic Raw Data 0 0 1 1 1 1 1. 1 1 1 1 1 1 1 1 1 2 2 2 2 3 3 3  D Adenoid Raw Data 0 0 0 0 1 1 2  E ME Raw Data 0 0 0 0 1 1 1 1 1 2 2  2 2 3 3 3 3 3 3 3 3 3  GST-re IN SALIVARY GLAND TUMOURS One Way Analysis of Variance  Source of variation  Degrees of freedom  Treatments (between columns) Residuals (within columns) S  =  S S S S  =  =  =  =  S  =  =  =  S:=E  =  =  =  S=  4 66 —  Total  —  —  Sum of squares  Mean square  33.345 38.204  8.3362 0.5789  —  70  71.549  F = 14.401 The P value is < 0.0001, considered extremely significant. Variation among column means is significantly greater than expected by chance.  Bartlett's test for homogeneity of variances. ANOVA assumes that all columns come from populations with equal SDs. The following calculations test that assumption. Bartlett statistic (corrected) = 5.8178 The P value is 0.2132. This test suggests that the difference among the SDs is not significant.  IN SALIVARY G L A N D TUMOURS Student-Newman-Keuls Multiple Comparison Test GST-TC  Mean  Difference  Comparison  -0.2468 -0.4286 -0.7764 -2.2468 -0.1818 -0.5296 -2.0000 -0.3478 -1.8182 -1.4704  Acinic vs Adenoid Acinic vs MA Acinic vs PA . Acinic vs ME Adenoid vs MA Adenoid vs PA Adenoid vs ME MAvsPA MA vs ME PAvsME  q  P value  0.9486 1.8017 3.3432 8.6376 0.8920 2.6856 8.7185 2.0855 8.9203 7.4554  ns P>0.05 ns P>0.05 ns P>0.05 * " P<0.001 ns P>0.05 ns P>0.05 * " P<0.001 ns P>0.05 " * P<0.001 •**' P<0.001 :  Lower Upper Mean Difference 95% CI 95% CI  Difference  -0.2468 -0.4286 -0.7764 -2.2468 -0.1818 -0.5296 -2.0000 -0.3478 -1.8182 -1.4704  Acinic-Adenoid Acinic - MA Acinic - PA Acinic - ME Adenoid - MA Adenoid - PA Adenoid - ME MA - PA MA - ME PA-ME  -0.9819 0.4883 -1.2361 0.3789 -1.6430 0.0902 -3.2797 -1.2138 -0.7578 0.3942 -1.1991 0.1399 -2.8561 -1.1439 -0.8192 0.1235 -2.5101 -1.1263 -2.0277 -0.9130  Summary of Data  Group .MA PA Acinic Adenoid ME  Group MA PA Acinic Adenoid ME  Number of Points . Mean 19 23 7 11 11  1.0000 1.3478 0.5714 0.8182 2.8182  Standard Deviation 0.8165 0.8317 0.7868 0.7508 0.4045  Standard Error of Mean 0.1873 0.1734 0.2974 0.2264 0.1220  Lower 95%Upper 95% Confidence Confidence MinimumMaximum Interval Interval 0.0000 0.0000 0.0000 0.0000 2.0000  3.0000 3.0000 2.0000 2.0000 3.0000  0.6064 0.9882 -0.1563 0.3138 2.5464  1.3936 1.7075 1.2991 1.3225 . 3.0899  Median 1.0000 1.0000 0.0000 1.0000 3.0000  /  S E X DIFFERENCES Raw Data  IN  GST-TC  Title Column ID Column Title Raw or Mean 1 2 3 4 5 • • • 8 '  7  8 9 10 1 1 12 13 14 15 1 6 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 ' 37 38 39 40 41 42 43 44 45  A  . B  MALE TUMOR EMAL TUMOR Raw Data Raw Data 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 2 1 2 2 1 1 2 2 1 3 1 3 2 2 2 2 2 2 2 • 2 2 2 2 2 3 3 3 3 3 3 :  SEX  D I F F E R E N C E S IN  GST-TC  Unpaired t-test  Unpaired t test Are the means of MALE TUMOR and FEMAL TUMOR equal? Mean difference = 0.2444 (Mean of FEMAL TUMOR minus, mean of MALE TUMOR) The 95% confidence interval of the difference: -0.2332 to 0.7220 t = 1.0205 with 71 degrees of freedom. The two-tailed P value is 0.3109, considered not significant. Test: Are the standard deviations equal? The t test assumes that the columns come from populations with equal SDs. The following calculations test that assumption. F = 1.1841 The P value is o:3249. This test suggests that the difference between the two SDs is hot significant. Summary of Data  Parameter: Mean: # of points: Std deviation: Std error: Minimum: Maximum: Median: Lower 95% CI: Upper 95% CI:  MALE TUMOR 1.0000 28 0.9428 0.1782 0.0000 3.0000 1.0000 0.6344 1.3656  FEMAL TUMOR 1.2444 45 1.0259 0.1529 0.0000 3.0000 1.0000 0.9360 1.5529  149 GST-TC  IN  CELL-RICH  VS.  