Louisiana State University LSU Digital Commons

LSU Historical Dissertations and Theses Graduate School

1995 Mechanisms of Chemically-Induced Hepatocarcinogenesis in Western Mosquitofish (Gambusia Affinis). Jerry Mchugh Law Jr Louisiana State University and Agricultural & Mechanical College

Follow this and additional works at: https://digitalcommons.lsu.edu/gradschool_disstheses

Recommended Citation Law, Jerry Mchugh Jr, "Mechanisms of Chemically-Induced Hepatocarcinogenesis in Western Mosquitofish (Gambusia Affinis)." (1995). LSU Historical Dissertations and Theses. 5966. https://digitalcommons.lsu.edu/gradschool_disstheses/5966

This Dissertation is brought to you for free and open access by the Graduate School at LSU Digital Commons. It has been accepted for inclusion in LSU Historical Dissertations and Theses by an authorized administrator of LSU Digital Commons. For more information, please contact [email protected]. INFORMATION TO USERS

This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer.

The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandardmargins, and improper alignment can adversely affect reproduction.

In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note wifi indicate the deletion.

Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand comer and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book.

Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.

A Bell & Howell Information Company 300 North Zeeb Road, Ann Arbor, Ml 48106-1346 USA 313/761*4700 800/521*0600

MECHANISMS OF CHEMICALLY-INDUCED HEPATOCARCINOGENESIS IN WESTERN MOSQUITOFISH (GAMBUSIA AFFINIS)

A Dissertation

Submitted to the Graduate Faculty of the Louisiana State University and Agricultural and Mechanical College in partial fulfillment of the requirements for the degree of Doctor of Philosophy

in

The Interdepartmental Program in Veterinary Medical Sciences through the Department of Veterinary Pathology

by Jerry McHugh Law, Jr. D.V.M., Louisiana State University, 1985 May, 1995 UMI Number: 9538746

UMI Microform 9538746 Copyright 1995, by UMI Company. All rights reserved.

This microform edition is protected against unauthorized copying under Title 17, United States Code.

UMI 300 North Zeeb Road Ann Arbor, MI 48103 DEDICATION

This work is dedicated to the memory of Harold W. Casey, D.V.M., PhD., and former Head, Department of Veterinary Pathology, LSU School of Veterinary

Medicine. Dr. Casey had a distinguished career as a Colonel in the United States

Air Force, serving for 22 years. He came to LSU in 1982 and served as

Department Head for 10 years. Dr. Casey loved pathology and always had time to sit down at the microscope with his graduate students. He made a number of major contributions to the field of cancer research throughout his career. He died on April

12, 1993, of the very disease he fought so hard to find answers to. ACKNOWLEDGEMENTS

"If I have seen further, it is by standing on the shoulders of giants." So said

Sir Isaac Newton in a letter to Robert Hook dated February 5, 1676. This axiom of scientific achievement remains valid today. Even the small contribution to scientific knowledge contained in the following pages could not have been possible without building on the work of many others, nor without the help of many people.

I would like to thank my co-major professors, Dr. H. Wayne Taylor and Dr.

Jay C. Means. Dr. Taylor helped to keep things organized and on time, and advised me on the nuances of liver pathology. Dr. Means graciously stepped in when we started looking for an aspect of small fish carcinogenesis that would complement the histopathology studies. He has contributed immensely to my understanding of environmental toxicology and analytical chemistry, helped me develop my grant writing skills, and helped me make some important professional connections.

Dr. David Swenson, a third committee member, was the truly brilliant "idea man" behind the DNA adduct work. He helped me see the value of this aspect of cancer research, provided laboratory space, and nudged me along even as my ignorance of chemical synthesis glared through. Thanks also to Dr. Jan

VanSteenhouse, another committee member, for the use of laboratory equipment and for helping me re-leam clinical pathology from a pathologist’s point of view.

It would to be difficult to say enough "thank you’s" to Dr. William E.

Hawkins of the Gulf Coast Research Laboratory in these few paragraphs. I developed an interest in small fish carcinogenesis when I started my research, and

Bill unselfishly opened his lab and its substantial resources to me. He has included

me in a number of meetings and pathology working groups, and his considerable

reputation has allowed me to gain the "inner circle" of this area of research. Many

thanks to Steve Manning, Rena Krol, Retha Edwards, Sue and Don Barnes, and

Nate Jordan for their excellent technical assistance. Steve, in particular, helped me

to understand exposure methodology and how to deal with aquatic species under

GLP conditions. Lisa Ortego helped me understand cell proliferation and developed

the PCNA methodology for use in fish tissue sections. Drs. William W. Walker and Robin M. Overstreet, also of GCRL, helped in many ways, including providing

fish and conceptualizing the exposure methods, and opened up their laboratories to

my use.

Dr. Joe Newton started out as my major professor, but then had to move on to "sweet home, Alabama." I owe him thanks for teaching me basic research lab skills and teaching me how to say "no" when confronted with numerous distractions.

Additionally, Dr. Newton saw the need for a molecular biology aspect to my project and connected me with Dr. Swenson. Thanks, Joe.

Back at LSU SVM, Mae Lopez deserves a special thanks. She diligently embedded and sectioned literally hundreds of fish specimens, taught me immunohistochemistry, cut numerous thin sections for EM, and printed photos. She is certainly multi-talented and a very valuable asset to the Department of Veterinary

Pathology. Thanks also to Cheryl Crowder and the rest of the histology lab. Pam Stratton provided always cheerful advice on clerical matters. I am further indebted

to all of the faculty in Veterinary Pathology that contributed to my graduate training

and taught me the value of diagnostic pathology.

Another special note of thanks goes to Debra McMillin. She patiently helped

me develop the GC/MS method for DNA adduct detection, and steadfastly refused

to allow sloppy work. Thanks also to my father-in-law, Bill Behrmann, for helping

me with the graphs and tables. Dr. Steve Barker and Connie David were always available for advice on analytical matters.

Lastly and most importantly is my family. Loving thanks to my wife,

Christine, and my daughter, Rebecca, for the patience they have shown throughout these extra years of schooling for Daddy. It hasn’t always been easy or smooth, but

I know it will be worth it. Thanks also to my parents, Jerry and Helen Law for their encouragement throughout my life’s choices, and to God, my source of help and strength in time of need.

v TABLE OF CONTENTS

DEDICATION...... ii

ACKNOWLEDGEMENTS ...... iii

LIST OF TABLES ...... vii

LIST OF FIG URES...... viii

ABSTRACT ...... xi

CHAPTER 1. INTRODUCTION, LITERATURE REVIEW, AND OBJECTIVES...... 1

CHAPTER 2. HEPATOCARCINOGENESIS IN WESTERN MOSQUITOFISH (GAMBUSIA AFFINJS) EXPOSED TO METHYLAZOXYMETHANOL ACETATE...... 58

CHAPTER 3. CELL PROLIFERATION, P53, AND APOPTOSIS IN HEPATOCARCINOGENESIS OF WESTERN MOSQUITOFISH (GAMBUSIA AFFINIS) EXPOSED TO METHYLAZOXYMETHANOL ACETATE...... 77

CHAPTER 4. QUANTIFICATION OF DNA ADDUCTS AT FEMTOMOLE LEVELS IN FISH EXPOSED TO ALKYLATING AGENTS...... 136

CHAPTER 5. SUMMARY AND CONCLUSIONS ...... 165

APPENDIX A. DATA FROM DEVELOPMENT OF DNA ADDUCT QUANTIFICATION METHOD USING GC/MASS SPECTROMETRY...... 176

APPENDIX B. LETTER OF PERMISSION ...... 180

VITA ...... 182

vi LIST OF TABLES

Table 2.1 Histopathological examination of western mosquitofish...... 63

Table 3.1 Experimental d esig n ...... 82

Table 3.2 Histopathological examination ...... 89

Table 3.3 Incidence of anti-p53 immunopositivity in western mosquitofish exposed to 1 or 10 ppm MAM-Ac...... 116

Table 4.1 Retention times and characteristic mass fragments of guanine adducts . . . 147

Table 4.2 Alkylguanine adduct concentrations in liver DNA of western mosquitofish 24 hr after exposure to 10 ppm MAM-Ac ...... 151

Table 4.3 Alkylguanine adduct concentrations in liver DNA of western mosquitofish before, mid-way through (-1), and after 2 hr pulse exposure to 10 ppm MAM-Ac ...... 153

vii LIST OF FIGURES

Figure 1.1 Stages of hepatocarcinogenesis...... 12

Figure 1.2 Possible pathways of differentiation ...... 15

Figure 1.3 Inter-relationships between cell proliferation, p53, and apoptosis ...... 20

Figure 2.1 Liver from control mosquitofish ...... 64

Figure 2.2 Hepatocellular adenoma in a mosquitofish exposed to MAM-Ac ...... 66

Figure 2.3 Hepatocellular carcinoma in a mosquitofish exposed to MAM-Ac ...... 67

Figure 2.4 Detail of liver from Fig. 2 .3...... 67

Figure 2.5 Cholangiocellular carcinoma in the liver of a mosquitofish exposed to MAM-Ac for 2 hr and killed 25 weeks later ...... 68

Figure 2.6 Cholangiocellular carcinoma in the liver of a mosquitofish exposed to MAM-Ac for 2 hr and killed 40 weeks later ...... 68

Figure 2.7 Electron micrograph of cholangiocellular carcinoma in a mosquitofish exposed to MAM-Ac for 2 hr and killed 40 weeks later ...... 70

Figure 2.8 Highly dense, roundish structure resembling an apoptotic b o d y ...... 70

Figure 3.1 Epitheliocystis sp...... 88

Figure 3.2 Diffuse cytotoxicity and numerous cytoplasmic inclusions in hepatocytes . 91

Figure 3.3 Cytotoxic changes in hepatocytes from a mosquitofish 2 weeks after exposure to M AM -Ac ...... 92

Figure 3.4 Tumor incidence by type in exposed adult Gambusia a ffin is...... 94

Figure 3.5 Tumor incidence by type in Gambusia affinis high dose f r y ...... 95

Figure 3.6 Spindle cell tumor with a whorling pattern ...... 96

Figure 3.7 Megalocytosis and spindle cell proliferation in liver of an adult mosquitofish 8 weeks after exposure to M A M -A c ...... 98

viii Figure 3.8 Apoptosis (arrows) and megalocytosis in liver of an adult mosquitofish 8 weeks after exposure to M AM -Ac ...... 99

Figure 3.9 Eosinophilic fo c u s ...... 100

Figure 3.10 Hepatocellular adenom a ...... 101

Figure 3.11 Hepatocellular carcinom a ...... 103

Figure 3.12 Highly anaplastic spindle cell tumor in an adult mosquitofish 24 weeks after exposure to M AM -Ac ...... 104

Figure 3.13 Well-differentiated exocrine pancreatic carcinoma in the liver of an adult mosquitofish 24 weeks after exposure to M AM -Ac ...... 106

Figure 3.14 Spindle cell tumor (possible intersitial cell tumor) in the testis of a mosquitofish fry 41 weeks after exposure to M A M -A c ...... 107

Figure 3.15 Electron micrographs of liver cells from an adult mosquitofish 24 weeks after exposure to M AM -Ac ...... 109

Figure 3.16 Electron micrographs of liver cells from an adult mosquitofish 24 weeks after exposure to M AM -Ac ...... 110

Figure 3.17 Typical PCNA staining in the liv e r ...... I l l

Figure 3.18 Mean liver cell labeling indices (Li’s) for control and exposed adult mosquitofish determined by PCNA immunohistochemistry ...... 112

Figure 3.19 Mean liver cell labeling indices (Li’s) for control and exposed low dose mosquitofish fry determined by PCNA immunohistochemistry ...... 114

Figure 3.20 Mean liver cell labeling indices (Li’s) for control and exposed high dose mosquitofish fry determined by PCNA immunohistochemistry ...... 115

Figure 3.21 Typical p53 staining in the liv e r ...... 118

Figure 4.1 Experimental design for DNA adduct quantification by GC/MS in small fish species...... 142

Figure 4.2 Extracted ion chromatograms of the four DNA adduct standards and their deuterated analogs ...... 148

Figure 4.3 Mass spectra of the four alkyl DNA adduct standards ...... 149

ix Figure 4.4 Methylguanine adduct concentrations in liver DNA of control and exposed western mosquitofish 24 hr after exposure to 10 ppm MAM-Ac . 152

Figure 4.5 Methylguanine adduct concentrations in liver DNA of western mosquitofish mid-way through (-1) and after 2 hr pulse exposure to 10 ppm MAM-Ac...... 154

Figure A. 1 Standard curve for adducts with duplication at nine le v e ls ...... , 177

Figure A.2 Selected ion chromatograms from a typical DNA adduct analysis ...... , 178

Figure A.3 Total ion chromatogram (TIC) from a typical DNA adduct analysis . . . . 179

X ABSTRACT

A need exists for fish sentinel species for monitoring environmental changes.

These studies were performed to examine the suitability of western mosquitofish

(Gambusia affinis) as a sentinel species and to explore basic mechanisms of

hepatocarcinogenesis.

In the first series of experiments, neoplastic lesions were induced in

laboratory-reared mosquitofish fry and adults using single pulse exposures in the ambient water to 1 or 10 ppm methylazoxymethanol acetate, a model alkylating carcinogen. Histopathology, electron microscopy, and immunohistochemistry were performed on groups of fish examined at 8 intervals from 24 hours to 41 weeks post-exposure. Early microscopic changes in the liver of exposed fish included severe cytotoxicity, megalocytosis, and widespread apoptosis. Neoplastic changes were found in the liver in approximately 20% of the exposed fish by 8 weeks after exposure and increased to approximately 70% by 41 weeks. These included foci of cellular alteration, hepatocellular carcinomas and cholangiocellular carcinomas.

There was also significant incidence of a spindle cell tumor, believed to be of hepatobiliary stem cell origin, which was often admixed with biliary ductular profiles. The only non-hepatic neoplasms observed were several exocrine pancreatic adenomas and carcinomas. Tumors were found in less than 1 % of controls.

Immunohistochemistry was performed on fixed, paraffin-embedded histologic sections of the fish from the histopathology study using avidin-biotin complex methodology. Randomly selected specimens were stained for proliferating cell nuclear antigen (PCNA, PC10), a protein intrinsically associated with the cell cycle, and a labeling index was developed for each specimen group. Significantly increased cell proliferation detected in liver cell populations of the high dose groups at early time points was attributed primarily to regenerative hyperplasia, and at later time points mainly to preneoplastic and/or neoplastic cell proliferation. Additional sections of the same fish specimens were stained for p53 protein, the product of the p53 tumor suppressor gene, using a commercially available monoclonal antibody.

Overexpression of p53, suggestive of p53 gene mutation, was detected as early as 48 hours after exposure and in approximately 38% of the tumors. However, at the early sampling time points, p53 immunohistochemistry did not appear to have good predictive value of later tumor formation. These data agree with findings in many human and other animal tumors by other workers and suggest a link between p53 gene mutation, failure of apoptosis, and inappropriate cell proliferation in hepatocarcinogenesis.

Levels of the DNA adducts 0 6-methylguanine and 7-methylguanine were measured in the liver in additional alkylating agent exposures using stable isotope dilution-GC/mass spectrometry. Adapted for small fish species, this assay gave detection limits of approximately 1.6 femtomoles per microgram of DNA, or about

1 adduct per 2 million DNA bases. Levels of 0 6-methylguanine ranging from approximately 55-185 picograms (330-1100 femtomoles) per microgram DNA, measured in the first 72 hours after exposure, were correlated with a 33% liver tumor incidence after 25 weeks in parallel histopathology studies. Although further studies are needed with this and other compounds, and with other fish species, this method appears to be useful for studies of carcinogens and mutagens in aquatic environments. Further, the carcinogen sensitivity, ease of laboratory culture, biology, and wide range of Gambusia affinis appear to indicate excellent potential for its use as a warmwater sentinel species and should allow direct validation of field studies by enabling substances to be tested under both laboratory and field conditions. CHAPTER 1. INTRODUCTION, LITERATURE REVIEW, AND OBJECTIVES

Introduction

The "war on cancer," initiated by President Richard M. Nixon when he signed the National Cancer Act in 1971, is far from over. Despite a massive assault on the problem, few true improvements in U.S. cancer death rates have been shown.

Although death rates from lung cancer have fallen in men, they have more than doubled in women since 1973, adding up to the largest contribution to increased cancer death rates (1.14). However, even if lung cancer deaths are excluded, and despite treatment advances, certain individual cancers have shown increasing death rates. These disturbing facts underscore an urgent need to understand the basic mechanisms of neoplasia and its causes so that effective treatment and prevention strategies can be implemented.

Carcinogenesis is a complex, multistage process involving carcinogen- induced genetic and epigenetic damage to susceptible cells that subsequently gain a selective growth advantage and undergo clonal expansion as the result of activation of protooncogenes and/or inactivation of tumor suppressor genes (1.68). Recent advances in cellular and molecular pathology have vastly broadened our horizons and provided new insights into the mechanisms of this process. A holistic approach to the study of these mechanisms should include not only investigations at several phyletic levels, but also investigations at all stages of the process, from initiation to the development of cancer. A review of the multistage nature of carcinogenesis is included in a later section of this chapter.

1 2 Cancer research is a multidisciplinary field that has a rich history of

accomplishments, ranging from discernment of causation to understanding processes

of clonal expansion and metastasis. Since there are numerous reviews covering the

subject (e.g., 1.17, 1.23, 1.47, 1.51, 1.61, 1.70, 1.77, 1.81, 1.86, 1.117, 1.137,

1.146, 1.178), it is not the purpose of this chapter to review the complete history of

cancer research. Rather, a brief overview will be given, emphasizing mechanistic

principles of hepatocarcinogenesis along with significant contributions from aquatic

pathobiology. Important work using methylazoxymethanol, the compound used in

the present studies, will also be included. Above all, the importance of a holistic

approach to carcinogenesis, integrating histopathological changes with molecular

biological and analytical data from all stages of the process, will be emphasized

throughout this work.

Cancer Risk from>the Environment

Environmental factors play a role in the occurrence of cancer, although the level of this role is the subject of much debate (1.166). Based on epidemiological evidence, Doll and Peto (1.45) estimated that environmental factors, including exposure to chemical carcinogens, account for nearly 80% of all human cancers in the United States. It is estimated that greater than 50,000 man-made chemicals are currently in use, and that between 500 and 1000 new chemicals are put on the market and introduced into our air, water, and soil each year (1.142). It is widely recognized that detection and appropriate regulation of these compounds is of great importance for the prevention of neoplasia in man (1.90). 3 Since "the proper study of mankind is man in relation to all things that surround him" (1.37) it is appropriate to extend our investigations to concern over impacts on a broader ecological scale. Thus, it would seem especially fitting for workers in the veterinary medical sciences to approach these problems from the standpoint of how chemical contaminants might affect all animals, including man.

Historically, the link between exposure to certain chemicals and cancer has been established using whole animal chronic bioassays. In the 1960’s and 70’s, U.S. government agencies were formed in response to concern over chemical safety. The

National Toxicology Program (NTP) was established in 1981 by the Secretary of

Health and Human Services to better coordinate and integrate toxicology research and applied studies, and was charged with: "1) broadening the spectrum of toxicologic information obtained on selected chemicals; 2) increasing the number of chemicals studies, within funding limits; 3) developing and validating assays and protocols responsive to regulatory needs; and 4) communicating Program plans and results to government agencies, the medical and scientific communities and the

Public" (1.88). With the advent of the Clean Air Act Amendments of 1990, the mandate of the NTP has become, essentially, to perform a greater number of tests on a sizable group of priority chemicals in a limited amount of time (5 years) for less money. Using present methodologies, this is a virtually unattainable goal (NTP administrators estimate it would take 19 years to test only the first group of chemicals on the list). Complete testing of one chemical takes 5 years and costs range from about a million dollars for a feeding study to several million dollars for 4 an inhalation study. Consequently, a need exists for alternative, less expensive and less time-consuming animal models for presumptive (stage one or tier one) or definitive tests. Short-term, in vitro assays of numerous types have been useful in mass screening of potentially carcinogenic compounds (e.g. the Ames test).

Although many of these assays are rapid and economical (1.63), their validity has been limited by large numbers of false negative and false positive determinations as well as an inherent inability to determine target organ-specific chemical carcinogenicity or promoting activity (1.90).

The Issue of Animal Rights

Vitally important to modem toxicology research is the issue of animal rights and justifying the use of large numbers of animals for chemicals testing. Cold­ blooded, "lower animals" seem to be less objectionable as test animals to a public increasingly sensitized to the use of animals for any kind of testing. The importance of animal rights issues in toxicity testing is illustrated by recent goals set forth in the

Strategic Plan of American Veterinary Medical Association (AVMA). Among seven recommendations sent to the AVMA Executive Board by the AVMA Animal

Welfare Committee at its November, 1991 meeting was an AVMA position statement on toxicity testing legislation. The adopted recommendation stated that,

"The AVMA will continue to support research, development, and validation of alternative testing methods that replace animals, reduce the numbers of animals used, or refine animal use to reduce pain...The AVMA opposes the regulatory approach to eliminating animal toxicity tests while it encourages the ongoing efforts of federal regulatory and research agencies, as well as private industry, to develop and validate alternatives. Proposed legislation 5 often assumes that animal-based testing is unnecessary and outmoded. Backed by current scientific knowledge, the AVMA rejects that assumption" (1.4).

The Use of Small Fish Species in Carcinogenicity Testing

For the above reasons as well as others, the use of fish as alternative test animals in carcinogenesis research has received considerable attention recently (e.g.,

1.19, 1.34, 1.38, 1.74, 1.75, 1.78, 1.85, 1.118, 1.123, 1.127, 1.140). Unlike in vitro methods, tests using small fish have the advantages of whole animal assays.

Many species of fish are available that are easily and economically bred and housed in a laboratory situation. They have been shown to be sensitive to a variety of known carcinogens with a short time to tumorigenesis, and yet have an exceedingly low spontaneous tumor rate in potential target organs. Additionally, most show little or no response difference between sexes and do not appear to be susceptible to non­ specific cultural conditions or "white noise" that can affect carcinogenesis testing using prevailing methods (1.74, 1.85).

Compared to their rodent counterparts, small fish species have a relatively small but expanding carcinogenicity database. Since small fish carcinogenicity testing has been sufficiently reviewed in the papers cited above, only a few selected studies will be mentioned here.

Although tumors in wild fish have been reported for decades, Dr. Mearl

Stanton appears to have pioneered the use of small aquarium fish for carcinogenicity testing in the controlled laboratory setting in the mid-1960s (1.36). A physician by training, Stanton had an interest in environmental causation of cancer. Using "zebra 6 dannies" ( Brachydanio rerio), he reported hepatic neoplasia in fish exposed to

diethylnitrosamine (1.164) and then cycasin (1.165).

Also in the mid-1960s, wild stocked rainbow trout ( Onchorhynchus mykiss)

were found to have an appreciable incidence of liver tumors (1.66, 1.153). It was

shown that aflatoxin, a mycotoxin occurring in moldy feeds, was the cause of the

liver neoplasia. Aflatoxins and their analogs have since been shown to be potent

liver carcinogens in many species, including man. Trout have subsequently been

used in numerous other carcinogenesis studies (e.g., 1.78-80).

Since that time, a number of laboratory studies have focused on small fish hepatocarcinogenesis, particularly involving the medaka ( Oryzias latipes) (1.24,

1.73, 1.82, 1.89), platyfish/swordtail hybrids ( Xiphophorus spp.) (1.2), topminnow

(Poeciliopsis spp.) (1.145), sheepshead minnow ( Cyprinodon variegatus) (1.33), and guppy (Poecilia reticulata) (1.59). There is a need, nonetheless, for a native, small freshwater fish that can be used to directly correlate laboratory studies with environmental field studies. Law et al (1.100) reported in a recent study that western mosquitofish {Gambusia affinis) are sensitive to chemical induction of liver neoplasia and are easily cultured in the laboratory.

Biomarkers of Environmental Damage

Besides their usefulness as test organisms, fish have been shown to be sensitive indicators of environmental contamination. Since exposure to toxic chemicals in the environment is difficult to assess because of the wide variety of potential exposure routes, differences in bioavailability of toxicants, and differences 7 in pharmacodynamic disposition of xenobiotics, many researchers have turned to the use of biological markers (biomarkers). Biomarkers are measurements of body fluids, cells, or tissues that indicate in biochemical or cellular terms the presence of contaminants or the magnitude of the host response (1.121). Data from biological systems provide vital information not readily available from chemical analyses of air, water, or soil. Environmental species have the potential of serving as 1) sentinels which demonstrate the presence of contaminants and the extent of exposure; 2) surrogates indicating potential human exposure and effects; and 3) predictors of long-term effects on populations or ecosystems (1.121).

Harshbarger and Clark (1.71) documented 41 geographic regions in North

America in which clusters, or epizootics, of cancer in wild fishes have occurred.

The occurrence of neoplasms involving epithelial tissues such as the liver, pancreas, gastrointestinal tract and some epidermal neoplasms appear strongly correlated with environmental contamination, that is, exposure to chemical carcinogens. For more exhaustive treatment, the reader is referred to recent reviews (1.20, 1.34, 1.38,

1.39, 1.71, 1.72, 1.127). However, several reports of tumors in wild fish have been pivotal and deserve special mention.

English sole from contaminated areas of the Puget Sound, Washington have high prevalences of liver lesions that range from megalocytosis to neoplasms

(1.131). Numerous detailed studies (e.g., 1.110-114) have established statistically significant associations between the presence of polycyclic aromatic hydrocarbons

(PAH) in the sediments and the prevalence of liver neoplasia. Recently, Malins et 8 al (1.115) identified a breakdown product of DNA guanine residues, 2,6-diamino-4-

hydroxy-5-formamidopyrimidine, in neoplastic livers of English sole from

carcinogen-impacted areas of the Puget Sound.

On the East Coast, epizootic hepatic neoplasia in winter flounder from

Boston Harbor, Massachusetts has been reported (1.129, 1.130). As in the case of

the Puget Sound sole but not as strictly established, the hepatic lesions in the winter

flounder were highly correlated with anthropogenic chemical contamination.

Although incidences of cancer epizootics have occurred in freshwater fishes

(1.20), none have been as well studied as the English sole and winter flounder epizootics. Epizootics of neoplasia in fish populations of brown bullhead catfish

(Ictalurus nebulosus) and Atlantic tomcod (Microgadus tomcod) should be noted.

Sediments rich in PAH have generally been considered the principal causes of skin and liver neoplasia in brown bullhead in the contaminated Black River (Ohio), a tributary of Lake Erie (1.11-1.13). In laboratory tests, medaka exposed to extracts and fractions of PAH-contaminated sediments from tributaries of the Great Lakes, including the Black River, developed liver neoplasia (1.50). Similarly, epizootics of hepatic neoplasia have been reported from Atlantic tomcod from the Hudson River

(1.31, 1.155). These liver neoplasms have been associated with elevated tissue levels of polychlorinated biphenyls (PCB’s) (1.93).

White suckers from industrially-polluted areas of Lake Ontario exhibited increased prevalences of hepatic and skin neoplasia (1.76, 1.157). As in other epizootics, the neoplasms have been associated with PAH contamination. Stalker et 9 al (1.163) showed that the progression of hepatocellular and bile duct neoplasms in

the white sucker is accompanied by a loss of immunoreactive glutathione S-

transferases which usually catalyze a major detoxification pathway.

Only a small number of reported cancer epizootics have dealt with small fish

species. Vogelbein and colleagues reported high prevalences of liver neoplasms in

mummichog (Fundulus heteroclitus) from a creosote-contaminated site in the

Elizabeth River, Virginia, (1.174). Pancreatic neoplasms have also been reported in

these fish recently (1.60, 1.173).

The Test Fish: Gambusia affinis

The western mosquitofish, Gambusia affinis, is a small, livebearing,

freshwater fish (order Atheriniformes; family Poeciliidae) that is native to the U.S.

It is found from New Jersey to northern Mexico and has been introduced world-wide

for mosquito control (1.41). The value of these fish for biological control of the

aquatic stages of the mosquito was first demonstrated in 1918 when they were used at Camp Hancock in Augusta, Georgia to control the spread of mosquito-borne diseases such as malaria. Worldwide distribution began in about 1920 when the

International Red Cross asked the U.S. government for brood stocks of Gambusia

for Italy and Spain. They have since become common in nearly all warm, fresh waters of the world. They are found in temperatures ranging from about 40° to

100° F. Their habitat includes fresh or brackish water, sluggish or standing water, ditches, pools, artesian well discharges, cisterns, water tanks, potholes, rain barrels, and even sewerage outfalls (1.41). 10 In the laboratory, mosquitofish are easily cultured using techniques developed

for other small fish species such as the guppy (Poecilia reticulata). Studies have

been performed on the acute toxicity of various chemicals for mosquitofish (e.g.,

1.43, 1.103, 1.132), but until recently there were no reports of neoplastic lesions in

either wild or laboratory-reared members of this species. They were recently shown

to be sensitive to the alkylating procarcinogen, methylazoxymethanol acetate, in a

controlled laboratory exposure (1.100).

The cytomorphology of mosquitofish is relatively well characterized (1.18,

1.26, 1.183). These fish have shown remarkable genetic plasticity with regard to

insecticide susceptibility or resistance (1.3, 1.55, 1.172, 1.181). The ultrastructure

of the Gambusia liver is also well characterized (1.176). In the wild, mosquitofish

occur in widely varied habitats and yet have a limited home range. They breed

prolifically and are apparently resistant to infectious diseases (personal

observations). Thus, they appear to hold great promise as a versatile model: once

toxicologic mechanisms in this fish are validated in the laboratory, future studies can

explore the use of mosquitofish as environmental sentinels.

The Multistage Nature of Hepatocarcinogenesis

Numerous volumes and review papers have been published on the stages of

carcinogenesis and, more specifically, hepatocarcinogenesis (e.g., 1.6, 1.51-53,

1.116, 1.138, 1.143, 1.154, 1.177, 1.179, 1.180). Several milestones are particularly noteworthy and will be highlighted here. 11 The first evidence that liver neoplasia occurs in sequential stages was

presented by Peraino and colleagues in 1971 (1.135), in which they showed that

phenobarbital promoted 2-acetylaminofluorene-induced hepatocarcinogenesis in rats.

Since then, a number of multistage hepatocarcinogenesis systems have been

developed, such as the Solt-Farber selection model or the Pitot partial hepatectomy

model (1.136). Farber and colleagues (1.53) showed that certain subpopulations of

cells in rat liver were resistant to cytotoxicity, mainly because of reduced or slowed

xenobiotic activation or enhanced detoxification systems. A small percentage of

cells which have acquired certain properties favoring survival and growth over their neighboring cells (i.e., neoplastic properties) will never fully redifferentiate to a state of normal growth control and thus will tend to be clonally expanded (1.53,

1.180). Because the liver comprises biochemically heterogeneous populations of hepatocytes (for example, albumin-containing cells constitute less than 1% of hepatocytes in adult mouse liver) (1.124) some small subpopulations of cells are likely to be resistant to the cytotoxic effects of a given chemical insult.

The stages of hepatocarcinogenesis are illustrated in Fig. 1.1. This traditional view has been derived primarily from studies of animal models. Farber and Sarma (1.53) defined initiation as "a change in a target tissue or organ, induced by exposure to a carcinogen, that can be promoted or selected to develop focal proliferations, one or more of which can act as sites of origin for the ultimate development of malignant neoplasia." This definition is best suited for models in which initiation, promotion, and progression can be observed and where a focal 12

Genotoxic INITIATION PROMOTION Carcinogen PROGRESSION Genetic Clonal \ Change Expansion Genetic Change * :* » w

Normal Initiated Hepalocyte Hepatocyte Proneoptastic Lesion Malignant Tumor

'Carcinogen dose Initiation Spectrum: 'Cell proliferation (PCNA) ‘Uptake and Distribution 'DNA adducts 'Apoptosls Kinetics ‘pS3 mutation ’pS3 protein overexpression 'P4S0 Induction 'Cytogenetic alterations 'Htstopathohgy of altered foci, neoplasms

Fig. 1.1: Stages of hepatocarcinogenesis, including some factors that should be considered in a holistic approach to the underlying mechanisms. 13 proliferation, such as a polyp or nodule, is produced. It also implies interaction

with the DNA of the target hepatocytes exposed to genotoxic carcinogens, with

which we are concerned in this work. However, a growing list of carcinogens are

being classified as having "epigenetic" modes of action, not requiring direct

interaction with DNA (1.104). These include agents which indirectly affect DNA or

are promoters.

