Advances in Cancer Research Volume 43

ADVANCESIN CANCER RESEARCH VOLUME 43

Contributors to This Volume D. Cohen

Philip Noguchi

David Colcher

Sidney Pestka

Paul B. Fisher

Varda Rotter

Judah Folkman

Volker Schirrmacher

John Greiner

Jeffrey Schlom

Patricia Horan Hand

Robert Szigeti

Donald Kufe

Maureen Weeks

Karin Moelling

David Wolf

David Wunderlich

ADVANCES IN CANCERRESEARCH Edited by

GEORGE KLEIN Department of Tumor Biology Karolinska lnstitutet Stockholm, Sweden

SIDNEY WEINHOUSE Fels Research Institute Temple University Medical School Philadelphia, Pennsylvania

Volume 43- 1985

ACADEMIC PRESS, INC. (Harcourt Brace Jovanovich, Publishers) Orlando San Diego New York London Toronto Montreal Sydney Tokyo

COPYRIGHT @ 1985, BY ACADEMIC PRESS. INC. ALL RIGHTS RESERVED. NO PART OF THIS PUBLICATION MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM OR BY ANY MEANS, ELECTRONIC OR MECHANICAL, INCLUDING PHOTOCOPY, RECORDING. OR ANY INFORMATION STORAGE AND RETRIEVAL SYSTEM, WITHOUT PERMISSION IN WRITING FROM THE PUBLISHER.

ACADEMIC PRESS, INC. Orlando. Florida 32887

United Kin dorn Edition ublished by

ACADEM~C PRESS I&. (LONDON) LTD. 24-28 Oval Road, London NW I 7DX

LIBRARY OF CONGRESS CATALOG CARDNUMBER: 52-1 3 3 6 0 I S B N 0-12-006643-2 FWNTED IN THE UNITED STATES OF AMERICA

85 86 87 88

9 8 7 6 5 4 3 2 I

CONTENTS

CONTRIBUTORS TO VOLUME 43 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

ix

Cancer Metastasis: Experimental Approaches, Theoretical Concepts, and Impacts for Treatment Strategies VOLKER SCHIRRMACHER Introduction . . . . . . . . . . . . . . . . , , . . . . . . . . . . . . . . . , , , . , , . , . , , . , , . , , 11. Metastasis Research. . , . . , . . , . ...................t. 111. Cascade Theory of Metastasis . . . . . , . , . . . . . . . . . . . . . . . . . . . , . , , . , , . , . , , , , , IV. Instability, Subpopulation Interactions, and Tumor Progression: Experimental Findings and Theoretical Concepts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , . , . . , . . V. Angiogenesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VI. Cancer Invasion ........................... . . . . . . . . . . . VII. Tumor Dormancy. . . . , . . . . . . , , . . . . . . . . . . . . . . . . . . . . . . . . . . . . , . VIII. Host Immune Responses in Metastasis . ........................ Ix. Impacts of Experimental Studies on Ca ment Strategies . , . X. Summary and Conclusions . . . . . . . . . . . . I . . I . . . References. . . . . . . . . . . . . . . . . . . , . . , , . . . . . . . . . . . . . . . . . . , . , , . , , . , . , . , . , . ,

1 4 14

23 33 36 46 48

55 63 64

The Canine Transmissible Venereal Tumor: A Unique Result of Tumor Progression D. COHEN I. . . . . . . . . . . . . . . 75 11. . ............................. 78 111. .............................. 80 IV. ................................. 82 V. ntation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 VI. The Immune Response against the TVT 86 VII. Etiology and the Mechanism of Transmission. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 VIII. 103 ....................... 107

V

vi

CONTENTS

Biological and Molecular Analysis of p53 Cellular-Encoded Tumor Antigen VAHD.4

ROTTER A N D DAVID W0l.F

I . Introduction I1. p53 Is Complexed with Viral-Encoded Protein . . . . . . . . . . . . . . . . . . . . . . . . . . . .

111. Expression of p53 in Nontransformed Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

IV . V. VI . VII . VIII . IX .

p53 Is Immunogenic . . . . . . . . . . . . . . ................................ Chromosomal Assignment of p53 Gene . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Subcellular Localization of p53 in Transformed and Nontransformed Cells The Use of p53 as a Marker for Cell Transformation . . . . . . . . . . . . . . . . . . . . . . . The Possible Function p53 Fulfills in Transformed Cells Molecular Mechanism Controlling the Expression of p53 . . . . . . . . . . . . . . . . . . . References . . . . . ...................................

113 114 116 117 120 123 127 130 133 139

Monoclonal Antibodies Reactive with Breast Tumor-Associated Antigens SCIILOM. DAVIDCOLCHER. PATRICIA HOHANHAND.JOllN GHEINER. DAVID WUNDERLICH. MAUREEN WEEKS. P A U I. B . FISIIER. PHILIPNocuclrI . S I D N E Y PESTKA. A N D DONALD KUFE

JEFFREY

I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... I1. Monoclonal Antibodies to Human Breast Carcinomas . . . . . . . . . . . . . . . . . . . . . 111. Generation and Characterization of Monoclonal Antibodies . . . . . . . . . . . . . . . . IV . Differential Reactivity of a Monoclonal Antibody (DF3) with Human

143 144 147

.... ....

....

151 155 160

....

161

....

166 169 171

Malignant versus Benign Breast Tumors . . . . . . . . . . . . . . . . . . . . . . . . . . . . V . Use of Radiolabeled Monoclonal Antibodies for Tumor Detection VI . Use of Monoclonal Antibodies in Tumor Therapy .................... VII . Antigenic Modulation and Evolution within Human Carcinoma Cell Populations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VIII . Use of Biological Response Modifiers to Enhance Detection of Human Carcinoma Antigens by Monoclonal Antibodies ......... IX . Future Directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . .

JUDAH

I. I1. I11. IV . V.

.... ....

FOLKMAN

......................................................... Methods for Studying Angiogenesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A Capillary Grows by Sequential Steps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Angiogenesis Is a Preneoplastic Marker . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Solid Tumors Are Angiogenesis Dependent ..............................

175 176 178 180 181

vii

CONTENTS

The Vascularized Tumor Continues to Alter Its Blood Supply., . , , . , . . . . . . . . Mast Cells and Heparin Can Potentiate Tumor Angiogenesis . . Angiogenesis Can Also Be Induced by Certain Nonmalignant Cells. . . . . . . . . . Angiogenic Factors and Endothelial Mitogens Have Been Isolated from Tumors and from Some Nonneoplastic Cells.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . X. Angiogenesis Inhibitors Are Found in Natural Sources XI. Role of Angiogenesis in Clinical Oncology. . . . . . . . . XII. Summa. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References. . .......................................

VI. VII. VIII. IX.

183 184 185 187

199

Fusion Proteins in Retroviral Transformation KARIN MOELLINC I. Introduction

..., .., ., . .. , ..... ... .. . .. . .. ..... .. ... . .... ... ... .. . ....

11. Fusion Proteins of Avian Retroviruses . . . . , . . . . . . . . . . . . . . . . . . . . , . . . . . . . . . 111. Fusion Proteins of Mammalian Viruses . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV. Discussion on the Effect of Viral Structural Proteins on Transformation . . . . . .

References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

205 209 227 233 234

Application of Migration Inhibition Techniques in Tumor Immunology ROBERTSZIGETI I. Introduction 111. IV. V. VI.

. .. . . . ..

..........................

Assay Systems and Techniques . . . . . . . . . ..................... Applications in Animal Tumors.. . . . . . . . . . . . . . Applications in Human Tumors.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Concluding Re ,......... References. . . . ................................

INDEX ..................................................................... CONTENTS OF RECENT VOLUMES... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

24 1 242 243 250 253 295 296 307 315

This Page Intentionally Left Blank

CONTRIBUTORS TO VOLUME 43 Numbers in parentheses indicate the pages on which the authors’ contributions begin.

D. COHEN(75), Department of Microbiology and Immunology, Faculty of Health Sciences, Ben-Gurion University of the Negev, 84105 BeerSheva, Israel DAVIDCOLCHER (143), Laboratory of Tumor Immunology and Biology, National Cancer Institute, National lnstitutes of Health, Bethesda, Maryland 20205 PAUL B. FISHER(143), Department of Microbiology, Cancer Centerllnstitute of Cancer Research, Columbia University College of Physicians and Surgeons, New York, New York 10032 JUDAH FOLKMAN (175), Department of Surgery, Children’s Hospital Medical Center, and Departments of Surgery and Anatomy, Harvard Medical School, Boston, Massachusetts 02115 JOHN GREINER(143), Laboratory of Tumor Immunology and Biology, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 20205 PATRICIA HORAN HAND(143), Laboratory of Tumor Immunology and Biology, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 20205 DONALD KUFE (143), Dana Farber Cancer Institute, Boston, Massachusetts 02115 KARIN MOELLINC(205), Max-Planck-Institutfur Molekulare Genetik, 1000 Berlin 33, Federal Republic of G e m n y PHILIPNOCUCHI(143), Office of Biologics Research and Review, Centerfor Drugs and Biologics, Food and Drug Administration, Bethesda, Maryland 20205 SIDNEYPESTKA(143), Roche Institute of Molecular Biology, Nutley, New Jersey 07110 VARDAROTTER (113), Department of Cell Biology, The Weizmunn Institute of Science, Rehovot 76100, Israel VOLKERSCHIRRMACHER (l),Institut fur Immunologie und Genetik, Deutsches Krebsforschungszentrum, 0-6900 Heidelberg, Federal Republic of Germany ix

X

CONTRIBUTORS TO VOLUME

43

SCHLOM(143), Laboratory of Tumor lmmunology and Biology, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 20205 ROBERTSZICETI (241), Department of Tumor Biology, Karolinska Institutet, S-104 01 Stockholm 60, Sweden MAUREENWEEKS(143), Laboratory of Tumor lmmunology and Biology, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 20205 DAVIDWOLF(113), Department of Cell Biology, The Weizmann Institute of Science, Rehovot 76100, lsruel DAVIDWUNDERLICH (143), Laboratory of Tumor lmmunology and Biology, National Cancer Institute, National lnstitutes of Health, Bethesda, Maryland 20205

JEFFREY

CANCER METASTASIS: EXPERIMENTAL APPROACHES, THEORETICAL CONCEPTS, AND IMPACTS FOR TREATMENT STRATEGIES Volker Schirrmacher InStRUt fur lmmunologie und Genetik. Deutsches Krebsforschungszentrum.Heidelberg, Federal Republic of Germany

I. Introduction . . . . . . . . . . ............................................ 11. Metastasis Research. . . . . . . . .

1

C. Important Issues and Quest

A. Clinical D a t a . . . . . . . . . . . B. Experimental Data ............................................ IV. Instability, Subpopulation Interactions, and Tumor Progression: Experimental

18

...........................................

36

...........................................

46

............................

48

VI. Cancer Invasion

I. Introduction

One of the most urgent problems in the management of cancer is the prevention of metastasis formation. The ability of malignant cells to disseminate from a locally growing tumor and to form secondary lesions at near or distant sites is the most life-threatening aspect of cancer, and yet there are no good tools available for either diagnosis or treatment of metastases. Until recently, radical cancer surgery was designed to ablate the primary tumor and its lymphatic drainage in the hope that dissemination would thus be prevented. However, in recent years this has been more critically evalu1 ADVANCES IN CANCER RESEARCH. VOL 43

Copyright 8 1985 by Academic Press, Inc All rights oi reproduction in any form reserved ISBN 0-12-006643-2

2

VOLKER SCHIRRMACHER

ated, although survery has improved technically to such an extent that even the most horrendous primary lesions can be quite successfully removed. Cancer surgeons are becoming aware that metastases are their main enemy and that a small number of disseminated cells could later give rise to secondary deposits that might defeat their efforts and kill the patient. In the past, the choice of cancer treatment has been almost entirely a matter of the physician’s personal opinion. There have rarely been sound scientific evaluations of the costs and the benefits to the patient. Meanwhile, as will be described, new scientific insight into the metastatic process is beginning to provide the kind of information necessary for an evaluation of therapeutic strategies. It is well known from clinical cancer that the most rapidly growing tumors are usually the most capable of producing local or distant metastasis (Sugar; baker, 1979, 1981). For example, children with Burkitt’s lymphoma have fast-growing tumor lesions which generate aggressive metastases and the patients succumb rather quickly. On the other hand, there are certain types of cancer such as basal cell carcinoma in which lesions are highly invasive but rarely metastatic. Another important aspect is the relationship between primary tumor size and incidence of metastasis. In general, the incidence of metastasis appears to increase statistically with increased tumor size, although the same type of cancer often produces quite diverse survival patterns among individual patients (Willis, 1973). There are, however, examples of rapidly spreading tumors in which there is little relationship between primary tumor size and incidence of metastasis such as small cell carcinoma of the lung. The locations of most initial metastases, at least of cancers with moderate metastatic potential (Sugarbaker, 1981), appear to be determined simply by regional anatomy (Ewing, 1928). Human cancer cells frequently travel from the primary lesion to regional nodes and from nodes to veins via lymphaticovenous communications (Weiss, 1976; Gilbert et al., 1980). However, in approximately 20% of patients with carcinomas that eventually metastasize to lung, liver, bone, or brain, no lymph node involvement can be detected clinically or histologically. About 90% of women with lymph node negative breast cancer will survive for 5 years or more after adequate local therapy, but there are still 10% who relapse. Similarly, about 60% of patients with extensive lymph node involvement in this disease will be dead within 5 years, whereas another 40% will remain well. These figures illustrate the level of prognostic significance of regional lymph node involvement. It is obvious that we need better markers for tumor prognosis. Also, we need better understanding of the pathways of metastases and their sequence of development with time based on a clear and coherent scientific concept. Spread and growth of cancer cells seem to depend to a great extent upon the venous pathways involved. Tumors rarely invade arteries larger than

CANCER METASTASIS

3

precapillary arterioles. The arterial circulation seems to be a particularly hostile environment for tumor cells, which is best exemplified by the rarity of arterial dissemination. Muscle, kidney, spleen, intestine, skin, and heart are involved in less than 10% of all metastases, although the arterial output to these organs is equivalent to about 70% of the total arterial output. In contrast, metastasis in the bone, a tissue that receives less than 5% of the total arterial output, is quite common. Of interest also is the difference between bone, which is a fairly common site of metastases, and cartilage, which is hardly ever involved (Eisenstein et al., 1975). The most frequent organ site of distant metastasis in many types of cancer appears to be the first organ encountered by circulating tumor cells. However, there are a number of examples of blood-borne clinical metastasis not explainable by lodgement in the first capillary system encountered (Sugarbaker, 1981). Reasons for this organ site predilection will be discussed later. Due to the potential of metastatic spread, cancer-at least from a certain stage on-must be considered as a generalized disease and thus cannot be treated only locally. It is a basic misconception of strategy if physicians try to treat a generalized disease with localized therapy. There is a growing awareness that new approaches to the treatment of cancer are needed and that these must be based upon an increased understanding of the pathogenesis of metastasis. New developments in gene technology, cell, tissue, and organ culture under defined conditions, as well as progress in cancer chemotherapy, hormone therapy, and immunotherapy, have given rise to considerable opportunities to investigate the biology of cancer cells and the effects of a large number of substances on their behavior. Considerable attention is now being focused on the most crucial processes that differentiate benign from malignant growth, namely, invasion, dissemination, and metastasis formation. A dramatic increase in the past few years of research in clinical and experimental metastasis has led to the foundation of several new journals which specialize on this subject, namely (1)Invasion and Metastasis (Karger, since 1981) (2) Cancer Metastasis Reviews, (Martinus Nijhoff Publishers, The Netherlands, since 1982), and (3)Clinical and Experimental Metastasis (Taylor and Francis Ltd., London, since 1982). The complexity of the phenomenon of metastasis is reflected in the diversity of disciplines engaged, which range from biophysics to clinical oncology and involve cell biologists, biophysicists, biochemists, immunologists, physicians, etc. Because of the complexity of the field and because research is being done by so many academic specialties, investigators often find it diffcult to keep up with all of the developments, to interpret individual findings, or to develop concepts consistent with the infarmation obtained from many different disciplines. It is therefore the aim of this article to familiarize those interested in the

4

VOLKER SCHIRRMACHER

subject with (1)important issues and questions that are being addressed at present, (2) new approaches developed to study such questions, and (3)new results and concepts developed. A previous article in this series on the biology of cancer invasion and mestastasis (Fidler et al., 1978) is recommended as a basis.

II. Metastasis Research

A. OVERVIEW: FROM ONCOGENES TO METASTASES Table I illustrates the different levels of research in the field of cancer metastasis. The table also indicates the kind of questions investigated at each level and the experimental approaches used. Recent advances in molecular biology have led to the identification of oncogenes as general and important determinants of carcinogenesis. Oncogenes can be defined as eukaryotic genes that have been conserved in evolution and therefore presumably fulfill an essential function in the cell. They code for a protein and have the potential to act as dominant genetic trait, for instance, after transfer into a normal cell environment. Although oncogenes were originally defined as transforming genes in retroviruses, incorporation into a virus vector is not an essential criterion of an oncogene. Oncogenes can become abnormally expressed by a variety of mechanisms (mutational change, promoter insertion, change in transcriptional or in translational control, chromosomal rearrangement with oncogene translocation, etc.). Recently there have been reports also .of concerted actions of several oncogenes. Whether oncogenes might also play a special role in tumor progression and in metastasis is unclear at present. In any case, the new techniques of molecular biology have enabled an approach to metastasis-related questions at the molecular level. What happens after oncogene activiation within a single cell of a multicellular organism is less clearly established. The misinformation carried by the cell must be reproduced several million times by clonal outgrowth before it can be clinically detectable. This, among other reasons, may explain the delay time or latent period. Failure of feedback tissue control systems and perhaps concomitant failures of host defense systems that may normally safeguard against this process may be involved. This level of cellular interaction is more difficult to study in precancerous stages than,it is in overt cancer and in metastasis. Metastasis research at the level of cellular interaction has been performed in uitro and in uiuo to study tumor subpopulation interactions, metabolic cooperation, immune cell-tumor cell interactions, and organ cell-tumor cell interactions.

5

CANCER METASTASIS

TABLE I CANCERMETASTASIS RESEARCH AT DIFFERENT LEVELS Level of research

Questions studied

Experimental approach

Role of onc genes Role of cell surface molecules Role of state of differentiation

DNA transfection Effect of genetic manipulation Effects of DNA demethylation

Tumor subpopulation interaction Metabolic cooperation Immune cell-tumor cell interactions Organ cell-tumor cell interactions

Comparison of clones and mixed cultures Test for drug sensitivity Cytotoxicity test Cytostasis test Immune escape mechanisms Adhesion and inhibition tests

Mechanisms of angiogenesis Mechanism of invasion Regulatory effects from tissue microenvironment

In oitro studies using cultures of tissue fragments or biological membranes Signal perception and tumor cell response

Mechanism of metastasis Mechanism of host resistance Improvements of cancer therapy

Comparison of selected subpopulations Immunological studies Monoclonal antibodies Monoclonal T cells Immunotoxins Drug-targeted liposomes Lymphokines

Tumor cell-tissue interactions are the next higher level of complexity in metastasis research. There are several critical steps of tissue interactions that probably determine whether the transformed clone will or will not remain under control of the host. One such critical point is the evolution of a blood supply for the clone. Without a special nutrient supply via angiogenesis, the

6

VOLKER SCHIRRMACHER

size of the clone will be strictly limited. Angiogenesis may also be taken as a starting point for hematogenous dissemination and visceral metastasis formation. Mechanisms of cancer metastasis formation, host resistance phenomena, and cancer therapy also have to be studied at the level of the whole organism. These important subjects require careful animal experimentation since even the most sophisticated in uitro system cannot mimic properly the situation within an intact organism. Immune defense reactions or cytotoxic anticancer drugs may be very effective outside the body. Their relative effectiveness for cancer therapy has to be tested, however, in uiuo in terms of prolongation of disease-free interval, reduction in number of metastases, or prolongation of life expectancy. All three kinds of control on the cell in a normal location are probably less effective at the metastatic site. The original genetic damage has diminished internal control of the cell; the transplantation to a new site has reduced the local control by adjacent cells, matrix, and soluble factors; and finally the host defense system for external control is deteriorating with further disease progression.

B. ANIMALMODELS Studies on human metastatic tumor cells have been hindered mostly by the lack of appropriate in viuo assay systems. An exception is the use of immunosuppressed animals, the most popular of which is the congenitally athymic nude mouse. There are a number of inherent problems, however, associated with such a human tumor xenograft model (Sordat and Wang, 1984). Since the nude mouse has a high level of natural defense mechanisms [for instance, natural killer (NK) cells and macrophages], there may be highly selective pressures exerted on the human tumor material so that the tumors that eventually arise may not represent the same cellular diversity as the original tumor. In addition, there may be particular tissue stroma and endocrine requirements for growth and spread of human tumors that may not be provided optimally in the foreign host. Another problem is the lack of comparable lines from both primary and metastatic lesions of the same patient, which is probably due to the difficulties encountered in obtaining fresh tissues immediately after surgical excision and the lack of reproducible methods for routine cultivation of human tumor cells. Investigations on the mechanism of cancer metastasis have therefore mostly depended on animal model systems, most of which represent rodent tumors transplanted in their syngeneic hosts. There are obvious and serious limitations with attempts to compare animal data from such tumor systems with human cancer because the models do not accurately reflect or mimic

CANCER METASTASIS

7

some of the events that occur when a spontaneous human neoplasm progresses and metastasizes in its host. Nevertheless, the extensive use of transplantable animal tumors that share a common genetic background with their host but differ in their metastatic phenotypes and cell properties has provided most of our current knowledge about tumor and host characteristics in metastasis. The available animal tumor models for studying metastasis and their respective limitations as well as advantages have been summarized and discussed in detail elsewhere (Fidler et al., 1978; Nicolson, 1982; Poste and Nicolson, 1983). Table I1 lists some representative examples of such tumor models, all of which consist of subpopulations with different metastatic properties. B16 melanoma and Lewis lung carcinoma, which have been used most extensively in metastasis research, are of spontaneous origin (both from the C57BL/6 mouse) but have been propagated in uiuo and in uitro for almost 30 years, which is approximately 15 times the average life span of their original host. Such tumors are highly selected for rapid growth and do not mimic the slowly growing human or animal primary tumors, which often take years to reach the size at which they are diagnosed and treated. Most of the longtransplanted animal cell lines are extremely anaplastic and can be considered as models only for rapidly growing, very aggressive human tumors. Newly induced tumors, such as chemically or virus-induced neoplasms, are seldom metastatic. Many spontaneous tumors are capable of metastasizing widely, but their unpredictable occurrence makes them difficult to work with and to obtain reproducible results. There are some kinds of spontaneous tumors that arise at relatively high incidence in certain strains of laboratory animals and in domestic or farm animals, such as mammary carcinomas, osteogenic sarcomas, and lymphomas in dogs and cats (Hewitt, 1978). The use of such tumors from larger nonrodent animals for experimental studies, however, is limited because of limitations in space, time, and money and also because most such animals are not sufficiently inbred to allow their use as recipients for tumor transplantation.

C. IMPORTANTISSUESAND QUESTIONS The choice of an animal tumor system depends most of all on the question to be studied. In any case, the origin and subsequent passage history of the tumor to be studied should be known before it is received in the investigator’s laboratory in order to avoid such artifacts as Hewitt et al. (1976) pointed out and criticized. Important questions that are presently under investigation in model systems can be formulated as follows:

TABLE I1 EXAMPLES OF ANIMAL TUMOR MODELSWITH SUBPOPULATIONS DIFFERING IN METASTATICPROPERTIES

Species

Tumor

Subpopulation

Major sites of metastases

References ~

Mouse Mouse

B16 melanoma B16 melanoma

Clones Organ-selected variants

Mouse

B16 melanoma

Mouse Mouse Mouse Mouse Mouse Mouse

B16 melanoma B16 melanoma 3LL Lewis lung carcinoma 3LL Lewis lung carcinoma Mammary carcinoma W 2237 fibrosarcoma

Immunoresistant variants Invasion selected Lectin resistant Lung selected Clones varying in H-2 Sublines Clones

Mouse

MDAY-D2 undifferentiated

Mouse

Eb/ESb lymphoma

Mouse Mouse Mouse Rat

Eb/ESb lymphoma EblESb lymphoma SV 3T3 sarcoma ARC-1-RT7 hepatocarcinorna HV-transf. AL.2 lymphoma Adenocarcinoma

Chicken Human

Lung>other sites Ovary, brain

Fidler and Kripke (1977) Brunson et al. (1978), Brunson and Nicolson (1979) Fidler et al. (1976)

Lung, lymph nodes

Plastic adherent Spleen selected

Spleen

Clones

Liver

Poste et al. (1980) Tao and Burger (1977) Fogel et al. (1979) Eisenbach et al. (1983) Heppner et al. (1978) Kripke et al. (1978), Raz et al. (1981) Kerbel et al. (1982), Dennis et al. (1981b), Frost and Kerbel (1981) Bosslet and Schirrmacher (1981), Schirrmacher et al. (1982~) Fogel et 01. (1983) Cbeingsong-Popov et al. (1983) Nicolson et al. (1978) Talmadge et al. (1979)

Ascites

Liver Lung

Takahashi et al. (1978)

Lectin resistant, immunoselected CTL resistant

Lung

Lung Liver, lung Liver, lung, spleen

Shearman and hngenecker (1980)

CANCER METASTASIS

9

1. Is there a metastatic phenotype and if so, what are its cellular properties? Can this explain why some tumors metastasize and others do not? 2. How stable are tumor cell subpopulations or clones, and what is the origin of tumor heterogeneity? 3. Is metastasis a random, a selective, or an adaptive process? To what extent do these phenomena influence metastatic spread? 4. How can the organ preference of certain tumors be explained? How and at what levels is this phenomenon determined by tumor cell properties or by properties of the host? 5. What host influences are there in cancer metastasis? What is the role of the microenvironment (cells, extracellular matrix, soluble factors)? Can metastatic cells receive signals from the microenvironment? How do metastatic cells escape immunological host control? 6 . How do some micrometastases remain dormant for prolonged periods of time? What makes them reactivate? 7. Does metastasis proceed in sequential steps? If so, is the whole process equivalent to the sum of the sequential steps or is it still more complex? 8. How does a growing primary tumor “condition” its host and what is the influence of the primary tumor on the growth of metastases? 9. Is metastasis a product of tumor-host interactions? If so, what is the mechanism of interaction between cancer cells and host tissues in sequential steps of the metastatic cascade, e.g., in angiogenesis, invasion, capillary adhesion, and immune escape. Can such processes be manipulated toward the advantage of the host? 10. What are the consequences of tumor heterogeneity, instability, and subpopulation interactions for cancer therapy? The list of questions is not complete but it may give some insight into the area and may show the nature and fundamental importance of the issues studied. It should be kept in mind that without appropriate model systems and test methods, these questions cannot be investigated at all.

D. ASSAYPROCEDURES A N D SELECTION OF METASTATIC SUBPOPULATIONS Before illustrating recent progress in the field, I would like to discuss some methodological aspects, namely (1)assay procedures to test for metastatic capacity and (2)selection procedures to obtain tumor subpopulations of different metastatic capacities. To test for metastatic capacity, transplantable tumors are inoculated locally (subcutaneously or intramuscularly) and then the formation of “spontaneous” metastases in various major organs is moni-

10

VOLKER SCHIRRMACHER

tored. In contrast to this spontaneous metastasis assay, in the “experimental” metastasis assay tumor cells are injected directly into the circulation, mostly by tail vein inoculation. The iv method circumvents the initial steps of metastasis and tests mostly for capacity to survive in the circulation and to implant in internal organs. This implantation phase includes arrest in capillaries, extravasation, and growth. Although tumor cells may successfully colonize organs after iv inoculation, they may have low spontaneous metastatic potential because of low rates of entry into the circulation after local inoculation. Welch et a2. (1983)claim that there is good agreement between experimental and spontaneous metastasis assays, whereas others (Stackpole, 1981) reported that this was not the case. Both the spontaneous and the experimental metastasis assays have been reviewed and discussed critically (Hagmar et al., 1983). Since the interaction of tumor cells with their environment is mediated by cell surface constituents, surface components are thought to play a major role in metastasis. Figure 1 illustrates two basically different approaches to correlate metastatic capacity with tumor cell surface properties (Burger, 1980). Both in uiuo selection of metastatic variants and testing for cell surface properties and in uitro selection of cell surface variants and testing for metastatic properties have been used in these studies. In uiuo sequential selections for metastatic subpopulations were already being performed several decades ago (Klein, 1955). They became popular when Fidler exploited this procedure more systematically in a series of elegant experiments. He started with sequential selections to enrich B16 melanoma cells for subpopulations with enhanced metastatic capacity or Wild type cell

Variant cell

surface change

FIG.1. Two different approaches to a search for correlations between alterations in metastatic capacity and cell surface alterations. (From Burger, 1980.)

C A N C E R METASTASIS

11

with enhanced potential to colonize specific organs, such as brain, liver, or ovary (see Table 11). In uitro selection procedures have been employed to isolate tumor cell variants with altered cell surface properties such as increased resistance to plant lectin toxicity (Tao and Burger, 1977; Dennis and Kerbel, 1981), lymphocyte-mediated cytotoxicity (Fidler et al., 1976), NK-cell-mediated killing (Hanna and Fidler, 1981), or antibody-complement-mediated lysis (Frost and Kerbel, 1981). Other procedures selected for changed adhesiveness to plastic (Fogel et al., 1983), substratum (Briles and Kornfeld, 1978), or endothelial cells (Nicolson, 1982) or for increased ability to invade bladder tissue (Hart, 1979) or blood vessels (Poste et al., 1980). When assayed in uiuo, such in uitro-selected variants showed alterations in their metastatic behavior. Another in uitro procedure employed for obtaining metastatic tumor variants has been single cell cloning. Recent studies, however, have pointed out that clonal tumor cell populations are not always stable. In fact, in the majority of cases analyzed systematically, cloned tumor cell lines turned out to have an inherent instability. The evidence for this and the impact of this finding will be discussed below (Section IV). In order to increase the frequency of variant generation, some investigators mutagenize their tumor lines, whereas others try to avoid the use of drugs because of their multiple effects and concentrate on spontaneous variants. Boon and associates (Boon and Kellermann, 1977; Boon, 1983) and recently also Frost et al. (1983) reported that mutagenesis of normally tumorigenic tumor cell lines followed by cloning of the surviving cells often gives rise to a very high frequency (10-90%) of clones which are unable to grow progressively in normal syngeneic mice. Such clones usually have acquired new strong tumor antigens that lead to their rejection in immunocompetent but not in immune incompetent mice. Similar results were recently obtained with the nonmutagenic drug 5-azacytidine, which causes undermethylation of DNA and deregulation of gene expression (Frost et al., 1984). These findings suggest that the high frequency of tumor variants was not generated by classical mutation events but rather by “epigenetic” mechanisms. The altered methylation patterns can be somatically inherited, although not with perfect fidelity (Wigler et al., 1981), thereby raising the distinct possibility that “methylation changes can masquerade as mutations” (Riggs and Jones, 1983). While the majority of clones derived from 5azacytidine-treated tumor cell cultures had an increased immunogenicity, there have also been clones isolated which showed increased metastatic capacity. These studies have important implications for our understanding of tumor immunogenicity and of tumor variant generation and will therefore be discussed again in the context of mechanisms of tumor progression (Section IV).

12

VOLKER SCHIRRMACHER

E. CELLSURFACE PROPERTIES AND CANCER METASTASIS Some of the findings derived from experimental studies comparing selected subpopulations with different metastatic properties provided strong evidence for the importance of cell surface properties in cancer metastasis. These are summarized in Table 111. The evidence comes from both types of approaches of Fig. 1 as well as from recent inhibition studies with monoclonal antibodies directed against important cell surface molecules. Several authors (Rapin and Burger, 1974; Chatterjee and Kim, 1977, 1978; Irimura et al., 1981) have suggested that cell surface carbohydrates are important in cancer metastasis. Yogeeswaran and Salk (1981) found that tumor lines with high metastatic capacity showed increased levels of sialylation of cell surface glycoconjugates when compared with related lines of low metastatic potential. We have suggested recently (Schirrmacher et al., 198213) that not only the amount but also the specific positioning of sialic acid at the cell surface may be important: Via changes in positioning, sialic acid could lead to either blocking or unblocking of cellular adhesion sites or antigenic determinants and could thus influence metastatic capacity. Lectin-carbohydrate interactions have been found to play a crucial role in many intercellular recognition processes. We recently provided the first molecular description of such an interaction between organ-derived normal TABLE 111 EVIDENCE 'I'llAT TUMOR CELL SURFACE PROPERTIES ARE IMPORTANT FOH METASTATIC CAPACITY Evidence Tumor wild type and metastatic variant show cell surface change Changes in glycoproteins and glycolipids Changes in lectin-binding characteristics Tumor wild type and surface variants show change in metastatic capacity Wheat germ agglutinin-resistant variants and metastatic revertants thereof

GP 70 cell surface variant Effects of tunicamycin Effects of fusion of membrane vesicles Antibodies against distinct cell surface components can have inhibitory effects on metastasis formation

Reference

For review see Nicolson (1982)

For review see Schirrmacher et al. (1982)

Tao and Burger (1977), Kerbel et al. (1982), Finne et al. (1980), Dennis and Kerbel (1981) Reading et al. (1980) Iriinura et al. (1981) Poste and Nicolson (1980) Nicolson (1982), Vollmers and Birchmeier (1983a,b), Vollmers et al. (1984)

13

C A N C E R METASTASIS

Tumor line

Eb Eb Eb ESb ESb ESb

Pretreatment

neurarninidase 0-galactosidase

-

neurarninidase P-gaiactosidase

Yo hepatocytes with tumor cell rosettes 5 89 0 74 92

26

FIG.2. Rosette formation between hepatocytes (central cells) and liver-metastasizing ESb tumor cells. (Reproduced from Schirrmacher et al., 1982.)

parenchymal cells-hepatocytes-and liver-metastasizing tumor cells (Schirrmacher et al., 1980; Cheingsong-Popov et al., 1983). Figure 2 shows this interaction in the form of rosettes between hepatocytes and highly metastatic ESb tumor cells. Removal of P-galactosyl groups from the ESb tumor cells by P-galactosidase treatment resulted in a reduction of spontaneous liver-rosette-forming capacity. The low metastatic parental type Eb cells had only a low liver cell rosette-forming capacity unless the cells were pretreated with neuraminidase. This treatment leads to exposure of free P-

14

VOLKER SCHIRRMACHER

galactosyl residues. The hepatocytes were found to bind ESb tumor cells through lectin-like hepatic binding proteins (HBPs) with molecular weights of 52, 56, and 110 kDa and specificity for D-galactosyl and N-acetyl-D-galactosaminyl residues (Cheingsong-Popov et al., 1983).Thomsen-Friedenreich (T) antigen, which expressed immunodominant Galpl3GalNac determinants, was found to be present on our tumor cells (Springer et al., 1983a). In soluble form, this T antigen was a powerful inhibitor of the spontaneous hepatocyte rosettes (Springer et al., 1983b). More than 10 different cell surface glycoproteins of ESb tumor cells and none of Eb-type tumor cells served as ligands in the hepatocyte adhesion. A possible relevance of such an interaction for the organotropism of cancer metastasis was suggested from the finding that spleen-selected ESb sublines differed from liver-selected ones in their organotropism as well as in their ability to form hepatocyte rosettes (Cheingsong-Popov et al., 1983). From these and other findings (Altevogt et al., 1983; Fogel et al., 1983)we recently formulated a new hypothesis that underlines the importance of sitespecific positioning of sialic acid (Schirrmacher et al., 1982b). We propose (1) that cell surface carbohydrate changes take place during tumor progression; (2) that selection favors changes that are not random but rather cell surface site specific, and (3) that glycosyltransferases may play an important role in these changes, leading to the masking of some sites (lectin receptors, adhesion sites, antigens) and exposure of others (other lectin receptors, adhesion sites, or antigens).

Ill. Cascade Theory of Metastasis

A. CLINICAL DATA Clinical autopsy data on visceral tumor metastases have been analyzed with respect to incidence and sequence in order to delineate pathways of metastasis. In one study more than 4700 consecutive autopsies were analyzed, first with standard statistical methods (“analysis of variance”) (Bross and Blumenson, 1976) and later by focusing on specific primary sites using the “cascade analysis” as a new statistical tool (Viadana et al., 1978a,b). The main purpose was to provide an objective procedure for determining patterns in the sequence of events of the metastatic process. This analysis together with others led to the foundation of the cascade theory of metastasis. It clearly suggested that the process of metastasis takes place step by step and involves both systematic processes of dissemination and chance mechanisms. In the terminology of biostatistics, it is a sequential as well as a stochastic process. Experimental animal data in addition suggest that the

15

CANCER METASTASIS

metastatic process is also a selective one (Poste and Fidler, 1980). The extent to which cancer metastasis is selective or random seems to depend on the parent tumor population (Talmadge and Fidler, 1982). Figure 3 illustrates major pathways of metastases from two clinically important types of cancer, namely, carcinoma of the lung and of colon. Lung cancer can disseminate via the left heart ventricle directly into the arterial circulation resulting in an arterial pattern of spread. Colon cancer tends to metastasize via the mesenteric lymphatics and the portal venous system into the liver as the first generalizing site. From there the cells can disseminate further via the right heart ventricle to the lung. Most types of cancer show a lymphaticovenous pattern of spread, like colon carcinoma. Major pathways of metastases of important types of human cancer can be summarized as follows (Gilbert et al., 1980):

B

A

Carotid artery

I

i \ + l , B

A

I) L I

L

\

@Primary @Metastases

1

h

\

FIG.3. Major pathways of metastasis from two clinically important types of cancer based on Gilbert et al. (1980).Notice the difference between lung cancer (A), which can disseminate via the left heart ventricle directly into the arterial circulation resulting in an arterial pattern of spread, and colon cancer (B), which tends to metastasize via the mesenteric lymphatics and the portal venous system into the liver as the first generalizing site.

16

VOLKER SCHIRRMACHER

1. Head and neck cancers metastasize to the lymphatics and remain confined in this area. Nasopharyngeal cancer tends to go to the liver and bone more frequently than other head and neck cancers. The spread to bone is via Batson’s plexus, a slow-moving valveless system of paravertebral veins. 2. Breast cancer metastasizes either directly to the lung via the lymph nodes and superior vena cava or passes directly to the bony skeleton via the paravertebral veins. 3. Lung cancer is the only tumor that has direct access to the arterial circulation via the pulmonary vein and the left heart ventricle. It can thus spread widely to many organs including the brain (see Fig. 3). 4. Colorectal carcinomas tend to metastasize via the mesenteric lymphatics and portal venous system into the liver as the initial resting place. From the liver these tumors can metastasize to the lungs via the inferior vena cava, the right ventricle, and the pulmonary artery. 5. Tumors of the testicle metastasize via the lymphatics to nodes of the periaortic area and then enter the subclavian veins to go to the right heart and finally to the lungs. Liver metastases occur late in the disease. 6. Ovarian cancer remains confined for long periods of time in the highly favorable environment of the abdominal cavity, especially the peritoneal surfaces, the posterior gutters, and the diaphragm. These tumors invade the liver only in a small percentage of cases at a very late stage, usually by direct invasion from omental disease or mesenteric venous emboli from omental implants. Lung metastases also occur, but late in the disease. 7. Gynecologic tumors tend to involve primarily lymphatic structures. Visceral organs are uninvolved even late in the course. 8. Prostate cancer can metastasize to the bone via Batson’s plexus of paravertebral veins. There is also involvement of the lung, liver, endocrine glands and the central nervous sytem.

The main points from the cascade analysis have been summarized by Viadana et al. (1978a) as follows: (1) The overwhelming majority of cancers are disseminated by a multistep process. (2) Generalized disease (such as brain metastasis) does not ordinarily occur directly from the primary tumor. (3)There are one or more key generalizing sites that are specific for a given primary tumor. (4) These sites depend largely on the drainage of the venous blood. (5)The generalized disease is produced by secondary metastases from the key sites. (6) The similarity of the cascade processes at so many different sites reflects a similar underlying process of generalization through the blood system. To all such general points there will be occasional exceptions. Given these data, one may ask why cancer cells metastasize step by step and do not generalize throughout the body directly from the primary tumor. It could be argued that cancer is a gradually developing disease. At the time

C A N C E R METASTASIS

17

when the cells leave the primary tumor and enter the blood stream, they may not yet have evolved to the point where they can function effectively as the free-living cells that can grow almost anywhere in the body. Another explanation is given by Weiss and co-workers who recently investigated cancer cell traffic from the lungs to the liver (Weiss, 1980) and from the liver to the lungs (Weiss et al., 1983). They propose that cancer cells which are temporarily arrested in the first organ encountered are “processed” by it, so that they die before or shortly after arrival in another organ. Mechanisms of tumor cell damage could include mechanical trauma, attack by NK cells, or damage due to cellular and humoral inflammatory responses to interactions of the cancer cells with vascular endothelium. It was suggested that metastases in second organs would, to a large extent, be generated by cancer cells from metastases in the “first organs” as distinct from direct seeding from cancer cells released from the primary tumor. These and other results to be discussed later (Section 111, B,4) emphasize the potential importance of metastasis from metastases in the natural evolution and spread of cancer. The underlying principle of the cascade hypothesis is that there are definite steps or stages in the evolution (Viadana et al., 1978b; Bross, 1980). Although a number of questions remain unanswered, the cascade theory represents at least the beginning of a comprehensive scientific concept of metastasis, which can take into account stepwise changes in cancer cells, changes in the host environment and its defense system, as well as physical factors such as the anatomical and biochemical aspects of the routes of dissemination. Apart from this, the findings of distinct metastatic pathways and of sequential development in human cancer raise a number of interesting questions that deserve intense future investigations:

1. Why are the lung and the liver such important key sites and not, for instance, the lymph nodes or other organs? 2. Are there changes occurring at the first metastatic site that make the tumor cells more adaptable for growth at other sites in the body? 3. Why are lung metastases so often the last step in the cascade process before the cancer generalizes to the endocrine glands, the central nervous system, and to other sites involved in the last stages of the disease? 4. What is the reason for organ predilection in metastasis? While for many years it was thought that metastases appeared in the organs in which they might be expected to occur on the basis of anatomy of the circulation, clinicians know of many exceptions. For instance, both Wilms’ tumor and neuroblastoma drain into the inferior vena cava yet the predominant metastatic site for Wilrns’ tumor is the lung and for neuroblastoma it is bone. Other examples have been reported (Kinsey, 1960). Concepts for mecha-

18

VOLKER SCHIRRMACHER

nisms have been developed from experimental studies and will be discussed under Section III,B (Table IV). B. EXPERIMENTAL DATA The cascade theory of metastasis has also formed the basis of a large number of experimental studies in the field. Here, analysis of functional aspects and of mechanisms are of primary concern. There is a surprisingly concordant view among experimentalists that the overall process of metastasis take place in sequential steps. Subsequent to the establishment of a primary tumor, there must be invasion of the surrounding tissue with the eventual penetration of blood and/or lymphatic vessels. Circulating tumor cells or tumor emboli that eventually survive in the blood stream can stop in capillaries of distant organs. To establish secondary lesions, they must penetrate the vessel wall, infiltrate the surrounding parenchyma, and be able to grow there. The whole process may then be repeated starting from the metastatic lesion (Sugarbaker et al., 1971; Hoover and Ketcham, 1975). An analysis of the various steps that lead from a primary tumor to local or distant secondary lesions has been presented in several excellent reviews which should be consulted for details (Weiss, 1976; Nicolson, 1978a, 1979; Poste and Fidler, 1980). Figure 4 shows two illustrations of the process reproduced from Burger (1980) and from Fidler and Poste (1982).

1. Metastatic lnefficiency It has been shown that only 1%or less of the millions of tumor cells that may escape from the primary tumor into the circulation survive to become a viable metastasis, while the majority of cells will die (Liotta et al., 1974). It is thus an apparent paradox that on one hand metastasis is a major cause of death in patients with cancer, but on the other hand in terms of cancer cells themselves, metastasis appears to be an inefficient process (Weiss, 1980). Since the vast majority of cells released into the circulation will die, the primary tumor must reach a critical population size and/or growth rate to deliver sufficient cells into the circulation to proceed to the next step in the cascade. Most clinical metastases do not occur until the primary tumor has reached at least 1 cm3, corresponding to about lo9 cells (Weiss, 1982). As causes of the high rate of cell death of circulating nonlymphoid tumor cells, the following have been suggested: (1) Mechanical shear forces, (2) loss of attachment substrate and spreading, (3) oxygen toxicity, and (4) destruction by host-derived circulating natural killer cells. In animal experiments, the number of final metastases has been found to be proportional to the number of circulating tumor cells but the best correlation was with clumps of tumor

19

C A N C E R METASTASIS

A

Embolization

Progressive Growth, Vascularization, and Invasion

I

Extraysation

4

3

5

2

Q

Blood Vessel

sive Metastatic Growth

B m Growth N o t Metastasizing

Metastasizing

I

Release from Primary Site

Establishment Secondary Site

FIG.4. Sequential (“cascade”)steps in metastasis. (A) The process of hematogenous metastasis formation is dissected into six steps. (From Fidler and Poste, 1982.) (B) The process of

metastasis is dissected basically into two topographically separate processes. The upper portion shows the process leading to penetration of a carrier system (blood, lymph, body cavities). After dissemination, the lower portion shows the processes that lead to the establishment of metastatic lesions in the periphery. (From Burger, 1980.)

20

VOLKER SCHIRRMACHER

cells. The experimental systems for studying the behavior of circulating tumor cells and their survival have been reviewed elsewhere (Fidler, 1976, 1978; Warren, 1981). Shedding of tumor cells at an adequate rate into the circulation is not sufficient per se for metastasis. The primary tumor may have to pass through several stages of increasing malignancy before its cells are capable of metastasizing. Thus, in 1973, Fidler proposed that survival and eventual metastasis of B16 melanoma cells are dependent upon properties unique to the tumor cells and that the metastatic process is not random (Fidler, 1973). Similar observations were made in some tumor systems but not in others (Giavazziet al., 1980; Ryd et al., 1983). The question of whether processes in metastasis are random or nonrandom is being discussed over and over again (Weiss, 1983). Most likely metastasis is neither an exclusively random nor an exclusively nonrandom process. A systematic analysis of the colonization potential of spontaneous mouse mammary tumors (Tarin and Price, 1981) after inoculation of one million viable tumor cells ip, sc, iv tail vein, or iv hepatic portal vein revealed that each tumor line had its intrinsic colonization potential. The expression of this, however, was influenced by the microenvironment of an organ, for instance, its circulatory anatomy. The degree and sites of colonization were thus the results of interactions between tumor- and organ-specific factors.

2 . Organ Site Predilection Apparently contradictory hypotheses have been put forward to explain the selective involvement of certain organs in certain types of cancer: The “seed and soil” hypothesis and the “mechanistic theory.” While Paget (1889) postulated that site-specific metastasis was the consequence of the provision of a fertile environment (the soil) in which compatible tumor cells (the seed) could proliferate, Ewing (1928) stated that site-specific metastasis was a direct consequence of the anatomical location of the primary tumor and of hemodynamic circumstances. Sugarbaker (1981) suggested that the mechanistic theory could explain phenomena at a relatively early stage of spread, whereas the seed and soil hypothesis might explain phenomena at later stages of disease progression. That tumor cell surface properties are important for organ selectivity was suggested by Hagmar (1972) and later by Fidler and Nicolson (1976). A theoretical analysis was performed by Weiss (1975), who suggested that the phenomenon of organ predilection is due to the ability of tumor cells to discriminate between various vascular beds during their travel in the circulation. Evidence for specific aggregation of organselected tumor cells with cell suspensions of the target organs-be they lung, liver, or ovary-was indeed obtained (Nicolson and Winkelhake, 1975).Similarly, using a cryostat section-binding assay, it could recently be shown that tumor lines would preferentially bind to sections from those

21

CANCER METASTASIS

organs to which they also metastasized in uivo (Netland and Zettler, 1984; Kieran and Longenecker, 1984). The specific site of organ-tumor cell interaction remains to be elucidated, however. Recent findings point toward a role of tumor cell recognition via organ-specific receptors (Schirrmacher et al., 1980; Cheingsong-Popov et al., 1982) or lymph node homing receptors (Stevens et al., 1982). In a series of elegant organ ectopic site experiments, it could be shown that tumors that preferentially colonize the lung but not the kidney would do so also when tissue fragments of these two organs were implanted at an ectopic site (Hart and Fidler, 1980). Organ-determined modulation of tumor growth was furthermore indicated by experiments of I. R. Hart (1982) and M . M . Burger (personal communication). Such experiments seem to corroborate Paget’s theory from nearly a century ago! The different concepts of mechanisms of site-specific metastases are summarized in Table IV. 3. Extravasation

Of importance for the process of extravasation seems to be the much higher affinity of circulating tumor cells to exposed matrix from the basal lamina as compared with the surface of blood vessel endothelium (Nicolson, 1978b; Vlodavsky et al., 1982, 1983a,b). A confluent sheet of vascular endothelium in culture was shown to separate in response to contact with a fibrin clot (Kadish et al., 1979) and to retract in response to tumor cells (Nicolson, 1978a). A similar retraction response to tumor cells was observed with mesothelium of the diaphragm (Haemmerli and Strauli, 1978; Granzow et al., 1980). Such a response may lead to exposure of a window of matrix to which tumor cells could stick and through which, after degradation, they could leave the circulation. Organ site predilection could be influenced by the composition of the subendothelial extracellular matrix in different organs. TABLE IV CONCEPTS OF MECHANISMS OF ORGAN SITE-SPECIFIC METASTASIS Concepts “Seed and soil” hypothesis Mechanical entrapment hypothesis Combinations of 1 and 2 Specific adhesive interactions

Organ-determined modulation of tumor growth

References Paget (1889) Ewing (1928) Sugarbaker (1981), Proctor (1976) Nicolson and Winkelhake (1975), Shearman et al. (1980). Schirrmacher et al. (1980), Cheingsong-Popov et al. (1983), Netland and Zetter (1984), Kieran and Longenecker (1983) Tarin and Price (1981), Hart (1982)

22

VOLKER SCHIRRMACHER

4 . Metastases from Metastases A clinically important question is whether metastases can further metastasize because the decision about whether and when to remove metastatic foci depends partly on their threat as a source of new metastases. To answer this question, Hoover and Ketcham (1975) performed paraboisis experiments. They amputated primary tumors from mice after pulmonary metastases were present and then sutured two animals together along their side so that skin, muscle, and peritoneum were united. Within 5-7 weeks, pulmonary metastases developed in the test mouse, indicating their origin from the pulmonary metastases of the donor mouse.

5. Znteractions between Primary and Secondary Lesions There seem to exist negative feedback mechanisms in otherwise uncontrolled tumor growth. Sugarbaker et al. (1977) showed (1) that a primary tumor can inhibit the growth of metastases, (2)that the strength of inhibition is proportional to tumor mass, and (3) that inhibitory influences of an expanding tumor mass are probably also exerted upon itself and on a second tumor implant. Such inhibitory influences could be partially due to concomitant tumor immunity and partially to nonimmunological mechanisms, as has been reviewed recently (Gorelik, 1983).

6. Important Properties in the Metastatic Cascade Factors of potential importance in the complex and dynamic process of metastasis are listed in Table V. They are subdivided according to tumor or host contribution and are by no means complete. Processes that promote the invasive capacity of a tumor and those that increase the resistance to natural and specific immune defenses will be advantageous for metastatic spread all along the metastatic cascade. A decrease in homotypic adhesion among the TABLE V SOMEFACTORS O F POTENTIAL IMPORTANCE I N T H E METASTATIC:CASC:ADE Contribution from Factor Growth rate and mechanical pressure Tumor cell surface properties Tumor cell adhesive properties Migratory capability and deformability Lytic enzymes Fibrous reaction (“barrier effect”), invasion, and tissue resistance Angiogenesis Host immune responses and tumor immune escape

Tumor

Host

CANCER METASTASIS

23

tumor cells will most probably promote release from the primary site. Conversely, an increase rather than a decrease in homotypic cell adhesion may promote tumor implantation and establishment in the periphery since cell aggregates trapped in a microvascular bed seem to escape the unsuitable environment of the blood stream much better than single cells. It thus seems that tumor adhesive properties have to change during different steps of the metastatic cascade. Host properties, in particular immunological reactions, could also have opposite effects resulting in either tumor inhibition or tumor enhancement. Release of lytic enzymes by inflammatory cells may be important for host defense reactions but it could also prepare host tissue for tumor invasion. Antibodies as well as T cells or macrophages have been shown to have tumor inhibitory effects in some situations and tumor-enhancing effects in others. The outcome may depend on subtle differences between subpopulations of the tumor cells as well as between subpopulations of the host’s responding cells. IV. Instability, Subpopulation Interactions, and Tumor Progression: Experimental Findings and Theoretical Concepts

One of the most insidious characteristics of cancer cells is their capacity to diversify and to create cellular variants. Even if a cancer originated from a single transformed cell as suggested by enzyme marker analysis (Fialkow, 1976), a tumor cell population at the time of clinical manifestation will be phenotypically and genotypically heterogeneous. Tumors may be heterogeneous in several ways: (1)heterogeneity among cancers of the same histological type coming from different individual patients, (2)heterogeneity arising over time by tumor progression within the same patient, i.e., from tumorigenic, noninvasive via invasive, nonmetastatic to invasive and metastatic subpopulations, (3) heterogeneity within a single tumor at any one time as revealed by histological examination; this includes differences due to microenvironmental conditions (stroma reaction, microvasculature, oxygen tension, pH, substrate supply, waste drainage) and due to stage of cell cycle. Studies in several tumor systems have meanwhile confirmed Fidler’s original observation (Fidler, 1973) that malignant tumors are heterogenous in metastatic capacity (Poste et ul., 1982). New observations have been made in the last few years concerning the stability of metastatic subpopulations or clones. When assessing the stability of the metastatic phenotype in clonal lines of B16 melanoma exhibiting high or low metastatic potential, it was found that the highly metastatic cells were often less stable (Fig. 5A). In the KV-2237 fibrosarcoma, a highly metastatic clone cultured for 60 days had become composed of subclones exhibiting a wide range of metastatic phenotypes, whereas a clone with low metastatic potential had retained this

24

VOLKER SCHIRRMACHER

,A

016 Melanoma

1 Drug resistance marker I

1-

High metastatic potential

A

+low clone

+intermediate clone

B B B

I 20

Passages I

/I\ with marker

I

with marker

Baa c c 1

/

I

1

1

c

1

/

\

Assay for lung colony forming potentialafter i v inoculation

c

Conclusion: single culture of clone results in instability

c

\

1

c

clone can be stabilized by addition of other clones

FIG. 5. Experimental evidence for clonal instability of the metastatic phenotype (lung colonization capacity) and of stabilization by subpopulation interactions. (Based on Poste et al., 1981.)

CANCER METASTASIS

25

phenotype (Poste et al., 1981). This, however, must not always be the case. The experience of several groups (Chambers et al., 1981; Neri and Nicolson, 1981; Miller et al., 1983) shows that individual subpopulations and clones thereof are heterogenous in their stability. Furthermore, tumor subpopulation changes can be sudden or gradual, and they can result in either more or less malignant phenotypes (Heppner and Miller, 1983). Changes can occur after in uitro or in uiuo passage. Clonal analysis of metastatic lesions produced by B16 melanoma populations containing clones with identifiable markers revealed that the majority of experimental metastases produced by intravenous injection of tumor cells were unicellular in origin (Talmadge et al., 1982; Poste et al., 1982). During the early stages of their growth (< 25 days), the majority of metastatic lesions contained cells with indistinguishable metastatic phenotypes (intralesional clonal homogeneity), while progressive growth (> 40 days) of metastatic lesions was accompanied by emergency of variant tumor cells with altered metastatic properties (intralesional clonal heterogeneity) (Poste et al., 1982; Talmadge et al., 1984). There is, thus, no question that tumors are heterogeneous for invasion, metastasis, and various other biological properties. What is not understood are the mechanisms responsible for the generation of heterogeneity in primary neoplasms and among and within metastases. This is an important issue because it directly relates to tumor progression, i. e., heterogeneity seen in the same tumor as a function of time. How can we be successhl in the long run in therapeutic intervention with metastasis if we do not understand mechanisms of progression? In the light of what we know so far about these processes, it seems quite possible that chemotherapeutic drugs, irradiation, and surgery, which all aim at reducing the overall tumor burden at the same time, affect the heterogeneity and change the biological behavior of the remaining tumor subpopulations. Different concepts of mechanisms of tumor progression which are under present discussion are summarized in Table VI. The concept that malignant cells appear as the result of progressive stepwise changes was first introduced by Foulds (1975). It was then refined by Nowell (1976), who suggested that the process of tumor progression was the result of acquired genetic lability within tumor cells allowing for continuous selection of variant sublines, a neo-Darwinian concept which could explain several experimental findings (Chow et al., 1983). Genetic errors could arise from classical genetic mechanisms or from the production of cellular variants as in normal tissue differentiation. The type of genetic errors in neoplastic cells could include mutations in structural genes, mutations in regulatory genes, or major genomic changes due to numerical or structural chromosomal alterations. Nowell (1982) proposed a special role of oncogenes in this process, suggesting that variability in number and place of insertion sites could result in position effects on gene regulation.

26

VOLKER SCHIRRMACHER

TABLE VI CONCEFTSOF MECHANISMSOF TUMORPROGRESSION Concepts 1. Progression due to stepwise neoplastic

development through qualitatively different stages 2. Progression due to acquired genetic lability allowing for continuous selection of variant sublines 3. Progression dependent on controlling elements from within (clone-dependent genetic programming for shifts) and from without (regulatory influences via tumor subpopulation interaction and via inductive signals from the microenvironment) 4. Progression due to changing environmental conditions; evidence for coexistence of hormone-dependent and -independent subpopulations prior to progression 5. Progression due to spontaneous somatic hybridization with host cells followed by chromosome segregation 6. Epigenetic mechanisms of tumor progression via changes in DNA methylation; induction of high-frequency heritable but phenotypically unstable changes

References Foulds (1975)

Nowell (1976, 1982)

Vaage (1980), Schirrmacher (1980), Neri and Nicolson (1981), Poste et al. (1981), Kiang et al. (1982). Katzav et al. (1983). Bennet (1983)

Sinha et al. (1977), Sinha (1981). Sluyser et al. (1981). Isaacs and Coffey (1981). Isaacs et al. (1982) De Baetselier et al. (1981), Dennis et a1 (1981b), Kerbel et al. (1983) Frost et al. (1983), Kerbel et al. (1984), Feinberg and Vogelstein (1983)

Two recent findings seem to have particular relevance for the evolution of cellular diversity and metastatic heterogeneity within neoplastic lesions. (1) Highly metastatic cells were found to have higher mutation rates than low metastatic cells (Cifone and Fidler, 1981). (2) Heterogeneity was found to develop rapidly in cultures containing only one or a few subpopulations of cells. The addition of other subpopulations from the same tumor caused stabilization of the metastatic phenotype as tested in an experimental metastasis assay (Poste et al., 1981). These findings as analyzed with the highly metastatic B16 melanoma are illustrated in Fig. 5B. Observations of clonal instability and clonal subpopulation interactions have been made in various tumor systems both in uitro and in uiuo (Miller et al., 1983; Poste, 1982; Poste and Nicolson, 1983). The observations of clonal instability and of high mutation rates in highly metastatic cell lines support Nowell’s hypothesis of progression being due to acquired genetic lability. The observations of subpopulation interactions introduce a controlling factor from the cells’ microenvironment. Tumor cell subpopulation interactions can influence each other’s growth behavior (Mil-

CANCER METASTASIS

27

ler et al., 1980), drug sensitivity (Miller et al., 1981),immunological properties (Nowotny and Grohsman, 1973; Miller and Heppner, 1980), and metastatic phenotype (Poste et al., 1982; Miller, 1983). They may be exerted via many mechanisms, including (1)metabolic cooperation, a process by which small molecules pass between cells in contact, presumably through gap junctions (Subak-Sharp et al., 1969; Loewenstein, 1979), (2) growth factors and chalones secreted by tumor cells which could influence growth of other cells, and (3)host-mediated mechanisms of immunological nature (Chow and Greenberg, 1980; Miller and Heppner, 1980). There have been other observations which suggest that genetic variability and selection may not be the whole story in tumor progression. For instance, Smith and Sanger (1982) described genomic rearrangements during tumor development that were nonrandom and were precisely reproducible from experiment to experiment. Table VII contains a list of observations where tumor cells were found to shift in their metastatic phenotype, often in a very characteristic manner, far from random. For instance, Vaage (1980) has shown in repeated serial transplantation experiments that individual C3H mammary tumors undergo “progression” in a highly reproducible way: Certain characteristics appeared in the same generations, as if on schedule. Similar observations of shifts or drifts were made by Neri and Nicolson (1981), Schirrmacher et al. (1982a-c, 1983), Dennis et al. (1981b), and Katzav et al. (1983). When studying stability of tumor antigen expression, we observed a clone-dependent variation of tumor antigen expression with time in tissue culture (Schirrmacher and Bosslet, 1982). During successive ip passage in uiuo of uncloned as well as cloned low metastatic parental Eb-type cells, we repeatedly observed a phenotypic shift toward the high metastatic ESb variant phenotype (Schirrmacher et al., 1982c, 1983). Such a shift was associated with changes in the expression of tumor antigens, differentiation antigens, and of Fcy receptors. Table VIII contains data of such a shift in metastatic properties involving multiple phenotypic traits. Another interesting observation of shifts in metastatic capacity associated with a cell surface phenotype change was recently reported from the T10 (H-2k x H-2b) F, sarcoma (De Baetselier et al., 1980; Katzav et al., 1983). Serial transfer of H-2k negative low metastatic clones eventually resulted in the appearance of high metastatic variants that concomitantly expressed H-2k antigens. These observations of shifts in metastatic phenotype could be interpreted in such a way that there is a clone-dependent genetic programing for shifts on one hand and regulatory environmental signals on the other that influence progression. Progression may thus depend on controlling elements both from within the cell and from without (Table VI,3.) Foulds original definition and rules of progression were based on extensive experimental and clinical documentation of how cancers behave. They

TABLE VII PHENOTYPIC SHIFTSI N METASTATTC PROPERTIES A N D POSSIBLE MECHANISMS Observation

Tumor system

C3H mouse mammary carcinoma

Progression of individual tumors during serial transplantation highly reproducible. Clone dependent genetic program? Cyclical variability during progression toward hormone independence. Regulatory mechanisms among various subpopulations? Genetic variability in oitro subject to regulatory influences involving tumor subpopulation interactions Clone-dependent phenotypic drift of metastatic and cell surface properties during growth in tissue culture Repeated observations of shifts during serial transplantation of low metastatic Eb-type cells toward high metastatic ESb-type cells. Shifts associated with the same changes in tumor antigen and differentiation antigen expression. Induction by signals from microenvironment? High-frequency shifts from lectin-resistant nonmetastatic to reexpression of lectin-sensitive and metastatic phenotype. Due to fusion with host cells followed by chromosome segregation. Serial transfer of H-2k-negative clones ends in appearance of H-2k-positive cells concomitantly with acquisition of high metastatic capacity.