CELL-POOR  PLEOMORPHIC  ADENOMAS  Mann-Whitney Test Are the medians of Cell-rich PA and Cell-poor PA equal? Mann-Whitney U-statistic = 23.500 IT = 66.500 Sum of ranks in Cell-rich PA = 68.500. Sum of ranks in, Cell-poor PA = 121.50. The two-tailed P value is 0.0789, considered not quite significant.  Summary of Data Parameter:  Mean:  # of points:  Std deviation: Std error: Minimum: Maximum: Median: Lower 95% CI: Upper 95% CI:  Cell-rich  PA  Cell-poor PA  9  10  1.0000  0.5000 0.1667 0.0000 2.0000 1.0000 0.6157 1.3843  * *  1.8000  2  1.0328 0.3266 0.0000 3.0000 2.0000 1.0612 -2.5388  - 1.5-0 IMAGE ANALYSIS - PRIMARY VS. RECURRENT PLEOMORPHIC ADENOMAS Column Column Column Column  1 2 3 4  -  mean optical density coefficient of variance optical density skewness optical density kurtosis  Primary Pleomorphic Adenoma 0. 228752 2*211862 -1. 354751 236068 p . 266149 2 .141454 0. 288194 2 .100970 -1. 157435 0: 307806 2 .158914 -1. 247325 0. 178352 2 .146496 -1. 273045 0 .249880 2 .119216 -1. 192077 0. 225746 2 .123096 -1. 208804 0. 250320 2 .107129 -1. 174880 0. 179153 2 .100417 -1. 173911 0. 296626 2 .164398 -1. 262625 0. 158097 2 .129569 -1. 216088 0.359848 2 .117234 -1. 178457 0. 214148 2 .118221 -1. 187536  2 .171126 1 .745883 1 .468434 1 .758410 1 .940092 1 .581858 1 .653563 1 . 539673 1 .554922 1 .812439 i .688909 l .522858 I .578117  Recurrent Pleomorphic Adenoma 0.440198 0. 478440 0. 243047 0.319771 0.450077 0.515297 0.369847 0.475250 0. 490157 0.298607 0. 260764 0.291717  2.037697 2 .035418 2 .069747 2 .098009 2 .063360 2 .057646 2 .041527 2 .060818 2 .062500 2 .067929 2 .059050 2 .110128  -1 .067636 -1 .069686 -1 .13983 3 -1 .165820 -1 .095549 -1 .087866 -1 .082128 -1 .095022 -1 .094996 -1 .134925 -1 .102237 -1 .188176  1. 188756 1. 191222 1. 455423 1. 529054 1. 265923 1. 241634 1. 237780 1. 263698 1. 261047 1. 401144 1. 310223 1. 573203  |Column ID jColumn Title [MEAN OD PRI IMEANODREC | Raw or Mean Raw Data Raw Data I ! 0.228752 ! 0.440198 I 0.266149 0.478440 0.288194 0.243047 0.307806 0.319771 0.178352 0.450077! 0.249880J 0.515297J 0.225746 0.369847 0.250320 0.475250 0.179153! 0.490157 0.296626] 0.298607! 0.158097] 0.260764] 0.359848 291717 0.214148| Q  O  m  <  CM  t  1^ 93  o  eg CO  [Title 1 Column ID| SKEW PRI | SKEWREC j KURT PRI I KURT REC J | CVREC Column Title MEANODPRI |MEAN OD REC | CV PRI Raw Data Raw Data j Raw Data j Raw Data Raw Data Raw Data { Raw Data Raw Data Raw or Mean 1 .2 2| -1.03J 1.691 1.33J 2 . 1 3 ) 2 . 4 3 1 0e + 04 Mean 0.246390} 2.0 6| 0.057776 0 . 0 3 1 1 5 3 0.022203 0 . 0 5 4 4 4 0 0.283421J 0.1 95804 0.1 30938 Sample size 0.016024 0.008640 0.00641 0.015099 0.081817 0.054306 0.037799 1.5 8 1.24 -1.25 -1 .21| 2.05 2.11 95% CI min 0.211474 2 . 0 8 1 . 1 9 1 . 8 1 1 .41 2.1.5 95% CI max 0 . 2 8 1 307 7.78168 + 04 -0.84691} 1 . 3 5 1 . 1 9 1 . 4 7 ) 1 . 19 2 . 0 4 2 . 1 0 Minimum 0.1 58097 0.243047I 1 . 1 6 2.17] 1 . 57 2 . 