Initiation has a minimum of 2 steps: 1) production of a biochemical or

molecular lesion or lesions and 2) the fixation of one or more of these lesions by a

round of cell proliferation (1.53). Thus, initiation can be more succinctly defined in

modem terms as a somatic mutation. The critical role of cell proliferation in

carcinogenesis is discussed below.

Promotion involves selective enhancement of neoplasm development from a

subpopulation of initiated cells by agents known as promoters (1.116) or by

manipulations or procedures such as partial hepatectomy. Depending on the model

system, promotion must always be defined in an operational sense with regard to

dosage and/or timing of administration. In the liver, microscopic islands or foci of hepatocytes are usually seen first, some or all of which may grow to form grossly visible nodules of altered hepatocytes (1.116).

Progression is the process by which one or more focal proliferations

(preneoplastic lesions), such as polyps or nodules, undergo a slow cellular evolution to a malignant neoplasm (1.53). This includes those changes that a malignant neoplasm undergoes as it becomes more malignant and develops greater tendencies 14 for invasion and/or metastasis. Preneoplastic lesions are usually considered

reversible if the promoting stimulus is removed. In fish, however, these lesions are

generally fewer in number than comparable lesions in rodent liver and may not

regress as do many of their rodent counterparts.

The histopathology of foci of cellular alteration (altered foci) has been

described in carcinogenicity studies with several fish species (e.g., 1.73, 1.79,

1.101). These lesions are generally similar histologically to those described in rats

(1.139, 1.162) exposed to hepatocarcinogens.

Liver Stem Cells in Carcinogenesis

The tremendous regenerative capacity of the liver in response to hepatocyte

loss is well known and widely studied (1.1). Liver regeneration is ordinarily

accomplished by entry of normally quiescent hepatocytes into the active phases of

the cell cycle. However, when normal hepatocyte regeneration is impaired, pluripotent cells called oval cells appear in histologic sections (1.105, 1.147).

Proliferation of oval cells or nonparenchymal epithelial cells has been reported in animals in response to hepatotoxicants (1.48, 1.92, 1.144), and in humans in association with hepatocellular carcinoma, necrotic lesions, and chronically damaged liver (1.144).

Evidence for liver stem cells is mounting (1.170). Whether the oval cell, or some precursor to it, is the true stem cell is still uncertain (Fig. 1.2). Most researchers now believe that a stem cell exists, but its exact role in the liver is still controversial: most support the concept of a small, dormant population of cells that 15

HEPATOCYTES TRANSITION TERMINAL CHOLANGIOCYTES/ DUCTt CELL DUCT CELL MATURE BILE DUCT , STEM CELL

OVAL CELLS Epithelial tissues from tha yok aao> derived primitive g u t PANCREATIC, HEPATOCYTES BILE INTESTINAL, other*? / \ DUCTS HEPATOMA CHOIANGIO- CARCINOMA

Fig. 1.2: Possible pathways of differentiation during chemical hepatocarcinogenesis [adapted from Sell and Pierce, 1994 (1.150)]. 16 proliferate only under conditions of extreme liver damage; others favor a "streaming

liver" hypothesis in which an active stem cell population constantly replaces mature

cells that are lost through senescence (1.150, 1.170). Nonetheless, recent

investigations have shown that the degree of differentiation of epithelial tumors

probably depends on the degree of differentiation of the target stem cells. Some

cells are induced to form, for example, well-differentiated hepatocellular

carcinomas. Others experience "maturation arrest," resulting in a whole spectrum

of neoplastic phenotypes (1.148-150). These neoplasms are believed to have

originated from liver stem cells that have the facultative capacity to differentiate into

hepatocytes, biliary epithelial cells, pancreatic acinar cells, and probably other

tissues that arise from invaginations of the yolk sac-derived primitive gut (1.65,

1.150).

DNA Adducts: Mutagenesis as a Mechanism of Carcinogenesis

An approach that integrates all of the various factors involved in chemical

exposure, uptake, and biotransformation is the comparison of levels of specific

covalent adducts of DNA in target tissues. It is generally accepted that most

chemical carcinogens act by interacting with the genetic material of the cell, in particular the DNA template (1.102). Chemical modification of DNA is the first in a series of steps that lead to mutation, cell transformation, and tumor development

(1.182). In fact, it has been suggested that any chemical that forms DNA adducts even at low levels has the potential to be mutagenic and carcinogenic (1.40).

Theodor Boveri is generally credited with the conceptualization of this "somatic 17 mutation theory" [reviewed in (1.8)]. He published his book (1.22), Zur Frage der

Entstehung Maligner Tumoren [On the Problem o f the Origin o f Malignant Tumors]

in 1914, and many of his original concepts have since been proven to be true (1.9).

That DNA adducts are critical to tumorigenesis is supported by several

observations, including the facts that 1) most carcinogens are also mutagens; 2) the

mutagenic and carcinogenic properties of most compounds depend on their in vivo

conversion to electrophilic derivatives that attack nucleophilic sites in DNA to form

adducts; 3) the degree of DNA adduct formation in a tissue can often be positively

correlated with tumorigenic response; and 4) the activation of protooncogenes has

been demonstrated through the interaction of chemical carcinogens with DNA

(1.15). DNA adduct determinations can provide crucial information on metabolic

pathways as well as chemical effects on DNA structure, transcription, synthesis, and

repair. Adduct analyses can also provide a direct test of the somatic mutation

theory. Furthermore, DNA adducts can be considered dosimeters of exposure to

chemicals in cancer risk assessments (1.15).

Analysis of DNA adducts in fish has been recently reviewed (1.108). Few

studies with small fish species have been reported. Sensitive methods for detection

of DNA adducts are essential for mechanistic studies of mutagenesis and carcinogenesis and for biomonitoring of populations at risk for environmentally caused cancer. Adducts are currently detected with such methods as 32P- postlabeling, immunoassays, HPLC with fluorescence detection, and gas chromatography/mass spectrometry (1.25). A new method in which specific DNA 18 adducts are isolated by the use of monoclonal antibodies and selective nitrocellulose

membranes, then quantitated by polymerase chain reaction, shows promise (1.84).

Alkylating agents are archetypal carcinogens, in that most other carcinogens

are active only after being metabolized to alkylating or aralkylating agents ( 1.102).

Alkylation (methylation, ethylation, etc.) at the N-7 position of guanine is a preferential site of attack for most alkylating agents, while attack at the O6 position

of guanine (less common) is most highly correlated with carcinogenesis (1.16,

1.102, 1.106, 1.167). Formation and persistence of 0 6-ethylguanine in rainbow trout exposed to DEN were reported by Fong and co-workers using HPLC with fluorescence detection (1.S8). Few studies, however, have attempted to measure specific adduct levels in small fish species exposed to small alkylating carcinogens.

This is perhaps because older methods were not sensitive enough to utilize such small amounts of tissue available from these species. 32P-postlabeling is extremely sensitive, but cannot identify exact chemical structures for specific adducts.

Immunochemical methods can be sensitive and very specific; however, they are limited to detection of adducts against which specific monoclonal antibodies are available. Additionally, these methods are not useful for detection of unknown adducts.

Gas chromatography/mass spectrometry (GC/MS) with single ion monitoring is a highly versatile method which can unequivocally identify specific adducts at extremely low detection limits (femtomole levels) (1.108). This method was used by Malins et al to detect hydroxyl radical-induced DNA adducts in liver neoplasms 19 of feral English sole from Puget Sound (1.109, 1.115). Highly accurate

measurements of adduct levels can be obtained using stable isotope-labeled analogs

of analytes as internal standards in a method known as isotope-dilution mass

spectrometry (1.44). This method has been adapted for use in small fish species in

our laboratory (see CHAPTER 4).

Cell Proliferation, p53, and Apoptosis

Recent investigations have uncovered important links between the p53 tumor

suppressor gene and the process of cell death known as apoptosis in the regulation of

cell proliferation (Fig. 1.3). This section will define these phenomena and discuss

how interrelationships between them relate to chemical carcinogenesis.

Apoptosis

The term apoptosis (derived from the Greek for "dropping off" or "falling

leaves") was first defined in the 1970’s, essentially renaming a distinctive

morphologic pattern of cell death that has long been recognized by pathologists.

Apoptosis usually involves single cells or clusters of cells which appear on HE

sections as round to oval masses of eosinophilic cytoplasm with dense chromatin

fragments which are rapidly phagocytosed by neighboring cells. Apoptodc bodies are still referred to as "Councilman bodies" in some descriptions of viral or toxic hepatitis (1.32).

The recent excitement over this morphologically drab phenomenon arose when discoveries were made that linked genetic controls to apoptosis. It was known that in normal development apoptosis was central to the destruction of superfluous 20

opto si

ZZ2ZZ

Cell Proliferation

Fig. 1.3: Inter-relationships between cell proliferation, p53, and apoptosis. 21 cells during embryogenesis, hormonal involution of endometrium, and thymic

involution. However, in the mid-1980’s, gene loci regulating apoptosis were

identified in nematodes. Inhibition of protein or RNA synthesis delayed or

prevented apoptosis in certain cell types (1.141, 1.185). In mammalian systems,

two proto-oncogene products have been shown to act as regulators in the pathway:

the myc gene product, when over-expressed in cells deprived of growth factors, induces apoptosis (1.49). Conversely, over-expression of bcl-2 makes cells resistant to apoptosis (1.171). Interleukin-1 /3-converting enzyme ("ICE"), the recently proposed vertebrate counterpart to the nematode enzyme CED-3, also promotes apoptosis (1.7).

Apoptosis is considered an active, coordinated event because ATP appears to be required for it to occur; conversely, necrotic cells exhibit declining ATP levels

(1.27). In contrast to necrosis, apoptosis induces no identifiable inflammatory response because cytoplasmic membrane integrity is maintained. Through the actions of an endogenous endonuclease, DNA is selectively digested to produce a

"ladder" on agarose gel electrophoresis. The nucleus and cytoplasm contract, chromatin becomes marginated, and the cell begins to break apart into small, membrane-bound morsels which are phagocytosed by neighboring cells (1.30,

1.184).

Apoptosis can also be induced by toxicants, particularly DNA damaging agents. Genotoxic anticancer therapies such as x-rays or alkylating agents may exert their effects through the induction of apoptosis (1.27). Based on this finding, the 22

frequency of apoptosis in the crypts of mouse bowel has been proposed as a

screening method for determining the efficacy of certain chemotherapeutic drugs

(1.87). Various cytotoxic agents have been shown to act selectively on a band of

cells along the side of the crypt. Assuming cellular hierarchies exist in tumors,

agents designed to kill tumor cells of a particular hierarchical status might be

screened by investigating whether they are able to induce apoptosis at a certain

position in the intestinal crypt. Those agents that kill the rapidly proliferating crypt

cells might have suitable palliative properties against tumors, whereas those that kill

stem cells in crypts might be curative (1.87).

p53

About half of all human cancers contain a mutation in a gene commonly

referred to as p53 or the p53 tumor suppressor gene (1.96). The normal or wild-

type p53 protein product (wt-p53) is a tumor suppressor, acting as a kind of G1

"checkpoint policeman" in the cell cycle. In 1989, workers studying the genetic

causes of colorectal cancer found a point mutation in the p53 gene. This type of

mutation is not easily detected on Southern blots and thus had not been discovered in

previous investigations. From that time, the search for associations with other types

of neoplasia began. Subsequent studies showed that only mutated p53 promotes

abnormal cell growth and the wild type suppresses tumor formation (1.35).

Soon, both transition and transversion mutations were found in tumors of the

bladder, liver, lung, larynx, breast, cervix, ovary, prostate, esophagus, stomach, pancreas, skin, thyroid gland, leukemias/lymphosarcomas, and brain. Clinical 23 studies have shown that p53 status is important in prognosis. Mutant forms of p53

protein are correlated with more aggressive tumors, greater metastatic potential, and

lower 5-year survival rates for certain neoplasms (1.35).

Mutations in p53 may arise by several possible routes. In addition to chemical carcinogen or radiation exposure, a possible mechanism involves endogenous mutagenesis. Transition mutations of C to T can be generated by spontaneous deamination of 5-methylcytosine to thymine. Alternatively, enzymatic deamination of cytosine by DNA (cytosine-5)-methyl transferase can occur when S- adenosylmethionine is in limiting concentration (1.69). Oxygen radicals enhance the rate of deamination of deoxynucleotides and thus could affect p53. Thus, this may be a reason that inhaled particulate materials that cause chronic inflammation but do not act directly on DNA can cause a high incidence of lung cancer in rats, and may be involved in the phases of tumor promotion in which oxygen radicals appear to play a role (1.69). p53 and Cell Proliferation

To determine the roles mutant p53 plays in the development of malignant neoplasia, we can begin by ascertaining how wt-p53 acts to regulate cell proliferation and thereby suppress tumors. One key to this is that the p53 protein can act as a transcription factor, binding specifically to other genes to control their expression (1.35). It was observed that several mutant forms of p53 bind DNA less avidly than does wt-p53. For example, it was reported in 1993 that the oncogene

MDM2 acts in a negative feed-back loop with p53. In other words, a sufficient level 24

of wt-p53 keeps the level of MDM2 sufficiently low until needed for cell replication.

If this loop is interrupted, MDM2 could be overexpressed. Increased levels of

MDM2 are found in about 30% of soft tissue sarcomas (1.35).

A second function of wt-p53 protein is to encode potent transcriptional

activating domains. The activating domain of wt-p53 has been localized to the first

42 amino acids of the N-terminus, a highly conserved region which contains a strong

negative charge and abundant proline residues. These features are characteristic of

the activating domains of many viral and cellular transcription factors (1.122). The

negatively charged regions are short stretches of amino acids known as acidic

domains that effect formation of amphipathic helical structures. These domains may

facilitate transcription by interacting with a general component of the initiation

complex, thereby stabilizing critical components such as TATA-binding factor

(1.126).

A third function of wt-p53 is to act as a checkpoint protein in the cell cycle.

In mammalian cells, two major checkpoints have been defined that determine the

rate of cell proliferation. One operates in early G1 to allow cells to exit the cell

cycle and enter GO (1.10). The other, called the restriction point, occurs in late G1 just before cells are committed to S-phase and potential fixation of DNA damage.

At this juncture in the cell cycle, p53 appears to arrest the cycle and prevents cells

with critical DNA damage from continuing the cycle until repair occurs.

Conformational changes may mediate these functions: p53 may be anchored in the cytoplasm under growth-stimulatory conditions, but may be translocated into the 25 nucleus under growth-restrictive conditions (1.134). Elegant experiments utilized

temperature-sensitive murine mutants with some unusual biological properties. At

32°C they act like wt-p53 and are antiproliferative, whereas at 39.5°C they act like

mutant p53 and are growth promoting. These workers demonstrated that mutant

forms of the protein are localized primarily to the cytoplasm ( 1.122).

The Cyclin Connection

Cyclins and their associated enzymes, cyclin-dependent kinases, are

intimately involved in the cell cycle. One form, called proliferating cell nuclear antigen (PCNA), is a 36 kd auxiliary protein of DNA polymerase 8 that is vital for

DNA replication. Monoclonal antibodies developed against PCNA are used in immunohistochemical staining to mark cells that have entered the active phases of the cell cycle. This has become a widely-used tool for analyzing and determining the prognosis of various neoplasms (1.57, 1.62).

Wt-p53 was shown recently to promote the expression of a newly characterized gene, W A F l/G pl. The protein product of this gene (p21) arrests cells in the middle of the replication cycle. It apparently does this by binding to cyclin- dependent kinases and inhibiting their actions. If the cell is arrested in G1 before it synthesizes new DNA and is committed to divide, this would give the cell time to repair DNA before dividing, thus preventing replication of damaged DNA. If the damage is too great, then endogenous endonuclease is triggered and the cell is committed to the apoptosis pathway (discussed below). Thus, a drug designed to 26 restore the function of W A F l/G pl could conceivably prevent certain tumors even if p53 is mutated (1.35).

Knock-out Experiments

Mutations in p53 are associated with neoplasia, but an oncogenic/mutant form is not necessary for the development of many types of tumors; instead, the mere loss of one or both p53 alleles is sufficient. This was demonstrated by the development of transgenic mice (C57BL/6 x 129) which are homozygous for null p53 allele. These mice look normally developed, but are prone to spontaneous development of numerous types of neoplasms; all are either dead or have tumors by

6 months of age (1.46, 1.97).

That the mere loss of p53 is sufficient is also supported by findings in a human disease. The Li-Fraumeni syndrome is a familial predisposition to cancer characterized by multiple early-onset malignancies. Most, but not all, of these families carry germ-line mutations in one allele of p53; in the neoplasm, expression of wild type p53 is completely abrogated and the loss of p53 appears to be sufficient to initiate tumorigenesis (1.97). One could argue that if loss of p53 were the initiating event in formation of these neoplasms, these people as well as the knock­ out mice should develop tumors much earlier in life. Therefore, it is possible that absence of proper p53 expression only predisposes the cell to further genetic damage. Maybe another hit or many more hits are required. This could explain why human patients with Li-Fraumeni syndrome as well as the p53 knock-out mice do not develop early tumors in all tissues at once. The requirement for a second 27 mutation or hit in a tissue implies that random events must occur and that many other factors may be involved.

Apoptosis, p53, and Cell Proliferation

Great interest was created when it was recently reported that over-expression of wild-type (normal/non-mutated) p53 could induce apoptosis in a range of cultured cell systems (1.151, 1.186). However, interpretation of these findings was uncertain because high levels of p53 in other types of cell systems induced growth arrest without apoptosis (1.125). Additionally, it was at first thought to be another set­ back to the p53-apoptosis connection hypothesis when it was found that the p53 knock-out mice develop normally and are fertile and immunologically competent.

This would seem to be contrary to the idea that p53 is involved in these apoptosis- associated processes. Notwithstanding, the finding that high levels of p53 are induced by DNA damage has led to the concept that p53 might be a part of a damage control pathway, rather than a critical part of the processes of normal development (1.97).

Two recent studies (1.29, 1.107) provided convincing evidence that this concept is true. It was known that excess thymocytes, including those that do not meet the major histocompatibility requirements of the body, are disposed of through apoptosis. Thus, workers at two independent labs investigated the role of p53 in thymocyte apoptosis in vitro and in vivo. Apoptosis of normal thymocytes can be accelerated by low doses of ionizing radiation or other DNA-damaging treatments, by glucocorticoids, or by application of calcium ionophore in the presence of 28

phorbol ester. The latter may mimic the natural signal responsible for the editing

out of self-reactive T-cells.

Interestingly, thymocytes from mice with absence of one or both p53 alleles

respond to glucocorticoids by undergoing apoptosis but are resistant to the induction

of apoptosis by radiation. Thymocytes of homozygous null mice are extremely

resistant to radiation, while those from heterozygous mice are only moderately

resistant. Thus, p53 appears to be essential for radiation-induced apoptosis, but not

for response to glucocorticoids. Glucocorticoids do not have a genotoxic mode of action. Along with other hormones, they act on cellular receptors. These results

show that there are at least two independent pathways to the induction of apoptosis: one involving DNA damage and mediated by p53, the other independent of p53

(1.97).

The "gene-dosage effect" of p53, i.e., homozygous thymocytes (2 hit kinetics) being more resistant to radiation than heterozygous thymocytes (1 hit kinetics) has important implications for the treatment of people with germ-line mutations in p53, such as those affected with Li-Fraumeni syndrome (1.175). This effect may also be important in tumor progression, since cells in which one p53 allele has been knocked out (somatically mutated) may have a survival advantage over normal cells when exposed to genotoxic agents, such as UV radiation to the skin. This could then give them a higher statistical chance of surviving a second hit to p53 or other locus that would otherwise kill the cell. The result is no apoptosis 29 and therefore a higher chance of sustaining a stable, heritable change in its genome.

Accordingly, a greater chance of neoplastic development results (1.97).

The exact mechanism by which p53 induces apoptosis in damaged cells is not

yet known. We do know that p53 levels rise in response to DNA strand damage.

This could specifically activate apoptosis-inducing genes downstream (1.91). Or,

p53 could interact more directly with the DNA repair machinery of the cell, possibly

increasing its fidelity, but at the same time playing its G1 "checkpoint policeman"

role and setting off apoptosis whenever sufficiently lethal damage is sustained. In

this sense, normal p53 is protecting the organism from propagation of cells that have

sustained a mutation by either helping them get repaired or, when damage is too

great, sending them into apoptosis. With loss of p53’s function of pushing damaged

cells toward apoptosis, more damaged cells may be resistant to apoptosis and higher

numbers of mutated cells may survive and replicate; hence, an increased chance of

neoplasia.

Molecular Dosimetry

Lastly, of interest for the future, researchers are discovering that the frequency and type of p53 mutations are specific for particular chemicals or classes of chemicals. For example, aflatoxin B! exposure has been correlated with G:C to

T:A transversions that lead to a serine substitution at residue 249 of p53 in hepatocellular carcinoma. Cigarette smoke exposure has been correlated with particular G:C to T:A transversions in lung, head, and neck carcinomas (1.69).

This has important implications for the molecular epidemiology of cancer risk. Not 30 only could it lead to better health recommendations and avoidance of cancer risk, but it could also lead to increased litigation involving putative causes of a person’s cancer. With the polymerase chain reaction, even p53 mutations in tumors within paraffin blocks from long dead relatives could be detected.

Methylazoxymethanol Acetate, a Model Alkylating Carcinogen

Methylazoxymethanol acetate (MAM-Ac) is an experimental carcinogen and teratogen that is derived from cycasin, an extractable component of cycad plants.

Several synonyms for MAM-Ac can be found in the literature, including methyl azoxymethyl acetate. The molecular formula is C 4HgN203; MW: 132.14.

Historical Overview

The history and toxicology of MAM-Ac are essentially linked with that of methylazoxymethanol (MAM), C 2H

They are of particular research interest because of the high incidence of human and animal neurological diseases in areas of the world where cycads are consumed.

Much of the literature focuses on two species in particular, Cycas circinalis (from

Guam) and Cycas revoluta (Japan), because of their use as a starch source by native 31 peoples in these areas (1.99, 1.158). These plants were invaluable in some areas as

famine food because they survived typhoons and drought when other food sources

were destroyed.

Although the toxicity of these compounds has only recently been scientifically

investigated, the toxic properties of these plants were known to populations that

utilized them for food long ago. According to the survey done by Dr. M.G.

Whiting in 1963, methods for detoxifying the seeds have been remarkably similar on

different continents that apparently had no intercommunications (1.158). The

dehusked, sliced or chopped seeds were repeatedly washed and dried in the sun

before being ground into a flour. Consuming improperly prepared seeds has led to

sickness and death of humans (1.158).

Nishida et al (1.133) isolated a glucoside from Cycas revoluta, determined its

structure, and named it "cycasin". It was toxic for mice and guinea pigs when given

enterically, but parenteral injections produced no symptoms. It was suggested that

its toxicity was due to the aglycone moiety, and that hydrolysis of the parent

glucoside was somehow accomplished in the digestive tract.

Once the toxicity (particularly the carcinogenicity) of this compound became

evident, the need arose for a readily available, purified, stable azoxy compound for research purposes. A series of studies by Matsumoto and colleagues (1.94, 1.95,

1.120) led ultimately to the synthesis of MAM-Ac. Bioassays were performed on the cycasin fractions. When ingested by rats, high doses produced acute liver 32 damage and death, whereas smaller doses over a longer period of time produced

hepatomas (1.120).

Distribution in the Environment

MAM-Ac is a synthetic compound available only for research purposes and

thus is probably not found in the environment. The reactive compound, MAM, is

relatively unstable and would also not be expected to be present in the environment

in significant amounts. However, as mentioned above, the parent compound,

cycasin, is widely distributed throughout many tropical and subtropical zones of the

world as a natural component of cycad plants, some of which are even grown as

ornamentals.

Metabolism

The carcinogenic, teratogenic and neurotoxic properties of MAM-Ac have prompted numerous studies on the metabolism of this intriguing compound.

Although complete metabolic pathways for activation and excretion have yet to be worked out, several reports will be mentioned which were key to our present-day understanding of the chemical.

As discussed above, both cycasin (the natural constituent from cycads) and

MAM-Ac are metabolized to the relatively unstable reactive species, MAM. The equation for the acid hydrolysis of cycasin as reported by Nishida et al (1.133) is as follows (Gl is the glucosyl residue): G/-C 2H3O2N2 + H20 = Glucose + N 2 +

HCHO + CH3OH. 33 In experiments involving cycasin, it was known that this compound was toxic

only when given orally whereas MAM was toxic by all routes. Previous

experiments had shown that there is a /3-glucosidase in the cells of the small

intestine; however, lack of uniform response to cycasin among animals and a

predilection for exclusive localization of intestinal neoplasms to the large intestine

suggested that bacterial flora might play an even greater role. When experiments

showed that germ-free rats did not break down cycasin, the role of bacterial flora

was proven (1.159, 1.160). Indeed, only strains of bacteria that possessed a /?- glucosidase were able to convert cycasin to MAM.

A perplexing exception to the requirement for intestinal hydrolysis of cycasin was reported in the late 1960’s (1.83). It was discovered that rats injected subcutaneously between the 1st and 17th day of postnatal life with cycasin died within 3-4 days; after that age, no toxicity by this route was apparent. It was first thought that the dams, in cleaning their young, were ingesting the cycasin that was excreted unchanged in the pup’s urine. Then, with metabolism in the dam’s gut,

MAM was possibly obtained by the pups through the dam’s milk. However, further experiments showed that artificially nursed rat pups still died after injection. It was later discovered that the skin of newborn and early postnatal rats contains a j 8- glucosidase capable of hydrolyzing cycasin. These studies have broader implications as to the nature and physiologic roles of transient biological enzyme systems.

Since MAM can spontaneously decompose to a reactive alkylating moiety

(1.64), it was unclear why the toxic effects of this chemical were limited to but a few organs. This led researchers to look for an enzyme system present in liver,

kidney, duodenum and colon that might activate the chemical. Studies by Zedeck

and collaborators led to the conclusion that MAM is enzymatically acted upon by

NAD+-dependent alcohol dehydrogenases which are present in the target tissues

(1.54, 1.64). Pyrazole, an inhibitor of alcohol dehydrogenase (ADH), blocked

NAD+ reduction in the presence of MAM and, when given to rats 2 hr prior to

MAM, prevented lethality; MAM was a substrate for purified horse liver alcohol dehydrogenase as well.

Feinberg and Zedeck reported the possible reactions leading to production of carbonium ions from MAM (1.54). [MC]acetate was used as a trapping agent to monitor the formation of carbonium ions. Experiments suggested that when the aldehyde of MAM is generated, it rapidly releases these reactive species. They were formed in target organs containing the NAD+-dependent dehydrogenases at much greater levels than would be expected to be found in those organs in which

MAM decomposes only spontaneously.

Using deer mouse strains both positive and negative for ADH, Fiala et al

(1.56) showed that other enzyme systems were capable of metabolizing MAM; i.e., metabolism of MAM was not exclusively due to ADH. The negative mouse strains were able to convert MAM to reactive species just as well as the positive strains.

The highly reactive alkylating moiety in these reactions was identified as the methyldiazonium ion (1.56). Its generation is believed to be the initiating event for tumor formation by MAM-Ac, cycasin, and MAM. 35 Toxic Effects in Man

Acute toxic effects given in the material safety data sheet for MAM-Ac

include, "May be fatal if inhaled, swallowed, or absorbed through skin...Extremely

destructive to tissue of the mucous membranes and upper respiratory tract, eyes and

skin. Inhalation may be fatal as a result of spasm, inflammation and edema of the

larynx and bronchi, chemical pneumonitis and pulmonary edema."

Epidemiologic evidence suggests that MAM is responsible for the high

incidence of the neurologic disorder, western Pacific parkinsonism

dementia/amyotrophic lateral sclerosis complex (P-D/ALS) in certain populations

that have utilized cycad plants for food (1.161). Clinically, the western Pacific

disease causes dementia, parkinsonism, motor neuron dysfunction and supranuclear

palsy. Alzheimer-like neurofibrillary degeneration represents the major pathological

picture. Since long periods (one to several decades) may elapse between cycad

exposure and the appearance of neurological disease in humans, Spencer et al

(1.161) proposed a "slow toxin" concept as the mechanism of action: DNA adducts

may be formed in the brain such that irreversible changes in gene expression are

caused in certain key neurons. This would lead to a slow build-up (or blockage) of

some substance in the brain such that the clinical effects are not noticed until much later. However, this concept is still very much in the theoretical stage and awaits further investigation. Although causation may never be completely proven for cycasin/MAM, heritable and infectious etiologies have been virtually ruled out, focusing on nontransmissible environmental factors. Zhang et al (1.188) provide 36 further evidence for cycasin as the causal factor in P-D/ALS in their report on the geographical distribution of this disease complex on Guam.

Nonhuman primates

Sieber et al (1.152) showed that chronic administration of cycasin and/or

MAM-Ac by the oral and intraperitoneal (ip) routes was hepatotoxic and carcinogenic in old-world monkeys. Dose ranges were relatively high. Toxic effects were highly variable in this study, perhaps because of variations in age and dose level. Several monkeys developed toxic hepatitis characterized by centrilobular necrosis or focal cellular degeneration and bile duct proliferation. A few developed pulmonary edema and/or pneumonia. Some monkeys developed single to multiple tumors including hepatocellular carcinomas, intrahepadc bile duct adenocarcinomas, renal carcinomas and adenomas, adenomatous polyps of the colon, squamous cell carcinomas of the esophagus, and adenocarcinomas of the small intestine.

Domestic animals

Cycad toxicity is a well known occurrence in parts of the world where these plants comprise a significant portion of the food source for livestock. Information was gathered by Whiting in 1954/55 on the frequency of paralytic conditions in cattle feeding on land where cycads grow, and in 1965, Hall and McGavin demonstrated that distinct demyelination of at least two of the spinal columns occurred in cattle showing paralysis after ingestion of cycads (1.99). These lesions, however, have not been seen in humans with neurological problems nor reproduced in rodents, and much remains to be investigated regarding effects in livestock. 37 Fish

Stanton reported hepatic neoplasms induced by cycasin in several small

aquarium fish in 1966 (1.165); Aoki and Matsudaira (1.5) and Hawkins et al (1.73)

reported liver tumors in medaka (Oryzias latipes) and several other small fish

species that were exposed to MAM-Ac. Harada and Hatanaka (1.67) described the

early changes in the liver of medaka caused by MAM-Ac. These consisted of PAS-

positive granules, basophilic and eosinophilic foci, and changes in mitochondria and

levels of endoplasmic reticulum by two weeks post-exposure to the toxicant. These

studies concluded that the relatively short time to tumor induction, the low

concentration of the chemical required to produce observable effects, and the

extremely low incidence of spontaneous or "background" tumors makes the medaka an excellent animal model for research on neoplasms and for level one or pre­ screening of potential carcinogens.

Law et al (1.100) used MAM-Ac to induce liver and bile duct origin tumors in the mosquitofish, Gambusia affinis (see CHAPTER 2). Borenfreund et al (1.21) used an epithelioid cell line derived from fin tissue of bluegill sunfish (BG/F) to demonstrate cell transformation by MAM-Ac. Changes included the induction of polyploidy, increased colony-forming efficiency, loss of contact inhibition, and formation of transformed foci.

Mutagenicity

MAM has been shown in a variety of test systems to be a potent mutagen.