GR mouse mammary tumors

B16 melanoma

13762 adenocarcinoma and B16 melanoma

Eb/ESb lymphosarcoma

MDW4 ( W G A h W G A S )

T10 (H-2b x H-2k) sarcoma

28

References Vaage (1980)

Kiang et al. (1982)

Poste et al. (1981)

Neri and Nicolson (1981), Miner et al. (1982)

Schirrmacher (1980), Schirrmacher et al. (1983)

Dennis et al. (1981)

Katzav et al. (1983)

TABLE VIII SHIFTS FROM E b TO ESb

DUFUNG

SUCCESSIVEip PASSAGEOF A TWICE-CLONEDEh CELLLINE

Specific cytotoxicity with CTL Anti-Eb

Tumor line

ce Eb control ESb control

PA4

Anti-ESh

2

71

25

3

75

7 77

Metastaticb capacity

Anti-H-2d

38

36 4

Surface markersc

(%)a

66 80

Thy 1

Lyt 1

Fcy R

Progression 0 TATAE~

I

v

1/40 35/40

+ + -

+

-

+

-

+

Percentage specific 51Cr release in a 4-hr assay at a 40:l effectortarget cell ratio. Frequency of metastases in liver, lung, spleen, or kidney 12 days after sc inoculation of 105 tumor cells into syngeneic DBA/2 mice. c Expression of lymphoid differentiation antigens (Thy 1, Lyt 1)analyzed by cytofluorograph; expression of Fcy R (receptor for IgC) determined in an EA rosette assay. a

30

VOLKER SCHIRRMACHER

were descriptive and did not suggest mechanisms. While later concepts of mechanisms (Table VI,2,3) suggest that progression occurs because new characteristics are acquired during tumor growth, there is still an extreme alternative possibility (Table VI,4), namely, that progression occurs because of changing environmental conditions. Such a situation may indeed exist in hormone-dependent cancers, where evidence was presented, even at the DNA level, for the coexistence of hormone-dependent and hormone-independent subpopulations prior to tumor progression (Michalides et ul., 1982). Isaacs and Coffey (1981) used fluctuation analysis to demonstrate the presence of androgen-independent cells within an androgen-dependent prostatic adenocarcinoma line. It may be interesting to note in this context that human prostatic cancer undergoes a morphological shift during transition to hormone independence (Sinha et al., 1977). High-frequency shifts from lectin-resistant nonmetastatic cells to reexpression of a lectin-sensitive and metastatic phenotype were reported by Dennis et al. (1981a,b), who studied the MDAY-D2 mouse tumor system. It could be shown later (Lagarde et al., 1983; Kerbel et al., 1983) that these shifts were due to a fusion of the tumor cells during local growth with host cells followed by chromosomal segregation. From the many possible segregants, those with metastatic properties were selected by the host and could be recovered from internal organs. The procedure for providing evidence for this type of mechanism is illustrated in Fig. 6. While several investigators have described fusion events between tumor and host cells in uiuo (Goldenberg et al., 1974; Lala et al., 1980), it is only relatively recently that evidence was presented that such a process could lead to the generation of highly metastatic tumor variants (De Baetselier et al., 1981, 1984; Lagarde et al., 1983). Such a process could also be responsible for the interesting finding of Kerbel and associates (Kerbel et aZ., 1980; Frost et aZ., 1981) who reported that injection of strain A mouse tumors into DBA/2 mice resulted in the production of high metastatic DBAI2-type tumors. Progression could thus also be mediated by spontaneous somatic hybridization (Table VI,5) followed by selection of highly malignant segregants. One could imagine that a tumor-host cell hybridization event could lead to major genomic changes involving chromosomal rearrangements, unequal distribution of chromosomes, and chromosome losses during cell division (Larizza and Schirrmacher, 1984). We are intensively investigating such a mechanism in our laboratory as a possible basis for tumor progression in the Eb/ESb model system (Larizza et d . , 1984a,b). In addition to genetic factors there seem to be microenvironmental factors involved in the generation of tumor heterogeneity. Based on observations of high-frequency variant generations in uiuo (Dennis et al., 1981b; Bosslet and Schirrmacher, 1982) and of microenvironmental influences (Schirrmacher et al., 1982a), we have proposed a model in which signals from the tumor

C A N C E R METASTASIS

"to

Drug-mrked

O 0, 0 "io

W G A ~tumor cells

o0ooo," o o

I (CBAIT6T6x DBA/2)Fj ('marked'hosl )

31

WGf] ThgR lumorcellmarkers Oua H-2d

T 6 chromosome

1 1 4 Examine cells for

(I) loss of drug markers (2) acquisltlon of C B A I T 6 T 6 markers

FIG.6. Experimental approach that was used to test whether metastatic tumor variants were derived from a spontaneous fusion between tumor cells and host cells. (Reproduced from Kerbel et al., 1982.)When drug-marked wheat germ-agglutinin-resistant low metastatic tumor cells were inoculated into genetically marked F1 hybrid mice, metastatic wheat germ-agglutinin-sensitive variants developed which carried genetic markers from both the tumor and the host. (Lagarde et al., 1983.)

microenvironment could activate genetic programs within tumor cell subpopulations causing phenotypic changes similar to those seen in normal tissue differentiation [ Schirrmacher, 1980; see Fig. 7B (Altevogt et al., 1982)l. In this model, the influence of the host is not only selective but also inductive. The model suggests (1) that multiple phenotypes can become expressed by one common genotype and (2) that microenvironmental soluble factors as well as cell-matrix contact (Reid, 1982) or cell-cell contactmediated signals could have a regulatory role on tumor cell phenotypes and expression of heterogeneity. It should be kept in mind that heterogeneity is not a property exclusive to tumors but is seen in normal tissues as well. That neoplastic cells could give rise to variants through a process resembling normal tissue differentiation was first suggested by Pierce (1974) who found that single cells isolated from a teratocarcinoma differentiated in uivo into a wide variety of tissues representing all three germ layers. Bennet (1983) recently reported on differentiation and dedifferentiation processes in a cloned mouse B16 melanoma line which may be relevant to this discussion. Differentiation (pigment production) was followed in inducing media by time-lapse films allowing observations within single cells. Differentiation appeared unrelated to cell division and could be reversed in a proportion of cells. Dedifferentiation was associated with cell proliferation, so that most

32

VOLKER SCHIRRMACHER

B

F

C

FIG.7. (A) Model for the differentiation of melanoma cells; 0 and 1 represent states in which a set of functions associated with melanin synthesis are turned off and on, respectively. Cells can change their state stochastically either way with rate constants ko and kl (cells per cell per unit time). These rates will depend on the extracellular and intracellular milieux. (Reproduced from Bennett, 1983.)(B) Model showing how microenvironmental signals could cause shifts in tumor cell phenotypes by activating preformed genetic programs (GP1-GPS). A-F, Phenotypic markers; S1-S6, signals from the microenvironment. (Reproduced from Schirrmacher, 1980.)

pigmented clones were small and most unpigmented clones were large. These findings were accomodated by a model (Fig. 7A) in which functions associated with differentiation could switch on and off. It was further suggested that an inhibition of the “off’ transition would build up in the “on” state. Also relevant in this context may be the findings by Raz and Ben-Ze’ev (1983) of modulation of the metastatic capability in B16 melanoma by cell shape. During normal cell differentiation many genes are apparently switched off by DNA methylation. Tumor cells on the other hand seem to reexpress some of these genes, as exemplified by expression of embryonic antigens or by

CANCER METASTASIS

33

production of “wrong” hormones. Gene activation by DNA demethylation could thus be yet another important factor in tumor progression (Table VI,6). This may be particularly relevant in the context of chemotherapy, where drugs are used, many of which are known to affect DNA methylation. The potential danger of facilitation of tumor progression by cancer therapy has been pointed out by Kerbel and Davies (1982). V. Angiogenesis

Several important aspects of metastasis have not been discussed yet: angiogenesis, invasion, dormancy, and host immune responses. Because of their importance, a separate paragraph is devoted to each of these phenomena. Most solid tumors appear to pass through two phases of growth. In the avascular phase, a tiny tumor nodule will grow up to a few millimeters in diameter only and usually will not be invasive. In the vascular phase, new capillary sprouts are induced from the host which grow toward the tumor. When these vessels penetrate the tumor, it begins to grow rapidly and to invade (Folkman, 1974a,b) (Fig. 8). The extensive vascularity of solid tumors has been recognized for over 100 years but little progress was made because of the lack of an experimental system with which to study angiogenesis. Only in the last 15 years has the importance of this phenomenon for autonomous tumor growth been demonstrated and only within the last 6 years has the possibility of chemical interference been apparent (Langer and Murray, 1982). This progress was made with the introduction of new model systems, such as (1) tumor implantation into the anterior chamber of the rabbit eye (Gimbrone et al., 1974) (where tumors remain small because new vessels cannot reach them through the aqueous humor), (2) tumor incubation inside a millipore chamber (demonstrating that vascularization is induced by a diffusible chemical substance) (Greenblatt and Shubik, 1968), and (3) cartilage implantation on the highly vascular chick chorioallantoic membrane (demonstrating antiangiogenetic factors in a tissue into which tumors cannot metastasize) (Eisenstein et al., 1975). Among human neoplasms in situ carcinoma is most analogous to the avascular phase. In situ carcinoma of the cervix or bladder may exist for years and this prolonged period is commonly free of metastases. Almost simultaneously with neovascularization, in situ carcinomas start to invade through the tough basement membrane, (BM) (Folkman, 1976).There is also experimental support for a close relationship between angiogenesis by the host and invasion by the tumor so that it seems difficult to decide which of the two processes comes first. As shown in Fig. 8, there are wide gaps between the endothelial cells in the advancing tips of proliferating capillaries and the

34

VOLKER SCHIRRMACHER

I

1

Tumor dlrwmlnatlon

"Antianglogonrrlr" I

FIG. 8. Illustration of three concepts basic to tumor growth and metastasis. (1) Solid tumors pass through an avascular and a vascular stage. (2) New host capillaries are stimulated b y a humoral angiogenetic factor (TAF) released by the tumor; rapid growth follows; upper left magnification shows endothelial cells at the tip of a new tumor-induced capillary, which originated from a venule; at the tip of the capillary sprout the basement membrane (BM) is fragmented so that tumor cells (T) from the primary mass may escape through these gaps into the circulation. (3)If angiogenesis is inhibited, further tumor growth is blocked and the tumor may enter a dormant phase. (Modified according to Folkman, 1974a.)

foremost part of the capillary has no basement membrane. This may explain the ease with which tumor cells can enter the circulation from the vascularized tumor (Fig. 8, inset). The processes of angiogenesis and invasion have been analyzed in detail in the rabbit cornea model (Folkman, 1976). The original tumor implant grows slowly during the avascular phase. When new capillaries grow into the cornea, they are able to intrude between tightly packed layers of collagen, an invasive property that was not seen with the tumor cells themselves. Once the tumor has become vascularized, it is very destructive and capable of invading through all layers of collagen to the outer aspect of the cornea. Folkman and associates have been able recently to culture human and bovine capillary endothelial cells, to clone them, and to study their growth and differentiation in long-term culture (Folkman and Haudenschild, 1980). The most interesting findings can be summarized as follows: (1)While aortic

CANCER METASTASIS

35

endothelial cells can be grown in regular culture medium, capillary endothelial cells grow only slowly and then die. In tumor-conditioned medium, however, these cells grow rapidly with a doubling time of 28 hr and continue to proliferate for as long as the tumor-conditioned medium is present. (2) Cloned capillary endothelial cells, cultured in tumor-conditioned medium, form capillary tubes similar to capillaries in uiuo, although there may be an inside-out type of conversion. The information necessary to develop an entire network in uitro with branch points and anastomoses thus seems to be contained within one cell type. It was recently shown that this whole process of capillary network formation is prompted also by collagen matrices (Montesano et al., 1983). Capillary endothelial cells will become very useful to distinguish between direct and indirect angiogenesis factors. Compounds like formic acid or silica which attract macrophages or other white cells might be recognized as indirect factors that are unable to directly induce capillary formation in uitro but which could stimulate other cells to do so. This in uitro procedure also will be likely to allow better characterization of antiangiogenesis factors and their mechanism(s) of action. With regard to angiogenesis inhibition and its effect on growth of primary tumors and their metastases, exciting results have been recently obtained and reported by Folkman and his associates (Folkman et al., 1983). They had previously demonstrated that angiogenesis in uiuo can be promoted by heparin and inhibited by protamine, an antagonist of heparin (Taylor and Folkman, 1982). Now they report that oral administration of heparin resulted in the release of a nonanticoagulant heparin fragment in the serum, a hexasaccharide, which when given together with cortisone brings about a very potent inhibition of angiogenesis. Neither the heparin fragment nor cortisone alone showed this effect. When tumor-bearing animals were given heparin and cortisone (dissolved in their drinking water), it was found that large tumor masses regressed and metastases were prevented. Although evidence was given that the antitumor effect was due to angiogenesis inhibition, the mechanism of the synergistic effects of heparin and cortisone was not resolved. Also unexplained was the finding that many but not all tumors tested responded to this kind of therapy. It was suggested that the nonresponder tumor types may perhaps be able to degrade heparin or in some way interfere with the effect of heparin and cortisone on endothelium. Another substance that can block tumor angiogenesis in uiuo, thereby restricting tumor growth, was recently reported to be present in shark cartilage (Lee and Langer, 1983). Advances in the study of capillary endothelial cell growth in uitro and of powerful angiogenesis inhibition in uiuo have thus demonstrated how re-

36

VOLKER SCHIRRMACHER

search focused on essential steps of the metastatic cascade can lead to a better understanding of mechanisms of metastasis formation and to new treatment strategies.

VI. Cancer Invasion

A. GENERALCONSIDERATIONS Higher organisms are composed of a number of tissue compartments that are separated from each other by extracellular matrices, such as basement membranes and interstitial stroma. During metastasis, tumor cells must traverse these matrix barriers as they cross tissue bounderies. The basic structures that serve as barriers to invasion can be divided into three broad categories: (1) organ parenchymal cells, which include epithelium, endothelium, and mesothelium; (2) basement membranes, which separate these cells from the underlying stroma; and (3) connective tissue consisting of connective tissue cells embedded in their extracellular matrix proteins (see Table IX). Carcinoma is by far the most commonly occurring form of cancer and is a neoplasm of epithelial cell origin. Direct invasion is the first and most crucial step in the malignant process and is defined in carcinomata by local disruption of basal lamina with tumor cell infiltration into the underlying interstitial stroma. After traversing the stroma, tumor cells gain access to lymphatics and blood vessels for further dissemination. Tumor cells also have to cross basement membranes when they move into or out of blood vessels (intravasation or extravasation) such as venules and capillaries. In the target organ, where metastases are initiated, tumor cells that have extravasated must migrate through the perivascular interstitial stroma before colony growth occurs in the organ parenchyma. It is obvious that the extracellular matrix (ECM) is a mechanical barrier of the host that has to be penetrated at multiple stages during the metastatic process. The mechanism of tumor invasion is therefore studied mostly on the level of the interaction of tumor cells with ECM. The basement membrane (BM)-probably the most important barrier-is a tough, elastic structure that is impermeable to colloidal carbon (0.5 pm). The insoluble nature of the BM is in part due to the unique arrangement of type IV collagen molecules (Miller, 1976; Timpl et al., 1978) which are interconnected at their end regions to form a hexagonal network. In addition to collagen fibers, the BM matrix contains specific glycoproteins, such as fibronectin or laminin; proteoglycans, such as heparin, heparan sulfate, or chondroitin sulfate; and elastin (Kefalides, 1978). The ECM is thus a network consisting of a mixture of soluble and insoluble cross-linked mac-

TABLE IX BIOCHEMICALPARAMETERS OF HOSTTISSUEPENETRATION BY METASTASIZING TUMOR CELLS" Host tissue

Structural element

Biochemical components

Degradative enzymes ~

~

~~~~

Desmosomes Tight junctions Anchoring fibrils

DisuKde-bonded proteins

Plasmin Plasminogen activators Serine proteases

Interface between different tissues

Basement membranes

Fibronectin laminin Type IV and V collagens Proteoglycans

Mast cell serine protease Type IV collagenase l(metd1oproteases) Type V collagenase Heparanase (endoglycosidase)

Connective tissue stroma

Anchoring fibers Collagen fibers Elastic fibers

Fibronectin, glycoproteins Type I, 111, and V collagens Elastin Proteoglycans

Epithelium Endothelium Mesothelium

p

Cathepsins and thiol proteases Leukocyte elastase Classical collagenase Bone Cartilage

Ground substance Mineralized matrix

.Osteonectin, elastin Type I and I1 collagens Proteoglycans and GAGS Hydroxyapatite ~

For reviews see Liotta et al. (1982a,b), Jones and De Clerck (1982), and Pauli et al. (1983).

~~~~~~

38

VOLKER SCHIRRMACHER

romolecules produced by specialized cell types. Detailed structural information of ECMs is an important prerequisite for understanding their function under normal and pathological conditions. The basal lamina has been shown to be a product of epithelial cells that serves to stabilize epithelial cell differentiation and orientation during organogenesis. In addition, this structure is most likely important for cell anchorage in uiuo and so may play a central role in growth regulation. As the early stages of oncogenesis involve deregulation of cell differentiation, orientation, and proliferation, it is possible that the gradual loss of basal lamina integrity that precedes its complete disruption during carcinoma development may be involved in neoplastic disorganization prior to the onset of malignant invasion. It is now known that the tumor can modify the matrix in the following ways: (1) tumor cell synthesis of matrix components, (2) degradation of matrix components associated with tumor invasion, and (3) stimulation of host cells to increased production of matrix components such as in desmoplasia (see Liotta, 1982). The matrix becomes locally permeable to cell movement not only during malignant invasion but also under normal physiological conditions, such as wound healing, tissue remodeling, neovascularization, and inflammation. Information on tumor cell invasive behavior in uiuo is mostly obtained from histologic and ultrastructural studies of fixed tissue sections because the tissues are normally opaque and thus exclude direct observations of tumor cells. There are a few exceptions, however, where the dynamics of tumor cell behavior have been followed in situ using transparent tissues such as mesentery, chorioallantoic membrane (CAM), rabbit ear chamber, hamster cheek pouch, and others (Amstrong, 1980). There are certain aspects of tumor invasion that can only be studied in uiuo, in particular those that deal with the interference of host inflammation and other defense reactions. Such inflammatory reactions at the site of tumor invasion could favor the process by the release of tissue lytic enzymes and mitogenic or angiogenic factors from lymphocytes or macrophages. On the other hand, invasion could be retarded by the recruitment of inflammatory cells with tumoricidal or tumoristatic activity exerted by activated macrophages, NK cells, or T cells. The correlation of the course of invasion with the type of host response is an important but very complex issue that obviously requires further improvement of technologies. Different in uitro assay systems for studies on cancer invasion are listed in Table X. Their respective advantages or disadvantages have been discussed elsewhere (Liotta and Hart, 1982). All of the methods in current use are simplified versions of in viuo processes and undoubtedly far away from the natural situation. However, most of these studies are performed to overcome some of the major obstacles of the in uiuo work. A main point concerns the

39

CANCER METASTASIS

ability to quantify the kinetics and the extent of tumor invasion in an accurate and reproducible manner. For this particular purpose tumor cells have been labeled radioactively, for instance with 51Cr, 991nT~,[3H]thymidine, [ 1251]iododeoxyuridine,and 35Se-labeled methionine. Another important advantage of some in vitro systems is the ability to recover the invasive tumor cells, for instance, after they have penetrated a basement membrane, and to compare them with noninvasive tumor cells from the same starting population. In view of the phenomena of clonal instability, tumor heterogeneity, and subpopulation interactions, it may become an important issue to quantify changes in the proportion of invasive cells at different stages of tumor development. TABLE X In Vitro ASSAYSFOR CANCEH INVASION Tumor cells cocultured with

References

Endothelial cell monolayers, hepatocyte cultures, subendothelial ECM, extracted bone cartilage

Liotta et al. (1977), Kramer and Nicolson (1979), Zamora et al. (1980), Jones and De Clerck (1980). Pauli et al. (1981), Roos et al. (1981), Vlodavsky et al. (1983a,b)

Organ cultures, tissue fragment cultures, rotation-mediated cell aggregates

Easty and Easty (1974), Schleich et al. (1974), Scher et a1 (1976), PourreauSchneider et a1 (1977). Mareel et a1 (1975, 1979). Noguchi et al. (1978). Schirrmacher et al. (1979), Lohmann-Matthes et a1 (1980), Poste et a1 (1980), Schirrmacher et al (1982d)

Membrane penetration systems, two-chamber systems

Hart et a1 (1978), Tchao et al. (1980), Thorgeirsson et a1 (1982)

Artificial vessel walls

Jones et al. (1981), Bogenmann et al. (1983)

40

VOLKER SCHIRRMACHER

With regard to the type ofnormal tissue substrate used, different types of in uitro assays have been established (Table X). These are, for instance, monolayer cell culture systems, three-dimensional cell aggregates mediated by rotation, organ cultures of a variety of host tissues, basement membranecontaining systems, and artificially constructed tissues grown on filters or nylon meshes. The utility of any particular experimental model obviously depends on the question being asked. An important point for the strategy of experimental analysis of tumor invasion is the requirement of parallel in uitro and in uiuo studies. The tumor cells to be used in uitro have to be shown to be invasive in uiuo. It has to be verified that the many changes that cells might undergo when initiated in culture do not affect the trait of interest, namely, their invasiveness in situ.

B. SEQUENTIAL STEPSI N CANCERINVASION A three-step hypothesis has been proposed (Liotta et al., 1982a) to describe the sequence of biochemical events during tumor invasion of ECM (Fig. 9). The steps are (1) attachment to the matrix (probably involving tumor

FIG.9. Three-step hypothesis of tumor cell invasion of extracellular matrix. (A) Attachment may be mediated by specific attachment factors such as laminin (in the case of basement membrane) and fibronectin. (B) Local degradation of matrix by tumor-associated proteases. (C) Tumor cell locomotion into the region of matrix modified by proteolysis. (Reproduced from Liotta et al., 1982a.)

C A N C E R METASTASIS

41

cell surface receptors binding to specific adhesive glycoproteins), (2) local proteolysis, and (3) tumor cell locomotion into the region of the matrix modified by proteolysis (for reviews see Liotta and Hart, 1982; Pauli et al., 1983; Jones and DeClerck, 1982; Strauli et al., 1980; Liotta and Hart, 1982). Three major theories have been proposed for describing the pathological interactions between transformed cells and host stroma during tumor invasion: (1) the mechanical pressure theory, (2) the enzymatic theory, and (3) the migratory theory. These concepts are not mutually exclusive, but may be interdependent and coordinated with one another. The relative contribution of each of these factors may vary depending on the type of tumor and the type of host stroma. The mechanical pressure theory proposes that the increased pressure of expanding primary tumors causes rapidly proliferating tumor cells to force their way into alternative locations within host tissue. These possibly seek for ways of least tissue resistance and often migrate along natural cleavage planes such as collagen fibers, blood or lymph vessels, or nerves. Large-scale studies on autopsy material, however, show no evidence of a correlation between tumor size, mechanical pressure, and invasion and metastasis: There are examples of early metastases from very small primary tumors which could only generate small expansion forces, as well as examples of large tumors with pressure atrophy of the parenchyma and lack of metastases. Pressure thus probably plays only a minor role in tumor invasion in most cases. The enzymatic and migratory theories gain support from a number of clinical and experimental observations. It appears reasonable to suppose that the translocation of a tumor cell from place A to place B in host tissues involves active locomotion. However, this can only be achieved when host barriers provided by extracellular matrices are altered to allow passage for the invading cells. Such alterations require the action of proteolytic enzymes. Table IX contains a list of the most common degradative enzymes involved in host tissue penetration by metastasizing tumor cells. Enzymatic action might be sufficient to focally lower the level of molecular organization and to reduce the physical resistance of host matrices. It has been shown for instance that proteinases that only act as collagen “cross-linkases” or “proteoglycanases” can transform the ECM from an insoluble into a more fluid state. A detailed model of matrix degradation has been proposed by Pauli et al. (1983). This is shown in Fig. 10. Proteolysis of the host ECM can be achieved either by enzymes derived from the tumor cells themselves or by enzymes from host cells, such as vascular endothelial cells (in angiogenesis), fibroblasts, macrophages, and mast cells. It has been suggested that some tumors may elaborate factors that directly stimulate fibroblasts to increased collagenase synthesis and secretion (Bauer et al., 1977).

42

VOLKER SCHIRRMACHER

A

3

B

D

FIG.10. Model of degradation of extracellular matrix (ECM) at the tumor invasion zone. (A) ECM consisting of collagen fibers (1)and GAGS (2) that provide swelling pressure. (B) Higher magnification of (A) showing covalent cross-links (4) between neighboring collagen molecules (3).Proteoglycans aggregated with hyaluronate (5) are restricted from swelling by the intact collagen network. (C) Collagen fibers are degraded by two enzymatic pathways: (i) proteinases (i.e,, cathepsins, elastase, plasmin, thrombin) act as cross-linkages (4) to liberate collagen monomers (6)which then denature (7), solubilize and become susceptible to many proteinases; (ii)collagenases specifically cleave the collagen triple helix (8);the resulting fragments denature (9) and become further degraded. (D) Collagen and proteoglycan degradation (10)transforms the matrix from a solid to a fluid state thus allowing locomotion and tumor cell penetration. (Reproduced from Pauli et ~ l . 1983.) ,

The most popular proteolytic enzymes studied in relation to tumor invasion are collagenases (Liotta et al., 1982a,b), plasminogen activators, cathepsins, elastases (Jones and DeClerck, 1980; Pauli et al., 1983), and proteoglycan-degrading enzymes (Kramer et d.,1982; Nakajima et al., 1983;

CANCER METASTASIS

43

Vlodavsky et al., 1983b). A summary of biochemical parameters of host tissue penetration and the involvement of respective degradative enzymes is given in Table IX. The migratory theory stresses the importance of cellular locomotion of tumor cells. This may be brought about by a creeping motion of isolated cells or by the moving boundary of a sheet of cells. Amoeboid deformability of single cells may allow them for instance, to insert an arm of cytoplasm (pseudopod) through endothelial cells as seen in uitro (Vlodavsky et al., 1983c)and to transfuse the rest of the cell body through this hole. Migration may be initiated by loss of intercellular adhesion, facilitating separation of tumor cells from each other and allowing tumor cells with increased motility to invade normal tissues. Common experiences with cell separation procedures show that enzymatic dissociation of intercellular bonds provided by fibronectin, heparan sulfate, hyaluronate, or divalent ions may contribute to detachment of individual cells or cell clusters from solid tumors. Of special importance may be cell surface-associated proteinases which lead to increased cell dissociation, increased doubling times, and abnormal intercellular junctions in carcinoma cells in uitro (Pauli and Weinstein, 1982). Cell locomotion requires sequential attachments and detachments of localized areas of the cell surface to a substrate (Toole, 1981). Based on a biochemical analysis of the attachment sites (footpads) of cells, Rollins and Culp (1979) have postulated that cell substrate adhesion is mediated by fibronectin and heparan sulfate, whereas subsequent detachment, which is necessary for cell movement, involves hyaluronate and chondroitin sulfate. A role for hyaluronate in cell movement within a cellular stroma has been suggested from various studies of normal and tumor cells. It seems that tissue swelling ahead of migrating cell populations depends on the presence of large concentrations of hydrated hyaluronate. The studies of Toole et al. (1979) in the V2 rabbit carcinoma model have shown that the hyaluronate concentrations are most elevated in the invasion zone. Tumor cell locomotion has been shown to be influenced by chemotactic factors (Romnaldez and Ward, 1975). Chemotactic factors can induce directional motility of tumor cells in the Boyden chamber assay. In addition, such factors can stimulate cell swelling and also foreign surface adhesiveness (Varani et al., 1981; Lam et al., 1981). In the Walker 256 carcinosarcoma, treatment with either the C5a-derived chemotactic peptide, the synthetic tripeptide NMFP (N-formylmethionylleucylphenylalanine),or with 12-0tetradecanoyl phorbol ester induces a rapid, transient adherence response. From drug inhibition studies it was suggested that this adherence response is mediated by lipoxygenase metabolites of arachidonic acid (Varani, 1982). In uiuo experiments in animals bearing circulating tumor cells demonstrated that chemotactic factor injection caused a localization of tumor cells

44

VOLKER SCHIRRMACHER

at the site of factor application (Ozaki et al., 1971; Orr et al., 1981). It has therefore been suggested that the chemotactic response of tumor cells could play an important role in cancer metastasis (Lam et al., 1981). Tumor cells localizing at metastatic sites in oiuo may be influenced in a similar way as leukocytes, which become localized at sites of inflammation. The sequence of events which characterize invasion can be summarized as follows: 1. Tumor cell shedding. Detachment of tumor cells from the primary mass could be brought about by loss of intercellular junctions (desmosomes), alterations in chemical composition and physical properties of the cell surface coat, and loosening of cell substrate interactions. 2. Tumor cell attachment to the matrix of interstitial connective tissue stroma or of basement membranes. Attachment may be mediated by specific adhesive glycoproteins such as collagens, laminin, fibronectin, and others. Laminin is a cross-shaped molecule (Timpl et al., 1978) that has a binding site for a cell surface laminin receptor (Lesot et al., 1983; Terranova et al., 1983; Malinoff and Wicha, 1983) as well as binding sites for fibronectin and heparan sulfate. Fibronectin is an S-S-bridged two-chain protein molecule with domains for binding of collagens, heparin, actin, fibrin, and for cells (see Vartio et al., 1983). Its possible role in cell adhesion and invasion has recently been discussed (Ruoslahti, 1984). 3. Local proteolysis. Increased proteolytic activities at the invasion front cause local alterations in the surrounding ECM making it easier to penetrate. Such proteases may be tumor cell or host cell derived and may degrade attachment proteins as well as the structural collagenous proteins of the matrix. Collagenases and cathepsins, elastase, and other neutral proteinases are the types of enzymes most frequently associated with matrix destruction and invasion. In some tissues this process appears to be regulated by natural inhibitors of proteases. 4. Tumor cell locomotion. Tumor cells migrate into the zones affected by proteolysis. They possibly move as aggregates along guidance tracks provided by host structures. Migration seems to be preceded by swelling of glycosaminoglycans, particularly hyaluronate, in the matrix of connective tissue (Pauli et al., 1983). C. TISSUERESISTANCETO INVASION A N D HOST DEFENSEREACTIONS While most loose connective tissues and bone are readily invaded by malignant tumors, cartilage and other avascular tissues such as aorta, heart valves, cornea lens, and epithelia are relatively resistant (Eisenstein et al., 1973, 1975; Kuettner et al., 1978; Waxler et al., 1982). This selective resistance of certain tissues to invasion may be determined by structural properties as well as by tissue-specific antiinvasive factors (AIFs) (Kuettner et al.,

CANCER METASTASIS

45

1977). Bovine nasal cartilage has been reported to contain at least three distinct proteinase inhibitors, an inhibitor of collagenases (MW 22,000), an inhibitor of thiol proteinases (MW 13,000), and an inhibitor of trypsin (MW 7000) (Rifiin and Crowe, 1975). Cartilage-derived AIFs also express inhibitory activities against neutral metalloproteinases that cleave collagen types IV and V and against human neutrophil elastase. In addition, AIF contains antiproliferative activity directed against vascular endothelial cells in culture and can inhibit tumor cell penetration of native connective tissue such as human amnion (Thorgeirsson et al., 1982). Another example of a host-tumor interaction is the deposition of fibrin which has been demonstrated within tumors by immunofluorescence and electron microscopy. It appears that tumors bring about fibrin deposition and remodeling in their own vicinity by releasing molecules which (1) render blood vessels permeable to plasma proteins; (2) coagulate extravasated fibrinogen to fibrin; and (3) remodel deposited fibrin by activating the fibrinolytic system (Dvorak et al., 1982; Cederholm-Williams, 1981). The cellular immune response may contribute to these processes. Fibrin deposits that have been planted in normal tissue have been found capable of inducing angiogenesis and desmoplasia. It has therefore been reasoned that the “tumor angiogenetic factor” of Folkman (1974a,b)may not be a single product unique to tumors but rather a series of mediators, closely linked to the physiological process of wound healing, which leads to deposition and modulation of tumor fibrin or fibronectin (Dvorak et al., 1982). Tumor angiogenesis and desmoplasia may be regarded as specialized examples of wound healing, a process whose cardinal features include fibrin and fibronectin deposition, angiogenesis, and fibrous connective tissue (scar) formation. Recruitment of cells of the host defense system to sites of tumor growth and invasion has also been found to limit tumor growth. Agents such as bacillus Calmette-Gukrin (BCG) and Corynebacterium paruum, for instance, which are used in cancer immunotherapy, seem to exert their effect via their ability to elicit inflammatory reactions and recruitment of tumoricidal activated macrophages. Figure 11 illustrates pathophysiological findings of tumor stroma development at different stages of tumor development using the immunogenic diethylnitrosamine-induced line 1 hepatocarcinoma, a transplantable tumor of strain 2 guinea pigs (Dvorak et al., 1982). Shortly after tumor transplanatation, the development of a fibrin-gel meshwork was observed which was followed by fibroblast invasion and collagen production. By day 8, a cellular antitumor immune response developed which involved lymphocytes, basophils, and monocytes forming prominent cuffs around venules at the periphery of the now fibrous connective tissue that enveloped tumor cell clumps. Macrophages were particularly prominent in peripheral zones of hemorrhage where they were engaged in pha-

46

VOLKEH SCHIRRMACHER

EARLY

PHASE 8-13 D A Y S

LATE

FIG.11. Kinetics of tissue reactions toward a transplanted immunogenic tumor in the guinea pig showing early fibrous reaction and later immunological reactions eventually leading to tumor necrosis. (Heproduced from Dvorak et al., 1982.)

gocytosing extravasated erythrocytes and debris. Extensive inflammatory

cell invasion of the inner tumor cell mass was never observed. These tumors were apparently rejected by a mechanism involving extensive damage of the microvasculature which led to tumor infarction (Dvorak et al., 1979a,b). VII. Tumor Dormancy

The dormant tumor state has been defined as one in which tumor cells persist in a clinically normal host for a prolonged period under growth restraint, with little or no increase in the size of the tumor cell population (Wheelock et al., 1981). There are numerous clinical reports that provide circumstantial evidence for the existence of dormant tumor states in man. One-third of the mortality from breast cancer, for instance, occurs more than 5 years after primary treatment (Adair et al., 1974). One-fifth of patients with cutaneous melanoma who develop recurrence do so after an interval greater than 5 years (Holland and Frei, 1973). A primary melanoma and a metastatic melanoma occurring years later were shown to have common chromosomal

CANCER METASTASIS

47

markers (Balaban et al., 1982). By means of a monoclonal antiidiotypic antibody, small numbers of B lymphoma cells were identified in the circulation of patients during a clinical phase of remission (Hatzubai et al., 1981). Both the preceding primary tumor and the succeeding secondary had the same idiotype. Three principally different mechanisms of tumor dormancy have been distinguished in animal models: (1)avascularity and sequestration of tumor cells, (2) constitutive dependency of tumor cells on growth factors, and (3) immunologic restraint. The main reasons for mentioning tumor dormancy in the context of metastasis are (1)that this phenomenon illustrates that tumorhost interactions can reach a state of balance fbr prolonged periods, (2) that this can possibly be exploited to improve host control of tumor cells that survive cancer treatment, and (3)that it is important to find out what terminates a dormant tumor state thus abrogating the state of balance with the consequence of destruction of either the tumor or the host. Since these have been reviewed in detail elsewhere (Wheelock and Robinson, 1983), they shall be discussed only briefly here. Small tumor nodules that do not develop a vascular network will be limited in size because of deficiency in nutrient diffusion. Such avascular tumor microspheres represent one type of tumor dormancy which has been demonstrated in uitro (Folkman and Hochberg, 1973) and in viuo (Gimbrone et al., 1972). Sequestration of tumor cells within a capsule elaborated by the host can also inhibit tumor cell proliferation or invasion (Dvorak et al., 1979a). A dormant tumor state can also be due to a constitutive dependency of tumor cells on growth factors, best known from hormone-dependent breast cancer (Noble and Hoover, 1975).The growth restraint will be broken, however, if tumor variants develop which elaborate their own growth factors and thus become autonomous (Yuhas and Tarleton, 1978). The first model of tumor dormancy by immunological restraint was demonstrated by Eccles and Alexander (1975). Fibrosarcoma cells that had metastasized to the lung prior to surgical excision of the primary tumor remained in the lung for many months in a dormant state. Immunosuppressive intervention, such as whole body irradiation or thoracic duct drainage, resulted in the outgrowth of lung metastases. The most intensively investigated animal model of tumor dormancy is probably that of Wheelock et al. (1982) who investigated the immunological restraint mechanisms of DBA/2 mice, which had been immunized ip with mitomycin C-treated syngeneic lymphoma cells (L5178Y) and subsequently challenged ip with viable L5178Y cells. Some mice remained clinically normal for many months but then suddenly developed ascitic tumors. It was found that a strong peritoneal cytotoxic T lymphocyte response was responsible for the establishment of the dormant tumor state. This then gradually waned but it could be reelicited in viuo or in uitro by reexposure to

48

VOLKER SCHIRRMACHER

tumor antigen. The maintenance of the dormant tumor state was ascribed to a synergistic cytolytic activity by peritoneal T lymphocytes and macrophages (Robinson and Wheelock, 1983). As the mouse proceeded through the dormant tumor state (2 months) and eventually developed an ascitic tumor, changes were observed in the tumor cells’ phenotypic characteristics as well as in the host response (Wheelock et al., 1982). In our own studies with the L5178YE lymphoma (Eb), we found that a state of tumor dormancy could be established without preimmunization if the cells were inoculated into the hind footpads of syngeneic mice, a site where the tumor cells did not grow (Schirrmacher et al., 1982a). When such tumor-inoculated mice were challenged with viable tumor cells sc in the back, a site where the tumor normally grows, no tumor growth was observed indicating that these animals had developed a status of antitumor immunity. However,in about 20% of thus challenged animals, an outgrowth of tumors was observed from their footpads. Termination of tumor dormancy in the footpad and tumor growth were also observed when animals previously inoculated with E b tumor cells in the footpad were challenged unspecifically with allogeneic normal cells sc in the back (Schirrmacher et al., 1982a). These findings indicate that dormancy due to immunological restraint can be “overcome” not only by immunosuppression (Eccles and Alexander, 1975) but also by immunostimulation (Schirrmacher et al., 1982a). Immunologically controlled tumor dormancy obviously represents a delicate state of balance that is dependent on properties of the host (local microenvironment, status of the immune system) and on properties of the tumor cells (immunogenicity, metastatic capacity). The dependency of tumor dormancy on tumor cell properties was obvious from the fact that a spontaneous metastatic variant from the E b tumor cells (ESb, Schirrmacher et al., 1979), when inoculated into the footpad, behaved differently in that they grew out and metastasized.

VIII. Host Immune Responses in Metastasis

A. SPONTANEOUS REGRESSIONSIN CANCERPATIENTS Host control of tumor cell growth is best illustrated clinically in cases of complete disappearance of a histologically identified tumor in the absence of therapy. Such spontaneous regressions have intrigued clinicians for many years. Cases of spontaneous regression have been reported for most types of human cancer (Everson and Cole, 1966; National Cancer Institute Monograph, 1976). Overall, however, it seems a rare event. Sixty percent of documented cases have occurred in four types of cancer: malignant melano-

CANCER METASTASIS

49

ma, hypernephroma, choriocarcinoma, and neuroblastoma. In some of the cases, tumor regression has been preceded by bacterial or viral infection, but in the majority this has not been the case. Malignant melanomas represent less than 1% of all cancers but account for 11% of reported cases of spontaneous regression. Approximately 40% of such patients with spontaneous regression were apparently cured with no recurrence on long-term follow-up. Regression or partial regression of cutaneous malignant melanoma was associated with a dense lymphocyte infiltrate (McGovern, 1975). After regression of the tumor, the lymphocytes disappear leaving behind scar tissue. In hypernephroma, cases have been reported of spontaneous regression of pulmonary metastases following nephrectomy for hypernephroma (Garfield and Kennedy, 1972). Such regressions appeared to be due to immunological rejection. Spontaneous tumor regression has also been observed in animal tumors whether spontaneous in origin (Miller and Olsson, 1971), virus induced (Dietz et al., 1977), or chemically induced (Rice, 1972). Such regressions were all mediated by immunological mechanisms. Nevertheless, there may be other mechanisms that could lead to spontaneous regression, for instance, induction of differentiation (as observed in neuroblastoma or in lymphomas) or deprivation of nutrients. Research on mechanisms of spontaneous regression of tumors, although difficult for logistical reasons, could lead to a better understanding of endogenous types of control of neoplastic growth and might lead to the development of ways to enhance these mechanisms for more effective cancer therapy.

B. EXPERIMENTAL STUDIES 1. Unspecific or Natural Immune Responses A close association between levels of natural killer (NK) cell activity and the ability of the host to eliminate circulating tumor cell emboli has been reported (Hanna and Fidler, 1980; Hanna, 1982).This came out from studies in hosts with low NK activity (such as beige mice) or hosts with high NK activity (such as nude mice). Furthermore, adoptive cell transfer studies demonstrated that NK cells were effective in viuo in destroying circulating tumor cells before their extravasation, whereas they exerted only a minimal inhibiting effect on already established micrometastases (Hanna and Fidler, 1980). Adoptive transfer of a cloned cell line with NK activity markedly inhibited lung colony formation of B16 melanoma cells when inoculated intravenously into NK-deficient syngeneic mice (Warner and Dennert, 1982). In the same study it was shown that the incidence of radiation-induced thymic leukemia in C57BL/6 mice was dramatically reduced by iv

50

VOLKER SCHIRRMACHER

inoculation of cloned NK cells 3 months before the normal onset of leukemia. That NK cells may exert selective pressure on disseminated tumor cells during spontaneous metastasis is suggested from findings in several tumor systems indicating that tumor cells isolated from metastatic lesions showed a higher level of NK-cell resistance than tumor cells from the primary lesion (Fogel et al., 1979; Gorelik et al., 1979b). It has to be kept in mind, however, that not all malignant cells are readily susceptible to NK-cell-mediated lysis. Whether NK cells play a role in the elimination of circulating tumor cells in patients is not known at present. Platelet aggregation and fibrin coating of the surface of tumor cells may be one of the mechanisms by which hematogenously spread tumor cells are protected from destruction by N K cells. Gorelik et al. (1984) recently showed that the antimetastatic effects of anticoagulant drugs, such as heparin and prostacyclin, depended on levels of NK activity in the host. It was therefore suggested that anticoagulant drugs may exert their antimetastatic effects by making tumor cells more vulnerable to NK-cell lysis rather than by blocking adherence of tumor cells to vascular endothelium. Polymorphonuclear leukocytes could also contribute to natural defense reactions against metastases. Such cells were recently reported to play a role in the pulmonary clearance of arrested B16 melanoma cells (Glaves, 1983). Natural antibodies appear to be important in host surveillance of at least some metastasizing tumors. It was shown by Vaage (1978) that normal mouse serum contains components, presumably natural antibodies and complement, which can kill mouse C3H mammary carcinoma and ovarian carcinoma cells in uitro as well as in duo and which can inhibit metastasis formation. Similar findings were reported by Chow et al. (1981)from murine lymphoid tumor systems. Another potentially important natural host defense system that could influence the outcome of metastasis is the mononuclear phagocyte system. The ease of monocyte mobilization to inflammatory sites, the widespread distribution of macrophages through the body, the exquisite synthetic machinery of the macrophage, and its potent antitumor activity imply that cells of this lineage could have a potentially important role in the control of neoplastic primary and metastatic growth. The mere presence of macrophages at the site of a growing tumor, however, does not appear sufficient for such a controlling function. The cells have to become appropriately activated to express tumor cytostatic and/or tumoricidal functions (Alexander, 1976; Hibbs, 1974; Evans, 1982; Key, 1983). Such activation can be achieved via priming and trigger signals delivered by nonspecific means, for instance by biological response modifiers (BRMs) or by T-cell-derived lymphokines, such as macrophage activating factor (MAF) (Meltzer, 1981). The most

CANCER METASTASIS

51

promising new protocols for in situ activation of tumoricidal macrophages make use of liposomes containing either MAF or muramyl dipeptides (Fidler, 1980). Results from Fidler and Hanna (1981) indicate that the application of such liposomes, which were designed to home to the lungs and to activate in situ the lung macrophages, might become an important new tool in antimetastatic therapy. Control animals with a sc growing B16B16 tumor were all dead by 80 days, whereas 60% of the mice that had received a regimen of liposome-encapsulated muramyl dipeptide were still alive 170 days postimplantation. It has been argued that tumor cell resistance to macrophage cytolysis is a rare phenomenon and that such variants if they exist at all, may arise with much lower frequency than phenotypes for resistance to other therapeutic modalities (Fogler et al., 1980). Macrophage-resistant progressor tumor variants have, however, been described recently (Urban and Schreiber, 1983). It has to be kept in mind also that cytotoxic macrophages can become suppressed in uiuo by T cells, immune complexes, or serum-blocking factors (Rao et al., 1979; Hellstrom et al., 1979). Furthermore, macrophages are themselves able to suppress immune T-cell functions (Ting and Rodrigues, 1980). In many animal systems, as well as in advanced cancer patients, there is evidence of impaired immunological functions, which may be related to such a kind of suppression.

2 . Spec+

Immune Responses and Tumor Immune Escape

Spontaneous pulmonary metastases from an antigenic methylcholanthrene-induced murine sarcoma were reported by Sugarbaker and Cohen (1972) to have an altered antigenicity compared to that of the primary tumor. Similar findings of differences in immunogenicity and/or antigenicity between primary and secondary tumor lesions were later reported from various other systems (Pimm et al., 1980; Fogel et al., 1979; Gorelik et al., 1979b; Schirrmacher et al., 1979). These findings could be interpreted as the result of random antigenic variation occurring during tumor growth and host immune selection of distinct subpopulations occurring during metastasis. This would result in the outgrowth of tumor variants able to escape the specific immune response against the primary tumor. Antigenic variation in cancer metastasis (Schirrmacher et al., 1982c) as observed with chemically induced tumors may be possible because of the enormous polymorphism of tumor-associated transplantation antigens (TATA) which can become expressed on chemically transformed cells. Such TATAs have similarities to normal histocompatibility antigens not only with respect to polymorphism but also with respect to their ability to induce strong T-cell-mediated immune responses. Chemicals that are either muta-

52

VOLKER SCHIRRMACHER

genic (Boon, 1983), or that only affect DNA methylation (Kerbel and Frost, 1982) have been shown to change a tumor cell’s immunogenicity as defined by immunization-protection assays. The strong immunoselective pressure exerted by the host onto antigenic tumor cells during their process of metastasis could lead to an alteration in antigenicity, as discussed already, or it could lead to antigen loss and thus complete resistance to T-cell-mediated immunity. Such a process has been described and analyzed in detail in the ESb lymphoma model system (Bosslet and Schirrmacher, 1981, 1982). Twelve days after sc implantation of twice cloned TATA-positive ESb lymphoma cells that could be recognized and lysed by anti-TATA cytotoxic T lymphocytes (CTL), tumor variants could be detected from metastatic deposits in the spleen which were completely resistant to these syngeneic tumor specific CTLs. The variant cells were not generally resistant to CTL lysis because they could be killed by allogeneic anti-H2 CTL. Such selectively immunoresistant tumor variants arose with a high frequency under specific in viva conditions. They were stable for over 100 cell generations in tissue culture and were not detectable in the parent cell population by either cloning or in uitro immunoselection procedures (Bosslet and Schirrmacher, 1982). A recent cytogenetic comparison between TATA+ and TATA- ESb-type cells revealed distinct chromosomal differences (Dzarlieva et al., 1982). Antigen loss may occur only under certain circumstances when strong antigens elicit immune responses capable of eliminating antigenic cells at primary or secondary lesions. The same TATA-positive ESb cells as described above remained TATA positive when inoculated directly into the spleen (Bosslet and Schirrmacher, 1982) and also when grown in the peritoneal cavity. In fact, it is remarkable that the TATA has not been lost in our standard ESb cell line which has been transplanted ip for over 10 years since it first appeared in 1968. Antigen loss variants have also been described for the MDAY-D2 tumor (Dennis et al., 1981c), but the frequency of occurrence during spontaneous metastasis from a sc site was lower than in the ESb tumor. The continued process of host selection, resulting eventually in the emergence of selectively immunoresistant variants, may also be a relatively common feature of slowly growing spontaneous neoplasms. Since T-cell-mediated immune reactivities are restricted by gene products of the major histocompatibility complex (MHC), Feldman, Segal, and associates investigated whether tumor cell diversity in metastatic competence correlated with diversity in the expression of MHC gene products. Variations in MHC antigenic profiles of tumor cells have been frequently observed (Parmiani et al., 1979) and their biological effects discussed (Festenstein and Schmidt, 1981). In the Lewis lung carcinoma, a correlation was found between the metastatic potential of subclones and an imbalance of

CANCER METASTASIS

53

H-2KblH-2Dbexpression. The lower the ratio, the more metastatic were the clones (Eisenbach et al., 1983). It is not yet resolved, exactly how H-2 gene products on tumor cells can influence metastatic capacity. In addition to specific T cells there are specific antibodies that can influence metastasis formation. Shearman and Longenecker (1980) described a monoclonal antibody that could partially block liver colonization of chick Al-2 lymphoma cells. Expression of the corresponding antigen on the lymphoma cells correlated with enhanced liver colonization. Similar results were described in the RAW 117 lymphoma with antibodies against fetal determinants (Nicolson, 1982). Vollmers and Birchmeier (1983a)reported on a monoclonal antibody against B16 melanoma cells that could block plastic adhesion of the tumor cells in uitro and that inhibited lung colony formation in uiuo. Such antibodies also interfered with adhesion of other tumor cell types including human carcinomas (Vollmers and Birchmeier, 1983b). Recent findings point toward an important role of tumor cell binding to laminin (L. A. Liotta and J. Varani, personal communications; Vollmers et al., 1984) which may open new approaches to selective interference in the metastatic cascade process (see below). New developments in hybridoma technology could lead to production of monoclonal antibodies that react preferentially with metastases of certain tumor types. Such antibodies may become of increasing importance for radioimmunodetection by radioimaging procedures (Farrands et al., 1982) and perhaps also, when combined with drug targeting, for new therapeutic approaches. The present state of the art has recently been exemplified by studies on the in uivo localization of antihuman-osteogenic sarcoma monoclonal antibody in human tumor xenografts in nude mice (Baldwin and Pimm, 1983). 3. Intratumoral Immunological Heterogeneity and Aspects of Zmmunoregulation Heterogeneity exists not only in host immune responses but also in the immunological properties of the tumor cells themselves. Tumors of the same histological origin differ from each other immunologically. Even within one and the same tumor there is immunological heterogeneity. This is reflected in variations of tumor-associated antigens, differentiation antigens, histocompatibility antigens, lectin-binding sites, and receptors for natural killer cells and natural antibodies (Heppner and Miller, 1983). Four antigenic subpopulations (A, B, C, and D) were, for instance, identified within a spontaneous AKR lymphoma (Olsson and Ebbesen, 1979). Treatment of tumor-bearing mice with a mixture of all four subpopulations was much more efficient in inducing protection than immunizations with mixtures lacking one or more of the subpopulations. Miller and Heppner (1979) described

54

VOLKER SCHIRRMACHER

antigenic heterogeneity of five tumor cell subpopulations derived from a single BALB/cf 3H mouse mammary tumor. Experiments in which mice were immunized with each subline separately or with mixtures thereof indicated that the specificities of the responses to the mixtures were not simply the sum of the responses to the individual sublines. Thus, the response to an immunogenic subpopulation may cause cross-protection of other subpopulations (Miller, 1982), whereas the response to a suppressogenic subpopulation could cause suppression of immune responses against other subpopulations. Basic immunology has provided evidence that immune responses underlie controlling elements that can exert either positive or negative regulatory functions such as “helper” functions or “suppressor” functions. T-cell subpopulations with either helper or suppressor function could be distinguished among others by differences in expression of Lyt differentiation antigens. The existence of immunological regulatory circuits with suppressor and contrasuppressor systems has been proposed and described by Gershon and colleagues (Gershon et al., 1981; Green et al., 1983). Lymphocyte subpopulation interactions can be mediated by antigen-specific and -nonspecific soluble factors and corresponding cell surface receptors. They could also involve specific receptor-antireceptor interactions as suggested in Jerne’s immunological network theory (Jerne, 1976). Such idiotype network interactions in antitumor immune reactions have been reviewed recently (Schreiber, 1984). From these considerations it becomes apparent that the dynamics of tumor progression with continuously evolving subpopulations may initiate complex immunological reaction cascades that develop into a specific immunological microenvironment. Heppner and Miller (1983) used the term “specific ecosystem” to describe the complex tumor-host network interactions. It is not surprising from such a point of view that it is virtually impossible to generalize on the role of host immunity in malignant processes and in metastasis in particular. In some tumor systems suppression of antitumor immune responses resulted in increased metastasis, whereas in other tumor systems similar manipulations decreased metastasis formation. Using C3H fibrosarcoma cell lines of differing immunogenicity, Fidler et al. (1979) showed that weakly immunogenic fibrosarcoma cells grew and metastasized more readily in immunocompetent as compared to immunoincompetent hosts, whereas the reverse was true for intermediate immunogenic tumor cells, which is consistent with Prehn’s idea of immune stimulation (Prehn, 1972). One way in which metastatic cells may successfully evade host surveillance is by suppression of host immune responses. Fujimoto et al. (1978) found that suppressor activity could be transferred from tumor-bearing ani-

C A N C E R METASTASIS

55

mals via spleen or thymus cells and could be abolished by anti-T-cell antibodies plus complement. Transfer of T suppressor cells resulted in the inhibition of the effector phase of T-cell-mediated immune cytolysis but only if the original sensitizing tumors were identical or very similar to the tumors being tested, indicating T suppressor specificity. Another interesting tumor system with regard to immunosuppression is that of UV light-induced skin tumors studied by Kripke and Fisher (1976). Chronic exposure of mice to UV light can produce fibrosarcomas and squamous cell carcinomas. These are highly immunogenic and are thus rejected after transplantation into syngeneic hosts. Pretreatment of such hosts, however, with intermittent doses of UV light could render them tolerant to the tumor transplants as if they were suppressed. The effect could be shown to be due to activation of T suppressor cells which prevented immune destruction of the highly antigenic tumors (Fisher and Kripke, 1978). In contrast to antitumor cytotoxic T lymphocytes (CTL) which could specifically distinguish individual antigenic determinants on different UV-induced tumors (Wortzel et al., 1982), the T suppressor cells seemed to recognize cross-reactive determinants which are apparently shared by all the UV-induced tumors but which are not expressed on chemically induced tumors (Urban et al., 1982). Naor (1983) has recently summarized findings from different tumor systems which suggest the coexistence of suppressogenic and immunogenic determinants within a tumor. Such determinants could be expressed on the same cell or on different tumor cell subpopulations. More refined technologies are required on the cellular and molecular level to unravel the nature of such determinants with opposite immunological effects. If the findings can be substantiated, however, they may become very important. It would then become quite rewarding to either eliminate the suppressogenic determinants or to alter them in such a way that they lose their suppressogenic capacity and perhaps acquire immunogenic capacity instead. In a recent report (Cianciolo et al., 1983), suppressogenic molecules have been identified on murine malignant but not on normal cells as a 19-kDa protein being antigenically related to the immunosuppressive retroviral protein, p15E.

IX. Impacts of Experimental Studies on Cancer Treatment Strategies

A. NEWSTRATEGIES FOR IMMUNOLOGICAL INTERVENTION The search for an immunological approach to the treatment of cancer derives as a logical extension of the effectiveness of the immune response in dealing with infectious diseases. Theoretically, the immune system should

56

VOLKER SCHIRRMACHER

be particularly effective against disseminated diseases and micrometastases as opposed to primary tumors because of the different numbers of tumor cells to be eradicated and the more favorable ratio of effector to target cells when dealing with small tumor foci. As discussed more extensively elsewhere (Frost and Kerbel, 1983), cycles of euphoria with regard to tumor immunotherapy have, to date, generally been followed by unsuccessful clinical trials. This was felt not to be surprising, however, when considering the diverse ways in which a host can respond to its tumor cells, the problem of tumor heterogeneity and progression and of subpopulation interactions, the weak immunogenicity of many “spontaneous” tumors, and the number of different escape routes available to highly malignant cells (Kim, 1979). Another reason for the discrepancy between the experimental and clinical data could be that most of the experimental animal tumor systems in which immunotherapy protocols proved to be effective may not have mimicked closely enough the human cancer and the course of its disease development (Eccles, 1982). Nevertheless, the possiblities of immunological intervention in metastasis should be exploited systematically. This should be done in appropriate experimental model systems first and should not be too hastily applied to the clinical situation as long as there is only a poorly defined immunological basis. While immunologic thinking has shifted in recent years toward the development of mainly unspecific immune stimulation modalities, at least some investigators feel that specific immunological intervention is still feasible and should not be forgotten. This hope is based on recent advances in our understanding (1) of the relationship of the immune response to metastasis, (2) of new ways to increase the immunogenicity of spontaneous tumors, (3) of insights into immunoregulatory phenomena as well as (4) of experimental antimetastatic-specific immunotherapy protocols, which already have proved to be effective. Some experimental approaches for immunological intervention in metastasis are listed in Table XI. They include efforts (1) to increase the tumor’s immunogenicity, (2) to interfere with host immune regulatory mechanisms, (3) to restore or increase immune competence, for instance, by adoptive cell transfer, and (4) to exploit monoclonal antibodies for interference at specific steps of the metastatic process. Both unspecific and specific components of the host defense system should be tested separately or in combination in order to design new strategies for immunotherapy of metastases. An increase in tumor immunogenicity can be obtained by various means, such as chemical modification, virus infection, treatment with mutagens, or somatic cell hybridization. Immunization against such modified tumor cells often resulted in the immune rejection of the nonmodified tumor cells. Protective immunity against an apparently noniminunogenic tumor line has

57

CANCER METASTASIS

TABLE XI IMMUNOLOGICAL INTERVENTION IN METASTASIS: EXPERIMENTAL APPROACHES Approach0

References

Increase in tumor immunogenicity Chemically modified cells Mutagenized tumor variants , Virus-infected tumor cells Somatic cell hybrids tumor cell

Hamaoka et al. (1979), Prager and Baechtel (1973), Boon (1983). Kobayashi (1979), Austin and Boone (1979). Kawashima et al. (1983)

-4 Interference with immunoregulation Elimination of suppressor cells Immune stimulation unspecific unselective Immune stimulator specific selective

Dye and North (1981), Naor (1979, 1983), Greene et 01. (1977), Hanna et a!. (1979), Hanna and Key (1982), Fidler (1980)

Adoptive cell transfer

2-

h

NK cells Syngeneic immune T cells Allogeneic H-2 identical immune T cells 11-2 expanded CTLs

Inhibition by monoclonal antibodies

Hanna and Fidler (1980), Warner and Dennert (1982), Treves et al. (1975), Schirrmacher et al. (1982), Frost and Kerbel (1983). Wiltrout et a!. (1979), Schirrmacher (1979), Dailey et al. (1982). Eberlein et'al. (1982), Kedar and Weiss (1983) Vollmers and Birchmeier (1983)

A a TA, Tumor antigen; NAD, new antigenic determinant; SD, suppressogenic determinant; TH. T helper cells; Ts, T suppressor cells.