2 1 Maximum 0.359848 -0.134925|  Raw Data 2.211862 | 2.141454 2.100970 2.158914 2.146496 2.119216! 2.123096! 2.107129 2.100417] 2.164398] 2.129569] 2.117234 2.118221]  UJ  SKEWREC | KURT PRI | KURTREC | Raw Data | Raw Data | Raw Data -1.067636 I 2.171126 I 1.188756| -1.069686 | 1.745883 1.191222| -1.139833 1.455423 1.468434 -1.165820 1.758410 1.529054 -1.095549 1.940092! 1.265923 -1.087866 1.581858! 1.241634 -1.082128 1.6535631 1.237780 1.263698 -1.095022! 1.539673! -1.094996 1.554922 1.261047 1.401144| 1.812439| -0.134925J 1.688909] 1.3102231 -1.1022371 1.522858 -1.188176 1.573203J 1.578117]  ul  SKEW PRI | Raw Data | -1.354741 | -1.236068 ! -1.157435 -1.247325 -1.273045 -1.192077! -1.208804 -1.174880 -1.173911 -1.262625] -1.216088] -1.178455 -1.187536]  CD  CVREC | Raw Data 2.037697 | 2.035418 2.069747 2.098009 2.063360 2.057646! 2.041527 2.060818 2.062500 2.067929| 2.059050] 2.110128  X  CVPRI  |  IMAGE ,ANALYSIS •• PRIMARY VS. RECURRENT PLEOMORPHIC ADENOMAS  151  OJ  T—  X  CO  CD  CM  u.  co  T-  LU  CM  T—  Q  m  CM  CO  O  -TT-  o •+'  m  5) CM  co  T-  CM  TT  O  <9 <D O  +  V—  CO  .+  o  O CM  «r o>  CM  CO  <  V) •w  o +  eg CM  at  152 IMAGE ANALYSIS - PRIMARY VS. RECURRENT PLEOMORPHIC ADENOMAS Mann-Whitney Test  Mann-Whitnev Test Are the medians of MEAN OD PRI and MEAN OD REC equal? Mann-Whitney U-statistic = 15.000 IT = 141.00 Sum of ranks in MEAN OD PRI = 106.00. Sum of ranks in MEAN OD REC = 219.00. The two-tailed P value is 0.0007, considered extremely significant. (The P value is an estimate based on a normal approximation.) Summary of Data Parameter: Mean: # of points: Std deviation: Std error: Minimum: Maximum: Median: Lower 95% CI: Upper 95% CI:  MEAN OD PRI 0.2464 1.3 0.0578 0.0160 0.1581 0.3598 0.2499 0.2115 0.2813  MEAN OD REC 24310 12 84211 24310 0.2430 291717 0.4451 -29196 77816  * * * * *  IMAGE ANALYSIS - PRIMARY VS. RECURRENT PLEOMORPHIC ADENOMAS Data - Individual Cases Primary Pleomorphic Adenoma Column 1 - stage x Column 2 - stage y Column 3 - area Column 4 - lntegrated(total) optical density Column 5 - mean optical density Column 6 - coefficient of variance of optical density Column 7 - optical density skewness Column 8 - optical density kurtosis mz211734 14242 13354 14396 15523 14004 16354 13405 13652 14654 14813 16255 14573  19518 34888 31194 19926 20362 ' 32508 20662 18357 20306 38397 30028 36899 79692 32637 29618 78383 32420 37271 32200 49130 33713 98617 132495 33524 55652.58  15911 15607.95 15243.77 8929.155 18796.72 17199.35 40546.21 43228.52 20831.11 22997.49 4960929 60569.43  0.456059 0.500351 0.468924 0.486417 0.489536 0.46612 0.508786 0.551504 0.558909 0.468095 0.50305 0.457145  2.073396 2.051795 2.070272 2.048862 2.073181 2.062172 2.041218 2.051877 2.063062 2.070282 2.062193 2.072093  -1.105758 -1.088214 -1.100178 -1.086365 -1.10436 -1.092954 -1.057631 -1.096745 -1.088446 -1.104967 -1.091125 -1.109101  1.291435 1.240437 1.273802 1.234985 1.285571 1.248121 1.153113 1.266969 1.237954 1.293945 1.251567 1.307857  -1.075756 -1.068172 -1.062288 -1.089115 -1.087169 -1.060489 -1.076435 -1.065351 -1.064145 -1.058811  1.208482 1.187337 1.169471 1.251002 1.245376 1.160636 1208565 1.176506 1.174451 1.160807  -1.069394 -1.07632 -1.064305 -1.098023 -1.100459 -1.107938 -1.092007 -1.093381 -1.080979 -1.091858  1.183842 1.203699 