Matsumoto and Higa (1.119) demonstrated with in vitro studies that MAM is a 38 methylating agent, most commonly forming 7-methylguanine in its reaction with

DNA or RNA. Smith in 1966 (1.156) reasoned that the genetic damage resulting

from such methylation should be reflected in an increased mutation rate. Indeed,

using histidine-requiring mutants of Salmonella typhimurium (a form of the Ames

test), he showed a significant increase above the spontaneous rate in reversion to

histidine independence.

Teas and Dyson (1.168) demonstrated a marked rise in sex-linked recessive

lethal mutations in Drosophila melanogaster after addition of either MAM or MAM-

Ac to the nutrient medium. Genetic damage to Allium (onion) seedlings (which have

/3-glucosidase activity) by cycasin was described by Teas et al (1.169) who showed

that chromosome breakage occurred at the same rate that could be produced with

200 R of gamma rays. This was referred to as a radiomimetic effect.

Carcinogenicity

Many reports of the carcinogenicity of cycasin, MAM and MAM-Ac have been referred to above, so only a few key reports will be mentioned here. Tumors have been induced by single and repeated doses of cycasin or MAM in rats, guinea pigs, mice, hamsters, non-human primates, and fish (1.99, 1.133, 1.152, 1.158,

1.165). A wealth of information exists on MAM/MAM-Ac effects in rodents.

Rodent studies provided the first evidence for the toxic properties of cycasin and

MAM, sparking tremendous interest in and numerous later studies involving this compound. 39 The first evidence for the carcinogenicity of crude cycad material was

obtained in 1962 when rats on a cycad meal diet had to be euthanized because of

ascites, anemia, and palpable abdominal tumor masses. At necropsy, tumors were

found to involve the liver, kidneys, and variably, intestine (1.98).

MAM-Ac is unique among carcinogenic substances in that a single dose is

often sufficient to produce liver tumors in most of the test animals. This concept of

the chemical as a "complete" carcinogen was strengthened with studies by Zedeck and Sternberg (1.187) who showed that a single dose produced nearly the same percentage of hepatic tumors in rats that had undergone partial hepatectomy as those that had not. Their findings suggested that initiation of tumor formation can occur in resting hepatic cells as well as in those that are actively proliferating.

MAM-Ac was found to enhance the oncogenic simian adenovirus 7 transformation of primary hamster embryo cells by 3- to 10-fold (1.28). Viral transformation increased linearly with concentration to a maximum, then decreased with further concentration increases. The chemical has been used to study the protective effects against colon cancer in rats by substances such as magnesium hydroxide (1.128) and dietary fish oil containing omega-3 fatty acids (1.42).

Objectives

A need exists for fish sentinel species for monitoring environmental changes.

Ideally, the species would be widely distributed, easily cultured in the laboratory, and would develop measurable biomarker alterations following contaminant exposure. Few native, freshwater, small fish species have been examined in field or 40 laboratory toxicology studies, particularly with respect to carcinogenesis. Further,

recent efforts to "reduce, replace, and refine" (1.4) with regard to the use of

laboratory animals in safety evaluations of industrial chemicals, pharmaceuticals, and

other new products has required a new approach. The effort to reduce and replace

the number of mammals used in carcinogenicity testing has spawned a search for the

ideal alternative test species. Refinement of status quo testing methodologies has

necessitated an integrated/holistic approach to the mechanisms of carcinogenesis, as

discussed above.

The following studies were undertaken to examine the suitability of western

mosquitofish (Gambusia afflnis) both as a sentinel species and as a laboratory test

animal for mechanistic studies in carcinogenesis. Mosquitofish are widely

distributed in warm waters and have a limited home range. In the laboratory,

mosquitofish are easily cultured using techniques developed for other small fish

species such as the guppy (Poecilia reticulata). They breed prolifically and appear

resistant to infectious diseases. However, when these investigations were begun, no

reports of neoplasia in Gambusia spp. were found in the existing literature.

These studies using MAM-Ac were designed to:

1) Establish the carcinogen sensitivity of mosquitofish.

2) Explore basic mechanisms of hepatocarcinogenesis in mosquitofish through:

a) Detailed histopathological analyses of the various stages of neoplastic

development in the liver, including cytotoxic changes, preneoplastic changes

such as foci of cellular alteration, and tumor development. 41 b) Correlation of morphological changes with quantitative

immunohistochemical analyses of cell proliferation in both juvenile and adult

fish.

c) Development and application of an immunohistochemical method for

detection of p53 overexpression in fish exposed to carcinogens.

d) Development and application of a method utilizing reverse isotope

dilution-GC/mass spectrometry for quantification and structural identification

of DNA adducts in fish tissues formed after exposure to a model alkylating

carcinogen.

3) Explore inter-relationships between cell proliferation, p53 overexpression, and apoptosis in a small fish model.

References

1.1. Alison MR (1986). Regulation of liver growth. Physiol. Rev. 66: 499-541.

1.2. Anders F, Schartl M, Bamekow A, and Anders A (1984). Xiphophorus as an in vivo model for studies on normal and defective control of oncogenes. Adv. Cancer Res. 42: 191-275.

1.3. Andreasen JK (1985). Insecticide resistance in mosquitofish of the lower Rio Grande valley of Texas-an ecological hazard? Arch. Environ. Contam. Toxicol. 14: 573-577.

1.4. Anonymous (1992). Committee shapes AVMA positions on animal welfare legislation. J. Am. Vet. Med. Assoc. 200: 430.

1.5. Aoki K and Mastsudaira H (1977). Induction of hepatic tumors in a teleost (Oryzias latipes) after treatment with methylazoxymethanol acetate: brief communication. J. Natl. Cancer Inst. 59: 1747-1749.

1.6. Bailey G, Selivonchick D, and Hendricks J (1987). Initiation, promotion, and inhibition of carcinogenesis in rainbow trout. Environ. Health. Perspec. 71: 147-153. 42 1.7. Barinaga M (1994). Cell suicide: by ICE, not fire. Science 263: 754-756.

1.8. Barrett JC (1991). Relationship between mutagenesis and carcinogenesis. In: Origins o f Human Cancer, J Brugge, T Curren, E Harlow, F McCormick (eds). Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, pp. 101- 112.

1.9. Barrett JC (1993). Mechanisms of multistep carcinogenesis and carcinogen risk assessment. Environ. Health Perspec. 100: 9-20.

1.10. Baserga R (1985). The Biology o f Cell Reproduction. Harvard University Press, Cambridge, Massachusetts.

1.11. Baumann PC (1989). PAH, metabolites, and neoplasia in feral fish populations. In: Metabolism o f Polycyclic Aromatic Hydrocarbons in the Aquatic Environment, U Varanasi (ed). CRC Press, Boca Raton, Florida.

1.12. Baumann PC, Harshbarger JC, and Hartman KJ (1990). Relationship between liver tumors and age in brown bullhead populations from two Lake Erie tributaries. Sci. Total Environ. 94: 71-87.

1.13. Baumann PC, Smith WD, and Parland WK (1987). Tumor frequencies and contaminant concentrations in brown bullheads from an industrialized river and a recreational lake. Trans. Amer. Fish. Soc. 116: 79-86.

1.14. Beardsley T (1994). A war not won. Scientific American Jan.: 130-138.

1.15. Beland FA and Poirier MC (1989). DNA adducts and carcinogenesis. In: The Pathobiology o f Neoplasia, AE Sirica (ed). Plenum Press, New York, pp. 57-80.

1.16. Beranek DT, Weis CC, and Swenson DH (1980). A comprehensive quantitative analysis of methylated and ethylated DNA using high pressure liquid chromatography. Carcinogenesis 1: 595-606.

1.17. Bishop JM (1991). Molecular themes in oncogenesis. Cell 64: 235-248.

1.18. Black DA and Howell WM (1979). The North American mosquitofish, Gambusia affinis: a unique case in sex chromosome evolution. Copeia 509-513.

1.19. Black JJ (1984). Aquatic animal neoplasia as an indicator for carcinogenic hazards to man. In: Hazard Assessment o f Chemicals: Current Developments (vol. 3), Jaxena (ed). Academic Press, New York, pp. 181-232. 43 1.20. Black JJ and Baumann PC (1991). Carcinogens and cancers in freshwater fishes. Environ. Health Perspec. 90: 27-33.

1.21. Borenfreund E, Babich H, and Martin-Alguacil N (1989). Effect of methylazoxymethanol acetate on bluegill sunfish cell cultures in vitro. Ecotoxicol. Environ. Safety 17: 297-307.

1.22. Boveri T (1914). Zur Frage der Entstehlung Maligner Tumoren, Gustave Fischer, Jena, Germany.

1.23. Brugge J, Curran T, Harlow E, and McCormick F, eds. (1991) Origins of Human Cancer, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

1.24. Bunton TE (1990). Hepatopathology of diethylnitrosamine in the medaka (Oryzias latipes) following short-term exposure. Toxicol. Pathol. 18: 313-323.

1.25. Cadet J and Weinfeld M (1993). Detecting DNA damage. Analyt. Chem. 65: 675A-682A.

1.26. Campos HH and Hubbs C (1971). Cytomoiphology of six species of gambusiine fishes. Copeia 566-569.

1.27. Carson DA and Ribeiro JM (1993). Apoptosis and disease. The Lancet 341: 1251-1254.

1.28. Casto BC, Pieczynski WJ, and DiPaolo JA (1974). Enhancement of adenovirus transformation by treatment of hamster embryo cells with diverse chemical carcinogens. Cancer Res. 34: 72-78.

1.29. Clarke AR, Purdie CA, Harrison DJ, et al. (1993). Thymocyte apoptosis induced by p53-dependent and independent pathways. Nature 362: 849-852.

1.30. Cohen JJ (1993). Overview: mechanisms of apoptosis. Immunology Today 14: 126-130.

1.31. Cormier SM, Racine RN, Smith CE, Dey WP, and Peck TH (1989). Hepatocellular carcinoma and fatty infiltration in the Atlantic tomcod, Microgadus tomcod (Walbaum). J. FishDis. 12: 105-116.

1.32. Cotran RS, Kumar V, and Robbins SL (1989). Robbins Pathologic Basis of Disease, 4th ed. W. B. Saunders Company, Philadelphia, Pennsylvania. 44 1.33. Couch JA and Courtney LA (1987). N-nitrosodiethylamine-induced hepatocarcinogenesis in estuarine sheepshead minnow (Cyprinodon variegatus) neoplasms and related lesions compared with mammalian lesions. J. Natl. Cancer Inst. 79: 297-321.

1.34. Couch JA and Harshbarger JC (1985). Effects of carcinogenic agents on aquatic animals: an environmental and experimental overview. Environ. Carcinogen. Revs. 3: 63-105.

1.35. Culotta E and Koshland DE (1993). p53 sweeps through cancer research. Science 262: 1958-1961.

1.36. Dawe CJ (1984). Dedication: an appreciation of Dr. Mearl F. Stanton. In: Use o f Small Fish in Carcinogenicity Testing, KL Hoover (ed). National Cancer Institute Monograph 65, Bethesda, Maryland, pp. 1-2.

1.37. Dawe CJ (1991). Letter to participants in the National Toxicology Program workshop on "Carcinogenesis Testing with Fish," National Institute of Environmental Health Sciences, Research Triangle Park, North Carolina, December 2-3, 1991.

1.38. Dawe CJ, Harshbarger JC, and Kondo S (1981). Phyletic Approaches to Cancer, Japan Scientific Societies Press, Tokyo, Japan.

1.39. Dawe CJ, Harshbarger JC, Wellings SR, and Stranberg JD (1995). Pathobiology of Spontaneous and Induced Neoplasms in Fishes: Comparative Characterization, Nomenclature, and Literature, Academic Press, New York, NY (in press).

1.40. de Serres FJ (1988). Banbury Center DNA adducts workshop, meeting report. Mutat. Res. 203: 55.

1.41. Dees LT (1961). The mosquitofish, Gambusia affinis. Fishery Leaflet 525, Bureau of Commercial Fisheries, Fish and Wildlife Service, U.S. Department of the Interior, September, 1961.

1.42. Deschner EE, Lytle JS, Wong G, Ruperto JF, and Newmark HL (1990). The effect of dietary omega-3 fatty acids (fish oil) on azoxymethanol-induced focal areas of dysplasia and colon tumor incidence. Cancer 66: 2350-2356.

1.43. Diamond SA, Newman MC, Mulvey M, Dixon PM, and Martinson D (1989). Allozyme genotype and time to death of mosquitofish, Gambusia affinis (Baird and Girard), during acute expsure to inorganic mercury. Environ. Toxicol. Chem. 8: 613-622. 45 1.44. Dizdaroglu M (1993). Quantitative determination of oxidative base damage in DNA by stable isotope-dilution mass sprectrometry. FEBS Letters 315: 1-6.

1.45. Doll R and Peto R (1981). The causes of cancer: quantitative estimates of avoidable risks of cancer in the United States today. J. Natl. Cancer Inst. 66: 1191-1308.

1.46. Donehower LA, Harvey M, Slagle BL, et al. (1992). Mice deficient for p53 are developmentally normal but susceptible to spontaneous tumours. Nature 356: 215-221.

1.47. Essex M, Todaro G, and zur Hausen H (eds) (1980) Viruses in Naturally Occurring Cancers, Book A. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

1.48. Eustis SL, Boorman GA, Harada T, and Popp JA (1990). Liver. In: Pathology o f the Fischer Rat, GA Boorman, SL Eustis, MR Elwell, CA Montgomery, WF MacKenzie (eds). Academic Press, Inc., San Diego, California, pp. 71-94.

1.49. Evan GI, Wyllie AH, Gilbert CS, et al. (1992). Induction of apoptosis in fibroblasts by c-myc protein. Cell 69: 119-128.

1.50. Fabacher DL, Besser JM, Schmitt CJ, Harshbarger JC, Peterman PH, and Lebo JA (1991). Contaminated sediments from tributaries of the Great Lakes: chemical characterization and carcinogenic effects in medaka ( Oryzias latipes). Arch. Environ. Contam. Toxicol. 20: 17-34.

1.51. Farber E (1984). The multistep nature of cancer development. Cancer Res. 44: 4217-4223.

1.52. Farber E, Phillips MJ, and Kaufman N (eds) (1987) Pathogenesis o f Liver Diseases, Williams & Wilkins, Baltimore, Maryland.

1.53. Farber E and Sarma DSR (1987). Hepatocarcinogenesis: a dynamic cellular perspective. Lab. Invest. 56: 4-22.

1.54. Feinberg A and Zedeck MS (1980). Production of a highly rective alkylating agent from the organospecific carcinogen methylazoxymethanol by alcohol dehydrogenase. Cancer Res. 40: 4446-4450.

1.55. Ferguson DE, Ludke JL, and Murphy GG (1966). Dynamics of endrin uptake and release by resistant and susceptible strains of mosquitofish. Trans. Amer. Fisheries Soc. 95: 335-344. 46 1.56. Fiala ES, Caswell N, Sohn OS, Felder MR, McCoy GD, and Weisburger JH (1984). Non-alcohol dehydrogenase-mediated metabolism of methylazoxymethanol in the deer mouse, Peromyscus maniculatus. Cancer Res. 44: 2885-2891.

1.57. Foley JF, Dietrich DR, Swenberg JA, and Maronpot RR (1991). Detection and evaluation of proliferating cell nuclear antigen (PCNA) in rat tissue by an improved immunohistochemical procedure. J. Histotechnology 14: 237-241.

1.58. Fong AT, Hendricks JD, Dashwood RH, Van Winkle S, and Bailey GS (1988). Formation and persistence of ethylguanine in liver DNA of rainbow trout (Salmo gairdneri) treated with diethylnitrosamine by water exposure. Food Chem. Toxicol. 26: 699.

1.59. Foumie JW, Hawkins WE, Overstreet RM, and Walker WW (1987). Exocrine pancreatic neoplasms induced by methylazoxymethanol acetate in the guppy Poecilia reticulata. J. Natl. Cancer Inst. 78: 715-725.

1.60. Foumie JW and Vogelbein WK (1994). Exocrine pancreatic neoplasms in the mummichog (Fundulus heteroclitus) from a creosote-contaminated site. Toxicol. Pathol. 22: 237-247.

1.61. Fry RJM (1990). Principles of carcinogenesis: physical. In: Cancer. Principles and Practice o f Oncology, V T DeVita, S Heilman, SA Rosenberg (eds). J. B. Lippincott Co., Philadelphia, Pennsylvania.

1.62. Garcia RL, Coltrera MD, and Gown AM (1989). Analysis of prolifrative grade using anti-PCNA/cyclin monoclonal antibodies in fixed, embedded tissues-comparison with flow cytometric analysis. Am. J. Pathol. 134: 733-739.

1.63. Goss LB and Sabourin TB (1985). Utilization of alternative species for toxicity testing: an overview. J. Appl. Toxicol. 5: 193-219.

1.64. Grab DJ, and Zedeck MS (1977). Organ-specific effects of the carcinogen methylazoxymethanol related to metabolism by nicotinamide adenine dinucleotide-dependent dehydrogenases. Cancer Res. 37: 4182-4189.

1.65. Groff JM, Hinton DE, McDowell TS, and Hedrick RP (1992). Progression and resolution of megalocytic hepatopathy with exocrine pancreatic metaplasia in a population of cultured juvenile striped bass Morone saxatilis. Dis. Aquat. Org. 13: 189-202. 47 1.66. Halver JE. Crystalline aflatoxin and other vectors of trout hepatoma. In: Halver JE, Mitchell IA (eds), Trout Hepatoma Research Conference Papers. Washington, D.C., Bureau of Sports Fisheries and Wildlife, 1967: 78-102.

1.67. Harada T and Hatanaka J (1988). Liver cell carcinomas in the medaka (Oryzias latipes) induced by methylazoxymethanol acetate. J. Comp. Path. 98: 441-452.

1.68. Harris CC (1991). Chemical and physical carcinogenesis: advances and perspectives for the 1990s. Cancer Res. 51 (Suppl.): 5023s-5044s.

1.69. Harris CC (1993). At the crossroads of molecular carcinogenesis and risk assessment. Science 262: 1980-1981.

1.70. Harris CC and Autrup HN (eds) (1983) Human Carcinogenesis, Academic Press, New York, NY.

1.71. Harshbarger JC and Clark JB (1990). Epizootiology of neoplasms in bony fish of North America. Sci. Total Environ. 94: 1-32.

1.72. Harshbarger JC, Spero PM, and Wolcott NM (1993). Neoplasms in wild fish from the marine ecosystem emphasizing environmental interactions. In: Pathobiology o f Marine and Estuarine Organisms, JA Couch, JW Foumie (eds). CRC Press, Boca Raton, Florida, pp. 157-176.

1.73. Hawkins WE, Overstreet RM, Foumie JW, and Walker WW (1985). Development of aquarium fish models for environmental carcinogenesis: tumor induction in seven species. J. Appl. Toxicol. 5: 261-264.

1.74. Hawkins WE, Overstreet RM, and Walker WW (1988). Carcinogenicity tests with small fish species. Aquatic Toxicol. 11: 113-128.

1.75. Hawkins WE, Overstreet RM, and Walker WW (1988). Small fish models for identifying carcinogens in the aqueous environment. Water Resources Bulletin 24: 941-949.

1.76. Hayes MA, Smith IR, Rushmore TH, et al. (1990). Pathogenesis of skin and liver neoplasms from industrially polluted areas in Lake Ontario. Sci. Total. Environ. 94: 105-123.

1.77. Henderson BE, Ross RK, and Pike MC (1991). Toward the primary prevention of cancer. Science 254: 1131-1138. 48 1.78. Hendricks JD (1982). Chemical carcinogenesis in fish. In: Aquatic Toxicology, U Weber (ed). Ravin Press, New York, pp. 149-211.

1.79. Hendricks JD, Meyers TR, and Shelton DW (1984). Histological progression of hepatic neoplasia in rainbow trout (Salmo gairdneri). Natl. Cancer. Inst. Monogr. 65: 321-336.

1.80. Hendricks JD, Meyers TR, Shelton DW, Casteel JL, and Bailey GS (1985). Hepatocarcinogenicity of benzo[a]pyrene to rainbow trout by dietary exposure and intraperitoneal injection. J. Natl. Cancer Inst. 74: 839-851.

1.81. Hiatt HH, Watson JD, and Winsten JA (eds) (1977) Origins o f Human Cancer. Book B: Mechanisms o f Carcinogenesis, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

1.82. Hinton DE, Lantz RC, and Hampton JA (1984). Effect of age and exposure to a carcinogen on the structure of the medaka liver: a morphometric study. In: Natl. Cancer Inst. Monogr. 65: 239-249.

1.83. Hirono I, Laquer GL, and Spatz M (1968). Tumor induction in Fischer and Osbome-Mendel rats by a single administration of cycasin. J. Natl. Cancer Inst. 40: 1003-1010.

1.84. Hochleitner K, Thomale J, Yu.Nikitin A, and Rajewsky MF (1991). Monoclonal antibody-based, selective isolation of DNA fragments containing an alkylated base to be quantified in defined gene sequences. Nucleic Acids Res. 19: 4467-4472.

1.85. Hoover KL (ed) (1984) Use o f Small Fish Species in Carcinogenicity Testing. Natl. Cancer Inst. Monogr. 65, NIH Publication No. 84-2653, National Cancer Institute, Bethesda, Maryland.

1.86. Hunter T (1991). Cooperation between oncogenes. Cell 64: 249-270.

1.87. Ijiri K (1989). Apoptosis (cell death) induced in mouse bowel by 1,2-dimethylhydrazine, methylazoxymethanol acetate, and gamma-rays. Cancer Res. 49: 6342-6346.

1.88. Irwin R (1991). National Toxicology Program workshop on "Carcinogenicity Testing with Fish," National Institute of Environmental Health Sciences, Research Triangle Park, North Carolina, Dec. 2-3, 1991. 49 1.89. Ishikawa T and Takayama S (1979). Importance of hepatic neoplasms in lower vertebrate animals as a tool in cancer research. J. Toxicol. Environ. Health 5: 537-550.

1.90. Ito N, Tatematsu M, Hasegawa R, and Tsuda H (1989). Medium-term bioassay system for detection of carcinogens and modifiers of hepatocarcinogenesis utlizing the GST-P positive liver cell focus as an endpoint marker. Toxicol. Pathol. 17: 630-641.

1.91. Kastan MB, Zhan Q, El-Deiry WS, et al. (1992). A mammalian cell-cycle checkpoint pathway utilizing p53 and GADD45 is defective in ataxia-telangiectasia. Cell 71: 587-597.

1.92. Kelly WR (1993). The liver and biliary system. In: Pathology o f Domestic Animals, KVF Jubb, PC Kennedy, N Palmer (eds). 4th ed. (vol. 2). Academic Press, Inc., San Diego, California, pp. 319-406.

1.93. Klauda RV, Peck TH, and Rice GK (1981). Accumulation of polychlorinated biphenyls in Atlantic tomcod ( Microgadus tomcod) collected from the Hudson River estuary, New York. Bull. Environ. Contam. Toxicol. 27: 829-835.

1.94. Kobayashi A and Matsumoto H (1964). Methylazoxymethanol, the aglycon of cycasin. Federation Proc. 23 (Pt.l): 1354-1355.

1.95. Kobayashi A and Matsumoto H (1965). Studies on methylazoxymethanol, the aglyconeof cycasin. Arch. Biochem. Biophys. 110: 373-380.

1.96. Koshland DE (1993). Molecule of the year. Science 262: 1953.

1.97. Lane DP (1993). A death in the life of p53. Nature 362: 786-787.

1.98. Laquer GL, Mickelsen O, Whiting MG, and Kurland LT (1963). Carcinogenic properties of nuts from Cycas circinalis L. indigenous to Guam. J. Nat. Cancer Inst. 31: 919-951.

1.99. Laquer GL and Spatz M (1968). Toxicology of cycasin. Cancer Res. 28: 2262-2267.

1.100. Law JM, Hawkins WE, Overstreet RM, and Walker WW (1994). Hepatocarcinogenesis in western mosquitofish ( Gambusia affinis) exposed to methylazoxymethanol acetate. 7. Comp. Path. 110: 117-127.

1.101. Law JM, Hawkins WE, Taylor HW, Overstreet RM, and Walker WW (1994). Cell proliferation, p53, and apoptosis in hepatocarcinogenesis of 50 western mosquitofish (Gambusia affinis) exposed to methylazoxymethanol acetate. Toxicol. Pathol, (in preparation).

1.102. Lawley PD (1984). Carcinogenesis by alkylating agents. In: Chemical Carcinogens, CE Searle (ed). American Chemical Society Monographs, Washington, DC., pp. 325-484.

1.103. Leung T-S, Naqvi SM, and Leblanc C (1983). Toxicities of two herbicides (Basagran, Diquat) and an algicide (Cutrine-Plus) to mosquitofish Gambusia affinis. Environ. Pollution (Series A) 30: 153-160.

1.104. Lijinsky W (1990). Non-genotoxic environmental carcinogens. Environ. Carcinogenesis Rev. 8: 45-87.

1.105. Lombardi B (1982). On the nature, properties and significance of oval cells. Recent Trends Chem. Carcinogen. 1: 37-56.

1.106. Loveless A (1969). Possible relevance of 0-6 alkylation of deoxyguanosine to the mutagenicity and carcinogenicity of nitrosamines and nitrosamides. Nature 223: 206-207.

1.107. Lowe SW, Schmitt EM, Smith SW, Osborne BA, and Jacks T (1993). p53 is required for radiation-induced apoptosis in mouse thymocytes. Nature 362: 847-849.

1.108. Maccubbin AE (1994). DNA adduct analysis in fish: laboratory and field studies. In: Aquatic Toxicology: Molecular, Biochemical, and Cellular Perspectives, DC Malins, GK Ostrander (eds). Lewis Publishers, Boca Raton, Florida, pp. 267-294.

1.109. Malins DC (1993). Identification of hydroxyl radical-induced lesions in DNA base structure: biomarkers with a putative link to cancer development. J. Toxicol. Environ. Health 40: 247-261.

1.110. Malins DC, Krahn MM, Brown DW, et al. (1985). Toxic chemicals in marine sediment and biota from Mukilteo, Washington: relationships with hepatic neoplasms and other hepatic lesions in English sole (Parophrys vetulus). J. Natl. Cancer Inst. 74: 487-494.

1.111. Malins DC, Krahn MM, Myers MS, et al. (1985). Toxic chemicals in sediments and biota from a creosote-polluted harbor: relationships with hepatic neoplasms and other hepatic lesions in English sole (Parophrys vetulus). Carcinogenesis 6: 1463-1469. 51 1.112. Malins DC, McCain BB, Brown DW, et al. (1984). Chemical pollutants in sediments and diseases of bottom-dwelling fish in Puget Sound, Washington. Environ Sci. Technol. 18: 705-713.

1.113. Malins DC, McCain BB, Landahl JT, et al. (1988). Neoplastic and other diseases in fish in relation to toxic chemicals: an overview. Aquatic Toxicol. 11: 43-67.

1.114. Malins DC, McCain BB, and Myers MS (1987). Field and laboratory studies of the etiology of liver neoplasms in marine fish from Puget Sound. Environ. Health Perspec. 71: 5-16.

1.115. Malins DC, Ostrander GK, Haimanot R, and Williams P (1990). A novel DNA lesion in neoplastic livers of feral fish: 2,6-diamino-4-hydroxy-5-formamidopyrimidine. Carcinogenesis 11: 1045-1047.

1.116. Maronpot RR (1991). Chemical carcinogenesis. In: Handbook o f Toxicologic Pathology, WM Haschek, CG Rousseaux (eds). Academic Press, Inc., San Diego, California pp. 91-130.

1.117. Marshall C (1991). Tumor suppressor genes. Cell 64: 313-326.

1.118. Masahito P, Ishikawa T, and Sugano H (1988). Fish tumors and their importance in cancer research. Jpn. J. Cancer Res. 79: 545-555.

1.119. Matsumoto H, and Higa HH (1966). Studies on methylazoxymethanol, the aglycon of cycasin: methylation of nucleic acids in vitro. Biochem. J. 98: 20c-22c.

1.120. Matsumoto H, aiid Strong FM (1963). The occurrence of methylazoxymethanol in Cycas circinalis L. Arch. Biochem. Biophys. 101: 299-310.

1.121. McCarthy JF, and Shugart LR (1990) Biomarkers o f Environmental Contamination, Lewis Publishers, Boca Raton, Florida.

1.122. Mercer WE (1992). Cell cycle regulation and the p53 tumor suppressor protein. Crit Rev. Eukaryotic Gene Expression 2: 251-263.

1.123. Metcalfe CD (1989). Tests for predicting carcinogenicity in fish. CRCRev. Aquatic Sci. 1: 111-129. 52 1.124. Michaelson J (1993). Cellular selection in the genesis of multicellular organization. Lab. Invest. 69: 136-151.

1.125. Michalovitz D, Halvey O, and Oren M (1990). Conditional inhibition of transformation and of cell proliferation by a temperature-sensitive mutant of p53. Cell 63: 671-680.

1.126. Mitchell PJ and Tjian R (1989). Transcriptional regulation in mammalian cells by sequence-specific DNA binding proteins. Science 245: 371-378.

1.127. Mix MC (1986). Cancerous diseases in aquatic animals and their association with environmental pollutants: a critical review of the literature. Mar. Environ Res. 20: 1-141.

1.128. Mori H, Morishita Y, Mori Y, Yoshimi N, Sugie S, and Tanaka T (1992). Effect of magnesium hydroxide on methylazoxymethanol acetate-induced epithelial proliferation in the large bowels of rats. Cancer Letters 62: 43-48.

1.129. Murchelano RA and Wolke RE (1985). Epizootic carcinoma in winter flounder, Pseudopleuronectes americanus. Science 228: 587-589.

1.130. Murchelano RA and Wolke RE (1991). Neoplasms and nonneoplastic lesions in winter flounder, Pseudopleuronectes americanus, from Boston Harbor, Massachusetts. Environ. Health. Perspec. 90: 17-26.

1.131. Myers MS, Landahl JT, Krahn MM, and McCain BB (1991). Relationships between hepatic neoplasms and related lesions and exposure to toxic chemicals in marine fish from the U.S. West Coast. Environ. Health Perspec. 90: 7-15.

1.132. Naqvi SM and Hawkins R (1988). Toxicity of selected insecticides (Thiodan, Security, Spartan, and Sevin) to mosquitofish, Gambusia affinis. Bull. Environ. Contam. Toxicol. 40: 779-784.

1.133. Nishida K, Kobayashi A, and Nagahama T (1955). Studies on cycasin, a new toxic glycoside, of Cycas revoluta thunb. Bull. Agr. Chem. Soc. Jap. 19: 77-84.

1.134. Oren M (1992). p53: the ultimate tumor suppressor gene? FASEB J. 6: 3169-3176.

1.135. Peraino C, Fry RJM, and Staffeldt E (1971). Reduction and enhancement by phenobarbital of hepatocarcinogenesis induced in the rat by 2-acetlyaminofluorene. Cancer Res. 31: 1506-1512. 53 1.136. Peraino C and Jones CA (1989). The multistage concept of carcinogenesis. In: The Pathobiology o f Neoplasia, AE Sirica (ed). Plenum Press, New York, pp. 131-148.

1.137. Pitot HC (1990). Principles of carcinogenesis: chemical. In: Cancer. Principles and Practice o f Oncology, VT DeVita, S Heilman, SA Rosenberg (eds). J.B. Lippincott Co., Philadelphia, Pennsylvania, pp. 116-135.

1.138. Plaa GL and Hewitt WR (1982). Toxicology o f the Liver, Raven Press, New York, NY.

1.139. Popp JA and Goldsworthy TL (1989). Defining foci of cellular alteration in short-term and medium-term rat liver tumor models. Toxicol. Pathol. 17: 561-568.

1.140. Powers DA (1989). Fish as model systems. Science 246: 352-358.

1.141. Raff MC (1992). Social controls on cell survival and cell death. Nature 356: 397-400.

1.142. Reines B (1986). Cancer Research on Animals: Impact and Alternatives, NAVS, Chicago, Illinois.

1.143. Rogler CE and Chisari FV (1992). Cellular and molecular mechanisms of hepatocarcinogenesis. Seminars in Liver Dis. 12: 265-278.