58

VOLKER SCHIRRMACHER

been observed in some tumor systems (Boon, 1983; Kobayashi, 1979; Kawashima et al., 1983). If such one-way cross-reactivities and possibly subpopulation interactions could be created andlor selectively enhanced, active immunotherapy might be beneficial even against nonimmunogenic cells. This might involve the isolation or activation of a “controlling” subpopulation by treatment with mutagens, manipulation of major histocompatibility antigen expression, infection with a virus, or chemical modification (Table XI). Efforts to interfere with immunoregulation might be directed toward the elimination of suppressor cells and/or toward stimulation of immune effector mechanisms. This may involve the use of irradiation regiments, drugs, and biological response modifiers. Such interference with immunoregulation should be as selective as possible and should have minimal side effects. One example of a selective interference is the use of macrophage activating liposomes that are designed to home to particular organs. A direct approach for specific immunotherapy of metastases was assessed by Treves et al. (1975) by adoptive transfer into tumor-bearing animals of T cells sensitized to irradiated tumor cells in uitro. The treatment was only partially successful, probably because of the interference of T suppressor cells. That T supressor cells generated in tumor-bearing mice could impair the effectiveness of adoptive immunotherapy has indeed been demonstrated (Dye and North, 1981). Later, a potent ability of T cells to inhibit metastases was demonstrated independently by two groups, both of which used allogeneic but H-2-identical T cells in adoptive immunotherapy protocols. Wiltrout et al. (1979) demonstrated that the adoptive transfer of spleen cells from immunoresistant BALB/c nu/+ tumor immune animals into BALB/c nulnu mice bearing the metastatic DBAI2 tumor MDAY-D2 could cause arrest and reversal of established visceral metastases. In a similar but syngeneic system, the author (Schirrmacher, 1979)could show that DBA/2 mice bearing the highly metastatic DBAl2 tumor ESb could be partially protected from death by metastases by the iv transfer of spleen cells from B10.D2 tumor immune animals. A single cell transfer led to a 100% increase of life expectancy. Recently Frost and Kerbel (1983) modified their immunotherapy protocol for application in a syngeneic system: First they used antitumor CTLs which were raised against an immunogenic variant of MDAY-D2 obtained by mutagen treatment. This variant did not induce suppressor cells and generated a much higher antitumor CTL response. Before adoptive cell transfer into syngeneic MDAY-D2 tumor-bearing mice, the primary tumor was surgically removed and eventual host suppressor cells (Bursuker and North, 1984) eliminated by irradiation. With this protocol, a large percentage of animals bearing metastases (about 75%) were cured, and the remaining 25% showed prolonged survival as compared to controls. Similar effective immunotherapy results were recently obtained by our group in the ESb

C A N C E R METASTASIS

59

tumor system, using a virus xenogenization approach (unpublished findings). Many parameters for an optimal and highly reproducible response in such an experimental immunotherapy protocol still have to be worked out. Furthermore, the mechanisms underlying such an effective protective immunity need to be unraveled in order to know in which direction to go if such protocols are to be tested for clinical applicability. In both highly metastatic tumor models, MDAY-D2 and ESb, effective protection against metastases could thus be obtained in spite of the fact that the tumors used are poorly immunogenic, are heterogeneous, and have a strong tendency to develop tumor antigen loss variants. It is possible that in both systems the specific cellular antitumor immune responses were flanked by nonspecific cytotoxic mechanisms mediated by macrophages and possibly NK cells. There are additional possibilities for further improvements of the therapy protocols: (1)the use of 11-2 and of 11-2 expanded long-term T-cell lines (Donohue et aE., 1984), (2) a combination of adoptive specific immunotherapy with macrophage activation via liposomes containing activating agents, and (3) combinations of immunotherapy with other treatment modalities. Synergistic effects of active specific immunotherapy and chemotherapy have been reported (Key et al., 1983). It appears that immunotherapy, if designed properly, still has a great potential. Antimetastatic immune T-cell therapy, although not yet practical, would be a means of achieving systemic, nontoxic, and nonmutagenic therapy. The potential role of T cells in cancer therapy has recently been reviewed (Fefer and Goldstein, 1982; Cheever et al., 1983).

B. OTHERIMPACTSFROM EXPERIMENTAL STUDIES The sequential and stepwise evolution of metastasis has a number of implications for treatment strategies. In addition to direct anticancer drugs there could be strategies developed to intervene at any one critical point during the multistep process of the metastatic cascade. Some examples, as derived from experimental studies, are given in Table XII. The treatment strategies listed aim at specific therapy targets which evolve from the analysis of mechanisms at each successive step of disease progression (e.g., angiogenesis, invasion, clumping, capillary adhesion, extravasation). Each step in the cascade is supposed to involve different tumor cell subpopulations and different mechanisms of tumor-host interactions. Such steps are illustrated in the left part of the table. References are given to the sections in this article where respective details have been discussed. The table is not meant to be complete but rather to illustrate possibilities of rational approaches to therapeutic intervention in cancer metastasis. Combinations of drugs may be found to prevent angiogenesis as ex-

TABLE XI1 Metastatic cascades

Event

1. Uncontrolled proliferation 2. Angiogenesis and local invasion

3. Survival in the circulation 4. Arrest and extravasion

5. Proliferation at new site into secondary foci

6. Evasion of host defense and resistance to therapy 7. Tumor dormancy

8. Angiogenesis and local invasion 9. 3” metastases from 2” metasta-

Tumorigenic noninvasive subpopulation

Invasive nonmetastatic subpopulation

Treatment strategy Metastatic subpopulation

Therapy target Activation of fibrous “barrier” formation Prevention of angiogenesis Antiinvasive treatment Prevention of clumping Stimulation of NK activity Inhibition of adhesion

ses a

+

Stimulation of M activity Antiproliferative antitumor agents Systemic activation and restoration of immune capacity New subpopulation diversification with emergence of resistant variants Prevention of “activation” of dormancy Prevention of angiogenesis Antiinvasive treatment Steps 3-7

Modified version from that of Poste (1982). BRM, Biological response modifiers, MAF, macrophage activating factor.

Methodb

?

Reference section V1

Use of combined drugs and AIF

V VI

Anticoagulant drugs BRM

VIII

Monoclonal antibody, laminin fragment Liposome-MAF

I1 VIII

Chemotherapy Radiotherapy Immunotherapy

V

Change treatment regiments in rapid succession

IV

?

VII

Systemic application of selected drugs (see Step 2 )

V

VI 111

CANCER METASTASIS

61

emplified by Folkman’s recent quite successful experimental studies (Folkmann et al., 1983). This could have an inhibitory effect not only in step 2 but also in step 8, which initiates the development of metastases from metastases. Cancer invasion may be prevented by antiinvasive drugs (Mareel, 1984) or antiinvasive tissue factors (AIF; Eisenstein et al., 1975). Although cancer invasion has to be considered an important target for therapy because this step leads to progression from a benign to a malignant state, the way that screening of anticancer drugs has been conducted so far has neglected this aspect and still does. In colon cancer, depth of invasion (Duke’s classification) is an important prognostic factor. The 5-year survival rate drops from about 70%for tumors that are limited to the mucosa to about 10% for tumors that have invaded all layers of the colonic wall. Potential inhibitors of invasion are ICRF-159 [ 1,2-bis(3,5 dioxopiperazin-l-yl)propan](which inhibits intravasation in some experimental tumors), protease inhibitors such as aprotinin, or microtubule inhibitors such as vindesine or podophyllotoxin (Mareel, 1982). The ideal antiinvasive agent should prevent invasion and transform a malignant into a quasi benign tumor without the dose-limiting toxicity of current antiproliferative or cytotoxic drugs. Such an ideal antiinvasive agent with high selectivity has not, however, yet been found. Agents acting on the coagulative-anticoagulative equilibrium and at the fibrin systek (Maat and Hilgard, 1981) could influence particular steps such as activation of a fibrous barrier, prevention of invasion or of tumor cell clumping, and capillary arrest in the circulation. Most experimental evidence for the antimetastatic effect of anticoagulants came from experimental metastasis assays after intravenous tumor cell inoculation. Coumarin and derivatives thereof produced a constant antimetastatic effect that was independent of the investigator and of the tumor model. Although antimetastatic effects of anticoagulants have failed so far in spontaneous metastasis assays, interest in coumarins as potential antiinvasive agents is supported by recent findings that warfarin treatment significantly retarded disease progression in small cell carcinoma of the lung (Zacharski et al., 1981). Specific reagents that can inhibit the organ implantation phase should also be looked for and developed. Blocking of distinct cellular receptors such as the receptor for laminin and use of selected monoclonal antibodies are two examples discussed in Section VIII. Immunological means of interference with metastases have been summarized and discussed in Section IX, A. Because of the problem of phenotypic instability of tumor cell subpopulations that survive a particular therapeutic treatment, the object of new strategies should be to prevent new diversification and outgrowth of resistant variants. Therefore it seems advisable to change treatment regimens in rapid succession and to use combinations of treatments that aim at different

62

VOLKEH SCHIRHMACHER

therapeutic targets and that make use of as many as possible modalities of treatment. The involvement of the body’s own defense systems, both natural and specific, should always be a part of the overall strategy because this can interfere in many different ways and in many different steps. Genetic instability and tumor subpopulation interactions (discussed in Section IV) could have profound impacts on treatment strategies. Most cancer treatments, if not complete, might influence phenotypic diversity among the remaining tumor cell subpopulations and might thereby affect tumor progression. Individual tumor cell subpopulations may differ significantly in their sensitivities to various therapeutic agents and this in itself represents a major obstacle to the development of new therapeutic strategies of cancer treatment. Advances in our knowledge on the complexities of tumor cell heterogeneity and diversity (Section IV) have rendered many of the traditional therapeutic strategies obsolete. Screening programs for potential anticancer agents have to take into account more than before the problems of phenotypic heterogeneity, the possible impact of the drugs themselves on the generation of instability and diversification, and the side effects on normal host tissues and on the immune system which again could influence metastatic progression. Obviously, the most successful therapies will be those that circumvent phenotypic diversity among tumor cells and those that alleviate the problem of survival of resistant subpopulations during therapy. The timing of therapy is also an important parameter because the cascade theory (Section III) suggests that the metastatic process develops sequentially by stepwise progression in time and space. There may be certain periods where antiproliferative anticancer agents can be expected to work most efficiently because a high proportion of cancer cells would be in a proliferative state. The removal of a large primary tumor mass, for instance, can be expected to lead to the removal of antiproliferative and other regulatory factors (Section III,B) which could stimulate the outgrowth of micrometastases. The application of antiproliferative anticancer drugs in this phase (Table XII,5) would thus be indicated. Some experimental studies indicate that certain common procedures in the clinic might have a stimulatory effect on metastases. Examples are reported where the use of certain anesthetic drugs (Shapiro et al., 1981) or of certain surgical procedures (Gorelik et aZ., 1979a; Keller, 1984)caused acceleration of the progression of postoperative metastases. Immunosuppressive effects have been documented for anesthetics and surgery (Walton, 1978; Lee, 1977) as well as for chemotherapeutic agents. More effort should be devoted to the development of methods to prevent or counteract the possible negative side effects that might be inherent in many of the present clinical treatment procedures.

CANCER METASTASIS

63

X. Summary and Conclusions

It has been the purpose of this article to describe recent advances in cancer metastasis research. Clinical realities and experimental approaches to the study of underlying basic mechanisms of metastasis formation were discussed. Wherever possible, results were reported which led to the development of theoretical concepts. Such results and concepts were finally evaluated in light of their possible impact for the design of new treatment strategies. Experimental findings from many diverse research fields were summarized with the help of tables, figures, and references. It was concluded that the process of metastasis is a dynamic event that can be described as a sequence of interrelated steps. Experimental results indicated that malignant cells that migrate and disseminate from the primary organ to distant sites and there eventually develop into metastases have to survive a series of potentially lethal interactions. Intimate tumor-host interactions were reported to take place all along the metastatic process. They were elucidated at the steps of angiogenesis, invasion, organ interaction, dormancy, tumor rejection, and tumor immune escape. The outcome of such tumor-host interactions seemed to depend on intrinsic properties of the tumor cells themselves as well as on the responsiveness of the host. Metastasis does not appear as a merely random process. Both clinical and experimental studies revealed that the whole process can be described more appropriately in terms of stochastic, sequential, and selective events, each of which is controlled and influenced by a number of mechanisms. With regard to therapeutic intervention, a selective event offers more possibilities than a random one because it is governed by rules that can be exploited experimentally. Various impacts from experimental studies for the design of antimetastatic cancer treatment strategies were discussed. Sequential steps of the metastatic cascade could become new therapy targets. Conventional empirically derived treatment modalities should become flanked by methods aimed more specifically at critical steps of cancer spread in order to prevent progression of the disease. This is where basic research on mechanisms could make significant contributions to therapy planning in the future. Furthermore, possible negative effects of surgery, radiotherapy, and adjuvant chemotherapy or immunotherapy that could result in enhancement of metastatic progression need to be critically evaluated to limit them as much as possible. The most formidable obstacle to the successful treatment of disseminated cancer may be the fact that the cells of a tumor are biologically heterogeneous and can generate phenotypic diversity even if they arose from a single clone. Genetic instability of malignant cells, subpopulation interac-

64

VOLKER SCHIRRMACHER

tions, and host selection seem to be important keys to understanding the evolution of tumor heterogeneity and progression. Experimental metastasis research could have a significant impact on the design of clinical treatment regimens if it were able to unravel the mechanisms controlling the generation of phenotypic diversity and progression. The potential danger that chemotherapeutic drugs themselves could facilitate tumor progression via generation of “mutants” or DNA methylation variants (Boehm and Drahovsky, 1983)could then possibly be limited. Screening of potential anticancer drugs for their abilities to inhibit the growth of tumors containing widely heterogeneous subpopulations of metastatic and nonmetastatic cells may not be sufficient to predict the efficacy against metastases. New methodologies and more specific antiinetastatic screening procedures are therefore urgently needed.

REFERENCES Adair, F., Berg, J . Joubert, L., and Robbins, G. F. (1974). Cancer 33, 1145. Alexander, P. (1976). In “Fundamental Aspects of Metastasis” (L. Weiss, ed.), pp. 227-239. North-Holland Publ., Amsterdam. Altevogt, P., Kuriiick, J. T., Kimura, A. K., Bosslet, K . , and Schirrmacher, V. (1982). Eur. J. Immunol. 12, 300-307. Altevogt, P., Fogel, M., Cheingsong-Popov, R., Dennis, J., Robinson, P., and Schirrmacher, V. (1983). Cancer Res. 43, 5138-5144. Amstrong, P. B. (1980).In “Cell Movement and Neoplasia” (M. De Brabander, M. Mareel, and L. De Ridder, eds.), pp. 131-151. Elsevier, Amsterdam. Austin, F. C., and Bootie, C. W. (1979). Ado. Cancer Res. 30, 301-345. Balaban, G., Herlyii, M., Guerry, D.,and Nowell, P. C. (1982). Proc. Am. Assoc. Cancer Res. 23, 130. Baldwin, R. W., and Pimm, M . V. (1983). Cancer Metastasis Reo., 2, 89-106. Bauer, E. A., Cordon, J. M., Reddick, M. E., and Eisen, A. Z. (1977).J . Invest. D e m t o l . 69, 363-367. Bennet, D. C. (1983). Cell 34, 445-453. Boehm, T. L. J . . and Ilrahovsky, D. (1983).J. Natl. Cancer Inst. 71 (Guest Editorial), 429-433. Bogenmann, E., Mark, C., Isaacs, H., Neustein, H. B., DeClerk, Y. A., Laug, W. E., and Jones, P. A. (1983). Cancer Res. 43, 1176-i186. Boon, T. (1983).Ado. Cancer Res. 39, 121-151. Boon, T., and Kellerniann, 0. (1977). Proc. Natl. Acad. Sci. U . S . A . ,74, 272-275. Bosslet, K . , and Schirrmacher, V. (1982). Int. J. Cancer 29, 195-202. Bosslet, K., and Schirrniacher, V. (1981).J. Exp. Med. 154, 557-562. Briles, E. B . , and Kornfeld, S. (1978).J , Natl. Cancer Inst. 60, 1217-1222. Bross, I . 1). J. (1980). In “Metastatic Tumor Growth” (E. Grundman, ed.) pp. 207-221. Fischer, Stuttgdrt. Bross, I. 1). J., and Blumenson, L. E (1976). In “Fundamental Aspects of Metastasis” (L. Weiss, ed.), pp. 359-375. North Holland Puhl., Amsterdam. Brunson, K. W., Beattie, G., and Nicolson, 6. L. (1978). Nature (London) 272, 543-545. Brunson, K. W., and Nicolson, C . L. (1979).J. Suprainol. Struct. 11, 517-528.

CANCER METASTASIS

65

Burger, M. M. (1980). In “Biology of the Cancer Cell” pp. 193-208. Kugler Publications, Amsterdam. Bursuker, I., and North, R. J. (1984).J . E x p . Med. 159, 1312-1321. Cederholm-Williams, S. A. (1981). Znoasion Metastasis 1, 85-98. Chambers, A. F., Hill, R. P., and Ling, V. (1981). CancerRes. 41, 1368-1372. Chatterjee, S. K., and Kim, U. (1977).J . Natl. Cancer Znst. 58, 278-280. Chatterjee, S. K., and Kim, U. (1978).J . Natl. Cancer Znst. 61, 151. Cheever, M. A., Greenberg, P. D., and Fefer, A. (1983).J . B i d . Response Mod$. 3, 113-127. Cheingsong-Popov, R., Robinson, P., Altevogt, P., and Schirrmacher, (1983). Int. J . Cancer 32, 356-366. Chow, D. A., and Greenberg, A. H. (1980). Znt. J . Cancer 25, 261-265. Chow, D. A., Wolosin, L. B., and Greenberg, A. H. (1981). Int. J . Cancer 27, 459-469. Chow, D. A., Ray, M., and Greenberg, A. H. (1983). Int. J . Cancer 31, 99-105. Cianciolo, G. J.. Lostrom, M. E., Milton, T., and Snyderman, R. (1983).1. E x p . Med. 158, 885-900. Cifone, M. A., and Fidler, I. J. (1981). Proc. Natl. Acad. Sci. U.S.A. 78, 6949-6952. Dailey, M. O., Pillemer. E., and Weissman, 1. L. (1982). Proc. Natl. Acad. Sci. U.S.A. 79, 534-5387, De Baetselier, P., Katzav, S., Gorelik, E., Feldman, M . , and Segal, S. (1980).Nature(London) 288, 179-181. De Baetselier, P., Gorelik, E., Eshar, Z., Ron, Y., Katzav, S., Feldman, M., and Segal, S. (1981).J . Natl. Cancer Znst. 67, 1079-1087. De Baetselier, P., Roos, E., Brys, L., Remels, L., Gobert, M . , DeKegel, D., Segal, S., and Feldman, M. (1984). Cancer Metastasis Reo. 3, 5-24. Dennis, J. W., and Kerbel; R. S. (1981). Cancer Res. 41, 98-104. Dennis, J. W., Donaghue, T. P., and Kerbel, R. S. (1981a).J . Natl. Cancer Znst. 66, 129-139. Dennis, J. W., Donaghue, T., Florian, M., and Kerbel, R. S. (1981b). Nature (London) 292, 242-245. Dennis, J. W., Donaghue, T. P., and Kerbel, R. S. (1981~).Znoasion Metastasis 2, 111-125. Dietz, M., Longley, C., Fouchey, S. P., Hall, L., Rich, M. A,, and Farmanski, P. (1977).J . Natl. Cancer Inst. 59, 957. Donohue, J. H . , Lotze, M. T., Robb, R. J. Rosenstein, M., Braziel, R. M., J d e , E. S., and Rosenberg, S. A. (1984). Cancer Res. 44, 1380-1386. Dvorak, H. F., Orenstein, N. S., Carvalho, A. C., Churchill, W. H., Dvorak, A. M., Galli, S. J., Feder, J., Bitzer, A. M . , Rypsysc, J., and Giovinco, P. (1979a).J . Zmmunol. 122, 166. Dvorak, H. F., Dvorak, A. M., Manseau, E. J., Wiberg, L., and Churchill, W. H. (197913).J . Natl. Cancer Znst. 62, 1459-1472. Dvorak, H . F., Senger, D. R., and Dvorak, A. M. (1982). Cancer Metastasis Reo. 2, 41-73. Dye, E. S., and North, R . J. (1981).J , E x p . Med. 154, 1033-1042. Dzarlieva, R . , Schirrmacher, V., and Fusenig, N. F. (1982). Znt. J . Cancer 30, 633-642. Easty, D. M., and Easty, G . C. (1974). Br. J . Cancer 29, 36-49. Eberlein, T. J., Rosenstein, M., and Rosenberg, S. A. (1982). J . Exp. Med. 156, 385-397. Eccles, S. A. (1982). In “Tumor Immunity in Prognosis” (S. Haskill, ed.), pp. 37-74. Dekker, New York. Eccles, S. A,, and Alexander, P. (1975). Nature (London) 257, 52. Eisenbach, L., Segal, S., and Feldman (1983). Znt. J. Cancer 32, 113-120. Eisenstein, G., Kuettner, K. E., Neapolitan, C., Soble, L. W., and Sorgente, N. (1975).Am. J . Pathol. 81, 337-347. Eisenstein, R . , Sorgente, N . , Soble, L. W., Miller, A., and Kuettner, K. E. (1973). Am. J . Pathol. 73, 765-774.

66

VOLKER SCHIRRMACHER

Evans, R. (1982). Cancer Metastasis Neo. 1, 227-239. Everson, T. C., and Cole, W. H. (1966). “Spontaneous Regression of Cancer.” Saunders, Philadelphia, Pennsylvania. Ewing, J. (1928). “Neoplastic Diseases,” 3rd ed. Saunders, Philadelphia, Pennsylvania. Farrands, P. A., Pimm, M. V., Embleton, M. J., Perkins, A. C., Hardy, J. D., Baldwin, R. W., and Hardcastle, J. D. (1982). Lancet August, 397-400. Fefer, A,, and Goldstein, A. L., eds. (1982). Prog. Cancer Res. Thera. 22, 1-297. Feinberg, A,, and Vogelstein, B. (1983). Nature (London) 301, 89-92. Festenstein, H., and Schmidt, W. (1981). Immunol. Reo. 60, 85. Fialkow (1976). Annu. Reo. Med. 30, 135-176. Fidler, I. J. (1973). Nature (London) 242, 148. Fidler, I. J. (1976). In “Fundamental Aspects of Metastasis” (L. Weiss, ed.), pp. 275-289. North Holland Puhl., Amsterdam. Fidler, I. J. (1978). Methods Cancer Res. 15, 399-439. Fidler, I. J. (1980). Science 203, 1469-1471. Fidler, I. J , , and Hanna, M. G. (1981). In “Fundamental Mechanisms in Human Cancer Immunology” (J. P. Saunders, J. C. Daniels, B. Serrou et al., eds.), pp. 425-437. Elsevier, Amsterdam. Fidler, I. J., and Kripke, M. L. (1977). Science 197, 893. Fidler, I. J., and Poste, G. (1982). Hosp. Pract. July, 57-64. Fidler, I. J. Gersten, D. M., and Budmen, M. B. (1976). Cancer Res. 36, 3160-3165. Fidler, I. J . , and Nicolson, G . L. (1976). J. Natl. Cancer Inst. 58, 1867-1872. Fidler, I. J . , Gersten, D. M., and Hart, I. R. (1978). Ado. Cancer Res. 28, 149-250. Fidler, I. J.. Gersten, D. M., and Kripke, M. L. (1979). Cancer Res. 39, 3816-3821. Finne, J., Tao, T. W., and Burger, M. M. (1980). Cancer Res. 40, 2580-2587. Fisher, M. S . , and Kripke, M. L. (1978). J. Immunol. 121, 1139-1144. Fogel, M., Gorelik, E., Segal, S., and Feldman, M. (1979). J. Natl. Cancer Inst. 62, 585. Fogel, M., Altevogt, P., and Schirrmacher, V. (1983). J. E x p . Med. 157, 371-376. Fogler, W. E., Raz, A , , and Fidler, I. J. (1980). Cell. Immunol. 53, 214-219. Folkman, J. (1974a). Ado. Cancer Res. 19, 331-358. Folkman, J. (1974b). Cancer Res. 34, 2109. Folkman, J. (1976). Sci. Am. 234, 58. Folkman, J., and Haudenschild, C. (1980). Nature (London) 288, 551-556. Folkman, J., and Hochberg, M. (1973). J . Exp. Med. 138, 745. Folkman, J., Langer, R., Linhardt, R. J., Haudenschild, C., and Taylor, S. (1983). Science 221, 719-725. Foulds, L. (1954). Cancer Res. 14, 327-339. Foulds, L. (1975). “Neoplastic Development.” Academic Press, New York. Frost, P., and Kerbel, R. S. (1981). Int. 1. Cancer 27, 381-385. Frost, P., and Kerbel, R. S. (1983). Cancer Met. Reo. 2, 239-256. Frost, P., Kerbel, R. S., and Tartamella-Biondo, R. (1981). Invasion Metastasis 1, 22-33. Frost, P., Kerbel, R. S., Bauer, F., Tartamella-Biondo, R . , and Celfalu, W. (1983). Cancer Res. 43, 125-131. Frost, P., Liteplo, R. G., Donaghne, T. P., and Kerbel, R. S. (1984).J . Exp. Med. 159, 14971507. Fujimoto, S., Matsuzawa, T., Nakagawa, K . , and Tada, T. (1978). Cell. Immunol. 38,378-387. Garfield, D. H., and Kennedy, B. J. (1972). Cancer 30, 190. Gershon, R. K., Eardley, D. D., Durum, S . , Green, D. R., Shen, F.-W., Yamauchi, K., Cantor, H., and Murphy, D. B. (1981). J. E x p . Med. 153, 1533-1546. Giavazzi, R . , Allesandri, G . , Spreafico, F., Garattini, S., and Mantovani (1980). Br. J. Cancer 42, 462-472.

CANCER METASTASIS

67

Gilbert, H. A., Kagan, A. R., Nussbaum, H., Hintz, B., and Chan, P. Y. M. (1980). In “Metastatic Tumor Growth” (E. Grundrnann, ed.), pp. 223-243. Fischer, Stuttgart. Gilbert, K., Chu, F . , Jones, E., and Diluzio, N. R. (1977).J . Reticuloendothel. Soc. 22, 319327. Gimbrone, M . A , , Leapman, S. B., Cotrane, R. S., and Folkman, J. (1972).J . Exp. Med. 136, 261. Girnbrone, M. A., Cotrane, R. S . , and Folkrnan, J. (1974).J . Natl. Cancer Znst. 52, 413. Glaves, D. (1983). Znoasion Metastasis 3, 160-173. Goldenberg, D. M., Pavia, R. A,, and Tsao, M. C. (1974). Nature (London) 250, 649-651. Gorelik, E. (1983). Ado. Cancer Res. 39, 71-120. Gorelik, E., Segal, S., and Feldman, M. (1979a). Znt. J . Cancer 21, 617-625. Gorelik, E., Fogel, M., Feldrnan, M., and Segal, S. (1979b). J . Natl. Cancer Znst. 63, 13971404. Gorelik, E., Bere, W. W., and Herberman, R. B. (1984). Znt. J . Cancer 33, 87-94. Granzow, C., Kopun, M., and Zimmermann, H.-P. (1980). In “Metastatic Tumor Growth” (E. Grundmann, ed.), pp. 43-51. Fischer, Stuttgart. Green, D. R., Flood, P. M., and Gershon, R. K. (1983). Annu. Reu. Zmmunol. 1, 439-464. Greenblatt, M., and Shubik, P. (1968). J . Natl. Cancer Znst. 41, 111. Greene, M. I., Dorf, M. E., Pierres, M., and Benacerraf, B. (1977). Proc. Natl. Acad. Sci. U.S.A. 74, 5118. Haemmerli, G., and Strauli, P. (1978). Virchows Arch. B Cell Pathol. 29, 167-177. Hagmar, B. (1972). Acta Pathol. Microbiol. Scand. Sect. A Pathol. 80, 357-366. Hagmar, B., Ryd, W., and Erkell, L. J. (1983). Znoasion Metastasis 3, 1-21. Hamaoka, T., Fujiwara, H., Teshima, K., Aoki, H., Yamamoto, H., and Kitagawa, M. (1979).J . E x p . Med. 1490, 185-199. Hanna, M. G., and Key, M. E. (1982). Science 217, 367-369. Hanna, M. G., Brandhorst, J . S., and Peters, L. C. (1979). Cancer Zmmunol. Zmmunother. 7, 165-173. Hanna, N. (1982). Cancer Metastasis Reo. 1, 45-64. Hanna, N., and Fidler, I. J. (1980). J . Natl. Cancer Znst. 65, 800-812. Hanna, N., and Fidler, I. J. (1981). J. Natl. Cancer Znst. 66, 1183-1190. Hart, I. R. (1979). Am. J . Pathol. 97, 587-600. Hart, I. R. (1982). Cancer Metastasis Reo. 1, 5-16. Hart, I. R., and Fidler, 1. J. (1978). Cancer Res. 38, 3218-3224. Hart, I. R., and Fidler, I. J. (1980). Cancer Res. 40, 2281-2287. Hatzubai, A,, Maloney, D. G . , and Levy, R. (1981).J. Zmmunol. 126, 2397. Hellstrom, I., Hellstrom, K. E., and Bernstein, I. D. (1979). Proc. Natl. Acad. Sci. U . S . A .76, 5294-5298. Heppner, G. H., and Miller, B. E. (1983). Cancer Metastasis Reo. 2, 5-23. Heppner, G . H., Dexter, D. L., DeNucci, T., Miller, F. R., and Calabresi, P. (1978). Cancer Res. 38, 3758. Hewitt, H. B. (1978). Ado. Cancer Res. 27, 149-200. Hewitt, H. B., Blake, E. R., and Walder, A. S. (1976). Br. J. Cancer 33, 241-259. Hibbs, J. B., Jr. (1974).J . Natl. Cancer Znst. 53, 1487-1492. Holland, J. F., and Frei, E. (1973). “Cancer Medicine,” p. 1841. Lea & Febinger, Philadelphia, Pennsylvania. Hoover, H. C., and Ketcham, A. S. (1975). Am. J . Surg. 130, 405-411. Irimura, T., Gonzales, R., and Nicolson, G . L. (1981). Cancer Res. 41, 3411-3418. Isaacs, J. T., and Coffey, D. S. (1981). Cancer Res. 41, 5070-5075. Isaacs, J. T., Wake, N., Coffey, D. S., and Sandberg, A. A. (1982). Cancer Res. 42,2353-2361. Jerne, N. K. (1976). ?Zaruey Lect. Ser. 70, 93-110.

68

VOLKER SCHIRRMACHER

Jones, P. A , , and DeClerck, Y. A. (1980). Cancer Res. 40, 3222-3227. Jones, P. A,, and DeClerck, Y. A. (1982). Cancer Metastasis Reu. 1, 289-317. Jones, P. A , , Neustein, H. B., Gonzales, F., and Bogenmann (1981). Cancer Res. 41, 46134620. Kadish, J. L., Butterfield, C. E., and Folkman, J. (1979). Tissue Cell 11, 99. Katzav, S., DeBaetselier, P., Tartakovsky, B., Feldman, M., and Segal, S. (1983). J . Natl. Cancer Inst. 71, 317-330. Kawashima, K., Nagura, E., Watanabe, E., Mizoguchi, K., Saga, S., Jsobe, K., Nakashima, I., Yamada, K., Oikawa, T., and Kojima, K. (1983). Int. J . Cancer 32, 507-514. Kedar, E., and Weiss, D. W. (1983). Ado. Cancer Res. 38, 171-287. Kefalides, N. E., ed. (1978). “Biology and Chemistry of Basement Membranes.” Academic Press, New York. Keller, R. (1983).J . Cancer Res. Clin. Oncol., in press. Kerbel, R. S., and Frost, P. (1982). Immunol. Today 3, 34-35. Kerbel, R. S . , and Davies, A. J. S. (1982). Lancet 2, 977-978. Kerbel, R. S., Florian, M., Man, M. S., Dennis, J., and McKenzie, I. F. C. (1980).J . Natl. Cancer Inst. 64, 1221-1230. Kerbel, R. S . , Dennis, J. W., Lagarde, A. E., and Frost, P. (1982). Cancer Metastasis Reo. 2, 99-140. Kerbel, R. S . , Lagarde, A , , Dennis, J. W., and Donaghue, T. P. (1983). Mol. Cell Biol. 3,523538. Key, M. E. (1983). Cancer Metastasis Reo. 2, 75-88. Key, M. E., Brandhorst, J . S., and Hanna, M. G. (1983).J . Immunol. 130, 2987-2992. Kiang, D. T., King, M., Zhang, H. J., Kennedy, B. J., and Wang, N. (1982). Science 216, 6870. Kieran, M. W., and Longenecker, B. M. (1983). Cancer Metastasis Reo. 2, 165. Kim, U. (1979). Breast Cancer 3, 1-49. Kinsey, D. L. (1960). Cancer 13, 674-676. Klein, E. (1955). E x p . Cell Res. 8, 188-212. Kobayashi, H. (1979). Ado. Cancer Res. 30, 279-299. Kramer, R. H., and Nicolson, G . L. (1979). Proc. Natl. Acad. Sci. U.S.A. 76, 5704-5708. Kramer, R. M., Vogel, K. C., and Nicholson, G . L. (1982).J . Biol. Chem. 257, 2678-2686. Kripke, M. L., and Fisher, M. S. (1976).J . Natl. Cancer Inst. 57, 211-215. Kripke, M. L., Gruys, E., and Fidler, I. J. (1978). Cancer Res. 38, 2962. Kuettner, K. E., Soble, L., Croxen, R. L., Marczynska, B., Hiti, J., and Harper, E. (1977). Science 196, 653-654. Kuettner, K. E., Pauli, B. U., and Soble, L. (1978). Cancer Res. 38, 277-287. Lagarde, A. E., Donaghue, T. P., Dennis, J. W., and Kerbel, R. S. (1983).J . Natl. Cancerlnst. 71, 183-191. Lala, P. K., Santer, V., and Rahl, K. S. (1980). Eur. 1. Cancer 16, 487-510. Lam, W. C., Delikatny, E. J., Orr, F. W., Wass, J.. Varani, J., and Ward, P. A. (1981).Am. 1. Pathol. 104, 69-76. Langer, R., and Murray, J. (1982). Appl. Biochem. Biotechnol. 8, 9-24. Larizza, L., and Schirrmacher, V. (1984). Cancer Metastasis Reo. in press. Larizza, L., Schirrmacher, V., Graf, L., Pfliiger, E., Perez, M., and Stohr, M. (1984a). Int. J . Cancer, November. Larizza, L., Schirrmacher, V., and Pfluger, E. (1984b). J . Exp. Med., in press. Lee, Y. T. N. (1977).J . Surg. Oncol. 9, 425-430. Lee, A , , and Langer, R. (1983). Science 221, 1185-1187. Lesot, H. Kiihl, U., and von der Mark, K. (1983). E M B O J . 2, 861-865.

CANCER METASTASIS

69

Liotta, L. (1982). Lab. Znoest. 47, 112-113. Liotta, L. A., and Hart, I. R., eds. (1982). “Tumor Invasion and Metastasis.” Nijhoff, The Hague. Liotta, L. A., Kleinerman, J., and Saidel, G . M. (1974). Cancer Res. 34, 997-1004. Liotta, L. A., Kleinerman, J., Catanzaro, P., and Rynbrandt, D. (1977).J . Natl. Cancer Znst. 58, 1427-1431. Liotta, L. A., Garbisa, S., and Tryggvason, K. (1982a). In “Tumor Invasion and Metastasis” (L. A. Liotta and I. R. Hart, eds.), pp. 319-333. Nijhoff, The Hague. Liotta, L. A., Thorgeirsson, U. P., and Garbisa, S. (1982b).Cancer Metastasis Reo. 1,277-288. Loewenstein, W. R. (1979). Biochim. Biophys. Acta 560, 165. Lohmann-Matthes, M.-L., Schleich, A., Shantz, G., and Schirrmacher, V. (i980). J . Natl. Cancer Znst. 64, 1413-1425. Maat, B., and Hilgard, P. (1981).J . Cancer Res. Clin. Oncol. 101, 275-283. McGovern, V. J, (1975). Pathology 7, 91. Malinoff, H. C., and Wicha, M. S. (1983).J . Cell B i d . 96, 1475-1479. Mareel, M. (1982). Drugs E x p . Clin. Res. VIII(5), 577-581. Mareel, M., De Ridder, L., De Brabander, M., and Vakaet, L. (1975).J . Natl. Cancer Znst. 54, 923-929. Mareel, M., Kint, J . , and Meyvisch, C. (1979). Virchows Arch. B Cell Pathol. 30, 95-111. Meltzer, M. S. (1981).J . Zmmunol. 127, 179. Michalides, R., Wagenaar, E., and Sluyser, M. (1982). Cancer Res. 42, 1154-1158. Miller, B. E., Miller, F. R., Leith, J., and Heppner, G. H. (1980).Cancer Res. 40,3977-3981. Miller, B. E., Miller, F. R., and Heppner, G . H. (1981). Cancer Res. 41, 4378-4381. Miller, E. J . (1976). Mol. Cell. Biochem. 13, 165-192. Miller, F. R. (1982). Cancer Metastasis Reu. 1, 319-334. Miller, F. R. (1983). Znoasion Metastasis 3, 234-242. Miller, F. R., and Heppner, G. H. (1979).J . Natl. Cancer Znst. 63, 1457. Miller, F. R., and Heppner, G . H. (1980). Proc. Am. Assoc. Cancer Res. 21, 201. Miller, F. R., Miller, B. E., and Heppner, G. H. (1983). Znuasion Metastasis 3, 22-31. Miller, L. D., and Olsson, C. (1971).J . Am. Vet. Med. Assoc. 158, 1536. Miner, K. M., Kawaguchi, T., Uba, G. W., and Nicolson, G. L. (1982).Cancer Res. 42,46314638. Montesano, R., Orci, L., and Vasselli, P. (1983).J . Cell Biol. 97, 1648. Nakajima, M., Trimura, T., Ferrante, D. D., Ferrante, N. D., and Nicolson, G. L. (1983). Science 220, 611-613. Naor, D. (1979). Ado. Cancer Res. 29, 45. Naor, D. (1983). Cancer Zmmunol. Zmmunother. 16, 1-10, National Cancer Institute Monographs (1976). “Conference on Spontaneous Regression of Cancer.” Vol. 44. Neri, A . , and Nicolson, G. L. (1981). Znt. J . Cancer 28, 731-738. Netland, P. A,, and Zetter, B. R. (1984). Science 224, 1113-1115. Nicolson, G. L. (1978a). Bioscience 28, 411-447. Nicolson, G . L. (1978b).J . Histochem. Cytochem. 30, 214-220. Nicolson, G. L. (1979). Sci. Am. 240, 50-61. Nicolson, G. (1982). Biochim. Biophys. Acta 695, 113-176. Nicolson, G . L., and Winkelhake, J. L. (1975). Nature (London) 255, 230-232. Nicolson, G. L., Brunson, K. W., and Fidler, I. J. (1978). Cancer Res. 38, 4105-4111. Noble, R. L., and Hoover, L. (1975). Cancer Res. 35, 2935. Noguchi, P.-D., Johnson, J. B., O’Donnel, R., and Petricciani, J. C. (1978).Science 199, 19801983.

70

VOLKER S C H I R R M A C H E R

Nowell, P. (1976). Science 194, 23-28. Nowell, P. (1982).In “Tumor Cell Heterogeneity. Origins and Implications” (A. H. Owens, D. S. Coffey, and S. B. Baylin, eds.), Vol. 22, pp. 351-365. Academic Press, New York. Nowotny, A., and Grohsman, J. (1973). Int. Arch. Allergy 44, 434-440. Olsson, L., and Ebl)esen, P. (1979).J. Natl. Cancer Inst. 62, 623-627. Orr, F. W., Lam, W. C., Delikatny, E. J . , Mokashi, S., and Varani, J. (1981). Invasion Metastasis 1, 239-247. Ozaki, T., Yoshida, K., Ushijima, K., and Hayashi, H. (1971). Int. J . Cancer 7, 93-100. Paget, S . (1889). Lancet 1, 571-573. Parmiani, G., Carhone, G., Invernizzi, G., Pierotti, M. A,, Sensi, M. L., Rogers, J. J., and Appella, E. (1979). Inimunogenetics 9, 1-24. Pauli, B. U., and Weinstein, R. S. (1982). Cancer Res. 42, 2289-2297. Pauli, B. U., Memoli, V. A , , and Kuettner, K. E. (1981). Cancer Res. 41, 2084-2091. Pauli, B. U., Schwartz, 11. E., Thonar, E. J. M., and Kuettner, K. E. (1983).Cancer Metastasis Reo. 2, 129-152. Pierce, G. B. (1974). In “World Symposium on Model Studies in Chemical Carcinogenesis” (T. ’so Pop, and J. A. DiPaolo, eds.), pp. 463-472. Dekker, New York. Pimm, M. V., and Baldwin, R. W. (1977). Int. /. Cancer 20, 37-43. Pimm, M. V., Embleton, M. J., and Baldwin, R. W. (1980). Int. J . Cancer 25, 621-629. Poste, G. (1982). Cancer Metastasis Reo. 1, 141-199. Poste, G . , and Fidler, I. J. (1980). Nature (London) 283, 139-146. Poste, G., and Nicolson, G . L. (1980). Proc. Natl. Acad. Sci. U . S . A . 77, 399-403. Poste, G., and Nicolson, G. L. (1983).In “Biomembranes” (A. Nowotny, ed.), Vol. 11, pp. 341364. Plenum, New York. Poste, G . , Doll, J., Hart, 1. R., and Fidler, I. J. (1980). Cancer Res. 40, 1636-1644. Poste, G., Doll, J., and Fidler, I. Y. (1981). Proc. Natl. Acad. Sci. U . S . A . 78, 6226-6230. Poste, G . ,Tzeng, J., Doll, J . , Greig, R., Rieman, D., and Zeidman, I. (1982).Proc. Natl. Acad. Sci. U . S . A . 79, 6574-6578. Pourreau-Schneider. N . , Felix, H., and Haenimerli, G. (1977). Virchows Arch. B Cell Pathol. 23, 257-264. Prager, M. D., and Baechtel, F. S. (1973). Methods Cancer Res. 9, 339-400. Prehn, R. T. (1972). Science 176, 170-171. Proctor, J. W. (1976). Br. J. Cancer 34, 652-654. Rao, V. S . , Bennett, J. A , , Grodzicki, R. L. et al. (1979). Cell. Immunol. 46, 227-238. Rapin, A. M., and Burger, M. M. (1974). Adc. Cancer Res. 20, 1-91. Raz, A , , and Ben-Ze’ev, A. (1983). Science 221, 1307-1310. Raz, A,, Haiina, N., and Fidler, I. J . (1981).J. Natl. Cancer Inst. 66, 183. Reading, C. L., Brunson, K. W., Torriani, M., and Nicolson, G . L. (1980). Proc. Natl. Acad. Sci. U . S . A . 77, 5943. Reid, L. C. M. (1982). In “From Gene to Protein: Translation into Biotechnology” (F. Ahmad, J. Schultz, E. C. Smith, W. J. Whelan, and S. Black, eds.), pp. 53-73. Academic Press, New York. Rice, J. M. (1972). Natl. Cancer Inst. Monogr. 35, 197. Rifkin, D. B . , and Crowe, R. M. (1975). lfoppe Seyler’s Z. Physiol. Chem. 358, 1525-1531. Riggs, A. D., and Jones, P. A. (1983). Ado. Cancer Res., 40, 1-30. Robinson, M . K . , and Wheelock, E. F. (1984). Cell. Immunol., in press. Rollins, B. J., and Culp, L. A. (1979). Biochemistry 18, 141-148. Romnaldez, A. G., Jr., and Ward, P. A. (1975). Proc. Notl. Acad. Sci. U . S . A . 72, 4128-4132. Roos, E., Van de Pavert, I. V., and Middelkoop, 0.P. (1981).J. Cell Sci. 47, 385-397. Ryd, W., Hagmar, B . , and Erkell, L. J. (1983). Submitted.

CANCER METASTASIS

71

Scher, C. D., Haudenschild, C., and Klagsbrun, M. (1976). Cell 8, 373-382. Schirrmacher, V. (1979). Int. J. Cancer 24, 80-86. Schirrmacher, V. (1980). Immunobiology 157, 89-98. Schirrmacher, V., and Bosslet, K. (1982). Cancer Immunol. Zrnmunother. 13, 62-68. Schirrmacher, V., Shantz, G., Clauer, K., Komitowski, D., Zimmermann, H.-P., and Lohmann-Matthes, M.-L. (1979). Znt. J . Cancer 23, 233-244. Schirrmacher, V., Cheingsong-Popov, R., and Arnheiter, H. (1980).J. E x p . Med. 151, 984989. Schirrmacher, V., Bosslet, K., Altevogt, P., Russmann, E., Beck, L., and Fogel, M. (1982a). Proc. Meet. EACR, 6th pp. 155-164. Schirrmacher, V., Altevogt, P., Fogel, M., Dennis, J., Waller, C. A., Barz, D., Schwartz, R., Cheingsong-Popov, R., Springer, G . F., Robinson, P. J., Nebe, T., Brossmer, R., Vlodavsky, I., Paweletz, N., Zimmermann, H.-P., and Uhlenbruck, G. (19821,). Inoasion Metastasis 2, 313-360. Schirrmacher, V., Fogel, M., Russmann, E . , Bosslet, K., Altevogt, P., and Beck, L. (1982~). Cancer Metastasis Rev. 1, 241-274. Schirrmacher, V., Waller, C., and Vlodavsky, I. (1982d).I n “Band T Cell Tumors: Biological and Clinical Aspects” (E. Vitetta and C. F. Fox, eds.). Academic Press, New York. Schirrmacher, V., Altevogt, P., and Bosslet, K. (1983). I n “Biochemical and Biological Markers of Neoplastic Transformation” (P. Chandra, ed.), pp. 121-131. Plenum, New York. Schleich, A. B., Frick, M., and Mayer, A. (1974). 2. Krebsforsch. 82, 247-255. Schreiber, H. (1984). Ado. Cancer Res. 41, 291-321. Shapiro, J., Jersky, J., Katzav, S., Feldman, M., and Segal, S. (1981).J. Clin. Incest. 68, 678685. Shearman, P. J., Gallatin, W. M . , and Longenecker, B. M. (1980). Nature (London)286, 267269. Sinha, A. A. (1981). In “Hormonal Management of Endocrine-related Cancer” (B. A. Stoll, ed.), pp. 13-19, Lloyd-Luke Medical Books, London. Sinha, A. A,, Blackard, C. E., and Seal, U. S. (1977). Cancer 4, 2836-2850. Sluyser, M., Degoeij, K. C. J., and Evers, S. C. (1981). Cancer Lett. 13, 71-77. Smith, B., and Sanger, R. (1982). Cancer Res. 42, 389-396. Sordat, B., and Wang, W. R. (1984). Behring Inst. Mitt. 74, 291-300. Springer, G . F., Desai, P. R., Tegtmeyer, H., Schirrmacher, V., and Cheingsong-Popov, R. (1983a). Naturwissetischaften 70, 98. Springer, G . F., Cheingsong-Popov, R., Schirrmacher, V., Desai, P. R., and Tegtmeyer, H. (1983b).J. Biol. Chem. 258, 5702-5706. Stackpole, C. W. (1981). Nature (London) 289, 798-800. Strauli, P., Barrett, A. J.. and Baici, A. (eds.) (1980). “Proteinases and Tumor Invasion.” Raven, New York. Stevens, S. K., Weissman, I. L., and Butcher, E. C. (1982).J . Inununol. 128, 844. Subak-Sharpe, H., Burk, R. R., and Pitts, I. D. (1969). J . Cell Sci. 4 , 353-361. Sugarbaker, E. V. (1979). Curr. Probl. Cancer 3, 3-59. Sugarbaker, E. V. (1981). In “Cancer Biology Reviews” (J. J. Marchalonis, M . G . Hanna, and I. J. Fidler, eds.), Vol. 2, pp. 235-278. Dekker, New York. Sugarbaker, E. V., and Cohen, A. M. (1972). Surgery 72, 155-161. Sugarbaker, E. V., Coben, A. M., and Ketcham, A. S. (1971). Ann. Surg. 174, 161-166. Sugarbaker, E. V., Thornthwaite, J., and Ketcham, A. S. (1977). In “Cancer Invasion and Metastasis” (S. B. Day et a l . , eds.). Raven, New York. Takahashi, S.. Yoichi, K., Nakatani, I., Innui, S., Kojima, K., and Shiratori, T. (1978).J . Natl. Cancer Inst. 60, 925.

72

VOLKER SCHIRRMACHER

Talmadge, J. E., and Fidler, I. J. (1982). Nature (London) 297, 593-594. Talmadge, J. E., Starkey, J . H., Davis, W. C., and Cohen, A. L. (1979).J. Suprarnol. Struct. 12, 227. Talmadge, J. E., Wolman, S. R., and Fidler, I. J. (1982). Science 217, 361-363. Talmadge, J., Benedict, K., Madsen, J., and Fidler, I. J. (1984). Cancer Res., in press. Tao, T. Mi., and Burger, M. M. (1977). Nature (London) 270, 437. Tarin, D., and Price, J. E. (1981). Cancer Res. 41, 3604-3609. Taylor, S., and Folknian, J. (1982). Nature (London) 297, 307. Tchao, R . , Schleich, A. B., Frick, M., and Mayer, A. (1980). In “Metastasis, Clinical and Experimental Aspects” (K. Hellmann, P. Hilgard, and S. Eccles, eds.), pp. 28-32. Nijhoff, The Hague. Terranova, V. P., Hao, C. N., Kalebic, T., Margulies, I. H., and Liotta, L. A. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 444-448. Thorgeirsson, U . P., Liotta, L. A., Kalebic, T., Margulies, I. M., Thomas, K., Rios-Candelore, M., and Husso, H. G. (1982).J . Natl. Cancer Inst. 69, 1049-1054. Timpl, H., Martin, G. R., Bruckner, P., Wick, G., and Wiedemann, H . (1978). Eur. J. Biochem. 84, 43. Ting, C. C., and Rodrigues, D. (1980). Proc. Natl. Acad. Sci. U.S.A. 77, 4256-4269. Toole, B. P. (1981). In “Cell Biology of Extracellular Matrix” (E. D. Hay, ed.), pp. 259-294. Plenum, New York. Toole, B. P., Biswas, C., and Gross, J. (1979). Proc. Natl. Acad. Sci. U.S.A. 76, 6299-6303. Treves, A. J., Cohen, I. H., and Feldman, M. (1975).J . Natl. Cancer Inst. 54, 777-780. Urban, J . L., and Schreiber, H. (1983).J . E x p . Med. 157, 642-656. Urban, J. L., Holland, J. M., Kripke, M. L., and Schreiber, H. (1982).J. Exp. Med. 156, 1025. Vaage, J. (1978). Cancer Zmtnunol. Initnunother. 4, 257-261. Vaage, J. (1980). Cancer Res. 40, 3495-3501. Varani, J. (1982). Cancer Metastasis Rev. 1, 17-28. Varani, J.. Wass, J., Piontek, G., and Ward, P. A. (1981). Cell Biol. Znt. Rep. 5, 525-530. Vartio, T., Vaheri, A., DePetro, G., and Barlati, S. (1983). Znoasion Metastasis 3, 125-138. Viadana, E., Bross, I. D. J . , and Pickren, J. W. (1978a).In “Pulmonary Metastases” (L. Weiss and H. Gilbert, eds.), pp. 142-167. Hall, Boston, Massachusetts. Viadana, E., Bross, I. D. J., and Pickren, J. W. (1978b). Oncology 35, 87-96. Vlodavsky, I., Fuks, Z . , and Schirrmacher, V. (1982).In “The Endotlielial Cell-A Pluripotent Control Cell of the Vessel Wall” (D. G. S. Thilo-Korner and H. I. Freshney, eds.), pp. 126-157. Karger, Basel. Vlodavsky, I., Schirrmacher, V., Ariav, Y., and Fuks, 2. (1983a).Inoasioti Metastasis 3, 81-97. Vlodavsky, I . , Fuks, 2.. Bar-Ner, M., Ariav, Y., and Schirrmacher, V. (1983b).Cancer Res. 43, 2704-271 1. Vollmers, P. H., and Birchmeier, W. (1983a). Proc. Natl. Acad. Sci. U.S.A. 80, 3729-3733. Vollmers, P. H . , and Birchmeier, W. (1983b). Proc. Natl. Acad. Sci. U.S.A. 80, 6863-6867. Vollmers, H. P., Imhof, B. A,, Braun, S., Waller, C. A , , Schirrmacher, V., and Birchmeier, W. (1984). FEBS Lett. 172, 17-20. Walton, B. (1978). Anaesthesia 33, 322-348. Warner, J . F., and Dennert, G. (1982). Nature (London) 300, 31-34. Warren, B. A. (1981). Cancer Biol. Reo. 2, 95-169. Waxler, B., Kuettner, K. E., and Pauli, B. U. (1982). Tissue Cell 14, 657-667. Weiss, L. (1975). In “Current Concepts in Cancer” (P. Rubin, ed.), Vol. 1, pp. 97-99. Pergainon, Oxford. Weiss, L. (1976). In “Fundamental Aspects of Metastasis” (L. Weiss, ed.), pp. 1-6. Elsevier, Amsterdam.

CANCER METASTASIS

73

Weiss, L. (1980). Znt. J . Cancer 25, 385-392. Weiss, L. (1982). In “Liver Metastasis” (L. Weiss and H. A. Gilbert, eds.), pp. 126-157. Hall, Boston, Massachusetts. Weiss, L. (1983). Znoasion Metastasis 3, 193-207. Weiss, L., Ward, P. M., and Holmes, J. C. (1983). Znt. /. Cancer 32. 79-83. Welch, D. R . , Neri, A., and Nicolson, G . L. (1984). Inuasion Metastasis, in press. Wheelock, E. F., and Robinson, M. K. (1983). Lob. Znoest. 48, 120-139. Wheelock, E. F., Weinhold, K. J., and Levich, J. (1981). Ado. Cancer Res. 34, 107. Wheelock, E. F., Robinson, M. K., and Truitt, G . A. (1982). Cancer Metastasis Reo. 1,29-44. Wigler, M., Levy, D., and Perucho, M. (1981). Cell 24, 33-40. Willis, R. A. (1973). “The Spread of Tumors in the Human Body,” 3rd ed. Butterworths, London. Wiltrout, R. H., Frost, P., Morrison, M. K., and Kerbel, R. S . (1979). Cancer Res. 39, 40344041. Wortzel, R. D., Urban, J. L., Phillips, C., and Schreiber, H. (1982). Fed. Proc., Fed. Am. SOC. Erp. Biol. 41, 726. Yogeeswaran, G., and Salk, P. L. (1981). Science 212, 1514-1516. Yuhas, J. M., and Tarleton, A. E. (1978). Cancer Res. 38, 3584. Zacharski, L. R . , Henderson, W. G. et al. (1981). Veterans Administration Study, No. 75, J. Am. Med. Assoc. 245, 831-835. Zamora, P. O., Danielson, K. G., and Hosick, H. L. (1980). Cancer Res. 40, 4631-4639.

This Page Intentionally Left Blank

THE CANINE TRANSMISSIBLE VENEREAL TUMOR: A UNIQUE RESULT OF TUMOR PROGRESSION D. Cohen Department of Microbiology and Immunology, Faculty of Health Sciences, Ben-Gurion University of the Negev, Beer-Sheva. Israel

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 Histology, Ultrastructure, and Cell Markers . . . . . . . . . . . . . . . . . 78 Clinical Aspects and Therapy.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 TVT-Induced Polycythemia 82 Experimental Transplantation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 The Immune Response against the TVT . . . . . . . . . . . . . . . . 86 A. In Vioo Studies.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 B. The Antigens of the TVT. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 C. In Vitro S .......................... 89 VII. Etiology and t sion. .......................... 94 A. Cytogenet 94 B. Lack of Pz-Microglobulin Expression on the Surface of TVT Cells. . . . . . . . . . 97 C. Canine Coitus and the Transmission of the TVT 103 VIII. Conclusion and Summary.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 References . . . . . . . . . . . . . . 107

I. 11. 111. IV. V. VI.

I.

Introduction

It has been established in recent years that a number of oncogenes are involved in the malignant transformation of normal cells (Bishop, 1983). Although cells expressing these oncogenes are potentially malignant, most probably they do not give rise to clinically detectable cancer soon after transformation. Presumably, host homeostatic mechanisms such as the immune and the endocrine systems prevent the unlimited expansion of the transformed clone(s). The transformed cells undergo, over the years, incidental changes some of which enhance their chances of survival and their ability to invade adjacent tissues. Once these cells undergo sufficient multiple changes that enable them to overcome host homeostatic mechanisms, they develop into clinical cancer. These multiple stepwise changes have been collectively designated by Foulds under the general term of “tumor progression’ (Foulds, 1958; Klein and Klein, 1977). The canine transmissible venereal tumor (TVT) is a naturally occurring neoplastic disease that affects the external genitalia of both sexes and is transmitted during coitus by the transplantation of cells. The significance of the TVT lies in the fact that this neoplasm evolved into a tumor that is 75 Copyright 8 1985 by Academic Press, Inc. ADVANCES IN CANCER RESEARCH, VOL. 43

All rights of reproduction in M Y Form reserved. ISBN 0-12-006643-2

76

D. COHEN

autonomous from the original host and became, in fact, a neoplastic cell that is a parasite. Such a tumor devleoped in the dog and not in other species, probably because of some unusual characteristics of canine coitus. Following coitus in the dog, injuries of the vaginal and penile mucosa occur, thus providing the bed for tumor transplantation. The tumor cells are not rejected as a first set allograft, probably because of certain alterations in the expression of histocompatibility antigens that we have recently described (Cohen et al., 1984), and that also occur in undifferentiated human tumors. Because the TVT may be a neoplasm that is as old as the family Canidae or the species Canis fami2iaris or, alternatively, one that arose repeatedly at an unknown frequency and has since been transplanted from one animal to the other, the TVT has probably been subjected to tumor progression in a more extensive manner than human tumors or experimentally transplanted neoplasia. It seems therefore reasonable to assume that some of the phenomena that are associated with tumor progression are expressed in the TVT in a more pronounced manner than in human tumors. Some of the characteristics of this neoplasm that have already been established, indeed, support this view. Investigations of various aspects of the tumor biology of the TVT may, therefore, provide clues to similar phenomena that occur in human tumors in a less pronounced manner. In addition to the natural mode of transmission, the TVT can also be induced in adult, immunocompetent, allogeneic dogs by inoculation with living tumor cells and this in fact was used in the first recorded successful experimental transmission of a tumor (Novinsky, 1876; also see Shimkin, 1955). At the turn of the century, the transmissibility of the TVT seemed to support the concept that cancer is “infectious” and prompted several investigations of this neoplasm (Wehr, 1888, 1889; Geissler, 1895; Smith and Washbourn, 1898a,b, 1899; Sticker, 1904, 1905, 1906; Bashford et al., 1905; Beebe, 1907; Crile and Beebe, 1908; Wade, 1908). The tumor was designated in these studies as infective sarcoma (Smith and Washbourn, 1899), transplantable lymphosarcoma (Sticker, 1904), and infective granuloma (Bashford et al., 1905). Sticker, a contemporary and co-worker of Paul Ehrlich, carried out detailed immunological studies of the diease and the TVT has since also been referred to as “Sticker’s sarcoma.” In the more recent literature, the most commonly used names are the transmissible venereal tumor and the transmissible venereal sarcoma. Because of the transmissibility of the TVT, numerous attempts have been made over the years to demonstrate a causative oncogenic viral agent, but to date, no reproducible evidence is available demonstrating a viral etiology for the tumor. In contrast, cytogenetic studies showed that the TVT,in different parts of the world, is characterized by extensive and specific chromosomal aberrations (Takayama and Makino, 1961; Weber et al., 1965; Barski and

C A N I N E TRANSMISSIBLE VENEREAL TUMOR

77

Cornefert-Jensen, 1966). Since tumor-specific karyotypic changes in virally induced tumors as well as in nonvirally induced tumors are, as a rule, not as extensive (reviewed by Klein, 1981; Miller and Miller, 1983), the results of the cytogenetic studies suggest that the TVT is a cell-transmitted neoplasm. Immunological investigations also support this view (Epstein and Bennett,

1974). Another aspect that has attracted attention to this unique tumor is the role played by the immune response in determining the course of the disease. Sticker (1906) already observed that the TVT often regresses spontaneously and demonstrated that spontaneous regression is followed by transplantation immunity. The antigens of the TVT and the immune response against this tumor have since been subjected to numerous investigations. The TVT has also been used as a tumor model in studies that require a large animal for the experimental manipulation or for simulating conditions that are similar to the conditions in humans. For example, tumor-imaging studies by use of radiogallium-transferrin complex uptake have been carried out using the TVT as a model (Wong et al., 1980; Terner et al., 1981). In other studies, autologous bone marrow rescue was investigated following high-dose chemotherapy of the TVT (Epstein and Sarpel, 1980) and plasmapheresis was carried out on TVT bearing dogs in an attempt to remove blocking factors and induce tumor regression (Zander et al., 1980). Yet another characteristic of the TVT that is of interest is the induction of polycythemia (erythrocytosis) in animals bearing a large tumor burden. In such animals, the level of erythropoietin is increased, and this hormone can be extracted from the tumor (Cohen, Jeglum, Goldwasser, and Herberman, unpublished data). Current interest in the TVT also stems from its clinical importance as a canine neoplastic disease. The naturally occurring TVT does not seem to regress spontaneously as often as the subcutaneously transplanted tumor, and metastatic spread is not uncommon. Recent therapeutic studies of the naturally occurring tumor show, that with appropriate treatment, a high rate of complete regression can be obtained (Brown et al., 1980; Calvert et al., 1982; Thrall, 1982). The TVT occurs throughout the world, but seems to be more prevalent in temperate zones (Rust, 1949; Higgins, 1966). In a recent study, the relationship between prevalence rates of the TVT in the United States and geographic and climatic conditions was analyzed (Hayes et al., 1983). This analysis shows that hospital prevalence of the TVT is inversely correlated to geographic latitude and positively correlated to higher mean annual temperature and increased rainfall. Hayes et al. (1983) point out that the TVT and Kaposi’s sarcoma have several common characteristics. These include increased susceptibility in immunosuppressed hosts, possible viral etiology, and association between high prevalence and climatic conditions.