1.170143 1.274609 1.276342 1.303024 1.255924 1.255855 1.221604 1.256465  mz214963 13788 15315 16211 10966 10915 13020 14222 15204 13556 12712  35829 36778 .34505 34325 32113 32197 34951 34744 32147 36037  76020 80492 129406 15739 26554 61980 13857 47805 22393 38002 51224.8  33632.59 38139.38 68722.37 7248.07 10828.17 32033.36 6616.404 24192.15 11443.01 17782.31  0.442418 0.473828 0.53106 0.460517 0.407779 0.516834 0.477477 0.506059 0.511008 0.467931  2.051132 2.045917 2.032692 2.030719 2.056834 2.043332 2.052564 2.028446 2.026154 2.010904  mz224215 16456 16483 16264 13271 14363 13876 13369 13380 15221 15219  27267 28104 27306 26729 27504 29036 30207 32181 33521 32722  37026 24321 27395 21048 15195 24691 31165 28556 24141 34578 26811.6  21325.65 14497.4 17220.6 10522.15 7673.397 11639.28 15723.96 14657.15 10869.6 15961.86  0.575964 0.596086 0.628604 0.499912 0.504995 0.471398 0.504539 0.513277 0.450255 0.461619  2.050705 , 2.055235 2.038016 2.043103 2.069478 2.072075 2.061956 2.064687 2.055804 2.061522  13518 12893 12893 12420 12783 13829 14926 15554 16406 17779 17779 16216 16535 15402 14516  32525 32564 32122 31503 31270 31684 31279 30867 30862 30034 30034 29577 31260 32075 33254  557815 222428 331168 191395 227481 180040 188206 272910 385189 146510 173724 296530 356605 69535 573116 278176.8  227343.6 119022.2 153642 90195,64 102179.6 98188.33 103222.9 1.18440.9 170191.4 50481.42 63015.69 103480.9 136324 3392127 235003  0.407561 0.535104 0.46394 0.471254 0.449179 0.54537 0.548457 0.433992 0.441839 0.34456 0.362734 0.348973 0.382283 0.48783 0.410044  2.050614 2.034062 2.041052 2.028911 2.045491 2.031821 2.037595 2.027493 2.048451 2.054906 2.000822 2.045438 2.048168 2.022794 2.042765  -1.075681 -1.05031 -1.060668 -1.050775 -1.068529 -1.068201 -1.053854 -1.06718 -1.074824 -1.089532 -1.080555 -1.070894 -1.074928 -1.066186 -1.06503  1.209599 1.137339 1.166185 1.139064 1.190156 1.183345 1.144837 1.188238 1.211171 1.261317 1.23338 12002 1.211814 1.181457 1.180998  -1.099682 -1.090292 -1.160223 -1.180661 -1.141217 -1.117746 -1.111542 -1.145056 -1.136191 -1.150783  1.278873 1250169 1.47489 1.546219 1.416293 1.34824 1.334127 1.440912 1.3993 1.461781  -1.075931 -1.074504 -1.107437 -1.103838 -1.106239 -1.117695 -1.098322 -1.123952 -1.123787 -1.096451  1.216284 1.220442 1.327622 1.316959 1.318184 1.358815 1.301161 1.39338 1.37309 1.293258  -1.111261 -1.146945 -1.18447 -1.203354 -1.190359 -1.18295 -1.194571 -1.262292 -1.165234 -1.253231 -1.181724  1.321743 1.448477 1.554588 1.627347 1.584349 1.528099 1.551069 1.7862 1.525476 1.844083 1.553504  mz227968 14202 14831 13677 14392 14097 15013 13203 11510 14757 13601  29015 29474 33642 30325 29073 26799 26330 27047 29663 22447  43510 48657 96094 126346 100840 66385 123079 136357 105703 87331 93430.2  16203.85 18159.99 28352.41 36151.13 30494.99 193342 30053.08 34961,76 32816.51 24154.39  0.372417 0.373225 0.295049 0286128 0.30241 0.291244 0.244177 0.256399 0.31046 0.276584  2.062188 2.061575 2.100593 2.111118 2.064793 2.025357 2.030183 2.087544 2.062178 2.070846  mz228129 19947 20186 20780 20905 7549 8242 8835 8963 9678 6800  32958 31864 31890 31346 21164 20861 21104 20840 20244 25798  160845 390970 625440 421422 734155 233425 593705 759564 269816 562428 475177  5681326 112990.9 146991.8 96661.34 199130.2 5561725 