1.144. Sarraf C, Lalani E, Golding M, Anilkumar TV, Poulsom R, and Alison M (1994). Cell behavior in the acetylaminofluorene-treated regenerating rat liver. Am. J. Pathol. 145: 1114-1126.

1.145. Schultz ME and Schultz RJ (1982). Induction of hepatic tumors with 7,12-dimethylbenz[a]anthracene in two species of viviparous fish (Genus Poeciliopsis). Environ. Res. 27: 337-351.

1.146. Searle CE (1984). Chemical Carcinogens, ACS Monograph 182, American Chemical Society, Washington, DC.

1.147. Sell S (1990). Is there a liver stem cell? Cancer Res. 50: 3811-3815.

1.148. Sell S (1993). Cellular origin of cancer: dedifferentiation or stem cell maturation arrest? Environ. Health Perspec. 101 (Suppl 5): 15-26. 54 1.149. Sell S and Dunsford HA (1989). Evidence for the stem cell origin of hepatocellular carcinoma and cholangiocarcinoma. Am. J. Pathol. 134: 1347-1363.

1.150. Sell S and Pierce GB (1994). Maturation arrest of stem cell differentiation is a common pathway for the cellular origin of teratocarcinomas and epithelial cancers. Lab. Invest. 70: 6-22.

1.151. Shaw P, Bovey R, Tardy S, Sahli R, Sordat B, and Costa J (1992). Induction of apoptosis by wild-type p53 in a human colon tumor-derived cell line. Proc. Natl. Acad. Sci. USA 89: 4495-4499.

1.152. Sieber SM, Correa P, Dalgard DW, Mclntire KR, and Adamson RH (1980). Carcinogenicity and hepatoxicity of cycasin and its aglycone methylazoxymethanol acetate in nonhuman primates. J. Natl. Cancer Inst. 65: 177-183.

1.153. Sinnhuber RO, Hendricks JD, Wales JH, and Putnam GB (1977). Neoplasms in rainbow trout, a sensitive animal model for environmental carcinogenesis. Ann. N.Y. Acad. Sci. 298: 389-408.

1.154. Sirica AE (1989) The Pathobiology o f Neoplasia, Plenum Press, New York, NY.

1.155. Smith CE, Peck TH, Klauda RJ, and McLaren JB (1979). Hepatomas in Atlantic tomcod Microgadus tomcod (Walbaum) collected in the Hudson River estuary in New York. J. Fish Dis. 2: 313-319.

1.156. Smith DWE (1966). Mutagenicity of cycasin aglycone (methylazoxymethanol), a naturally occurring carcinogen. Science 152: 1273-1274.

1.157. Sonstegard RA (1977). Environmental carcinogenesis studies in fishes of the Great Lakes of North America. Ann. N.Y. Acad. Sci. 298: 261-269.

1.158. Spatz M (1969). Toxic and carcinogenic alkylating agents from cycads. Ann. N.Y. Acad. Sci. 163: 848-859.

1.159. Spatz M, McDaniel EG, and Laquer GL (1966). Cycasin excretion in conventional and germfree rats. Proc. Soc. Exp. Biol. Med. 121: 417-422.

1.160. Spatz M, Smith DWE, McDaniel EG, and Laquer GL (1967). Role of intestinal microorganisms in determining cycasin toxicity. Proc. Soc. Exp. Biol. Med. 124: 691-697. 55 1.161. Spencer PS, Kisby GE, and Ludolph AC (1991). Slow toxins, biologic markers, and long-latency neurodegenerative disease in the western Pacific region. Neurology 41 (suppl 2): 62-66.

1.162. Squire RA, and Levitt MH (1975). Report of a workshop on classification of specific hepatocellular lesions in rats. Cancer Res 35: 3214-3223.

1.163. Stalker MJ, Kirby GM, Kocal TE, Smith IR, and Hayes MA (1991). Loss of glutathione S-transferases in pollution-associated liver neoplasms in white suckers ( Catastomus commersoni) from Lake Ontario. Carcinogenesis 12: 2221-2226.

1.164. Stanton MF (1965). Diethylnitrosamine-induced hepatic degeneration and neoplasia in the aquarium fish, Brachydanio rerio. J. Natl. Cancer Inst. 34: 117-130.

1.165. Stanton MF (1966). Hepatic neoplasms of aquarium fish exposed to Cycas circinalis. Federation Proceedings 25: 661.

1.166. Swenberg JA, Hoel DG, and Magee PN (1991). Mechanistic and statistical insight into the large carcinogenesis bioassays on N-nitrosodiethylamine and N-nitrosodimethylamine. Cancer Res 51 (suppl.): 6409-6414.

1.167. Swenson DH, and Lawley PD (1978). Alkylation of deoxyribonucleic acid by carcinogens dimethyl sulphate, ethyl methanesulphonate, N-ethyl-N-nitrosourea, and N-methyl-N-nitrosourea. Biochem. J. 171: 575-587.

1.168. Teas HJ, and Dyson JG (1967). Mutation in Drosophila by methylazoxymethanol, the aglycone of cycasin. Proc. Soc. Exp. Biol. Med. 125: 988-990.

1.169. Teas HJ, Sax HJ, and Sax K (1965). Cycasin: radiomimetic effect. Science 149: 541-542.

1.170. Travis J (1993). The search for liver stem cells picks up. Science 259: 1829.

1.171. Vaux DL, Cory S, and Adams JM (1988). Bel-2 gene promotes haemopoietic cell survival and cooperates with c-myc to immortalize pre-B cells. Nature 335: 440-442.

1.172. Vinson SB, Boyd CE, and Ferguson DE (1963). Resistance to DDT in the mosquito fish, Gambusia affinis. Science 139: 217-218. 56 1.173. Vogelbein WK, and Foumie JW (1994). Ultrastructure of normal and neoplastic exocrine pancreas in the mummichog, Fundulus heteroclitus. Toxicol Pathol 22: 248-260.

1.174. Vogelbein WK, Foumie JW, Van Veld PA, and Huggett RJ (1990). Hepatic neoplasms in the mummichog Fundulus heteroclitus from a creosote-contaminated site. Cancer Res 50: 5978-5986.

1.175. Vogelstein B (1990). A deadly inheritance. Nature 348: 681-682.

1.176. Weis P (1972). Hepatic ultrastructure in two species of normal, fasted and gravid teleost fishes. Am. J. Anat. 133: 317-332.

1.177. Weisburger EK (1989). Chemical carcinogenesis in experimental animals and humans. In: The Pathobiology o f Neoplasia, Sirica AE, (ed). Plenum Press, New York, pp. 39-56.

1.178. Weisburger JH, and Williams GM (1982). Chemical carcinogenesis. In: Cancer Medicine, Holland JF, Frei E, (ed). 2nd ed. Lea and Febiger, Philadelphia, PA, pp. 42-95.

1.179. Williams GM (1980). The pathogenesis of rat liver cancer caused by chemical carcinogens. Biochimica Biophysica Acta 605: 167-189.

1.180. Williams GM (1987). The significance of preneoplastic liver lesions in experimental animals. Adv Vet Sc Comp Med 31: 21-44.

1.181. Wise D, Yarbrough JD, and Roush RT (1986). Chromosomal analysis of insecticide resistant and susceptible mosquitofish. J. Heredity 77: 345-348.

1.182. Wogan GN (1988). Detection of DNA damage in studies on cancer etiology and prevention. In: Methods fo r Detecting DNA Damaging Agents in Humans: Applications in Cancer Epidemiology and Prevention, Bartsch H, Hemminki K, O’Neill IK, (ed). IARC Scientific Publication No. 89, International Agency for Research on Cancer, Lyon, France, pp. 32-51.

1.183. Wooten MC, Scribner KT, and Smith MH (1988). Genetic variability and systematics of Gambusia in the Southeastern United States. Copeia 283-289.

1.184. Wyllie AH (1993). Apoptosis (The 1992 Frank Rose Memorial Lecture). Brit J Cancer 67: 205-208. 57 1.185. Wyllie AH, Morris RG, Smith AL, and Dunlop D (1984). Chromatin cleavage in apoptosis: association with condensed chromatin morphology and dependence on macromolecular synthesis. J Pathol 142: 67-77.

1.186. Yonish-Rouach E, Resnitzky D, Lotem J, Sachs L, Kimchi A, and Oren M (1991). Wild-type p53 induces apoptosis of myeloid leukaemic cells that is inhibited by interleuken-6. Nature 353: 345-347.

1.187. Zedeck M, and Sternberg SS (1977). Tumor induction in intact and regenerating liver of adult rats by a single treatment with methylazoxymethanol acetate. Chemico-Biological Interactions 17: 291-296.

1.188. Zhang Z, Anderson DW, and Mantel N (1990). Geographic patterns of paridnsonism-dementia complex on Guam. Arch. Neurol. 47: 1069-1074. CHAPTER 2. HEPATOCARCINOGENESIS IN WESTERN MOSQUITOFISH (GAMBUSIA AFFINIS) EXPOSED TO METHYLAZOXYMETHANOL ACETATE*

Summary

To assess the potential for western mosquitofish (Gambusia affinis) for the

detection of environmental carcinogens, laboratory-reared specimens were exposed

to 10 mg/L methylazoxymethanol acetate for 2 h then examined periodically for the

onset of neoplastic lesions. Approximately 33% of the exposed fish developed liver

neoplasms within 25 weeks of exposure and 52% within 40 weeks. The lesions

were mostly hepatocellular carcinomas and cholangiocarcinomas. No neoplastic

lesions were detected in other organs. The carcinogen sensitivity and the

widespread distribution of the mosquitofish suggest that this species would be useful

as a warmwater sentinel for environmental contamination.

Introduction

Epizootics of liver neoplasia in fishes from geographically isolated sites in

the United States indicate that wild fish are sensitive monitors of environmental

contaminants, especially hepatocarcinogens (2.12, 2.13). Liver neoplasms have

been reported mainly from marine or brackish water species, for example, Atlantic

tomcod (Microgadus tomcod) from the Hudson River, New York, (2.26), winter

flounder (Pleuronectes americanus) from the northeastern United States (2.9, 2.21),

and English sole (Pleuronectes vetulus) from the Puget Sound, Washington (2.22).

•Reprinted by permission of the JOURNAL OF COMPARATIVE PATHOLOGY (1994, Vol. 110, 117-127).

58 59 Except in the brown bullhead catfish ( Ictalurus nebulosus) from the Great Lakes

area, few epizootics of hepatic neoplasms have been reported from freshwater

species (2.2).

Epizootic liver neoplasia has been reported in only a single small fish

species, namely mummichog ( Fundulus heteroclitus) from brackish waters of the

Elizabeth River, Virginia (2.30). There are no reports of epizootic liver neoplasia

in small freshwater species in the United States. In the laboratory, however, small

fishes are being used with increasing frequency for testing chemical substances and

for examining mechanisms of carcinogenesis (2.18). Laboratory studies have

focused on several small fish species such as the medaka ( Oryzias latipes) (2.3,

2.14, 2.17, 2.19), platyfish/swordtail hybrids (Xiphophorus sp.) (2.1), topminnow

(Poeciliopsis sp.) (2.24), sheepshead minnow ( Cyprinodon variegatus) (2.5) and guppy (Poecilia reticulata) (2.8).

However, because many of the studies in small fish species have been made with exotic or non-native species or those having limited natural ranges, correlation with field studies has usually been indirect. There is a need for a widely distributed freshwater small fish species that could serve as an environmental sentinel and also be amenable to laboratory culture for validation of biological markers. In this study, we investigated tumor induction in the western mosquitofish ( Gambusia affinis) with methylazoxymethanol acetate (MAM-Ac), a well-studied alkylating carcinogen. The mosquitofish is a freshwater livebearer (order Atheriniformes; family Poeciliidae) that is widely distributed and amenable to laboratory culture 60 (2.6), personal observations). Studies have been performed on the acute toxicity of

various chemicals for mosquitofish (2.7, 2.20), but there are no reports of neoplastic

lesions in either wild or laboratory-reared members of this species.

Materials and Methods

Specimens used in this study were the 2- to 3-day-old progeny of laboratory-

maintained western mosquitofish which were caught in Ocean Springs, Mississippi,

U.S.A. The progeny were collected daily and removed to separate aquaria.

One hundred test fish were placed in 1-L pyrex beakers containing aged well

water (pH 9.0, dissolved oxygen 7.8 p.p.m.) maintained at 26°C in a heated

recirculating waterbath within a ventilated laboratory hood. The lengths of five

randomly selected fish ranged from 7.5 to 9.0 mm. Forty specimens served as non­

exposed controls. Sixty specimens were exposed for 2 h to methylazoxymethanol

acetate (MAM-Ac; Sigma Chemical Company, St Louis, Missouri) 10 mg/L, rinsed

and removed to clean water. Exposed and control fish were held in 40-L aquaria

maintained at 27 ± 1°C by a recirculating water bath. The fish were fed

commercial flake food (Prime Tropical Flakes-Yellow, Zeigler Brothers, Inc.,

Gardners, Pennsylvania) three times daily and brine shrimp larvae aged 24 h once

daily.

Twenty-five weeks after the exposure, 18 exposed and 20 control specimens were anesthetized with tricaine methane sulphonate (MS-222), examined macroscopically, and fixed in ice-cold Lillie’s fixative for routine processing. Most specimens were fixed, embedded and sectioned whole. However, for some large 61 females only the viscera were embedded and sectioned. The specimens were

embedded in paraffin wax, sectioned at S /im in three planes (left parasagittal, mid­

plane, and right parasagittal) and stained with hematoxylin and eosin. At 40 weeks

after exposure the survivors, 21 exposed and 7 control fish, were examined. Fish

that died naturally were not examined because of rapid autolysis. For electron

microscopy, macroscopically visible liver nodules from two MAM-Ac-exposed fish

were dissected, fixed in cold 3% glutaraldehyde in 0.1 M sodium cacodylate buffer

for 3 h, rinsed in buffer overnight, postfixed in 1 % osmium tetroxide in 0.1 M

sodium cacodylate buffer, and embedded in Epon 812. Thin sections were cut on a

Reichert Ultracut ultramicrotome, stained with uranyl acetate and lead citrate, and

examined with a JEOL 100 SX transmission electron microscope.

Results

Macroscopical Findings

No significant abnormalities were observed in the specimens examined 25

weeks after exposure. At 40 weeks after exposure, six MAM-Ac-exposed fish had

moderately swollen, pale tan livers, four of which contained several focal, raised,

whitish nodules 4-7 mm in diameter. Two nodules that were dissected and processed for electron microscopy were identified microscopically as cholangiocellular carcinomas. No other significant abnormalities were detected in the specimens. Twenty-one of the MAM-Ac-exposed specimens and 13 controls died during the grow-out period and were not examined. 62 Histopathological Findings

Table 2.1 summarizes the histopathological observations from control and exposed fish examined 25 and 40 weeks after exposure. No significant microscopical abnormalities were found in any of the control fish (Fig. 2.1). The pathological effects of MAM-Ac exposure were limited to the liver. Neoplastic lesions in the liver included foci of cellular alteration (altered foci), hepatocellular adenomas, hepatocellular carcinomas, and cholangiocellular carcinomas. Several other microscopical changes were frequently found in exposed fish with and without neoplastic changes. Some specimens contained areas of hepatocellular swelling and scattered foci of hepatocellular necrosis characterized by nuclear pyknosis, karyorrhexis, or karyolysis. The cytoplasm of hepatocytes often contained one to several variously sized eosinophilic droplets or inclusions, some of which resembled apoptotic bodies. In some specimens, there were scattered focal accumulations of mononuclear inflammatory cells and occasional small aggregates of pigmented macrophages in the liver. No parasitic lesions were found.

The foci of cellular alteration were small areas of basophilic cells characterized by mild loss of cytoplasmic storage products (presumably both glycogen and lipid), enlarged nuclei and prominent nucleoli. These foci exhibited no significant thickening of hepatic cords, compression of surrounding tissue, or encapsulation. Rather, at the edges of the lesion, the altered hepatocytes tended to blend with the surrounding hepatocytes while maintaining the parenchymal organization. 63

Table 2.1: Histopathological examination of western mosquitofish exposed to MAM-Ac.

Weeks Number (and percentage) of fish with______between exposure and any hepatocellular cholangiocellular altered examination tumor tumor tumor foci

25 6/18 (33) 4/18 (22) 2/18 (11) 3/18 (17)

40* 11/21 (52) 5/21 (24) 7/21 (33) 1/21 (5)

At 25 and 40 weeks, 20 and 7 control fish examined, respectively, had no lesions. Numerator, number of fish with lesion; denominator, number examined. *One fish had both types of tumor. 64

Fig. 2.1: Liver from control mosquitofish. The hepatocytes are arranged in branching, two-cell thick plates or laminae separated by sinusoids. A branch of the hepatic portal vein is surrounded by acini of exocrine pancreas (arrowheads). HE. x 110. 65 Hepatocellular adenomas frequently compressed surrounding tissue and often abutted hepatic vessels (Fig. 2.2). Adenomas were well demarcated from the surrounding hepatocytes with disruption of the parenchymal pattern. Cytologically, adenomas consisted of enlarged, basophilic cells with enlarged nuclei, prominent nucleoli, and a decrease in normal storage products. Hepatic cords showed variable, mild thickening.

Hepatocellular carcinomas exhibited a trabecular pattern and were characterized by cords of basophilic hepatocytes several cells thick which were discontinuous with the adjacent hepatic tissue (Fig. 2.3). Carcinomas usually compressed the surrounding normal tissue. The neoplastic cells were enlarged with an increase in nuclear-to-cytoplasmic ratio and a decrease in cytoplasmic storage products (Fig. 2.4). Frequent mitotic figures and occasional binucleate cells were noted. No evidence of metastasis or local vascular invasion was found.

Cholangiocellular neoplasms appeared to be aggressive lesions and showed varying degrees of anaplasia (Fig. 2.5). They often occupied more than 80% of a liver section. The neoplastic cells formed gland-like structures of varying size and occasionally larger cystic spaces lined by neoplastic biliary epithelial cells. Dense connective tissue stroma was often interposed between the glands, along with scattered packets of neoplastic cells embedded in the stroma. The neoplastic cells were primarily cuboidal and well differentiated. However, some sections contained anaplastic, strap-like cells organized in whorling patterns within a fibrous connective tissue stroma (Fig. 2.6). Adjacent to several of the cholangiocellular carcinomas 66

mmBw

Fig. 2.2: Hepatocellular adenoma in a mosquitofish exposed to MAM-Ac for 2 hr and killed 25 weeks later. HE. x 380. 67

Fig. 2.3: Hepatocellular carcinoma in a mosquitofish exposed to MAM-Ac for 2 hr and killed 25 weeks later. Trabecular cords of deeply basophilic, neoplastic hepatocytes occupy more than two-thirds of the liver parenchyma. HE x 120.

Fig. 2,4: Detail of liver from Fig. 2.3. The large neoplastic cells have an increased nuclear-to-cytoplasmic ratio and a decreased amount of cytoplasmic storage products. HE x 480. 68

Fig. 2.5: Cholangiocellular carcinoma in the liver of a mosquitofish exposed to MAM-Ac for 2 hr and killed 25 weeks later. Note gland-like structures lined by neoplastic biliary epithelial cells. HE x 110.

Fig. 2.6: Cholangiocellular carcinoma in the liver of a mosquitofish exposed to MAM-Ac and killed 40 weeks later. Note numerous anaplastic, strap- like cells in a dense fibrous connective tissue stroma. HE x 250. 69 were multilocular, pale, eosinophilic areas infiltrated with small numbers of histiocytic cells; these lesions resembled spongiosis hepatis as described in other small fish species.

Electron Microscopy

Ultrastructural examination of two of the liver nodules confirmed the epithelial origin of the cholangiocarcinomas. The cells were irregular in shape, arranged in tubules, and rested on a basal lamina (Fig. 2.7). Desmosomes occurred together with abundant rough endoplasmic reticulum which was condensed into concentric layers in some cells. Many of the nuclei had bizarre shapes and highly folded nuclear membranes. Occasionally, a dense, rounded structure consisting of concentric electron-dense laminations was seen (Fig. 2.8). This structure resembled apoptotic-like bodies seen with light microscopy.

Discussion

The neoplasms induced in mosquitofish by MAM-Ac were histologically similar to those described in carcinogenicity studies with other fish species (2.11,

2.14, 2.16). However, the only foci of cellular alteration observed in the present study resembled the basophilic type described in rats (2.23, 2.29) and rainbow trout

(2.16) exposed to hepatocarcinogens. These lesions are considered to be pre­ neoplastic foci and probably do not regress as do many of the altered foci seen in rodent studies. Although both the mosquitofish and the guppy (Poecilia reticulata) belong to the Family Poeciliidae, the response of the two species to MAM-Ac differs. In addition to having hepatic neoplasms, approximately 9% of guppies 70

Fig. 2.7: Electron micrograph of cholangiocellular carcinoma in a mosquitofish exposed to MAM-Ac for 2 hr and killed 40 weeks later. Uranyl acetate and lead citrate stain (UL). x 1500.

Fig. 2.8: Highly dense, roundish structure resembling an apoptotic body (a) found within cholangiocellular carcinoma in a mosquitofish exposed to MAM-Ac for 2 hr and killed 40 weeks later. UL x 2500. 71 exposed to 2, 4, or 10 mg/L MAM-Ac for 2 h developed exocrine pancreatic

adenomas, acinar cell carcinomas, or adenocarcinomas (2.8). In contrast, no

pancreatic neoplasms or neoplasms of any tissue other than liver were found in

mosquitofish in the present study. The relatively high incidence of hepatic lesions in

the mosquitofish, nonetheless, was similar to that found in the guppy (2.15).

A lesion associated with cholangiocellular carcinomas in mosquitofish

resembled one diagnosed as spongiosis hepatis in medaka exposed to MAM-Ac

(2.17) and in sheepshead minnows exposed to N-nitrosodiethylamine (2.4).

Spongiosis hepatis was suggested by Couch (2.4) to be a pre-neoplastic lesion that

develops after carcinogen exposure. Although these lesions were seen in MAM-Ac-

exposed mosquitofish, there was no evidence of progression to clearly neoplastic

forms. Whether spongiosis hepatis is a pre-neoplastic, neoplastic, or toxic lesion in

the mosquitofish and other small fish species requires further investigation.

For this study, MAM-Ac was chosen because it is a well-characterized,

known carcinogen that induces neoplasms in non-human primates, rodents, and fish.

Metabolism of MAM-Ac in some species is thought to be primarily through the action of NAD+-dependent alcohol dehydrogenases (2.10). Target organ specificity for the carcinogenic effects of MAM-Ac, therefore, appears to depend on the presence of these enzymes. Prolonged administration of MAM-Ac to Old-World monkeys resulted in hepatocellular carcinomas, intrahepatic bile duct adenocarcinomas, renal carcinomas and adenomas, squamous cell carcinomas of the oesophagus, adenocarcinomas of the small intestine, and adenomatous polyps of the 72 colon (2.25). In man, epidemiological evidence has linked methylazoxymethanol, the naturally occurring metabolite of cycad flour used traditionally by certain native

Pacific islanders, to the western Pacific parkinsonism dementia/amyotrophic lateral sclerosis complex (2.28). In rodents, neoplasms are induced in the liver, kidneys, and intestinal tract (2.31), and various teratogenic effects such as microencephaly are observed in offspring (2.27). In a study in which seven species of small fish were exposed to MAM-Ac, hepatic neoplasms developed in the medaka, guppy, sheepshead minnow, Gulf killifish, inland silverside, rivulus, and fathead minnow

(2.14). Additionally, neoplasms occurred in the retina, mesenchymal tissues, exocrine pancreas, and kidney of medaka and guppies, and in ocular and nervous tissues of medaka and sheepshead minnows. Although no non-hepatic neoplasms were observed in mosquitofish in the present investigation, the high incidence of cholangiocellular neoplasms suggests that some of the ultimate carcinogenic metabolites are concentrated in the bile.

This study showed that mosquitofish have a rather short latency period for neoplasm development after exposure to a chemical carcinogen. Although further studies are needed to determine whether mosquitofish respond to carcinogens representing different mechanistic classes, the species appears to be useful for environmental biomarker studies. With regard to hepatic carcinogen sensitivity, the response of mosquitofish was equal to or greater than that of the seven aquarium fish species (see above) exposed to MAM-Ac under similar conditions (2.14).

Furthermore, unlike exotic, non-native species, mosquitofish occupy a wide range of 73 freshwater habitats including drainage ditches, sewerage outfalls, and treatment

ponds (2.6). Because mosquitofish have a limited home range, they might be more

reliable sentinels for a particular site than would be a migratory species. Because no

lesions were detected in control mosquitofish, interference from background tumors,

infectious conditions, or other confounding conditions appears to be minimal. These

attributes, along with ease of laboratory culture, would facilitate the direct validation

of field studies by enabling substances to be tested under both laboratory and field

conditions.

References

2.1. Anders F, Schartl M, Bamekow A, and Anders A (1984). Xiphophorus as an in vivo model for studies on normal and defective control of oncogenes. Adv. Cancer Res. 42: 191-275.

2.2. Black JJ and Baumann PC (1991). Carcinogens and cancers in freshwater fishes. Environ. Health Perspec. 90: 27-33.

2.3. Bunton TE (1990). Hepatopathology of diethylnitrosamine in the medaka (Oryzias latipes) following short-term exposure. Toxicol. Pathol. 18: 313-323.

2.4. Couch JA (1991). Spongiosis hepatis: chemical induction, pathogenesis, and possible neoplastic fate in a teleost fish model. Toxicol. Pathol. 19: 237-250.

2.5. Couch JA and Courtney LA (1987). N-nitrosodiethylamine-induced hepatocarcinogenesis in estuarine sheepshead minnow ( Cyprinodon variegatus) neoplasms and related lesions compared with mammalian lesions. J. Natl. Cancer Inst. 79: 297-321.

2.6. Dees LT (1961). The mosquitofish, Gambusia affinis. Fishery Leaflet 525 (September, 1961), Bureau of Commercial Fisheries, Fish and Wildlife Service, U.S. Department of the Interior.

2.7. Diamond SA, Newman MC, Mulvey M, Dixon PM, and Martinson D (1989). Allozyme genotype and time to death of mosquitofish, Gambusia 74

affinis (Baird and Girard), during acute expsure to inorganic mercury. Environ. Toxicol. Chem. 8: 613-622.

2.8. Foumie JW, Hawkins WE, Overstreet RM, and Walker WW (1987). Exocrine pancreatic neoplasms induced by methylazoxymethanol acetate in the guppy Poecilia reticulata. J. Natl. Cancer Inst. 78: 715-725.

2.9. Gardner GR, Pruell RJ, and Folmar LC (1989). A comparison of neoplastic and non-neoplastic disorders in winter flounder (Pseudopleuronectes americanus) from eight areas in New England. Marine Environ. Res. 28: 393-397.

2.10. Grab DJ and Zedeck MS (1977). Organ-specific effects of the carcinogen methylazoxymethanol related to metabolism by nicotinamide adenine dinucleotide-dependent dehydrogenases. Cancer Res. 37: 4182-4189.

2.11. Harada T and Hatanaka J (1988). Liver cell carcinomas in the medaka (Oryzias latipes) induced by methylazoxymethanol acetate. J. Comp. Path. 98: 441-452.

2.12. Harshbarger JC and Clark JB (1990). Epizootiology of neoplasms in bony fish of North America. Sci. Total Environ. 94: 1-32.

2.13. Harshbarger JC, Spero PM, and Wolcott NM (1993). Neoplasms in wild fish from the marine ecosystem emphasizing environmental interactions. In: Pathobiology o f Marine and Estuarine Organisms, JA Couch, JW Foumie (eds). CRC Press, Boca Raton, Florida, pp. 157-176.

2.14. Hawkins WE, Overstreet RM, Foumie JW, and Walker WW (1985). Development of aquarium fish models for environmental carcinogenesis: tumor induction in seven species. J. Appl. Toxicol. 5: 261-264.

2.15. Hawkins WE, Overstreet RM, and Walker WW (1988). Carcinogenicity tests with small fish species. Aquatic Toxicol. 11: 113-128.

2.16. Hendricks JD, Meyers TR, and Shelton DW (1984). Histological progression of hepatic neoplasia in rainbow trout (Salmo gairdneri). Natl. Cancer Inst. Monogr. 65: 321-336.

2.17. Hinton DE, Lantz RC, and Hampton JA (1984). Effect of age and exposure to a carcinogen on the structure of the medaka liver: a morphometric study. Natl. Cancer Inst. Monogr. 65: 239-249. 75 2.18. Hoover KL (1984). Use o f Small Fish Species in Carcinogenicity Testing. Natl. Cancer Inst. Monogr. 65, NIH Publication No. 84-2653, National Cancer Institute, Bethesda, Maryland.

2.19. Ishikawa T and Takayama S (1979). Importance of hepatic neoplasms in lower vertebrate animals as a tool in cancer research. J. Toxicol. Environ. Health 5: 537-550.

2.20. Leung T-S, Naqvi SM, and Leblanc C (1983). Toxicides of two herbicides (Basagran, Diquat) and an algicide (Cutrine-Plus) to mosquitofish Gambusia affinis. Environ. Pollution (Series A) 30: 153-160.

2.21. Murchelano RA and Wolke RE (1991). Neoplasms and nonneoplastic lesions in winter flounder, Pseudopleuronectes americanus, from Boston Harbor, Massachusetts. Environ. Health Perspec. 90: 17-26.

2.22. Myers MS, Landahl JT, Krahn MM, and McCain BB (1991). Relationships between hepatic neoplasms and related lesions and exposure to toxic chemicals in marine fish from the U.S. West Coast. Environ. Health Perspec. 90: 7-15.

2.23. Popp JA and Goldsworthy TL (1989). Defining foci of cellular alteration in short-term and medium-term rat liver tumor models. Toxicol. Pathol. 17: 561-568.

2.24. Schultz ME and Schultz RJ (1982). Induction of hepatic tumors with 7,12-dimethylbenz[a]anthracene in two species of viviparous fish (genus Poeciliopsis). Environ. Res. 27: 337-351.

2.25. Sieber SM, Correa P, Dalgard DW, Mclntire KR, and Adamson RH (1980). Carcinogenicity and hepatoxicity of cycasin and its aglycone methylazoxymethanol acetate in nonhuman primates. J. Natl. Cancer Inst. 65: 177-183.

2.26. Smith CE, Peck TH, Klauda RJ, and McLaren JB (1979). Hepatomas in Atlantic tomcod Microgadus tomcod (Walbaum) collected in the Hudson River estuary in New York. J. Fish Dis. 2: 313-319.

2.27. Spatz M (1969). Toxic and carcinogenic alkylating agents from cycads. Ann. N.Y. Acad. Sci. 163: 848-859.

2.28. Spencer PS, Kisby GE, and Ludolph AC (1991). Slow toxins, biologic markers, and long-latency neurodegenerative disease in the western Pacific region. Neurology 41 (suppl 2): 62-66. 76 2.29. Squire RA and Levitt MH (1975). Report of a workshop on classification of specific hepatocellular lesions in rats. Cancer Res. 35: 3214-3223.

2.30. Vogelbein WK, Foumie JW, Van Veld PA, and Huggett RJ (1990). Hepatic neoplasms in the mummichog Fundulus heteroclitus from a creosote-contaminated site. Cancer Res. 50: 5978-5986.