78

D. COHEN

Hayes et al. (1983)propose that the TVT may serve as an animal model for Kaposi’s sarcoma and they raise the possibility that both tumors are more prevalent in temperate climates because such areas favor the survival of a hypothetical oncogenic virus or an insect vector. It has, however, been observed that the TVT is prevalent in areas where stray dogs are abundant and serve as a reservoir for the disease (Rust, 1949; Higgins, 1966). Since stray dogs are often more common and probably survive better in warm areas, the high prevalence of the TVT in temperate zones may well be associated with large numbers of stray dogs and lack of control of the breeding activity of the dog population. Little information is available about the prevalence of the TVT in different countries, but the occurrence of the TVT has been recorded in recent years in the United States (Brown et al., 1980; Calvert et al., 1982; Thrall, 1982) and several European countries including France (Woimant and Chdaux, 1977), the USSR (Osipov and Golubeva, 1976), Poland (Wasecki and Mazur, 1977), Yugoslavia (Cermak et al., 1976), and Italy (Muzetto et al., 1975). Transmissible venereal tumor cases have also been recorded in recent years in Australia (Locke et al., 1975), Japan (Oshimura et al., 1973), Iran (Ivoghli, 1977), India (Mohanty and Rajya, 1977), Israel (Cohen et al., 1978), Nigeria (Idowu, 1977), and Kenya (Ndiritu et al., 1977). Several brief reviews dealing with the canine TVT have been published in recent years (Cohen, 1978, 1980a; Brown et al., 1981). This article represents an attempt to review the available information about this unique tumor in a more comprehensive manner, and discuss the relevance of some of the data to current concepts in cancer research.

I I . Histology, Ultrastructure, and Cell Markers Although the histology and ultrastructure of the TVT have been studied extensively, little is known about the histogenesis of this tumor. The TVT is defined histologically as an undifferentiated round-cell neoplasm of reticuloendothelial origin (Bloom et al., 1951; Samso, 1966; Lombard and Yulzari, 1967; Drommer and Schulz, 1969; Duncan and Prasse, 1979). In formalin-fixed, hematoxylin and eosin-stained sections, the tumor cells are round, slightly oval, or polyhedral, and display uniformity in size arid appearance. The diameter of TVT cells ranges from 14 to 30 p,m (Duncan and Prasse, 1979). The centrally located, large oval or round nuclei contain coarsely aggregated chromatin and usually one eccentrically located prominent nucleolus. The abundant pale blue cytoplasm is finely granular and often contains vacuoles. The cytoplasmic outline of the cells can be seen in areas where the cells are less densely packed. Several mitotic figures can lie observed in most high-power fields. The cells are closely packed and ar-

CANINE TRANSMISSIBLE VENEREAL TUMOR

79

ranged in diffuse masses, clusters, or cords along fibrous connective tissue trabeculae that contain blood vessels and transverse the tumor tissue in an irregular fashion. Fibrous stroma is usually scanty in tumors that have developed in the host for only several weeks and is more abundant in wellestablished TVT. Regressing tumors are infiltrated by lymphocytes, plasma cells, and macrophages (Wade, 1908; Bloom et al., 1951; Weir et al., 1978; Chandler and Yang, 1981).The TVT can be distinguished on morphological grounds from other canine cutaneous neoplasia such as histiocytoma (Drommer and Schulz, 1969), cutaneous lymphosarcoma, and mast cell tumors (Duncan and Prasse, 1979). Histocheinical studies of the TVT have been carried out in an attempt to define the histogenesis of the tumor. Transmissible venereal tumor cells do not contain metachromatic granules that are characteristic of mast cell tumors (Bloom et d., 1951) and they are negative for alkaline phosphatase (Bloom et al., 1951), a-naphthyl butyrate and anaphthyl acetate, nonspecific esterase, myeloperoxidase, and P-glucuronidase (Cohen, unpublished data). Transmissible venereal tumor cells are characterized by strong acid phosphatase activity (Bloom et a l . , 1951; Hernandez-Jauregui et al., 1973) that is tartrate-sensitive (Cohen, unpublished data). In humans, acid phosphatase activity in reticulum cells, leukemic reticulosis (hairy cell leukemia), and cutaneous histiocytic lymphomas is tartrate-resistant and represents isoenzyme 5 activity (Yam et al., 1971; Krishnan et al., 1983), whereas cutaneous T-cell and B-cell lymphomas are acid-phosphatase positive but tartrate-sensitive. Catovsky et al. (1978) showed that 90% of human T-cell acute lymphoblastic leukemias (ALL) are acid-phosphatase positive whereas only 2% of common ALL and 10% of null ALL are positive. Ultrastructural characteristics of the TVT also suggest that the tumor is of reticuloendothelial origin (Carteaud et al., 1963; Lombard and Cabanie, 1967; Droinmer and Schulz, 1969; Murray et al., 1961; Cabanie et aZ., 1973; Hernandez-Jauregui et a l . , 1973; Battistaci and Moriconi, 1974; Cockrill and Beasley, 1975; Kennedy et d . ,1977). In electron microscopic studies, TVT cells appear as round cells, with a high nuclear/cytoplasmic ratio and the nucleus usually contains a large nucleolus. The cytoplasm contains an extensive network of granular endoplasmic reticulum, few swollen mitochondria, vacuoles, a Golgi apparatus, and numerous free and clustered ribosomes. The cytoplasm also contains lamellar arrays of unknown nature (Cockrill and Beasley, 1975; Kennedy et al., 1977). In areas where TVT cells are tightly clustered, electron microscopic studies revealed numerous microvilli and extensive interdigitations between adjacent cells. The ultrastructure of regressing TVT is characteristic of degenerating cells and consists of clumping of nuclear chromatin along the nuclear envelope and a widened perinuclear space, swelling and vacuolation of mitochondria, and a decrease in the

80

D. COHEN

amount of endoplasinic reticulum and ribosomes (Cabanie et al., 1973; Cockrill and Beasley, 1975; Kennedy et al., 1977). Kennedy et al. (1977)observed that the number of irregularly shaped round cells and fibroblast-like cells increases with increasing tumor mass and suggested that the characteristic round TVT cells may transform into fibroblast-like cells. Few attempts have been made to define TVT cell markers. Yang et al. (1976a)have shown that 3 to 35% of TVT cells in mechanically disaggregated single-cell suspensions form rosettes with human erythrocytes. This finding was confirmed by Zarrouk et al. (1980). In the dog, a subpopulation of T lymphocytes (Bowles et al., 1975; Zander et al., 1975; Onions, 1977), as well as monocytes (Esser et al., 1977), neutrophils (Krakowa and Guyot, 1977), and eosinophils (Yang and Kennedy, 1976), can form rosettes with human erythrocytes. In addition to erythrocyte rosette formation, up to 30% of the cells in disaggregated TVT cell suspensions form rosettes with antibodycoated sheep erythrocytes (Cohen et al., 1971). Most of these rosette-forming cells are probably lymphocytes and macrophages, but a proportion of the rosette-forming cells may be TVT cells. The available information is obviously insufficient to determine the cellular origin of the TVT. A systematic search for lymphoid cell markers may lead to a better classification of this undifferentiated neoplasm. Ill. Clinical Aspects and Therapy

The canine TVT usually affects the external genital organs of both sexes, occurs most frequently in adult animals, and may develop in any breed of dog (Karlson and Mann, 1952; Brown et al., 1980). Transmission of the tumor occurs by transplantation of viable tumor cells during coitus from the affected male or female to the other breeding partner. This mode of transmission is highly efficient. Smith and Washbourn (1898a)observed transmission of the TVT by coitus from one affected male dog to 11of 12 bitches. The early symptoms of the disease include serous or sanguineous discharge from the vulva or the prepuce and licking of the affected organs. In the early stages of the disease, the tumor appears as single or multiple small gray-red firm nodules involving in the male the bulbus glandis, but may also involve the glans penis and the prepuce (Karlson and Mann, 1952). In the female, the TVT may develop at any site within the vagina. At later stages of the disease, the tumor appears as a cauliflower-like, pedunculated or multilobulated mass which frequently bleeds and ulcerates. The naturally occurring TVT can reach a diameter of 10 cm or more and may invade adjacent tissues. The clinical diagnosis of the disease can be confirmed by cytologic examination of impression smears and by histologic examination of formalin-fixed tissue. The naturally occurring TVT may regress spontaneously (Higgins,

CANINE TRANSMISSIBLE VENEREAL TUMOR

81

1966) but the frequency of regression has not been determined. It seems, however, that spontaneous regressions of the naturally occurring disease are not as frequent as the regressions observed in the experimentally induced TVT (Brown et al., 1980). The TVT usually develops as a local tumor, but numerous cases of metastatic TVT have been recorded. In a recent study, metastatic spread of the tumor was observed in 2 of 30 naturally occurring TVT cases (Brown et al., 1980). Metastatic TVT has been observed most frequently in the inguinal and iliac draining lymph nodes (Prier and Johnson, 1964; Ndiritu et al., 1977; Brown et al., 1980), but the TVT may also metastasize to the spleen (Prier and Johnson, 1964; Oduye et al., 1973), liver (Oduye et al., 1973), brain (Adams and Slaughter, 1970), pituitary (Spence et al., 1978), and the lungs (Oduye et al., 1973). The TVT may also involve extragenital sites with or without involvement of the external genitalia. It seems that in some of these cases, the tumor is transmitted by licking of tumor-affected genital organs. Extragenital TVT without genital involvement has been observed in the skin (Higgins, 1966), eyelid (Abbott, 1966), and oral and nasal cavities (Ndiritu et al., 1977; Weir et al., 1978; Gorman et al., 1983). Several approaches to the treatment of the TVT have been explored, but only in recent years have effective treatment regimens been defined. In resectable TVT cases, surgical removal of the tumor may be curative (Brodey and Roszel, 1967; Muzetto et al., 1975). However, the tumor is often not resectable and recurrence following surgical excision has been reported (Brodey and Roszel, 1967; Brown et al., 1980; Thrall, 1982). Radiotherapy of the TVT is highly effective (Nanta et al., 1951; Osipov and Golubeva, 1976). In a recent study, 18 TVT cases were treated with orthovoltage radiotherapy. Complete regression of the tumor in all 18 dogs was induced by radiotherapy and no recurrence was observed within 1 year after completion of treatment. A dose of 10 Gy (1000 rad) was applied in each treatment, and in most cases, one to three treatments were required to induce complete remission of the tumor (Thrall, 1982). The TVT can also be treated successfully with chemotherapy. Attempts to treat the TVT with cyclophosphamide alone were only partially successful (Hernandez-Jauregui et al., 1973), but a combination of busulfan and cyclophosphamide induced regression in all eight experimentally induced tumors (Epstein and Sarpel, 1980). In another study, 30 dogs with the naturally occurring TVT were treated with a combination chemotherapy regimen consisting of vincristine, cyclophosphamide, and methotrexate. In 28 of these animals the tumor regressed completely; 1 tumor showed partial response to treatment, and in one animal, only a minimal response was observed. No tumor recurrence was observed following complete regression of the TVT (Brown et al., 1980). In a more recent study, it has been estab-

82

D. COHEN

lished that vincristine treatment alone is as effective as combination chernotherapy and results in fewer side effects. Of 41 naturally occurring TVT cases, 39 dogs had complete tumor regression without relapse. The treatment consisted of weekly intravenous administration of 0.025 mg vincristine/kg body weight. In most animals, tumor regression was initiated already after one treatment. A mean number of 3.3 treatments was required to induce complete regression in the 39 dogs that responded to chemotherapy (Clavert et al., 1982). Animals in which the TVT had regressed following busulfan and cyclophosphamide treatment were immune to tumor challenge (Epstein and Sarpel, 1980). It seems, therefore, that the immune response toward the TVT is involved in the prevention of tumor recurrence following chemotherapy. It is possible that the immune response is also involved in the induction of tumor regression during and following radio- and chemotherapy. Immunotherapy of the TVT by intralesional BCG administration (Hess et ul., 1977) is discussed in Section VI. IV. TVT-Induced Polycythemia

Nonendocrine neoplasia often synthesize and release large quantitites of hormones. Such ectopic hormone production may lead to the development of paraneoplastic syndromes by inducing endocrine imbalance. Chronic production of adrenocorticotropic hormone by human sinall-cell carcinoma of the lung, for example, may lead to the development of Cushing’s syndrome (Amatruda and Upton, 1974; Hansen and Hummer, 1979; Yallow, 1979; Brown, 1981). Erythropoietin, a glycoprotein hormone that regulates red blood cell (HBC) formation in mammals, is produced mainly by as yet unidentified cells in the kidney in response to anemia or hypoxia (Hainmond and Winnick, 1974; Goldwasser, 1975; Miyake et al., 1977). Physiologically inappropriate erythropoietin production is often associated with the presence of a variety of tumors in humans (Haminond and Winnick, 1974) as well as in animals (Peterson and Zanjani, 1981; Wardrop et al., 1982; Nelson et al., 1983). Several mechanisms may be involved in inappropriate erythropoietin production in cancer patients. While some renal tumors secrete erythropoietin (Sytowski et al., 1983), it has been suggested that the growth of nonerythropoietin-secreting tumors in the kidney may cause hypoxia thus inducing an increase in erythropoietin production by normal kidney cells (Hammond and Winnick, 1974). It has also been shown that erythropoietin can be extracted from nonrenal tumors such as hepatomas, cerebellar hemangiosarcoinas, and uterine fibroadenomas (Hammond and Winnick, 1974). These findings suggest that nonrenal tumors can be involved in ecto-

CANINE TRANSMISSIBLE VENEREAL TUMOR

83

pic erythropoietin production. Such inappropriate erythropoietin release may lead to the development of paraneoplastic polycythemia (erythrocytosis), an increase in RBC formation, which is manifested as an increase in RBC count, hematocrit, hemoglobin concentration, and red cell mass. We have recently investigated the effect of TVT growth on RBC production by the host and found that dogs with a large tumor burden develop polycythemia (Cohen, Jeglum, Goldwasser, and Herberman, unpublished data). In this study, beagle dogs were transplanted with TVT at six different subcutaneous sites and tumor volume, RBC counts, serum erythropoietin levels, and blood chemistry profiles were monitored before and at intervals after tumor transplantation. We detected a statistically significant positive correlation between RBC concentration and total tumor volume at 28, 56, and 71 days after tumor transplantation. During this observation period, RBC counts did not exceed 8.5 x lo6 RBC/pl, the upper normal value for the dog (Schalm et al., 1975). Also at 112 days after transplantation, RBC concentration was positively correlated with total tumor burden and in some dogs with tumor burden in excess of 300 cm3, RBC counts were higher than 9 X 106/p1. While hematocrit and hemoglobin concentration were also elevated in animals with a large tumor burden, the values of mean corpuscular volume, mean corpuscular hemoglobin, and mean corpuscular hemoglobin concentration were not correlated with total tumor volume. To establish whether the high RBC counts reflected an increase in RBC production and were not a result of hemoconcentration, we determined the blood volume and the red cell mass of the TVT-bearing dogs. We found that the blood volume as well as the red cell mass were increased in animals with large tumor burdens. The analysis of bone marrow biopsies from these dogs revealed an increase in erythroid precursors. These findings demonstrate that erythropoiesis is increased in dogs with a large TVT burden. In addition, we observed that while the blood urea nitrogen (BUN) levels in all TVT-bearing dogs were within the normal range, animals with a large tumor burden had decreased BUN levels. Polydipsia and polyuria are often observed in canine polycythemia cases and presumably represent compensatory mechanisms to decrease blood viscosity (MacGrath, 1974). Since food intake of dogs with large tumor burdens did not seem to be reduced and liver function was not altered, it seemed possible that the low BUN levels are a result of increased glomerular filtration rate due to polydipsia. Indeed, we could demonstrate that dogs with a large tumor burden consumed more water and excreted more urine than animals with a smaller tumor burden. We also determined serum erythropoietin levels of the TVT-bearing dogs using a radioimmunoassay developed for the detection of human erythropoietin (Shenvood

84

D. COHEN

and Goldwasser, 1979; Koeffler and Goldwasser, 1981). Overall, erythropoietin levels were increased in most dogs following TVT transplantation although they were not significantly correlated with total tumor volume. The increase in serum erythropoietin levels in TVT-bearing dogs could be a result of erythropoietin synthesis by the TVT or of increased erythropoietin secretion by the host as a response to TVT-generated stimuli. Postmortem examination of the polycythemic dogs did not reveal metastatic spread of the TVT to the kidney that could lead to kideny hypoxia thus stimulating erythropoietin synthesis by normal cells. On the other hand, we found that TVT extracts contain high levels of erythropoietin. It seems likely that TVT cells synthesize erythropoietin and could, therefore, be used for the study of erythropoietin synthesis in oitro. The tumor could also be utilized as an animal model for investigations of paraneoplastic polycythemia. V. Experimental Transplantation

The TVT is the only canine tumor that is transplantable to adult immunocompetent allogeneic dogs. Tumor transplantation is usually carried out by subcutaneous inoculation of enzymatically disaggregated tumor cells (Cohen, 1972; Epstein and Bennett, 1974). The minimal number of TVT cells required for successful transplantation of the tumor has not been determined. In most studies, about 1 x lo8 viable tumor cells were transplanted (Cohen, 1972; Yang and Jones, 1973; Epstein and Sarpel, 1980). The proportion of adult, nonimmunosuppressed dogs in which the tumor develops, and the course of the experimentally induced disease, may depend on the size of the inoculuin and the number of sites injected with TVT cells. Karlson and Mann (1952)transplanted the TVT through 40 passages to a total of546 dogs. Each animal was transplanted with TVT tissue fragments at two different subcutaneous sites. Overall, 68% of the animals developed at least one tumor. While the proportion of “takes” was generally between 50 and go%, the highest rate of takes was 95% and the lowest 30%.The sex of the animals did not seem to affect the rate of tumor takes. Recently, we inoculated 25 male beagle dogs with 1 X los TVT celldsite, at six subcutaneous sites (Cohen, Jeglum, Goldwasser, and Herberman, unpublished data). Of the 25 animals, 22 developed six tumor nodules, 1 animal developed five tumor nodules, and 2 dogs developed four tumor nodules. In another study, 89%of adult dogs that were inoculated at one site with 1 x lo8 cells developed tumors (Yang and Jones, 1973). Using an isotope labeling technique, Holmes (1981a) estimated that only about 13%of the injected TVT cells survive in the host to develop a tumor. Following subcutaneous tumor transplantation, a palpable tumor nodule can be detected in most animals within 2 to 3 weeks (Karlson and Mann, 1952).

C A N I N E TRANSMISSIBLE VENEREAL TUMOR

85

In the first few weeks after transplantation, the volume doubling time of the TVT is between 4 and 7 days, with an estimated cell loss factor of less than 50% (Cohen and Steel, 1972;Holmes, 198lb).The subsequent growth rate of the tumor is substantially decreased and is characterized by a volume doubling time of more than 20 days and an estimated cell loss factor that exceeds 80 or 90% (Cohen and Steel, 1972). The clinical course of the experimentally induced disease ranges from spontaneous regression to metastatic spread. Karlson and Mann (1952)followed the course of tumor growth and regression in 328 dogs. They observed spontaneous tumor regression in 42% of the animals within 60 days, and in another 12% of the animals the tumor had regressed within 80 days after first tumor detection. Overall, in 87% of the dogs, the TVT had regressed within 6 months after first tumor detection. Tumor recurrence was not observed in dogs in which the TVT had regressed spontaneously. Epstein and Sarpel (1980)observed spontaneous regression in 40% of 54 dogs within 6 months after transplantation. In beagle dogs transplanted with the TVT at six different sites, we observed spontaneous regression in 4 of 25 animals within 110 days after transplantation (Cohen, Jeglum, Goldwasser, and Herberman, unpublished data). Spontaneous regression of the TVT can occur in animals with a total tumor burden of 100 cm3 or more and the tumor often disappears completely within 2 to 3 weeks. Metastatic spread of the TVT was observed by Karlson and Mann (1952)in 5 of 328 dogs that developed the tumor and were followed for the whole duration of the study. Epstein and Sarpel(l980)reported that in their studies 10-20% of the animals developed metastatic disease. In addition to subcutaneous tumor transplantation, TVT cell suspensions have also been inoculated intravenously and intraperitoneally. DeMonbreun and Goodpasture (1934)inoculated 2 dogs intravenously with an unspecified number of TVT cells. In one of these animals, numerous cutaneous tumor nodules were detected 83 days after transplantation. Autopsy was performed 30 days after tumor detection and revealed in addition to cutaneous nodules, numerous tumor nodules in the subcutaneous fat and skeletal muscles. Tumor nodules were also detected in the omentum, thyroid and liver, but not in the heart, the lung, or any other internal organs. A similar distribution of tumor dissemination with cutaneous involvement was observed in the second dog. Recently, we inoculated each of 12 dogs intravenously with 4 x lo9 TVT cells. One animal developed more than 100 cutaneous tumor nodules that were detected 89 days after transplantation. In each of 3 other dogs, one cutaneous tumor was detected at about 100 days after inoculation. Cutaneous tumors also developed at about 100 days after inoculation in a dog that was injected intraperitoneally with 1.9 x lo9 TVT cells. These cutaneous tumors developed at sites that were remote from the inoculation site and were, therefore, a result of tumor dissemination. Autopsy of this dog did

86

D. COHEN

not reveal tumor development in other tissues. Four other dogs that were inoculated intraperitoneally with a similar number of tumor cells did not develop any tumor (Cohen and Krup, unpublished data). DoMonbreun and Goodpasture (1934) observed tumor growth in the abdominal cavity following intraperitoneal inoculation. Cutaneous tumors appeared at the site of inoculation but may have been the result of inadvertent subcutaneous tumor transplantation. The development of cutaneous tumors following intravenous and intraperitoneal tumor inoculation is a highly unusual tumor dissemination pattern and may be related to the histogenesis of the TVT. Heterotransplantation of the TVT has also been attempted. The tumor can be transplanted to fox (Sticker, 1906; Wade, 1908), jackal (Samso, 1966), and coyote (Cockrill and Beasley, 1979). The TVT also grows in nude mice (Oughton and Owen, 1974; Holmes, 1981a), in the cheek pouch of cortisonetreated hamsters (Samso, 1966), and temporarily in X-irradiated mice (Stubbs and Furth, 1934). The TVT that develops in nude mice can be transplanted to dogs (Holmes, 1981a). Transplantations of the TVT into rats, mice, chicken, monkeys, guinea pigs, rabbits, cats, and opossums did not result in tumor growth (Sticker, 1905; DeMonbreun and Goodpasture, 1934; Samso, 1966; Cockrille and Beasley, 1979). VI. The Immune Response against the TVT

The frequent spontaneous regressions of the experimentally transplanted TVT suggested to investigators who studied this neoplasm at the turn of the century that the immune response against the tumor plays a major role in determining the course of the disease. Sticker (1906) already demonstrated that spontaneous regression of the TVT is followed by transplantation immunity. This finding was later confirmed by DeMonbreum and Goodpasture (1934)and Powers (1968). Wade (1908), who studied the histology of regressing TVT, reported, “The final disappearance of the tumor is associated with a flooding oT the tumor tissue with lymphocytes, polyblasts, and plasma cells. The tumor is borne away on a lymphocyte tide which may leave a few of its number as sessile adventitia cells.” Wade (1908) concluded “that the tumors disappear by gradual cytolysis associated with the presence of, and probably caused by, lymphocytes and polyblasts.”

A. I N VIVOSTUDIES More recent in vivo experiments extend the observations made by Sticker (1906) and Wade (1908) and prove that immunological mechanisms inhibit tumor growth and induce the spontaneous regressions of the TVT. Passive transfer of postregression sera to TVT-bearing dogs was shown to inhibit

C A N I N E TRANSMISSIBLE VENEREAL TUMOR

87

tumor growth and prevent tumor development if administered simultaneously with tumor transplantation (Powers, 1968). Moreover, the TVT is a highly malignant neoplastic disease in immunosuppressed animals as well as in newborn puppies. In most dogs that were subjected to 200 rads wholebody X-irradiation prior to tumor transplantation, the TVT grew progressively and metastasized. The course of the disease in the untreated animals was more benign and irregular and ranged from growth that was followed by regression in some animals, to metastatic spread in one of seven dogs (Cohen, 1973). Transplantation of the TVT to newborn puppies also resulted in progressive tumor growth and metastatic spread in most animals. The growth of the tumor was inhibited in puppies born to dams that were immunized with TVT before or during pregnancy (Yang and Jones, 1973). These experiments indicate that the TVT is potentially a highly malignant disease, but immunological mechanisms in adult immunocompetent animals limit the growth and spread of the tumor. Since the TVT is probably an allograft, the degree of incompatibility between the canine major histocompatibility complex (DL-A) antigens of the tumor and those of the host may determine the intensity of the immune response and consequently the course of the disease. The growth pattern of the TVT (early regression or prolonged persistance) would therefore be expected to be similar in littermates that are identical for DL-A antigens. Bennett et al. (1975a) transplanted the TVT to littermates of four canine families that were typed for DL-A antigens. They showed that sib pairs with identical DL-A haplotypes had concordant TVT growth patterns, whereas, sib pairs that differed by two haplotypes had discordant growth patterns. Pairs that differed by only one haplotype had both concordant and discordant tumor growth patterns. Whether these results are interpreted as a similar recognition pattern of TVT DL-A antigens by DL-A identical littermates and/or as regulation of the intensity of the immune response against TVT antigens by immune response genes of the host, the segregation of TVT growth patterns with DL-A segregation is compatible with the hypothesis that immunological mechanisms determine the course of the disease. Further evidence that the immune response is involved in inducing TVT regression is derived from immunotherapy studies with bacillus CalmetteGuCrin (BCG). Pairs of littermates, identical for DL-A antigens were transplanted with the TVT in two different subcutaneous sites, and BCG was inoculated into one of the tumor nodules of each pair. The BCG-injected tumor nodules, as well as the noninjected tumors in the same hosts, regressed within 63 days as compared to tumor persistance beyond 100 days in the controls (Hess et al., 1977). The results of these in uivo studies clearly demonstrate that the clinical course of the TVT is determined by the immune response toward the tumor.

88

D. COHEN

B. THE ANTIGENSOF THE TVT The immunologically mediated rejection of established TVT suggests that membrane-associated transplantation antigens are expressed on the surface of TVT cells. Although such antigens have not yet been isolated and characterized, serological studies indicated that at least part of the anti-TVT antibody response is directed against major histocompatibility complex (MHC) gene products. Anti-TVT sera react with normal dog lymphocytes (Cohen, 1974) and the pattern of reactivity of some postregression sera with lymphocytes from 105 normal dogs suggests that the antibodies are directed against DL-A specificities (Epstein and Bennett, 1974). Moreover, the results of tissue typing of TVT cells with a panel of antisera that are used for canine DL-A typing studies indicate that DL-A antigens are expressed on the surface of the tumor cells (Epstein and Bennett, 1974). We have recently demonstrated that p,-microglobulin (P2-m) is not expressed on the surface of the tumor cells (Cohen et al., 1984). Since P2-m expression is required for the expression of class I MHC (MHC-I) antigens on the cell surface, these results suggest that MHC-I antigens are not expressed on the surface of TVT cells. The anti-DL-A typing sera used by Epstein and Bennett (1974) may, therefore, have reacted with class I1 MHC antigens. The expression of class I1 MHC antigens would be compatible with the reticuloendothelial origin of the tumor as well as with results of mixed lymphocyte tumor culture (MLTC) studies (Hess et al., 1975). The expression of DL-A antigens on the surface of TVT cells is discussed in more detail in Section VII,B, and the results of MLTC studies are described in Section V1,C. In addition, an antigen that does not seem to be associated with the DL-A system can be detected in 3 M KCI and saline extracts of the TVT (Palker and Yang, 1981). A rabbit antiserum was raised against such extracts and extensively absorbed with cross-linked normal dog serum proteins as well as extracts of normal canine tissues. In an enzyme-linked immunosorbent assay, reactivity of this antiserum with the TVT extracts was significant and specific as compared to minimal reactivity with similarly prepared extracts of normal canine tissues and other canine tumors. The antigenic activity of the TVT extract was detected in samples with estimated molecular weight greater than 70,000 and was greatly reduced by trypsin digestion, exposure to heat, as well as pH 2.8 and 11.0. In immunofluorescence studies, this antigen can be detected by use of the rabbit antiserum in the cytoplasm and nucleolus of acetone-fixed TVT cells. The results of leukocyte adherence inhibition (LAI) studies suggest that this TVT antigen is immunogenic in the dog (Harding and Yang, 1981a,b). It has not yet been established, however,

CANINE TRANSMISSIBLE VENEREAL TUMOR

89

whether the immune response toward this antigen plays a role in the induction of the spontaneous regression of the TVT. C. I N VITROSTUDIES Although DL-A antigens are expressed on TVT cells, and the tumor is immunogenic in the dog, in most cases the TVT is not rejected several days after transplantation but grows progressively for several weeks or months and then often regresses spontaneously. Numerous in vitro studies have been carried out in an attempt to elucidate the immune mechanisms that are involved in the anti-TVT response during the various phases of the disease. The recognition of TVT antigens by canine peripheral blood lymphocytes (PBL) has been investigated by use of the mixed lymphocyte tumor cell culture (MLTC). Hess et al. (1975)demonstrated that DNA synthesis of PBL from both normal and regressor dogs was equally stimulated by TVT cells, whereas it was substantially suppressed in tumor-bearing animals. The response of PBL from normal, tumor-bearing, and regressor dogs to PHA, was, however, essentially identical, suggesting that the suppression MLTC response in tumor-bearing animals was not due to nonspecific suppression of T lymphocyte functions. To exclude the possibility that host lymphocytes contaminating the TVT suspensions induce the MLTC response, one member of a DL-A-identical sibling pair was transplanted with the tumor. Peripheral blood lymphocytes from the tumor-carrying dog did not stimulate the lymphocytes of the other member of the DL-A-identical sibling pair in the MLTC assay, whereas the tumor induced significant stimulation. Furthermore, the magnitude of the M LTC response during the progressive phase of tumor growth correlated with the eventual course of the disease. High blastogenic responses were detected in dogs in which the tumor eventually regressed, whereas low or no responses were observed in dogs in which the TVT eventually metastasized. In addition, the already suppressed MLTC response of lymphocytes from dogs carrying progressively growing tumors could be further suppressed by the presence of autologous serum. Such sera, however, did not influence the response of the same lymphocytes to PHA. Hess et al. (1975) proposed that the stimulation of normal dog lymphocytes by TVT cells which is comparable in intensity to the response in allogeneic mixed lymphocyte culture, indicates that class I1 MHC antigens are expressed on the surface of the tumor cells. Also, the similar intensity of the MLTC response in normal and regressor dogs is taken as indirect evidence that it is primarily directed against MHC gene products. Hess et al. (1975) also proposed that the suppressed response of lymphocytes in tumor-bearing animals may be one of the mechanisms that enable tumor

90

D. C O H E N

growth in the initial stages of the disease. Zander et al. (1980) investigated, therefore, the effect of plasmapheresis on the growth of established TVT and correlated the effect on the tumor with the suppressive effect of the serum of the animal in the MLTC assay. Six pairs of DL-A-identical dogs were transplanted with the TVT and one member of each pair was treated by extensive plasmapheresis during the progressive phase of tumor growth. Plasmapheresis did not have a statistically significant effect on tumor growth but in two of five animals, the growth rate was lower. Sera removed from the tumor-bearing dogs before plasmapheresis suppressed the MLTC response, whereas sera removed from animals after plasmapheresis were significantly less suppressive. Zander et al. (1980) proposed several possible reasons for the failure of plasmapheresis to induce early regression in spite of reduced suppressive effect by the sera in MLTC. The effect of plasmapheresis could be short-lived, plasmapheresis may remove factors that suppress the MLTC response as well as tumor inhibitory factors, and finally, the suppressive effect of serum factors on the MLTC response may not be the essential mechanism that acts in uiuo and enables tumor growth. The MLTC response has also been studied in TVT-bearing dogs that were treated with BCG (Hess et al., 1977). Intralesional BCG administration enhanced tumor regression and the MLTC response increased significantly in these dogs during the period of tumor regression. Thus, the MLTC studies indicate that canine lymphocytes respond to class I1 MHC antigens expressed on the surface of TVT cells. The nature of the responding cells has not yet been determined. From the results of mixed lymphocyte culture studies in mouse and human, it seems likely that the MLTC response against the TVT represents recognition of TVT MHC antigens by helper T lymphocytes (Klein, 1982). In uitro methods have also been used to characterize immunological effector mechanisms against TVT antigens. The detection and characterization of the antibody response was the subject of most of the earlier studies. Antibodies to TVT homogenates have first been detected by use of the tanned red cell hemagglutination technique (Dozza and Torlone, 1960a), the conglutinin-complement adsorption test (McKenna and Prier, 1966), and the Ochterloney technique (Powers, 1968). These initial studies were followed by the demonstration of antibodies against TVT-cell membrane-associated antigens using the indirect membrane immunofluorescence test. These predominantly IgG antibodies could be detected in the serum of most tumor-bearing dogs at about 40 days after transplantation. Simultaneously with the detection of antibody in the serum, in uiuo coating of TVT cells with antibody could also be demonstrated. Moreover, these studies showed that sera from clinical TVT cases in Ireland react with TVT cells that originated in Malaya (Cohen, 1972). The anti-TVT antibodies could be detected during

C A N I N E TRANSMISSIBLE VENEREAL TUMOR

91

progressive tumor growth as well as during and following tumor regression. In addition, antibodies were detected in the serum of an animal in which the tumor did not grow, in an animal with metastatic TVT, as well as in animals that were irradiated prior to tumor transplantation (Cohen, 1973). The detection of anti-TVT antibody in the serum of tumor-bearing dogs by use of indirect membrane immunofluorescence was confirmed by Epstein and Bennett (1974) who also demonstrated that IgG antibody can be eluted at low pH from in uiuo antibody-coated tumor cells as soon as 7 days after transplantation. Furthermore, sera from dogs in which the tumor had regressed were shown to be cytotoxic to TVT cells in the presence of complement (Bennett et al., 1975b). The effect of anti-TVT sera on tumor cell colony formation in semisolid agar was also investigated (Bennett et al., 1975b). Sera from dogs in which the tumor was regressing (regressors) inhibited tumor cell colony formation in the absence of complement, whereas sera from dogs in which the tumor was growing progressively (progressors) or normal dog serum did not. The inhibitory effect of the regressor sera could be blocked by preincubation of the TVT cells with progressor sera or sera from metastatic TVT cases. Beschorner et al. (1979) identified both the inhibitory factors in regressor sera as well as the blocking factors in progressor sera, as IgG,,. Low pH eluates from progressively growing tumors and from regressing TVT were also identified as IgG,, and exhibited inhibitory and blocking activity that correlated with the clinical course of the disease. Attempts to demonstrate TVT antigens or immune complexes in the blocking IgG fraction were unsuccessful. It seems of interest to establish the mechanisms involved in the inhibitory and blocking activity of the anti-TVT sera and to determine the differences between the IgG,, antibody with blocking activity and the IgG,, antibody with the inhibitory activity. The leukocyte adherence inhibition (LAI) assay has also been used to detect immune reactivity against the TVT. In this assay, blood leukocytes or cells from the spleen or lymph nodes are placed in glass tubes and incubated in the presence of the test antigen. The nonadherent cells are then removed and counted and the results expressed as nonadherent index (Harding and Yang, 1981a). The 3 M KCI extract of the TVT that has been described by Palker and Yang (1981), was used as the test antigen and a similarly prepared extract of normal canine tissue was used as a control antigen. Harding and Yang (1981a) first determined the LA1 reactivity of blood leukocytes, spleen cells, and cell from draining and nondraining lymph nodes at different stages of tumor growth and regression. They found that the LA1 reactivity of peripheral blood leukocytes from dogs in which the tumor was regressing was significantly increased as compared to the reactivity of luekocytes form normal dogs and animals with progressive tumor growth. The LA1 reactivity of

92

D. COHEN

leukocytes from dogs in which the tumor was stationary was intermediate. The pattern of LA1 reactivity of draining and nondraining lymph node cells as well as spleen cells was essentially the same as that of the peripheral blood leukocytes. The sequential changes in LA1 reactivity of blood leukocytes were also investigated during the course of the disease (Harding and Yang, 1981b). In most animals, an initial increase in LA1 reactivity was observed 10 to 15 days after tumor transplantation. This increase lasted for another 10 to 15 days and-was followed by a rapid decline during the progressive phase of growth. Reactivity remained at a low level in animals in which the tumor did not regress. In contrast, it increased significantly during and following tumor regression. Within 3 to 7 days of surgical removal of progressively growing tumors, an increase in LA1 reactivity that lasted for 4-6 weeks was observed. Furthermore, challenge of dogs in which the tumor was removed surgically with TVT cells was followed within 24 hr by a substantial increase in LA1 reactivity. Although the LA1 assay has been used extensively in tumor immunology studies (reviewed by Harding and Yang, 1981a), the mechanism involved in adherence inhibition of different types of leukocytes to glass is poorly understood. It has been proposed that the responding cells are glass-adherent, Fc receptor-bearing cells that are coated with cytophilic antibody. The addition of antigen that interacts with the cytophilic antibody is thought to prevent the adherence of the cells to glass (Grosser and Thomson, 1976; Marti et al., 1976). The kinetics of the LA1 reactivity against the TVT antigen (Harding and Yang, 1981b) and the similar reactivity of blood leukocytes and cells from lymphoid organs (Harding and Yang, 1981a) are indeed compatible with the view that this assay reflects an antibody response. Since the antigen that was used in these studies is localized primarily in the nucleolus and cytoplasm of TVT cells (Palker and Yang, 1981), the LA1 reactivity may reflect changes in TVT growth, rather than an effector mechanism that determines the course of the disease. Histological studies demonstrated that regressing TVT nodules are infiltrated by lymphocytes, plasma cells, and macrophages (Wade, 1908; DeMonbreun and Goodpasture, 1934; Bloom et al., 1951; Yang et al, 1976b). Moreover, animal-derived TVT cell suspensions contain up to 30% Fc receptor-bearing cells, most of which are probably lymphocytes and macrophages (Cohen et al., 1971). Chandler and Yang (1981) separated host leukocytes from TVT cells by use of a gradient and showed that about 20% of cells derived from regressing TVT nodules were lymphocytes, monocytes, neutrophils, and eosinophils as compared to only about 5% of the cells from progressively growing tumors. Of the 20% infiltrating cells in regressing tumors, 17% were lymphocytes, and less than 2% were monocytes. Using a heterologous rabbit anticanine thymocyte serum and an anticanine IgG re-

CANINE TRANSMISSIBLE VENEREAL TUMOR

93

agent in the membrane immunofluorescence assay, Chandler and Yang (1981) characterized the tumor derived lymphocytes. In regressing TVT, about 60% of the lymphocytes were T cells, 26% were €3 cells, and about 14% were null cells. In contrast, in progressively growing tumors only 34% of the cells were T cells, 37% were B cells, and about 26% were null cells. Infiltration of the TVT with lymphocytes and macrophages and in uiuo coating of the tumor cells with antibody (Cohen, 1972; Epstein and Bennett, 1974; Bennett et al., 1975b; Beschorner et al., 1979) may provide the conditions for antibody-dependent cellular cytotoxicity (ADCC) to occur in v i m . The ADCC of blood leukocytes against TVT target cells was investigated, using a 51Cr release assay (Cohen, 1980b). The target cells were treated either with normal dog serum or postregression sera and the effector cells were prepared by separation of heparin-treated blood on commercially available Ficoll-Hypaque gradients (density: 1.077 g/ml). While in humans, interface cells obtained by this method are hardly contaminated by polymorphonuclear (PMN) cells, up to 25% of canine interface cells are PMN cells (Largiader et aE., 1972; Kuramochi, 1974; Ho and Babiuk, 1978). Leukocytes from normal dogs as well as leukocytes from tumor-bearing animals and dogs in which the tumor had regressed were cytotoxic to antibodycoated TVT target cells. Moreover, we could demonstrate (Cohen et al., 1982) that purified lymphocytes, monocytes, as well as PMN cells from normal dogs were cytotoxic to antibody-coated TVT target cells. In addition, leukocytes from dogs in which the TVT had regressed were cytotoxic to TVT target cells that have not been coated with antibody, whereas leukocytes from normal dogs were not (Cohen, 1980b). Overall, the immunological studies of the TVT clearly demonstrate that the tumor is antigenic in the dog and that the immune response against the tumor plays a major role in determining the course of the disease. The immune response is directed, at least in part, against DL-A antigens, and antibody against membrane antigens can be detected. The results of MLTC, LAI, and colony inhibition studies suggest that suppressive serum factors may facilitate tumor growth in the initial stages of the disease. The demonstration that passive transfer of immune serum can induce tumor regression indicates that antibody-dependent mechanisms are involved in the induction of tumor regression. Antibodies from regressor animals are cytotoxic to TVT cells in the presence of complement and may therefore destroy the tumor cells in uiuo without the involvement of cellular effector mechanisms. However, the infiltration of regressing TVT by Fc receptor-bearing lymphoid cells and the in uiuo coating of TVT cells with antibody may provide the conditions for ADCC to occur. Leukocytes from normal and tumor-bearing animals have indeed been shown to mediate ADCC against postregression serum-coated TVT target cells. Whether the cytotoxic activity of leukocytes

94

D.

COHEN

from immune animals against TVT cells that have not been coated with antibody represents direct cytotoxicity, or is in fact ADCC due to cytophilic antibody that attaches to Fc-bearing effector cells, has not been determined. VII. Etiology and the Mechanism of Transmission

The facility with which the TVT can be transmitted to outbred dogs, either by coitus or by the inoculation of tumor cells, suggested that the tumor is transmitted and induced by an infectious agent (see Gross, 1983). Several histological studies were carried out at the beginning of this century to determine whether a TVT induced by inoculation of tumor cells is derived from host cells or from the transplanted tissue (Bashford et al., 1905; Beebe, 1907; Wade, 1908). These studies gave conflicting results. In a more recent study, [3H]thymidine-labeled TVT cells were inoculated and the developing tumors were analyzed by autoradiography. The results of this study suggest that the tumors developed by multiplication of the transplanted cells (Kudo et al., 1974). The demonstration of an oncogenic viral agent that transmits and induces the tumor was attempted by the inoculation of cell-free extracts of the TVT. Some investigators claimed success in transmitting the TVT with cell-free material (Ajello, 1960; Dozza and Torlone, 1960b; Ajello and Gimbo, 1965a), but the methods employed in these studies raise doubts regarding the validity of these experiments. DeMonbreun and Goodpasture (1934) failed to transmit the disease by the inoculation of lyophilized or hypotonically disrupted TVT cells. Samso (1966) and Cabanie et al. (1973) have also attempted to transmit the disease with cell-free extracts of the TVT but have failed to do so. The search for an oncogenic virus by electron microscopic studies was also inconclusive. While virus-like particles were observed in some studies (Ajello, 1961; Ajello and Gimbo, 1965b,c; Lombard et al., 1967; Sapp and Adams, 1970; Battistacci and Moriconi, 1974; Kennedy et al., 1977), such particles could not be detected in other studies (Carteaud et al., 1963; Murray et al., 1969; Cockrill and Beasley, 1975). To date, no reproducible evidence is available to indicate that an oncogenic virus is involved in the induction or the transmission of the TVT. Experimentally, the tumor can only be transmitted by inoculation of living tumor cells. Moreover, since most of the available evidence suggests that the TVT is transmitted by the transplantation of cells, there is no compelling reason to postulate a viral etiology for the disease.

A. CYTOGENETIC AND IMMUNOLOGICAL STUDIESSUGGESTING CELLULAR TRANSMISSION Cytogenetic studies of the TVT suggest that the tumor is transmitted by the transplantation of cells. The normal diploid chromosome number of the

CANINE TRANSMISSIBLE VENEREAL TUMOR

95

dog is 78; all autosomes are acrocentric, the X chromosome is a large submetacentric a i d the Y is a small metacentric (Oshimura et al., 1873). The first detailed studies of the karyotype of the TVT have been c a d out by Takayama and Makino (1961) and Makino (1963), who demonstrated that TVT from difierent locations in Japan are characterized by a modal chromosome number of 59. Of these 59 chromosomes, 16 were metacentric and 43 were acrocentric. Weber et al. (1965) observed a modal number of58 chromosomes with 15 metacentrics or submetracentrics and 44 acmcentrics in tumors in the United States. Barski and Cornefert-Jensen (1966)h n d that the modal chromosome number of three different TVT cases in France was between 57 and 59 with a modal number of 17 metacentrics. Essentially the same karyotype was found in Jamaica (Thorburn et al., 1968), the USSR (Kakpakova d a / ., 1968),Uganda (Wright et al., 1970), and Nigeria (Idowo, 1977). The reduction in the number of chromosomes of the TVT as compared to the number of chromosomes in normal dog cells is not accompanied by a reduction in the number of chromosomal arms or the DNA content (Barski and Cornefert-Jensen, 1966; Wright et al., 1970). It has been proposed that the increased number of metacentric chromosomes in the TVT arose from multiple centric fusions (Barski and Cornefert-Jensen, 1966). Banding pattern analysis of normal dog and TVT chromosomes revealed extensive structural rearrangcmrtts in the TVT. Since several large biarmed and acrocentric chromosomes were identical in different tumors, it is to be assumed that the tumors had a common origin (Oshimura et al., 1973). The karyotype of the TVT was also studied in TVT cells grown in tissue culture. Prier (1966) showed that fresh TVT cells and tumor cells grown in tissue culture for 24 hr had the karyotype that is characteristic of the TVT, whereas TVT cells after 72 or 192 hr in culture had the normal dog karyotype. Samso (19666) also showed that the karyotype of TVT cells propagated in tissue culture was that of normal dog cells. The most likely interpretation of these results is that the TVT cells stopped dividing or died and the cells that were grown in tissue culture were actually normal dog cells. Adams et at. (1968, 1981) succeeded in growing TVT cells in tissue culture for prolonged periods and showed that the karyotype of the cells grown in tissue culture was indeed the TVT karyotype. Inoculation of these cells into the anterior chamber of the eye of a dog resulted in tumor growth (Adams et al., 1968), but tumors did not develop following parenteral inoculation of these cell lines into newborn or adult dogs (Adams et al., 1981). Thus, the cytogenetic studies demonstrate that the TVT is characterized by a karyotype that is significantly different from the karyotype of normal dog cells and that the chromosome aberrations of the TVT are almost identical in tumors originating in different parts of the world. Cytogenetic studies of spontaneous murine and human tumors also revealed specific chromosomal aberrations, however, these are usually not as extensive as they are

96

D. COHEN

in the TVT. Trisomy 15, for example, is often found in inurine T-cell leukemia and specific translocations occur in murine plasmacytomas and Burkitt’s lymphoma. There is evidence to suggest that these chromosomal aberrations are involved in the expression of malignancy of these neoplasms (reviewed by Klein, 1981; Miller and Miller, 1983). Little is known about chromosomal aberrations in canine tumors other than the TVT. Chromosome analysis of a canine fibrosarcoma showed that the tumor had 54-56 chromosomes including 13-24 biarmed chromosomes (Sonoda et al., 1970). Cytogenetic studies of 10 canine lyinphosarcomas revealed irregular chromosomal aberrations and the modal number was between 77 and 81 (Basur and Gilman, 1966; Benjamin and Noronha, 1967). Miles et al. (1970) also analyzed the chromosomes of canine lymphosarcomas. Of 12 cases that they studied, 10 had the normal dog karyotype, 1 tumor had a modal number of 69 or 70 with 3 to 9 metacentric or submetracentric chromosomes, and 1 canine lymphosarcoma cell line had a modal number of 44 chromosomes with 26 metacentric or submetacentric chromosomes. In addition, a canine thyroid adenocarcinoma cell line was found to have a modal chromosome number of 60 with 5 metacentric chromosomes. A canine melanoma cell line that was analyzed in the same study had a hyperploid chromosome pattern with 20 to 34 metacentric chromosomes (Pakes et al., 1965). Moreover, a radiation-induced canine osteosarcoma was characterized by a hypodiploid stem line that contained 45-55 chromosomes with 10-15 abnormal metacentric and submetacentric chromosomes and a hyperdiploid stem line that contained 90-105 chromosomes with 20-30 abnormal metacentric and submetacentric chromosomes (Taylor et al., 1975). Miles et al. (1970) and Oshimura et al. (1973) suggested that in animals that possess predominantly telocentric or acrocentric chromosomes, there is a tendency for centric fusions to occur. Although the chromosomal changes that have been observed in various canine neoplasia bear some resemblance to the karyotype of the TVT, they are not as consistent and they are not characteristic of the tumor. It seems unlikely that the chromosomal aberrations of the TVT could be induced repeatedly at every transmission by an oncogenic virus. The cytogenetic characteristics of this tumor rather indicate that the TVT in different geographical locations developed from a common origin and has since been transmitted by cell transplantation. In addition to the cytogenetic evidence, immunological studies also suggest that the TVT is transmitted by the transplantation of cells. In one of these studies, it was assumed that if the TVT develops in an inoculated dog by malignant transformation of the host cells, the DL-A antigens of the host should be expressed on the surface of the tumor cells. Using isoimmune sera directed against host tissue antigens, it was shown that the alloantigens of the host cannot be detected on the surface of TVT cells (Cohen, 1974). The

C A N I N E TRANSMISSIBLE VENEREAL TUMOR

97

rationale for these experiments seemed to be correct at the time (Klein and Klein, 1958; Klein and Moller, 1963). However, in view of recent evidence that spontaneous tumors often do not express MHC-I antigens (Section VII,B), the results of this study cannot be regarded any more as evidence for the cellular transmission of the TVT. In the same study, it was also shown that anti-TVT sera react in some cases with normal dog lymphocytes. Epstein and Bennett (1974) showed that some sera from dogs in which the TVT had regressed could be used to identify DL-A specificities on normal canine lymphocytes. In the same study, two TVT lines originating from Malaya and Chicago were typed with a panel of 64 canine lymphocyte cytotoxic antisera that were used for canine histocompatibility typing studies. Both tumor lines were reactive with antisera against the DL-A 3 and 10 specificities and were weakly reactive with DL-A 8 antisera. Identical positive reactions were also obtained in the two lines with unclassified antisera mostly of broad specificity. Overall, the pattern of reactivity with 62 of the 64 antisera was identical in the two tumor lines. These results suggest a high degree of similarity between the two tumor lines for specificities that were recognized by the typing panel. In the same study Epstein and Bennett (1974)also showed that antigenic specificities recognized by the typing sera remained constant in TVT transplanted to 40 different dogs. If the TVT were induced in each host by the malignant transformation of normal host cells, those DL-A antigenic specificities that are expressed on the TVT should be of host origin. The findings that the same DL-A specificities are expressed on the surface of TVT cells in different animals and on tumors from different origins, and that anti-TVT antibodies are directed, at least in part, against DL-A specificities, clearly demonstrate that the tumor is transmitted by cell transplantation. It should be emphasized, however, that although the cytogenetic and immunological studies can hardly be reconciled with viral transmission of the TVT, these studies do not exclude the possibility that the tumor was originally induced by an oncogenic virus. EXPRESSION B. LACKOF &-MICROGLOBULIN ON THE SURFACE OF TVT CELLS Since tumors cannot, in most cases, be transplanted across MHC barriers, the allotransplantability of the TVT raises the question of whether the expression of MHC antigens on the surface of the tumor cells is modified. In the allotransplantable TA3 murine mammary carcinoma, for example, the expression of MHC antigens is masked by the cell surface glycoprotein, epiglycanin (Miller et aZ., 1982).Allotransplantability could also be the result of lack of MHC antigen expression on the surface of tumor cells. Previous serological studies showed that TVT cells express DL-A antigens on the cell

98

D. COHEN

surface and that anti-TVT sera recognize DL-A antigens on normal canine lymphocytes (Epstein and Bennett, 1974). It was not known, however, what classes of MHC antigens are expressed on TVT cells and whether the allotransplantability of this tumor is associated with a modification of MHC antigen expression. In the dog (Van Der Feltz et al., 1981), as in other mammals, class I MHC (MHC-I) molecules consist of two-chain structure: a polymorphic glycoprotein heavy chain of 44,000 MW that carries the alloantigenic sites, and the monomorphic microglo globulin (P2-m) light chain of 12,000 MW. &-Microglobulin is maintained on the cell surface by noncovalent association with the membrane-bound heavy chain (Strominger et al., 1977). The synthesis of P2-m seems to be essential for the expression of the heavy chains on the cell surface. Daudi, a lymphoblastoid cell line derived from a Burkitt’s lymphoma patient (Klein et al., 1967), cannot synthesize P2-m and does not express MHC-I antigens on the cell surface (Nilsson et al., 1974; Raf€, 1975; Ploegh e t al., 1979; Sege et a l . , 1981; Rosa et al., 1983). This cell line does, however, synthesize the MHC-I heavy chains (Ploegh et al., 1979) and somatic cell hybridization studies showed that Daudi MHC-I antigens are reexpressed upon fusion with P,-m-synthesizing human and murine cell lines (Klein et al., 1977; Arce-Gomez et al., 1978). The presence of P2-m on the cell surface can, therefore, be used as a marker for MHC-I antigens expression on the cell surface. In an attempt to define possible modifications in the expression of MHC-I antigens on TVT cells that could account for allotransplantability, we have recently studied the expression of canine Pz-m on the surface of the tumor cells (Cohen et a l . , 1984). Using the immunofluorescence technique, we showed that TVT frozen sections treated with rabbit antidog Pe-m (kindly provided by Dr. M. D. Poulik) and FITC-conjugated goat antirabbit IgG, were essentially negative for P2-m with sporadic cells showing fluorescence. Lymph node sections were stained brightly and uniformly. We could also demonstrate with indirect membrane immunofluorescence that 7 0 4 0 % of collagenase-disaggregatedTVT cells were negative for P2-m, whereas 2030%of the cells were clearly positive. Most of the collagenase-disaggregated lymph node cells were positive for P2-m. Since tumors may be heterogeneous and often consist of subpopulations of cells with widely disparate phenotypes (Trope, 1975; Fidler, 1978; Kerbel, 1979), we tried to determine whether the P,-m-positive cells in TVT cell suspensions are a subpopulation of the tumor cells or infiltrating host cells. Using an immunoabsorption technique, TVT cells reacting with anticanine P2-m serum were separated from cells that do not react with this serum, and cells reacting with autologous anti-TVT sera were separated from cells that do not react with such sera. The use of autologous anti-TVT sera enabled the distinction between

C A N I N E TRANSMISSIBLE VENEREAL T U M O R

99

tumor and infiltrating host cells. Such sera react with autologous TVT cells, but do not react with the normal cells of the animal from which the tumor and the serum are derived (Cohen, 1974). Following the separation of the cell population, the cell fractions were tested for reactivity with anti-p,-m and autologous anti-TVT serum by indirect membrane immunofluorescence and cell radioimmunoassay. The results of these experiments show that cell fractions enriched for p,-m-positive cells were depleted of anti-TVT serum reactive cells, whereas cell fractions depleted of p,-m-positive cells were enriched for anti-TVT serum reactive cells. Enrichment for anti-TVT serumreactive cells resulted in depletion of &-m-positive cells and depletion of anti-TVT serum-reactive cells was associated with enrichment for Pz-mpositive cells. These results suggest that cells reactive with autologous antiTVT serum and cells reactive with anticanine pz-m are two independent cell populations. Thus, most of the cells in animal-derived TVT cell suspensions are positive for TVT markers and negative for P2-m, whereas 20-30% of the cells express pz-m and are negative for TVT markers. We concluded, therefore, that TVT cells do not express &-m on the cell surface, and the pz-m positive cells in the cell suspensions are probably infiltrating host cells. A previous study has indeed shown that infiltrating host lymphocytes and macrophages comprise about 20% of the animal-derived TVT cell suspension (Chandler and Yang, 1981). Cytoplasmic &-m could not be detected in the immunofluorescence studies of frozen TVT sections. Our studies suggest, therefore, that lack of Pz-m expression on the surface of TVT cells is not a result of masking of MHC antigens by other cell surface molecules as seems to be the case in the allotransplantable TA3 murine mammary carcinoma (Miller et d., 1982). Since MHC-I heavy chains cannot be expressed on the cell surface without the expression of P2-m, the lack of Pz-m expression on the surface of TVT cells implies that MHC-I antigens are not expressed on the surface of TVT cells. Such a modification in the expression of MHC antigens could account for the allotransplantability of this cell-transmitted neoplasm. However, in view of the data demonstrating that TVT cells express DL-A antigens and that anti-TVT sera can be used to identify DL-A antigens on the surface of normal canine lymphocytes (Epstein and Bennett, 1974), it is necessary to determine at the molecular level, the class of DL-A antigens that is expressed on TVT cells and characterize the cell surface antigens that induce the antibody response in the dog. The results of the MLTC studies by Hess et al. (1975), suggest that class I1 MHC antigens are expressed on the surface of TVT csells. Is lack ofpz-m expression on the surface of tumor cells a characteristic of a few exceptional tumors such as Daudi and the TVT, or is it a more general

100

D. COHEN

P

Z

-

M

Tumor Embryonal carcinoma Teratocarcinoma

Gestational choriocarcinoma

Daudi lymphoblastoid cell line

Basal cell papilloma Basal cell carcinoma

Keratoacanthoma, basal cell papillom a Actinic keratosis

Squamous cell carcinoma, melanoma, basal cell carcinoma Actinic keratosis Basal cell carcinoma

Breast carcinoma Well differentiated

TABLE I ~A N D~MHC ~ CLASS ~ ~I ANTIGEN ~ ~ EXPRESSION - ~ ~O N TUMOR ~ ~ CELLS ~ ~ Results

Method

Reference

No H-2 and Pz-m mRNA expression Low levels of HLAA, -8, -C, and Pz-m No Pz-m and no Pz-m RNA

cDNA hybridization

Morello et al. (1982)

RIA

Andrews et al. (1981)

SDS gel electrophoresis, cellfree mRNA translation RIA, SDS gel electrophoresis RIA

Tanaka et al. (1981)

Cell-free mRNA translation; immunoprecipitation, SDS gel electrophoresis Immunofluorescence

Ploegh et al. (1979)

NO HLA-A, -B, -C; reduced Pz-m No surface Pz-m; no Pz-m secretion No Pz-m expressed; HLA-A, -B chains synthesized, not expressed on cell surface Pz-m expression In 15 of 20 tumors, lack of Pz-m expression In 1 of 12 tumors no Pz-m expression; in 1/12 partial expression In 1 of 10 tumors no Pz-m expression; in 7 of 10 partial expression In 19 of 23 tumors no Pz-m expression; in 3 of 23 partial expression In 2 of 14 tumors no Pz-m expression In 14 of 16 tumors no Pz-m expression Microheterogeneity in Pz-m expression

Immunofluorescence

Trowsdale et al. (1980) Nilsson et al. (1974)

Tjernlund and Forsum (1977) Tjernlund and Forsum (1977)

Immunoperoxidase (formalin-fixed sections)

Turbitt and Mackie (1981)

Immunoperoxidase (formalin-fixed sections)

Turbitt and Mackie (1981)

Immunoperoxidase (formalin-fixed sections)

Turbitt and Mackie (1981)

Immunofluorescence

Korthals-Altes et al. (1981) Korthals-Altes et al. (1981)

Immunofluorescence

Immunofluorescence and immunoperoxidase

Weiss et al. (1981), DiPersio et al. (1982)

101

CANINE TRANSMISSIBLE VENEREAL TUMOR

TABLE I (Continued) Tumor

Poorly differentiated

Results

Method

Reference

Lack of &-m expression

Immunofluorescence and immunoperoxidase

Weiss et al. (1981). DiPersio et al. (1982)

183 ng P2-m/mg protein 17.8 ng Pz-m/mg protein In 20 of 53 tumors, lack of P2-m and HLA-A, -8, -C expression HLA-A, -B, -C, and P2-m expression Lack of HLA-A, -B, -C, and &-m expression Lack of P2-m expression P2-m expression

RIA

DiPersio et al. (1980) DiPersio et al. (1980) Bhan and DesMarais (1983)

Colon carcinoma Well differentiated Poorly differentiated Breast carcinoma

Melanoma Nevomelanocytic hyperplasia, nevocellular nevi Malignant porocarcinoma Benign promota

RIA Immunoperoxidase

Immunoperoxidase

Ruiter et al. (1982)

Immunoperoxidase

Ruiter et al. (1982)

Immunoperoxidase

Sanderson and Beverley (1983) Sanderson and Beverley (1983)

Immunoperoxidase

phenomenon? Nilsson et al. (1974) studied over 50 human tumor cell lines and tumors in short-term culture for P2-m expression and found that only the Daudi cell line did not express P2-m on the cell surface. Rosa et al. (1983) recently detected a point mutation in the initiation codon of Daudi Pe-m mRNA and suggest that because of this alteration, the P2-m mRNA cannot be translated. Murine embryonal carcinoma (Morello et al., 1982) and human teratocarcinoma (Andrews et al., 1981) were also shown not to express P2-m on the cell surface. In humans, the villous trophoblast (reviewed by Searle, 1982) as well as gestational choriocarcinoma cells do not express MHC-I antigens, whereas Pz-m expression was found to be low in two cell lines (Trowsdale et al., 1980) and completely absent in two different cell lines (Tanaka et al., 1981). Moreover, in recent years, a series of benign, well-differentiated human tumors and undifferentiated neoplasia of the same histogenetic origin have also been studied for P2-m and MHC-I antigen expression. The results of these studies and the data concerning P2-m and MHC-I molecule expression on teratocarcinoma, gestational choriocarcinoma, and the Daudi cell line are presented in Table I. The results of studies by Tjernlund and Forsum (1977), DiPersio et d.(1980, 1982), Turbitt and Mackie (1981), Krothals-Altes et d . (1981), Weiss et d . (1981),and

102

D. COHEN

Sanderson and Beverley (1983), show that a variety of well-differentiated tumors express Pz-m and MHC-I antigens, whereas the malignant undifferentiated histologic counterparts often do not express these cell surface molecules. These findings raise the possibility that suppression or permanent mutation of the B2-m gene or the MHC-1 heavy chain gene is not an uncommon event in the course of tumor progression toward malignancy. Since cell surface antigens are recognized by cytotoxic T lymphocytes in the context of MHC-I molecules (Zinkernagel and Doherty, 1979), lack of expression e€ these antigens may reduce the susceptibility of tumor cells to T lymphocytes and may enhance their chances for survival. lysis by *toxic Sandcrnm and Beverley (1983) recently proposed that a major function of interferons is to regulate the expression of MHC-I antigens. They proposed that tumor cells with high MHC-I antigen expression are eliminated by cytotoxic T cells, whereas tumor cells that fail to respond to interferons or are unable to synthesize or express MHC-I molecules escape immunologic effector mechanisms. Recent studies by Schrier et al. (1983)and Bernards et al. (1983) also show that lack of MHC-I molecule expression may play an important role in tumorigenesis. While adenovirus 12 as well as adenovirus 5 can transform rat cells in oitro, adenovirus 12-transformed cells are highly tumorigenic in the syngeneic host, whereas adenovirus 5-transformed cells are rarely tumorigenic in immunocompetent syngeneic hosts. These investigators showed that the MHC-I heavy chain is not synthesized in cells transformed by plasmids containing the transforming adenovirus 12 genes, whereas the MHC-I heavy chain is synthesized in cells transformed by plawnids containing the transforming adenovirus 5 genes. They also demonstrated that adenovirus 12-transformed cells are not susceptible to cytotoxic T-)yarphocyte killing directed against their alloantigens, whereas adenovirus S&mrEamed cells are susceptible to such T-cell kiiling. These experiments s u w drat adenovirus Wtransformed cells are tumorigenic in syngeneic hortr bemuse the suppreuion of MHC-I heavy chain expression does not enrMe tbe elimination of the cells by cytotoxic T lymphocytes. Im contrast, ademwirus $transformed cells express MHC-I molecules and are not turndgemic in syngeneic hosts, probably because such cells are killed by cytaQlic T lymphocytes. The studies on the expression of P2-m and MHC-I heavy chains on the surface of tumor cells show that lack of P2-m expression probably is a widespread phenomenon among undifferentiated tumors. The TVT may have originally developed on the mucosa of the external genitalia of a dog as a spontaneous tumor that does not express P2-m on the ceH s u h and was, b e r e h e , ahtransplantable. The preservation of the TVT as a coitally transmitted neoplasm in the dog was probably made possible by some unusual characteristics of canine external genitalia and canine coitus.