138607.9 170685.4 82735.42 137087.3  0.353217 0289001 0235021 0.229369 0271237 0238266 0.233463 0.224715 0.306636 0.243742  2.045707 2.041287 2.062578 2.060293 2.064416 2.069364 2.050497 2.068821 2.074507 2.056041  mz231167 12840 13680 13963 13278 12418 14390 14824 14605 14970 17465 17464  13682 13386 16507 16505 16670 29653 28288 27795 43698 41289 40830  259090 129773 233059 303023 211648 39574 56215 21274 49870 44818 95380 131247.6  73841.68 37779.25 67464.34 86697.4 57132.71 14166.71 21845.65 7312.387 11802.08 11052.11 24102.04  0.285004 0.291118 0.289473 0.286108 0.269942 0.35798 0.388609 0.343724 0.236657 0.2466 0.252695  2.060878 2.054583 2.114795 2.123062 2.114744 2.099651 2.138215 2.173474 2.077866 2.140417 2.111923  mz234490 11348 12705 13319 14055 13719 13294 12804 13228 13213 15983  39831 30097 30097 30316 31132 31676 31073 30295 30607 35117  114351 60309 71653 53336 62493 146325 261492 04451 94430 49706 100854.6  47836.94 26251.62 28902.29 26159.79 30818.97 66341.33 99789.94 41514.9 44661.78 23926.76  0.418334 0.435285 0.403365 0.490472 0.493159 0.453383 0.381618 0.439539 0.472962 0.481366  2.072643 2.072817 2.066292 2.046661 2.063287 2.054565 2.061259 2.061353 2.069585 2.064249  -1.109304 -1.108099 -1.097423 -1.085752 -1.091364 -1.083084 •1.093312 -1.087006 -1.100616 -1.097537  1.307661 1.302524 1.270605 1.24039 1.250567 1.226219 1.263189 1.235556 1.276442 1.276001  -1.129785 -1.145047 -1.113382 -1.425039 -1.142129 -1.064558 -1.098435 -1.109607 -1.096478 -1.106854  1.41015 1.450446 1.351606 2.557432 1.501666 1.193847 1.2982 1.340433 1.28303 1.311424  -1.227151 -1.385078 -1.075811 -1.375534 -1.090824 -1.114399 -1.077496 -1.152484 -1.081727 -1.121756  1.709232 2.366202 1.221973 2.348865 1.260827 1.336523 1.218109 1.449395 1.223641 1.342414  -1.078021 -1.073 -1.082898 -1.08944 -1.084312 -1.069096 -1.080856 -1.070092 -1.105901 -1.084896  1.219307 1.209656 1.233509 1.262285 1.248654 1.192079 1.236115 1.206359 1.314833 1.24639  mz235848 12495 16342 15977 14663 13710 11786 10679 10130 8873 6960  21563 22661 23192 24023 24814 24856 24558 22548 21618 23888  22724 65522 111018 65442 34105 118746 58986 39178 14939 21991 55265.1  4431.709 15191.94 30332.87 6356.586 7581.656 28495.63 14103.07 9418.571 4847.873 , 6707.164  0.195023 0.2*186 0.273225 0.097133 0.222303 0.239971 0.239092 0.240405 0.324511 0.304996  2.072011 2.071601 2.056275 2.210371 2.069052 2.023541 2.057218 2.062044 2.011352 2.066876  mz237130 7934 7934 7545 10513 9725 9118 8913 7557 5185 6090  23517 24236 25041 25378 29192 29374 28597 28768 29381 34013  91372 60084 109543 131911 209498 15917 49212 16073 29182 39685 75247.7  8854.065 6311.839 29690.05 13396.24 78634.68 5094.872 22293.33 6290.729 15193.97 17509.89  0.096901 0.10505 0.271036 0.101555 0.375348 0.32009 0.453006 0.391385 0.520662 0.441222  2.140468 2.211496 2.031198 2.208133 2.057744 2.071105 2.05087 2.095949 2.046901 2.08212  -  mz29419a 12766 13379 12683 13352 13953 14582 15216 15486 15824 14179  32318 32318 33484 33484 33206 32599 32335 31240 31174 . 33305  224622 386207 191630 515100 441610 127902 441169 424637 263410 418516 343480.3  99987.43 130788.7 79966.82 176859.4 145390.2 57997.87 149297.2 143009.7 96353.52 146524.5  . 