2.31. Zedeck M and Sternberg SS (1977). Tumor induction in intact and regenerating liver of adult rats by a single treatment with methylazoxymethanol acetate. Chemico-Biological Interactions 17: 291-296. CHAPTER 3. CELL PROLIFERATION, P53, AND APOPTOSIS IN HEPATOCARCINOGENESIS OF WESTERN MOSQUITOFISH (GAMBUSIA AFFJNIS) EXPOSED TO METHYLAZOXYMETHANOL ACETATE

Summary

To investigate relationships among the induction of neoplastic lesions,

overexpression of the p53 tumor suppressor gene, cell proliferation, and apoptosis,

approximately 600 specimens of western mosquitofish (Gambusia affinis) were

exposed ambiently for 2 hr to methylazoxymethanol acetate (MAM-Ac). Two

groups of fry, 15-22 days old, were exposed to 1 and 10 mg/L MAM-Ac and one

group of adult fish, 13-14 months old, to 10 mg/L only. Specimens were examined

at 8 sampling times from 24 hr to 41 weeks post-exposure (PE). Early histologic

changes included cytotoxicity, apoptosis, and stem cell proliferation. Cell

proliferation determined by proliferating cell nuclear antigen (PCNA)

immunohistochemistry was significantly increased in fry and adults exposed to 10

mg/L group possibly because of regenerative hyperplasia. At later time points, increased PCNA labeling in exposed specimens resulted from preneoplastic and neoplastic cell proliferation. Although the 1.0 mg/L fry exposure was carcinogenic, labeling was not significantly increased over controls. Hepatocellular carcinomas, cholangiocellular carcinomas, and a type of spindle cell neoplasm probably originating from a liver stem cell were observed as early as 8 weeks PE. Total tumor incidence generally was time and dose dependent but apparently independent of age at exposure. The overexpressed product of p53 gene was detected in exposed fish immunohistochemically as early as 48 h post-exposure but its presence did not

77 78 predict eventual tumor formation under the staining conditions employed. These findings suggest that hepatocarcinogenesis in mosquitofish is a multistage process with underlying mechanisms not unlike their mammalian counterparts, making this species potentially useful in carcinogenicity testing and as a warmwater sentinel for environmental contamination.

Introduction

Recent advances in cellular and molecular pathology have provided new insights into the mechanisms of neoplasia. Chemical carcinogens may cause mutational disturbances in the normal events of the cell cycle and the fine-tuning involved in its control. Consequently, this may impart selective growth advantages to certain clones of cells over normal, neighboring cells. Links between the p53 tumor suppressor gene and apoptotic cell death in the regulation of cell proliferation have been recently established (3.14, 3.39, 3.45, 3.68, 3.70). Few studies, however, have investigated relationships among histopathological changes, cell proliferation, p53 overexpression, and apoptosis.

There is increasing interest in the use of fish in cancer research: 1) to examine basic mechanisms from a comparative standpoint 2) as models for carcinogenicity testing; and 3) as environmental sentinels (3.13, 3.27, 3.33, 3.49,

3.51). As test animals, several small fish species have been shown to have good carcinogen sensitivity with extremely low background tumor incidences. Most species are easily maintained in the laboratory and have substantial cost savings over their mammalian counterparts (3.27, 3.33). A number of laboratory studies have 79 focused on small fish hepatocarcinogenesis, particularly involving the medaka

(Oryzias latipes) (3.9, 3.26, 3.32, 3.34), platyfish/swordtail hybrids ( Xiphophorus

spp.) (3.3), topminnow ( Poeciliopsis spp.) (3.61), sheepshead minnow ( Cyprinodon

variegatus) (3.12), and guppy ( Poecilia reticulata) (3.23). There is a need,

nonetheless, for a native, small freshwater fish that can be used to directly correlate

laboratory studies with environmental field studies. We have shown in a previous

study that the western mosquitofish ( Gambusia affinis), native to a large portion of

North America and introduced worldwide as a mosquito control agent, is sensitive to

chemical induction of liver neoplasia and is easily cultured in the laboratory (3.42).

The endpoint of most carcinogenesis studies with small fish has been

confined to tumor formation. Comparatively few studies have focused on early

histopathological changes, such as cytotoxicity (3.41). Proliferating cell nuclear

antigen (PCNA) has been used to study cell proliferation in neoplastic and

nonneoplastic lesions of several aquatic species, including medaka, flounder, and

shrimp (3.57). And, although p53 has been identified and characterized in several

fish species (3.1, 3.11), correlation with histopathological changes has been lacking.

A holistic view of neoplasia is essential if we are to understand the process in alternative animal models and effectively correlate it with human carcinogenesis.

The major target organ reported in most field and laboratory studies with fish has been the liver. Thus, in the present series of experiments, we have attempted to correlate histopathological changes with several new slide-based tools of molecular 80 biology to encompass both early and later steps in the process of

hepatocarcinogenesis.

Materials and Methods

Animals and Experimental Design

Western mosquitofish {Gambusia affinis) used in this study were progeny of

laboratory-maintained mosquitofish caught near Ocean Springs, Mississippi, USA.

The studies incorporated two experimental groups: young fish (fry) 15 to 22 days

old at the time of exposure and adult fish 13 to 14 months old.

Three hundred adults and four hundred twenty fry were randomized into 1-L

pyrex beakers containing aged well water (pH 8.8, dissolved oxygen 7.0 ppm)

maintained at 26°C in a heated recirculating waterbath and held within a vented

laboratory hood. Ten randomly selected adults ranged from 22 to 33 mm in length.

Ten randomly selected fry were 10 to 17 mm in length. One hundred fifty adult

specimens and 140 fry served as controls. One hundred fifty adults and 140 fry

were exposed for 2 hr to 10.0 mg/L methylazoxymethanol acetate (MAM-Ac; Sigma

Chemical Company, St. Louis, MO, USA) dissolved in the ambient water. One hundred forty fry were exposed to 1.0 mg/L MAM-Ac. The specimens were then rinsed and removed to clean water for the duration of the grow-out period. Exposed and control specimens were held in 120 L aquaria fitted with hinged, removable covers and maintained at 27±1°C by a recirculating water bath. The grow-out room was maintained on a 16-hr light/8-hr dark cycle with fluorescent lighting.

Fish were fed commercial flake food (Prime Tropical Flakes-Yellow, Zeigler 81 Brothers, Inc., Gardners, PA, USA) three times daily and 24-hour-old brine shrimp

larvae once daily. Water quality monitoring and tank maintenance were performed

twice weekly.

Fish from each grow-out tank were sampled at 24 h, 48 h, 1 wk, 2 wk, 8

wk, 16 wk, 24 wk, and 41 wk post exposure (Table 3.1). Fish that died naturally

were not examined because of rapid autolysis at this ambient temperature.

Specimens were anesthetized with tricaine methane sulphonate (MS-222), examined

grossly, and fixed in buffered zinc formalin (Z-FIX, ANATECH, LTD., Battle

Creek, MI, USA) for 15 to 17 h. This fixative was chosen because the ionized zinc

purportedly preserves antigenic sites in the tissues by inhibiting crosslinking by the

formaldehyde thereby optimizing immunohistochemical staining. Specimens were

decalcified in a formic acid/sodium citrate solution (TBD-2, Shandon Inc.,

Pittsburgh, PA, USA) which facilitated sectioning of whole fish. Scales were gently

removed from most specimens, a ventral slit was made from anal fin to gills, and

the fish were fixed, embedded, and sectioned whole. With some of the larger

females only the viscera were embedded and sectioned. The specimens were

embedded in paraffin, sectioned at 5 pm in two planes (left parasagittal and mid­ plane) and stained with hematoxylin and eosin. Remaining tissues in the paraffin blocks were reserved for immunohistochemical procedures.

For electron microscopy, tissues from selected exposed and control fish were dissected, fixed in cold 3% glutaraldehyde in 0.1 M sodium cacodylate buffer, rinsed in buffer, postfixed in 1 % osmium tetroxide in 0.1 M sodium cacodylate 82

Table 3.1: Experimental Design

Test Group Control Exposed Control Low Dose High Adult Adult Fry Fry Dose Fry Age at exposure 13-14 13-14 15-22 15-22 15-22 months months days days days MAM-Ac exposure 0 10.0 0 1.0 10.0 concentration (mg/L) Number o f control or 150 150 140 140 140 exposed specimens Specimen length 22-33 22-33 10-17 10-17 10-17 ranges (mm) SAMPLING TIMES: (Number of specimens sampled at each sampling time) 24 hours 10 10 10 10 10 48 hours 10 10 10 10 10 1 week 10 10 10 10 10 2 weeks 10 10 10 10 10 8 weeks 20 20 20 20 20 16 weeks 20 20 20 20 20 24 weeks 20 20 20 20 20 41 weeks 20 11* 20 29 24

'Specimens that died prior to sampling times were not examined 83 buffer, and embedded in Spurr’s resin. One micron thick, toluidine blue-stained

sections were prepared and examined light microscopically. Thin sections were cut

on a Reichert Om U3 ultramicrotome, stained with uranyl acetate and lead citrate,

and examined with a Zeiss 109 transmission electron microscope.

PCNA, p53, and Cytokeratin Immunohistochemistry

Ten paraffin blocks from each control and exposed group were selected from

the 48 h, 2 wk, 8 wk, and 24 wk groups for both proliferating cell nuclear antigen

(PCNA) and p53 immunohistochemistry, and from the 40 wk group for p53

immunohistochemistry only. However, with some of the 24 and 40 week groups,

only 4 specimens were chosen in order to focus on the tumor-bearing specimens.

With groups that contained more than 10 individuals, blocks were chosen using a

table of random digits. Tissue sections were cut at 5 fim , mounted on silanized

glass slides (ProbeOn Plus, Fisher Scientific, Pittsburgh, PA, USA), air-dried

overnight at 37°C, deparaffinized in xylene, and rehydrated through a graded series

of ethyl alcohols to phosphate buffered saline (PBS) with 0.05% Tween 20 (Sigma

Chemical Co., St. Louis, MO, USA). All slides were stained using a FisherBiotech

MicroProbe Staining System (Fisher Scientific, Pittsburgh, PA, USA) for consistency of reagent application.

PCNA immunohistochemistry was performed using a procedure adapted for aquatic animals (3.57), with some modifications. After deparaffinization and rehydration, slides were immersed in 3% hydrogen peroxide with 0.1% sodium azide for 15 min to quench endogenous peroxidase, then rinsed in distilled water. 84 Antigen retrieval processing was employed to unmask PCNA epitopes. For this, slides were immersed in 1 % zinc sulfate in distilled water then heated in a microwave oven for two, 2 min periods, with a 1 min cooling period in between.

After rinsing, slides were fitted into a MicroProbe slide holder assembly and a protein block consisting of one part 1 % powdered milk in distilled water and one part 1% bovine serum albumin (Sigma Chemical Co., St. Louis, MO, USA) in PBS was applied for 20 min. Slides were drained, and monoclonal anti-PCNA antibody

(PC10, mouse anti-human PCNA, Dako Corp., Carpinteria, CA, USA) diluted

1:100 was applied for one hour at room temperature. Negative control sections received mouse IgG (Vector Laboratories, Inc., Burlingame, CA, USA) to approximate the protein concentration in the primary antibody. Slides were rinsed 6 times in PBS and then biotinylated anti-mouse secondary antibody was applied

(StrAviGen Super Sensitive kit, BioGenex, San Ramon, CA, USA) for 30 min. The

PBS rinse was repeated, then horseradish peroxidase-conjugated streptavidin

(BioGenex) was applied for 30 min. After further rinsing, diaminobenzidine (Sigma

Chemical Co.) with hydrogen peroxide was applied as the chromogen. Tissue sections were lightly counterstained with hematoxylin, dehydrated, and coverslipped with a resinous mounting medium.

Adjacent tissue sections were used for immunohistochemical detection of p53 protein using the same basic procedure described above. A monoclonal mouse anti­ human p53 protein antibody (DO-7, Dako Corp., Carpinteria, CA, USA) was used for the primary antibody at a dilution of 1:50. This antibody reacts with an epitope 85 on the N-terminus of the p53 protein which is common to both the wild type and mutant type of the p53 protein.

Selected sections were also stained for cytokeratins using anti-cytokeratin monoclonal antibody AE1/AE3 at the recommended 1:200 dilution (BioGenex

Laboratories, San Ramon, California, USA). The same basic protocol as above was used except that in place of the microwave antigen retrieval step, enzyme digestion was performed for 30 min at 37°C using 0.1 % trypsin.

Cell Proliferation Quantification

A labeling index (LI) was determined for each PCNA stained specimen.

Since many proliferative foci were composed not of typical hepatocytes but of small, spindloid cells determined to be of epithelial origin by cytokeratin staining, all epithelial-like cells were counted. In livers with obvious neoplastic lesions or obviously highly proliferative foci, the counts were concentrated on those foci.

Otherwise, random fields were counted at high power. Criteria for PCNA labeling followed those of Foley et al (3.21, 3.22). Liver cells with distinct, light golden brown to dark brown nuclear staining were scored as positive late G,- to S-phase cells. The LI was calculated by dividing the number of positively stained liver cells by the total number of hepatocytes counted, which was at least 1000 hepatocytes, then multiplying by 100 for expression on a percentage basis.

Statistical Analysis

The estimate of LI has a binomial variance p(l-p)/c, where p is the probability of label and c is the number of cells scored (53). The variance of the 86 estimate of p decreases as c increases, giving a more precise estimate of p. Morris

(S3) plotted the power of the statistical test, which takes into account the variables

involved in detecting a treatment effect with cell proliferation data. In most cases, it

was determined that optimum reliability could be achieved by counting at least 1000

hepatocytes in sections from 10 or more animals per experimental group. It was

further determined that, under these conditions, nearly equivalent power calculations

could be obtained using computer simulations between the conventional Student’s t-

test and the more rigorous rank sum test. In this study, therefore, Student’s Mest

was used to test for significant differences in LI between control and exposed

animals at each time point as well as between selected exposed groups which had

inferred relationships through graphical analysis.

Results: Pathology

Gross Findings

Two exposed fry and one control adult specimen in the study had mild

scoliosis, a common skeletal abnormality in some laboratory reared small fish.

These fish, however, fed and behaved normally.

Few remarkable gross abnormalities were detected in exposed specimens. In

the adult test group at 16 weeks post-exposure (PE), one female had a pale, whitish,

slightly swollen liver and another female had a dark brown to black liver at 24

weeks PE. A second 24 week PE female had multifocal pinpoint raised whitish areas over the liver surface and a third female at 24 weeks had multifocal clear, 1-2 mm in diameter gelatinous areas in the liver. In each of those three specimens, 87 hepatic spindle cell neoplasms were identified microscopically. Furthermore, in the

specimen with the darkly pigmented liver the neoplasm had invaded the body cavity

and infiltrated both the ovaries and kidneys. At 41 weeks PE, two high dose female

fry had pale, whitish, moderately swollen livers that were later determined to be

hepatocellular carcinomas.

Histopathology

Several non-neoplastic lesions unrelated to exposure were found in the

specimens, with and without neoplastic changes. Gills of six specimens harbored

cysts identified as Epitheliocystis sp. (Fig. 3.1), a rickettsial-like organism that

typically does not affect the host. No other parasites were seen. Small numbers of

lymphocytes and pigmented macrophages, sometimes occurring in small aggregates,

were present in liver, spleen, and kidney of control and exposed fish. Although

scattered, small numbers of mononuclear leukocytes are considered to be a normal

background finding, exposed fish with cytotoxic changes described below had

significantly increased numbers of lymphocytes, pigmented macrophage aggregates,

and course eosinophilic granulocytes in the liver.

Table 3.2 summarizes neoplasm incidences. A single neoplastic lesion was

seen in control specimens, a poorly differentiated cholangiocellular carcinoma with

adjacent spindle cell proliferation in an adult female examined at 41 weeks. The

incidence of neoplasia in all controls was therefore approximately 0.4%. The liver of one other female from that same group contained an eosinophilic focus. Both the 88

■< m j I ■iV ■ w*'*a **■ 4 .WtfvVA ‘i,A j? V*V. ■** *

Fig. 3.1: Epitheliocystis sp. (arrows) in the gill of a mosquitofish. HE x 860. Table 3.2. Histopathological examination of western mosquitofish exposed to MAM-Ac*

Legend of Tumor Types: HC Hepatocellular Carcinoma HA Hepatocellular Adenoma C Cholanglocellular S Spindle Cell P Pancreatic

Time Post Exposure, Weeks 8 16 24 41

Adults:

Control (CA) Number of Fish 20 20 20 20 % With Tumors 0 0 0 5 Tumor Incidence, by Tumor Type (%) HC 0 0 0 0 HA 0 0 0 0 C 0 0 0 5 S 0 0 0 0 P 0 0 0 0

Exposed (EA) Number of Fish 20 20 20 10 % With Tumors 25 75 65 100 Tumor Incidence, by Tumor Type (%) HC 0 35 25 70 HA 0 30 25 10 C 0 5 5 20 S 25 20 25 40 P 0 0 5 0

Fry:

Control (CF) Number of Fish 20 20 20 20 % With Tumors 0 0 0 0 Tumor Incidence, by Tumor Type (%) HC 0 0 0 0 HA 0 0 0 0 C 0 0 0 0 S 0 0 0 0 P 0 0 0 0

Low Dose (EF) Number of Fish 20 20 20 29 % With Tumors 0 0 0 10 Tumor Incidence, by Tumor Type (%) HC 0 0 0 0 HA 0 0 0 0 C 0 0 0 10 S 0 0 0 0 P 0 0 0 0

High Dose (EE) Number of Fish 20 20 20 24 % With Tumors 15 35 65 96 Tumor Incidence, by Tumor Type (%) HC 0 5 35 58 HA 0 0 10 4 C 0 20 50 75 S 15 10 5 4 P 0 5 0 8 * Note: In several specimen groups, some fish had more than one type of tumor. 90 cholangiocellular carcinoma and the eosinophilic focus were immunohistochemically

positive for overexpression of p53 protein.

The major pathological effects of MAM-Ac exposure were limited to the

liver in both the fry and adult test groups. As early as 8 weeks PE, preneoplastic

and neoplastic lesions observed in the liver included foci of cellular alteration

(altered foci), hepatocellular adenomas, hepatocellular carcinomas, cholangiocellular

carcinomas, exocrine pancreatic carcinomas, and hepatic spindle cell neoplasms.

Cytotoxicity and Apoptosis

No remarkable cytotoxic changes were noted in exposed fish sampled at 24

or 48 hr PE or in control fish at any sampling time. At 1 week PE, hepatocytes of

exposed adults and high dose fry showed single cell necrosis, increased vacuolation

and cytoplasmic inclusions. The inclusions were of various sizes, round, hyalinized

and eosinophilic and stained periodic acid-Schiff (PAS) positive (Fig. 3.2). Some

areas of perivascular spindle cell proliferation were noted. A moderate infiltration

of mixed inflammatory cells was especially prominent around necrotic cells.

By 2 weeks PE, cytotoxic changes increased in severity and consisted of

widespread macrovesicular vacuolation, cell swelling, anisokaryosis,

hepatocytomegaly, and occasional areas of coagulative or lytic necrosis (Fig. 3.3).

The vacuolation was PAS negative suggesting fatty deposition. Hepatocytomegaly

was characterized by marked enlargement of cell diameter either with or without apparent nuclear enlargement. Some giant binucleate and multinucleate hepatocytes 91

Fig. 3.2: Diffuse cytotoxicity and numerous cytoplasmic inclusions in hepatocytes from a mosquitofish 1 week after exposure to MAM-Ac. HE x 540. 92

Fig. 3.3: Cytotoxic changes in hepatocytes from a mosquitofish 2 weeks after exposure to MAM-Ac. A. Necrosis, cell swelling, and infiltration of numerous macrophages, many of which contain phagocytosed cell fragments. HE x 540. B. Abundant, PAS-positive, intracytoplasmic material within macrophages and greatly swollen hepatocytes. PAS x 540. 93 were noted. Nuclear inclusions or pseudoinclusions (i.e., cytoplasmic invaginations)

were not a prominent feature.

In exocrine pancreatic tissue located along large veins within the liver, there

was a moderate to severe reduction of zymogen granules. Many specimens had a

lacy meshwork of spindle cells with small, round to oval, hyperchromatic nuclei.

Spindle cells often surrounded islands of megalocytic hepatocytes which were

probably regenerative nodules. Numerous hepatocytes contained hyalinized,

eosinophilic, PAS-positive, cytoplasmic inclusions, many of which were probable

apoptotic bodies. Inflammatory infiltrates were severe at this sampling period.

Neoplastic Lesions

Figs. 3.4 and 3.5 illustrate the distribution of neoplasms by type in exposed

adult fish and high dose fry at 8, 16, 24, and 41 weeks PE. By 8 weeks PE, 25%

of the exposed adults had spindle cell tumors, 15% of the high dose fry had spindle

cell tumors, and low dose fry had no tumors. Spindle cell tumors were characterized by elongated spindle-shaped cells as described above. The neoplastic cells were arranged in small groups closely applied to vessels or surrounding hepatic regenerative nodules. In some specimens, the spindle cells were arranged in haphazard flowing or whorling patterns (Fig. 3.6) or in solid sheets. Spindle cells in these neoplasms were shown to be positive for cytokeratins by immunohistochemistry. Occasional ductular profiles lined by cuboidal epithelial cells were seen within these spindle cell areas. Spindle cell tumors were often 140 B Pancreatic ■ Spindle Cell 120 - □ Cholangiocellular 6 i B Hepatocellular Adenoma g* 100 £ • B Hepatocellular Carcinoma

• ° 80 cTu s ■ q 60 *5 s u 40 - o

h 20- ■

n ----- 16 24 41 Time Post Exposure (Weeks)

Fig. 3.4: Tumor incidence by type in exposed adult Gambusia affinis. Fig. 3.5: Tumor incidence by type in type by incidence Tumor 3.5: Fig. Tumor Incidence, by Type (% ) mePs psr Weeks) (W xposure E Post e im T 1 2 41 24 16 8 Gambusia affinis Gambusia B B Pancreatic B □ Cholangiocellular Cholangiocellular □ B Hepatocellular Carcinoma Hepatocellular B Adenoma Hepatocellular B high dose fry. dose high Spindle Cell Spindle 95 96

Fig. 3.6: Spindle cell tumor with a whorling pattern in the liver of an adult mosquitofish 8 weeks after exposure to MAM-Ac. HE x 340. 97 infiltrated with small to moderate numbers of macrophages, course eosinophilic

granulocytes, or lymphocytes. In some of the 8 week high dose specimens,

cytotoxicity remained high. Islands of megalocytic hepatocytes surrounded by

spindle cells were infiltrated with mixed inflammatory cells (Fig. 3.7). There was

loss of the trabecular pattern with disruption of hepatic parenchymal patterns and

loss of intrahepatic pancreatic tissue. Numerous hepatocytes contained apoptotic

bodies (Fig. 3.8). Necrodc areas were infiltrated with scattered aggregations of plump macrophages with abundant eosinophilic cytoplasm.

At 16 weeks PE, there was more heterogeneity in types of neoplasms seen in the high dose groups (see Figs. 3.4 and 3.5). In addition, several foci of cellular alteration were noted in all exposed groups. These typically consisted of small areas of cells which exhibited either basophilic (basophilic foci) or eosinophilic

(eosinophilic foci) tinctorial change. These cells had mild loss of cytoplasmic storage products (presumably glycogen and lipid) and prominent nucleoli. However, no significant thickening of hepatic cords, compression of surrounding tissue, or encapsulation was present. Altered foci tended to blend with the surrounding hepatocytes, while maintaining the parenchymal organization (Fig. 3.9).

Several of the exposed adults had hepatocellular adenomas at 16 weeks PE.

These neoplasms were usually well demarcated and compressed the surrounding normal hepatocytes, often abutting hepatic vessels (Fig. 3.10). Adenomas consisted of large, basophilic cells with enlarged nuclei, prominent nucleoli, and a decrease in 98

Fig. 3.7: Megalocytosis and spindle cell proliferation in liver of an adult mosquitofish 8 weeks after exposure to MAM-Ac. HE x 540. 99

Fig. 3.8: Apoptosis (arrows) and megalocytosis in liver of an adult mosquitofish 8 weeks after exposure to MAM-Ac. HE x 540. Fig. 3.9: Eosinophilic focus in the liver of an adult mosquitofish 8 weeks after exposure to MAM-Ac. HE x 135. Fig. 3.10: Hepatocellular adenoma in the liver of an adult mosquitofish 16 weeks after exposure to MAM-Ac. HE x 55. 102 normal cytoplasmic storage products. There was variable, mild thickening of

hepatic cords.

Hepatocellular carcinomas occurred in both high dose MAM-Ac groups at 16

wk PE and their incidence increased in the 24 and 41 wk samples. Most carcinomas

had a trabecular pattern with basophilic hepatocytes arranged in cords several cells

thick which were discontinuous with the surrounding tissue (Fig. 3.11). Carcinomas

compressed the surrounding tissue and had infiltrative leading edges in which neoplastic cells projected into the surrounding parenchyma. The neoplastic cells had an increased nuclear-to-cytoplasmic ratio, decreased cytoplasmic storage products, frequent mitotic figures and occasional binucleate or large, multinucleate cells.

Necrosis and apoptosis were common in the adjacent hepatic parenchyma along with moderate, mixed inflammatory infiltrates as described above.

Cholangiocellular carcinomas were highly variable in morphologic presentation, ranging from groups of well-differentiated ductular profiles to highly invasive, anaplastic lesions with predominantly spindle cell morphology (Fig. 3.12).

Some of these tumors were solid with only a few scattered ductular profiles scattered within an abundant connective tissue stroma. The neoplastic cells within ductular profiles were usually cuboidal except in anaplastic lesions in which they were flattened and spindle-shaped with oval to cigar-shaped nuclei. Combinations of cholangiocellular carcinomas with spindle cell tumors were not uncommon, making discrimination between the two difficult. 103

Fig. 3.11: Hepatocellular carcinoma in the liver of a mosquitofish fry 24 weeks after exposure to MAM-Ac. A. Well-differentiated hepatocellular carcinoma. HE x 340. B. Heterogeneous, anaplastic hepatocellular carcinoma. Note large "ground glass" area (arrows) and severely cystic area (arrowheads). HE x 135. 104

Fig. 3.12: Highly anaplastic spindle cell tumor in an adult mosquitofish 24 weeks after exposure to MAM-Ac. A. Liver. Note ductular profiles at one edge of the neoplasm (arrows). HE x 340. B. Spindle cell invasion into the caudal body cavity, with infiltration of ovaries. HE x 340. 105 Exocrine pancreatic neoplasia was found in the liver of one high MAM-Ac

fry at 16 weeks PE, one exposed adult at 24 weeks PE, and one high MAM-Ac

ffyat 41 weeks PE (Fig. 3.13). These three lesions were large, solid masses

morphologically similar to the adenomas or well-differentiated acinar cell

carcinomas described in guppies exposed to MAM-Ac by Foumie et al (3.23).

Overall, there was an increased amount of exocrine pancreatic tissue in the liver.

The acinar cells generally retained their normal shape and complement of zymogen

granules. However, a focal area of atypical, flattened cells with decreased zymogen

granules was present in one fish. These neoplasms had few mitotic figures.

At 41 weeks PE, several control and exposed adult female specimens had

gonadal changes that were probably related to aging. Ovaries were moderately to

severely atrophied with a decrease in the size, number and staining affinity of

oocytes, a proliferation of connective tissue stroma. Whereas many of the fish exposed as fry had well developed embryos at this stage, none of the adults did.

One of the low dose, MAM-Ac-exposed, male fry had a focal proliferation of spindle cells which surrounded spermatogonia and spermatocytes (Fig. 3.14). These were shown to be highly proliferative by proliferating cell nuclear antigen (PCNA) immunohistochemistry. The origin of these cells was uncertain but was possibly from interstitial (Leydig) cells.

Electron Microscopy (EM)

Samples from livers of three exposed fish were examined ultrastructurally.

Grossly, the liver of one specimen was dark brown to black while the other two 106

Fig. 3.13: Well differentiated exocrine pancreatic carcinoma in the liver of an adult mosquitofish 24 weeks after exposure to MAM-Ac. HE x 340. 107

Fig. 3.14: Spindle cell tumor (possible intersitial cell tumor) in the testis of a mosquitofish fry 41 weeks after exposure to MAM-Ac. HE x 340. 108 specimens had multifocal, pale whitish nodules. Spindle cell neoplasms were

diagnosed by light microscopy in each of the three specimens. With EM the

hepatocytes were highly irregular in shape and size (Figs. 3.15 and 3.16). Most had

multiple cytoplasmic surface projections or processes and abundant rough

endoplasmic reticulum. Mitochondria were increased in number and had a high

concentration of cristae. Many nuclei had atypical shapes or highly folded nuclear

membranes. Some cells contained large lysozomes filled with an amorphous,

electron dense material. Spindle cells had elongated, fusiform nuclei and a relative

paucity of cytoplasmic organelles, including endoplasmic reticulum (Fig. 3.16).

Results: PCNA Labeling

Adult Fish

Examples of PCNA staining are shown in Fig. 3.17. Mean liver cell

labeling indices (Li’s) for control and exposed adult mosquitofish determined by

PCNA immunohistochemistry at each of four post-exposure (PE) time points are

plotted against time in Fig. 3.18. The Li’s for control adults was rather constant

over time, ranging from 2% to 4.3%. LI for exposed adults was not significantly

different than control LI at 48 hr PE. However, a major increase in LI to 18.7%

occurred by 2 weeks PE. This increased LI was sustained at 8 weeks and reached

23.3% at 24 weeks PE. LI in exposed adults for each of these latter three time points was significantly greater than LI for the corresponding control fish (p <

0.05); however, they were not significantly different from each other. With onset of tumorigenesis, complications ensue with regard to determination of LI. Note: focal 109

Fig. 3.15: Electron micrographs of liver cells from an adult mosquitofish 24 weeks after exposure to MAM-Ac. Note highly folded nuclear membranes (N), abundant rough endoplasmic reticulum (R), and irregular cytoplasmic surface projections (arrows). A. UL x 6060. B. ULx 11,400. i ’.

Fig. 3.16: Electron micrographs of liver cells from an adult mosquitofish 24 weeks after exposure to MAM-Ac. A. Abundant rough endoplamic reticulum and numerous electron dense cytoplasmic bodies. UL x 6060. B. Two spindle cells from fish diagnosed with "spindle cell tumor." Note elongated, fusiform nuclei and paucity of cytoplasmic organelles. UL x 9000. Fig. 3.17 Typical PCNA staining in the liver of mosquitofish after exposure to MAM-Ac. A. Note positive staining of intestinal crypt epithelial cell nuclei (arrows). This serves as an internal, positive control for proliferating cells, B. Nuclear staining (DAB chromogen) showing "leading edge" of proliferating liver tumor cells. Labeling Index (%) 30 20 25 15 10 0 5 0 Fig. 3.18: Mean liver cell labeling indices (Li’s) for control and exposed adult exposed and control for (Li’s) indices labeling cell liver Mean 3.18: Fig. mosquitofish determined by PCNA immunohistochemistry. PCNA by determined mosquitofish 5 Time Post Exposure (Weeks) Exposure Post Time 10 Exposed Adults Exposed Control Adults Control 15 20 25 113 areas of neoplastic change are expected to have a high concentration of labeled cells.

Variations in plane of section will expose different cross-sectional areas of tumors

and introduce variability in comparisons between one exposed specimen and another

or between exposed and control specimens.

Low Dose Fry

Liver cell Li’s for fry exposed to 1 ppm MAM-Ac are shown in Fig. 3.19.

In the low dose groups, no statistical difference in. mean Li’s was found between

exposed and control animals at any time point. LI at 24 weeks PE was increased

over controls (10.1% to 6.2%, respectively) but not significantly so.

High Dose Fry

Liver cell Li’s for fry exposed to 10 ppm MAM-Ac are presented in Fig.

3.20. LI for control fry decreased from 18.3% at 48 h to a low of 1.8% at 8

weeks. The LI was not significantly different from the control LI at 48 h PE. At 2

weeks PE, the exposed fry LI was higher but not statistically different from the

control LI. By 8 weeks PE, the LI of exposed fish had declined compared to the 2

week exposed LI but was significantly increased over the corresponding control LI.

At 24 weeks PE, there was a significant increase in LI over the controls to 27.2%.