C A N I N E TRANSMISSIBLE VENEREAL TUMOR

103

C. CANINECOITUSAND THE TRANSMISSION OF THE TVT

The transmission of the TVT under natural conditions is highly efficient. Smith and Washbourn (1898a,b) reported the transmission of the disease by coitus from one affected male dog to 11 of 12 bitches. It seems likely that unique anatomical characteristics of the canine penis and some features of canine coitus may have contributed to the development of a coitally transmitted, cell-transplanted neoplasm in this species. The ejaculate of the dog is passed in three distinct portions that eventually mix in the uterus (Siegmund et al., 1961). To ensure the mixing of the spermatozoa with the other two fluids of the ejaculate, certain anatomical features that are unique to the family Canidae have evolved. The canine penis contains a bone, the 0s penis, which is believed to be a part of the corpus cavernosum that has ossified. In addition, behind the pars longa of the glans penis there is a rounded enlargement, the bulbus glandis, that is composed of erectile tissue (Sisson and Grossman, 1956). These unusual characteristics of the canine penis have a profound effect on the course of coitus in the dog. Following insertion of the penis into the vagina, the male dog and the bitch are “tied” to each other for up to 30 min. The tie is caused by the vascular engorgement of the bulbus glandis and contraction of vaginal muscles distal to the bulbus glandis. The male dog and the bitch usually attempt to separate from each other before the contraction of the vaginal muscles and the engorgement of the bulbus glandis have subsided. These mechanical strains on the external genitalia often cause injuries of the vaginal and penile mucosa, thus providing the bed for tumor transplantation. Indeed, it has been shown that TVT can be transmitted by placing the tumor cells on the scarified penis of a dog (Stubbs and Furth, 1934). VIII. Conclusion and Summary

The canine TVT is the only tumor known to be transmitted in nature by the transplantation of cells. Under natural conditions, the TVT is transmitted by coitus and experimentally the tumor can be induced by the inoculation of living tumor cells to immunocompetent allogeneic dogs. Evidence that the tumor is transmitted by cells is derived from cytogenetic and immunological studies. Cytogenetic studies show that TVT from different geographical origins is characterized by extensive and characteristic chromosomal aberrations. It seems unlikely that the extensive chromosomal changes could be repeatedly induced by an oncogenic virus and therefore support the view that the TVT is transmitted by the transplantation of cells. With the characterization of the banding patterns of normal dog chromosomes (Selden et al., 1975; Manolache et al., 1976), the background information for comparative

104

D. COHEN

banding pattern analysis of TVT chromosomes in different parts of the world is now available. Such studies may enable investigators to determine whether this unique tumor has a common origin or arose repeatedly in the domestic dog. Moreover, in recent years it has been shown that cellular oncogenes are localized near the break points of tumor-specific chromosomal aberrations, suggesting an involvement of these genes in the expression of malignancy (Klein, 1981; Leder et al., 1983; Heisterkainp et al., 1983). Since the TVT is one of the tumors that is characterized by tumor-specific chromosomal aberrations, this neoplasm could be used for investigations of the translocation of cellular oncogenes and their involvement in the expression of malignancy. The results of immunological studies also suggest that the TVT is transmitted by the transplantation of cells. It was shown that the same DL-A specificities are expressed on the surface of TVT cells that originated from different dogs and from different locations. In addition, the anti-TVT immune response is mounted, at least in part, against MHC antigens. Since those MHC antigens that are expressed by tumors are of host origin, these findings demonstrate that the TVT is transmitted by the transplantation of cells. It has also been shown that @,-microglobulin is not expressed on the surface of TVT cells. Since @,-microglobulin is required for the expression of class I MHC antigens, it seems likely that the DL-A antigens that are expressed on the surface of TVT cells are not class I MHC antigens but class I1 MHC molecules. The results of mixed lymphocytes turnor culture studies indeed suggest that class I1 MHC antigens are expressed on the surface of TVT cells. In recent years it has been established that undifferentiated tumors often do not express @,-microglobulin and class I MHC antigens on the cell surface, whereas differentiated tumors of the same histological origin express these cell surface antigens. The lack of class I MHC antigen expression on the surface of turnor cells is probably an expression of one of the stepwise changes that occur in tumors in the course of tumor progression toward increased autonomy (Foulds, 1958). Since cytotoxic T lymphocytes recognize cell surface antigens in the context of class I MHC antigens, tumor cells that do not express these antigens are probably less suceptible to killing by this effector mechanism. In the case of the TVT, this modification of the cell surface could account for the allotransplantability of the tumor and may explain why the TVT is not rejected, in most cases, several days after transplantation. The development of the TVT as a coitally transmitted neoplasm in the dog is probably facilitated by some unique characteristic of sexual intercourse in this species that leads to injuries of the vaginal and penile mucosa and thus provides the bed for tumor transplantation. The TVT is defined histologically as an undifferentiated round-cell neo-

CANINE TRANSMISSIBLE VENEREAL TUMOR

105

plasm of reticuloendothelial origin. The localization of the tumor under natural conditions and the development of cutaneous TVT following intravenous and intraperitoneal inoculation of the tumor cells indicate that the TVT may have developed originally as a cutaneous neoplasm. Further characterization of the TVT could be achieved by testing TVT cells for lymphoid cell markers such as receptors for the third component of complement and the Fc portion of immunoglobulins as well as membrane and cytoplasmic y and p immunoglobulin heavy chains. The interpretation of the results of such studies could be complicated by the in vivo coating of TVT cells with antibody and the infiltration of the tumor by host lymphoid cells. Since tissue culture lines of this tumor are not available, characterization of the TVT for lymphoid cell markers could be carried out on TVT grown in nude mice. Moreover, the use of recently developed monoclonal antibodies against surface antigens of canine lymphocyte subpopulations and class I1 MHC antigens (McKenzie and Fabre, 1981; Deeg et al., 1982; Wulf et al., 1982a,b; Iwaki et al., 1983; Urban et al., 1983) could also enable a more precise classification of the TVT. The demonstration that animals with a large TVT burden develop polycythemia and that TVT extracts contain erythropoietin is yet another interesting characteristic of this tumor. Transmissible veneral tumor cells probably synthesize and secrete erythropoietin and could, therefore, be used for the study of erythropoietin synthesis in uitro. Moreover, the TVT is the only available experimental animal model for the study of the kinetics of paraneoplastic polycythemia. Investigations of the immune response against the TVT have clearly demonstrated that the tumor is antigenic in the dog and that immunological mechanisms are involved in the induction of the spontaneous regressions of this neoplasm. While it has not yet been established whether tumor-associated transplantation antigens are expressed on the surface of TVT cells, it has been shown that the antibody response is directed, at least in part, against MHC antigens. Anti-TVT antibody against membrane-associated antigens can be detected in the serum of most animals during the whole course of the disease. In the initial stages of the disease, suppressive serum factors seem to facilitate tumor growth, but further increase in tumor volume is often followed by spontaneous regression. Since passive transfer of postregression serum can induce tumor regression, it seems likely that humoral antibody is involved in the induction of the spontaneous regression of the tumor. Postregression sera can inhibit TVT colony formation in agar, are cytotoxic to TVT cells in the presence of complement, and can mediate antibody-dependent cellular cytotoxicity against the TVT. Further progress in the analysis of the immune response against the TVT could be achieved by the molecular

106

D. COHEN

FIG.1. Black frames, TVT characteristics and products; white frames, host organ systems that have been shown to respond to TVT products; gray frames, interactions between the host and the TVT. Abbreointions used: TVT, transmissible venereal tumor; MHC, major histocompatibility complex; NEG, negative; POS, positive; Nuclear/Cytoplasmic, antigen detected by the LA1 assay; TATA, tumor associated transplantation antigen; MLTC, mixed lymphocyte tumor culture; LAI, leukocyte adherence inhibition; IF, indirect membrane immunofluorescence; C', complement; ADCC, antibody dependent cellular cytotoxicity; EPO, erythropoietin; BM, bone marrow; BUN, blood urea nitrogen. Recently, we have detected in the kidneys of TVT-bearing dogs mesangial proliferative glomerulonephritis. This type of glomerulonephritis is indicative of the presence of large circulating immune complexes in the serum. These data are currently being prepared for publication.

characterization of TVT antigens. Such studies could also enable investigators to redefine the relevance of the TVT as a model for tumor and transplantation immunology studies. Investigations of various aspects of this unique neoplasm require the establishment of TVT tissue culture cell lines. Attempts to propagate the TVT in culture resulted in most cases in the death of the tumor cells and the outgrowth of normal canine fibroblasts. Only Adams et al. (1981)succeeded in growing the TVT in tissue culture for prolonged periods but these cultures have not been preserved (E. W. Adams, personal communication). The development of a reproducible method for the propagation of the TVT in tissue culture could open new avenues for the investigation of this tumor and thus contribute to the understanding of malignant disease.

CANINE TRANSMISSIBLE VENEREAL TUMOR

107

In the present article the available information about this unique tumor has been reviewed and the relevance of some of the data to current concepts in cancer research has been discussed. The essential characteristics and the products of the TVT as well as the interactions of TVT products with the host and its organ systems are depicted in a schematic manner in Fig. 1.

ACKNOWLEDGMENTS Part of the research of the author was supported by the Israel Cancer Association, a grant in memory of Ada and Ernest Klein, the Israel Cancer Research Fund, The Leukemia Research Foundation, Inc., Chicago, IL, and The Richard Cohn Cancer Research Fund.

REFERENCES Abbott, P. K. (1966). Aust. Vet. J . 42, 29. Adams, E. W., and Slaughter, L. J. (1970). Pathol. Vet. 7, 498-502. Adarns, E. W., Carter, L. P., and Sapp, W. J. (1968). Cancer Res. 28, 753-757. Adams, E. W., Sapp, W. J., and Carter, L. P. (1981). Cornell Vet. 71, 336-346. Ajello, P. (1960). Nuova Vet. 36, 179-183. Ajello, P. (1961). Bull. SOC. Ztal. Biol. Sper. 37, 247-248. Ajello, P., and Gimbo, A. (1965a). Arch. Vet. Ztal. 16, 275-286. Ajello, P., and Gimbo, A. (1965b). Atti Soc. Ztal. Sci. Vet. 19, 736-739. Ajello, P., and Gimbo, A. (1965~).Biol. Lat. 18, 311-317. Amatruda, T. T., and Upton, G. V. (1974). Ann. N.Y. Acad. Sci. 230, 168-180. Andrews, P. W., Bronson, D. L., Wiles, M. V.,and Goodfellow, P. N. (1981). Tissue Antigens 17, 493-500. Arce-Gomez, B., Jones, E. A,, Barnstable, C. J., Solomon, E., and Bodmer, W. F. (1978). Tissue Antigens 11, 96-112. Barski, G . , and Cornefert-Jensen, Fr. (1966).J . Natl. Cancer Znst. 37, 787-797. Bashford, E. F., Murray, J. A., and Cramer, W. (1905). Sci. Rep. Znuest. Imp. Cancer Res. Fund. 2, 33-37. Basur, P. K., and Gilrnan, J. P. W. (1966). Cornell Vet. 56, 451-469. Battistacci, M., and Moriconi, F. (1974). Nuooa Vet. 50, 226-236. Beebe, S. P. (1907).J . Am. Med. Assoc. 49, 1492-1493. Benjamin, S. A., and Noronha, F. (1967). Cornell Vet. 57, 526-542. Bennett, B. T., Taylor, Y.,and Epstein, R. (1975a). Transplant. Proc. 7, 503-505. Bennett, B. T., Debelak-Fehir, K. M., and Epstein, R. B. (1975b).Cancer Res. 35,2942-2947. Bernards, R., Schrier, P. I., Houwellig, A., Bos, J. L., van der Eb, A. J., Zijlstra, M., and Melief, C. J. M. (1983). Nature (London) 305, 776-779. Beschorner, W. E., Hess, A. D., Nerenberg, S. T., and Epstein, R. B. (1979). Cancer Res. 39, 3920-3927. Bhan, A. K., and DesMarais, C. L. (1983).1.Natl. Cancer Znst. 71, 507-516. Bishop, J. M. (1983). Annu. Reu. Biochem. 52,301-354. Bloom, F., PaiT, G. H., and Noback, C. R., (1951). Am. J . Pathol. 27, 119-139. Bowles, C. A., White, G . S., and Lucas, D. (1975).J . Zmmunol. 114, 399-402. Brodey, R. S., and Roszel, J. F. (1967).J . Am. Vet. Med. Assoc. 151, 1294-1307. Brown, N. 0. (1981).J . Am. Anim. Hosp. Assoc. 17, 911-916. Brown, N. O., Calvert, C., and MacEwen, E. G. (1980).J.Am. Vet. Med. Assoc. 176,983-986.

108

D . COHEN

Brown, N . O., MacEwen, E. G., and Calvert, C. (1981). CaliJ Vet. 35, 6-10. Cabanie, P., van Haverbeke, G., and Magnol, J. P. (1973). Reo. Med. Vet. (Toulouse) 124, 1239-1253. Calvert, C . A , , Leifer, C. E., and MacEwen, E. G., (1982).J. Am. Vet. Med. Assoc. 181, 163164. Carteaud, J. P., Lagneau, F., and Fontain, M. (1963). Bull. Soc. Fr. D e m t o l . Syphiligr. 70, 793-799. Catovsky, D., Greaves, M. F., Pain, C., Cherchi, M . , Janossy, G., and Kay, H. E. M. (1978). Lancet 1, 749-751. Cermak, K . , Marolt, J., Herceg, M., and Maticic, Z. (1976). Prax. Vet. 24, 279-287. Chandler, J. P., and Yang, T.-J. (1981). Br. J . Cancer 44, 514-521. Cockrill, J. M., and Beasley, J. N. (1975). Am. J. Vet. Res. 36, 677-681. Cockrill, J. M . , and Beasley, J. N. (1979). Am. J . Vet. Res. 40, 409-410. Cohen, D. (1972). Int. J. Cancer 10, 207-212. Cohen, D. (1973). Eur. J. Cancer 9, 253-258. Cohen, D. (1974). Transplantation 17, 8-11. Cohen, D. (1978). Isr. J. Med. Sci. 14, 14-19. Cohen, D. (1980a). In “The Canine as a Biomedical Research Model: Immunological, Hematological, and Oncological Aspects” (M. Shifrine and F. D. Wilson, eds.), pp. 364-479. Technical Information Center/U.S. Dept. of Energy. Cohen, D. (1980b).J. Natl. Cancer Inst. 64, 317-321. Cohen, D., and Steel, G. G. (1972). Br. J . Cancer 26, 413-419. Cohen, D., Gurner, B. W., and Coombs, R. R. A. (1971). Br. J. E x p . Pathol. 52, 447-451. Cohen, D., Yagil, R., and Waner, T. (1978). Refuah. Vet. 35, 12-13. Cohen, D., Greenzweig, C., Rabani, R., and Bargum, R. (1982). Proc. Int. Cancer Congr., 13th, Seattle 684. Cohen, D., Shalev, A , , and Krup, M. (1984).J. Natl. Cancer Inst., 72, 395-401. Crile, G. W., and Beebe, S. P. (1908).J. Med. Res. 18, 385-406. Deeg, H. J., Wulff, J. C., DeRose, S., Sale, G. E., Braun, M., Brown, M. A , , Springmeyer, S. C., Martin, P. J . , and Storb, R. (1982). Zmniunogenetics 16, 445-457. DeMonbreun, W. A., and Goodpasture, E. W. (1934). Am. J. Cancer 21, 295-321. DiPersio, L., Dingle, S., Michael, J. G., and Pesce, A. J. (1980). Exp. Cell Biol. 48, 429-438. DiPersio, L., Weiss, M. A., Michael, J. G . , and Pesce, A. J . (1982). Am. J. Clin. Pathol. 77, 700-704. Dozza, G., and Torlone, V . (1960a).Atti Soc. Ital. Vet. 14, 531-536. Dozza, C . , and Torlone, V. (1960h). Vet. Ztal. 11, 647-658. Drommer, W., and Schulz, L.-CL. (1969). Pathol. Vet. 6, 273-286. Duncan, J. R., and Prasse, K. W. (1979). Vet. Pathol. 16, 673-679. Epstein, R. B., and Bennett, B. T. (1974). Cancer Res. 34, 788-793. Epstein, R. B., and Sarpel, S. C. (1980). E x p . Hemutol. 8, 683-689. Esser, R . E., Cosgrove, M., Scott, G., and Cosimi, A. B. (1977). Transplantation 24,223-225. Fidler, I . J. (1978). Cancer Res. 38, 2651-2660. Foulds, L. (1958).J. Chronic Dis. 8, 2-37. Geissler. (1895). Verh. Dtsch. Ges. Chir. 24, 87-92. Goldwasser, E. (1975). Fed. Proc., Fed. Am. SOC. Exp. Biol. 34, 2285-2292. Gorman, N. T., Probst, C. W., and Goring, R. L. (1983).J . Am. Vet. Med. Assoc. 183, 893895. Cross, L. (1983). In “Oncogenic Viruses,” 3rd Ed., pp. 117-122. Pergamon, Oxford. Grosser, N., and Thomson, D. M. P. (1976). Znt. J. Cancer 18, 58-66. Hammond, D., and Winnick, S. (1974). Ann. N.Y. Acad. Sci. 230, 219-226.

CANINE TRANSMISSIBLE VENEREAL TUMOR

109

Hansen, M., and Hummer, L. (1979). Prog. Cancer Res. Ther. 11, 199-207. Harding, M. W., and Yang, T.-J. (1981a). Znt. J . Cancer 27, 349-355. Harding, M. W., and Yang, T.-J. (1981b). Znt. J . Cancer28, 361-366. Hayes, M. H., Biggar, R. J., Pickle, L. W., Hoover, R., and Toft, J. D. (1983). Am. J. Epidemiol. 117, 108-109. Heisterkamp, N., Stephenson, J. R., Groffen, J., Hansen, P. F., deKlein, A,, Bartram, C. R., and Grosveld, G . (1983). Nature (London) 306, 239-242. Hernhdez-JBuregui, P., Gonzalez-Angulo, A., and De La Vega, G. (1973). J. Natl. Cancer Znst. 51, 1187-1196. Hess, A., Cunningham, B., Bennett, B. T., and Epstein, R. (1975). Transplant. Proc. 7,507511. Hess, A. D., Catchatourian, R., Zander, A. R., and Epstein, R. B. (1977). Cancer Res. 37, 3990-3994. Higgins, D. A. (1966). Vet. Rec. 79, 67-71. Ho, C. K., and Babiuk, L. A. (1978). Zmmunology 35, 733-740. Holmes, J. M. (198la). Res. Vet. Sci. 30, 248-250. Holmes, J. M. (1981b). Res. Vet. Sci. 31, 306-311. Idowo, L. (1977). Res. Vet. Sci. 22, 271-273. Ivoghli, B. (1977). Vet. Pathol. 14, 289-290. Iwaki, Y., Terasaki, P. I., Kinukawa, T., Thai, T. H., Root, T., and Billing, R. (1983). Transplantation 36, 189-191. Kakpakova, E. S., Pogosyanz, E. E., and Ponomarkov, V. I. (1968). Vopr. Onkol. 14, 43-50. Karlson, A. G., and Mann, F. C. (1952). Ann. N.Y. Acad. S c i . 54, 1197-1213. Kennedy, J. R . , Yang, T. J., and Allen, P. L. (1977). Br. J. Cancer 36, 375-385. Kerbel, R. S. (1979). Nature (London) 280, 358-360. Klein, E., and Moller, E. (1963).J . Natl. Cancer Znst. 31, 347-364. Klein, E., Klein, G . , Nadkarni, J. S., Nadkarni, J. J., Wigzell, H., and Clifford, P. (1967). Lancet 2, 1068-1070. Klein, G. (1981). Nature (London) 294, 313-318. Klein, G., and Klein, E. (1958).J . Cell. Comp. Physiol. 52 (Suppl.), 125-160. Klein, G., and Klein, E. (1977). Transplant. Proc. 9, 1095-1103. Klein, G . , Terasaki, P., Billing, R., Honig, R., Jondal, M., Rosen, A., Zeuthen, J., and Clements, G. (1977). Znt. J . Cancer 19, 66-76. Klein, J. (1982). In “Immunology, the Science of Self, Non-Self Discrimination,” pp. 453-458. Wiley, New York. Koeffler, H. P., and Goldwasser, E . (1981). Ann. Znt. Med. 94, 44-47. Krakowa, S . , and Guyot, D. J. (1977). Infect. Zmmun. 17, 73-79. Krishnan, Y., Li, C. Y., and Daniel, W. P. (1983). Am. J . Clin. Pathol. 79, 157-165. Krothals-Altes, H. R . , Van Vinninghe, L., and Neumann, H. A. M. (1981). Br. J . Dermutol. 104, 515-519. Kudo, T., Koike, T., Otomo, K., and Sakai, T. (1974).Jpn. J . Vet. Res. 22, 105-110. Kuramochi, T. (1974).J . Zmmunol. Methods 5 , 65-69. Largiader, F., Illi, E . , and Uhlschmid, G. (1972). Eur. Surg. Res. 4, 377-388. Leder, P., Battey, J., Lenoir, G., Moulding, C., Murphy, W., Potter, H., Stewert, T., and Taub, R. (1983). Science 222, 765-771. Locke, K. B., Yeh, F. J., and Hooper, P. T. (1975). Aust. Vet. J . 51, 449. Lombard, C., and Cabanie, P. (1967). Bull. Cancer 54, 357-365. Lombard, Ch., and Yulzari, J. J. (1967). Rev. Med. Vet. 118, 85-93. Lombard, C., Cabanie, P., and Izard, J. (1967).J. Microsc. (Paris) 6, 81-84. MacGrath, C. J . (1974).J. Am. Vet. Med. Assoc. 164, 1117-1122.

110

D. COHEN

McKenna, J. M., and Prier, J. E. (1966). Cancer Res. 26, 137-142. McKenzie, J. L, and Fabre, J. W. (1981). Transplantation 31, 275-282. Makino, S. (1963).Ann. N.Y. Acad. Sci. 108, 1106-1122. Manolache, M., Ross, W. M., and Schmid, M. (1976). Can. J . Genet. Cytol. 18, 513-518. Marti, J. H., Grosser, N., and Thomson, D. M. P. (1976). Znt. J. Cancer 18, 48-57. Miles, C. P., Moldavannu, G., Miller, D. G., and Moore, A. (1970).Am. J . Vet. Res. 31, 783790. Miller, D. A., and Miller, 0. J. (1983). Ado. Cancer Res. 39, 153-182. Miller, S. C., Codington, J. F., and Klein, G. (1982). J. Natl. Cancer Znst. 68, 981-988. Miyake, T., Hung, C. K. H., and Goldwasser, E. (1977). J . B i d . Chem. 252, 5558-5564. Mohanty, G. C., and Rajya, B. S. (1977). Vet. Pathol. 14, 420-425. Morello, D., Daniel, F., Baldacci, P., Cayre, Y., Gachelin, G., and Kourilsky, P. (1982). Nature (London) 296, 260-262. Murray, M., James, Z. H., and Martin, W. B. (1969). Res. Vet. Sci. 10, 565-568. Muzetto, P., Petruzzi, V., and Lepori, S. (1975). Clin. Vet. 98, 124-128. Nanta, Marques, Lasesrre, Bazex, Bru, and Puget. (1951). Reo. Med. Vet. (Toulouse)102,321337. Ndiritu, C. G., Mbogwa, S. W., and Sayer, P. D. (1977). Mod. Vet. Pract. 58, 940-946. Nelson, R. W., Hager, D., and Zanjani, E. D. (1983).J . Am. Vet. Med. Assoc. 182, 1396-1397. Nilsson, K., Evrin, P. E., and Welsh, K. T. (1974). Transplant. Reo. 21, 53-84. Novinsky, M. A. (1876). Zentralbl. Med. Wiss. 14, 790-791. Oduye, 0. O., Ikede, B. O., Esuruoso, G. O., and Akpokodje, J. U. (1973).J. Small Anim. Pract. 14, 625-637. Onions, D. (1977).J~ Natl. Cancer Znst. 59, 1001-1006. Oshimura, M . , Sasaki, M., and Makino, S. (1973). J . Natl. Cancer Znst. 51, 1197-1203. Osipov, N. E., and Golubeva, V. A. (1976). Veterinariya 7, 97-98. Oughton, S. M. J., and Owen, L. N. (1974). Res. Vet. S c i . 17, 414-416. Pakes, S. P., Griesemer, R. A,, and Kasza, L. (1965). Am. J. Vet. Res. 26, 837-843. Palker, T. J.. and Yang, T.-J. T. (1981). J . Natl. Cancer Inst. 66, 779-787. Peterson, M. E., and Zanjani, E. D. (1981).J. Am. Vet. Med. Assoc. 179, 995-996. Ploegh, H. L., Cannon, L. E., and Strominger, J. L. (1979). Proc. Natl. Acad. Sci. U.S.A. 76, 2273-2277. Powers, R. D. (1968). Am. J. Vet. Res. 29, 1637-1645. Prier, J. E. (1966). Nature (London) 212, 724-726. Prier, J. E., and Johnson, J. H. (1964). J . Am. Vet. Med. Assoc. 145, 1092-1094. R&, M. (1975). Nature (London) 254, 287-292. Rosa, F., Berissi, H., Weissenbach, J., Maroteaux, L., Fellous, M . , and Revel, M. (1983). EMBO J. 2, 239-243. Ruiter, D. J., Bhan, A. K., Harrist, T. J., Sober, A. J., and Mihm, M. C. (1982).J . Zmmunol. 129, 2808-2815. Rust, J. H. (1949).J . Am. Vet. Med. Assoc. 114, 10-14. Samso, A. (1966). Bibl. France 155 (Suppl. D), 3650. Sanderson, R. A., and Beverley, P. C. L. (1983). Zmmunol. Today 4, 211-213. Sapp, W. J., and Adams, E. W. (1970). Am. J. Vet. Res. 31, 1321-1323. Schalm, 0. W., Jain, N. C., and Carroll, E. J. (1975). In “Veterinary Hematology,” 3rd Ed., p. 87. Lea & Febiger, Philadelphia, Pennsylvania. Schrier, P. I., Bernards, R., Vaessen, R. T. M. J., Houwellig, A., and van der Eb, A. J. (1983). Nature (London) 305, 771-775. Searle, R. F. (1982). Zmmunol. Today 3, 63-64 Sege, K., Rask, L., and Peterson, P. A. (1981). Biochemistry 20, 4523-4530.

CANINE TRANSMISSIBLE VENEREAL TUMOR

111

Selden, J. R., Moorehead, P. S., Oehlert, M. L., and Patterson, D. F. (1975). Cytogenet. Cell Genet. 15, 380-387. Shenvood, J. B., and Goldwasser, E. (1979). Blood 54, 885-893. Shimkin, B. M. (1955). Cancer 8, 653-655. Siegrnund, 0. H. (1961). “The Merck Veterinary Manual,” p. 1005. Merck and Co., Inc., Rahway, New Jersey. Sisson, S., and Grossman, J. D. (1956). In “The Anatomy of the Domestic Animals,” 4th Ed. Saunders, Philadelphia, Pennsylvania. Smith, G. B., and Washbourn, J. W. (1898a).J . Pathol. Bacteriol. 5, 99-100. Smith, G . B., and Washbourn, J. W. (1898b). Br. Med. J. 2, 1346-1347. Smith, G. B . , and Washbourn, J. W. (1899). Br. Med. J . 2, 1807-1810. Sonoda, M., Niiyama, M., and Mori, M. (1970).Jpn. J. Vet. Res. 18, 145-151. Spence, J. A., Holt, P. E . , Sayer, P. D., Ruttcher, D., and Cooper, J. E. (1978).J . Small Anim. Pract. 19, 175-184. Sticker, A. (1904). 2. Krebsforsch. 1, 413-444. Sticker, A. (1905). Bed. Tierarzt. Wschr. 21, 353-358. Sticker, A. (1906). Z. Krebsforsch. 4, 227-314. Strominger, J. L., Mann, D. L., Parham, P., Robb, R., Springer, T., and Terhorst, C. (1977). Cold Spring Harbor Symp. Quant. Biol. 41, 323-329. Stubbs, E. L., and Furth, J. (1934). Am. J . Pathol. 10, 275-286. Sytowski, A. J., Richie, J. P., and Bicknell, K. A. (1983). Cancer Res. 43, 1415-1419. Takayama, S . , and Makino, S. (1961). 2. Krebsforsch. 64, 253-261. Tanaka, K . , Naheshima, Y., Takahashi, H., Takeuchi, S., Nabeshima, Y., and Ogata, K. (1981). Cancer Res. 41, 3639-3641. Taylor, N., Shifirine, M., Wolf, H. G., and Trommershausen-Smith, A. (1975). Transplant. Proc. 7, 485-493. Terner, U. K., Noujaim, A. A., Lentle, B. C., andTurner, C. J. (1981). Znt. J. Nucl. Med. Biol. 8, 357-362. Thorburn, M. J., Gwynn, R. V. R., Ragbeer, M. S., and Lee, R. I. (1968). Br. J. Cancer 23, 720-727. Thrall, D. E. (1982). Vet. Radiol. 23, 217-219. Tjernlund, U. M., and Forsum, U. (1977). Acta. D e m t o l . Venereol. 57, 503-506. Trope, C. (1975). Neoplasms 22, 171-180. Trowsdale, J., Travers, P., Bodmer, W. F., and Patillo, R. A. (1980).J . E r p . Med. 152, 11s17s. Turbitt, M. L., and Mackie, R. M. (1981). Br. J. Dermutol. 104, 507-513. Urban, C., Wulff, J. C., Deeg, H. J., RaK, R. F., Durkopp, N., and Storb, R. (1983). Transplantation 35, 469-473. Van Der Feltz, M. J. M., Van Der Korput, J. A. G. M., Giphart, M. J., and Westbrook, D. L. (1981). Transplantation 32, 253-255. Wade, H. (1908).J . Pathol. Bacteriol. 12, 384-425. Wardrop, K. J., Nakamura, J., and Giddens, W. E. (1982). Lab. Anim. Sci. 32, 280-282. Wasecki, A,, and Mazur, 0. (1977). Med. Weter. 33, 142-143. Weber, W. T., Nowell, P. C., and Hare, W. C. D. (1965).J. Natl. Cancer Znst. 35, 537-547. Wehr, (1888). Zentralbl. Chir. 15, 8-9. Wehr, (1889). Arch. Klin. Chir. 39, 226-228. Weir, E. C., Pond, M. J . , Duncan, J. R., and Polzin, D. J. (1978).J . Am. Anim. Hosp. Assoc. 14, 532-536. Weiss, M. A., Michael, J. G., Pesce, A. J., and DiPersio, L. (1981). Lab. Znuest. 45, 46-57. Woirnant, X., and ChaEaux, S. (1977). Rec. Med. Vet. 153, 331-338.

112

D. COHEN

Wong, H., Terner, U. K . , English, D., Noujaim, A. A., Lentle, B. C., and Hill, J. R. (1980). Int. J. Nucl. M e d . Biol. 7, 9-16. Wright, D. H., Peel, S., Cooper, E. H., and Hughes, D. T. (1970). Reo. Eur. E t u d . Clin. Riol. 15, 155-160. WulfF, J. C., Deeg, H. J.. and Storb, R. (1982a). Transplantation 33, 616-620. Wulff, J. C., Durkopp, N., Aprile, J., Tsoi, M. S., Springmeyer, S. C., Deeg, H. J., and Storb, R. (1982b). E r p . Hematol. 10, 609-619. Yallow, R. S. (1979). Prog. Cancer Res. Ther. 11, 209-216. Yam, L. T., Li, C. Y.,and Lam, K. W. (1971). N . Engl. J. M e d . 284, 357-360. Yang, T. J., and Jones, J. B. (1973).1.Nod. Cancer Znst. 51, 1915-1918. Yang, T.J., and Kennedy, J. R. (1976). Lab. Znoest. 35, 302-305. Yang, T.-J., Kennedy, J. R., and Andrews, R. B. (1976a).Am. J . Pathol. 83, 359-366. Yang, T.-J., Roberts, R. S., and Jones, J. B. (1976b). Virchows Arch. B Cell Path. 20, 197-204. Zander, A. R . , Boopalan, N., and Epstein, R. B. (1975). Transplant. Proc. 7, 369-373. Zander, A. R . , Sarpell, S., Cunningham, B., Hess, A. D., Johnston, D. A., and Epstein, R. B. (1980). Exp. Hemutol. 8, 1123-1128. Zarrouk, K., Huard, M., Mathlouthi, A,, and Ben Khlifa, A. (1980).Arch. Znst. Pasteur Tunis 57,53-59. Zinkernagel, R. M . , and Doherty, P. C. (1979). Ado. Zmmunol. 27, 51-177.

BIOLOGICAL AND MOLECULAR ANALYSIS OF p53 CELLULAR-ENCODED TUMOR ANTIGEN

Varda Rotter and David Wolf Department of Cell Biology, The Weizmann Institute of Science, RehOVQt, Israel

I. Introduction . . . . . . . . . . . . . . 11. p53 Is Complexed with Viral111. Expression of p53 in Nontran

..................... 113 . . . . . . . 114 . . . . . . . 116 s......

VII. The Use of p53 as a Marker for Cell Transformation. . . . . . . . VIII. The Possible Function p53 Fulfills in Transformed Cells ..................... IX. Molecular Mechanism Controlling the Expression of p53. . . References. . . . . . . . .......................

123 130 139

I. Introduction

Transformation of normal cells into malignant cancer cells involves a cascade of cellular events; one of the initial steps involves the activation of a quiescent cellular oncogene, leading to the synthesis of a functional oncogene protein. The activation of cellular oncogenes can be mediated by various molecular mechanisms including chromosomal translocation (Klein, 1981), integration of the cellular oncogene sequences into viral genomes (Witte et al., 1978; Reynold et aZ., 1978; Hanafusa et al., 1980), insertion of viral-like promoter elements next to cellular oncogenes (Hayward et al., 1981), point mutations (Parda et al., 1982; Santos et al., 1982), and probably other yet uncovered mechanisms. The initial step of oncogene(s) activation may then be followed by the overproduction of cellular proteins which are no longer regulated by normal control mechanisms. One such protein is the cellular-encoded p53 tumor antigen, which we have been studying for several years. Augmented synthesis of p53 molecules was observed in a wide range of cell lines (Lane and Crawford, 1979; Linzer and Levine, 1979; DeLeo et al., 1979; Rotter et al., 1980, 1981) and in primary tumors in mice (Rotter, 1983). This protein was found in a variety of tissue types including sarcoma and leukemia cells of several species (Kress et al., 1979; Simmons et d . , 1980; Dippold et al., 1981; Crawford et al., 1981; Rotter et al., 1983~).Tumor cells induced by very different techniques express this protein; cells transformed by RNA or 113 ADVANCES IN CANCER RESEARCH, VOL. 43

Copyright Q 1885 by Academic Press, Inc. All rights of reproduction in any form resewed. ISBN 0-12-w6643-2

114

VARDA R O V E R AND DAVID WOLF

DNA tumor viruses or chemical carcinogens contain the same p53 molecule (DeLeo et al., 1979; Linzer and Levine, 1979; Rotter et al., 1981), suggesting that augmented p53 synthesis is a common secondary event following a primary signal that induces malignant transformation.

II. p53 Is Complexed with Viral-Encoded Protein

p53 was originally observed as a cellular product that formed a stable complex with the viral large T antigen expressed in Simian Virus 40 (SV40)transformed cells (Lane and Crawford, 1979; Linzer and Levine, 1979). Immunoprecipitation of cell lysates with polyclonal antibodies indicated the presence of the nonviral-encoded p53, in addition to the large T and small t proteins synthesized in SV4O-transformed cells. The conclusion that p53 is complexed with the large T antigen was initially derived from the observation that anti-large T antigen monospecific antibodies, prepared against a gel-purified T antigen fraction, efficiently coimmunoprecipitated the nonviral p53 (Lane and Crawford, 1979). It should be mentioned that the large T antigen and p53 do not share any homology in structure and represent two immunological epitopes expressing different antigenic determinants. The existence of a stable complex between these two proteins was further confirmed when anti-p53 monoclonal antibodies were used instead of anti-large T antigen. In this case again, the large T antigen coimmunoprecipitated with the p53 (Gurney et al., 1980; Harlow et al., 1981; Rotter et al., 1981). The same pattern was observed when cell lysates were immunoprecipitated with anti-p53 monoclonal antibodies prepared against p53 of SV40 or non-SV40transformed cell origin (Rotter et al., 1981). The existence of a stable p53-large T antigen complex implies high affinity binding between these proteins. Indeed, McCormick et al., (1981) showed that mixing these two proteins under in uitro conditions yields the formation of a stable complex between the two. It should be noted, however, that the stability of this complex varies in the different animal species. In the mouse, the complex is probably the most stable and apparently most of the protein is found in the complex. On the other hand, however, in monkey nonpermissive SV40-infected cells, very little of the p53 protein is complexed with the large T antigen. Separation of monkey nonpermissive SV4O-infected cell lysates over a glycerol gradient indicated only a low percentage of p53 molecules complexed with the large T; most of the p53 was found in the uncomplexed form (McCormick and Harlow, 1980). In humans, an intermediate situation exists; a complex is formed, but some of the p53 is unbound. Variations in stability of the p53-large T antigen complex could be due to subtle alterations in the structure of the p53 molecule found in various

p53 CELLULAR-ENCODED

TUMOR ANTIGEN

115

species, which may be responsible for the variations in affinity to the large T antigen. Close association of p53 with a tumor antigen was also found in cells transformed by adenovirus. Sarnow et al. (1982)have found a stable complex between Elb-58 kDA adenovirus-encoded tumor antigen and the cellularencoded p53. They found that specific antibodies directed against the Elb-58 kDa or the p53 proteins both coimmunoprecipitated the two proteins. Although all of the cellular-encoded p53 protein is found in a complex with the Elb-58 kDa, in this case some of the viral Elb-58 kDa protein was detected in a form not associated with the cellular-encoded p53 (Sarnow et al., 1982). Prompted by the observation that two different DNA tumor virus antigens bind the cellular-encoded p53, Luka et al. (1980)experimented to determine whether the 53K protein that copurifies with the Epstein-Barr virus (EBV) antigen of transformed human and mouse cells is the cellular-encoded p53. Although the initial results seemed to suggest an identity between two proteins, detailed analysis employing several established anti-p53 monoclonal antibodies showed clearly that the 53K cellular protein that is complexed with the EBV antigen differs structurally from the p53 cellular protein complexed with the large T antigen. The 53K protein that copurified with the EBV-encoded protein could very well be another member of the class of cellular-encoded proteins that is overproduced in the EBV-transformed cells and forms a complex with the EBV-encoded antigen (Jornvall et al., 1982). Although several laboratories have documented the existence of p53DNA viral-encoded complexes, the function of these complexes and the possible reason for their formation remain unknown. It has been speculated that by forming a complex with the large T antigen, the short-lived p53 avoids being degraded (Oren et al., 1981). Such a protein stabilization was suggested to be the main cause for accumulation of the p53 in SV40-transformed cells. Unlike in DNA virus-transformed tumor cells, p53 in several other systems apparently does not form any detectable stable complexes with other proteins. In the Abelson murine leukemia virus (A-MuLV)-transformed cells, we detected the p53 as an unbound single molecule immunoprecipitating with polyclonal or monoclonal anti-p53 antibodies (Rotter et al., 1980). The A-MuLV-transformed cells express concomitantly the viral-encoded p120 oncogene protein (Witte et al., 1979), which could be a potential partner for complexing with p53. Careful analysis of proteins extracted with different detergents sedimented through glycerol gradients did not reveal any possible complexes between the ab p12O or other cellular products (Rotter et al., unpublished data). Likewise, DeLeo et al. (1979) observed

116

VARDA ROlTEH A N D D A V I D WOLF

p53 in chemically transformed cells, again in its unbound form. In both of these examples, the amount of p53 detected was much higher than that observed in nontransformed cells. The accumulation of p53 in these cells is probably mediated by a mechanism other than stabilization through complex formation with other proteins. p53 is found in tumor cells in its phosphorylated form. The addition of radioactive phosphorus yielded a phosphorylated p53 that immunoprecipitated with anti-p53 monoclonal antibodies. Under in vivo conditions, p53 was phosphorylated onto a serine amino acid (Rotter et al., 1981; Jay et al., 1981). Jay et al. (1981) showed that immunoprecipitation of p53 with a specific anti-p53 monoclonal antibody (200.47) yielded a p53 protein that was also autokinased under in vitro conditions. In our experiments, p53 immunoprecipitated with several other anti-p53 monoclonal antibodies apparently did not bind the radioactive phosphate of y-ATP under in uitro conditions (Rotter et al., 1981). These conflicting results, concerning autokinase of p53 in vitro after immunoprecipitation with one reagent and not with another could be explained by the assumption that some monoclonal antibodies bind directly to the site of phosphorylation. It was suggested that in SV40-transformed cells, in vivo phosphorylation of the p53 is a function of its being complexed with the large T antigen (Melero et al., 1980). 111. Expression of p53 in Nontransformed Cells

The assumption that p53 is encoded by the normal cellular genome is based on the observation that it is expressed in several types of nontransformed cells. Judged by size and binding to anti-p53 monoclonal antibodies, it appears that p53 of normal cells is structurally similar to that found in transformed cells. The major difference observed so far is quantitative; while tumor cells synthesize high levels of p53, in normal cells it is hardly detected. In addition to an increased rate of synthesis, the amount of steadystate p53, evaluated by radioimmunoassay, is also higher in tumor cells than in normal cells (Benchimol et al., 1982). Screening of normal lymphoid cells indicated the presence of p53 in normal thymocytes (Rotter et al., 1980; Jay et al., 1980). In other normal lymphoid or nonlymphoid cell populations, we did not detect any biosynthetically labeled p53 immunoprecipitating with specific anti-p53 antibodies. The relatively augmented synthesis of p53 in normal thymocytes compared to other normal cells could be due to the fact that this population consists of highly dividing cells. Normal embryonic cells are another example of nontransformed cells synthesizing detectable amounts of p53. Levels of this protein peaked in primary cultures of 12- to 14-day-old embryos and declined toward the sixteenth day (Mora et al., 1980, 1982; Chandrasekaran et aZ., 1981). A reduction in p53 level was also

p53

CELLULAR-ENCODED TUMOR ANTIGEN

117

observed in F-9 embryonal carcinoma cells undergoing retinoic acid differentiation (Oren et al., 1982) and in Friend erythroleukemia cells induced to differentiate (Shen et al., 1983; Ben-Dori et al., 1983).These results suggest that normal cell differentiation may be accompanied by a reduction in the level of p53 expression and support the notion that p53 is an embryonic protein that is apparently suppressed in adult differentiated cells. Malignant transformation of normal cells may be therefore regarded as a dedifferentiation process leading to de novo expression of p53. Several investigators suggested that the basis for the quantitative difference of p53 is transformed and nontransformed cells is due to posttranslational regulatory mechanisms. p53 is apparently synthesized at about the same rate in normal and transformed cells. However, the p53 in nontransformed NIH-3T3 cells is unstable; it is rapidly degraded and has a short half-life of about 20-60 min. The level of p53 in SV40-transformed cells is higher because the protein is not rapidly turned over and its half-life is greater than 20 hr in these cells (Oren et al., 1981). However, it seems that this is not the only way p53 protein is stabilized in transformed cells. The fact that p53 is well conserved implies that the protein fulfills a function in normal cells. Studies by Milner and Milner (1981) showed that p53 is induced in cells treated with mitogens and suggested that it starts functioning early during the transition from Go to G,. Using quiescent nontransformed NIH-3T3 cells which were induced to enter a round of cell division by serum stimulation, Mercer et al. (1982) showed that the injection of anti-p53 monoclonal antibodies into the cells inhibited DNA synthesis. They suggested that p53 is synthesized as a late G, protein. Others suggested that p53 may be the R protein that was shown to play a role in initiating events leading to DNA synthesis (Campisi et al., 1982). IV. p53 Is Immunogenic

Although p53 is a cellular-encoded product and can be defined in immunological terms as a “self’ protein, there are occasions when it induces the production of specific antibodies against itself. In our early studies, we observed that mice bearing a syngeneic A-MuLV-induced tumor produced anti-p53 antibodies at a late stage of tumor development (Rotter et al., 1980). Likewise, a survey of human serum showed that cancer patients exhibit a certain incidence of anti-p53 antibodies, whereas sera of healthy controls do not contain anti-p53 antibodies (Crawford et al., 1982). The immunization of mice and humans against the p53 could be the result of the breakdown of immunologic tolerance against the p53 self protein caused by overproduction of this product. Another possibility is that p53 of tumor cells exhibits subtle structural changes compared to that of normal cells, which elicited an

118

VARDA ROIITER A N D DAVID WOLF

immunologic response. It should be mentioned that in addition to anti-p53 antibodies, tumor-bearing mice produce at a certain frequency antibodies against the oncogene expressed concomitantly in the tumor cells. For example, mice bearing an A-MuLV-induced tumor produce anti-abl antibodies (Witte et al., 1979). Likewise, mice or rats bearing an Avian sarcoma virusinduced tumor produce anti-src antibodies. This suggests that overproduction of transformation-related protein in tumor cells stimulates production of specific antibodies in the host. It is not clear, however, whether these antibodies play a role in the rejection or enhancement of the growing tumors. The fact that p53 is immunogenic under certain conditions has made possible the production of anti-p53 monoclonal antibodies. Figure 1 illustrates the immunoprecipitation pattern of mouse p53 employing several types of anti-p53 monoclonal antibodies. The RA3-2C2 was established by immunization of rats with mouse A-MuLV-transformed lymphoid cells (Rotter et al., 1980; Coffman and Weissman, 1981). In this instance, immunization over an allogenic barrier stimulated the production of anti-p53 antibodies which were directed against a mouse specific determinant. This antigenic determinant was not detected in the p53 molecule of other species, including the rat that served as the immunized host. The absence of a RA3-2C2 site in the human p53 was also observed by others (Benchimol et aZ., 1982). RA3-2C2 antibodies were selected for their ability to bind to the cell surface of pre-B lymphocyte through a determinant other than p53 (Coffman, 1982). The p53 molecule, which is found in the cytoplasmic and nuclear compartments of tumor cells, apparently shares a common antigenic determinant with the pre-B lymphocyte surface membranal molecule (Rotter and Coffman, unpublished data). Further indication supporting the idea that RA3-2C2 binds two independent proteins comes from our study with L12 cells. These cells are A-MuLV-transformed pre-B lymphoid cells that lack the expression of p53 yet exhibit positive immunofluorescent cell surface staining with RA3-2C2 anti-p53 monoclonal antibodies (Rotter and Coffman, unpublished data). It is not yet clear whether, in addition to their sharing a common antigenic determinant, these two molecules have other common properties. A third molecule which binds the RA3-2C2 is 10 kDa in size and was found in several species (Rotter et d . , unpublished data). Monoclonal antibodies are highly specific reagents, yet occasionally they cross-react with proteins which share only a determinant recognized by the particular monoclonal antibody studied. It is therefore possible that p53 is wholly unrelated to the cell surface molecule or the 10-kDa product, although all three bind the RA3-2C2 monoclonal antibodies. PAb122 and PAb421 were prepared by immunization of mice with syngeneic SV40-transformed cells (Gurney et al., 1980; Harlow et al., 1981, respectively). These two independently established hybridomas secrete

p53

CELLULAR-ENCODED TUMOR ANTIGEN

119

FIG. 1. Immunoprecipitation of mouse p53 molecules wit11 monoclonal and polyclonal antibodies. Equal aliquots of [35S]methionine-labeled 2M3 cell lysate were immuno-precipitated with (a) normal serum; (b) RA3-2C2; (c) PAb122; (d) PAb421; and (e)tumor-bearer (TB) serum containing anti-p53 antibodies. Samples were processed and separated by gel electrophoresis.

antibodies directed to closely associated p53 antigenic determinant(s).These antigenic sites differ completely from the RA3-2C2 determinant. The PAb122/PAb421 sites were better conserved during evolution and are found in the p53 molecules of mouse, human, and rat origin. It has been reported that transformed human cells contain p53 molecules that are structurally indistinguishable from mouse p53 (Crawford et al., 1981). More recently, however, several human specific anti-p53 monoclonal antibodies were established (Leppard and Crawford, 1983). The hamster p53 molecule represents a third category. Similar to transformed mouse cell p53, hamster p53 coimmunoprecipitated with the large T antigen expressed in SV40-transformed hamster cells (Linzer and Levine,

120

VARDA RO'lTER A N D DAVID WOLF

1979). However, all three aforementioned anti-p53 monoclonal antibodies failed to bind hamster p53. A fourth type of monoclonal hybridoina cell line, 200.47 established by Dippold et al. (1981), against the mouse p53 also recognizes the hamster p53 product. In the case of the hamster, we observed a heterogeneity of at least three species of p53 varying slightly in size (Rotter et al., unpublished data). Previously, it was observed that p53 of hamster origin and p53 of mouse origin exhibit different partial proteolytic digestion peptide maps (Linzer and Levine, 1979). Mouse and hamster p53, although of similar size, manifest different antigenic determinants and molecular structure. This implies that mouse and hamster p53 molecules are products of diverse evolutionary processes. The monoclonal antibodies described here were all prepared against p53 of tumor cells (Rotter et d., 1980; Coffman and Weissman, 1981; Gurney et al., 1980; Harlow et al., 1981; Dippold et al., 1981), yet they bind p53 determinants of transformed and nontransforined cells equally well, suggesting that antigenic determinants recognized by these anti-p53 monoclonal antibodies are species specific, rather than tumor specific (see Table I). p53 molecules of the different species seem to vary with regard to the content of the specific antigenic determinants recognized by the various established monoclonal antibodies (Rotter et al., 1983~).The fact that each species has retained a different group of antigenic determinants strongly suggests that these monoclonal antibodies bind regions in the molecule that do not lie within the functional site of the p53 molecule. It is expected that the functional regions of p53 should be conserved during evolution. V. Chromosomal Assignment of p53 Gene

The striking difference between the antigenic determinant in p53 of mouse and hamster origin made it a convenient tool for the chromosomal mapping of p53 in hamster/mouse hybrid cell lines. Analysis of mouse x Chinese hamster hybrids retaining different subsets of mouse chromosomes on a constant Chinese hamster background indicated that the active p53 gene is located on mouse chromosome 11 (Rotter et al., 1984). All of the hybrids described in Table I1 which retained mouse chromosome 11 and expressed the mouse p53 were produced from hamster E36 x mouse embryo fibroblast crosses. It is known that the level of p53 in normal cells is very low or undetectable (Mora et al., 1980; Rotter et aZ., 1981), possibly because of very rapid degradation of the protein (Oren et al., 1981). The easy detection of the mouse p53 in the hybrids indicates that some factors contributed by the Chinese hamster parent, which has the growth characteristics of a transformed line, either increased the production of, or stabilized, the mouse p53. The amount of p53 synthesized in these hybrids is

PRESENCE OF p53

IN

TABLE I TRANSFORMED A N D NONTRANSFORMED CELLSOF SEVERALSPECIES~ Antigenic determinants of p53

Species Mouse

Human

Rat

Hamster

Cell type

RA3-2C2

PAb122

PAb421

Polyclonal

Ab-MuLV-transformed lymphoid cell lines Chemically transformed fibroblasts (Meth A fibroblasts) NIH-3T3 nontransformed fibroblasts Primary embryonic fibroblasts C57BL/6 and (C3H/eb X BALB/c)FI SV80-transformed fibroblasts Daudi, Burkitt’s lymphoma Normal skin fibroblasts 45-9, Moloney sarcoma-transformed fibroblasts B77, Avian sarcoma-transformed fibroblasts MRK, Normal rat kidney cells Normal thymocytes ( L e w i s ) G&, SV40-transformed Chinese hamster fibroblasts Primary embryonic Chinese hamster fibroblasts Primary embryonic golden hamster fibroblasts

+ + + +

+ + + +

+ + +

+ + + + + + + + + + + + +

+

+

a Antigenic determinants were evaluated by immunoprecipitation of [35S]methionine-labeled cell lysates with each of the anti-p53 monoclonal antibodies or the polyclonal serum. Equal amounts of [35S]methionine-labeled proteins were compared. The amount of p53 in nontransformed cells was reduced in comparison to that observed in transformed cells. (+) denotes the presence of a radioactive p53 band observed as late as 30 days after exposure of the autoradiogram; (-) absence of radioactive p53 band.

TABLE I1 HYBRIDCLONES TESTEDFOR MURINEp53" Mouse chromosomes retained Hybrid Clone Clone Clone Clone Clone Clone Clone Clone

1 2 3 4

5 6 7 8

1

86" 100

74 68 95 41

2 93 100 80 100 100 76 89

3

4

5

6

7

8

9

43 92 100 95 100 100

48 92 93 81 70 73

75 92 93 95 82

82

93

100 93 100 90

100

96 78

86 100

96 81 75

100 78

+

+

+

10

12

13

14

15

86

96

100

100 64 86 100

79 92 100 90

17 100 95 60 47 90 89

75 100 100 100 100 94 100 100

63 75

90 100

11

+

95 100

a5 61 100 94

16

17

18

19

X

89 100 70 63 95 64 95 89

75 92

75 100 90 96 95

82 92 93 95 95 97 95 94

90 50

32 90

-

75 89

Y

pSb

+ + -

a The Chinese hamster parent of all the hybrids in the HPRT- and ouabain-resistant line E36 (Gillin et al., 1972).Hybrid clones 1-6 were isolated from crosses with BALBlc primary fibroblasts and clones 7-8 are from a hybridization with an in uiuo-maintained BALB/c sarcoma, CMS4 (DeLeo et al., 1977). The chromosome composition of the hybrids was determined by karyotyping at least 20 G-banded (Wang and Federoff, 1971)metaphases of each clone and by isozyme analysis (Nichols and Ruddle, 1973). b p 5 3 was determined by immunoprecipitation of [~S]methionipe-labeledcell lysates with anti-p53 monoclonal antibodies. In order to improve the sensitivity of p53 detection in these hybrid cell lines we performed a second immunoprecipitation step. Cell lysates were immunoprecipitated with anti-pS monoclonal antibodies. The complex of antigen-antibody was dissociated and the supernatant containing p53 was further diluted in phospholysis buffer. Anti-p53 monoclonal antibodies were added for a second incubation. cThe numbers indicate the percentage of mouse cells canying a particular mouse chromosome. d ( + ) indicates that the chromosome was rearranged and was identified only by isozymes.

p53

CELLULAR-ENCODED TUMOR ANTIGEN

123

still lower than that found in the CMS4-transformed mouse parent cells studied as controls. The elevated level of mouse p53, however, was not associated with the transformed phenotype of the chromosome ll-retaining hybrids: clone 2, for example (see Table 11), containing the mouse chromosome 11, consisted of flat, contact-inhibited, slowly growing cells. In this respect, it is interesting to note that chromosome 11 is known to be rapidly lost by mouse/hamster (E36) hybrids and is usually absent from them after a few passages (Kozak and Ruddle, 1977). It is believed that this chromosome encodes some unknown factors that are detrimental to the growth of the hybrids (Kozak and Ruddle, 1977). It would be of interest to determine whether these growth-inhibiting factors are in any way connected to the synthesis of p53. Chromosome 11 is not among the chromosomes known to be consistently affected by structural or numerical aberrations in mouse tumors (Klein, 1981), and thus changes at the chromosome level are not likely to be involved in causing the elevated levels of p53 in transformed cells. Given the almost universal presence of high levels of p53 in transformed cells and the lack of any chromosome aberration specific for the transformed state per se is therefore not surprising. This is also consistent with the notion that, at least in the case of SV40-transformed cells, posttranslational events are responsible for the increased level of p53 (Oren et al., 1981). The method we have originally used to map the functionally active p53 employed anti-p53 monoclonal antibodies. Recent studies using p53-cDNA-specific probes have confirmed our observation that the principal active p53 gene indeed maps to chromosome 11 (M. Oren, personal communication). VI. Subcellular Localization of p53 in Transformed and Nontransformed Cells

The p53 molecules of transformed and nontransformed cells are indistinguishable in their size and antigenic determinants, as evaluated by binding to several types of monoclonal anti-p53 antibodies (see Table I). The major difference between p53 in transformed and nontransformed cells appears to be quantitative; transformed cells produce higher levels of this protein than do nontransformed cells. In addition to the differences in the rate of synthesis of the protein, the p53 molecule seems to be located in different subcellular compartments in transformed and nontransformed fibroblasts. Immunofluorescent staining (Fig. 2 ) (Rotter et al., 1983a; Gurney et al., 1980; Dippold et al., 1981), as well as specific immunoprecipitation of metabolically labeled p53 molecules from various cell fractions (see Fig. 3), showed that p53 is mainly concentrated in the nuclei of transformed cells and only a minor fraction is found in the cytoplasmic compartments (Rotter et al., 1983a). In the nontransformed cells, however, p53 is detected only in

124

VARDA ROlTER A N D D A V I D WOLF

FIG. 2. Localization of p53 in transformed and nontransformed established cell lines. Cells were grown on glass coverslips and either fixed (a, b, c, d, g, h) and permeabilized or first extracted (e, f) with Triton X-100to prepare cytoskeletons prior to fixation with formaldehyde. The cells were processed for iminunofluorescence with PAb122 anti-p53 antibody. (a) Longterm primary mouse fibroblasts: (b) Swiss 3T3; (c) BA/c 3T3; (d) SV-101; (e) UV-2237 P fibrosarcoma; (0 B16-F1 melanoma; (g) Meth-A; (h) SV-101 with nonimmune serum.

I

Iz

m

FIG.3. Distribution of p53 in the subcellular fractions of a transformed established cell line. Meth A-transformed cells were labeled with [aSImethionine and fractionated into a Triton X-100 soluble (I), a cytoskeletal fraction (11),and a nuclear fraction (111). Control of total cell lysate was also included (IV). Equal amounts of radioactive proteins were immunoprecipitated with (a) normal serum; (b) RA3-2C2; (c) PAb122; (d) goat anti-Moloney serum.