0.445136 0.338649 0.417298 0.34335 0.329228 0.453456 0.338413 0.336781 0.365793 0.350105  2.02467 2.045574 2.053955 2.054587 2.0508 1.999605 2.04943 2.009689 2.063538 2.052477  156  mz210922  12684 10749 G846 9846 10751 9611 9312 10888 11048 10062  22833 23019 23019 23586 24516 25284 24808 23812 23013 23143  107702 40663 136803 103519 06809 63862 85163 118567 39818 156628 04953.4  58364.12 2029228 63701.58 45427.73 47023.40 20065.72 38430.74 48821.00 18824.16 77506.38  0.541904 0.409035 0.466303 0.438835 0.485735 0.469226 0.451367 0.411750 0.472755 0.405418  2.0642 2.046356 2.057036 2.066601 2.0761 2.063605 2.051033 2.075264 2.059067 2.054778  -1.093077 -1.070013 •1.09812 -1.000852 -1.111601 -1.003626 -1.086028 -1.113542 -1.084575 -1.002411  1.256505 1217263 1272663 1280177 1.311505 1260226 1235084 1.321822 1230723 1256815  IMAGE ANALYSIS - PRIMARY VS. RECURRENT PLEOMORPHIC ADENOMAS Data - Individual Cases Recurrent Pleomorphic Adenoma Column Column Column Column Column Column Column Column  1 2 3 4 5 6 7 8  - stage x - stage y • area • Integrated(total) optical density - mean optical density - coefficient of variance of optical density - optical density skewness - optical density kurtosis  Z47S7 26134 26322 25424 25608 26410 25893 25289 26495 25927 25403  33838 34487 32832 30287 31456 32787 32616 31279 33725 33062  158268 181204 123375 330348 259563 228644 248452 554126 351375 184837 262019.2  45629.57 53337.03 38510.48 97201.18 75619.66 64386.18 57293.75 139383.9 86573.92 40400.03  0.288306 0.294348 0.312142 0.294239 0.291335 0.2816 0.230603 0.251538 0.246386 0.218571  2.123956 2.141916 2.09882 2.139065 2.13613 2.11729 2.164434 2.142262 2.193517 2.149301  -1.198167 -1.237375 -1.158695 -1.213103 -1.224472 -1.185002 -1.291254 -1.2472 -1.31575 -1.262863  1.58394 1.740664 1.476077 1.639088 1.706596 1.562418 1.968606 1.801993 2.018696 1.853315  2.089851 2.104633 2.058333 2.07039 2.091564 2.07514 2.109678 2.110607 2.123551 2.138459  -1.140269 -1.169705 -1.08725 -1.113307 -1.152552 -1.112335 -1.168769 -1.167538 -1.191787 -1.214138  1.405882 1.509901 1.245226 1.337146 1.449724 1.319382 1.503256 1.494648 1.577118 1.655248  Z6508  29813 30429 31230 31470 31514 30800 29844 28543 29381 29946  22065 22033 22468 23518 22787 21886 22298 13786 15148 14448  656065 215171.5 651309 174506.4 650198 226348.1 725166 195674.1 331899 101197.8 501420 159381.9 766516 211796.9 563305 146460.1 594843 181420.6 531750 130863.7 597247.1  0.327973 0.267932 0.348122 0.269834 0.304906 0.317861 0.276311 0.260001 0.304989 0.2461  Z6917  29599 30148 30276 29061 28828 30055 29713 29863 28642 29033  23747 24188 24984 25158 25558 25803 27550 28343 28885 29860  113513 229955 212193 215932 247503 184437 531030 440586 380400 388109 294365.8  31530.73 80425.21 60729.85 59976.43 57502.18 54437.62 173015.5 148437.1 133821.1 126601  0.277772 0.349743 0.286201 0.277756 0.232329 0.295156 0.325811 0.336908 0.351791 0.3262  2.202347 2.128867 2.132551 2.207758 2.307463 2.181202 2.142161 2.11609 2.116264 2.106674  -1.322466 -1.19188 -1.216733 -1.320792 -1.475595 -1.26936 -1.213752 -1.179439 -1.190248 -1.170153  2.018816 1.559898 1.673249 1.998025 2.556883 1.804938 1.631483 1.5354 1.566027 1.498894  z33802b  25118 23858 21820 20368 24065 27864 26926 27915 26416 24505  35110 34444 34008 33873 32139 34504 28774 33340 33182 34089  188055 89777 469866 342955 419848 417570 380340 517284 158563 