Results: p53 Immunohistochemistry

Table 3.3 summarizes the incidence of positive anti-p53 staining for each experimental group. Except for a cholangiocellular neoplasm and an altered focus seen in control fish at 41 weeks, immunoreactivity to anti-p53 was confined to the liver of MAM-Ac-exposed individuals. In neoplasm-bearing specimens, 8 weeks PE Labeling Index (%) 20.0 16.0 - 18.0 10.0 12.0 14.0 4.0 0.0 6.0 2.0 8.0

-1 0 Fig. 3.19: Mean liver cell labeling indices (Li’s) for control and exposed low exposed and control for (Li’s) indices labeling cell liver Mean 3.19: Fig. 5 dose mosquitofish fry determined by PCNA immunohistochemistry. PCNA by determined fry mosquitofish dose Time Post Exposure (Weeks) Exposure Post Time 10 Exposed Fry—Low Dose Exposed Control Fry Control IS 20 25 H-* 4^ Labeling Index (%) 25.0 35.0 20.0 30.0 10.0 15.0 0.0 5.0 0 Fig. 3.20: Mean liver cell labeling indices (Li’s) for control and exposed high exposed and control for (Li’s) indices labeling cell liver Mean 3.20: Fig. 5 dose mosquitofish fry determined by PCNA immunohistochemistry. PCNA by determined fry mosquitofish dose Time Post Exposure (Weeks) Exposure Post Time 10 — Exposed Fry—High Dose Fry—High Exposed — 525 15 Control Fry Control 20 116

Table 3.3: Incidence of anti-p53 immunopositivity in western mosquitofish exposed to 1 or 10 ppm MAM-Ac.

Time between (percentage). exposure and examination Exposed adults Low dose fry High dose fry

48 hours 1/10 (10) 0/10 (0) 2/10 (20)

2 weeks 6/10 (60) 2/10 (20) 3/10 (30)

8 weeks 5/10 (50) 4/10 (40) 0/10 (0)

24 weeks 3/10 (30) 2/10 (20) 6/10 (60)

41 weeks 4/10 (40) 1/7 (14) 4/10 (40)

"‘Except for one apparently spontaneous neoplasm and one altered focus in the livers of two control adult specimens at 41 weeks, no positive anti-p53 staining was detected in controls. 117 and later, positive p53 staining was limited mainly to preneoplastic (altered foci) and

neoplastic cell populations as determined by adjacent HE sections. Normal liver

cells adjacent to neoplasms were negative, as were non-neoplastic proliferating cell

populations such as intestinal crypt cells. In nontumor-bearing specimens, staining

was usually limited to a few focal liver cells. In the 48 h and 2 week PE groups,

p53 staining occurred in hepatocytes and occasionally in small, oval to stellate to

spindle-shaped cells believed to be liver stem cells. In all cases, the pattern of

staining was exclusively cytoplasmic and varied from light golden brown to dark

brown, diffuse, granular staining (Fig. 3.21). Staining intensity was generally

greater with increasing time post-exposure.

The earliest detection of p53 over-expression was in 1 of 10 exposed adult

fish and in 2 of 10 high dose fry at 48 h PE. At 2 weeks PE, staining was observed in individuals from all three exposed groups, the highest being in 6 of 10 exposed adults. At 8 weeks PE, positive staining was seen in 5 of 10 exposed adults, 4 of which had marked cytotoxicity but no tumors. Four of 10 low dose fry at this time point were positive, but no tumors were diagnosed in the group. No p53 staining was detected in high dose fry at 8 weeks PE, even though 3 were diagnosed with tumors.

At 24 weeks PE, 3 of 10 exposed adults had tumors, all of which were positive for p53 over-expression. An eosinophilic focus in the low dose fry group was positive. Six of 10 high dose fry were positive, 3 of which had tumors. Fig. 3.21: Typical p53 staining in the liver of mosquitofish after exposure to MAM-Ac. Note that staining is limited to the cytoplasm. 119 In total, of 25 exposed adult specimens diagnosed with liver tumors, 8

stained positively for p53 over-expression whereas 7 of 18 high dose fry with tumors

were positive for p53. None of the tumors found in low dose fry were positive for

p53.

Discussion

Cell proliferation, p53 overexpression, and apoptosis are basic processes in

normal as well as neoplastic cells. Normal p53 appears to control cell proliferation

in several ways, thus establishing a link between mutated p53 and carcinogenesis.

Normal or wild-type p53 protein can act as a transcription factor, binding

specifically to other genes such as MDM2 to control their expression until needed for

cell proliferation (3.14). p53 suppresses cell proliferation by selective down-

regulation of PCNA expression (3.50). p53 also acts as a checkpoint protein in the

cell cycle by arresting cell replication in the G1 phase before cells are committed to

the S-phase and potential fixation of DNA damage. When DNA is damaged

substantially, an endogenous endonuclease is triggered and the cell is committed to apoptosis (3.56). Thus, cells in which a p53 allele has been mutated may have a survival advantage over neighboring cells when exposed to genotoxic agents (3.39).

The mosquitofish is indeed sensitive to the carcinogenic effects of MAM-Ac and the lesions appear to progress through a series of moiphologically detectable stages from cytotoxicity to foci of cellular alteration to neoplasia (3.42). Alkylating agents are archetypal carcinogens, in that most other carcinogens are active only after being metabolized to alkylating or aralkylating agents (3.43). MAM-Ac, the 120 aglycone of cycasin, was chosen for this study because its carcinogenic effects in

non-human primates, rodents, and several small fish species are relatively well-

characterized. It is a potent alkylating agent that kills dividing cells by methylating

purine bases of nucleic acids (3.36). Several target organs are involved. Old World

monkeys developed hepatocellular carcinomas, intrahepatic bile duct

adenocarcinomas, renal carcinomas and adenomas, squamous cell carcinomas of the

esophagus, adenocarcinomas of the small intestine, and adenomatous polyps of the

colon in response to prolonged MAM-Ac administration (3.63). Rodents developed

liver, kidney, and intestinal tract neoplasia (3.40, 3.78), along with central nervous

system disorders and teratogenic effects such as microencephaly (3.66). Neoplasia

has been reported in several small fish species exposed to MAM-Ac, including the

medaka, guppy, sheepshead minnow, Gulf killifish, inland silverside, rivulus, and

fathead minnow (3.4, 3.13, 3.23, 3.25, 3.26, 3.51). Although most developed

hepatic neoplasia, species specific differences were observed. In medaka, the

compound also caused ocular medulloepitheliomas (3.26). MAM-Ac-induced carcinogenesis is dependent on both exposure concentration and time, and is accompanied by cell proliferation (3.5-6).

Early neoplastic changes in the liver of MAM-Ac-exposed mosquitofish reflect the carcinogenic mechanisms of that compound. A single, brief exposure to

MAM-Ac initiates tumorigeneis suggesting that irrepairable cellular changes necessary for neoplastic initiation occurred during or soon after exposure (3.80). In adult liver, one round of cell proliferation is necessary for "fixation" of the initiating 121 event (3.19, 3.48). Thus, interaction of an electrophilic methyl group from MAM-

Ac with DNA alone would not produce an initiated cell. The severe cytotoxicity at

10 ppm MAM-Ac provided the stimulus for regenerative proliferation of surviving

hepatocytes, biliary epithelial cells, liver stem cells, and others. Conversely, in young growing animals the liver normally contains numerous hepatocytes which are in the active phases of the cell cycle and thus the need for cell proliferation is already fulfilled (3.19). This may explain how low MAM-Ac doses produced neoplasms in the young mosquitofish and is supported by cell proliferation data from this study. The decrease in the PCNA L i’s in both control and exposed fry indicates a higher level of actively dividing liver cells early in the study, followed by an age- related "minimum" which was reached 8 weeks later. Labeling indices for control fry plateaued as their hepatocytes matured while that for high dose fry increased because of regenerative hyperplasia and neoplastic proliferation. Statistically significant increases in Li’s of exposed versus controls were not seen, however, in low dose fry, probably because 1 ppm MAM-Ac did not cause remarkable cytotoxicity. Concomitantly, there was a sharply reduced incidence of tumor formation in this group consistent with a non-linear dose-response.

A complicating factor in MAM-Ac-induced carcinogenesis following a single pulse of the chemical is its dual proliferative and anti-proliferative effects. MAM-

Ac caused nucleolar structural abnormalities as early as IS min following intravenous administration in rats (3.79-80), suggesting that the compound may interfere with cell replication. Opposing this are the cytotoxic effects of MAM-Ac 122 which stimulate cell proliferation in the liver through regenerative hyperplasia. No

significant cytotoxic changes were noted in the mosquitofish of the present study

until after 48 hr PE. In mammalian systems, cell proliferation is no longer effective

in initiation after 96 hr and most lesions are repaired by this time (3.35, 3.77). In

mosquitofish, therefore, the formation of critical mutagenic lesions such as DNA

adducts, along with critical degrees of necrosis or apoptosis followed by regenerative

proliferation, would likely occur between 48 and 96 hr after exposure.

Alternatively, DNA repair mechanisms may be comparatively slower in these fish.

This awaits further investigation.

Hepatocytomegaly characterized by polyploid hepatocytes with anisokaryosis

or multinucleated hepatocytes is a histopathologic biomarker of toxicant exposure in

fish from field and laboratory studies (3.24, 3.30). For example, megalocytic

hepatosis is a frequently encountered lesion in the liver of English sole (Parophrys

vetulus) from contaminated sites in the Puget Sound, Washington (3.54, 3.55). It

has also been induced in trout {Oncorhynchus mykiss) by exposure to pyrrolizidine

alkaloids (3.29) and in medaka (Oryzias latipes) by diethylnitrosamine exposure

(3.31). In the present study, hepatocytomegaly occurred both as hepatocellular

hypertrophy attributed to organelle hyperplasia and as megalocytosis with concurrent

cellular and nuclear enlargement. Megalocytosis is stimulated by necrogenic agents

such as dimethylnitrosamine, pyrrolizidine alkaloids and aflatoxins, consistent with inhibition at the late S or G2 phase of the cell cycle (3.38). Interference with RNA transcription by alkylation of DNA (3.37) may delay expression of critical cell 123 cycle-associated proteins. Cell cycle delay may also help explain the lag in increased cell proliferation index in this study as determined by PCNA immunohistochemistry. At the later time points, the increased PCNA labeling index appears to reflect the proliferative capacity and behavior of the preneoplastic and neoplastic cells. Thus, the increased cell proliferation affected the spectrum of stages in carcinogenesis by allowing less time for DNA repair, permitting gene amplification, and enabling clonal expansion of initiated cells. Because the liver comprises biochemically heterogeneous populations of hepatocytes (for example, albumin-containing cells constitute less than 1 % of hepatocytes in adult mouse liver)

(3.52) some small subpopulations of cells are likely to be resistant to the cytotoxic effects of a given chemical insult. Farber and colleagues (3.19) showed that certain subpopulations of cells in rat liver were resistant to cytotoxicity, mainly because of reduced or slowed xenobiotic activation or enhanced detoxification systems. A small percentage of cells which have acquired certain properties favoring survival and growth over their neighboring cells (i.e., neoplastic properties) will never fully redifferendate to a state of normal growth control and thus will tend to be clonally expanded (3.19, 3.73).

Hyalinized eosinophilic inclusions in hepatocytes were prominent in some mosquitofish and have been described as a response to hepatotoxicants in other fish species. These "cytosegresomes" may be formed by masses of cytoplasmic organelles from in the cell and condensed within lysosomes, or they may be derived from neighboring cells that have undergone apoptosis (3.37). Groff et al (3.24) 124 described eosinophilic cytoplasmic inclusions at several stages of hepatocyte injury

in cultured juvenile striped bass (Morone saxatilis) exposed to an unidentified

toxicant. Increased numbers of eosinophilic, apoptotic bodies were seen within

hepatocytes and resident macrophages during the cytotoxicity phase of

diethylnitrosamine-induced hepatic neoplasia in medaka (Oryzias latipes) (3.41).

Globular, eosinophilic bodies similar to those described in the mosquitofish of the

present study were also noted in medaka in a preliminary, large-scale,

diethylnitrosamine dose-response study (Dr. William Hawkins, Gulf Coast Research

Laboratory, Ocean Springs, Mississippi, personal communication). Whether these

structures are predominantly apoptotic bodies, masses of cytoplasmic organelles, or

protein accumulation within the rough endoplasmic reticulum secondary to

obstruction between the rough endoplasmic reticulum and the Golgi apparatus (3.24)

awaits further study.

Infiltration of necrotic areas and neoplastic foci by inflammatory cells

occurred in many of mosquitofish specimens. The presence of macrophages and

eosinophilic granule cells (EGC) is a well documented response to cell damage in

fish (3.2, 3.15, 3.16, 3.41, 3.72, 3.74). However, new relationships between

neoplastic cells, extracellular matrix, and "tumor-associated macrophages" (TAM)

are now being realized. Cytokines produced by TAM produce growth factors,

promote angiogenesis, and have procoagulant activity that may outweigh the

antitumor effects of TAM. In turn, neoplastic cells produce factors that attract and promote the survival of TAM, thus benefitting the neoplasm (3.47). 125 Most of the neoplastic lesions in mosquitofish were histologically similar to

those described in carcinogenicity studies with other fish species (3.25-26, 3.28)

including our earlier study with mosquitofish and MAM-Ac (3.42). Altered foci

were similar to those in rats (3.58, 3.67) and rainbow trout (3.28) exposed to hepatocarcinogens. These lesions, however, were fewer in number than comparable lesions in rodent liver and probably do not regress as do many of their rodent counterparts.

The low incidence of exocrine pancreatic neoplasms in mosquitofish contrasts with that seen in another poeciliid fish, the guppy (Poecilia reticulata), which had a high sensitivity to this lesion from MAM-Ac exposure (3.23). However, MAM-Ac failed to induce exocrine pancreatic neoplasms in six other species of small fish under similar exposure conditions (3.27).

Two types of spindle cell lesions occurred in MAM-Ac-exposed mosquitofish. The first type occurred soon after exposure and involved proliferation of small, oval to spindle-shaped cells whereas the second type appeared to be neoplasms composed of spindle cells that were diagnosed as early as 8 weeks post­ exposure. Proliferation of oval cells or nonparenchymal epithelial cells in response to hepatocarcinogens has been reported in other animal models (3.17, 3.37, 3.46,

3.60) and liver stem cells appear to be involved in some hepatocellular carcinomas and cholangiocellular carcinomas (3.69). Whether the oval cell or its precursor is the true stem cell is not known. Nevertheless, the degree of differentiation of epithelial tumors probably depends on the degree of differentiation of the target stem 126 cells. For example, some cells form well-differentiated hepatocellular carcinomas

whereas others undergo maturation arrest resulting in a whole spectrum of neoplastic

phenotypes (3.62-64). The epithelial origin of spindle cells in mosquitofish was

supported by positive cytokeratin staining and by electron microscopy. Similarly,

bile ducts but not hepatocytes, were positive with a comparable monoclonal antibody

to cytokeratins, in normal adult striped bass (Morone saxatilis) and medaka (Oryzias

latipes) (3.10). Assuming the mosquitofish spindle cells are analogous to the

putative liver stem cells and probably reside in some component of the biliary tree

(3.44), this hypothesis could explain the range of proliferative lesions and levels of

differentiation of the neoplasms seen in this study. Thus, these neoplasms are

believed to have originated from liver stem cells that have the facultative capacity to

differentiate into hepatocytes, biliary epithelial cells, pancreatic acinar cells, and

probably other tissues that arise from invaginations of the yolk sac-derived primitive

gut (3.24, 3.64).

Hepatotoxic changes occurred earlier and were more severe more severe in

adult mosquitofish than in similarly treated fry probably because of differences in

metabolism. Because significant apoptosis was not observed until one week after exposure in any group, inhibition of gene expression and, hence, transduction of signals or protein production for apoptosis may have been delayed until some of the cells recovered from the initial exposure. In other models inhibition of protein or

RNA synthesis can delay or prevent apoptosis in some cell types (3.59, 3.75). Two proto-oncogene products have been shown to act as regulators in the pathway. 127 When over-expressed in cells deprived of growth factors, the myc gene product

induces apoptosis (3.18) whereas over-expression of bcl-2 makes cells resistant to

apoptosis (3.71). Interleukin-1 /8-converting enzyme ("ICE"), the recently proposed vertebrate counterpart to the nematode enzyme CED-3, also promotes apoptosis

(3.7).

Mosquitofish exposed to MAM-Ac were shown immunohistochemically to overexpress p53 protein, but whether a specific point mutation was involved is not known and interpretation of these findings must be made cautiously. The significance of detecting p53 in tissue sections with antibodies that recognize both wild type and mutant forms is debated (3.8). Nevertheless, detection of increased levels of p53 protein might be attributed to interruption of normal degradative pathways even in the absence of mutations (3.20, 3.76) and overexpression could serve as a marker of disrupted p53 Junction. The p53 protein is typically detected in the nucleus and its occurrence in the cytoplasm of mosquitofish hepatocytes is unusual. In a recent study, however, the intracellular location and staining patterns of p53 protein were shown to be dependent on several factors including the type of fixative used (3.20). In different sections from a single tumor specimen, nuclear staining occurred with one type of fixative and cytoplasmic staining with another fixative. Standardization of techniques and methods of interpretation among different animal species is needed before further conclusions can be drawn.

In mosquitofish, p53 overexpression occurred in almost a third of the liver tumors in both high dose groups, similar to the levels reported in some human 128 tumor types (3.14). Also, p53 protein was detected in several fish before any

tumors were detected, but there was not close agreement between levels of

neoplastic development and positive p53 staining in mosquitofish. Further, the

significance of p53 overexpression in exposed specimens with no tumor at 8 to 41

weeks is uncertain. Additional studies are needed before immunohistochemical

detection of p53 can be considered a reliable biomarker of early or pre-neoplastic

changes in fish.

References

3.1. Ahmed FE (1993). The Xiphophorus fish: a model for molecular mechanisms of environmental carcinogenesis. Environ. Carcino. & Ecotox. Revs. C ll: 125-161.

3.2. Ainsworth AJ (1992). Fish granulocytes: morphology, distribution, and function. Annual Rev. o f Fish Diseases 2: 123-148.

3.3. Anders F, Schartl M, Bamekow A, and Anders A (1984). Xiphophorus as an in vivo model for studies on normal and defective control of oncogenes. Adv. Cancer Res. 42: 191-275.

3.4. Aoki K and Mastsudaira H (1977). Induction of hepatic tumors in a teleost (Oryzias latipes) after treatment with methylazoxymethanol acetate: brief communication. J. Natl. Cancer Inst. 59: 1747-1749.

3.5. Aoki K and Matsudaira H (1981). Factors influencing tumorigenesis in the liver after treatment with methylazoxymethanol acetate in a teleost, Oryzias latipes. In: Phyletic Approaches to Cancer, CJ Dawe (ed). Japan Scientific Societies Press, Tokyo, pp. 205-216.

3.6. Aoki K and Matsudaira H (1984). Factors influencing methylazoxymethanol acetate initiation of liver tumors in Oryzias latipes: carcinogen dosage and time of exposure. Natl. Cancer Inst. Monogr. 65: 345-351.

3.7. Barinaga M (1994). Cell suicide: by ICE, not fire. Science 263: 754-756.

3.8. Battifora H (1994). p53 immunohistochemistry: a word of caution. Human Pathol. 25: 435-436. 129 3.9. Bunton TE (1990). Hepatopathology of diethylnitrosamine in the medaka (Oryzias latipes) following short-term exposure. Toxicol. Pathol. 18: 313-323.

3.10. Bunton TE (1993). The immunocytochemistry of cytokeratin in fish tissues. Vet. Pathol. 30: 418-425.

3.11. Caron de Fromentel C, Pakdel F, Chapus A, Baney C, May P, and Soussi T (1992). Rainbow trout p53: cDNA cloning and biochemical characterization. Gene 112: 241-245.

3.12. Couch JA and Courtney LA (1987). N-nitrosodiethylamine-induced hepatocarcinogenesis in estuaiine sheepshead minnow ( Cyprinodon variegatus) neoplasms and related lesions compared with mammalian lesions. J. Natl. Cancer Inst. 79: 297-321.

3.13. Couch JA and Harshbarger JC (1985). Effects of carcinogenic agents on aquatic animals: an environmental and experimental overview. Environ. Carcinogen. Revs. 3: 63-105.

3.14. Culotta E and Koshland DE (1993). p53 sweeps through cancer research. Science 262: 1958-1961.

3.15. Ellis AE (1977). The leucocytes of fish: a review. J. Fish Biol. 11: 453-491.

3.16. Ellis AE (1985). Eosinophilic granular cells (EGC) and histamine responses to Aeromonas salmonicida toxins in rainbow trout. Dev. Comp. Immunol. 9: 251-260.

3.17. Eustis SL, Boorman GA, Harada T, and Popp JA (1990). Liver. In: Pathology o f the Fischer Rat, GA Boorman, SL Eustis, MR Elwell, CA Montgomery, WF MacKenzie (eds). Academic Press, Inc., San Diego, California, pp. 71-94.

3.18. Evan GI, Wyllie AH, Gilbert CS, et al. (1992). Induction of apoptosis in fibroblasts by c-myc protein. Cell 69: 119-128.

3.19. Farber E and Sarma DSR (1987). Hepatocarcinogenesis: a dynamic cellular perspective. Lab. Invest. 56: 4-22.

3.20. Fisher CJ, Gillett CE, Vojtesek B, Barnes DM, and Millis RR (1994). Problems with p53 immunohistochemical staining: the effect of fixation and variation in the methods of evaluation. Br. J. Cancer 69: 26-31. 130 3.21. Foley J, Ton T, Maronpot R, Butterworth B, and Goldsworthy TL (1993). Comparison of proliferating cell nuclear antigen to tritiated thymidine as a marker of proliferating hepatocytes in rats. Environ. Health Perspect. 101 (Suppl 5): 199-206.

3.22. Foley JF, Dietrich DR, Swenberg JA, and Maronpot RR (1991). Detection and evaluation of proliferating cell nuclear antigen (PCNA) in rat tissue by an improved immunohistochemical procedure. J. Histotechnology 14: 237-241.

3.23. Foumie JW, Hawkins WE, Overstreet RM, and Walker WW (1987). Exocrine pancreatic neoplasms induced by methylazoxymethanol acetate in the guppy Poecilia reticulata. J. Natl. Cancer Inst. 78: 715-725.

3.24. Groff JM, Hinton DE, McDowell TS, and Hedrick RP (1992). Progression and resolution of megalocydc hepatopathy with exocrine pancreatic metaplasia in a population of cultured juvenile striped bass Morone saxatilis. Dis. Aquat. Org. 13: 189-202.

3.25. Harada T and Hatanaka J (1988). Liver cell carcinomas in the medaka (Oryzias latipes) induced by methylazoxymethanol acetate. J. Comp. Path. 98: 441-452.

3.26. Hawkins WE, Overstreet RM, Foumie JW, and Walker WW (1985). Development of aquarium fish models for environmental carcinogenesis: tumor induction in seven species. J. Appl. Toxicol. 5: 261-264.

3.27. Hawkins WE, Overstreet RM, and Walker WW (1988). Carcinogenicity tests with small fish species. Aquatic Toxicol. 11: 113-128.

3.28. Hendricks JD, Meyers TR, and Shelton DW (1984). Histological progression of hepatic neoplasia in rainbow trout ( Salmo gairdneri). Natl. Cancer Inst. Monogr. 65 65: 321-336.

3.29. Hendricks JD, Sinnhuber RO, Henderson MC, and Buhler DR (1981). Liver and kidney pathology in rainbow trout (Salmo gairdneri) exposed to dietary pyrrolizidine (senecio) alkaloids. Exper. Mol. Pathol. 35: 170-183.

3.30. Hinton DE, Baumann PC, Gardner GR, et al. (1992). Histopathologic biomarkers. In: Biomarkers: Biochemical, Physiological, and Histological Markers o f Anthropogenic Stress, RJ Huggett, RA Kimerle, PM Mehrle, HL Bergman, (eds). Lewis Publishers, Boca Raton, Florida, pp. 155-209. 131 3.31. Hinton DE, Couch JA, Teh SJ, and Courtney LA (1988). Cytological changes during progression of neoplasia in selected fish species. Aquatic Toxicol. 11: 77-112.

3.32. Hinton DE, Lantz RC, and Hampton JA (1984). Effect of age and exposure to a carcinogen on the structure of the medaka liver: a morphometric study. Natl. Cancer Inst. Monogr. 65: 239-249.

3.33. Hoover KL (1984) Use o f Small Fish Species in Carcinogenicity Testing, Natl. Cancer Inst. Monogr. 65, NIH Publication No. 84-2653, National Cancer Institute, Bethesda, Maryland.

3.34. Ishikawa T and Takayama S (1979). Importance of hepatic neoplasms in lower vertebrate animals as a tool in cancer research. J. Toxicol. Environ. Health 5: 537-550.

3.35. Ishikawa T, Takayama S, and Kitagawa T (1980). Correlation between time of partial hepatectomy after a single treatment with diethylnitrosamine and induction of adenosine triphosphatase-deficient islands in rat liver. Cancer Res. 40: 4261-4264.

3.36. Johnston MV, Grzanna R, and Coyle JT (1979). Methylazoxymethanol treatment of fetal rats results in abnormally dense noradrenergic innervation of neocortex. Science 203: 369-371.

3.37. Kelly WR (1993). The liver and biliary system. In: Pathology o f Domestic Animals (4th ed.), KVF Jubb, PC Kennedy, N Palmer (eds). Academic Press, Inc., San Diego, California, vol. 2, pp. 319-406.

3.38. Kim HY, Stermitz FR, Molyneux RJ, Wilson DW, Taylor D, and Coulombe RA Jr. (1993). Structural influences on pyrrolizidine alkaloid-induced cytopathology. Tox. Appl. Pharm. 122: 61-69.

3.39. Lane DP (1993). A death in the life of p53. Nature 362: 786-787.

3.40. Laquer GL and Spatz M (1968). Toxicology of cycasin. Cancer Res. 28: 2262-2267.

3.41. Lauren DJ, Teh SJ, and Hinton DE (1990). Cytotoxicity phase of diethylnitrosamine-induced hepatic neoplasia in medaka. Cancer Res. 50: 5504-5514. 132 3.42. Law JM, Hawkins WE, Overstreet RM, and Walker WW (1994). Hepatocarcinogenesis in western mosquitofish ( Gambusia affinis) exposed to methylazoxymethanol acetate. 7. Comp. Path. 110: 117-127.

3.43. Lawley PD (1984). Carcinogenesis by alkylating agents. In: Chemical Carcinogens, CE Searle (ed). American Chemical Society Monographs, Washington, DC, pp. 325-484.

3.44. Lenzi R, Liu MH, Tarsetti F, et al. (1992). Histogenesis of bile duct-like cells proliferating during ethionine hepatocarcinogenesis: evidence for a biliary epithelial nature of oval cells. Lab. Invest. 66: 390-402.

3.45. Levine AJ, Momand J, and Finlay CA (1991). The p53 tumour suppressor gene. Nature 351: 453-456.

3.46. Lombardi B (1982). On the nature, properties and significance of oval cells. Recent Trends Chem. Carcinogen. 1: 37-56.

3.47. Mantovani A (1994). Tumor-associated macrophages in neoplastic progression: a paradigm for the in vivo function of chemokines. Lab. Invest. 71: 5-16.

3.48. Maronpot RR (1991). Chemical carcinogenesis. In: Handbook o f Toxicologic Pathology, WM Haschek, CG Rousseaux (eds). Academic Press, Inc., San Diego, pp. 91-130.

3.49. Masahito P, Ishikawa T, and Sugano H (1988). Fish tumors and their importance in cancer research. Jpn. J. Cancer Res. 79: 545-555.

3.50. Mercer WE, Shields MT, Lin D, Appella E, and Ullrich SJ (1991). Growth suppression induced by wild-type p53 protein is accompanied by selective down-regulation of proliferating-cell nuclear antigen expression. Proc. Natl. Acad. Sci. USA 88: 1958-1962.

3.51. Metcalfe CD (1989). Tests for predicting carcinogenicity in fish. CRCRev. Aquatic Sci. 1: 111-129.

3.52. Michaelson J (1993). Cellular selection in the genesis of multicellular organization. Lab. Invest. 69: 136-151.

3.53. Morris RW (1993). Analysis of cell proliferation data. Environ. Health Perspect. 101 (Suppl 5): 73-78. 133 3.54. Myers MS, Landahl JT, Krahn MM, Johnson LL, and McCain BB (1990). Overview of studies on liver carcinogenesis in English sole from Puget Sound: evidence for a xenobiotic chemical etiology I: pathology and epizootiology. Sci. Total Environ. 94: 33-50.

3.55. Myers MS, Rhodes LD, and McCain BB (1987). Pathologic anatomy and patterns of occurrence of hepatic neoplasms, putative preneoplastic lesions, and other idiopathic hepatic conditions in English sole ( Parophrys vetulus) from Puget Sound, Washington. J. Natl. Cancer Inst. 78: 333-363.

3.56. Oren M (1992). p53: the ultimate tumor suppressor gene? FASEB J. 6: 3169-3176.

3.57. Ortego LS, Hawkins WE, Walker WW, Krol RM, and Benson WH (1994). Detection of proliferating cell nuclear antigen in tissues of three small fish species. Biotechnic & Histochemistry (in press).

3.58. Popp JA and Goldsworthy TL (1989). Defining foci of cellular alteration in short-term and medium-term rat liver tumor models. Toxicol. Pathol. 17: 561-568.

3.59. Raff MC (1992). Social controls on cell survival and cell death. Nature 356: 397-400.

3.60. Sarraf C, Lalani E, Golding M, Anilkumar TV, Poulsom R, and Alison M (1994). Cell behavior in the acetylaminofluorene-treated regenerating rat liver. Am. J. Pathol. 145: 1114-1126.

3.61. Schultz ME and Schultz RJ (1982). Induction of hepatic tumors with 7,12-dimethylbenz[a]anthracene in two species of viviparous fish (Genus Poeciliopsis). Environ. Res. 27: 337-351.

3.62. Sell S (1993). Cellular origin of cancer: dedifferentiation or stem cell maturation arrest? Environ. Health. Perspec. 101 (Suppl 5): 15-26.

3.63. Sell S and Dunsford HA (1989). Evidence for the stem cell origin of hepatocellular carcinoma and cholangiocarcinoma. Am. J. Pathol. 134: 1347-1363.

3.64. Sell S and Pierce GB (1994). Maturation arrest of stem cell differentiation is a common pathway for the cellular origin of teratocarcinomas and epithelial cancers. Lab. Invest. 70: 6-22. 134 3.65. Sieber SM, Correa P, Dalgard DW, Mclntire KR, and Adamson RH (1980). Carcinogenicity and hepatoxicity of cycasin and its aglycone methylazoxymethanol acetate in nonhuman primates. J. Natl. Cancer Inst. 65: 177-183.

3.66. Spatz M (1969). Toxic and carcinogenic alkylating agents from cycads. Ann. N.Y. Acad. Sci. 163: 848-859.

3.67. Squire RA and Levitt MH (1975). Report of a workshop on classification of specific hepatocellular lesions in rats. Cancer Res. 35: 3214-3223.

3.68. Stewart BW (1994). Mechanisms of apoptosis: integration of genetic, biochemical, and cellular indicators. J. Natl. Cancer Inst. 86: 1286-1296.

3.69. Travis J (1993). The search for liver stem cells picks up. Science 259: 1829.

3.70. Ullrich SJ, Anderson CW, Mercer WE, and Appella E (1992). The p53 tumor suppressor protein, a modulator of cell proliferation. J. Biol. Chem. 267: 15259-15262.

3.71. Vaux DL, Cory S, and Adams JM (1988). Bel-2 gene promotes haemopoietic cell survival and cooperates with c-myc to immortalize pre-B cells. Nature 335: 440-442.

3.72. Vogelbein WK, Foumie JW, and Overstreet RM (1987). Sequential development and morphology of experimentally induced hepatic melano-macrophage centres in Rivulus marmoratus. J. Fish Biol. 31 (Supplement A): 145-153.

3.73. Williams GM (1987). The significance of preneoplastic liver lesions in experimental animals. Adv. Vet. Sc. Comp. Med. 31: 21-44.