SUBCELLULAR b C A L l Z A n O N

OF

TABLE 111 p53 MOLECULESIN TRANSFORMED A N D NONTRANSFORMEDFIBROBLASTS Subcellular location of p530

Fibroblast cell lines Primary embryonic fibroblasts: Passage 12 Passage 21 3T3 A31 Carcinogen-treated embryonic fibroblasts ANN-1 sv-101 Meth A BKF1 uv-2237-P

Method of transformation

Genetic origin of mouse

Cytoplasmic

-

(C57BL/6 X BALB/c)FI (C57BL/6 X BALB/c)Fi NIH Swiss BALBIc (C57BL/6 X BALBlc)Fl

+ + + +

In oitro treatment with methylcholanthrene Abelson murine leukemia virus SV40 virus In vioo treatment with methylcholanthrene Spontaneous melanoma UV-induced fibrosarcoma

2

NIH Swiss NIH Swiss BALB/c C57BL16 C3H/eb

* 2

Nuclear

-

-

+++ +++ +++ +++ +++ +++

0 p53 was estimated by both immunoprecipitation of [SSImethionine-labeled proteins with specific monoclonal antibodies from the various cell fractions and immunofluorescent staining with the same monoclonal antibodies.

p53

CELLULAR-ENCODED TUMOR ANTIGEN

127

the cytoplasmic compartment (Rotter et al., 1983a). The differential subcellular distribution of p53 molecules we detect in transformed and nontransformed cells appears to be restricted to the fibroblastic cell population. Table I11 summarizes the subcellular localization of p53 in a number of transformed and nontransformed cells (Rotter et al., 1983a). In A-MuLVtransformed lymphoid cells, however, p53 was detected by immunofluorescent staining in the cytosol fraction (Rotter et al., 1981). More recently, we observed that the great majority of p53 is easily released from these cells by even mild treatment with Triton X-100 (Rotter et al., 1983a). Accumulation of p53 in the nuclei of transformed fibroblasts may suggest a possible mechanism for the stabilization of this protein by direct binding to nuclear elements such as the chromatin. Indeed, we found that when the chromatin is solubilized by digestion with DNase and high salt, leaving behind a nuclear matrix (Berezney and Coffey, 1974; Ben-Ze’ev et al., 1982), p53 is removed together with the chromatin. Treatment with DNase alone, which does not remove the chromatin from nuclei, also does not remove p53. In the case of cells transformed by SV40, it has been suggested that the nuclear localization of the p53 molecule is mediated by its binding to the large T antigen (Lane and Crawford, 1979; Linzer and Levine, 1979; Gurney et al., 1980). However, the same nuclear localization of p53 was also found in carcinogen-treated primary fibroblasts or in the other established non-SV40transformed cell lines tested (Table 111)where no complex with other protein was detected. It is suggested, therefore, that the migration of p53 into the nuclei of transformed cells may involve a procedure leading to the stabilization and subsequent accumulation of this protein. It is not clear, however, whether the accumulation of p53 in the nucleus is the result or the cause of cellular transformation. VII. The Use of p53 as a Marker for Cell Transformation

The fact that p53 is synthesized in a wide range of tumor cells at concentrations above that of normal cells suggested the possibility of using this protein as a specific marker for tumor cells. Since the protein was detected in tissue culture cell lines and primary tumors, it was assumed that augmented synthesis of the protein was not a tissue culture artifact. The use of p53 as a marker for cancer cells could be implemented in several ways. The first is based on the direct detection of p53 protein in tumor cells by either search for the presence of biosynthetically labeled protein or by measuring steady-state amounts of this protein by radioimmunoassay. Employing a specific immunoprecipitation assay, we detected accentuated p53 synthesis in primary tumors in mice (Rotter, 1983). Table IV summarizes results obtained with several primary induced tumors in

TABLE IV p53 TRANSFORMATION-RELATED PROTEIN I N PRIMARY MOUSETUMORS" Mouse strain 2M3 cell line Tumorb 148-54 148-153 148-171 148-172 148-181 148-192 148-193 148-50 147-708 68-127-70 59-127-4 50-136-91 147-800 MT-6 38C-14 1-5

Tumor type

Method of induction

BALBlc

Pre-B leukemia

A-MuLV

C57BLl6 C57BLl6 C57BLl6 C57BLl6 C57BLl6 C57BLl6 C57BL16 C57BL/6d AKRc C57BLl6 C57BL16 C57BLJ6 AKR BALBlc C3Hleb AKR

T-cell lymphoma T-cell lymphoma T-cell lymphoma T-cell lymphoma T-cell lymphoma T-cell lymphoma T-cell lymphoma Null lymphoma Null lymphoma T-cell lymphoma T-cell lymphoma T-cell lymphoma Mammary adenocarcinoma Mammary adenocarcinoma B-cell lymphoma T-cell lymphoma

MNUA" (iv) MNUA (iv) MNUA (feed) MNUA (feed) MNUA (feed) MNUA (feed) MNUA (feed) MNUA (iv) MNUA (iv) D-RadLV (400 rads of X-irradiation) D-RadLV (400 rads of X-irradiation) A-RadLV Spontaneous Spontaneous Spontaneous Spontaneous

Latency (days)

PS

100 110

110 95 104 98 118 119 76 110 90

80 95 250 150 150 93

106 98 286

94 112 118 121 97 115 136 124 93 112 83 312 3

a p53 was determined by immunoprecipitation of a constant amount of radioactive [35S]methionine-labeled tumor cell extract with monoclonal (RA3-2C2 and PAb122) and with polyclonal anti-p53 antibodies. In each individual experiment, the intensity of the radiogram and the counts per minute of the p53 band were compared with those of the 2M3 cell lysate. p53 in each individual tumor is expressed as percentage of that found in 2M3 cell line. b Donated by N. Haran-Ghera, Department of Immunochemistry, The Weizmann Institute, Israel. c Two-month-old animals were treated with 1 mg of methylnitrosourea (MNUA) by iv injection of a saline solution or oral administration of a polyethylene glycol 400 solution. d Animals thymectomized at age 1 month.

$3

CELLULAR-ENCODED TUMOR ANTIGEN

129

mice. Measurerent of p53 by immunoprecipitation of a metabolically labeled protein seems cumbersome; therefore, it became important to develop an easy assay for the detection of this protein. A radioimmunoassay measuring the steady-state levels of p53 was recently developed by Benchimol et al. (1982). This quantitative radioimmunological solid phase assay detects p53 from mouse and human tumors (Benchimol et al., 1982). The mouse p53 assay depends on the existence of two monoclonal antibodies reacting with the different p53 determinants. The assay for human p53 detection took a different approach and made use of anti-p53 antibodies detected in sera of some breast cancer patients. Using these methods, Benchimol et al. (1982) confirmed the observation that human tumor cells contain more p53 than normal tissue. Modifications in this human p53 radioimmunoassay can be introduced in the future, due to the development of new anti-p53 monoclonal antibodies reacting specifically with antigenic sites of the human molecule (Leppard and Crawford, 1983). A slightly different approach, which again involves the measurement of steady-state p53 amounts, can be based on immunofluorescent staining of either fixed tumor cells or frozen sections of tumor tissues. We have observed that the subcellular localization of p53 varies between normal and transformed fibroblasts (Rotter et al., 1983a). Immunofluorescent staining of fixed cells with anti-p53 monoclonal antibodies have shown that p53 was mainly found in the nucleus of transformed cells, whereas in parent nontransformed enbryonic cells, p53 was detected in the cytoplasm (Rotter et al., 1983a). We suggest, therefore, that this method, based on the subcellular localization of p53, may provide a cytological diagnostic method for the detection of malignant transformed cells. A second method of using p53 as a malignant transformation marker can be approached by measuring anti-p53 antibodies in the serum of malignant tumor patients. Previously, we observed that tumor-bearing mice develop specific antibodies against p53. We found that serum obtained from mice bearing an A-MuLV-induced tumor contains antibodies against both the p120 ab oncogene product and the p53 cellular-encoded tumor antigen (Rotter et al., 1980). In accordance with these observations, Crawford et al. (1982) did a survey of serum from cancer patients and found antibodies reacting with p53 in serum of patients with primary or secondary carcinoma of the breast. Nine percent of patients tested exhibited positive titers of antip53 antibodies in their serum. No positives were detected among control sera obtained from normal women. They found that localization of the first metastasis in patients with positive sera were unusual, with more lung metastases and fewer bone metastases than expected (Crawford et al., 1982). These results, which are based on a limited number of cancer patient studies, suggest a certain correlation between the presence of anti-p53 antibodies and the manifestation of tumor growth.

130

VARDA RO'lTER A N D D A V I D WOLF

VIII. The Possible Function p53 Fulfills in Transformed Cells

One way of understanding the function p53 fulfills in either transformed or nontransformed cells can be achieved by a comparative study between variant cell lines that do or do not express this protein. In searching for such a variant cell line we came across a unique A-MuLV-transformed cell line, L12, which lacks detectable p53 protein (Rotter et al., 1981, 1983b) (Fig. 4).

FIG. 4. Absence of p53 in L12 and its presence in the 230-23-8 A-MuLV-transformed cell lines. Equal amounts of [35S]methionine-labeledproteins of L12 or 230-23-8 cell lysates were immunoprecipitated with the following antibodies: (a) normal serum; (b) RA3-2C2; (c) PAb 122; (d) PAb 421; and (e) goat anti-Moloney containing antiviral encoded p120 and anti-Moloneyrelated products.

$3

131

CELLULAR-ENCODED TUMOR ANTIGEN

I 1 c

b

s g c

J

ap

5

10 15 20 25 Days

30 35

5

10 15 20 25 30 35 Days

5 10 15 20 25 30 35 Days

FIG.5 . Development of tumors in mice. Cells (5 x 104) were washed repeatedly in PBS and injected sc in the left lower flank of syngeneic C57L mice. L12T cells were obtained about 4 weeks after cessation of TPA treatment. Tumor appearance was followed by daily monitoring of the mice. Rejection of tumor was defined as absence of any local tumor in mice previously manifesting a local tumor. (0-0),Incidence of tumor takes; (0-O), incidence of lethal tumor takes. For each cell line three groups of about 10 mice were tested. Each point therefore represents the average result of about 30 mice.

Injection of L12 cells into syngeneic mice led to the development of tumors that were subsequently rejected, whereas other A-MuLV-transformed cells, overproducing p53, induced lethal tumors (Witte et al., 1979; Rotter et al., 1983b) (Fig. 5). This suggests that expression of p53 in tumor cells may be correlated with their capacity to exhibit a fully transformed phenotype as expressed in the development of lethal tumors in mice. Figure 4 illustrates the specific [35S]methionine-labeled products detected when L12 and 230-23-8 A-MuLV-transformed cell lysates were immunoprecipitated with anti-pS3 and anti-pl2O antibodies. It is clear that L12 cells lacked a detectable p53 product, whereas A-MuLV-transformed 230-23-8 cells produced measurable quantities of p53. p53 was also absent in L12 cells labeled with [3H]-leucine or orthophosphate, radioactive isotopes which are efficiently incorporated into newly synthesized p53 molecules (Rotter et al., 1980, 1981). However, L12 cells contained a viral-encoded p120 that immunoprecipitated with specific antibodies (Fig. 4), suggesting that they had already acquired some malignant characteristics. The p120 of these cells seemed to be a functional oncogene product. Addition of radioactive y-ATP to the in uitro-immunoprecipitated molecule yielded a tyrosine-phosphorylated p120 product (Rotter et al., 1983b), as detected in other A-MuLVtransformed cell lines (Witte et al., 1980). A-MuLV-transformed cell lines

132

VAHDA HO’ITEH A N D D A V I D WOLF

were established by infecting bone marrow cells with a mixture of A-MuLV and Moloney murine leukemia virus (Mo-MuLV) stocks (Rosenberg and Baltimore, 1976). As expected, both cell lines exhibited Moloney-related products immunoprecipitating with goat anti-Moloney serum (Fig. 4e); most prominently detected were the envelope and the “gag” Moloney precursor molecules. A-MuLV-transformed lymphoid cells of BALB/c and C57L/ J origin exhibiting abundant amounts of the viral-encoded p120 and the cellular-encoded p53 are lethal in syngeneic mice. The L12 cells express a viral-encoded p12O oncogene but lack the cellular-encoded p53 protein (Rotter et al., 1980, 1983b), as well as the specific polyadenylated mRNA directing the synthesis of p53 (Wolf et al., 1984). In syngeneic mice, L12 cells developed into tumors which, in contrast to other A-MuLV transformed induced tumors, were rejected by the host. Based on these criteria, we assumed that L12 cells are partially malignant and thus may be further transformed. Tumor cell promoters were shown to increase the probability of a neoplastic transformation initiated by previous exposure of cells to a carcinogenic agent (Berenblum, 1975). These reagents were shown to enhance the transformation of cell cultures by DNA tumor viruses (Fisher et al., 1978; Yamamoto and Zur Hausen, 1979) and to induce the increase of plasminogen activator in Rous sarcoma virus-transformed fibroblasts, a phenomenon that is associated with malignant transformation (Weinstein et al., 1977). It was reported that 12-O-tetradecanoyl-phorbol ester acetate (TPA) is involved in the induction of newly synthesized proteins (Balmain, 1978; Laskin et al., 1981).We therefore decided to test whether treatment with a tumor cell promoter such as TPA would further transform the L12 cells. The results indicate that TPA treatment indeed led to the establishment of variant cells, L12T, exhibiting a lethal tumor phenotype and a stable expression of p53. L12T cells injected into syngeneic mice developed into local tumors that were lethal for the hosts (see Fig. 5). L12T cells expressed p53 molecules which, by protein analysis, seemed to be identical to those of other AMuLV-transformed cells (Rotter et al., 1983b). The viral-encoded, p120transforming Abelson protein as well as the Moloney helper virus were apparently unaffected by the TPA treatment. This is in agreement with studies showing that phorbol ester action is independent of viral and cellular src kinase levels (Goldberg et al., 1981). L12T cells exhibiting a stable p53-producing phenotype were treated with TPA for about 1 month. It is not clear, however, whether under such prolonged treatment we induced de nooo cellular production of p53 in the L12 cells or whether we selected a p53-producing subpopulation. Our preliminary studies of p53 gene organization in the L12T cells, using cDNA p53 cloned probe, favor the first possibility. Data presented here support the notion that in this group of A-MuLV-transformed cells the presence of p53 in

p53

CELLULAR-ENCODED TUMOR ANTIGEN

133

the cells correlates with their capability to develop into lethal tumors in syngeneic mice. In the course of these experiments, we followed the expression of p53; whether other cellular modification occurs in the L12T cells following treatment with TPA is being studied. IX. Molecular Mechanism Controlling the Expression of p53

Recently p53-specific cDNA was molecularly cloned (Oren and Levine, 1983), thus permitting the direct analysis of specific p53 mRNA species as well as DNA sequences. The p53 cDNA was prepared from a cDNA library constructed from an enriched p53 mRNA of a SV40-transformed mouse cell. The specific recently transcribed p53 mRNA was obtained by immunoprecipitation of polysomes with anti-p53 monoclonal antibodies. The criterion employed to identify the p53-specific cDNA clones was specific hybrid selection of p53 that translates the protein under in vitro conditions (Oren and Levine, 1983). The original cDNA clone was used to screen for additional p53 cDNA which seemed to correspond to most of the mature p53 mRNA molecules (Oren et al., 1983). Analysis of genomic DNA sequences hybridizing to p53 cDNA probes has indicated the presence of two noncontiguous p53 genes (Oren et al., 1983; Wolfet al., 1984; Zakut et al., 1983).The principal active gene was detected in a 16-kilobase (kb) EcoRI fragment. The second p53 gene residing in a 3.3kb EcoRI fragment is considered to be a nonfunctional pseudogene due to its paucity of introns, close resemblance in restriction map to the cDNA, and lack of part of the first exon (Oren et al., 1983; Zakut et al., 1983). mRNA molecules hybridizing with specific p53 cDNA probe have indicated that all p53 producer cell lines, including the nontransformed NIH-3T3 fibroblasts, contain a 2.0-kb mRNA consisting of the mature p53 coding species. Analysis of the variant L12 cells that lack detectable p53 protein have indicated the absence of the 2.0-kb mature p53-specific mRNA molecule. Instead, L12 cells displayed substantial levels of two larger polyadenylated RNA species, of approximately 6.5 and 3.5 kb, that hybridized with the p53 cDNA (p53-271) (Fig. 6). Very low amounts of a 6.5-kb mRNA were also detected in all p53 producers. Very low amounts of a 6.5-kb mRNA were also detected in all p53 producers. This product may represent a precursor mRNA molecule, which is not unexpected, since total (nuclear and cytoplasmic) RNA was used in these experiments. It is unclear whether this putative precursor is identical to the major 6.5-kb RNA found in L12 cells. The 3.5-kb mRNA hydridizing with the p53 cDNA probe is unique to the L12 cell line and is not detected in other transformed cells studied here. In order to further define the specific regions of homology between the p53-271 cDNA probe and the various polyadenylated mRNA species observed above, p53-271 DNA was restriction enzyme digested with PstI into

FIG.6. Analysis Analysis of polyadenylated polyadenylatedmRNA mRNAfrom from various variouscellular cellularsources sourcesfor forp53-specific p53-specificsequences. sequences.RNA RNAwas washybridized hybridizedto toeither either(a) (a)whole whole plasmid p53-271 containing the p53 p53 or or (b) (b)CDNA CDNA probe probe restriction restrictionenzyme enzymedigested digestedfragments fragmentsof ofp53-271 $3-271 cDNA cDNAprobe. probe.32P-labeled 32P-labeled18 18SSand and containingthe 28 28 SS ribosomal RNA RNA were were used used as as molecular molecular weight weight markers markers (M). (M).

p53 CELLULAR-ENCODED

T U M O R ANTIGEN

135

three fragments: Pst A, B and C (see Fig. 6b. The order of the PstI fragments corresponds to their 5’ and 3’ orientation within the p53-271 plasmid). These fragments were labeled and utilized for RNA hybridization. A representative result is illustrated in Fig. 6b. Clearly, the 5‘ Pst A fragment [320 base pairs (bp)] contains most of the sequences homologous to the aberrant polyadenylated mRNA unique to the L12 cell line. Furthermore, an internal XhoIPu !I fragment of the p53-271 cDNA probe (approximately 350 bp) spanning from the middle of Pst A through Pst C (see Fig. 6b) only hybridizes to the p53 2.0-kb polyadenylated mRNA. This suggests that the extreme 5’ end of p53-271 (approximately 130 bp) corresponds to the novel polyadenylated mRNA found in the L12 cell line. The fragments of Pst B (12Obp) and C (160 bp) appear to be specific for mature polyadenylated p53 mRNA. Primer extension assays have indicated that the p53-271 cDNA clone used in the above experiments is derived from the 5’ proximal region of the mature p53 mRNA molecule, suggesting that the aberrant polyadenylated p53 mRNA species produced in the L12 cell line is homologous only in the 5’ part to the normal mature polyadenylated p53 mRNA. Previously analyzed systems displayed a quantitative regulation of cellular p53 levels due to their modulation of the amount of translatable p53 mRNA (Oren d al., 1982) or to posttranslational mechanisms (Oren et al., 1981).The L12 cells represent a unique system in which p53 synthesis is completely absent, apparently due to the lack of mature p53 mRNA. Analysis of genomic DNA of the variant L12 A-MuLV-transformed cells by specific p53 cDNA probes has indicated that the principal active p53 gene contained in a 16-kb EcoRI fragment went through major alterations, leading to the appearance of a substantially larger 28.0-kb p53-specific EcoRI fragment. However, the pseudogene of these cells, contained in a 3.3-kb EcoRI fragment was intact (Fig. 7). The fact that L12 cells lack the 16-kb EcoRI fragment implies that these cells are devoid of the norma1 p53 homolog chromosome. Preliminary analyses of the altered p53 gene in L12 cells using p53 cDNA probes have mapped the rearrangement to the 5’ proximal part of this gene (Wolf et aE., 1984) (Fig. 7). Definition of the rearranged region and characterization of its nature was achieved by using genomic p53 DNA probes. Hybridization of DNA fragments with clone p8RH.2, which corresponded to the 5’ proximal part of the first intron of the active p53 gene, indicated the presence of a 12-kb DNA insert. On the basis of a detailed analysis using probes spanning throughout the whole p53 gene, we concluded that the L12 p53 gene contained all exons and the principal introns of the normal p53 16-kb gene. However, its structure was interrupted by the integration of a novel DNA insert into the noncoding sequences (Wolf and Rotter, 1984a). The structure of the p53 genes in a producer cell compared

FIG.7, Southern blot analysis of p53-speci6c sequences in A-MuLV-transformed moiise lymphoid cells. Aliquots (5 pg) of hi&-molecularwei&t DNA from A-MuJ,V-transformed mutise lymphoid cell lines L12 (a) and 2 3 0 2 3 8 (b) were electrophcrresed and hvhidimd with (I) whole clone p53-271DNA and (11) Pst A and Psl C fragments of ps-271 cDNA p 53-speciGc probe. The llindIII digests of h b d a DNA were used as molecular weight markers.

p53 CELLULAR-ENCODED

TUMOR ANTIGEN

137

to the variant L12 is illustrated in Fig. 8. Subcloning of the inserted DNA fragment and its analysis have revealed a close homology to the Mo-MuLV (Wolf and Rotter, 1984b). Further supportive evidence that the novel inserted sequence was a Mo-MuLV-like particle was based on the analysis of the nucleic acid sequence at the 5’ LTR junction region. It appeared that the Moloney particle was integrated into the first intron of the L12 p53 gene; the typical inverted repeat sequences of the Mo-MuLV LTR were retained and there was no appreciable rearrangement in the p53 intron-flanking sequence. This Moloney-like viral particle resided in a 5’ to 3’ transcriptional orientation similar to the p53 gene. A 3.3-kb region of polymorphism was detected in map units 5.4 to 5.6, bordering at the envelope gene. The modifications observed in the inserted Moloney-like particle, compared to the functional Mo-MuLV, may be attributed to a rearrangement that occurred, possibly during integration into the p53 sequences. However, it remains to be determined whether the Moloney-like insert in the p53 intron is a functional virus. Establishment of A-MuLV-transformed cell lines involved the infection of bone marrow cells with a stock containing the AMuLV and Mo-MuLV which yielded tumor cells usually expressing both viruses (Rosenberg and Baltimore, 1976). It is possible, therefore, that this initial step of viral infection, while establishing the L12 cells, included a rare event whereby Mo-MuLV was integrated into the first intron of the p53 gene. Results obtained utilizing a Moloney-specific DNA probe (Ms-4, hybridizing specifically to Mo-MuLV 4.95-5.4 map units obtained from I. Verma, The Salk Institute) have clearly indicated that L12 cells contain, in addition to the rearranged inserted Mo-MuLV, a normal size Moloney provirus which is found in most A-MuLV-transformed cells. Inactivation of cellular genes by retrovirus insertion has been reported in several instances. Kuff et al. (1983) have demonstrated the presence of an Atype particle in the intervening sequences of mutated immunoglobulin genes. Jenkins et al. (1981) have shown that insertion of retroviral genome into germ line of mice was associated with spontaneous mutations. In culture, Varmus et al. (1981) induced reversion of transformed cells by integration of Mo-MuLV into the gag or envelope region of the Avian sarcoma virus, considered as intervening sequences with respect to the src transforming gene. Furthermore, Schnieke et al. (1983)showed that integration of Moloney proviral genomes into the a (I) collagen gene led to its complete block. These observations, in agreement with our own, support the notion that modifications in the coding or intervening sequences by insertion of various DNA elements may alter eukaryotic gene expression. In summary, we have described an A-MuLV-transformed variant cell line which expresses a functional oncogene product, p120 (see Fig. 4, but lacks detectable concentrations of p53 cellular-encoded protein. It is conceivable

gene No. 2 230- 23- 8

H

BOI

( * 2 K b ) I, , ,

I

5'

PRSH I I f

I

I

PXBHSKP I I

X.P.S

If((

I

i

5'5

S

K

B

SHP

K

H

BglR

P

* *

I

1

I l l

I

I

I I

1

1

*

B 1

3'

3'bA*

L 12

75'

&PS 1 S I K1 (*ZKb)j

~

BI

SHP I l l

K1

H I

BglR I 1

PI

B 1

sp2

3' BgI

H ! / I 1

I

I

I

B 1

S K

PXBH

IU

n I Kb

FIG.8. A structural comparison of the p53 genes contained in L12 nonproducer and 230-23-8 producer A-MuLV-transformed cell lines. The structure of p 5 3 genes was deduced from genomic blot restriction enzyme analysis using p53-specific cDNA and genomic probes.

p53 CELLULAR-ENCODED T U M O R

ANTIGEN

139

that the absence of detectable amounts of this protein in L12 cells is caused by the insertion of Moloney-like particle into the first intron noncoding sequences of the principal active p53 gene. However, because the entire p53 gene remains basically intact, it is still conceivable that L12 cells may produce minute amounts of p53 that are not detected by in vivo-labeled immunoprecipitating protein or by specific p53 mRNA hybridization, thus enabling these cells to enter into the cell cycle and replicate as suggested by others (Milner and Milner, 1981; Mercer et al., 1982). It should be noted that insertion of an A particle into intervening sequences of immunoglobulin gene did not completely block protein production and the mutated cell line produced 10% of immunoglobulin normally synthesized (Kuff et al., 1983). Our future goal is to introduce a functional p53 gene into L12 cells and test whether expression of a functional p53 gene will change the phenotype of these cells from those that develop rejectable tumors in syngeneic or nude mice into cells that are lethal for the host. This approach may give a clue as to the role p53 fulfills in the cancer cell.

ACKNOWLEDGMENTS This work was supported by grants from the Israel Cancer Research Fund and the Leo and Julia Forchheimer Center for Molecular Genetics, the Weizmann Institute of Science. Varda Rotter is the incumbent of the Norman and Helen Asher Career Development Chair at the Weizmann Institute. We wish to thank Mrs. S. Admon for her excellent technical assistance and Mrs. M. Baer for her excellent editorial assistance.

REFERENCES Balmain, A. (1978). In “Carcinogenesis” (A Sivak and R. K. Bourtwell, eds.), pp. 153-172. Raven, New York. Benchimol, S., Pim, D., and Crawford, L. (1982). EMBOJ. 1, 1055-1062. Ben-Dori, R., Resnitzki, D., and Kimchi, A. (1983). FEBS Lett. 162, 84-389 Ben-Ze’ev, A., Ahulafia, R., and Alony, Y. (1982). EMBOJ. 1, 1225-1231. Berenblum, I. (1975). In “Cancer: A Comprehensive Treatise” (F. F. Becker, ed.), pp. 323344. Plenum, New York. Berezney, R., and Coffey, D. S. (1974). Biochem. Biophys. Res. Commun. 60, 1410-1417. Campisi, J., Medrano, E. E., Morreo, G . , and Pardee, A. B. (1982). Proc. Natl. Acad. Sci. U.S.A. 79, 436-440. Chandraseka, K., McFarland, V. W., Simmons, D. T., Dziadek, M., Gurney, E. G . , and Mora, P. T. (1981). Proc. Natl. Acad. Sci. U.S.A. 78, 6953-6957. Coffman, L. (1982). Immunol. Reu. 69, 5-23. Coffman, L., and Weissman, I. L. (1981).J. E r p . Med. 153, 269-279. Crawford, L. V., Pim, D. C., Gurney, E. G., Goodfellow, P., and Taylor-Papadimitriou, J. (1981). Proc. Natl. Acad. Sci. U.S.A. 78, 41-45. Crawford, L. V., Pim, D. C., and Bulbrook, R. D. (1982). Int. J. Cancer 30, 403-408. DeLeo, A. B., Shiku, H., Takahosi, T., John, M., and Old, L. J. (1977).J. E r p . Med. 146, 720734.

140

VARDA RO'ITER A N D DAVID WOLF

DeLeo, A. B., Jay, E., Appella, G. C., Dubois, C. G., Law, L. W., and Old, L. J. (1979).Proc. Natl. Acad. Sci. U.S.A. 76, 2420-2424. Dippold, W. G., Jay, G., DeLeo, A. B., Khoury, G., and Old, L. J. (1981). Proc. Natl. Acad. Sci. U.S.A. 78, 1695-1699. Fisher, P. B., Weinstein, B., Eisenberg, D., and Ginsberg, H.S. (1978). Proc. Natl. Acad. Sci. U.S.A. 75, 2311-2314. Gillin, F. D., Roufa, D. J., Beaudet, A. L., and Caskey, C. T. (1972). Genetics 72, 239-252. Goldberg, A. R., Delclos, P. B., and Blumberg, P. M. (1981). Science 208, 191-192. Gurney, E. G., Harrison, R. O., and Fenno, J. (1980).J . Virol 34, 752-763. Hanafusa, T., Wang, L.-H., Anderson, S. M., Karess, R. E., Hayward, W. S. and Hanafus, H. (1980). Proc. Natl. Acad. Sci. U.S.A. 77, 3009-3013. Harlow, B., Crawford, L. V., Pim, D. C., and Williamson, N . H. (1981).J . Virol39, 861-869. Hayward, W. S . , Neel, B. G., and Astrin, S. M. (1981). Nature (London)290, 475-479. Jay, G., DeLeo, A. B., Appella, E., Dubois, G. C . , Law, L. W., Khoury, G., and Old, L. J. (1980). Cold Spring Harbor Symp. Quant. B i d . 44, 659-664. Jay, G . , Khoury, G., DeLeo, A. B., Wolfgang, G . , Diffold, W. G., and Old, L. J. (1981).Proc. Natl. Acad. Sci. U.S.A. 78, 2932-2936. Jenkins, N. A., Copeland, N. G., Taylor, B. A,, and Lee, B. K. (1981). Nature (London) 293, 370-374. Jornvall, H., Luka, J . , Klein, G., and Appella, E. (1982). Proc. Natl. Acad. Sci. U.S.A.79,287291. Klein, G . (1981). Nature (London)294, 313-318. Kozak, C. A,, and Ruddle, R. F. (1977). Somatic Cell Genet. 3, 121-133. Kress, M., May, E., Cossing-Ena, R., and May, P. (1979).J . Virol. 31, 472-483. Kuff, E. L., Feenstra, A., Lueders, K., Smith, L., Hawley, R., Hozumi, N., and Shulman, M. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 1992-1996. Lane, D. P., and Crawford, L. V. (1979). Nature (London) 278, 261-263. Laskin, F. D., Mufson, R. A., Piccinini, L., Engelhardt, D. H., and Weinstein, I. B. (1981). Cell 25, 441-449. Leppard, K., and Crawford, L. (1983). EMBOJ. 2, 1457-1464. Linzer, D. I. H., and Levine, A. J. (1979). Cell 17, 43-52. Luka, J., Jornvall, H., and Klein, G. (1980).J . Virol. 35, 592-602. McCormick, F., and Harlow, E. (1980).J . Virol. 34, 213-224. McCormick, F., Clark, R., Harlow, E . , and Tjian, R. (1981). Nature (London) 292, 63-65. Melero, J. A., Tur, S., and Carroll, R. B. (1980). Proc. Natl. Acad. Sci. U.S.A. 77, 97-101. Mercer, W. E., Nelson, D., DeLeo, A. B., Old, L. J., and Baserga, R. (1982). Proc. Natl. Acad. Sci. U.S.A. 79, 6309-6312. Milner, J., and Milner, S. (1981). Virology 112, 785-788. Mora, P. T., Chandrasekaran, K., and McFarland, W. (1980). Nature (London) 288, 722-724. Mor , P. T., Chandrasekaran, K., Hoffman, J. C., and McFarland, V. W. (1982). Mol. Cell. Biol. 2, 763-771. Nicols, E. A., and Ruddle, F. H. (1973).J . Histochem. Cytochem. 21, 1066-1068. Oren, M., and Levine, A. J. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 56-59. Oren, M . , Malzman, M., and Levine, A. J. (1981). Mol. Cell. Biol. 1, 101-110. Oren, M., Reich, N. C., and Levine, A. J. (1982). Mol. Cell. B i d . 2, 443-449. Oren, M., Bienz, B., Givol, D., Rechavi, G., and Zakut, R. (1983). E M B O ] . 2, 1633-1639. Parda, L. F., Tabin, C. J., Shih, C., and Weinberg, B. (1982). Nature (London)297, 474-479. Reynolds, F. H., Sacks, T. L., Dedsagkar, D. H., and Stephenson, J. R. (1978). Proc. Natl. Acad. Sci. U.S.A. 75, 3974-3978. Rosenberg, N., and Baltimore, D. (1976). J . Exp. Med. 143, 1453-1463.

$3

CELLULAR-ENCODED TUMOR ANTIGEN

141

Rotter, V. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 2613-2617. Rotter, V., Witte, 0. N., Coffman, R., and Baltimore, D. (1980).J . Virol. 36, 547-555. Rotter, V., Boss, M. A,, and Baltimore, D. (1981).J . Virol. 38, 336-346. Rotter, V., Abutbul, H., and Ben Ze’ev, A. (1983a). EMBOJ. 2, 1041-1047. Rotter, V., Abutbul, H., and Wolf, D. (1983b). Znt. J. Cancer 31, 315-320. Rotter, V., Friedman, H., Katz, A., Zerivitz, K., and Wolf, D. (1983~). J . Zmmunol. 181, 329333. Rotter, V., Wolf, D., Pravtcheva, D., and Ruddle, F. H. (1984). Cell Mol. Biol. (in press). Santos, E., Tronick, S . , Aaronson, S. A., Pulciani, S., and Barbacid, M. (1982). Nature (London) 298, 343-347. Sarnow, P., Ho, Y. S., Williams, J., and Levine, A. J. (1982). Cell28, 387-394. 304, 315-320. Shen Ding-We, Real, F. X., DeLeo, A. B., Old, L. J . , Marks, P. A., and Ritkind, R. A. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 5919-5922. Simmons, D. T., Martin, M. A., Mora, P. T., and Chang, C. (1980). J . Virol. 34, 650-657. Varmus, H . E., Quintrell, N., and Ortiz, S. (1981). Cell 25, 23-26. Wang, H. C., and Federoff, S. (1971). Nature (London)New Biol. 235, 52-54. Weinstein, I. B., Wigler, M., and Pietropaolo, C. (1977).In “Origins ofHuman Cancer” (H. H. Hiatt, J . D. Watson, and J. A. Weinstein, eds.), pp. 751-772. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York. Witte, 0. N., Rosenberg, N., Paskin, M., Shields, A,, and Baltimore, D. (1978). Proc. Natl. Acad. Sci. U . S . A . 75, 2488-2492. Witte, 0. N., Rosenberg, N., and Baltimore, D. (1979).J . Virol. 31, 776-784. Witte, 0. N . , Dasgupta, A., and Baltimore, D. (1980). Nature (London) 283, 826-831. Wolf, D., and Rotter, V. (1984a). Cold Spring Harbor Symp. Cell Prol$ Cancer Cells 2,403409. Wolf, D., and Rotter, V. (1984b). Mol. Cell. Biol. 4, 1402-1410. Wolf, D., Admon, S., Oren, M., and Rofter, V. (1984). Mol. Cell. Biol. 4, 552-558. Yamamoto, N., and Hausen, H., zur (1979). Nature (London) 280, 244-245. Zakut-Houri, R., Oren, M., Bienz, B., Lavie, V., Hazum, S., and Givol, D. (1983). Nature (London) 306, 594-597.

This Page Intentionally Left Blank

MONOCLONAL ANTIBODIES REACTIVE WITH BREAST TUMOR-ASSOCIATED ANTlG ENS Jeffrey Schlorn,' David Colcher,* Patricia Horan Hand,' John Greiner,' David Wunderlich,' Maureen Weeks,' Paul B. Fisher,t Philip Noguchi,* Sidney Pestka,§ and Donald Kufe" Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Monoclonal Antibodies to Human Breast Carcinomas Generation and Characterization of Monoclonal Antib Differential Reactivity of a Monoclonal Antibody (DF3) with Human Malignant versus Benign Breast Tumors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. Use of Radiolabeled Monoclonal Antibodies for Tumor Detection ... VI. Use of Monoclonal Antibodies in Tumor Therapy ........................... VII. Antigenic Modulation and Evolution within Human Carcinoma Cell Populations VIII. Use of Biological Response Modifiers to Enhance Detection of Human Carcinoma Antigens by Monoclonal Antibodies ............................. IX. Future Directions ........................ .......... References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I. 11. 111. IV.

143 144 147 151 155 160 161 166 169 171

1. Introduction

The advent of hybridoma technology in 1975 represented a quantum leap in the field of tumor immunology (Kohler and Milstein, 1975). For the first time, single cell populations of B lymphocytes from immunized hosts could be immortalized via fusion with drug-selected nonimmunoglobulin secretor murine myeloma cells. These fused cell products could now be cloned and propagated indefinitely. The supernatant fluids of cultures from literally thousands of different cloned cell populations (termed hybridomas) could now be assayed to select for homogeneous populations of immunoglobulins with the desired reactivity. As a direct result of this technology, numerous monoclonal antibodies (MAbs) have been generated that have led to the identification of novel tumor-associated antigens (TAAs)from various human carcinomas, melanomas, leukemias, and lymphomas. To date, well over 100 MAbs against human carcinomas have been reported in the literature; each one has been characterized to a unique degree with respect to range of reactivity and reactive antigen. An attempt will be made here to identify and *Laboratory of Tumor Immunology and Biology, National Cancer Institute, National Institutes of Health, Bethesda, Maryland, tDepartment of Microbiology, Cancer Center/Institute of Cancer Research, Columbia University College of Physicians and Surgeons, New York, New York, $Office of Biologics Research and Review, Center for Drugs and Biologics, Food and Drug Administration, Bethesda, Maryland, BRoche Institute of Molecular Biology, Nutley, New Jersey, and **Dana Farber Cancer Institute, Boston, Massachusetts.

143 Copyright 0 1985 by Academic Press, Inc. ADVANCES IN CANCER RESEARCH, VOL. 43

All rights of reproduction in any form reserved. ISBN 0-12-006643-2

144

JEFFREY SCHLOM ET AL.

TABLE I POTENTIAL USESFOR MAbs I N THE MANAGEMENT OF MAMMARY CANCER Diagnosis of primary and metastatic lesions Blood tests for tumor-associated antigens Nuclear scanning with radiolabeled MAb IV to detect distal metastases Lymphoscintigraphy to detect regional node involvement Immunohistochemistry for detection of occult tumor cells Aspiration cytology Lymph nodes and bone marrow Pleural effusions Treat ine n t Direct cytotoxity of MAb Mediated via complement Cell mediated Drug conjugation of MAb (e.g.. adriamycin) Toxin conjugation of MAb (e.g., ricin) Radionuclide conjugation of MAb (e.g., alpha or beta emitters) Prognosis (tumor typing) Immunohistopathology on sections of primary tumor masses Possible projections based on tumor antigenic phenotype Metastatic potential of cell populations Specific sites of metastasis Response (or lack of) to specific therapeutic regimens

briefly describe the approximately two dozen most characterized monoclonals reactive with human breast tumors. Several of these will be chosen for further description to elaborate on those phenomena that are most relevant for the use of MAbs in the management of human breast cancer, as well as in the study of the biology of human mammary carcinoma cell populations. As can be seen in Table I, these reagents may ultimately find use in virtually all arms of patient management. The progress being made in many of these areas as well as the problems that have been encountered will be discussed below. 11. Monoclonal Antibodies to Human Breast Carcinomas

Numerous MAbs that are reactive with human mammary carcinomas have been described in the literature. In general, they can be classified into four groups based on the immunogen used to generate the MAb. These include using (1) breast tumor cell lines, (2) milk fat globule membrane, and (3) membrane-enriched extracts of breast carcinoma metastases as immunogen, or (4)lymph nodes from mastectomy patients. Each of the MAbs thus far

MONOCLONAL ANTIBODIES TO BREAST TAAS

145

described (see Table 11), including those prepared by several different groups to milk fat globule membrane, appears to be unique with respect to percentage of reactive mammary tumors, percentage of reactive cells within tumors, location of reactive antigen within the tumor cell, or degree of reactivity with nonmammary tumors as well as normal tissues. Among the monoclonal antibodies that have been developed using the human mammary carcinoma cell lines is monoclonal antibody 10-3D-2 which is reactive to five of five breast carcinoma cell lines. However, it is also reactive to neoplastic lung, colon, placenta, and melanoma cells (Soule et al., 1983). Other monoclonal antibodies developed in the same manner (Table 11) include C11, G3, and H7, made to a 24,000-MW protein purified from a TABLE I1 MONOCLONALANTIBODIESREACTIVE WITH H U MA N BREAST CARCINOMA Immunogen

MAb

Breast cancer cell lines

MBr1,2,3

Unknown

24.17.1 24.17.2 F36l22 M7/105 C11, G3, H7 H59

95K lOOK Unknown

Milk fat globule membranes

Mammary carcinoma metastases

Human lymph nodes

a

Reactive antigen

24K Unknown

10-3D-2 HMFGl HMFG2 3.15.C3 M3, M8 M18, M24 67-D- 11 B72.3

Unknown 220-400K glycoprotein

B6.2 B1.l DF3 MBE6

90K Glycoprotein 180K Glycoprotein 290K Glycoprotein Unknown

CLNH5 HmAbl,2 NDa

Unknown Unknown Unknown

ND, No designation given.

126K Glycoprotein High-molecular-weight glycoprotein Unknown

Reference Menard et 01. (1983); Canevari et al. (1983) Thompson et al. (1983) Papsidero et al. (1983) Croghan et al. (1983) Adams et al. (1983) Hendler et al. (1981); Yuan et al. (1982) Soule et al. (1983) Taylor-Papadimitriou et al. (1981) Arklie et al. (1981) Foster et a!. (1982) Rasmussen et al. (1982) Colcher et al. (1981a); Nuti et al. (1982); and Colcher et al. (1983a) Colcher et al. (1981a, 1983a) Colcher et al. (198313) Kufe et a!. (1984) Schlom et al. (1980); Wunderlich et al. (1981); and Teramoto et al. (1982) Glassy et al. (1983) Imam et 01. (1984) Cote et al. (1983)

146

JEFFREY SCHLOM ET AL.

mammary carcinoma line. These antibodies are reactive with the 24K estrogen-regulated protein (positive in 56 of 94 cytosol preparations from breast tumors) and to cytosol preparations from the MCF-7 cell line (Adams et al., 1983). Three monoclonal antibodies with distinct reactivities were produced using delipidated human milk fat globule membrane as immunogen (Table 11). Monoclonal antibody HMFGl reacts with seven of eight breast cancer cell lines and with primary carcinomas, but not with metastases of breast carcinoma. Monoclonal antibody HMFG2 demonstrates an overall specificity for breast epithelium of primary carcinoma and lactating breast; it is also reactive with metastases to lymph nodes and adenocarcinomas of the lung, uterus, and ovary (Epenetos et al., 1982a,b). Antibody 3.15.C3 is reactive with two human breast carcinoma cell lines and with the LS-174T colon carcinoma cell line (Taylor-Papadimitriou et al., 1981; Arklie et al., 1981). Other monoclonal antibodies made to human milk fat globule membranes include M3, M8, M18, M24, and 67-D-11 (Arklie et al., 1981; Foster et al., 1982; Rasmussen et al., 1982; Epenetos et al., 1982a,b; see Table 11). An extensively characterized group of MAbs reactive with human breast carcinomas is that generated using membrane-enriched or crude extracts of metastatic breast lesions as immunogens. The rationale for generating these antibodies was to utilize immunogens with determinants that would be present on metastatic human mammary carcinoma cells. One hypothesis that was considered was that primary tumor masses may contain a subpopulation of cells with a predefined metastatic potential; the possibility may therefore exist that antibodies made against determinants present on the vast majority of breast carcinoma cells in primary lesions may not be present on cells in metastatic lesions. Similarly, the use of mammary tumor cell lines as immunogens was avoided. The hypothesis set forth here is that cell lines are the products of great selective pressure on cell populations found in vivo; the antigenic phenotype of those cells selected for their ability to grow in culture may therefore be quite different from the antigenic phenotype of those cells in primary or metastatic masses. Multiple assays using tumor cell extracts, tissue sections, and live cells in culture must be employed to reveal the range of reactivities and diversity of each monoclonal antibody generated (Colcher et al., 1981a, 1983b,c; Nuti et aZ., 1982; Nuti et al., 1981; Horan Hand et al., 1983). Table I1 lists four of the antibodies generated from breast metastases. Monoclonal antibody B72.3, which is reactive with a novel 220,000-400,000 high-molecular-weight glycoprotein complex, demonstrates an extremely strong selective reactivity for carcinoma versus normal adult tissues. Monoclonal antibody B72.3 reacts with 50% of breast carcinomas and over 80% of colon carcinomas. It will be discussed in detail below. Monoclonal antibody B1.1 was generated using a breast metastasis as

MONOCLONAL ANTIBODIES TO BREAST TAAS

147

immunogen, and it has been shown to be reactive with carcinoembryonic antigen (CEA). Monoclonal antibody B6.2 is reactive with a 90K glycoprotein that is usually, but not always, coordinately expressed with CEA. Monoclonal antibody D F 3 (Table 11)has recently been characterized and appears to be reactive with a differentiation antigen since mammary carcinomas display a cytoplasmic reactivity, whereas benign breast lesions have the reactive antigen concentrated on apical borders; this will be discussed in detail later. Several groups have also used lymph nodes from mastectomy patients to generate human MAbs (Table 11). The rationale for these studies was that breast cancer patients may contain B lymphocytes in axillary lymph nodes that have been primed by antigens shed by breast carcinoma cells. These nodal B lymphocytes, obtained sterilely at the time of mastectomy and nodal dissection, were fused with either murine nonimmunoglobulin secretor myeloma cells or human myeloma cells to produce human monoclonal antibodies. To date, several human antibodies have been generated that show varying degrees of selective reactivity for human tumor versus normal tissue (Table 11). The major obstacles using this methodology have been (1) the apparent low affinity of many of MAbs generated, most being of the IgM class, (2) the shut off of human immunoglobulin production from humanmouse hybrid cells, and (3) the lack of a suitable nonimmunoglobulin secretor human myeloma fusion partner. Many of the innovations possible in this line of research will be discussed later. Ill. Generation and Characterization of Monoclonal Antibodies

It is important to consider those factors involved in the selection and characterization of any MAb. The steps involved in the generation of murine MAbs to breast carcinoma will therefore be used as a prototype. In a typical set of experiments, mice were immunized with membrane-enriched fractions of human metastatic mammary carcinoma cells, in this case from either of two involved livers from two different patients (Colcher et al., 1981a). Spleens from immunized mice were then fused with nonimmunoglobulinsecreting NS-1 murine myeloma cells to generate 4250 primary cultures. All hybridoma methodology and assay methods have been described in detail elsewhere (Colcher et al., 1981b; Herzenberg et al., 1978). Supernatant fluids from hybridoma cultures were first screened in solid phase radioimmunoassays (RIAs) for the presence of immunoglobulin reactive with extracts of metastatic mammary tumor cells from involved livers and not reactive with extracts of apparently normal human liver. Following passage and double cloning by endpoint dilution of cultures secreting immunoglobulins demonstrating preferential reactivity with breast carcinoma cells, the MAbs

148

JEFFREY SCHLOM ET AL.

from 11 hybridoma cell lines were chosen for further study. The isotypes of all 11 antibodies were determined: 10 were IgG of various subclasses and 1 was an IgM. The 11 MAbs could be divided into three major groups by solid phase RIAs based on their differential reactivity to the two metastases of different patients. All 11 antibodies were negative when tested against similar extracts from normal human liver, a rhabdomyosarcoma cell line, the HBL-100 cell line derived from cultures of human milk cells, mouse mammary tumor and fibroblast cell lines, disrupted mouse mammary tumor virus and mouse leukemia virus, purified CEA, and ferritin. Two monoclonal antibodies were used as positive controls in all these studies: (1)W6/32, an antihuman histocompatibility antigen (Barnstable et aZ., 1978), and (2) B139, generated in our laboratory against a human breast tumor metastasis, which demonstrates reactivity to all human cells tested. To determine whether the monoclonals bound cell surface determinants, each antibody was tested for binding to live cells in culture, i.e., established cell lines of human mammary carcinomas. The various MAbs could be further separated into three different groups on the basis of their differential binding to cell surface determinants. None of the 11 monoclonal antibodies, however, bound to the surface of sarcoma or melanoma cell lines, nor to the surface of 24 cell lines derived from apparently normal human tissues (Colcher et d . , 1981a; Kufe et d . , 1983). Control monoclonals W6/32 or B139 did bind all of these cells. To further define the specificity and range of reactivity of any MAbs, immunohistochemical techniques, e.g., the immunoperoxidase method, should be employed on tissue sections. This technique can be used as it rapidly distinguishes between reactivity to tumor tissue versus reactivity to normal cellular counterparts from the same patient. Furthermore, numerous normal tissue types in the body can be assayed which are otherwise untestable by conventional means. In this case, all the MAbs reacted with mammary carcinoma cells of primary mammary carcinomas, both infiltrating ductal and lobular. The percentage of primary mammary tumors that were reactive varied for the different monoclonals. In many of the positive primary and metastatic mammary carcinomas, not all tumor cells stained. Both of these phenomena will be discussed in detail below. A high degree of selective reactivity with mammary tumor cells-and not with apparently normal mammary epithelium (except in the area of the tumor)-stroma, blood vessels, or lymphocytes of the breast was also observed. Two MAbs, B72.3 and B6.2, were chosen for further study since they appeared to be noncoordinately expressed in different breast tumor cells and reacted with approximately 50 and 75 percent, respectively, of 39 formalin-fixed, infiltrating ductal carcinomas tested via the immunoperoxidase technique. The major cross-reactivity to normal tissues of monoclonal B6.2 was subsets of circulat-

149

MONOCLONAL ANTIBODIES TO BREAST TAAS

ing polymorphonuclear leukocytes. Monoclonal B72.3 demonstrates the most selective degree of reactivity for tumor tissues of any MAb thus far generated in that no reactivity with adult normal human tissues has been detected. However, fetal tissue has not yet been carefully examined nor has absolutely every tissue type in the body. It would be naive to assume that any antigenic determinant consisting of a few amino acids would be expressed only in carcinoma cells and at no time during development in the embryo or during various stages of cell differentiation within the spectrum of adult tissues. Monoclonal B72.3 (an IgG, reactive with a 220-4OOK glycoprotein complex) has been used for studies to determine the effect of antibody concentration on the staining intensity and the percentage of tumor cells stained in immunoperoxidase assays. Since one cannot titrate antigen in the fixed tissue section, an antibody dilution experiment may be performed to give an indication of the relative titer of reactive antigen within a given tissue. As seen in Table 111, a 500-fold range of antibody concentration, varying from 0.02 to 10 p g of purified B72.3 immunoglobulin (per 200 pl) per tissue section, was used on each of four mammary carcinomas from different patients. The results demonstrate that (1)different mammary tumors may vary in the amount of the antigen detected by B72.3, (2) a given mammary tumor may contain tumor cell populations which vary in antigen density, and (3) some mammary tumors may score positive or negative depending on the dose of antibody employed. These studies point out a concept that must be kept in mind if tissue sections are to be used in prospective or retrospective prognostic studies using monoclonal antibodies. As seen in Table 111, when 4 TABLE 111 DOSEOF MONOCLONALANTIBODYB72.3 vs REACTIVITY OF HUMAN MAMMARY CARCINOMA CELLSI N IMMUNOPEROXIDASE ASSAY Tumor staining intensityb

B72.3 (pg)"

Tumor 1

Tumor 2

Reactive tumor cells (8) Tumor 3

Tumor 4 ~~

10

4 2 1 0.2 0.02

1 + (90) 2+(10)

3+(100)

3+(80)

NEG

1+(5)

2+(100) 1+(80) NEG NEG NEG

3+(80) 3+(70) 3+(70) 2+(50) 2+(30)

NEG NEG NEG NEG NEG

NEG NEG NEG NEG

A dose of0.02 pg of B72.3 is equivalent to a 1:1OO,OOO dilution of B72.3 produced in mouse ascites fluid. Staining intensity: 1 + , weak; 2+, moderate; 3 + , strong; NEG, negative.

150

JEFFREY SCHLOM ET AL.

FIG.1. Immunoperoxidase staining of fixed tissue sections of primary and metastatic mammary carcinomas of four different patients with monoclonal antibody B72.3. (A) Infiltrating ductal

carcinoma: at the center of the field is a negative large normal duct (N) surrounded by positively . Infiltrating ductal carcinoma; note the intense staining infiltrating tumor cells (T). ~ 5 4 (B) membrane and faint cytoplasmic staining of the tumor cells (T). The broad arrow indicates a negative tumor cell flanked by positive tumor cells. x540. (C) In situ element (T)ofan infiltrat. ing ductal carcinoma; note t h e stroma and lymphocytes (L) which are negative. ~ 1 3 0 (D) Breast tumor metastasis in the pleura. This is an example of the focal pattern of staining: Intense stain is concentrated in the cytoplasm of tumor cells (T). The stroma (S) is negative. X330.

MONOCLONAL ANTIBODIES TO BREAST TAAS

151

pg of B72.3 is used, Tumor 4 appears to have an antigenic phenotype different from that of the other three. However, when 1 p g of B72.3 is used in the assay, Tumor 3 now appears to have the unique phenotype. Thus, studies on tissue sections should always include antibody titrations and should be evaluated on the basis of dose of antibody employed. The immunoperoxidase technique has also been used to test a variety of malignant, benign, and normal mammary tissues. Using 4 p g of monoclonal per slide, the percentage of positive primary breast tumors was 46%(19/41); 62% (13/21) of the metastatic lesions scored positive (Nuti et al., 1982). Several histologic types of primary mammary tumors scored positive; these were infiltrating ductal (Fig. 1A and B), infiltrating lobular, and comedo carcinomas. Many of the in situ elements present in the above lesions also stained (Fig. 1C). None of the six medullary carcinomas tested was positive. Approximately two-thirds of the tumors that showed a positive reactivity demonstrated a cell-associated membrane and/or diffuse cytoplasmic staining (Fig. lB), approximately 5% showed discrete focal staining of the cytoplasm (Fig. lD), and approximately one-fourth of the reactive tumors showed an apical or marginal staining pattern. Metastatic breast carcinoma lesions that were positive occurred in axillary lymph nodes and at the distal sites of skin, liver, lung, pleura (Fig. lD), and mesentery. Fifteen benign breast lesions were also tested; these included fibrocystic disease, fibroadenomas, and sclerosing adenosis. Only 2 of these 15 specimens showed positive staining: one case of fibrocystic disease where a few cells in some ducts were faintly positive and a case of intraductal papillomatosis and sclerosing adenosis with the majority of cells staining strongly. Monoclonal B72.3 was also tested against normal breast tissue from noncancer patients and showed no reactivity. Several nonbreast tissues were tested and thus far show little or no degree of reactivity; these included uterus, liver, spleen, lung, bone marrow, colon, stomach, salivary gland, and lymph node. IV. Differential Reactivity of a Monoclonal Antibody (DF3) with Human Malignant versus Benign Breast Tumors

Monoclonal antibody DF3 was generated using a membrane-enriched fraction from a human metastatic mammary carcinoma (Kufe et al., 1984). The molecular weight of the reactive antigen is approximately 290,000 and it is found on the surface of mammary carcinoma cells. While MAb DF3 is reactive with the vast majority of several histologic types of malignant and benign human breast tumors, it shows a strong differential reactivity to cytoplasmic antigen in carcinomas versus reactivity to antigen concentrated on apical borders in benign breast lesions (Kufe et al., 1984).

152

JEFFREY SCHLOM ET AL.

Several histologic types of human malignant mammary carcinomas were examined for reactivity with MAb DF3 using the ABC immunoperoxidase method and 5-pm sections of formalin-fixed tissues. Monoclonal DF3 was reactive with 78% of 32 infiltrating ductal carcinomas (IDC), with the proportion of tumor cells staining within each of the 25 IDCs ranging from a few to over 90% (Fig. 2). Furthermore, the staining pattern for these carcinomas was primarily cytoplasmic (Fig. 2). A similar staining pattern was observed for mammary carcinomas containing both infiltrating ductal and intraductal elements. The percentage reactive cells varied for individual tumors, and the infiltrating elements of each tumor displayed a higher degree of cytoplasmic reactivity than the intraductal component. Similar staining patterns were observed with in situ, medullary, and infiltrating lobular carcinomas. The three infiltrating lobular carcinomas and one medullary carcinoma showed only cytoplasmic staining. In contrast, the in situ carcinomas had varying degrees of cytoplasmic and apical border reactivity. In contrast to the cytoplasmic reactivity observed with the malignant breast tissues, MAb DF3 reacted principally with the apical borders of all benign breast lesions. A similar pattern of reactivity was observed with a lactating breast tissue. Approximately 10-20% of normal ducts from breast tumor patients also showed a slight apical staining. However, none of five

loo

r

A

6

h60

a

"t

100

FIG.2. Reactivity of MAb D F 3 with 5-pm sections of infiltrating ductal carcinomas using the ABC immunoperoxidase method. Each bar represents a different patient. The dark areas represent the percentage of tumor cells with a cytoplasmic reaction; the striped areas represent the percentage of tumor cells displaying MAb DF3 reactivity on apical borders.

MONOCLONAL ANTIBODIES TO BREAST TAAS

153

3

FIG. 3. Reactivity of MA11 DF3 with 5-pm sections of benign mammary tumors using the immunoperoxidase method. Panel A represents five fibroadenomas, and Panel B represents eight fibrocystic disease specimens. The dark areas represent the percentage of tumor cells with a cytoplasmic reaction; the striped areas represent the percentage of tumor cells displaying MAb DF3 reactivity on apical borders.

fibroadenomas (Fig. 3A), only one of eight fibrocystic disease specimens (Fig. 3B), and no normal duct cells showed any evidence of cytoplasmic MAb DF3 reactive antigen. In scoring the reactivity (cytoplasmic and apical) of MAb D F 3 with human breast tumors, 87% of 52 malignant lesions and 100% of 13 benign lesions were positive. If reactivities were scored on the basis of only cytoplasmic staining, however, 78%of malignant lesions were positive, whereas only 1 of 13 benign lesions displayed this pattern of reactivity. Primary and metastatic breast carcinomas from three patients were also examined for reactivity with MAb DF3; all 14 metastases to axillary lymph nodes and distal sites were positive. The pattern of reactivity for distal metastatic disease was uniformly cytoplasmic as was observed with the primary breast carcinomas. One exception was one lung metastasis obtained from one patient which also revealed some degree of apical staining. The reactivity of MAb DF3 with breast tumor metastasis is demonstrated in Fig. 4. Figure 4A shows the strongly staining breast carcinoma cells metastatic to

FIG.4. Reactivity of MAb DF3 with metastatic mammary carcinoma lesions. (A) Strong cytoplasmic reactivity with a mammary tumor metastasis (T) in the ovary; S is stroma; F is follicular cyst, x 130. (B) Positive metastatic mammary tumor cells (T) in the bone marrow (M). X220.

MONOCLONAL ANTIBODIES TO BREAST TAAS

155

ovary. Figure 4B reveals strongly reactive metastatic disease in the bone marrow of the same patient. Since the immunogen used to prepare MAb DF3 was a membrane-enriched fraction of a human mammary carcinoma metastatic to liver, we compared reactivity of this antibody with the similarly prepared MAbs B6.2, B72.3, and B1.l. Monoclonal B72.3 is reactive with a 220-400K glycoprotein complex, whereas MAb B1.l is reactive with the 180K glycoprotein CEA. In contrast, MAb D F 3 reacts with a 290K antigen detectable in metastatic tumor but not in uninvolved liver. Furthermore, there was no correlation in terms of reactivity of MAbs DF3 and B72.3 with 20 primary breast carcinomas. Similarly, there was noncoordinate expression of antigens reactive with MAbs DF3, B6.2, and B1.1 in 10 primary breast carcinomas (Kufe et al., 1984). It was of further interest to determine whether the immunoperoxidase staining of human breast tumors with MAb DF3 is related to cell surface expression of the reactive antigen. The reactivity of MAb DF3 with MCF-7 breast carcinoma cells was demonstrated using indirect immunofluorescence. MCF-7 and the BT-20 mammary carcinoma cells were also assayed for surface antigen using a live cell radioimmunoassay. Monoclonal DF3 bound to the surface antigen using a live cell radioimmunoassay. Monoclonal DF3 bound to the surface of both cell lines. In contrast, there was no evidence of MAb DF3 binding to the surface of two human colon carcinoma cell lines (Kufe et al., 1984). The above results have shown that the DF3 antigen is present in high levels on apical borders of differentiated secretory mammary epithelial cells and in the cytosol of less differentiated cells. This finding could have prognostic significance. For example, the prognostically more favorable in situ carcinomas demonstrated predominantly apical reactivity with MAb DF3. Tubular breast carcinomas, another pathologic category of favorable outcome, also reacted with MAb DF3 along apical borders. Immunoperoxidase staining of large numbers of primary breast tumors as a correlate of diseasefree internal state and survival probability will now be required to resolve the prognostic value of reactivity with this antibody. V. Use of Radiolabeled Monoclonal Antibodies for Tumor Detection

Radioactively labeled antibodies from hyperimmune sera have been used to detect the presence of tumors in both experimental animals and humans by gamma scintigraphy. The majority of the antibodies used for clinical trials have been constituents of goat or rabbit antisera and were directed against antigens such as CEA (Ege et al., 1980; Pettit et al., 1980; Kim et al., 1980; Deland et al., 1979). The studies using anti-CEA antibodies have demon-

156

JEFFREY SCHLOM ET AL.

strated the localization of malignant breast tumors (Deland et al., 1979; Goldenberg et al., 1980a,b, 1981; Primus and Goldenberg, 1980; Mach et al., 1980); some of these immunoglobulins have been partially purified using affinity chromatography with an increase in iinmunoreactivity of the IgG (Primus and Goldenberg, 1980). With the development of the hybridoina technology, homogenous populations of defined MAbs to TAAs can now be utilized in radioimmunolocalization studies. Several groups have recently attempted to use radiolabeled monoclonal antibodies generated against a variety of immunogens including colon carcinomas, melanomas, milk fat globules, and carcinoembryonic antigen to localize tumors in patients bearing a variety of tumors. Some of these monoclonal antibodies exhibit a broad range of reactivity but still show some promise as imaging agents (Epenetos et al., 1982a). Other antibodies with a more restricted range of reactivity have been successfully used to localize melanomas and colon carcinomas (Larson et al., 1983; Mach et al., 1983; Moldofsky et al., 1983). The majority of these early clinical localization studies were performed using antibodies radiolabeled with 1311, and some work is being done with 1231,ll1In, and et al., 1983). g g m T(Morrison ~ There are numerous parameters that must be considered in the effective localization of tumors using radiolabeled MAbs; a summary is presented in Table IV. Virtually none of the parameters listed in this table have been analyzed in clinical trials; moreover, few have even been carefully evaluated in a well-controlled experimental model. Thus, the use of radiolabeled MAbs in tumor localization can be considered to be in its infancy. Below is a summary of several studies that have recently been conducted to help evaluate some of the crucial parameters outlined in Table IV for the use of radiolabeled MAbs in the localization of human breast tumors. Monoclonal antibodies can be utilized in breast cancer patients in either lymphoscintigraphy, to detect mammary tumor lesions in nodes of the axilla and internal mammary chain, or to detect distal metastases. In the studies described here, MAb B6.2, which may be useful in lymphoscintigraphy procedures, was chosen as a prototype. B6.2 IgG was purified and F(ab’), and Fab’ fragments were generated by pepsin digestion. All three forms of the antibody were radiolabeled and assayed to determine their utility in the radioimmunolocalization of transplanted human mammary tumor masses (Colcher et al., 1983~).The IgG and its fragments were labeled with lZ5I using the Iodogen method to specific activities of 15 to 50 pCi/pg. The labeled antibody was shown to bind to the surface of live breast tumor cells and retained the same specificity as the unlabeled antibody. Better than 70% of the antibody remained immunoreactive in sequential saturation solid phase RIAs after labeling.