144994 312925.2  38367 15399.27 8656227 71339.13 39496.03 6806125 32504.8 81613.19 25322.15 22566.95  0.20402 0.171528 0.184228 0.208013 0.094072 0.162994 0.085462 0.157772 0.159698 0.155641  2.076315 2.104514 2.123674 2.120223 2.138206 2.130958 2.212531 2.126138 2.111846 2.140426  -1.12099 -1.164018 -1.195246 -1.21959 -1.266646 -1.20829 -1.3931 -1.201838 -1.174763 -1.216242  11355728 1.492886 1.590418 1.670532 1.876427 1.629059 2.511057 1.619614 1.510886 1.660694  z16945e  25211 25607 22323 21608 22333 21346 21210 21465 22827 24824  21044 21216 34837 31752 32593 31694 26565 27643 27885 27112  319738 486919 173319 334275 554684 524873 401198 317938 267932 277159 365803.5  75111.15 82183.21 22058.44 55149.53 95242.62 91735.74 86177.94 62608.43 51643.21 43649  0.234915 0.168782 0.127271 0.164982 0.171706 0.174777 0.214802 0.19692 0.192747 0.157487  -1.20927 -1.298631 -1.325692 -1.210188 -1.227761 -1.177816 -1.246458 -1.296128 -1.232469 -1.389543  1.665322 2.01241 2.151475 1.743518 1.802726 1.608251 1.797561 2.063584 1.78379 2.356185  2.11749 2.109895 2.148229 2.128839 2.117679 2.125553 2.088696 2.112169 2.117911 2.143796  -1.230138 -1.176016 -1.284747 -1.202389 -1.191949 -1.222794 -1.155161 -1.193782 -1.201053 -1.21941  1.758709 1.534011 1.945279 1.60944 1.589829 1.739889 1.468247 1.622821 1.616055 1.652458  2.135828 2.114629 2.157313 2.116015 2.124213 2.112807 2.073122 2.099315 2.072358 2.086409  -1.227165 -1.179411 -1.294812 -1.188917 -1.192816 -1.173547 -1.116839 -1.167045 -1.120349 .^1.13139  1.729224 1.536623 2.041862 1.578317 1.577254 1.510486 1.34604 1.505614 1.353488 1.37875  2.122684 2.165184 2.175503 2.106473 2.116046 2.08563 2.143245 2.137807 2.124394 2.217244  z17758g  25530 26350 26835 25765 23154 22257 22756 23908 24132 29040  31347 31347 32140 32434 33188 33482 36947 37444 38466 36846  51240 213889 64295 95678 79575 100433 214435 192499 254971 78987 134600.2  9959.511 45075.63 14953.88 25412.05 16426.1 19128.62 52893.37 39699.76 469102 21569.3  0.19437 0210743 0.232582 02656 0206423 0.190462 0246664 0.206234 0.183983 0.273074  z20760f  24598 25581 27423 28221 30720 30460 29284 28078 26921 25820  30257 30112 30104 29354 27720 26224 26224 26068 25364 24993  256864 186474 316879 213975 88388 133071 312203 256856 415303 374255 255426.8  57968.75 41346.07 59809.72 52346.32 24852.86 39270.86 85881.5 57234.27 95659.86 106325.5  0.225679 0.221726 0.188746 0.244638 0.281179 0.295112 0.275082 0.222826 0.230338 0.284099  z23783f 18440 25302 27962 23239 19955 18684 19183 19035 20254 26821  28234 20867 21850 22750 31123 31907 30736 32218 35818 40996  188786 145111 136888 111460 93975 152075 158656 143686 96697 179531 140686.5  52676.54 23632.1 24499.7 15968.81 24826.35 44373.15 35840.82 35080.68 21723.03 30595.33  0.279028 0.162855 0.178976 0.143269 0.26418 0.291785 0.225903 0.244148 0.224651 0.170418  2.089331 2.091067 2.103702 2.111897 2.092507 2.077936 2.106477 2.089727 2.123915 2.206937  -1.13116 -1.151362 -1.155193 -1.207734 -1.137528 -1.118684 -1.15131 -1.130814 -1.177233 -1.350937  1.363672 1.445539 1.450736 1.663182 1.380492 1.327074 1.426227 1.363334 1.505398 2216818  2.103172 2.089532 2.098581 2.076868 2.074652 2.090885 2.081657 2.095188 2.086548 2.153754  -1.175223 -1.163426 -1.161274 -1.137487 -1.123816 -1.16369 -1.139581 -1.156149 -1.17959 -1.256395  1.553532 1.544243 1.489736 1.451571 1.37548 1.534983 1.435542 1.476047 1.598935 1.822037  2.189465 2269143 2.328735 2.177711 2.185463 2.195299 2233885 2.185012 2.140377 2212698  -1.308458 -1.419288 -1.567752 •1.300323 -1.31125 -1.334365 -1.375498 -1.374262 -1.224439 -1.343315  1.984344 2.353714 3.024052 1.92727 2.012475 2.100806 2.228477 2.365959 1.663737 2.110775  2.146586 2.083698 2.10727 2.111928 2.127013 2.097156 2.110144 2.156448 2.120029 2.125551  -1.215617 -1.128877 -1.182968 -1.223623 -1.186913 -1.149323 -1.184505 -1.248809 -1.190622 -1.200797  1.629739 1.360179 1.543343 1.756187 1.540083 1.417528 1.581069 1.769201 1.576981 1.613065  Z25392f  19729 24792 25307 27061 25727 29134 26624 27819 25485 25259  33293 34529 35361 36286 37542 21748 20828 20654 19747 20916  157868 503582 525448 632161 586949 452824 439374 443899 196699 40814 397961.8  26514.12 68258.81 103171.8 88849.87 94079.63 64623.11 69407.74 81234.81 36130.4 10257.15  0.167951 0.135547 0.19635 0.140549 0.160286 0.142711 0.15797 0.183003 0.183684 0.251315  225572b  29624 20524 19163 21792 23405 22868 23603 26038 26627 23205  26441 35455 34717 36857 32704 28345 27595 26765 24037 27061  68301 92519 187824 51070 138969 339132 243726 364957 31331 237220 175504.9  13990.94 21239.9 36915.7 15951 3390228 65990.33 51266.89 50555.53 10241.08 54405.72  0.204842 0.229573 0.196544 0.312336 0.243956 0.194586 0210346 0.138525 0.326867 0.229347 Z28642I  29374 29247 27365 25038 27338 24508 26277 25794 25817 26837  24178 25425 20285 23612 23952 22738 24591 19369 17296 17095  60515 66947 195274 155182 52159 10121 363834 107533 158497 195521 136558.3  2011221 25048.05 41457.21 25126.73 16840.05 3673.626 63480.59 25039.53 29765.54 37833.51  0.332351 0.374147 0.212303 0.161918 0.32286 0.362971 0.174477 0.232854 0.187799 0.193501  /  z29073d  19119 22953 23740 22992 25246 23579 24546 19824 23796 22435  27256 26694 26435 25682 29128 31755 33361 36004 37009 37306  9450 . 3964.874 12408 4408.567 207730 48955.68 20147 9662.278 121812 29590.49 271004 79060.79 115313 28883.3 23965 9190.286 88704 17696.59 74743 15122.04 94527.6  0.419563 2.105484 0.3553 2.13453 0.23567 2.182646 0.479589 2.102036 0.242919 2.146455 0.291733 2.146275 0.250477 2.200838 0.383488 2.157716 0.199502 2.218667 0.202321 2.206867  -1.161894 -1.211553 -1.300276 -1.152012 -1.249693 -1.223724 -1.321211 -1.225579 -1.344467 -1.349234  1.450312 1.616312 1.963732 1.416527 1.784814 1.671331 2.018051 1.649454 2.106018 2.130141  -1.161458 -1.288224 -1.158646 -1.090363 -1.270442 -1.213224 -1.158425 -1.168657 -1.153756 -1.149914  1.455411 1.895797 1.440438 1.240616 1.844156 1.637835 1.465572 1.486277 1.444045 1.420645  Z35368I  25375 28148 28142 28656 25142 25895 27949 28910 29334 28266  30498 23538 29167 28746 32772 33781 38737 36739 38374 33028  11393 20319 39380 36879 77786 95496 109899 92240 80367 30649 59440.8  4081.601 5081.118 18680.94 21283.36 23708.75 28502.47 25727.99 25213.54 24566.02 13518.11  0.358255 0.250067 0.474376 0.577113 0.304795 0.298468 0.234106 0.273347 0.305673 0.441062  2.106302 2.169107 2.114873 2.064711 2.16025 2.140145 2.104366 2.11476 2.103369 2.105848  

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:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0086945/manifest

Comment

Related Items