3.74. Wolke RE (1992). Piscine macrophage aggregates: a review.Ann. Rev. Fish Dis. 2: 91-108.

3.75. Wyllie AH, Morris RG, Smith AL, and Dunlop D (1984). Chromatin cleavage in apoptosis: association with condensed chromatin morphology and dependence on macromolecular synthesis. J. Pathol. 142: 67-77.

3.76. Wynford-Thomas D (1992). p53 in tumour pathology: can we trust immunocytochemistry? J. Pathol. 166: 329-330. 135 3.77. Ying TS, Enomoto K, Sarma DSR, and Farber E (1982). Effects of delays in the cell cycle on the induction of preneoplastic and neoplastic lesions in rat liver by 1,2-dimethylhydrazine. Cancer Res. 42: 876-880.

3.78. Zedeck M and Sternberg SS (1977). Tumor induction in intact and regenerating liver of adult rats by a single treatment with methylazoxymethanol acetate. Chemico-Biological Interactions 17: 291-296.

3.79. Zedeck MS, Sternberg SS, McGowan J, and Poynter RW (1972). Methylazoxymethanol acetate: induction of tumors and early effects on RNA synthesis. Fed. Proc. 31: 1485-1492.

3.80. Zedeck MS, Sternberg SS, Yataganas X, and McGowan J (1974). Early changes induced in rat liver by methylazoxymethanol acetate: mitotic abnormalities and polyploidy. J. Natl. Cancer Inst. 53: 719-724. CHAPTER 4. QUANTIFICATION OF DNA ADDUCTS AT FEMTOMOLE LEVELS IN FISH EXPOSED TO ALKYLATING AGENTS

Summary

It is widely believed that most chemical carcinogens act by binding to cellular genetic material and causing somatic mutations. Chemical modification of DNA

(i.e., the formation of DNA adducts) is the first in a series of stages that lead to mutation, cell transformation, and tumor development.

Sensitive methods for detection of DNA adducts are essential for mechanistic studies of mutagenesis and carcinogenesis and for biomonitoring of populations at risk for environmentally caused cancer. DNA adducts are currently detected with such methods as 32P-postlabeling, immunoassays, HPLC with fluorescence detection, and gas chromatography/mass spectrometry. Alkylation at the N-7 position of guanine is a preferential site of attack for most alkylating agents, but may be of little biological consequence, while adduct formation at the O6 position of guanine (less common) is most highly correlated with carcinogenesis.

We have recently begun to examine the suitability of western mosquitofish

(Gambusia affinis) as a sentinel species for monitoring environmental exposures to mutagens and carcinogens. Mosquitofish are a species native to the U.S. which are also widely distributed in warm waters throughout the world. These fish are ideal for monitoring local exposures because they have a limited home range. In the laboratory, mosquitofish are easily cultured using techniques developed for other small fish species. They breed prolifically and appear resistant to infectious diseases.

136 137 In these studies, mosquitofish were exposed to a model alkylating agent,

methylazoxymethanol acetate, in the ambient water for 2 h. Fish were then

transferred to clean water and held for up to 72 h. Livers were excised, pooled for

5 fish, and DNA was extracted and prepared for analysis. Liver DNA extracts were

assessed for levels of N-7 and 0 6-methylguanine by isotope dilution GC/mass

spectrometry, using deuterated analogs of methyl guanine adducts as standards.

Detection limits for the assay were approximately 1.6 femtomoles of adduct per f i g

DNA (about 1 adduct per 2 million bases). Approximately 55-185 pg (330-1100 femtomoles) 0 6-methylguanine per fig DNA were measured in exposed fish in the first 72 hr after exposure. These lesions were correlated with a 33% liver tumor incidence after 25 weeks in parallel experiments.

Introduction

Chemical carcinogenesis is a complex, multistage process thought to involve genetic and epigenetic damage to susceptible cells that subsequently gain a selective growth advantage over neighboring cells as the result of activation of protooncogenes and/or inactivation of tumor suppressor genes (4.19). An approach to assessing genotoxic damage that integrates all of the various factors involved in chemical exposure, uptake, biotransformation, etc., is to compare levels of specific covalent adducts of DNA in the target tissue. It is generally accepted that most chemical carcinogens act by interacting with the genetic material of the cell, in particular the DNA template (4.31). Chemical modification of DNA is the first in a series of steps that lead to mutation, cell transformation, and tumor development 138 (4.47). In fact, it has been suggested that any chemical that forms DNA adducts, even at low levels, has the potential to be mutagenic and carcinogenic (4.8). DNA adduct determinations can therefore provide crucial information on metabolic pathways as well as chemical effects on DNA structure, transcription, synthesis, and repair. Adduct analyses can also provide a direct test of the somatic mutation theory, and, DNA adducts can be considered dosimeters of exposure to chemicals in cancer risk assessments (4.4).

Alkylating agents are archetypal carcinogens, in that most other carcinogens are active only after being metabolized to alkylating or aralkylating agents (4.31).

Substitution at the N-7 position of guanine is a preferential site of attack for most alkylating agents, while attack at the O6 position of guanine (less common) is most highly correlated with carcinogenesis (4.5, 4.31, 4.32, 4.46). Agents that cause this type of DNA damage can be divided into two major classes: (1) those that form bulky chemical adducts of the purine and pyrimidine bases, represented by such classical chemical carcinogens as benzo(a)pyrene, aflatoxin, 7,12- dimethylbenz(a)anthracene, and 2-acetylaminofluorene; and (2) agents that form simple base modifications, differing from bulky adducts in that they do not distort the DNA helix (4.11). This second group includes such classical alkylating mutagens as N-methyl-N-nitrosourea (MNU), methyl methanesulfonate (MMS), N- methyl-N’-nitro-N-nitrosoguanidine (MNNG), N-ethyl-N-nitrosourea (ENU), ethyl methanesulfonate (EMS), and diethylnitrosamine (DEN) or their metabolites. 139 Although simple base modifications can block DNA replication, more often they

directly produce promutagenic adducts in DNA (4.11, 4.38, 4.41).

Methylazoxymethanol acetate (MAM-Ac), a model alkylating carcinogen,

forms methyl DNA adducts following esterase-mediated release of a highly reactive

carbonium ion (4.14). MAM-Ac is carcinogenic in several mammalian animal

models, and has been shown to cause liver neoplasia in several fish species (4.21,

4.43). For example, a 2 hr pulse exposure to 10 mg/L MAM-Ac in ambient water

results in a 33% incidence of liver tumors in western mosquitofish by 25 weeks after

exposure and more than a 50% incidence after 40 weeks (4.29). There is increasing

interest in the use of small fish models for carcinogenicity testing, for investigation

of basic carcinogenic mechanisms, and as sentinel species for monitoring health risks

associated with aquatic environments (4.22, 4.23, 4.25). Unlike many rodent

models, fish have an extremely low background incidence of tumors and are

relatively inexpensive to breed and maintain. However, because most published

studies to date have been performed with exotic or non-native species or with larger cold water fish species difficult to rear in the laboratory, there is a need for a

freshwater, temperate-zone small fish species that could serve as an environmental

sentinel and also be amenable to laboratory culture for validation of biological markers.

The present study was designed to measure DNA adduct formation and persistence in the liver of western mosquitofish (Gambusia afflnis) exposed to

MAM-Ac in order to further validate the use of this sentinel species in both 140 laboratory and field studies. The western mosquitofish is a small, native, freshwater

fish that is distributed from New Jersey to northern Mexico and has been introduced

world-wide as a mosquito control agent (4.9). Mosquitofish are easily cultured

using techniques developed for other small fish species such as the guppy (Poecilia

reticulata). They breed prolifically and appear resistant to many infectious diseases.

They have been shown in previous studies to express measurable changes in a

variety of biomarkers following contaminant exposure and, thus, may prove useful

for direct validation of field studies by enabling substances to be tested under both

laboratory and field conditions (4.30).

Direct measurement of DNA adducts may be employed as a biomarker of

both exposure and effects in organisms exposed to genotoxic agents (4.39).

Biomarkers such as DNA adducts may afford a temporally and spatially integrated

measure of bioavailable pollutants not provided by chemical analyses of water, soils,

or tissues, thus giving information that is more relevant to evaluating ecological or

health risks (4.37). The application of DNA adduct analysis in fish has been

reviewed recently by Maccubbin (4.33), and it was apparent that few studies with

small fish species have been reported. Formation and persistence of 0 6-ethylguanine

in rainbow trout exposed to DEN were reported by Fong and co-workers using

HPLC with fluorescence detection (4.15). Few studies, however, have attempted to

measure specific adduct levels in small fish species exposed to small alkylating carcinogens. This is a result of the fact that older methods were not sensitive enough to utilize the small amounts of tissue available from these species. Accurate, 141 sensitive measurements of DNA adduct levels have been obtained using stable

isotope-labeled analogs of analytes as internal standards with analysis by selected ion

monitoring GC/MS (4.10). The present report describes the application of these

techniques for use in small fish species.

Methods

Experimental Design

Experiments were performed to investigate the extent and kinetics of DNA

alkylation and subsequent DNA repair following a single, short, pulse exposure to

MAM-Ac. The experimental design employed is shown schematically in Fig. 4.1.

All specimens used in this study were the progeny of laboratory-maintained western

mosquitofish which were originally caught in Ocean Springs, Mississippi, USA. In

both variants of the experimental protocol, adult fish ranging from 9 to 14 months

old were used. The fish were acclimated for a minimum of two weeks in 40-liter aquaria containing purified, HPLC-grade water maintained at 27±1°C by a recirculating water bath. The aquarium room was maintained on a 12-hr light/12-hr dark cycle with incandescent lighting. The fish were fed commercial flake food

(Prime Tropical Flakes-Yellow, Zeigler Brothers, Inc., Gardners, PA, USA) two times daily. Water quality was monitored and tank maintenance performed once weekly.

Exposure

Two basic protocols, differing only in post-exposure sampling times (Fig. 1) and numbers of specimens used, were followed in these experiments. Test fish were 142

DNA ADDUCT ANALYSIS Experimental Design: Experiment 1 2 hour puise exposure Experiment 2'

Excise vers, pool

Extract & purify DNA

Acid/thermal hydrolysis

Spike samples with deuterated analogs

Lyophil ze samples •Excl 11 Expose Ash. sample 24 hra. later •Expt. 2: Expos* Ash, sample mM-way through exposure, immadiataly attar exposure, TMS derivatization then 4,8,24,48, and 72 hra. later to assess brmotion and repair of DNA adducts GC/MS-SIM analysis for DNA adducts

Fig. 4.1: Experimental design for DNA adduct quantification by GC/MS in small fish species. 143 randomly distributed into 4-L pyrex beakers containing aged, purified, HPLC-grade

water (pH 8.7, dissolved oxygen 9.0 mg/L) maintained at 27±1°C in a heated

recirculating waterbath contained within a ventilated laboratory hood. Five

randomly selected adult fish ranging from 30 to 35 mm in length were used for each

treatment. Ten fish of similar size served as unexposed controls during each

experiment. Specimens were exposed for 2 hr to methylazoxymethanol acetate

(MAM-Ac; Sigma Chemical Company, St. Louis, Missouri, USA) at 10 mg/L in

test water, then rinsed and transferred to clean water until sampled. This protocol

was similar to that used in an independent study (4.29) in which liver tumors were

induced after 25 weeks in mosquitofish exposed to MAM-Ac. Exposed and control

fish were held in separate 40-liter aquaria maintained at 27±1°C. At each

designated sampling time, five fish were randomly selected and were euthanatized

by stunning in ice water followed by decapitation.

DNA Extraction and Purification

Livers from 5 fish were immediately excised, pooled, weighed, and flash- frozen in liquid nitrogen. DNA was extracted and purified according to standard methods (4.44), with some modifications. Briefly, the frozen liver tissue was pulverized in 2 ml polypropylene microtubes, then suspended in digestion buffer containing proteinase-K and incubated with shaking at 50°C for 14-16 hrs. A reagent blank which contained all reagents and materials except liver tissue was also prepared. Organic extraction was accomplished using high purity, biotechnology grade phenol:chloroform:isoamyl alcohol (25:24:1, pH 8.0, Amresco, Solon, Ohio, 144 USA). The extracted DNA was then precipitated by addition of isopropanol and 7.5

M ammonium acetate (to reduce residual RNA) followed by centrifugation at 13,000

x g. The DNA pellet was then rinsed in 70% ethanol and briefly air-dried.

Hydrolysis and Derivatization

Isolated DNA samples were hydrolyzed in 0.5 ml of 60% formic acid

(Mallinckrodt, Paris, Kentucky, USA) in evacuated, PFTE-sealed, conical bottom,

glass vials at 75°C for 30 min. After cooling, all samples were spiked with a

mixture of four deuterium-labeled adduct standards (see below). Standard blanks

were also prepared at this time by adding both the deuterated and a non-deuterated

guanine adduct standard mixture to clean vials. HPLC grade water (2 ml) was

added to each tube except the standard blanks. Samples and blanks were lyophilized

and trimethylsilyl derivatives were produced using a mixture of 80 /zL of

bis(trimethylsilyl)-trifluoroacetamide (BSTFA)/1% trimethylsilylchlorosilane

(Alltech, Deerfield, Illinois, USA) and 20 /xL of acetonitrile under nitrogen at 75°C

for 30 min. After cooling, derivatized mixtures were injected directly onto the gas chromatograph (GC). Lyophilized mixtures of standard methylated and ethylated guanine bases and their deuterium-labeled analogs were silylated in the same fashion and used to create standard response curves on the GC.

Stable Isotope-Labeled Standards

7-methylguanine (7-MG) was purchased from Sigma. 7-ethylguanine (7-EG),

0 6-methylguanine (O-MG), and 0 6-ethylguanine (O-EG) were prepared using procedures in the literature (4.5, 4.13). Procedures specific for the synthesis of 145 methyl derivatives were modified to yield ethyl derivatives by selecting the

corresponding ethyl donating agent. Deuterium-labeled analogs of the four analytes

were prepared from the corresponding fully-deuterated methyl or ethyl donating

agents. Briefly, for the synthesis of deuterium-labeled 7-alkylguanines, 2’-

deoxyguanosine 5'-monophosphate (Aldrich Chemical Company, Milwaukee,

Wisconsin, USA) was dissolved in DMSO and stirred overnight with iodomethane-d’j

(Sigma) or iodoethane-tf, (Aldrich). The resulting product was precipitated in ethyl acetate and purified by recrystalization. For the synthesis of the deuterated O6- alkylguanines, either methyl-d, alcohol-d (Aldrich) or ethyl-dj alcohol-d (Aldrich) was first reacted with sodium metal to form Na+0'CD3 or Na+0'C2Ds. These products were heated with 2-amino-6-chloropurine (Aldrich) at 70°C overnight in sealed vials. The solutions were neutralized with 3.5 mM glacial acetic acid, solvents removed by evaporation, and the resulting products purified by recrystalization.

The purity of the methylated and ethylated guanine bases and their stable- labeled analogs was confirmed by GC/MS (see below), HPLC with fluorescence detection (4.5), and UV spectrophotometry (4.2). The identity of the derivatized compounds was verified on the basis of reference mass spectra (NIST/EPA/NIH

Spectral Library, U.S. Department of Commerce), where available.

Quantitative Analysis

The derivatized adduct standards and corresponding alkylated bases isolated from control and exposed biological samples described earlier were analyzed using a 146 Hewlett-Packard GC/MS system. A HP 5890 Gas Chromatograph equipped with a

30 m by 0.25 mm ID DB-5 column with a liquid film of 0.25 /zm was directly

interfaced with a HP 5970 Mass Selective Detector. The column was initially held

at 100°C for 3 min and then temperature programmed to 300°C at a rate of

7°C/min, following the method of Dizdaroglu (4.10). The transfer line was held at

300°C. Initial spectral identifications and confirmations were carried out using the

scanning mode (50-550 amu, at 1.3 scan/sec). Three ions were selected for each of

the 8 analytes based upon relative abundance and the lack of mass spectral

interferences such as column bleed. A selected ion monitoring (SIM) method was

then developed based on retention times and characteristic ions of each component.

Samples and standards were introduced to the GC using a HP 7673 liquid

autosampler set to deliver 2 /xL aliquots into a split/splitless injection liner packed

with deactivated glass wool and held at 250°C. The electron multiplier was operated at 400 volts above the tune value.

Results

Table 4.1 shows the chromatographic and mass spectral parameters for each analyte. Trimethylsilyl-guanine adducts and their deuterated analogs were initially analyzed separately by GC/MS. Examples of extracted ion chromatograms within selected retention time windows for the non-deuterated adduct standards are given in

Fig. 4.2. Their corresponding spectra are shown in Fig. 4.3. Gas chromatographic retention times of the derivatized adduct standards were not significantly different

(< 0.05 min) from their respective deuterium-labeled analogs. 147

Table 4.1: Retention times and characteristic mass fragments of guanine adducts.

Retention : Relative Adduct m lz Time Abundance* 0 6-methylguanine 21.86 294 100% 309 35% 295 24% d3 0 6-methylguanine 21.83 297 100% 312 33% 298 24% 7-methylguanine 22.56 294 100% 309 22% 295 21% d3 7-methylguanine 22.54 297 100% 312 21% 298 25% 0 6-ethylguanine 22.38 308 *67% 323 79% 324 20% d5 0 6-ethylguanine 22.31 313 *93% 328 93% 329 25% 7-ethylguanine 22.87 308 100% 323 20% 324 52% d3 7-ethylguanine 22.84 313 100% 328 22% 329 13%

T he most abundant mass fragments for the deuterated and non-deuterated 0 6-ethylguanines were 280 and 281, respectively, the same as seen in liquid phase background bleed of the GC column. Therefore, these ions were not used for SIM analyses.

fSpectra obtained under standard tune conditions. 148

06- & N7-METHYLGUANINE 06-4 N7-ETHYLGUANINE Abundancelon 294.00 (293.70 co 294.70) ■ AbundanceZon 308.00 (307.70 co 308.70) r - I 1000000 2000000-j 22154 22 36 4 i I 1500000

500000 j 0 — l \L. rime— >______22.00 22.50______ritne--> 22.50 23.00 AbundanceIon 309.00 (308.70 co 309.70) j Abundance Ion 323.00 (322.70 CO 323.70) | 21 86 22136 22.54 1000000 -] 400000 1

500000 200000

22.86 US _A_ iTime--> 22.00 22.50 rime--> 22.50 23.00 d5-06- 4 d5-N7-ETHYLGUANINE______Abundance Ion 29 7.00 (296.70 CO 297.70) Abundance Ion 313.00 (312.70 co 313, 70) 21 82 1 22130 ! 600000 ■* 1000000 j

j 400000 22.82 500000 -j 1 2 0 0 0 0 0 -i 4 1 22.53 J u ■ ■ .... I 0 Time—> 22.00 22.50 rim e—>______22.50 23.00 AbundanceIon 31 2.00 (311.70 CO 312.70) Abundance Ion 328.00 (327.70 co 328. 70) j 21 82 j 22131 400000 ■) i 600000 300000 -j 1 400000 200000 -1 200000 100000 - 1 22.53 . 22.82 J J 0 k ^ Time—> 22! 00 2 2 ! 50 fTime- 22.50 23.00

Fig. 4.2: Extracted ion chromatograms of the four DNA adduct standards and their deuterated analogs. 149

iundance Scan 878 (21.856 min): 4273A3.D (-,*)

O

59 73 89 191 220

Scan 917 (22.382 min): 4273A3.D (-,*) 280 323 O 6-ethylguanine 308 500

73

Scan 930 (22.557 min): 4273A3.D (-,*) 294 N^.methylguanine

500

309 99 113725140 220 rc/z--> Scan 953 (22.867 min): 4273A3.D (-,*) 308

500 -

323 56 73 84 1Q6 1321*6 171 1?4 218 234 2Sg4 280

Fig. 4.3: Mass spectra of the four alkyl DNA adduct standards. 150 Standard curves were prepared by plotting amount ratios (non-deuterated

versus deuterated) over a range representing an average of 30 femtomoles to 6.1

nanomoles of alkylguanines on column. Duplicate dilutions at nine levels of the

non-deuterated compound containing a constant amount of its deuterated analog were

prepared by lyophilization and derivatization. Regression statistics were generated

by plotting amount ratios against response ratios. The standard curves resulted in

linear responses (R2 = 0.998 or greater) across all data points and indicated

detection limits as low as femtomoles (1015 moles) of DNA adduct.

Results of DNA adduct analyses from the first set of experiments, in which

levels of 7-MG and O-MG were quantified in fish after exposure to a single pulse

dose of MAM-Ac, are shown in table 4.2 and Fig. 4.4. On average, approximately

10 pg (61 femtomoles) 0 6-methylguanine and approximately 50 pg (303 femtomoles)

7-methylguanine per fig DNA were obtained at 24 hr post-exposure. Levels of 7-

MG were significantly higher in the exposed fish than the controls. No O-MG was

detected in the control fish in this experiment.

Results from the second set of experiments which were designed to demonstrate and quantify the kinetics of DNA adduct formation and repair over time are shown in Table 4.3 and Fig. 4.5.

Discussion

The present studies with western mosquitofish (Gambusia affinis) were performed in order to develop and apply a GC/MS method for the identification and quantification of DNA adducts in tissues of a small fish species exposed to a known Table 4.2. Alkylguanine adduct concentrations in liver DNA of western mosquitofish 24 hr after exposure to 10 ppm MAM- Ac. Percent recovery of deuterium-labeled standards is also given.

Recovery of deuterated surrogates FINAL VALUES, corr. for recovery SAMPLE d06-MG d7-MG d06-EG d7-EG 06-MG 7-MG 06-EG 7-EG TISS WT. % % % % ng/mg ng/mg ng/nig ng/mg ng/mg Reag. Blank 85% 28% 79% 46% ncl nd nd nd Control 1 27% 25% 22% 32% nd 0.3 nd nd 133 Control 2 30% 27% 24% 32% nd 0.3 nd nd 167 Exposed 1 22% 33% 17% 35% 0.172 1.0 nd nd 176 Exposed 2 44% 61% 33% 67% 0.089 0.7 nd nd 152 Exposed 3 52% 77% 40% 83% 0.107 0.8 nd nd 120 N7-MG N 06-MG (x3)

SAMPLE

Fig. 4.4: Methylguanine adduct concentrations in liver DNA of control and exposed western mosquitofish 24 hr after exposure to 10 ppm MAM-Ac. Table 4.3: Alkylguanine adduct concentrations in liver DNA of western mosquitofish before, mid-way through (-1), and after 2 hr pulse exposure to 10 ppm MAM-Ac. The "0" time point denotes sampling immediately following removal of fish specimens to carcinogen-free water. Percent recovery of deuterium-labeled standards is also given.

RECOVERY FINAL VALUES, corr. for recovery Time post- d06-MG d7-MG d06-EG d7-EG 06-MG 7-MG 06-EG 7-EG TISS. WT. exposure (hrs.) % % % % ng/mg ng/mg ng/mg ng/mg_____ mg REAG. BLK.A 90% 6% 81% 20% 0.030 iid OOltT iid 66*" REAG. BLK. B 40% 3% 31% 7% 0.051 0.227 0.050 0.165 79* Control a 21% 4% 31% nd 0.033 0.279 nd nd 87 Control b 17% 27% 17% 43% nd 0.397 nd 0.017 87 -la 16% nd 23% 4% 0.369 nd 0.021 nd 59 -lb 20% 31% 17% 60% 0.146 1.170 0.032 0.090 83 0a 16% 5% 31% nd 0.123 0.907 0.028 nd 60 Ob 17% 23% 14% 38% 0.248 1.538 nd 0.050 74 4a 35% 3% 48% 8% 0.113 3.756 0.043 nd 65 4b 16% 14% 17% 14% 0.157 2.697 nd 0.025 95 8a 16% 16% 27% 21% 0.132 2.694 0.033 nd 76 8b 21% 45% 17% 65% 0.296 3.450 nd 0.051 70 24a 2% 3% 4% 5% nd 2.072 nd 0.163 62 24b 15% 36% 15% 46% 0.364 4.115 nd 0.048 65 48a 18% 2% 26% 0% 0.131 1.714 nd nd 52 48b 40% 36% 29% 61% 0.296 2.705 0.246 0.460 60 72a 19% 2% 29% nd 0.109 1.241 nd nd 65 72b 9% 16% 9% 28% 0.279 1.558 nd 0.038 99

♦average of liver sample weights from each specimen group used for calculation of reagent blank adduct concentrations.

in Oi 06-M G (x3)

□ N7-MG ng/mg ng/mg liver tissue

N7-MG \ N 06-M G (x3) 8 24 Time Post-Exposure (hrs)

Fig. 4.5: Methylguanine adduct concentrations in liver DNA of western mosquitofish mid-way through (-1) and after 2 hr pulse exposure to 10 ppm MAM-Ac.

5 i 155 carcinogen. Although MAM-Ac has been shown in histopathology studies to be

carcinogenic for several small fish species, including mosquitofish, medaka, guppy,

sheepshead minnow, Gulf killifish, inland silverside, rivulus, and fathead minnow

(4.1, 4.7, 4.16, 4.18, 4.21, 4.29), there are no previous reports of in vivo alkylation

of fish liver DNA by this compound. Determination of baseline levels of DNA

adducts resulting from exposures to small alkylating carcinogens, such as

methylating or ethylating agents, are useful for mechanistic analyses of

carcinogenesis in alternative test animals such as small fish species (4.22, 4.25,

4.39).

Most studies available on carcinogen-DNA adduct formation in aquatic

species have relied upon the use of radiolabeled chemicals (4.33). However, studies designed to measure DNA adducts in whole animals on multiple tissues, especially involving previous and/or possibly unknown exposures to environmental mutagens or carcinogens, require alternative methods that do not involve pre-administration of radiolabeled compounds. Since stable isotope labeled compounds were used in the isotope dilution GC/MS method applied in these studies, this method could be used in both field and laboratory studies, allowing for direct measurements of environmental exposures. GC/MS with single ion monitoring is a highly versatile method which can unequivocally identify and quantify specific adducts at extremely low detection limits (femtomole levels) (4.33). A similar method was used by

Malins et al to detect hydroxyl radical-induced DNA adducts in liver neoplasms of feral English sole from Puget Sound (4.34, 4.35). 156 Other methods that have been used to measure adducts in aquatic species

include 32P-postlabeling, immunoassays, and HPLC with fluorescence detection (4.6,

4.33). 32P-postlabeling, while extremely sensitive, cannot identify the exact

chemical structures for specific adducts detected. Immunochemical methods can be

sensitive and very specific; however, they are by nature limited to the detection of

adducts against which specific monoclonal antibodies have been developed. In

addition, these two methods are not useful for detection of unknown adducts. A

newer immunochemical method in which specific DNA adducts are isolated by the

use of monoclonal antibodies and selective nitrocellulose membranes, then

quantitated by polymerase chain reaction, shows some promise (4.24).

MAM-Ac, the aglycone of cycasin, was chosen for this study because its

carcinogenic effects in non-human primates, rodents, and small fish species are

relatively well-characterized. It is a potent alkylating agent that can kill crucial,

dividing cells in developing embryos or neonates as well as cause neoplasia (4.27).

Further, the compound, like many environmental carcinogens, requires metabolic

activation for maximum carcinogenic effectiveness. Several target organs are involved in MAM-Ac carcinogenesis. Old World monkeys developed hepatocellular carcinomas, intrahepatic bile duct adenocarcinomas, renal carcinomas and adenomas, squamous cell carcinomas of the esophagus, adenocarcinomas of the small intestine, and adenomatous polyps of the colon in response to chronic MAM-Ac administration

(4.40). Rodents developed liver, kidney, and intestinal tract neoplasia (4.28, 4.49), along with central nervous system disorders and teratogenic effects such as 157 microencephaly (4.42). Metabolic activation of MAM-Ac in mammalian species is

thought to occur primarily through the action of NAD+-dependent alcohol

dehydrogenases (4.17).

Fish respond to aqueous MAM-Ac exposure with cytotoxic and neoplastic

lesions, the majority of which have been reported in the liver (4.21, 4.29).

However, exocrine pancreatic neoplasms (4.16) and eye neoplasms, thought to be

related to retinoblastoma of humans (4.20), have been reported. Liver lesions

include cytotoxicity, megalocytosis, spindle cell proliferation, foci of cellular

alteration, hepatocellular carcinomas, and cholangiocellular carcinomas (4.30).

Because of this background information, we chose to limit our adduct analyses to the

liver.

Table 4.3 and Fig. 4.5 show general trends of carcinogen-DNA adduct formation and repair with time following an aqueous pulse exposure to MAM-Ac.

Maximal concentrations of 7-methylguanine were detected in liver samples at 4 hr post-exposure (PE) and decline through the last sampling period, suggesting that repair of this lesion is initiated within 72 after cessation of exposure. As in previous studies with mammalian species (4.31, 4.45), 7-MG is probably not a[ critical mutagenic lesion in fish. However, its detection could be useful in biomonitoring studies since this adduct is formed early after exposure and at higher levels.

Concentrations of 0 6-methylguanine in DNA averaged 12% of the corresponding 7-MG concentrations in DNA at each of the 7 sampling times, which is similar to 0-MG:7-MG ratios reported in vitro for another methylating agent 158 (N-methyl-N-nitrosourea, 9 to 11%) by Lawley (4.31). Alkylation at the O6 position

is thought to have considerably greater biological consequences. This lesion fixes

guanine in an unfavorable tautomeric configuration that tends to base-pair with

thymine rather than cytosine, leading to transition mutations (4.45). Once formed,

O-MG adducts did not appear to undergo significant repair by 72 hr PE. Other

workers have also concluded that fish have relatively slow rates of repair of

alkylguanine adducts at the O6 position (4.33). Fong ef al (4.15) found in rainbow

trout exposed to 250 ppm diethylnitrosamine that, although Os-alkylguanine-DNA

alkyltransferase activity was present, the rate and extent of removal of O-EG was

low. They suggested that the enzyme was depleted rapidly and was probably not

induced in the liver. The same study demonstrated that 7-ethylguanine and O6-

ethylguanine were removed from liver DNA in a biphasic manner. That is, concentrations of these adducts decreased rapidly from 0 to 12 hr PE, but increased

slightly between 24 and 48 hr PE; no significant decreases in adduct concentrations were found from 12 to 96 hr after DEN exposure (4.15). Compounds that must be metabolically activated in vivo can show complicated kinetics with regard to DNA adduct formation and repair (4.3). Depending upon the exposure method, target organ enzyme levels, and depuration from tissue depots such as fat or blood, the opposing processes of adduct formation and repair may occur simultaneously.

Another complicating factor in MAM-Ac-induced carcinogenesis following a single pulse of the chemical is its dual proliferative and anti-proliferative effects.

Tumor initiation requires at least one round of cell proliferation for "fixation" of an 159 initiating event such as a mutation arising from DNA adduct formation (4.12, 4.36).

MAM-Ac caused nucleolar structural abnormalities as early as 15 min following

intravenous administration in rats (4.50, 4.51), suggesting that the compound may interfere with cell replication. Opposing this are the cytotoxic effects of MAM-Ac which stimulate cell proliferation in the liver through regenerative hyperplasia. In

mammalian systems, cell proliferation is no longer effective in tumor initiation after

96 hr and most lesions are repaired by that time (4.26, 4.48). In mosquitofish, therefore, the formation of critical mutagenic lesions such as alkylguanine adducts, followed by a round of cell proliferation, would likely occur before 96 hours for fixation of tumor initiation to occur. Alternatively, DNA repair mechanisms may be comparatively slower in this species. This awaits further investigation.

Considerable differences in concentrations of methylguanine adducts at 24 hr

PE were found in identically treated groups of fish in the first set of experiments

(Fig. 4.4). These differences may be attributed to method/instrumental error from both sample preparation/extraction and instrument variability, as well as to biological variability (i.e., individual fish will have different rates of carcinogen metabolism). Carcinogen-DNA adducts formed by aqueous exposure to the alkylating agent, MAM-Ac, were detected at femtomole levels in a small fish model using an isotope dilution GC/mass spectrometry method. Levels of O-MG ranging from approximately 55-185 pg (330-1100 femtomoles) per ng DNA, measured in the first 72 hr after exposure, were correlated with a 33% liver tumor incidence after 25 weeks in parallel experiments. Although further studies are needed with this and 160 other carcinogens representing different mechanistic classes and with other small fish

species, the method appears to be useful for studies of carcinogens and mutagens in

aquatic environments. These findings, along with additional mechanistic studies,

should also serve to further demonstrate the utility of the western mosquitofish as

both an environmental sentinel organism and as a model for carcinogenicity testing

in the laboratory.