MONOCLONAL ANTIBODIES TO BREAST TAAS

157

TABLE IV FACTORS IN THE USE OF MONOCLONALANTIBODIES FOR TUMOR LOCALIZATlON A N D THERAPY Size of the tumor mass Number of cells expressing the reactive antigen in the tumor mass Number of antigen molecules per cell surface Degree of vascularization of the tumor Degree of infiltration and necrosis in the tumor mass Presence and reactivity of circulating antigen in the blood Duration of MAb binding to cell surface Avidity of MAb Fate of antigen-antibody complex Internalization Capping Shedding Isotype of immunoglobulin (IgC subtypes or IgM) Whole immunoglobulin or fragments: Fab, Fab’, F(ab’)z Clearance of MAb from blood Excretion Reticuloendothelial system Dose of MAb used Route of inoculation of MAb Species of MAb (i.e., mouse, rat, human) Cocktails of MAbs If radiolabeled MAb used Choice of isotope (half-life and energy) Ability of MAb to be labeled with a specific isotope Specific activity of radiolabeled MAb Atfinity of radiolabeled MAb (e.g., radiolabeling at the binding site will reduce affinity) Depth of tumor from body surface (for localization only) Time of scanning (for localization only) Half-life of isotope Clearance rate of MAb from blood Time bound to tumor

Athymic mice bearing transplantable human mammary tumors were injected with 0.1 pg of 125I-labeled MAb B6.2. The ratio of radioactivity/milligram of tissue in the tumor compared to that of various tissues rose over a 4-day period in the liver, spleen, and kidney at day 4. Ratios of the counts in tumors to those found in the brain and muscle were greater than 50:l and as high as 11O:l. When the mammary tumor-bearing mice were injected with ‘=I-labeled F(ab’), fragments of B6.2, higher tumor to tissue ratios were obtained. The tumor to tissue ratios in the liver and spleen were 15-2O:l at 96 hr. This was probably due to the faster clearance of the F(ab’), fragments as opposed to the IgG. Athymic mice bearing a human melanoma, a tumor that

158

JEFFREY SCHLOM ET AL.

shows no surface reactivity with B6.2, were used as controls and were negative for nonspecific binding of the labeled antibody or antibody fragments to tumor tissue. Similarly, no localization was observed when either normal murine IgG or MOPC-21 IgG, (the same isotype as B6.2 from a murine myeloma) or their F(ab’), fragments were used. Athymic mice bearing mammary tumors were also injected with 1251-labeledB6.2 Fab’. The clearance rate of the Fab’ fragment was considerably faster than the larger F(ab’), fragment and the intact IgG. Acceptable tumor to tissue ratios were obtained, but the fast clearance rate resulted in a large amount of the labeled Fab’ being found in the kidney and bladder, and in low tumor to kidney ratios. These studies therefore indicate that F(ab’), fragments may be superior to Fab’ or intact IgG in some radiolocalization protocols. Studies were undertaken to determine whether the localization of the 1251-labeledantibody and fragments in the tumors was sufficient to be de. mice bearing tected using a gamma scanner (Colcher et al., 1 9 8 3 ~ )Athymic the mammary tumor or the human melanoma were injected intravenously with approximately 30 pCi of 12”I-labeled B6.2 IgG. The mice were scanned and then sacrificed at 24-hr intervals. The mammary tumors were easily detected at 24 hr (Fig. 5) using radiolabeled B6.2 IgG; a small amount of activity was detectable in the blood pool. The tumor remained strongly positive over the 4-day period with background activity decreasing until it was barely detectable at 96 hr (Fig. 5). The 0.5-cm-diameter tumors localized in Fig. 5A and B appear bigger than their actual size; this is due to

FIG.5. Gamma camera scanning of athymic mice bearing transplanted human tumors using 1BI-labeled 86.2 IgG. Athymic mice hearing a transplantable human mammary tumor [Clouser, (A) and (B)] or a human melanoma [A375 (C)] were inoculated with approximately 30 pCi of ‘25I-labeled B6.2 IgG. The mice were scanned after various time intervals [(A, C) 24 hr; (B) 96 hr], until an equal number of counts were detected in each field.

MONOCLONAL ANTIBODIES TO BREAST TAAS

159

FIG.6. Gamma camera scanning of athymic inice bearing transplanted human tumors using 1251-labeled B6.2 F(ab')z. Athymic mice bearing a transplantable human mammary tumor [Clouser, (A) and (B)] or a human melanoma [A375, (C)] were inoculated with approximately 30 FCi of '2sI-labeled B6.2 F(ab'),. The mice were scanned after various time intervals [(A,C), 24 hr; (B), 96 hr] until an equal number of counts were detected in each field (T, tumor; K, kidney; B, bladder).

the dispersion of rays through the pinhole collimator. No tumor localization was observed using radiolabeled B6.2 IgG in mice bearing the control human melanoma transplants (Fig. 5C). Human-tumor-bearing mice were also injected with 1251-labeled B6.2 F(ab'), fragments. The mice cleared the fragments faster than the intact IgG. A significant amount of activity was observed in the kidneys and bladder at 24 hr (Fig. 6A), but tumors were clearly positive for localization of the lZ5Ilabeled B6.2 F(ab'), fragments. The activity was cleared from the kidneys and bladder by 48 hr. The tumor to background ratio increased over the 4day period of scanning, resulting in low background levels; good tumor localization was observed at 96 hr (Fig. 6B). No localization of activity was observed with the radiolabeled B6.2 F(ab'), fragments in the athymic inice bearing the A375 melanoma (Fig. 6C). 1251-Labeledantibody and fragments have all been shown to be successfully localized, with tumor localization using the F(ab'), fragments giving the highest overall tumor to tissue ratios. Fab' or F(ab'), fragments may eventually prove more useful because of the potential problem of Fc receptors on a variety of cells binding the labeled IgG and yielding a higher nonspecific distribution of the antibody. The use of an antibody without the Fc portion should also reduce its immunoreactivity in patients and thus minimize an antimurine immunoglobulin immune response. Smaller fragments also clear the body faster than intact immunoglobulin and should

160

JEFFREY SCHLOM ET AL.

therefore result in a lower whole body radiation absorbed dose to the host and provide greater “signal to noise” ratios. Radiolabeled monoclonal antibodies that are reactive with the surface of human mammary carcinoma cells may eventually prove useful in several areas in the management of human breast cancer. The detection of occult metastatic lesions at distal sites via gamma scanning could serve as an adjunct in determining which patients should receive adjuvant therapy, and subsequent scanning could reveal which tumors respond to therapy. At present, only axillary lymph nodes removed at mastectomy are examined for tumor involvement for use in staging; the extent of nodal tumor involvement in the internal mammary chain is not determined. The use of radiolabeled monoclonal antibodies in lymphoscintigraphy of the internal mammary chain may thus eventually increase the reliability of staging of nodal involvement as a prognostic indicator. Along with their potential in the diagnosis and prognosis of breast cancer, monoclonal antibodies coupled with isotopes decaying via high-energy transfer with short-range radiation kill several cell diameters, therefore, only one cell of a cluster of tumor cells need express a target antigen. This approach could be quite useful in light of the heterogeneity of antigen expression in some tumors. Furthermore, a monoclonal antibody coupled to an isotope need not be internalized by the tumor cell for its therapeutic potential to be realized. This approach, however, is obviously quite complex and would first require extensive studies in an experimental system. VI. Use of Monoclonal Antibodies in Tumor Therapy

As outlined in Table I, several modes of presentation of MAbs can be considered for cancer therapy. They include (1)MAbs administered alone and thus aided via complement- or cell-mediated mechanisms, (2) drug- or toxin-conjugated MAbs or (3) MAbs conjugated with killer isotopes. The use of MAbs in tumor therapy becomes quite complex when one considers just some of the parameters involved (see Table IV). Monoclonal antibodies directed against human carcinomas have been coupled with the toxin ricin (A chain) and the drug daunomycin (Gilliland et a l . , 1980; Krolick et al., 1981; Belles-Isles and Page, 1983).All three conjugated MAbs have been used in in uitro to show selective killing of carcinoma cells in culture. To date, phase I clinical trials have been conducted with carcinoma patients only with one MAb (MAb 17-1A) directed against colon cancer. Most patients received a single injection of the IgG,, MAb in a dose range of 15 to 1000 mg per patient. No untoward immediate or delayed reaction to the initial injection was observed in any of the patients. Mouse immunoglobulin circulated in patients’ blood for 2-50 days depending on the dose of MAb

MONOCLONAL ANTIBODIES TO BREAST TAAS

161

injected and was detected in tumor tissue within 1 week after its administration. Eight of 9 patients who received 366- to 1000-mg doses of MAb did not develop antimouse antibodies, whereas 8 of 9 who received less than 200 mg antimouse immunoglobulin antibody did. Three out of this heterogeneous group of 20 patients had no detectable disease at 10, 12, and 22 months after immunotherapy. These are obviously preliminary results, but demonstrate in carcinoma patients (1)the relative lack of toxicity of MAbs, (2)the problem of development of human antimurine Ig, and (3) that larger doses of MAb may be less immunogenic than smaller doses. VII. Antigenic Modulation and Evolution within Human Carcinoma Cell Populations

Heterogeneity among tumor cells within carcinoma masses must have been noticed by those who were among the first to utilize microscopic techniques to examine such specimens. In 1954, Foulds formally documented the existence of distinct morphologies in different areas of a single mammary tumor (Foulds, 1954). Since then, several investigators have reported the occurrence of heterogeneity in a variety of tumor cell populations (Kerbel, 1979; Hart and Fidler, 1981). Using a variety of methods and reagents including heterologous antisera, heterogeneity has also been observed with respect to the antigenic properties of various tumor cell populations (Kerbel, 1979; Hart and Fidler, 1981; Prehn, 1970; Pimm and Baldwin, 1977; Miller and Heppner, 1979; Poste et al., 1981). Monoclonal antibodies have now been used to (1) define the extent of antigenic heterogeneity and modulation of specific TAAs that exists among human carcinomas as well as within a given carcinomatous mass; (2) determine some of the parameters that mediate the expression of various antigenic phenotypes; and (3) develop model systems in which to study and perhaps eventually control these phenomena. Recent studies have revealed a wide range of antigenic phenotypes present in mammary carcinoma masses (Horan Hand et al., 1983). For example, different mammary carcinomas have been shown to differ in their pattern of staining with a given MAb. Table V shows the result of reacting 39 different infiltrating ductal carcinomas with 4 MAbs. As can be seen, 10 antigenic phenotypes emerge, ranging from those that express all 4 antigens to those that express none. Correlations of antigenic phenotype with biologic behavior of individual tumors are now being attempted by several groups. Patterns of reactivity of a MAb with different tumors may also differ. These may include focal staining (representing dense foci of TAA in the cytoplasm), diffuse cytoplasmic staining, or membrane apical staining (representing a concentration of TAA on the luminal borders of cells).

162

JEFFREY S C H L O M ET AL.

TABLE V DIFFERENTIAL REACTIVITY OF MONOCLONAL ANTIROUIILS WIT11 D I F I > E H E N T INFILTHATIN(: DUCTAL MAMMARY CARCINOMAS“ Tumor phenotype

Number of patients”

5 3 1 11 1 6 1 2 5 -4 39

Group A Group B Group C Group D Group E Group F Group G Group H Group I Group J Total ~~