References

4.1. Aoki K and Mastsudaira H (1977). Induction of hepatic tumors in a teleost (Oryzias latipes) after treatment with methylazoxymethanol acetate: brief communication. J. Natl. Cancer Inst. 59: 1747-1749.

4.2. Balsiger RW and Montgomery JA (1960). Synthesis of potential anticancer agents. XXV. Preparation of 6-alkoxy-2-aminopurines. J. Am. Chem. Soc. 25: 1573-1575.

4.3. Becker RA and Shank RC (1985). Kinetics of formation and persistence of ethylguanines in DNA of rats and hamsters treated with diethylnitrosamine. Cancer Res. 45: 2076-2084.

4.4. Beland FA and Poirier MC (1989). DNA adducts and carcinogenesis. In: The Pathobiology o f Neoplasia, AE Sirica (ed). Plenum Press, New York, pp. 57-80.

4.5. Beranek DT, Weis CC, and Swenson DH (1980). A comprehensive quantitative analysis of methylated and ethylated DNA using high pressure liquid chromatography. Carcinogenesis 1: 595-606.

4.6. Cadet J and Weinfeld M (1993), Detecting DNA damage. Analyt. Chem. 65: 675A-682A.

4.7. Couch JA and Harshbarger JC (1985). Effects of carcinogenic agents on aquatic animals: an environmental and experimental overview. Environ. Carcinogen. Revs. 3: 63-105.

4.8. de Series FJ (1988). Banbury Center DNA adducts workshop, meeting report. Mutat. Res. 203: 55. 161 4.9. Dees LT (1961). The mosquitofish, Gambusia affinis. Fishery Leaflet 525 (September, 1961), Bureau of Commercial Fisheries, Fish and Wildlife Service, U.S. Department of the Interior.

4.10. Dizdaroglu M (1993). Quantitative determination of oxidative base damage in DNA by stable isotope-dilution mass sprectrometry. FEBS Letters 315: 1-6.

4.11. Dresler SL (1989). DNA repair mechanisms and carcinogenesis. In: The Pathobiology o f Neoplasia, AE Sirica (ed). Plenum Press, New York, NY, pp. 173-197.

4.12. Farber E and Sarma DSR (1987). Hepatocarcinogenesis: a dynamic cellular perspective. Lab. Invest. 56: 4-22.

4.13. Farmer PB, Foster AB, Jarman M, and Tisdale MJ (1973). The alkylation of 2’-deoxyguanosine and of thymidine with diazoalkanes. Biochem. J. 135: 203-213.

4.14. Feinberg A and Zedeck MS (1980). Production of a highly rective alkylating agent from the organospeciflc carcinogen methylazoxymethanol by alcohol dehydrogenase. Cancer Res. 40: 4446-4450.

4.15. Fong AT, Hendricks JD, Dashwood RH, Van Winkle S, and Bailey GS (1988). Formation and persistence of ethylguanine in liver DNA of rainbow trout ( Salmo ga.ird.neri) treated with diethylnitrosamine by water exposure. Food Chem. Toxicol. 26: 699.

4.16. Foumie JW, Hawkins WE, Overstreet RM, and Walker WW (1987). Neoplasms of the exocrine pancreas in guppies (Poecilia reticulata) after exposure to methylazoxymethanol acetate. J. Natl. Cancer Inst. 78: 715-725.

4.17. Grab DJ and Zedeck MS (1977). Organ-specific effects of the carcinogen methylazoxymethanol related to metabolism by nicotinamide adenine dinucleotide-dependent dehydrogenases. Cancer Res. 37: 4182-4189.

4.18. Harada T and Hatanaka J (1988). Liver cell carcinomas in the medaka (Oryzias latipes) induced by methylazoxymethanol acetate. J. Comp. Path. 98: 441-452.

4.19. Harris CC (1991). Chemical and physical carcinogenesis: advances and perspectives for the 1990s. Cancer Res. 51 (Suppl.): 5023s-5044s. 162 4.20. Hawkins WE, Foumie JW, Overstreet RM, and Walker WW (1986). Intraocular neoplasms induced by methylazoxymethanol acetate in Japanese medaka (Oryzias latipes). J. Natl. Cancer Inst. 76: 453-365.

4.21. Hawkins WE, Overstreet RM, Foumie JW, and Walker WW (1985). Development of aquarium fish models for environmental carcinogenesis: tumor induction in seven species. J. Appl. Toxicol. 5: 261-264.

4.22. Hawkins WE, Overstreet RM, and Walker WW (1988). Carcinogenicity tests with small fish species. Aquatic Toxicol. 11: 113-128.

4.23. Hawkins WE, Overstreet RM, and Walker WW (1988). Small fish models for identifying carcinogens in the aqueous environment. Water Resources Bulletin 24: 941-949.

4.24. Hochleitner K, Thomale J, Yu.Nikitin A, and Rajewsky MF (1991). Monoclonal antibody-based, selective isolation of DNA fragments containing an alkylated base to be quantified in defined gene sequences. Nucleic Acids Res. 19: 4467-4472.

4.25. Hoover KL (1984) Use o f Small Fish Species in Carcinogenicity Testing, Natl. Cancer Inst. Monogr. 65, NIH Publication No. 84-2653, National Cancer Institute, Bethesda, Maryland.

4.26. Ishikawa T, Takayama S, and Kitagawa T (1980). Correlation between time of partial hepatectomy after a single treatment with diethylnitrosamine and induction of adenosine triphosphatase-deficient islands in rat liver. Cancer Res. 40: 4261-4264.

4.27. Johnston MV, Grzanna R, and Coyle JT (1979). Methylazoxymethanol treatment of fetal rats results in abnormally dense noradrenergic innervation of neocortex. Science 203: 369-371.

4.28. Laquer GL and Spatz M (1968). Toxicology of cycasin. Cancer Res. 28: 2262-2267.

4.29. Law JM, Hawkins WE, Overstreet RM, and Walker WW (1994). Hepatocarcinogenesis in western mosquitofish ( Gambusia affinis) exposed to methylazoxymethanol acetate. 7. Comp. Path. 110: 117-127.

4.30. Law JM, Hawkins WE, Taylor HW, Overstreet RM, and Walker WW (1995). Cell proliferation, p53, and apoptosis in hepatocarcinogenesis of western mosquitofish (Gambusia affinis) exposed to methylazoxymethanol acetate. Toxicol. Pathol, (submitted). 163 4.31. Lawley PD (1984). Carcinogenesis by alkylating agents. In: Chemical Carcinogens, CE Searle (ed). American Chemical Society Monographs, Washington, DC, pp. 325-484.

4.32. Loveless A (1969). Possible relevance of 0-6 alkylation of deoxyguanosine to the mutagenicity and carcinogenicity of nitrosamines and nitrosamides. Nature 223: 206-207.

4.33. Maccubbin AE (1994). DNA adduct analysis in fish: laboratory and field studies. In: Aquatic Toxicology: Molecular, Biochemical, and Cellular Perspectives, DC Malins, GK Ostrander (eds). Lewis Publishers, Boca Raton, Florida, pp. 267-294.

4.34. Malins DC (1993). Identification of hydroxyl radical-induced lesions in DNA base structure: biomarkers with a putative link to cancer development. J. Toxicol. Environ. Health 40: 247-261.

4.35. Malins DC, Ostrander GK, Haimanot R, and Williams P (1990). A novel DNA lesion in neoplastic livers of feral fish: 2,6-diamino-4-hydroxy-5-formamidopyrimidine. Carcinogenesis 11: 1045-1047.

4.36. Maronpot RR (1991). Chemical carcinogenesis. In: Handbook o f Toxicologic Pathology, WM Haschek, CG Rousseaux (eds). Academic Press, Inc., San Diego, pp. 91-130.

4.37. McCarthy JF and Shugart LR (1990). Biomarkers o f Environmental Contamination, Lewis Publishers, Boca Raton, Florida.

4.38. Miller JH (1983). Mutational specificity in bacteria. Amu. Rev. Genet. 17: 215-238.

4.39. Shugart L, Bickham J, Jackim G, et al. (1992). DNA alterations. In: Biomarkers: Biochemical, Physiological, and Histological Markers o f Anthropogenic Stress, RJ Huggett, RA Kimerle, PM Mehrle, Jr., HL Bergman (eds). Lewis Publishers, Boca Raton, Florida, pp. 125-154.

4.40. Sieber SM, Correa P, Dalgard DW, Mclntire KR, and Adamson RH (1980). Carcinogenicity and hepatoxicity of cycasin and its aglycone methylazoxymethanol acetate in nonhuman primates. J. Natl. Cancer Inst. 65: 177-183.

4.41. Singer B and Kusmierek JT (1982). Chemical mutagenesis. A m u. Rev. Biochem. 52: 655-693. 164 4.42. Spatz M (1969). Toxic and carcinogenic alkylating agents from cycads. Annals o f the New York Academy o f Science 163: 848-859.

4.43. Stanton MF (1966). Hepatic neoplasms of aquarium fish exposed to Cycas circinalis. Federation Proceedings 25: 661.

4.44. Strauss WM (1987). Preparation of genomic DNA from mammalian tissue. In: Current Protocols in Molecular Biology, FM Ausubel, R Brent, RE Kingston (eds). John Wiley and Sons, Inc., Media, Pennsylvania, pp. 2.2.1-2.2.3. (vol. 1).

4.45. Swenson DH (1983). Significance of electrophilic reactivity and especially DNA alkylation in carcinogenesis and mutagenesis. In: Developments in the Science and Practice o f Toxicology, AW Hayes, RC Schnell, TS Miya (eds). Elsevier Science Publishers B.V., London, pp. 247-254.

4.46. Swenson DH and Lawley PD (1978). Alkylation of deoxyribonucleic acid by carcinogens dimethyl sulphate, ethyl methanesulphonate, N-ethyl-N-nitrosourea, and N-methyl-N-nitrosourea. Biochem. J. 171: 575-587.

4.47. Wogan GN (1988). Detection of DNA damage in studies on cancer etiology and prevention. In: Methods for Detecting DNA Damaging Agents in Humans: Applications in Cancer Epidemiology and Prevention, H Bartsch, K Hemminki, IK O’Neill (eds). IARC Scientific Publication No. 89, International Agency for Research on Cancer, Lyon, France, pp. 32-51.

4.48. Ying TS, Enomoto K, Sarma DSR, and Farber E (1982). Effects of delays in the cell cycle on the induction of preneoplastic and neoplastic lesions in rat liver by 1,2-dimethylhydrazine. Cancer Res. 42: 876-880.

4.49. Zedeck M and Sternberg SS (1977). Tumor induction in intact and regenerating liver of adult rats by a single treatment with methylazoxymethanol acetate. Chemico-Biological Interactions 17: 291-296.

4.50. Zedeck MS, Sternberg SS, McGowan J, and Poynter RW (1972). Methylazoxymethanol acetate: induction of tumors and early effects on RNA synthesis. Federation Proceedings 31: 1485-1492.

4.51. Zedeck MS, Sternberg SS, Yataganas X, and McGowan J (1974). Early changes induced in rat liver by methylazoxymethanol acetate: mitotic abnormalities and polyploidy. J. Natl. Cancer Inst. 53: 719-724. CHAPTERS. SUMMARY AND CONCLUSIONS

Introduction

"holistic \hd-’lis-tik\ adj (1926) I: of or relating to holism 2: relating to or concerned with wholes or with complete systems rather than with the analysis of, treatment of, or dissection into parts < * medicine attempts to treat both the mind and the body> < * ecology views man and the environment as a single system > " (5.7).

It is obvious that simply looking at the histopathology of chemically-induced tumors in small fish species, which has already received much attention in the recent scientific literature, would not add substantially to an understanding of the carcinogenic process. The process should be viewed as a continuum or whole.

When investigating the underlying mechanisms of a pathophysiological process such as neoplasia, the logistics of scientific investigation necessarily require artificial division of the process into steps or stages so that each one may be thoroughly examined. However, successful integration of these stages into a logical process is a different ball game. It requires stepping back and seeing inter-relationships that are sometimes unique to biological systems. This is where a holistic approach comes into play. Drawing upon several different disciplines, we investigated the initial stages of genetic damage by measuring the formation and loss of DNA adducts; next, early cellular/tissue reactions to such damage were studied through the use of histopathology and immunohistochemistry of cell proliferation; mutations in a critical tumor suppressor gene were assessed by the immunohistochemical detection of p53 overexpression; and, finally, tumor progression/development were characterized using histopathology and electron microscopy.

165 166 Summary

Mechanisms of chemically-induced carcinogenesis in the liver of a small fish model were explored using a model alkylating carcinogen, methylazoxymethanol acetate (MAM-Ac). The experiments were designed to examine hepatocarcinogenesis in western mosquitofish (Gambusia affinis) at several stages:

Neoplasm development

In the first experiment, laboratory-reared adult mosquitofish were exposed in the ambient water to 10 ppm MAM-Ac for 2 hours and then examined periodically for the onset of neoplastic lesions using light and electron microscopy.

Approximately 33% of the exposed fish developed liver neoplasms within 25 weeks of exposure and 52% within 40 weeks. The lesions were predominantly hepatocellular carcinomas and cholangiocarcinomas, many of which were highly scirrhous and/or anaplastic. No neoplastic lesions were found in other organs. This was apparently the first published report of neoplasia in this particular species.

Cytotoxicity and early markers of neoplasia

The next set of experiments involved a similar exposure to MAM-Ac, only this time on a much larger scale. Approximately 600 juvenile and adult, laboratory- reared specimens of Gambusia were examined at 8 sampling times which ranged from 24 hours to 41 weeks after exposure. Early changes included cytotoxicity, apoptosis, and proliferation of a spindle-shaped cell, believed to be of stem cell origin. Cell proliferation was measured using PCNA immunohistochemistry; over- expression of p53 was detected with immunohistochemistry as well. As in the first 167 experiment, hepatocellular carcinomas and cholangiocellular carcinomas were found.

However, a spindle cell neoplasm which stained positively for cytokeratins and often occurred in conjunction with ductular profiles, also occurred in many of the exposed fish. Only one neoplasm and one altered focus were detected in unexposed controls.

DNA adducts

The third set of experiments involved development and application of an isotope dilution GC/mass spectrometry method for measurement of DNA adducts in the liver of western mosquitofish using an exposure methodology designed to mimic that which produced high levels of liver tumors within 40 weeks. We hypothesized that formation and persistence of high levels of 0 6-methylguanine in the liver would lead to mutations and eventual neoplasia. Standards for 0 6-methyl and ethylguanines, 7-methyl and ethylguanines, and their deuterium-labeled analogs were synthesized here in our laboratories at LSU and their purities confirmed with HPLC,

IR, UV, and GC/MS. Several methods of DNA isolation, DNA hydrolysis, and derivatization were attempted before a satisfactory methodology was achieved.

After many trials, phenol/chloroform DNA isolation, formic acid hydrolysis, and trimethylsilyl derivatization methods were determined to be the best.

Fish were exposed to 10 ppm MAM-Ac for 2 hours and sampled at several intervals during and after exposure for up to 72 hours post-exposure. Concentrations of O6- methylguanine in DNA averaged 12% of the corresponding 7-methylguanine concentrations at each sampling period, which is similar to the levels reported in 168 rodent studies with other methylating agents. However, the levels of 7-

methylguanine declined rapidly, suggesting that this lesion is repaired quickly. O6-

methylguanine, on the other hand, appeared to be repaired much more slowly and

may also have greater biological consequences; concentrations ranging from approximately 55-185 ng 0 6-methylguanine per mg DNA were correlated with greater than a 50% liver tumor incidence by 40 weeks after exposure in parallel histopathology studies.

Conclusions

A proper discussion of the conclusions should address the stated objectives:

Carcinogen sensitivity

These experiments suggest that the western mosquitofish (Gambusia affinis) has good hepatic carcinogen sensitivity. The response of Gambusia was equal to or greater than that of seven aquarium fish species exposed to MAM-Ac under similar conditions, including medaka, guppy, sheepshead minnow, Gulf killifish, inland silverside, rivulus, and fathead minnow (5.3, 5.6). The carcinogen sensitivity of

Gambusia affinis, together with its widespread occurrence in warmwater environments, low background incidence of tumors, apparent disease resistance

("hardiness"), and ease of laboratory culture, would facilitate the direct validation of field studies by enabling substances to be tested under both laboratory and field conditions, thus making this fish a good candidate for an environmental sentinel species. 169 Histopathology and immunohistochemistry

The neoplasms induced in these experiments were histologically similar to those described in carcinogenicity studies with other fish species (5.2, 5.3, 5.5).

The foci of cellular alteration observed in the present study resembled those described in rats (5.8, 5.11) and rainbow trout (5.5) exposed to hepatocarcinogens.

These lesions are considered to be pre-neoplastic foci in fish and probably do not regress as do many of the altered foci seen in rodent studies since fewer foci are found in fish liver and they are usually correlated with the same number of neoplasms. Although both the mosquitofish and the guppy (Poecilia reticulata) belong to the Family Poeciliidae, the response of the two species to MAM-Ac differs. In addition to having hepatic neoplasms, approximately 9% of guppies exposed to 2, 4, or 10 mg/L MAM-Ac for 2 h developed exocrine pancreatic adenomas, acinar cell carcinomas, or adenocarcinomas (5.1). In contrast, less than

1 % of the mosquitofish in the present study developed pancreatic neoplasms. The relatively high incidence of hepatic lesions in the mosquitofish, nonetheless, was similar to that found in the guppy (5.4). Additionally, it is possible that if these experiments were repeated but the fish were allowed to grow out for a longer period

(e.g., a full year), a higher incidence of pancreatic and other non-hepadc lesions would be found.

The spindle cell lesions in the present study deserve special mention. Since these cells stained positively for cytokeratins and often occurred in conjunction with ductular profiles, they were thought to be of liver stem cell origin. Many new 170 findings on liver stem cells are being reported, and much controversy still exists.

However, agreement with Dr. Stewart Sell’s hypothesis that these tumors can

undergo various levels of "maturation arrest" and give a variable histological

appearance seems to be warranted (5.9, 5.10). Obviously, much more work needs

to be done before any of this can be proven in fish species.

PCNA immunohistochemistry proved to be quite valuable in determining

which cell types were actively proliferating in a lesion (sometimes with surprising

results) and in correlating proliferation rates among different experimental groups

with other histopathological parameters. The fact that PCNA labeling indices

increased dramatically in concert with cytotoxicity is consistent with the hypothesis

that regenerative cell proliferation provides the necessary round of cell proliferation

for fixation of DNA damage produced by a genotoxic carcinogen. The young,

growing fish in the expanded study already had the necessary proliferation of

hepatocytes. Thus, a lower dosage of MAM-Ac (1 ppm) was able to produce

tumors. The data generated in this study should serve as a useful baseline for future

comparisons of cell proliferation. p53 is currently a "hot topic" in molecular biology. However, the utility of p53 immunohistochemistry in these experiments was less clear. Positive staining over background was achieved on exposed animals versus unexposed controls, but correlation with eventual tumor formation was difficult. That is, it did not appear to have positive predictive value in this species under these conditions. Nonetheless, the fact that p53 protein was detected at all in a small fish species using immunohistochemistry was gratifying, since there are no 171 previous reports of this. Further work in this field is needed in order to fully assess

this important mechanism.

Inter-relationships between ceil proliferation, p53 overexpression, and apoptosis

Apoptosis should also be mentioned, not only because it too is a current "hot topic," but also because it is essentially linked to p53 and cell proliferation. One of the objectives of this study was to explore the inter-relationships between these three phenomena. In these fish specimens, high levels of what appeared by light microscopy to be eosinophilic cytosegrosomes or apoptotic bodies were present in liver sections of exposed fish. This process should be regarded as a type of individual cell death — a death pathway invoked in cells not so completely blasted by the alkylating agent so as to cause outright necrosis, but having sustained enough genetic damage to need to perish for the good of the organism. Since many different phenotypic subtypes of hepatocytes exist at any one time in the liver and in different locations with regard to blood supply, it is expected that some would undergo necrosis, some would undergo apoptosis, and some would survive.

However, with such a high level of apoptosis occurring at one time in the exposed specimens, it is not surprising that the process would fail to occur in some of the cells that it should have (e.g., those with p53 mutations or ineffective DNA repair mechanisms). Hence, the hypothesis that p53 mutations may lead to failure of apoptosis and subsequent uncontrolled cell proliferation appears to hold true for this species under these conditions. 172 DNA adducts

Isotope dilution GC/mass spectrometry, while not perfect, is a highly sensitive and specific method for detecting and quantifying DNA adducts. With

GC/MS, sensitivities equal to or greater than the most common methods in current use such as 32P-postlabeling, HPLC with fluorescence detection, or radioimmunoassay were achieved. Additionally, GC/MS has the potential to detect multiple adducts simultaneously, or to even identify unknown adducts, which could be quite useful in field studies.

Although 7-methylguanine was formed at higher levels in the liver of mosquitofish exposed to MAM-Ac, it appeared to be rapidly repaired and of less biological consequence. 0 6-methylguanine was formed at much lower levels, but appeared to be repaired much more slowly in fish. With its tendency to form point mutations (C to T transitions), Ofi-methylguanine is highly significant in hepatocarcinogenesis of fish exposed to alkylating agents.

Personal Recommendations for Future Work

One of the responsibilities that goes along with scientific writing is communicating to future workers some of the pitfalls encountered so that they may avoid repeating the same mistakes. Little scientific progress would be made without this contribution. What could have been done better in this study?

For the histopathology studies, a better fixative could have been used for the specimens. Zinc formalin is good for immunohistochemistry, but some cellular 173 detail was lost which compromised the clarity of the photomicrographs. An ideal

fixative would be good for both.

I would like to have had some known positive control tissues for the p53

immunohistochemistry, preferably with fish tissues. However, I do not know if

these are currently available. Experimentation with different fixatives might solve

the dilemma over nuclear versus cytoplasmic staining against p53. I would also like

to have had time to try more dose levels for both p53 and PCNA, so as to develop a

dose-response type curve. Better planning at the beginning, particularly in the presence of a statistician, could have saved me a lot of time. Future work should include a range of doses, representatives from several carcinogen classes, and several small fish species. The medaka, in particular, is the "gold standard" against which all other small fish species work should be compared.

Regarding the DNA adduct work, obviously I would like to have had more successful trials so as to be able to get better at quantifying the adducts. Two sets of data do not allow for very good statistical correlations. However, much time and effort went into the "development" stage of this method. My first goal was just to be able to detect the alkylguanine adducts. When the first experiment with live fish worked so well, we were elated. After that, we had a series of set-backs, which were attributed mainly to improper lyophilization of the hydrolyzed DNA bases (of this I’m not so sure). After numerous attempts at more complicated solutions, a conversation with Larry Reily led to a very simple one; that is, since the 60% 174 formic acid solution had too low of a freezing point for easy freeze-drying, all we had to do was add water to each sample vial to raise the freezing point substantially.

In the Appendix, several examples of pitfalls in the GC/MS method for DNA adduct detection are given. Don't repeat our mistakes! This is a potentially valuable method, but it can be very tedious and frustrating at times. For example, proteinase-K was left out of the DNA isolation step during one experiment, which apparently left too much protein in the sample. Not only did these samples not get derivatized properly, but they seemed to "gum up" the GC column.

Cleanliness/thoroughness of each prep, proper ramp temperatures on the GC, and repeated injections of TMS-derivatized standards to obtain consistency of active column sites seem to be very important with this method.

Several other things may improve the utility and precision of this method.

The deuterium-labeled internal standard could be added earlier in the sample prep to get a truer estimate of recovery, possibly just after the liver tissue is pulverized in the sample tube and before the organic extraction. It remains to be seen if the standards will survive the formic acid hydrolysis. Perhaps a better way to express the concentrations of alkylguanine adducts may be nanograms of adduct per microgram of guanine present in each sample; this expression would normalize the data and would not be affected by the imprecise weighing of wet liver tissue. Other goals to strive for include better recovery, better peak separation, and at least 3 confirming ions on each sample. Lastly, for future work, this method should be attempted with other fish species, including the medaka (

5.1. Foumie JW, Hawkins WE, Overstreet RM, and Walker WW (1987). Neoplasms of the exocrine pancreas in guppies (Poecilia reticulata) after exposure to methylazoxymethanol acetate. J. Natl. Cancer Inst. 78: 715-725.

5.2. Harada T and Hatanaka J (1988). Liver cell carcinomas in the medaka (Oryzias latipes) induced by methylazoxymethanol acetate. J. Comp. Path. 98: 441-452.

5.3. Hawkins WE, Overstreet RM, Foumie JW, and Walker WW (1985). Development of aquarium fish models for environmental carcinogenesis: tumor induction in seven species. J. Appl. Toxicol. 5: 261-264.

5.4. Hawkins WE, Overstreet RM, and Walker WW (1988). Carcinogenicity tests with small fish species. Aquatic Toxicol. 11: 113-128.

5.5. Hendricks JD, Meyers TR, and Shelton DW (1984). Histological progression of hepatic neoplasia in rainbow trout ( Salmo gairdneri). Natl. Cancer Inst. Monogr. 65: 321-336.

5.6. Law JM, Hawkins WE, Overstreet RM, and Walker WW (1994). Hepatocarcinogenesis in western mosquitofish (i Gambusia affinis) exposed to methylazoxymethanol acetate. J. Comp. Path. 110: 117-127.

5.7. Mish FC (1986) Webster's Ninth New Collegiate Dictionary, Merriam-Webster Inc., Springfield, Massachusetts.

5.8. Popp JA and Goldsworthy TL (1989). Defining foci of cellular alteration in short-term and medium-term rat liver tumor models. Toxicol. Pathol. 17: 561-568.

5.9. Sell S (1990). Is there a liver stem cell? Cancer Res. 50: 3811-3815.

5.10. Sell S (1993). Cellular origin of cancer: dedifferentiation or stem cell maturation arrest? Environ. Health Perspec. 101 (Suppl 5): 15-26.

5.11. Squire RA and Levitt MH (1975). Report of a workshop on classification of specific hepatocellular lesions in rats. Cancer Res. 35: 3214-3223. APPENDIX A. DATA FROM DEVELOPMENT OF DNA ADDUCT QUANTIFICATION METHOD USING GC/MASS SPECTROMETRY

176 □ 06M G y = 1.095x + 0.012 r2 = 0.999 o N7MG y = 1.071x + 0.128 r2 = 0.998 o 06EG y = 0.719x + 0.041 r2 s 0.999 A N7EG y = 1.220x + 0.001 r2 = 1.000

1 1 1------0 5 10 15 20

AMOUNT RATIO (ng adduct/ng deuterated adduct)

Fig. A.I. Standard curve for adducts with duplication at nine levels. _ 06- & N7-METHYLGUANINE 06- & N7-ETHYLGUANINE Ion 294.00 {293.70 to 294.70): Abung^ijige WfBPSfiS6, Ion 308.00 (307.70 to 308.70): 21/^91 / 22.50

f ^ ------/V n n - 1 1 1 1 1 1 1 1 1 1 I1 ' 1 - J — " 1 rime--:21 .50 22.00 22.50 rime--^ 2 .00 22.50 23.00 Ion 309.00 (308.70 to 309.70): Ion 323.00 (322.70 to 323.70): Abun?IS§e

21.91 22.51 n - « « • « | 4 1 1 ‘ “ ' I 1 1 I n - f * • • 1 » 1 1 ■» I ' I 1 " l ..f ritne--:21 .50 22.00 22.50 Time--^2 .00 22.50 23.00 d3-06- & d3-N7-METHYLGUANINE d5-06- & d5-N7-ETHYLGUANINE Ion 297.00 (296.70 to 297.70): Ab^g|9 £ e Ion 313.00 (312.70 to 313.70): 2lfl77 22.26 22A77 j \ 22.48 ____ nu - 1 i i j f l i i | i i i i * un . i ki i l j----1 i ,r—r 1V ----j----1----i—l----1---- rime--:21 50 22.00 22.50 rim e--£2 .00 22.50 23.00 Ion 312.00 (311.70 to 312.70): Sbj9 gi9£e Ion 328.00 (327.70 to 328.70): 22.27 21.77 A 22.76 n n . .— —/ >—. /V U ^ 1 1 I | I ■ T---1-' I rjl ^ ---1---1---1--- u l . i l J----1---i * -1 *" i -i— r----1----1----1---- rime--:21 .50 22.00 22.50 Time---2.2 .00 22.50 23.00

Fig. A.2. Selected ion chromatograms from a typical DNA adduct analysis via GC/MS of liver from a mosquitofish after exposure to 10 ppm MAM-Ac. Note that the biological samples give broader peaks with more background "noise" than the standards shown in Chapter 4.

00 TIC: 4178CG2S.D

Guanine Adducts(TMS)

3 .5 e+ 0 7 Adenine(TMS)

3e+07 Guanine(TMS)

Thymine(TMS) 2 . 5e+07

2e+07 Cytosine(TMS)

1 .5 e+ 0 7

rim e — >

Fig. A.3. Total ion chromatogram (TIC) from a typical DNA adduct analysis via GC/MS of liver from a mosquitofish experiment. This demonstates that multiple purine and pyrimidine adducts could be monitored simultaneously from one sample, provided that suitable standards were available. APPENDIX B. LETTER OF PERMISSION 181

H jf to u r t Brace & Company Limited HARCOURT 24*28 O j I RojJ BRACE London NVC'I 7DX Tel 071*267 4466 Fax 0 '1 -482 2293 Si 071*484 4752

2 September 1994

J McHugh Law, DVM Louisiana State U niversity Dept of Veterinary Pathology School of Veterinary Medicine Baton Rouge Louisiana 70803 USA

Dear Mr Law

R e . "Hepatocarcinogenesis in Western Mosquitofish (Gambusia affinis) Exposed to Methylazoxymethanol Acetate" i n t h e JOURNAL OF COMPARATIVE PATHOLOGY, 9 3 : 189 (1994 VOl 110, 117-127)

Thank you for your le tte r of 11 August requesting permission to use the above m aterial from the Journal of Comparative Pathology. We are happy to grant permission for this use of your material provided that (1) complete credit is given to the source, including the Academic Press copyright notice (2) the m aterial to be used has appeared in our publication without credit or acknowledgment to another source and (3) if commercial publication should resu lt, you must contact Academic Press again.

He realize that University Microfilms must have permission to sell copies of your thesis, and we agree to this. He would point out, however, that this does not apply to separate sale of your a r t i c l e .

Thank you for approaching uc in this m atter.

Yours sincerely

CHRISTINE MORTON (MS) ff Rights/Perm issions Manager AP London and Bailli& re Tindall

Rffiurrtd No. 59HSN A Division o f Harcourt Brace & Company VITA

Jerry McHugh ("Mac") Law, Jr. was bom on April 15, 1959 in Baton

Rouge, Louisiana to Jerry and Helen Law as the youngest of three children. Mac graduated from Robert E. Lee High School in Baton Rouge in 1977. He received the Doctor of Veterinary Medicine (DVM) degree in 1985 from Louisiana State

University and entered private practice in the Houston, Texas area where he practiced companion animal veterinary medicine for five years. He is married to

Christine Behrmann Law and has a daughter, Rebecca Marion Law.

182 DOCTORAL EXAMINATION AND DISSERTATION REPORT

Candidate: Jerry McHugh Law, Jr.

Major Field: Veterinary Medical Sciences

Title of Dissertation: Mechanisms of Chemically-induced Hepatocarcinogenesis in Western Mosquitofish (Gambusia affin is)

Approved:

j or Pifofess'or and Chairman

Dean of th< iduate School

EXAMINING COMMITTEE:

y Co-Chairman

Date of Examination:

March 3, 1995