~~

Monoclonal antibody

B72.3

B1.l

B6.2

+ + + NEG +

+ +

+ + + + + + NEG +

NEG NEG NEG NEG NEG

NEG

+ NEG + + NEG NEG NEG

NEG NEG

B38.1

+ NEC,

+ +

NEC, NEG NEG NEG

+

NEG

~~~

Serial sections of formalin-fixed mammary tumors were tested for expression of TAAs detected by monoclonal antibodies using the immunoperoxidase method. Tumors were scored positive if antigen was present on any carcinoma cells. b Number of patients with tumor specimens displaying the indicated pattern of reactivity with monoclonal antibodies. a

Phenotypic variation has also been observed in the expression of TAAs within a given tumor. One pattern sometimes observed was that one area of a tumor contained cells with TAAs reactive with a particular MAb, whereas another area of that tumor contained cells that were unreactive with the identical antibody (Fig. 7A). A more common type of antigenic heterogeneity was observed among cells in a given area of a tumor mass. This type of antigenic diversity, termed “patchwork,” reveals tumor cells expressing a specific TAA directly adjacent to tumor cells negative for the same antigen (Fig. 1B). Patterns of reactivity with a specific monoclonal antibody were also observed to vary within a given tumor mass, i.e., antigen was detected in the cytoplasm of cells in one part of the tumor mass and on the luminal edge of differentiated structures in a different part of the same mass. In an attempt to elucidate the phenomenon of antigenic heterogeneity in human carcinomas, model systems have been examined. The MCF-7 human mammary tumor cell line has been tested for the presence of TAAs using the cytospin/immunoperoxidase method (Nuti et al., 1982; Horan Hand et al., 1983). This cell line was shown to contain various subpopulations of cells as defined by variability in expression of TAA reactive with several MAbs, i.e., positive MCF-7 cells were seen adjacent to cells that scored negative (Fig. 7B). Studies were conducted to determine if the antigenic heterogeneity

FIG. 7. Heterogeneity of antigenic expression of TAAs. In (A), an infiltrating ductal mammary carcinoma was reacted with monoclonal antibody B6.2 using the immunoperoxidase technique. Note the stained (T) and unstained (t) tumor cells. Normal mammary cells (N) do not react with the antibody. X130. In (B), using the cytospinlimmunoperoxidase technique as described (Nuti et al., 1982; Horan Hand et al., 1983), MCF-7 cells were stained with monoclonal antibody B6.2. x540.

164

JEFFHEY SCHLOM ET AL.

observed in MCF-7 breast tumor cells was (1) the result of at least two stable genotypes or phenotypes, (2) the reflection of a modulation of cell surface antigen expression of a single phenotype, or (3) both phenomena. In experiments designed to monitor cell surface antigen expression in different phases of cell growth, it was observed that MCF-7 cells, at contact inhibition, expressed less antigen on their surface as detected by monoclonal B6.2 than cells in active proliferation. Using fluorescence-activated cell sorter analyses, it has been shown that this particular MAb is most reactive with the surface of MCF-7 cells during S phase of the cell cycle (Fig. 8). This may explain the immunohistochemical observations of TAA positive cells adjacent to TAA negative cells. Studies were also undertaken to determine if any change in antigenic phenotype occurs during extended passage of cells in culture. The BT-20 human breast cancer cell line was serially passaged and assayed at each passage level during logarithmic growth. While cell surface HLA antigen expression was detected at all passage levels, the antigen detected by MAb B6.2 was not evident after each passage. This phenomenon was repeatedly observed in several experiments with antigen presentation disappearing at approximately the same passage levels. A similar phenomenon was also observed in other carcinoma cell lines. As a result of these antigenic changes observed after passage in culture, the MCF-7 mammary tumor cell lines obtained from four sources were examined for the presence of several cell surface TAAs. Karyotype and isoenzyme profiles of the four lines were tested

DNA CONTENT

FIG. 8. Analysis of effects of cell cycle on the expression of cell surface tumor-associated antigens. MCF-7 cells were grown to confluency and 24 hr after passage were stained with Hoechst dye (5 pglml for 1 hr at 37°C). Approximately 1-2 x lo6 cells were treated with 0.1 ml of monoclonal antibody B6.2 at a concentration of 0.2 pg/ml and then reacted with 0.1 ml of a 1:40 dilution of goat antimouse fluorescein isothiocyanate. Cells were then analyzed by tlow cytometry and indirect immunofluorescence (Kufe et al., 1983).

yirn& : 165

MONOCLONAL ANTIBODIES TO BREAST TAAS

a - '

-

y

6-

z -

i

4:

2-

-

'

' '

'

I

A--

I '

'

I

B--

'

' I

I

'

"C-

ng IMMUNOGLOBULIN

FIG.9. Reactivity of monoclonal antibodies with the surface of the parent MCF-7 mammary adenocarcinoma cell line and cloned MCF-7 cell populations. Using a live cell radioimmunoassay, increasing amounts of monoclonal antibodies B1.l (O), B6.2 (H),and B72.3 were tested for binding to the parent MCF-7 cell line (A), MCF-7 clone 6F1 (B), and MCF-7 clone 10B5 (C).

A),

and were all identical and characteristic of the MCF-7 cell line. Using a live cell RIA capable of detecting the reactivity of antigens at the cell surface, antigenic profiles of the four MCF-7 cell lines were determined. Using three MAbs, four discrete antigenic phenotypes emerged for the four MCF-7 cell lines (Horan Hand et al., 1983). This observation should serve as a caveat to investigators who are utilizing established cell lines in their studies and attempting to correlate their results with those of other laboratories. To further understand the nature of antigenic heterogeneity of human mammary tumor cell populations, MCF-7 cells were cloned by endpoint dilution and 10 different clones were obtained and assayed for cell surface TAAs (Horan Hand et al., 1983). As seen in Fig. 9, the parent MCF-7 culture reacts most strongly with monoclonal antibody B1.1 and least with monoclonal B72.3. Clone 6F1 (Fig. 7B) exhibits a similar phenotype to that of the parent (Fig. 9A). At least three additional major phenotypes were observed among the other clones. For example, clone 10B5 was devoid of detectable expression for any of the antigens assayed (Fig. 9C), although it does express normal levels of HLA (30). To determine the stability of the cell surface phenotype of the MCF-7 clones, each line was monitored through a 4-month period and assayed during log phase at approximately every other passage. While some of the MCF-7 clones maintained a stable antigenic phenotype throughout the observation period, a dramatic change in antigenic phenotype, i.e., antigenic evolution, was observed in others (Fig. 10). Antigenic variability of TAAs among and within human mammary tumor cell populations thus presents a potential problem in the development and optimization of immunodiagnostic and therapeutic procedures for carcinomas. The studies described here now enable one to demonstrate the

166

JEFFREY SCHLOM ET AL.

6

5 "

4

2

e 2 1

---

A

I

l

r

l

Is

I

1

1

B

500 100 20 4 0.80.16 O.(

500 100 20 4 0.8 0.16 0.03 500 100 20 4 0.8 0.16 0.03 ng IMMUNOGLOBULIN

FIG. 10. Detection of TAAs in an MCF-7 cloned population with continued cell passage. Using a live cell RIA to detect cell surface antigen expression, increasing amounts of monoclonal antibodies 91.1 (A), B6.2 (B), and B72.3 (C) were tested for binding to MCF-7 clone 10B5 at passage 9, (B);passage 12, and passage 15, passage 6, (0);

(A);

(v).

extent of specific antigenic variability that may exist within human carcinoma cell populations. The variability observed in cell surface expression of some antigens at different stages of the cell cycle indicates that the patchwork antigenic heterogeneity observed in tissue sections may reflect, in part, cells in different phases of growth. The studies involving the MCF-7 cell clones indicate that some stable antigenic phenotypes may exist, but that other clones exhibit an antigenic evolution, i.e., the gradual expression of specific cell surface TAAs with extended passage. One can thus hypothesize that the antigenic phenotype of a given tumor mass in situ, as defined by specific MAbs, may differ at the time of detection versus later stages of tumor progression. This concept has serious implications if monoclonal antibodies are to play a useful role in the diagnosis, prognosis, and therapy of breast cancer. VIII. Use of Biological Response Modifiers to Enhance Detection of Human Carcinoma Antigens by Monoclonal Antibodies

The high degree of antigenic heterogeneity that exists in many human mammary carcinoma masses may compromise the effectiveness of some MAbs for the detection and/or therapy of such lesions. In fact, if MAbs are to be used successfully for the in situ therapy of human breast cancer, the phenomenon of antigenic heterogeneity must be dealt with so that most, if

MONOCLONAL ANTIBODIES TO BREAST TAAS

167

not all, cells of a tumor mass express a given TAA. One approach would be to investigate whether well-defined substances that (1)alter states of cell differentiation and (2) have potential clinical applicability may also possess the capability of modifying the expression of cell surface TAAs. Native interferon (Attallah et al., 1979; Imai et al., 1981; Grant et al., 1982; Fellous et al., 1982) and recombinant human leukocyte interferon (IFN-aA) (Fisher et al., 1983; Pestka, 1983) meet such criteria. Because of its homogeneity, stability, and extensive degree of characterization, recombinant human leukocyte (clone A) interferon (IFN-aA) (Pestka, 1983; Staehlin et al., 1981a,b) has recently been evaluated for its ability to enhance the detection of TAAs on the surface of human carcinoma cells by MAbs (Greiner et al., 1984). Recombinant human leukocyte interferon was first titrated on the human mammary and colon carcinoma cell lines for its ability to enhance the detection of tumor-associated cell surface antigens (Greiner et al., 1984). Ten, 100, or 1000 U/ml of IFN-aA was shown to enhance the detection of several TAAs in a dose-dependent manner. These were the 90K, 180K, and 220400K antigens detected by MAbs B6.2, B1.l, and B72.3, respectively. Studies were then carried out to determine if various concentrations of IFN-aA would induce the expression of these antigens on normal as well as noncarcinoma neoplastic cells not normally expressing these surface antigens. Three such cell lines were chosen for these studies: WI-38 (normal human embryonic lung), Flow 4000 (normal human embryonic kidney), and A375 (human melanoma). All three cell lines were previously shown to be negative for the expression of the three TAAs (Colcher et d , 1981a, 1983) and also remained so following exposure to various concentrations (10- 1000 U/ml) of IFN-aA. However, IFN-aA did increase the expression of the normal cell surface antigens such as HLA. This latter finding confirms the previously reported observation of the enhanced expression of HLA on the surface of human melanoma cells by partially purified leukocyte interferon (Attallah et al., 1979; Braylon et al., 1982). The enhanced TAA expression on the surface of human breast and colon carcinoma cells mediated by recombinant human leukocyte interferon could be due to a variety of cellular and/or molecular changes that include (1)the increased expression of TAAs on a subpopulation of cells already expressing the TAAs, (2) the induction of the expression of a given TAA on a population of cells not previously expressing the antigen, (3) a change in the surface areas of the cell, or (4) a combination of any of these phenomena. To explore such possibilities MAb B1.1 was reacted with WiDr human colon carcinoma cells in the presence or absence of 1000 U/ml IFN-aA, and the cells were analyzed with a cell sorter. The background analysis of WiDr cells in the absence of addition of MAb B1.l, with or without IFN-aA, is shown in Fig. 11A. Figure 11B shows a heterogeneous population of cells, reacting with

168

JEFFREY SCHLOM ET AL.

A

1

20

40

60

80

100

a z n

FLUORESCENCE INTENSITY FIG.11. Flow cytometric analyses of MAh B1.1 binding the surface of WiDr cells following treatment with IFN-aA. Each figure is a three-dimensional isometric display of DNA content (y axis), fluorescence intensity, i.e., cell surface 180K CEA expression as reflected by MAb B1.1 binding (x axis), with the z axis representing the number of cells. WiDr cells (107) were incubated for 24 hr at 37°C with or without 1000 Ulml IFN-aA. The cells were then harvested, washed, and incubated in medium containing 2 pg MAh B 1 . l per 106 cells. After incubation the cells were washed and incnhated with fluorescein-conjugated sheep antimouse antibody. The cells were washed, centrifiiged at 500 g and resuspended at a concentration of 106 cellsl0.3 nil. The cells were then fixed, pelleted, resuspended, and stained for 4 hr at room temperature, 106 cells/ml in PBS with propidiiim iodide and ribonuclease A. The stained cells were analyzed on

MONOCLONAL ANTIBODIES TO BREAST TAAS

169

MAb B 1 . l , depicted as a spectrum of fluorescence intensities (x-axis) representing the expression of various levels of the cell surface 180K protein. Following a 24-hr incubation with 1000 U/ml IFN-aA, however, a dramatic shift is observed (Fig. 11C) in both the percentage of cells expressing the 180K antigen (vertical z axis) and the fluorescence intensity per cell (x axis). No difference in cellular DNA content (Fig. 11, y axis) or relative size of the individual WiDr cells was observed after interferon treatment. Thus, following the addition of IFN-aA, computer analysis revealed both an increase in mean fluorescence intensity per cell and binding of MAb by more than 98% of tumor cells. These findings demonstrate that the increase of TAA expression induced by recombinant IFN-aA is a result of both more antigen being expressed per cell and a “recruitment” of new cells to express the antigen. One potential clinical application of MAbs is the use of radiolabeled immunoglobulins to detect micrometastases in regional nodes and at distal sites, and a major consideration of radiolocalization studies is the reduction in the amount of radiolabeled MAb required to bind and detect a given tumor mass in situ. The use of lower levels of radioactively labeled MAb should increase signal to noise ratios and thus make detection of smaller lesions more efficient. A reduction in the amount of MAb required to give an equally efficient signal for surface binding of a MAb could be facilitated by prior or concomitant treatment with interferon. Interferon would therefore have potential clinical applications for the in situ detection of carcinoma lesions with radiolabeled MAbs or in the use of MAbs for immunotherapy. The ability of biologic response modifiers such as recombinant human leukocyte interferon to somewhat selectively enhance the expression of monoclonal-defined TAAs in carcinoma cell lines could also eventually prove useful in defining the role of specific TAAs in the expression of the transformed phenotype. IX. Future Directions

At least two dozen MAbs to human mammary tumors have now been generated and characterized (Table 11).Without exception all of these MAbs an Ortho Cytofluorograf System 50H with blue laser excitation of 200 mW at 488 nrn. Under these conditions PI bound to nuclear DNA fluoresces red, whereas surface iinmunofluorescence bound to the 180K antigen fluoresces green. Data from 25,000 cells were stored on an Ortho Model 2150 computer system and used to generate the figures shown here. (A) WiDr cells stained for DNA content (with or without IFN), but with no MAb B1.1. (B) WiDr cells stained for both nuclear DNA and MAb B 1 . l binding. (C) WiDr cells treated with 10oO U IFN-aA for 24 hr and stained as in (B).

170

JEFFREY SCHLOM ET AL.

have thus far revealed a “pancarcinoma” reactivity, i.e., binding to broad classes of carcinomatous tissues. Future studies will no doubt focus on attempts to develop reagents with selective reactivity toward breast tumors only. Also, it is now clear that rarely does one MAb react with all carcinomas from a particular organ site. Thus, the need for “cocktails” or combinations of MAbs in many situations is becoming obvious. One approach toward this end is the examination of biopsy material with several MAbs via immunohistochemical methods to define those specific MAbs that are most reactive with a given patient’s tumor. If one considers the use of MAbs for tumor therapy, two major problem areas have emerged from studies over the past several years: (1) the antigenic heterogeneity of many mammary carcinoma masses and (2) the development of human antimurine immunoglobulin immune response. The problem of heterogeneity may be dealt with at three levels: (1)cocktails of MAbs may be used so that all turnor cells are reactive with at least one of the MAbs; (2) biologic response modifiers, such as recombinant interferon, which alter states of differentiation of tumor cells may be employed; or (3) “killer” isotopes such as a or p emitters coupled to MAbs may be used. Since some of these isotopes can kill 5-cell diameters, a MAb need bind only those tumor cells demonstrating cell surface TAA expression to also kill those adjacent tumor cells not expressing the antigen. The use of radiolabeled MAbs in tumor therapy also has the advantage that a MAb need not enter a cell to kill it. It has been shown that some of the carcinoma-associated antigens detected are stable components of the cell membrane, thus, antigen-antibody complexes formed at the surface were shown not to internalize (Kufe et al., 1983). The problem of patients receiving murine MAbs may also be dealt with at several levels. First, fragments of immunoglobulins devoid of Fc regions may be less immunogenic in patients than whole immunoglobulin. Since relatively low toxicity has been associated with the administration of MAbs, perhaps initial inoculations of high doses of MAbs will be shown to induce “immune tolerance” in patients. Finally, human MAbs are expected, but not guaranteed, to be less immunogenic than murine MAbs. To date, several groups have developed human MAbs using human lymph nodes from mastectomy patients fused with either murine or human myeloma cells. As expected, most or all of these appear to be of low affinity since they were derived from patients with no active immunization. Thus, when protocols are developed involving active specific immunization of carcinoma patients, human MAbs (developed from peripheral blood lymphocytes) may be an extremely valuable “spin off.” Genetically engineered human MAbs should also be considered. These could be constructed by first isolating the gene coding for the binding region

MONOCLONAL ANTIBODIES TO BREAST TAAS

171

of a specific murine MAb (based on amino acid analysis of the Fab region). This gene would then be linked with previously isolated human genes encoding the remainder of the immunoglobulin molecule. This recombinant hybrid molecule could then be propagated in either procaryotic or eucaryotic systems. The MAbs thus far developed against human breast tumors have led to a more systematic study of this disease. For the first time, one now has tools other than morphology or the presence of hormone receptor to distinguish between two different tumors or among different cell populations within a given tumor mass. The number of these tumor “fingerprints” is now in the process of being expanded. Many new genes including “onc” genes have now been identified in a variety of human carcinomas. The knowledge of the DNA sequence of these genes has led to the production of synthetic peptides reflecting these sequences. Monoclonal antibodies directed against these extremely specific immunogens will undoubtedly provide insight into the conditions governing their expression in various cell types. Conversely, the MAbs thus far developed are now being used to purify large quantities of previously unattainable tumor-associated antigens. These are being examined for amino acid composition so that their encoding genes can be identified and analyzed. Only when this is achieved will one be better able to understand some of the basic mechanism underlying mammary carcinoma cell initiation and progression.

REFERENCES Adarns, D. J., Hajj, H., Edwards, D. P., Bjercke, R. J., and McGuire, W. L. (1983). Cancer Res. 43, 4297-4301. Arklie, J., Taylor-Papadimitriou, J., Bodmer, W., Egan, M., and Millis, R. (1981). Int. J. Cancer 28, 23-29. Attallah, A., Needy, B., Noguchi, P., and Elisberg, B. (1979). Znt.J. Cancer 24, 49-55. Barnstable, C. J., Bodmer, W. F., Brown, G . , Galfre, G., Milstein, C., Williams, A. F., and Ziegler, A. (1978). Cell 14, 9-20. Belles-Isles, M., and Page, M. (1983).Znt. J . Zmmunophamcol. 3, 97-102. Braylon, R., Bensen, N., Nourse, V., and Kouth, H. (1982). Cytometry 2, 337-343. Canevari, S., Fossati, G., Balsari, A,, Sonnino, A,, and Colnaghi, M. 1. (1983). Cancer Res. 43, 1301-1305. Colcher, H., Horan Hand, P., Nuti, M., and Schlom, J. (1981a). Proc. Natl. Acad. Sci. U . S . A . 78, 3199-3203. Colcher, D., Horan Hand, P., Teramoto, Y. A , , Wunderlich, D., and Schlom, J. (1981b). Cancer Res. 41, 1451-1459. Colcher, D., Hand, P., Wunderlich, D., Nuti, M., Teramoto, Y., Kufe, D., and Schlom, J. (1983a). In “Monoclonal Antibodies and Cancer” G . L. Wright, Jr., ed. Dekker, New York. Colcher, D., Horan Hand, P., Nuti, M., and Schlom, J. (1983b). Cancer Inoest. 1, 127-138. Colcher, D., Zalutsky, M., Kaplan, W., Kufe, D., Austin, F., and Schlom, J. (1983~).Cancer Res. 43, 736-742.

172

JEFFREY SCHLOM ET AL.

Cote, R. J., Morrissey, D. M . , Houghton, A. N., Beattie, J . R., Oettgen, E. J., and Old, L. J. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 2026-2030. Croghan, G . A., Papsidero, L. D., Valenzuela, L. A,, Nemoto, T., Penetrante, R., and Chu, T. M. (1983). Cancer Res. 43, 4980-4988. Deland, F. H., Kim, E. E., Corgan, R. L., Casper, S., Primus, F. J , , Spremulli, E., Estes, N., and Goldenberg, D. M. (1979).J. Nucl. Med. 20, 1243-1250. Ege, G . N., Bronskell, M. J . , Deland, F. H., Kim, E. E., and Goldenberg, D. M. (1980).J . Nucl. Med. 21, 804-807. Epenetos, A. A , , Britton, K. E., Mather, J . , Shepherd, J., Granowska, M., Taylor-Papadimitriou, J., Nimnon, C. C., Durbin, H., Hawkins, L. R., and Malpas, J. S. (1982a). Lancet Nov. 6, 1000-1004. Epenetos, A. A,, Canti, G., Taylor-Papdimitriou, J., Curling, M., and Bodmer, W. F. (1982b). Lancet Nov. 6, 1004-1006. Fellous, M., Nir, U . , Wallach, D., Merlin, G., Rubinstein, M., and Revel, M. (1982). Proc. Natl. Acad. Sci. U.S.A. 79, 3082-3086. Fisher, P. B., Miranda, A. F., Babiss, L. E., Pestka, S., and Weinstein, I. B. (1983). Proc. Natl. Acad. Sci. U . S . A . 80, 2961-2965. Foster, C. S . , Dinsdale, E. A,, Edwards, P. A. W., and Neville, A. M. (1982). Virchows Arch. Pathol. Anat. 394, 295-305. Foulds, L. (1954). Cancer Res. 14, 327-339. Gilliland, D. G., Steplewski, Z., Collier, R. J., Mitchell, K. F., Chang, T. H., and Koprowski, H. (1980). Proc. Natl. Acad. Sci. U.S.A. 77, 4539-4543. Glassy, M . C., Handley, H. H . , Hagiwara, H., and Royston, I. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 6327-6331. Goldenberg, 1).M., Kim, E. E . , Deland, F. H., Bennett, S., and Primus, F. J. (1980a).Cancer Res. 40, 2984-2992. Goldenberg, D. M., Deland, F. H., and Kim, E. E. (1980b).N . Engl. J . Med. 303, 1237-1242. Goldenberg, D. M . , Kim, E. E., and Deland, F. H. (1981). Proc. Natl. Acad. Sci. U.S.A. 78, 7754-7758. Grant, S . , Bhalla, K., Weinstein, I. B., Pestka, S., and Fisher, P. B. (1982). Biochem. Biophys. Res. Comniun. 108, 1048-1055. Greiner, J. W., Horan Hand, P., Noguchi, P., Fisher, P. B., Pestka, S., and Schlom, J. (1984). Cancer Res. 44, 3208-3214. Hart, 1. E. , and Fidler, I. J. (1981). Biochim. Biophys. Acta 651, 37-50. Hendler, F. J.. Yuan, D., and Vitetta, E. S. (1981). Trans. Assoc. Am. Physicians 94,217-224. Herzenberg, L. A,, Herzenberg, L. A., and Milstein, C. (1978). Handb. E x p . Immunol. 3rd Ed. Chap. 25.1-25.7. Horan Hand, P., Nuti, M., Colcher, D., and Schlom, J. (1983). Cancer Res. 43, 728-735. Imai, K., Ng, A.-K., Glassy, M. C., and Ferrone, S. (1981). Immunology 127, 505-508. Imam, A., Taylor, C. R., and Tokes, Z. A. (1984).J. Zmmunobiol., in press. Kerbel, R. S. (1979). Nature (London)280, 358-360. Kim, E. E., Deland, F. H., Nelson, M. O., Bennett, S., Simmons, G., Alpoert, E., and Goldenberg, D. M. (1980). Cancer Res. 40, 3008-3012. Kohler, M., and Milstein, C. (1975). Nature (London)256, 494-497. Krolick, K. A,, Yuan, D., and Vittetta, E. S. (1981). Cancer Immunol. Zmmunother. 12, 39-41. Kufe, D. W., Nadler, L., Sargent, L., Shapiro, L., Hand, P., Austin, F., Colcher, D., and Schlom, J. (1983). Cancer Res. 43, 851-857. Kufe, D., Inghirami, G., Abe, M., Hayes, D., Justi-Wheeler, H., and Schlom, Jr. (1984). Hybriduma 3, 223-232. Larson, S. M . , Carrasyuillo, J , A , , Krohn, K. A , , McCufin, R . W., Williams, 1). L., Hellstroin, I . , and Lyster, 1). (1983).J. Am. Med. Assoc. 249, 811-812.

MONOCLONAL ANTIBODIES TO BREAST TAAS

173

Mach, J. P., Carrel, S., Forni, M., Ritschard, J., Donath, A,, and Alberto, P. (1980).N . Engl. J . Med. 303, 5-10. Mach, J. P., Chatal, J. F., Lumbroso, J. D., Buchegger, F., Forni, M., Ritschard, J., Berche, C., Douillard, J. Y., Carrel, S., and Herlyn, M. (1983). Cancer Res. 43, 5593-5600. Menard, S., Tagliabue, E., Canevari, S., Fossati, G., and Colnaghi, M. I. (1983). Cancer Res. 43, 1295-1300. Miller, F. R., and Heppner, G. H. (1979).J . Natl. Cancer Znst. 63, 1457-1463. Moldofsky, P. J., Powe, J., Mulhern, C. B., Hammond, N., Sears, H. F., Catenby, R. A., Steplewski, Z., and Koprowski, H. (1983). Radiology 149, 549-555 Morrison, R. T., Lyster, D. M., Szasz, I., Alcorn, L. N., Rhodes, B. A., Breslow, K., Burchiel, S. W., and Adams, R. (1983). In ”Radioimmunoimaging and Radioirnmunotherapy” (S. W. Burchiel and B. A. Rhodes, eds.), pp. 369-383.Elsevier, Amsterdam. Nuti. M., Colcher, D., Horan Hand, P., Austin, F., and Schlom, J. (1981). In “Monoclonal Antibodies and Developments in Immunoassay” (A. Albertini and R. Ekins, eds.), 87-97. Elsevier, Amsterdam. Nuti, M., Terarnoto, Y. A., Mariani-Costantini, R., Horan Hand, P., Colcher, D., and Schlom, J. (1982). Znt. J . Cancer 29, 539-545. Papsidero, L. D., Croghan, G. A., O’Connell, M. J., Valenzuela, L. A., Nemoto, T., and Chu, T. M. (1983). Cancer Res. 43, 1741-1747. Pestka, S. (1983). Arch. Biochern. Biophys. 221, 1-37. Pettit, W. A., Deland, F. H., Bennett, S. J., and Goldenberg, D. M. (1980). Cancer Res. 40, 3043-3045. Pimm, M. V., and Baldwin, R. W. (1977). Znt. J . Cancer 20, 37-43. Poste, G., Doll, J., and Fidler, I. J. (1981). Proc. Natl. Acad. Sci. U.S.A. 78, 6226-6230. Prehn, R. T. (1970). J . Natl. Cancer Znst. 45, 1039-1105. Primus, F. J., and Goldenberg, D . M. (1980). Cancer Res. 40, 2979-2983. Rasmussen, B. B., Hilkens, J., Hilgers, J., Nielsen, H. H., Thorpe, S. M., and Rose, C. (1982). Breast Cancer Res. Treat. 2, 401-405. Schlom, J., Wunderlich, D., and Terarnoto, Y. A. (1980). Proc. Natl. Acad. Sci. U.S.A. 77, 6841-6845. Soule, H. R., Linder, E . , and Edgington, T. S. (1983). Proc. Natl. Acad. Sci. U.S.A.80, 13321336. Staehlin, T., Hobbs, D.S., Kung, H.-F., Lai, C.-Y., and Pestka, S. (1981a).J . Biol. Chem. 256, 9750-9754. Staehlin, T., Hobbs, D. S., Kung, H.-F., and Pestka, S. (1981b). In “Interferons, Methods in Enzymology” (S. Pestka, ed.), Vol. 78, pp. 505-522. Academic Press, New York. Taylor-Papadimitriou, J., Peterson, J. A., Arklie, J., Burchell, J., Ceriani, R. L., and Bodmer, W. F. (1981). Int. J . Cancer 28, 17-21. Teramoto, Y. A,, Mariani, R., Wunderlich, D., and Schlom, J. (1982). Cancer 50, 241-249. Thompson, C. H., Jones, S. L., Whitehead, R. H., and McKenzie, I. F. C. (1983). J . Natl. Cancer Znst. 70, 409-419. Wunderlich, D., Terarnoto, Y. A., and Schlom, J. (1981). Eur. I. Cancer Clin. Oncol. 17, 719730. Yuan, D., Hendler, F. J., and Vitetta, E. S. (1982).J . Natl. Cancer Znst. 68, 719-728.

This Page Intentionally Left Blank

TUMOR ANG IOGENESIS Judah Folkman Department of Surgery. Children's Hospital Medical Center and Departments of Surgery and Anatomy, Harvard Medical School. Boston. Massachusetts

I. Introduction . . . . . . . . . . . . . . . . . . . . . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , . . , 11. Methods for Studying Angiogenesis A. The Corneal Micropocket . . . . .

111.

IV. V. VI. VII. VIII.

IX.

X. XI. XII.

B. Sustained-Release Polymer Implants C. Chick Embryo Chorioallantoic D. Cloned Capillary Endothelial Cells A Capillary Grows by Sequential S te p s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Angiogenesis Is a Preneoplastic Marker. . . . . . . . . . . . . . . . . . . . . . . . . Solid Tumors Are Angiogenesis Dependent . . . . . . . . . . . . . . . . . . . . . . . . . . , . . . . . The Vascularized Tumor Continues to Alter Its Blood Supply.. . . . . . . . . . . . . . . . Mast Cells and Heparin Can Potentiate Tumor Angiogenesis Angiogenesis Can Also Be Induced by Certain Nonmalignant Angiogenic Factors and Endothelial Mitogens Have Been Isolated from Tumors and from Some Nonneoplastic Cells. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , . . , . , , . A. Tumor-Derived Angiogenic Factors B. Angiogenic Factors Derived from Nonneoplastic Cells and Tissues. . . . . . . . . Angiogenesis Inhibitors Are Found in Natural Sources Role of Angiogenesis in Clinical Oncology.. . . . . . . . . . Summary References. . . . . . , . , . . , . . . . . . . . . . . . , . . , . , . . . . . . , . . . . , . . . . . . . . . , . , , . , . , , ,

175 176 176 177 177 178 178 180 181 183 184 185 187 188 189 191 196 198 199

I. Introduction

Angiogenesis is the process of generating new capillary blood vessels and leads, therefore, to neovascularization. Angiogenesis occurs during embryonic development (Wagner, 1980; Bar, 1980) and during several physiological and pathological conditions in adult life. For example, ovulation and wound healing could not take place without angiogenesis (Jakob et al., 1977; Gospodarowicz and Thakral, 1978; Hunt et al., 1981). Angiogenesis is also associated with chronic inflammation (Fromer and Klintworth, 1975) and with certain immune reactions (Sidky and Auerbach, 1975). Many nonmalignant diseases of unknown cause are dominated by angiogenesis. For example, the neovascularization associated with retrolental fibroplasia or with diabetic retinopathy may lead to blindness in both cases. New capillaries may invade the joints in arthritis (Folkman et al., 1980). Solid tumors induce angiogenesis. However, tumor angiogenesis differs at 175 ADVANCES IN CANCER RESEARCH, VOL 43

Copyright 6 1985 hy Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12-006643-2

176

JUDAH FOLKMAN

least in a temporal way from the other types of angiogenesis described. In physiological situations, as for example in the development of the corpus luteum or in ovulation, angiogenesis subsides, or is turned off, once the process is completed. In certain nonmalignant processes, angiogenesis is abnormally prolonged, although still self-limited, as in pyogenic granuloma or keloid formation. By contrast, tumor angiogenesis is not self-limited. Once tumor-induced angiogenesis begins, it continues indefinitely until all of the tumor is eradicated or until the host dies. Recent progress in understanding the role of angiogenesis in progressive tumor growth will be discussed in this article. It is important to recognize that the phenomenon of angiogenesis is less accessible to investigation than, for example, blood coagulation. Progress in this field was hampered until new methods were developed for the study of capillary growth in vivo and in vitro. II. Methods for Studying Angiogenesis

The early literature on tumor angiogenesis is mainly descriptive, and most investigators took advantage of transparent chambers that could be implanted in animals. The first such chamber, developed by Sandison in 1928 (Sandison, 1928), was implanted in the ear of a rabbit. New vessels grew into the chamber in response to the wound of implantation. Later, tumors were inserted into the chambers. (For an excellent review see Peterson, 1979.) For example, Algire et uZ. (1945) concluded from chambers implanted in mice that tumors could continuously induce the growth of new blood vessels. In 1968, Greenblatt and Shubik (Greenblatt and Shubik, 1968) used a similar chamber in the hamster cheek pouch to demonstrate that capillary proliferation was still induced by a tumor even when it was separated from the host’s vascular bed by a millipore filter. In the 1970s, new techniques were needed to quantitate capillary growth and to test multiple fractions of tumor extracts for angiogenic activity. Four methods developed by Folkman and his associates are now employed by many investigators who study angiogenesis. These are described below. A. THE CORNEAL MICROPOCKET The corneal micropocket technique permits the linear measurement of individual growing capillaries. Tumor implants (1 mm3) are inserted into a pocket made in the cornea of a rabbit at a distance of 1 to 2 mm from the edge of the cornea and the normal vascular bed (Gimbrone et al., 1974). New capillaries grow at right angles from the edge of the cornea and elongate toward the tumor at approximately 0.2 mm/day. They are measured with a

TUMOR ANGIOGENESIS

177

slit-lamp stereoscope. Other tissues can be interposed between the tumor and its vascular bed. For example, the angiogenesis inhibitory property of cartilage was demonstrated in this way (Brem and Folkman, 1975). A disadvantage is that tumors other than those of rabbit origin may excite an immune response once they are vascularized. Subsequent immunological neovascularization may confuse the original experiment. This problem has been overcome at least for mouse tumors by the implantation of tissues into the cornea of inbred mice (Muthukkaruppan and Auerbach, 1979). The rat cornea has also been used (Fournier et al., 1981). Another pitfall is nonspecific inflammation. Inflammatory agents usually can be defined by their ability to attract neutrophils, lymphocytes, or macrophages into the cornea. These cells may themselves elicit neovascularization. For example, solutions that are hyperosmotic or of abnormally low or high pH are inflammatory. Whenever a new substance is tested for angiogenic activity, it is important to obtain histological sections to determine if an inflammatory response is present.

POLYMERIMPLANTS B. SUSTAINED-RELEASE After the corneal micropocket technique was developed, it became feasible to substitute soluble tumor extracts for tumor implants. However, most of these extracts rapidly diffused away. The problem was how to release these extracts over a sustained period of time so that a concentration gradient could be established within the cornea. There were two further requirements: (1)the implantable sustained-release vehicle had to be inert and could not itself cause inflammation in the cornea, and (2) the implant had to be capable of releasing substances of large molecular weights. After empirical testing of a variety of polymers, two were found to satisfy these conditions: poly(2-hydroxyethyl methacrylate) and ethylene-vinyl acetate copolymer (Langer and Folkman, 1976). From such implants it has been possible to release proteins and other macromolecules at nearly constant rates at micrograms/day and, more recently, nanogramstday (Murray et al., 1983a)for periods of weeks to months. Implants can be as small as 1mm3 and are well tolerated in the cornea and other tissues.

C. CHICKEMBRYO CHORIOALLANTOIC MEMBRANE As biochemists attempted to purify angiogenic activity from extracts of tumor and normal tissue, a more rapid assay was needed to screen numerous fractions. A method of dropping the chorioallantoic membrane and opening the egg shell to reveal a large expanse of vascular membrane was previously described by Leighton (Leighton, 1967). The vessels of the chorioallantoic

178

JUUAH FOLKMAN

membrane grow rapidly until day 11, after which endothelial proliferation rapidly decreases (Ausprunk and Folkman, 1977). Tumors or tumor fractions implanted on day 9 induce an angiogenic response within 48-72 hr. This response can be recognized under a stereoscope as new capillary loops converging on the implant (Knighton et al., 1977; Klagsbrun et al., 1976). Two vehicles are commonly used for the delivery of small quantities (1-50 pg) of test material over these relatively brief periods of time. The test substance can be dissolved in 0.5% methyl cellulose, which is then dried to make disks of about 2 mm diameter (Taylor and Folkman, 1982). An alternative method is to dry the test substance in 5-10 pl of distilled water on a Thermanox plastic coverslip. Either disk is then applied to the chorioallantoic membrane. Recently it was found that the addition of 1-2 units of heparin (6-12 Fg) to the test substance potentiated the angiogenesis so that the assay could be read in 1 or 2 days (Taylor and Folkman, 1982; Folkman et al., 1983). Another recent finding is that angiogenesis inhibitors can easily be tested on the 4-day yolk sac vessels or on the 6-day chorioallantoic membrane vessels of shell-less embryos cultured in petri dishes (Taylor and Folkman, 1982; Folkman et al., 1983).

D. CLONEDCAPILLARY ENDOTHELIAL CELLS Capillary endothelial cells have been cloned and passaged in long-term culture (Folkman et al., 1979). These cells are useful for detecting endothelial cell mitogens. They form tubes and branches in uitro (Folkman and Haudenschild, 1980; Madri and Williams, 1983) and further buttress in vivo observations that suggest that capillary growth is a multistep process requiring an orderly sequence of events. Tube formation in uitro has also been observed with endothelial cells from umbilical vein and fetal aorta (Maciag et al., 1982; Feder et al., 1983).

Ill. A Capillary Grows by Sequential Steps

By using all of these techniques to study angiogenesis, it is becoming clear that capillary growth takes place by a series of sequential steps that are similar regardless of' the type of angiogenic stimulus. These steps can be summarized as follows: 1. New capillaries originate from small venules or from other capillaries. Larger vessels with layers of smooth muscle do not usually give rise to capillary sprouts. 2. Local degradation of the basement membrane on the side of the venule closest to the angiogenic stimulus (Ausprunk and Folkman, 1977) is one of the earliest events. Capillary endothelial cells stimulated in uitro by an-

TUMOR ANGIOGENESIS

179

giogenic substances secrete high concentrations of collagenase and plasminogen activator (Moscatelli et al., 1981; Gross et aZ., 1983). These findings suggest that the local degradation of basement membrane seen in viuo is carried out directly by endothelial cells once they receive an angiogenic stimulus. Additional evidence for the availability of stimulated endothelial cells to degrade basement membrane is reported by Madri and Williams (1983) and by Kalebic et al. (1983). 3. Through this opening in the basement membrane, endothelial cells begin to migrate toward the angiogenic stimulus (Ausprunk and Folkman, 1977). 4. Endothelial cells that follow the leaders align in a bipolar fashion as the first sprout begins to form. This alignment has also been demonstrated in uitro, in capillary sprouts growing in plasma clots (Nicosia et al., 1983). 5 . Lumen formation begins. In most examples of postembryonic angiogenesis, a lumen appears to be produced by a curvature that develops in the capillary endothelial cell as if the cytoskeleton itself were undergoing curvature. This phenomenon has been observed in viuo in the cornea (Ausprunk and Folkman, 1977) and also in uitro (Folkman and Haudenschild, 1980). However, during early embryonic angiogenesis, lumen formation may take place by vacuole formation. This has also been observed in viuo (Bar and Wolff, 1972) and in uitro (Folkman and Haudenschild, 1980). 6. Endothelial cells in the midsection of the sprout begin to undergo mitosis. The leading capillary endothelial cells at the very tip of the sprout continue to migrate, but usually do not divide. 7. Loop formation occurs next, as individual sprouts join or anastomose with each other. These loops then elongate and may be the origin of additional sprouts. How individual sprouts find each other is not known. The loops continue to converge upon the angiogenic target. 8. Flow begins slowly after loops have formed. 9. Pericytes emerge along the length of the capillary sprout. 10. Synthesis of new basement membrane follows. When cloned capillary endothelial cells were studied in uitro, most of these same events were observed. Entire capillary networks developed in culture dishes and included lumina, branches, and multiple layers (Folkman and Haudenschild, 1980; Madri and Williams, 1983; Montesano et al., 1983). Maciag, using umbilical vein endothelial cells, has shown that these in uitro “tubes” are hollow and will carry fluid (Maciag et al., 1982). Pericytes were not present in uitro. When the observations from these studies are taken together with in uiuo observations, they suggest that the vascular endothelial cell expresses a defined program of events to generate a capillary network. The program seems to be the same regardless of whether the angiogenic signal is from a tumor or from an inflammatory agent or an immunological stimulus.

180

J U D A H FOLKMAN

IV. Angiogenesis Is a Preneoplastic Marker

In a series of carefully designed experiments, Gullino and his associates demonstrated that angiogenic activity is acquired (or markedly increased) during progression of normal cells to the neoplastic state (Gimbrone and Gullino, 1976; Brem et al., 1977; Maiorana and Gullino, 1978). In mice and rats, the resting adult mammary gland has practically no angiogenic activity when tested in the anterior chamber of the rabbit eye. However, mammary carcinomas acquire the capacity almost consistently. Certain hyperplastic lesions develop in the mammary gland of both species (DeOme et al., 1959). These lesions can be transplanted into the mammary fat pad and grow in it to a limited extent. For some of these transplants it is possible to predict quite accurately the frequency of neoplastic transformation (Medina, 1973). The lesions with a high frequency of neoplastic transformation induced angiogenesis at a much greater rate than did lesions of low frequency of transformation. This elevated angiogenic capacity was observed long before any morphologic sign of neoplastic transformation was apparent (Gimbrone and Gullino, 1976; Maiorana and Gullino, 1978). In fact, hyperplastic lesions of the human breast behaved similarly, showing the appearance of strong angiogenic activity long before the onset of malignancy (Brem et al., 1978). Preneoplastic lesions of human bladder mucosa also display high angiogenic activity in contrast to benign lesions that have little or no angiogenic activity (Chodak et al., 1980). Two recent reports provide additional evidence that expression of angiogenic activity (or amplification of angiogenic activity) may appear in cells long before they reach the stage of tumor formation. In one such experiment, the appearance of angiogenesis predicted the later formation of sarcomas around plastic implants in rodents (Ziche and Gullino, 1981). In a second study, normal mouse diploid fibroblasts were carried in culture (see Fig. 1).At each passage, i.e., approximately once per week, the cells were tested for angiogenic activity in the rabbit eye and for tumorgenicity by reimplantation into the mouse strain that donated the fibroblasts (Ziche and Gullino, 1982). Angiogenic activity first appeared at the fifth passage in some dishes and was present in virtually all cells by the seventh passage. Angiogenic activity persisted in all cells thereafter. However, tuinorigenicity did not occur until the fifteenth passage. It should be recognized that angiogenic capacity and neoplastic transformation are probably not interdependent events. Each can be expressed in the absence of the other (Brem et al., 1978). Certain benign tumors can stimulate intense angiogenesis, but do not become malignant, for example, adrenal adenomas. Furthermore, “tumor take” does not require angiogenesis.

TUMOR ANGIOGENESIS

PASSAGE NUMBER

MOUSE FIBROBLASTS

c

181

I

FIG. 1. Summary of experiment by Ziche and Gullino (1982) demonstrating that angiogenic capacity precedes tumorigenicity. Normal mouse diploid fibroblasts are harvested from inbred mice and cultured. At each weekly passage, the cells are reimplanted into the same mouse strain to test for turnorigenicity. The cells are also implanted into the rabbit cornea to test for angiogenic capacity. (The rabbits were pretreated with corticosteroids to prevent inflammatory or immune angiogenesis.) By 5 weeks, cells in some of the culture plates are angiogenic hut none is tumorigenic. By the seventh passage, virtually all of the cells are angiogenic. However, tumorigenicity does not appear until the fifteenth week.

V. Solid Tumors Are Angiogenesis Dependent

A clue that tumor growth might be dependent upon neovascularization came from observing the growth of tumors implanted into organs maintained by isolated perfusion in glass chambers (Folkman et al., 1963, 1966; Folkman, 1970). New capillaries could not proliferate from the vascular network ofthese isolated organs. This was due to gradual endothelial degeneration, a common problem of isolated, perfused organs (Gimbrone et al., 1969). The tumor implants remained “avascular” and stopped growing at diameters less than 23 mm. Nevertheless, when these avascular tumors were transplanted back into mice, they grew, became vascularized, and eventually killed their host. From these experiments taken together with those of Algire et al. (1945),a hypothesis was proposed that tumors are angiogenesis dependent (Folkman, 1972). In its simplest terms, this hypothesis can be stated: Once tumor take has occurred, every increase in tumor cell population must be preceded by an increase in new capillaries that converge upon the tumor. The hypothesis also implies that angiogenesis is a control point common to most, if not all, solid neoplasms. Evidence in support of this idea was accumulated from four types of experiments:

1. The prevascular phase can be measured and prolonged in uivo. A discrete avascular phase of tumor growth exists during which tumors are usually microscopic or at least not palpable. In the conventional transplanta-

182

JUDAH FOLKMAN

tion of tumor into experimental animals, this phase is brief (3 to 6 days) and generally indiscernible. However, the avascular phase can be studied comfortably when the tumor is separated from its vascular bed. The severe limitation of tumor growth in the absence of angiogenesis is then immediately appreciated. For example, tumors implanted into the anterior chamber of the rabbit eye will float in the aqueous humor and reach mean volumes of approximately 0.5-0.6 mm3 by 14 days. In contrast, the same tumors subsequently seated on the iris become vascularized rapidly and attain a mean volume of approximately 330 mm" during the same period of time (Gimbrone

et al., 1972). 2. The prevascular phase can be simulated by an in vitro model. Multicellular tumor spheroids suspended in soft agar (Sutherland et al., 1971) can be used as an approximate model of the prevascular phase of tumor growth by changing the medium frequently (Folkman and Hochberg, 1973; Folkman et aZ., 1974). Spheroids from a variety of different tumor types at first enlarge exponentially, then gradually slow their growth, and reach a critical diameter beyond which there is no further enlargement. For example, B16 melanoma reaches a mean diameter of approximately 2.4 min. At this steady state, viable cells proliferate in the periphery while cells in the center undergo necrosis. The mechanism of this dormancy appears to be a combination of limited inward diffusion of oxygen (once the spheroid's radius exceeds approximately 150-200 pm) and limited outward diffusion of metabolic wastes. (For a detailed study of the mechanism of growth arrest in these spheroids see Carlsson et d.,1979.) 3. The effects of the vascular phase upon tumor growth can be observed. Once the vascular phase has commenced, tumor cells lying closest to an open capillary have the highest [3H]thymidine-labeling index. The labeling index in tumor cells decreases as they increase their distance from an open capillary (Tannock, 1968). This relationship of tumor cell proliferation to proximity of capillaries within the tumor correlates well with the estimated oxygen diffusion lengths from the center of a given capillary. Thus, it appears that solid tumors are made up of cylinders or cords of tumor cells surrounding individual capillary units. Denekamp (1982) has shown that this relationship also holds for human tumors. That tumor cells prefer contiguity to capillaries may not wholly depend on diffusion of oxygen and nutrients. Nicosia et al. (1983) implanted small fragments of rat aorta into plasma clots, and capillaries grew out radially. When tumor cells were inoculated into the periphery of the plasma clot, they grew slowly as a small spheroidal focus. However, when a capillary sprout approached the tumor focus, tumor cells grew rapidly around the capillary as a cylindrical cuff. Tumor growth extended toward the center of the plasma clot (where the aortic segment had been placed). There is no blood flow in this system. One interpretation is

TUMOR ANGIOGENESIS

183

that tumor growth can be directly facilitated by capillary endothelial cells or perhaps by the matrix which they produce. 4. The administration of angiogenesis inhibitors stops tumor growth. Certain angiogenesis inhibitors now available can be administered locally, regionally, or systemically and can prevent tumors from becoming vascularized. Thus, some tumors can now be maintained in the prevascular phase or eradicated by angiogenesis inhibition alone. These experiments will be discussed in Section X (see also Langer et al., 1980; Taylor and Folkman, 1982; Folkman et al., 1983). In summary, it should be remembered that the dependency of solid tumors on angiogenesis is related to the growth of their cells in a tightly packed population of high density (10R-109 cells/cm3)).Wherever malignant cells have developed the capacity to grow separately from each other (for example, as ascites or as infilitrates in certain leukemias), angiogenesis is not required. (There are, of course, many ascites tumors that retain the capacity to induce angiogenesis, and these cells can grow in both the solid and ascites configurations. Examples are Walker carcinoma in rats and ovarian carcinoma in humans.)

VI. The Vascularized Tumor Continues to Alter Its Blood Supply

At first glance, a growing tumor whose cells are stimulating new capillary sprouts and then accumulating around them appears analogous to an organ in the embryo. But the analogy does not hold because tumors continue to alter their intrinsic vasculature in a way that normal organs do not. For example, tumors tend to increase their vascular volume. Thus, in rats where vascular volume is 20%of normal tissue weight, it is 50% of tumor weight in the early stage of hepatoma (Yamaura and Sato, 1974). While the increase in vascular volume is different from one tumor type to another, specific tumor types tend to maintain a characteristic increase in vascular space (Gullino and Grantham, 1964). Second, tumors eventually begin to compress their own capillaries. In most experimental tumors of a size less than 0.5 cm3, new vessels are open throughout the tumor, and there is no necrosis. Beyond this size, there may be gradual compression of capillaries. Because this extravascular pressure can close capillaries and stop the microcirculation (Warren, 1970), it is more accurate to envisage that tumors compress their blood supply rather than to perpetuate the notion that tumors outgrow their blood supply. This compression eventually leads to prolonged cessation of flow in the core of the tumor followed by central necrosis. Central necrosis begins to appear for a variety of experimental tumors after

184

JUDAH FOLKMAN

they have grown beyond approximately 1-2 om3 Goldacre and Sylven, 1962). What accounts for this compressive effect? Mitosis contributes. After each cell division, the added pressure from daughter cells must be attenuated throughout the tumor mass. Leaky capillaries also contribute. New capillaries in the tumor leak fluorescein and colloidal carbon which normal capillaries do not. This increased permeability of new capillaries is compounded by the lack of lymphatics in most tumors (Swabb et al., 1974). We have never found lymphatics in tumor implants in the rabbit cornea, even after these implants have become heavily vascularized. As a result, extravascular tissue pressure is higher in a tumor than in its normal counterpart (Young et al., 1950; Peters et al., 1980; Paskins-Hulburt et al., 1982). The increased leakiness of capillaries within a tumor bed may be facilitated by a tumor factor that increases permeability (Dvorak et al., 1981). Also, in certain tumors (e.g., brain tumors), structural alterations appear in the new capillaries. Fenestrations arise in the endothelial cells, pinocytotic vesicles are increased, and occasionally there are interruptions in the basement membrane (Hirano and Matsui, 1975; Pousa et al., 1979; Deane and Lantos, 1981). VII. Mast Cells and Heparin Can Potentiate Tumor Angiogenesis

There is increasing evidence that the mast cell may act as a helper to the capillary endothelial cell during certain types of angiogenesis, especially tumor angiogenesis. In his first description of mast cells, Ehrlich (1879) noted that they seemed to be more frequent in the neighborhood of carcinomas. While increased mast cell populations are found in a variety of pathological conditions including immediate hypersensitivity or chronic inflammation, they have also been shown to congregate in areas of neovascularization. Increased mast cells have been reported in the most highly vascularized areas of certain tumors (Giani, 1964), at the periphery of carcinoma in situ (Dunn and Montgomery, 1957), and in the axillary nodes of patients with breast cancer metastatic to the nodes (Thoresen et al., 1982). These correlations led to the notion that mast cells might somehow be associated with the growth of new capillaries, although just what this association might be was not known (Ryan, 1970; Smith and Basu, 1970). In an attempt to quantitate the relationship of mast cells to neovascularization, tumors were implanted on the chorioallantoic membrane of the chick embryo (Kessler et al., 1976). There was a 40-fold increase in mast cell density in the area of the tumor implant. These mast cells appeared in the neighborhood of the tumor about 24 hr before new capillary sprouts. While mast cells alone were unable to induce angiogenesis, cultured mast cells or the media in which they incubated stimulated migration of capillary endo-

TUMOR ANGIOGENESIS

185

thelial cells in uitro (Azizkhan et al., 1980). Heparin released by the mast cells was responsible for this stimulation of capillary endothelium (Azizkhan et al., 1980). Furthermore, heparin potentiated tumor angiogenesis. When the chorioallantoic membrane is exposed to tumor extracts with angiogenic activity, neovascularization is usually observed in about 3 days. However, if a small amount (6-12 pg) of heparin is added to the tumor extract, angiogenesis is potentiated and neovascularization appears within 1day (Taylor and Folkman, 1982; Folkman et al., 1983). Heparin by itself does not induce angiogenesis. However, heparin does become angiogenic when it is complexed to copper (Ziche et al., 1981). The mechanism of this phenomenon is unknown. These findings will be discussed in more detail in Section IX. VIII. Angiogenesis Can Also Be Induced by Certain Nonmalignant Cells

Several types of normal cells are known to induce angiogenesis under appropriate conditions. For example, macrophages, when properly activated, can release angiogenic activity in the cornea (Polverini et al., 1977). Macrophages are activated in uiuo by injecting paraffin oil or endotoxin into the peritoneal cavity of experimental animals or in uitro by adding latex beads to the culture medium. Macrophages collected from wounds display similar angiogenic activity (Thakral et al., 1979; Hunt et al., 1981). Recently it has been shown that wound macrophages are oxygen sensitive in relation to angiogenic activity. When these macrophages are activated by wound debris or fibrin products, they produce maximal angiogenic activity if oxygen concentration is minimal. As local oxygen tension rises (for example, after new capillaries appear in the wound), macrophage angiogenic activity is decreased and eventually turned off (Banda et al., 1982; Knighton et al., 1983). Macrophages are attracted to some, but not all, tumors (Mantovani, 1982). Polverini and Leibovich (1984) have recently reported that macrophages embedded in a tumor may also contribute to its angiogenic activity. The idea that other types of'leukocytes might also have angiogenic activity arose from the observation of corneal neovascularization associated with leukocyte infiltrates (Fromer and Klintworth, 1975). Lymphocyte-induced angiogenesis was observed when allogeneic immunocompetent lymphocytes were injected intradermally into immunosuppressed or irradiated host mice. The phenomenon has been most clearly defined by Auerbach and his associates and by subsequent workers (Sidky and Auerbach, 1975; Fromer and Klintworth, 1975; Kaminski et al., 1975; Auerbach, 1981). Not all classes of lymphocytes are able to induce angiogenesis, however. For example, angiogenesis is evoked by effector T cells (i.e., corticosteroid-resistantthymocytes or by spleen or lymph node cells), but not by spleen cells from

186

J U D A H FOLKMAN

athymic (nulnu)donor animals (Sidky and Auerbach, 1975; Kaminiski et aZ., 1978a). The most dramatic lymphocyte-induced angiogenesis is achieved by donor lymphocytes that differ at the H - 2 locus from the recipient animal (Auerbach and Sidky, 1979). Taken together, these results indicate that the cell responsible for lymphocyte-induced angiogenesis belongs in the general class that includes most of the lymphokine-producing T cells (Adelman et al., 1980). However, the mechanism of lymphocyte-induced angiogenesis is not yet clear. It is not known whether foreign lymphocytes can stimulate endothelial cells directly or whether lymphocyte angiogenic activity is mediated through another cell such as an activated macrophage. Fat cells or adipocytes derived from 3T3 fibroblasts that have undergone differentiation in uitro (Green and Kehinde, 1976) are also capable of inducing angiogenesis (Castellot et al., 1980). Angiogenic activity is differentiation-dependent because adipocytes secrete more angiogenic activity than preadipocytes. These 3T3 adipocytes secrete in a differentiation-dependent fashion a factor that stimulates chemotaxis of vascular endothelial cells in uitro and neovascularization in uiuo (Castellot et aZ., 1982). Conditioned medium from adipocytes, but not from preadipocytes, strongly stimulates both plasminogen activator and collagenase release from capillary endothelial cells, but not from aortic endothelial cells (Rifkin et al., 1982). Taken together, these findings provide the first demonstration of an angiogenic activity secreted as a consequence of the dqferentiation of the producer cell type. At least six normal tissues so far have been demonstrated to express angiogenic activity during some period of their adult life. For some tissues, this angiogenic activity is expressed briefly or in a cyclic manner. For each of the tissues, the cells responsible for angiogenesis are as yet unknown. For example, during the period when a follicle forms in the ovary and the corpus luteum develops, the corpus luteurn is quickly invaded by new capillary sprouts. At this stage, the corpus luteum can express angiogenic activity when transplanted, for example, to the hamster cheek pouch or to the rabbit cornea (Jakob et al., 1977; Gospodarowicz and Thakral, 1978). Human follicular fluid is also angiogenic (Frederick et d.,1984). When the corpus luteum regresses, the new vessels also regress. Testicular fragments from 1day-old mice implanted into the subcutaneous tissue of castrated mice of the same strain can also induce transient angiogenesis (Huseby et al., 1975). Fragments of kidney from newborn hamsters produce a weak angiogenesis response in the hamster cheek pouch even when enclosed by a millipore filter (Warren et al., 1972). Fragments of embryonic or adult mouse kidney also give a weak angiogenic response on the chorioallantoic membrane of the chick embryo (Folkman and Cotran, 1976).

TUMOR ANGIOGENESIS

187

Under certain conditions epidermis, but not dermis, is angiogenic (Nishioka and Ryan, 1972; Wolf and Harrison, 1973). Fragments of retina placed into a corneal pocket in rabbits also induce neovascularization, whereas other ocular tissues such as sclera have no activity or are weakly angiogenic (G. C. Brown, 1980). Glaser et al. (1980a) isolated a vasoproliferative factor from mammalian retina. Further purification of this activity from bovine retina has been reported and will be discussed in the following section on angiogenic factors (D’Amore et al., 1981). Extracts of the male mouse salivary gland were found to induce angiogencsis in the chick embryo (Folkman and Cotran, 1976). However, this observation has not been pursued because of the inflammatory reaction associated with it.

IX. Angiogenic Factors and Endothelial Mitogens Have Been Isolated from Tumors and from Some Nonneoplastic Cells

In the previous discussion we have seen that a better understanding of the phenomenon of angiogenesis has been achieved by the study of its component events. It is more difficult to describe angiogenesis in terms of its chemical mediators. However, progress is gradually being made from the contributions of many laboratories. In some respects, the attempt to understand the chemical mediators of angiogenesis is analogous to the current effort to understand classic immunological phenomena in terms of the chemical factors that mediate them. However, the elucidation of angiogenesis factors is formidable because the bioassay for angiogenesis is carried out only in uiuo. Furthermore, even if neovascularization results from the application of a test substance to such an in uiuo assay, the substance cannot immediately be labeled as an angiogenic factor. It could, by causing injury or inflammation, act as a chemotactic factor for other cells such as macrophages, which themselves might then be the source of angiogenic activity. Formic acid is an example. To avoid this confusion the concept of direct and indirect angiogenic activity was introduced (Folkman and Haudenschild, 1980). A “direct” angiogenic factor could stimulate capillary proliferation without having to depend on intervening cell types. Careful histological monitoring, at present, is about the only way of distinguishing between direct and indirect angiogenesis. Another difficulty is that components of angiogenesis, such as endothelial proliferation, migration, or enzyme production, can be studied individually in uitro, but they do not necessarily predict angiogenic activity in uiuo. Thus, a factor that stimulates endothelial mitosis in uitro may or may not also be angiogenic.

188

JUDAH FOLKMAN

With these caveats in mind, the following discussion summarizes angiogenic factors that have been reported to date. A. TUMOR-DERIVED ANCIOGENICFACTORS

The first angiogenic factor was isolated from Walker 256 carcinoma grown in rats (Folkman et al., 1971). The angiogenic activity was isolated from cells in the ascites form, and this activity was destroyed by proteases or by heating. In a subsequent report, similar activity was isolated from the nuclei of these tumor cells and was associated with the nonhistone proteins (Tuan et al., 1973). Phillips et ul. (1976) extracted angiogenic activity from solid Walker tumor cells and found two high-molecular-weight fractions, one of which (35,000 to 100,000) gave the most intense angiogenic activity. Rat liver or regenerating liver handled in the same way was not active. Angiogenic activity was also found in a human Wilms’ tumor and in a hypernephroma. Fenselau and Mello (1976; Fenselau et al., 1981a) also used Walker tumor cells, but guided purification by in vitro assays based on proliferation of fetal bovine aortic endothelial cells. Periodically the fractions were tested for angiogenic activity in the chick embryo and in the rat cornea. In their early reports, angiogenic activity from homogenates of ascites tumor cells appeared in fractions of high molecular weight when analyzed by gel filtration and neutral pH. However, at p H 4, angiogenic activity was associated with materials of molecular weight less than 800. Further purification of these materials by silica gel chromatography of ethanol extracts of the lyophilized tumor homogenates revealed a low-molecular-weight material with an ultraviolet absorption maximum near 260 nm which was not composed of protein or peptides (Watt, 1981; Fenselau, et al., 1981b; Fenselau, 1984). A factor isolated from the media of cultured neural cells and neural tumor cells stimulated proliferation of umbilical vein endothelial cells in culture (Suddith et ul., 1975). In a similar experiment, an endothelial mitogen for aortic endothelial cells was isolated from cultures of Walker tumor cells (McAuslan and Hoffman, 1979). In another study, a factor isolated from Walker tumor cells was mitogenic for microvascular endothelial cells derived from cow brain and was also angiogenic in the chick embryo (Schor et al., 1980).This factor and the one reported by McAuslan were both of low molecular weight. Angiogenic activity was found in the cultured supernate of seven cell lines derived from a variety of spontaneous human tumors and grown in largescale suspension culture. Activity was not further purified (Tolbert et al., 1981). The low-molecular-weight angiogenic factor (approximately 200) pre-

TUMOR ANGIOGENESIS

189

viously found in Walker tumor cells by Schor et aZ. (1980) stimulated capillary endothelial cells but not aortic endothelial cells, and then only when the capillary cells were growing on native collagen substratum (Keegan et aZ., 1982). Most recently, this group in Kumar’s laboratory has grown a human lung tumor in serum-free medium for 12 months and has isolated an angiogenic factor from the medium (Kumar et al., 1983a). The highest angiogenic activity was in the range of approximately 80,000 MW. Also, a dialyzable, low-molecular-weight angiogenic factor was found in this medium. These authors concluded that the high-molecular-weight fraction possibly contained a carrier protein and the low-molecular-weight fraction was carrier-free. The low-molecular-weight component did not absorb at 260 nm and was felt to be different from that isolated by Fenselau et al. (1981b). This study also ruled out the possibility that angiogenic activity could be a component of serum because the cells were grown in serum-free media. Angiogenic activity has also been isolated from human central nervous system tumors in culture (Matsuno, 1981) and from cultures of human malignant melanoma cells (Stenzinger et aZ., 1983). However, neither of these angiogenic activities has been purified. Recently, a tumor-derived endothelial mitogen that is angiogenic has been purified to homogeneity (Shing et al., 1983, 1984). This factor was obtained from chondrosarcoma grown in the rat. The factor was found in extracts of the tumor cell but also from extracts of the chondrosarcoma matrix. It was purified one million-fold to a single-band preparation by a two-step procedure that utilized cationic exchange on Biorex 70 and affinity chromatography by heparin-Sepharose. The purified factor is a cationic peptide with an isolectric point of approximately 9.8 and a molecular weight of about 18,000 (Fig. 2). It stimulates capillary endothelial proliferation halfmaximally at a concentration of 1ng/ml. It stimulates strong angiogenesis on the chorioallantoic membrane of the 9-day-old chick embryo at concentrations of 120 ng within 24 hr. Histologic sections reveal that neovascularization takes place in the virtual absence of inflammatory cells. It is too early to say which of these tumor-derived endothelial mitogens and angiogenic factors are, in fact, responsible for the angiogenesis induced by growing tumors. It is also premature to predict which factors may be structurally similar to each other.

FACTORS DERIVEDFROM NONNEOPLASTIC CELLS B. ANGIOGENIC AND TISSUES Angiogenic activity has been isolated and partially purified from synovial fluid (R. A. Brown et al., 1980),wound fluid (Banda et aZ., 1982), and largescale cultures of granulocytes and monocytes (for review see Wissler, 1982).

190

JUDAH FOLKMAN

43 K-

25 7 KI8.4K14.3K-

6.2K-

i m

,s ,z ,_, ~,

3,910

J ,),,3 ,

~,

1 2 SLOT NUMBER

0

4 0 12 16 F R A C T I O N NUMBER

20

FIG.2. A purified endothelial mitogen derived from rat chondrosarcoma. Chondrosarcoma extracellular matrix was digested with collagenase and purified by Bio-Rex 70 chromatography followed by heparin-Sepharose chromatography as described by Shing et al. (1984).On the left, SDS-polyacrylamide gel electrophoresis. (Slot 1) Molecular weight markers (BRL). (Slot 2) The peak fraction of growth activity (200 ng = lo00 U). This purified growth factor was added at various concentrations to cultures of bovine capillary endothelial (BCE) cells and proliferation measured. Half-maximal stimulation is induced by growth factor concentrations of about 1 ng/ml. About 600 U activity (120 ng) are required to produce strong angiogenesis within 24 hr on the 9-day-old chick chorioallantoic membrane. Histological sections reveal that inflammation is virtually absent. The recovery of activity from the tumor is about 5% and the yield of pure growth factor is about 1 pg from 5 g of crude chondrosarcoma matrix.

The mammalian retina has also been a rich source of angiogenic activity, the first report of which was by Glaser et al. (1980a,b). In subsequent reports, this retinal angiogenic factor has been further characterized and further purified (D'Amore et al., 1981). These factors obtained from bovine retina were of large molecular weight. However, a low-molecular-weight angiogenic factor was isolated from cat retina (Kissun et al., 1982). Also, an angiogenic factor in bovine retina was found to possess at least one common antigenic determinant with an angiogenic factor from a tumor extract (Shahabuddin and Kumar, 1983). A factor has been derived from the medium of cultured 3T3 adipocytes that stimulates neovascularization as well as chemotaxis of capillary endothelial cells (Castellot et al., 1982). Angiogenic activity has also been extracted recently from human myocardial infarcts taken postmortem (Kumar et al., 1983b). Partial purification demonstrated a low-molecular-weight factor of approximately 300, analogous to the molecular weight of tumor an-

TUMOR ANGIOGENESIS

191

giogenesis activity obtained from Walker tumor cells by this group of investigators. Prostaglandin E, (PGE,) has also been correlated with angiogenic activity in the cornea (BenEzra, 1978; Ziche et al., 1982), and prostaglandin E, (PGE,) has been shown to have angiogenic activity on the chorioallantoic membrane (Form and Auerbach, 1983). There is also reported to be a relationship between copper ions and angiogenesis. McAuslan (1980) first suggested that copper might play a role in mediating angiogenesis. Subsequently, Raju et al. (1982) showed that either heparin or the tripeptide glycylhistidyllysine, when bound to copper became angiogenic. Copper-free molecules were not angiogenic. At this writing, only one angiogenic factor has been purified to homogeneity. Much work remains to be done (1) to purify each of the reported angiogenic factors; (2) to determine which factors act directly on endothelial cells (or pericytes) and which act as chemotactic agents for other cells (such as mast cells); and (3) to ascertain which factors might actually be participating in tumor angiogenesis or in the angiogenesis of injury. A central question is, if it were possible to neutralize any of these activities (for example, with antibodies), for which factor(s) would such neutralization lead to the inhibition of angiogenesis? Hopefully, in the future it will be possible to sort out the bewildering diversity of factors that influence endothelial and capillary growth. Perhaps we will learn that some angiogenic factors are first synthesized as highmolecular-weight species from which smaller active units are cleaved. X. Angiogenesis Inhibitors Are Found in Natural Sources

The concept of “antiangiogenesis” as a potential therapeutic approach was put forward in 1972 (Folkman, 1972). At the time, however, there was no known angiogenesis inhibitor. Clinical observations suggested that if an angiogenesis inhibitor existed at all, cartilage would be a good place to look for it. For example, osteogenic sarcoma of the bone rarely spreads to adjacent cartilage; breast cancer metastatic to the vertebral bones rarely invades adjacent cartilage in the vertebral disc. ‘The first experimental evidence for this notion was the demonstration by Eisenstein et al. (1973) that cartilage extracted with guanidine lost its resistance to vascular invasion. Subsequently, Brem and Folkman (1975) showed that an implant of cartilage placed adjacent to a tumor in the rabbit cornea inhibited growth of new blood vessels toward the tumor. A factor was isolated and partially purified from cartilage that when locally administered by a sustained-release polymer, inhibited tumor angiogenesis in the cornea (Langer et al., 1976). This cartilage-derived factor was also infused into the carotid artery of mice and

192

JUDAH FOLKMAN

rabbits. The factor inhibited gowth of tumor blood vessels and of the tumor itself (Langer et al., 1980). Eisenstein’s group demonstrated that resistance to invasion by both vascular and connective tissue could be diminished if the cartilage were extracted with guanidine hydrochloride (Sorgente et al., 1975). These guanidine extracts contained protease inhibitor activity (Sorgente et al., 1976). The extracts suppressed the growth of fibroblasts and aortic endothelial cells in culture. Kuettner and his associates showed that cartilage also contained inhibitors of collagenase that blocked bone resorption in vitro (Kuettner et al., 1978). The term “antiinvasive factor” (AIF) was used to categorize these inhibitors. Subsequently, Kuettner’s group found that the collagenase secreted into the media by tumor cells was also inhibited by AIF (Kuettner et al., 1977). When human bone explants were cocultured with tumor cells, the tumor cells invaded and eroded the bone but not the cartilage. In fact, tumor cells infiltrated the cartilage matrix only where it had been previously occupied by capillary loops of the growth plate and nutrient canal (Kuettner and Pauli, 1983). These findings were interpreted to mean that vascular endothelial cells and tumor cells may each generate collagenase and that cartilage could resist invasion by both cell types. Recently, Kaminski et al. (1978b) showed that extracts of human cartilage administered intravenously inhibited vasoproliferation in mice. Also, Cawston et al. (1981) and Bunning et al. (1984) have isolated a series of metalloproteinase inhibitors from cartilage that may play an important role in the antiangiogenic properties of cartilage. Murray et al. (1983) and Gabrielides and Rifkin (1983) have reported further on the purification of a collagenase inhibitor from cartilage. The antiinvasion factor of Kuettner et al. is not purified, nor is the angiogenesis inhibitor of Langer et al. It is too early to say whether these two moieties will turn out to be similar; each group uses different bioassays to guide purification. Nevertheless, data from both groups point to cartilage as a major source of angiogenesis inhibitor and suggest that angiogenesis inhibitors may be found in other tissues. For example, Eisenstein et a2. (1979) have reported an angiogenesis inhibitor isolated from the wall of the aorta. This material can inhibit inflammatory angiogenesis in the cornea Medroxyprogesterone acts as an angiogenesis inhibitor when released locally in the cornea in the presence of tumor-induced vessels (Gross et al., 1981). However, when administered systemically, it does not inhibit angiogenesis or tumor growth. Protamine sulfate was the first angiogenesis inhibitor to be effective when administered systemically (Taylor and Folkman, 1982). This discovery arose from a series of experiments on the role of mast cells in angiogenesis which demonstrated the following (see Section VII). (1) Mast cells accumulated at a tumor site before the ingrowth of new capillary sprouts (2) Heparin from

TUMOR ANGIOGENESIS

193

mast cells increased the migration of capillary endothelial cells in uitro. (3) Tumor angiogenesis was potentiated by heparin in uiuo (4) Protamine blocked the ability of heparin to stimulate migration of capillary endothelial cells. This suggested that protamine or other heparin antagonists might also inhibit angiogenesis in uiuo. Protamine prevented tumor-induced angiogenesis on the chorioallantoic membrane of the chick embryo and it also inhibited growth of embryonic vessels. In the presence of protamine, avascular zones appeared in the 3-day-old yolk sac membranes and in the 6- to 8-day-old chorioallantoic membrane. Protamine had no effect on nongrowing vessels in the chorioallantoic membrane after day 10. Furthermore, protamine-polymer pellets implanted into the rabbit cornea inhibited capillary growth whether induced by tumors, inflammation, or an immunologic reaction (Taylor and Folkman, 1982). When administered systemically to mice, protamine reduced the volume of lung metastases. Many lung tumors remained avascular and stopped growing at mean tumor volumes of 2-3 mm3. Tumor cells were not directly affected by protamine. Tumors implanted in other regions such as the subcutaneous tissue were less responsive to protamine than tumors growing in the lung, probably due to the fact that the uptake of protamine by lung is at least 5 times higher than by subcutaneous tissue and 44 times higher than uptake by skin. It was not possible to raise the dose of protamine sufficiently to inhibit growth of primary tumors or to cause tumor regression because of the toxicity of protamine (lethargy, hypocalcemia, and occasionally sudden death) (Potts et al., 1984). The toxicity is unrelated to antiangiogenesis activity. Despite its toxicity, protamine was the first angiogenesis inhibitor with a known structure. Protamine is an arginine-rich basic protein of 4300 MW found only in sperm. Its amino acid sequence has been determined (Ando and Watanabe, 1969). The experience with protamine suggested that, at best, therapy with angiogenesis inhibitors could reduce a tumor to the avascular stage but could probably not eradicate it. This notion was soon discarded because of the discovery of a more potent form of antiangiogenesis. A new way to inhibit angiogenesis developed as follows. On the basis of the previous finding that heparin potentiated tumor angiogenesis, it was thought that tumor angiogenesis enhanced by heparin might be made more conspicuous in the chick embryo by adding cortisone to suppress background inflammation that occasionally arose as a result of eggshell dust. The result was unexpected. While heparin alone enhanced tumor angiogenesis and cortisone alone had little or no effect, angiogenesis was inhibited by the combination of heparin and cortisone (Folkman et al., 1983). It was further found that heparin administered with cortisone was a potent inhibitor of capillary growth that occurred during embryogenesis (Fig. 3) or that was observed in the cornea from inflammatory or immunological stimuli. Heparin mixed with cortisone in a sustained-release pellet suppressed tumor

194

J U D A H FOLKMAN

Fic. 3. Inhibition of embryonic angiogenesis. Histologic cross sections (A) of day 8 chorioallantoic membranes. Fertilized chick embryos were removed from their shell on day 3 (or 4) and incubated in a petri dish in high humidity and 3-58 COZ. On day 6 a inethylcellulose disk of approximately 2 mm diameter, previously dried from 10 pl of 0.5% methylcellulose, was implanted on the chorioallantoic membrane. Each disk contained either (1)cortisone acetate, (2) heparin, (3)cortisone + heparin, or (4) cortisone + hexasaccharide. Forty-eight hours later, a clear avascular zone appeared in the membranes exposed to heparin + cortisone but not in the membranes exposed to either compound alone (see below). In the avascular zones, capillaries

TUMOR ANGIOGENESIS

195

angiogenesis in the rabbit cornea. The anticoagulant function of heparin was not responsible for this antiangiogenic function because a nonanticoagulant fragment, which was a hexasaccharide, substituted for the activity of the parent molecule. The hexasaccharide has a molecular weight of about 1600. Oral administration of heparin to mice and rats resulted in the release of nonanticoagulant heparin fragments in the serum, which in the presence of cortisone administration had similar antiangiogenic effects. Subcutaneous tumors of B16 melanoma, reticulum cell sarcoma, Lewis lung carcinoma, and bladder carcinoma (MB49) regressed in animals drinking heparin and receiving cortisone acetate injections but not in animals receiving heparin or cortisone alone. In fact, it was possible to eradicate tumors completely in more than 50% of mice. The mice remained tumor-free after treatment was stopped. The number of lung metastases in all mice was reduced to 0.1%of the controls. However, in the case of four tumors, neither angiogenesis nor tumor growth was suppressed by optimal combinations of heparin and cortisone. These were sarcoma 1509A, meth A sarcoma, glioma 26, and glioma 261B. It is interesting that these tumors were all induced by the carcinogen 3-methylcholanthrene. Furthermore, when a nonresponding tumor (sarcoma 1509A) was implanted in the left flank of nude mice and a responding tumor such as reticulum cell sarcoma was implanted in the right flank, the reticulum cell sarcoma regressed in each mouse treated with the heparinwere absent, whereas the ectodermal and endodermal layers were present. ~ 5 0 0 (From . Folkman et al., 1983, with permission of the publisher.) Subsequent work has shown that hydrocortisone 21-phosphate (Sigma) provides more reproducible avascular zones than cortisone acetate, presumably because of higher solubility. An optimum concentration is approximately 50 pg. Heparin antiangiogenic activity varies greatly by manufacturer and by batch, and over wide concentration ranges. For example, Abbott Panheprin is optimally effective (with hydrocortisone) at 6-12 pg; Hepar heparin (Franklin, Ohio) is effective from at least 6 to 200 pg with an optimum at approximately 50 pg; Sigma heparin is effective from 6 to 200 pg with an optimum at about 100-150 pg. (B) Avascular zone in 8 day chick embryo chorioallantoic membrane after application of hydrocortisone 21-phosphate (50 kg) and Panheprin (Abbott) (2 units, i.e., approximately 12 pg). X6. After this study was completed in 1983, Panheprin became unavailable because of an earlier decision by the Abbott Company to stop manufacturing heparin. Other beparins such as Hepar, or Sigma, are effective inhibitors of angiogenesis (with corticosteroids) in the chick embryo or the rabbit cornea. However, when administered orally to mice, only Panheprin can bring about the tumor regressions described in the text. The next most potent heparin (Hepar, Inc.), caused regression only of reticulum cell sarcoma, and it and other heparins were generally ineffective against other tumor types. The lack of Panheprin is of no consequence for the demonstration of angiogenesis inhibition in the chick embryo or the rabbit cornea, because a variety of other heparins are effective in these systems. However, until a heparin of antiangiogenic potency equivalent to Panheprin becomes available, or until large quantities of the appropriate hexasaccharide can be easily produced, tumor regression cannot be attained with currently available heparins except in the case of reticulum cell sarcoma.

196

JUDAH FOLKMAN

cortisone, whereas the sarcoma 1509A continued to grow in the same mouse. Serum from heparin-cortisone or hexasaccharide-cortisone-treated mice was not cytotoxic to tumor cells alone. The inhibitory effect of heparincortisone was specific for growing microvessels; mature, nongrowing vessels remained unaffected. This was demonstrated in the chick embryo at different stages. It is not understood how angiogenesis inhibition results in complete regression of a large tumor mass. However, serial histological studies on a daily basis show progressive loss of capillaries. Residual tumor cells cluster around the remaining capillaries, until finally the entire tumor disappears. If the doses of heparin and cortisone are lowered so that regression is very slow, tumors can be held at a nearly constant size. In some cases, tumors can remain in the avascular state. This implies that with rapid regression, there may be “bystander” killing of residual tumor cells. The mechanism of angiogenesis inhibition by heparin-cortisone or hexasaccharide-cortisone is unknown. However, it has recently been shown (Crum and Folkman, 1984) that neither the glucocorticoid nor the mineralocorticoid activity of cortisone is necessary for antiangiogenesis. For example, a compound such as lla-epicortisol (Upjohn), which has the identical structure to hydrocortisone except that the 11-hydroxyl group is in the a position instead of the f3 position, has no glucocorticoid or mineralocorticoid activity. Yet in the presence of heparin or a hexasacharide fragment of heparin, 11a-epicortisol is a strong angiogenesis inhibitor. The fact that a hexasaccharide fragment of heparin with no previously assigned function, and a corticosteroid with no known biological function can suppress angiogenesis when administered together suggests that regardless of the complexity of the angiogenic phenomenon, it may be governed by simple, naturally occurring molecules. These experimental findings offer a potential fbture role for angiogenesis inhibitors as a new class of pharmacologic agents, of possible use in antitumor therapy, or in other diseases dominated by abnormal neovascularization. XI. Role of Angiogenesis in Clinical Oncology

There are a variety of seemingly unconnected clinical observations that may now be better understood because they are based on angiogenic phenomena. For example, many human tumors seem to exist at first in a prevascular state. An in situ carcinoma may stay for years at a small size of a few millimeters. Examples are carcinoma in situ of the bladder, breast, and cervix (Farrow et al., 1977; Hicks, 1977; Stafl and Mattingly, 1975). The onset of vascularization of a tumor marks the transition to more rapid growth, local invasion, and distant metastasis. The progression of superficial melanoma, for example Clark level I, to its more invasive, faster growing

TUMOR ANGIOGENESIS

197

counterpart also correlates with the arrival of new blood vessels in the tumor. Another example of this transition is observed in ovarian carcinoma that has metastasized to the peritoneal lining. Tiny peritoneal implants may remain white, avascular, and fairly uniform in size until new capillaries penetrate from beneath the peritoneum. The implants then grow and ascites fluid in the abdomen becomes bloody. In fact, bleeding, when it represents a clinical sign of malignancy, generally indicates that a tumor no matter where it is located has become vascularized. Sometimes a vascularized tumor may be revealed because it can stimulate angiogenesis in a remote location. For example, neovascularization of the iris is commonly associated with neoplasms of the retina, especially the retinoblastoma. Another example is the appearance of neovascularization in an old mastectomy scar that precedes the recurrence of tumor beneath the scar. Metastasis is also influenced by angiogenesis. Prior to vascularization, tumors are generally unable to shed cells into the circulation. For this reason, prevascular tumors have a low probability of metastasizing compared to their vascularized counterparts. For example, melanomas that are less than 0.7 mm thick reside entirely above the basement membrane and are avascular. They are rarely, if ever, associated with metastasis (Seigler and Setter, 1977). As a tumor becomes vascularized, the number of cells released into the circulation correlates with the density of blood vessels in the primary tumor. Thus, in experimental animals Liotta et al. (1976)found malignant cells in the effluent of tumors implanted in the mouse thigh, but only after new blood vessels had appeared in the tumor. The number of cells shed from the primary tumor correlated with the density of tumor blood vessels and with the number of lung metastases observed later. Another clinical phenomenon probably related to angiogenesis is the breakdown of the blood-brain barrier observed in primary tumors of the brain as well as in metastases. The protein leakage and local edema are thought to be due to the uniquely permeable ultrastructure of tumor capillaries (Long, 1979). These capillaries tend to remain undifferentiated and immature and generally lack smooth muscle cells (Ausprunk and Folkman, 1977). For this reason the vascular bed of a large tumor may fail to respond to epinephrine or other vasoconstrictors; it is sometimes difficult to control bleeding from a large tumor bed at the time of surgery. Tumor dormancy is another clinical phenomenon in which angiogenesis may play a role, although there is less evidence. For example, a metastasis may appear in the lung 5 years after removal of a rapidly growing Wilms’ tumor in a child. A metastasis may appear 15-20 years after removal of a primary breast cancer. Where were these tumor cells during the long quiescent period? The exact nature of this dormancy is one aspect of the meta-

198

JUDAH FOLKMAN

static process that is most obscure. Several hypotheses have been proposed. One is that tumor cells could be arrested in a subendothelial position. In the rabbit ear chamber, this has actually been observed; endothelial cells sealed over a gap after tumor cells had passed through it. Another possibility is that once tumor cells have escaped from the vascular lumen, they might grow to a small spheroidal mass of a few millimeters, but remain in the prevascular phase for a prolonged period. During this period, they might, for various reasons, be unable to stimulate angiogenesis and thus a small population of tumor cells could lie dormant even though the cells themselves had not stopped proliferating. This is an area where speculation is plentiful but where experimental data would be most welcome. XII. Summary

The hypothesis that tumors are angiogenesis dependent has, in the past decade, generated new investigations designed to elucidate the mechanism of angiogenesis itself. Many laboratories are now engaged in this pursuit. Some are studying angiogenesis that occurs in physiological situations, whereas others are interested in angiogenesis that dominates pathological conditions. These efforts have led to (1)the development of bioassays for angiogenesis; (2) the partial purification and, in one case, the complete purification of angiogenic factors from neoplastic and non-neoplastic cells; (3)the development of new polymer technology for the sustained release of these factors and other macromolecules in uiuo; (4)the cloning and long-term culture of capillary endothelial cells; (5)the demonstration of the role of nonendothelial cells, such as mast cells in modulating angiogenesis; (6) the discovery of angiogenesis inhibitors; and (7)the demonstration that certain animal tumors will regress when angiogenesis is inhibited. The effects of angiogenesis inhibitors provide perhaps the most compelling evidence for the role of angiogenesis in tumor growth. It is conceivable that the original effort to understand the role of angiogenesis in tumor growth will also lead to the use of angiogenesis inhibitors as a new class of pharmacologic agents in a variety of non-neoplastic diseases such as arthritis, psoriasis, and ocular neovascularization. However, much work remains to be done before it will be possible to understand (1)the regulatory systems that govern capillary density in normal tissues; (2) the factors that maintain the viability of microvascular endothelium; (3)the development of the vascular system itself; and (4)the mechanism by which vascular regression occurs, both in the embryo and in the postnatal organism. A knowledge of the mechanisms which underlie these normal processes may help to enlarge our comprehension of tumor angiogenesis.

TUMOR ANGIOGENESIS

199

ACKNOWLEDGMENTS This work was supported by USPHS Grant RO1-CAI4019 from the National Cancer Institute, by a grant to Harvard University from the Monsanto Company, by USPHS Grants GM-25810 and EY04002 (to Robert Langer), and by contributions from the Franzheim synergy Trust.

REFERENCES Adelman, N. E., Ksiazek, J., Yoshida, T., and Cohen, S. (1980).J . Zmmunol. 124, 825-830. Algire, G. H . , Chalkley, H. W., Legallais, F. Y., and Park, H. D. (1945).1. Natl. Cancerlnst. 6, 73-85. Ando, T., and Watanabe, S. (1969). Int. J . Protein Res. I, 221-224. Auerbach, R. (1981). Lymphokines 4, 69-88. Auerbach, R., and Sidky, Y. A. (1979).J . Zmmunol. 123, 751-754. Ausprunk, D. H . , and Folkman, J. (1977). Microuasc. Res. 14, 53-65. Azizkhan, R. G., Azizkhan, J. C., Zetter, B. R., and Folkman, J. (1980).J . Erp. Med. 152,931934. Banda, M. J., Knighton, D. R., Hunt, T. K., and Werb, Z. (1982). Proc. Natl. Acad. Sci. U.S.A. 79, 7773-7777. Bar, Th. (1980).In “Advances in Anatomy, Embryology, and Cell Biology” (A. Brodal, W. Hild, and J. van Limborgh, eds.), pp. 1-62. Springer-Verlag, Berlin and New York. Bar, Th., and Wolff, J. R. (1972). Z . Zelljiorsch. 133, 231-248. BenEzra, D. (1978). Am. J . Ophthalmol. 86, 455-461. Brem, H., and Folkman, J. (1975).J . E r p . Med. 141, 427-439. Brem, S. S., Gullino, P. M., and Medina, D. (1977). Science 195, 880-882. Brem, S. S . , Jensen, H. M., and Gullino, P. M. (1978). Cancer 41, 239-244. Brown, G. C., Magargal, L. E., and Federman, J. L. (1980). Trans. Ophthalmol. Soc. U . K . 100, 377-380. Brown, R. A., Weiss, J. B., Tomlinson, I. W., Phillips, P., and Kumar, S. (1980). Lancet 1, 682-685. Bunning, R. A. D., Murphy, G., Kumar, S., Phillips, P., and Reynolds, J. J. (1984). Eur. Biochem. 139, 75-80. Carlsson, J., Stalnacke, C.-G., Acker, H . , Haji-Karim, M., Nilsson, S., and Larsson, B. (1979). Int. J . Radiat. Oncol. Biol. Phys. 5, 2011-2020. Castellot, J. J., Jr., Karnovsky, M. J.. and Spiegelman, B. M. (1980). Proc. Natl. Acad. Sci. U.S.A. 77, 6007-6011. Castellot, J . J., Jr., Karnovsky, M. J., and Spiegelman, B. M. (1982). Proc. Natl. Acad. Sci. U.S.A. 79, 5597-5601. Cawston, T. E., Galloway, W. A., Mercer, E., Murphy, G., and Reynolds, J. J. (1981). Biochem. J . 195, 159-165. Chodak, G. W., Haudenschild, C., Gittes, R. F., and Folkman, J. (1980).Ann. Surg. 192, 762771, 1980. Crum, R., and Folkman, J. (1984).J . Cell. B i d . 199 (Abst. 581), 158. D’Amore, P., Glaser, B. M., Brunson, S. K., and F’enselau, A. H. (1981).Proc. Natl. Acad. Sci. U.S.A. 78, 3068-3072. Deane, B. R., and Lantos, P. L. (1981).J. Neurol. Sci. 49, 67-77. Denekamp, J. (1982). Br. J . Cancer 45, 136-139. DeOme, K. B., Faulkin, L. J., Jr., Bern, H. A,, and Blair, P. B. (1959). Cancer Res. 19, 515520. Dunn, M. R., and Montgomery, P. O’B. (1957). Lab. Znoest. 6, 542-546.

200

JUDAH FOLKMAN

Dvorak, H. F., Quay, S. C., Orenstein, N. S., Dvorak, A., Hahn, P., and Bitzer, A. (1981). Science 212, 923-924. Ehrlich, P. (1879). Arch. Anat. Physiol. (Leipzig 3, 166-69. Eisenstein, R., Sorgente, N., Soble, L. W., Miller, A,, and Kuettner, K. E. (1973). Am. J . Pathol. 73, 765-774. Eisenstein, R . , Goren, S. B., Shumacker, B., and Choromokos, E. (1979). Am. /. Pathol. 88, 1005-1012. Farrow, G . M., Utz, D. C., Rife, C. C., and Greene, L. (1977). Cancer Res. 37, 2794-2798. Feder, J., Marasa, J. C., and Olander, J . V. (1983).J . Cell. Physiol. 116, 1-6. Fenselau, J. (1984).In “Growth and Maturation Factors” (G. Guroff, ed.), Vol. 2. Wiley, New York. Fenselau, A,, and Mello, R. J. (1976). Cancer Res. 36, 3269-3273. Fenselau, A,, Kaiser, D., and Wallis, K. (1981a).J . Cell. Physiol. 108, 375-384. Fenselau, A,, Watt, S.,and Mello, R. J. (1981b).J . B i d . Chem. 256, 9605-9611. Folkman, J . (1970).In “Cancer of the Colon and Antecedent Epithelium” (W. J. Burdette, ed.), pp. 113-127. Thomas, Springfield, Illinois. Folkman, J. (1972).Ann. Surg. 175, 409-416. Folkman, J , , and Cotran, R. S. (1976). Int. Reo. E x p . Pathol. 16, 207-248. Folkman, J., and Haudenschild, C. (1980). Nature (London) 288, 551-556. Folkman, J., and Hochberg, M. (1973).J . E x p . Med. 138, 745-753. Folkman, J., Long, D. M., and Becker, F. F. (1963). Cancer 16, 453-467. Folkman, J., Cole, P., and Zimmerman, S. (1966). Ann. Surg. 164, 491-502. Folkman, J., Merler, E., Abernathy, C., and Williams, G . (1971).J . E x p . Med. 133, 275-288. Folkman, J., Hochberg, M., and Knighton, D. (1974). In “Control of Proliferation in Animal Cells” (B. Clarkson and R. Baserga, eds.), pp. 833-842. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York. Folkman, J., Haudenschild, C. C., and Zetter, B. R. (1979). Proc. Natl. Acad. Sci. U.S.A. 76, 5217-5221. Folkman, J., Ausprunk, D., and Langer R. (1980). In “Textbook of Rheumatology” (W. N. Kelly, E. D. Harris, Jr., S. Ruddy, and C. B. Sledge, eds.), pp. 210-220. Saunders, Philadelphia, Pennsylvania. Folkman, J., Langer, R., Lindhardt, R. J., Haudenschild, C., and Taylor, S. (1983). Science 221, 719-725. Form, D. M . , and Auerbach, R. (1983). Proc. Soc. Exp. Biol. Med. 172, 214-218. Fournier, G . A., Lutty, G. A., Watt, S., Fenselau, A., and Patz, A. (1981).Znoest. Ophthalmol. Visual Sci. 21, 351-354. Frederick, J. L., Shimanaki, T., and DiZevega, G. S. (1984). Science 224, 389-390. Fromer, C. H., and Klinhvorth, G. K. (1975). A m J . Pathol. 79, 537-550. Gabrielides, C., and Rifkin, D. B. (1983).J . Cell B i d . 97, 460. Giani, E. (1964). G . Ital. Chir. (Naples) 20, 95-109. Gimbrone, M. A., Jr.. and Gullino, P. M. (1976). Cancer Res. 36, 2611-2620. Gimbrone, M. A., Jr., Aster, R. H., Cotran, R. S., Corkery, J., Jandl, J., and Folkman, J. (1969). Nature (London) 222, 33-36. Gimbrone, M. A., Jr.. Leapman, S. B., Cotran, R. S., and Folkman, J. (1972).J . Exp. Med. 136, 261-276. Gimbrone, M. A., Jr., Cotran, R. S., Leapman, S. B., and Folkman, J. (1974).J . Natl. Cancer Inst. 52, 413-427. Glaser, B. M., D’Amore, P. A., Michels, R., Patz, A., and Fenselau, A. (1980a).J . Cell Bid. 84, 298-304. Glaser, B. M., D’Amore, P. A., Michels, R., Brunson, S., Fenselau, A., Rice, T., and Patz, A. (1980b). Ophthalmology 87, 440-446.

TUMOR ANGIOGENESIS

201

Goldacre, R. J., and Sylven, B. (1962). Br. J . Cancer 16, 306-322. Gospodarowicz, D., and Thakral, K. K. (1978). Proc. Natl. Acad. Sci. U.S.A. 75, 847-851. Green, H., and Kehinde, 0. (1976) Cell 7, 105-133. Greenblatt, M., and Shubik, P. (1968).J. Natl. Cancer Znst. 41, 111-124. Gross, J., Azizkhan, R. G., Biswas, C., Bruns, R., Hsieh, D., and Folkman, J. (1981). Proc. Natl. Acad. Sci. U.S.A. 78, 1176-1180. Gross, J. L., Moscatelli, D., and Rifkin, D. B. (1983). Proc. Natl. Acad. Sci. U.S.A.80, 26232627. Gullino, P. M . , and Grantham, F. H. (1964). Cancer Res. 24, 1727-1732. Hicks, R. M. (1977). Cancer Res. 37, 2822-2823. Hirano, A,, and Matsui, T. (1975). Hum. Pathol. 6, 611-621. Hunt, T. K., Andrews, W. S., Halliday, B., Greenberg, G., Knighton, D., Clark, R., and Thakral, K. (1981). I n “The Surgical W o u n d (P. Dineen and G. Hildick-Smith, eds.), pp. 1-18. Lea & Febiger, Philadelphia, Pennsylvania. Huseby, R. A., Currie, C., Lagerborg, V. A., and Garb, S. (1975). Microoasc. Res. 10, 396413. Jakob, W., Jentzsch, K. D., Mauersberger, B., and Oehme, P. (1977). E x p . Pathol. Bd. 13, 231-236. Kalebic, T., Garbisa, B., Glaser, B., and Liotta, L. A. (1983). Science 221, 281-283. Kaminski, M., Majewski, S., Kaminska, G., Bem, W., and Szmurlo, A. (1975).Arch. Immunol. Ther. E x p . 26, 1075-1078. Kaminski, M., Kaminska, G., and Majewski, S. (1978a). Folio Biol. 24, 104-110. Kaminski, M., Kaminska, G., and Majewski, S. (1978b). Experientia 34, 490-491. Keegan, A., Hill, C., Kumar, S., Phillips, P., Kumar, S., and Weiss, J. (1982).J . Cell Sci. 55, 261-276. Kessler, D. A., Langer, R. S., Pless, N. A., and Folkman, J. (1976). Znt. J . Cancer 18,703-709. Kissun, R. D., Hill, C. R., Garner, A., Phillips, P., Kumar, S., and Weiss, J. B. (1982). Br. J . Opthalmol. 66, 165-169. Klagsbrun, M., Knighton, D., and Folkman, J. (1979). Cancer Res. 36, 110-114. Knighton, D., Ausprunk, D., Tapper D., and Folkman, J. (1977). Br. J . Cancer 35, 347-356. Knighton, D. R., Hunt, T. K., Scheunenstuhl, H., Halliday, B. J., Werb, Z., and Banda, M. J. (1983). Science 221, 1283-1285. Kuettner, K. E . , and Pauli, B. U. (1983).In “Cartilage Structure, Function, and Biochemistry” (B. K. Hall, ed.), Vol. 1, pp. 281-312. Academic Press, New York. Kuettner, K. E., Soble, L., Croxen, R. L., Marczynska, B., Hiti, J . , and Harper, E. (1977). Science 196,653-654. Kuettner, K. E., Pauli, B. U., and Soble, L. (1978). Cancer Res. 38, 277-287. Kumar, S., West, D., Daniel, M., Hancock, A., and Carr, T. (1983a).Znt. J . Cancer 32, 461464. Kumar, S., Shahabuddin, S., Haboubi, N., West, D., Arnold, F., Reid, H., and Carr, T. (1983b). Lancet 2: 364-368. Langer, R., and Folkman, J. (1976). Nature (London) 263, 797-800. Langer, R., Brem, H., Falterman, K., Klein, M., and Folkman, J. (1976). Science 193, 70-72. Langer, R., Conn, H., Vacanti, J., Haudenschild, C., and Folkman, J. (1980). Proc. Natl. Acad. Sci. U.S.A. 77, 4331-4335. Langer, R., and Folkman, J. (1976). Nature (London) 263, 797-800. Leighton, J. (1967). In “The Spread of Cancer.” Academic Press, New York. Liotta, L. A,, Kleinerman, J., and Saidel, G. M. (1976). Cancer Res. 36, 889-894. Long, D. M. (1979).J. Neurosurg. 51, 53-58. McAuslan, B. R. (1980). In “Control Mechanisms in Animal Cells” (L. Jimenez de Ausua, R. Levi-Montalcini, and R. Shields, eds.), pp. 285-292. Raven, New York.

202

J U D A H FOLKMAN

McAuslan, B. R., and Hoffman, H. (1979). Exp. Cell Res. 119, 181-190. Maciag, T., Kadish, J., Wilkins, L., Stemerman, M. B., and Weinstein, R. (1982). J . Cell B i d . 94, 511-520. Madri, J.. and Williams, S. K. (1983). J . Cell B i d . 97, 153-165. Maiorana, A., and Gullino, P. M. (1978). Cancer Res. 38, 4409-4414. Mantovani, A. (1982). Ado. Pharmucol. Chemother, 19, 35-66. Matsuno, H. (1981). Nerrrol. Med. Chir. (Tokyo) 21, 765-773. Medina, D. (1973). In “Methods in Cancer Reserch” (H. Busch, ed.), pp. 3-53. Academic Press, New York. Montesano, R . , Orci, L., and Vassalli, P. (1983). J . Cell B i d . 97, 1648-1652. Moscatelli, D., Cross, J. L., and Rifkin, D. B. (1981). J . Cell Biol. 91, 201a. Murray, J. B., Brown, L., Langer, R., and Klagsbrun, M. (1983a). In Vitro 19, 743-748. Murray, J. B., Krochin, N . , Hill, C., Zetter, B., and Langer, R. (1983b). Fed. Proc. Fed. Am. SOC.E x p . Biol. 42, 1889. Muthukkaruppan, Vr., and Auerbach, R. (1979). Science 205, 1416-1417. Nicosia, R. F., Tchao, R., and Leighton, J. (1983). Cancer Res. 43, 2159-2166. Nishioka, K., and Ryan, T. J. (1972). J . Znoest. Denmtol. 58, 33-45. Paskins-Hurlburt, A. J., Hollenberg, N . K., and Abranis, H. L. (1982). Microoasc. Res. 24, 1524. Peters, W., Teixera, M., Intaglietta, M., ad Cross, J. F. (1980). J . Natl. Cancer Znst. 65, 631642. Peterson H.-I., ed. (1979). “Tumor Blood Circulation, Angiogenesis, Vascular Morphology and Blood Flow of Experimental Human Tumors,” pp. 103-135. CRC Press, Boca Raton, Florida. Phillips, P., Steward, J . K., and Kumar, S. (1976). Znt. J . Cancer 17, 549-558. Polverini, P. J.. and Leibovich, S. J. (1984). Laboratory Znuestigation, December. Polverini, P. J., Cotran, R. S , Cimbrone, M. A,, Jr., and Unanue, E. R. (1977). Nature (London) 269, 804-806. Potts, M.,Doppelt, S., Taylor, S., Folkman, J., Neer, R., and Potts, J. T., Jr. (1984). CalciJ Tissue Znt. 36, 189-193. Pousa, S. L., Acebes, J . , and Ferrer, I. (1979). Molfol. Normal Pathol. Sect. B 3, 541-550. Raju, K. S., Alessandri, G., Ziche, M., and Gullino, P. M. (1982). J . Natl. Cancer Znst. 69, 1183- 1188. Rifitin, D. B., Cross, J. L., and Moscatelli, D. (1982). In “Pathobiology ofthe Endothelial Cell” (H. L. Nossel and H. J. Vogel, eds.), pp. 191-197. Academic Press, New York. Ryan, T. J. (1970). Br. J. Dermatol. 82, 99-111. Sandison, J. C. (1928). Am. J . Anat. 41, 447-473. Schor, A. M., Schor, S. L., Weiss, J. B., Brown, R . , Kumar, S . , and Phillips, P. (1980). Br. 1. Cancer 41, 790-799. Seigler, H. F., and Setter, B. F. (1977). Ann. Surg. 186, 1-12. Shahabuddin, S., and Kumar, S. (1983). Br. J . Ophthalmol. 67, 286-291. Shing, Y., Folkman, J., Murray, J., and Klagsbrun, M. (1983). J . Cell B i d . 97, 295a. Shing, Y., Folkman, J., Sullivan, R., Butterfield, C., Murray, J., and Klagsbrun, M. (1984). Science, 223:1296-1298. Sidky, Y. A., and Auerbach, R. (1975). J . Exp. Med. 141, 1084-1100. Smith, S. S., and Basu, P. K. (1970). Can. J . Ophthalmol. 5, 175-183. Sorgente, N., Kuettner, K. E., Soble, L. W., and Eisenstein, R. (1975). Lab. Znoest. 32, 217222. Sorgente, N., Kuettner, K., and Eisenstein, R. (1976). Proc. Colloq., 23rd, Brugge, 1975 23, 227-230.

TUMOR ANGIOGENESIS

203

Stan, A., and Mattingly, R. F. (1975). Am. J. Obstet. Gynecol. 121, 845-852. Stenzinger, W., Bruggen, J.. Macher, E., and Sorg, C. (1983). Eur. J. Clin. Oncol. 19, 649656.

Suddith, R. L., Kelly, P. J., Hutchinson, H. T., Murray, E., and Haber, B. (1975).Science 190, 682-684. Sutherland, R. M., McCredie, J. A., and Inch, W. R. (1971).J. Natl. CancerInst. 46, 113-117. Swabb, E. A., Wei, J., and Gullino, P. M . (1974). Cancer Res. 34, 2814-2822. Tannock, I. F. (1968). Br. J . Cancer 22, 258-273. Taylor, S., and Folkman, J. (1982). Nature (London) 297, 307-312. Thakral, K. K . , Goodson, W. H., and Hunt, T. K. (1979).J. Surg. Res. 26, 430-436. Thoresen, S . , Tangen, M., and Hartveit, F. (1982). Diagn. Histopathol. 5, 65-67. Tolbert, W. R . , Kuo, M. J., and Feder, J. (1981). In Vitro 17, 259 (abstract 242). Tuan, D., Smith, S., Folkman, J., and Merler, E. (1973). Biochemistry 12, 3159-3165. Wagner, R . C. (1980). Ado. Microcirc. 9, 45-75. Warren, B. A. (1970). Microvasc. Res. 2, 443-453. Warren B. A . , Greenblatt, M., and Kommineni, V. R. C. (1972). Br. J. E x p . Pathol. 53, 216224. Watt, S. L. (1981). Ph.D. thesis, Johns Hopkins University. Wissler, J. H, (1982). In “Biochemistry of Differentiation and Morphogenesis” (L. Jaenicke, ed.), pp. 257-274. Springer-Verlag. Berlin and New York. Wolf, J. E., and Harrison R. G. (1973). J. Invest. Dermatol. 61, 130-141. Yamaura, H., and Sato, H. (1974).J. Natl. Cancer Znst. 53, 1229-1240. Young, J. S . , Lumsden, C. E., and Stalker, A. L. (1950). J. Pathol. Bacteriol. 62, 313-333. Ziche, M., and Gullino, P. M. (1981). Cancer Res. 41, 5060-5063. Ziche, M., and Gullino, P. M. (1982).J. Natl. Cancer Znst. 69, 483-487. Ziche, M., Jones, J.. and Gulling, P. M. (1982).J. Natl. Cancer Znst. 69, 457-482.

This Page Intentionally Left Blank

FUSION PROTEINS IN RETROVIRAL TRANSFORMATION Karin Moelling Max-Planck-lnstitutfur Molekulare Genetlk. Berlin, Federal Republic of Germany

I. Introduction .................... ................................... 11. Fusion Proteins of Avian Retroviruses ..................................... A. Myelocytomatosis Virus MC29. . . . . . . .............................

.............................

C. HB-1 and MH2

............................................... .................................

205 209 209 216 217 219 22 1 224 227 227 228 228 229 231

E. Myeloblastosis Virus and Erythroblastosis Virus ......................... F. Sarcoma Viruses .................................................... 111. Fusion Proteins of Mammalian Viruses .................................... A. Abelson Murine Leukemia Virus ......................... B. Murine Osteosarcoma Viruses. ........................... C. 3611 Fibrosarcoma Virus . . . . . . . . . . . D. Feline Sarcoma Viruses.. ............................................ E. Simian Sarcoma Virus. . . . . . ........................ IV. Discussion on the Effect of Viral Structural Proteins on Transformation . . . . . . . . 233 References.. ........................................................... 234

I. Introduction

Retroviral transforming proteins are coded for by oncogenes. These are dominant genes that cause carcinogenesis. If they are transferred into a normal cell environment, they are able to induce oncogenic transformation. Oncogenes are closely related to normal eukaryotic genes which are highly conserved during evolution and therefore presumably fulfill an essential normal function in the cell. In all animal species, genes homologous to oncogenes have been identified, including Drosophila and even yeast (Shilo and Weinberg, 1981; Gallwitz et al., 1983; DeFeo-Jones et al., 1983). Originally oncogenes were identified as transforming genes of retroviruses (v-oncs) where they were detected first. Retroviruses can be considered as vectors or carriers of oncogenes which allow the transfer of an oncogene into a cell. Oncogenes become incorporated into retroviral genomes almost always at the expense of the viral replicative genes gag, pol, and env. The resulting tumorigenic agent can transform cells, but for the production of infectious viral progeny the same cell must be infected with a complete set of the three replicative viral genes that can complement the defects of the transforming virus. While the cellular progenitor genes of the oncogenes (c-oncs) usually consist of introns and exons, viral oncogenes never contain a 205 ADVANCES IN CANCER RESEARCH, VOL. 43

Copyright 0 1985 by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12-006643-2

SUMMARY

Viral strain

Animal species

myc

MH2. OKlO

Chicken

Agag-myc Agag-myc' e Agag- Apol-myc Agog-mil myb Agag-myb-ets Agog-erbA erbB

MC29, CMII HB-1 OKlO MH2 AMV E26-E4 AEV, AEV-H AEV

src

RSV, PR-C, BH FSV, PRCII, URl Y73, Esh UR2

Oncogene

Agag- yes Agag-ros re1

Chicken

Protein

Modificationb

Locationc

Propertied

Relationship

p55

PP

P110, P90 p108 P200 PlOO P45 p135 P75 p62-p74

PP PP PP PP

DNA bp DNA bp DNAIRNA bp DNA/RNA bp, PK' DNA bp(?) DNA bp

gP, PP

PK(?)

src, EGFR

p60

PP

PK

p130/140

PP

p80/90 P a Turkey

TABLE I OF ONCOGENES"

N

C

Disease

E1A

M yelocytomas, sarcomas, carcinomas Rapid lymphomas

raf croII, A rep

EGFR

Carcinoma M yeloblastosis Erythroblastosis Erythroblastosis Sarcomas, carcinomas Sarcomas

PK

src

Sarcomas

PK PK

src src

Sarcomas Sarcomas Reticuloendotheliosis

DNA bp

VWS

Agag-abl Agag- fos Agag-raf ras

Agog-fes Agag-fgr Agag-fm Agog-abl Agag-sis Aenv-sis

Moloney SV Mouse Abelson FBR osteosarcoma 3611 KirstenRat Harvey MuSV ST, GA-FeSV Cat GR-FeS SM-FeS HZ-2 FeS PI-FeSV Primate

p37 PW-p160 P55

PP PP, m PP

src N

p90/75 p21/29

gplpplm?

C

p85-110 P70 p180-120 P98 P75 p28

pp PP gp PP PP

PK

RNA/DNA bp, PK* GTP bp

mil, src

PK

fps src

M

Fibrosarcoma

PK PDGF PDGF

The names of oncogenes and the viral strains are explained in the text. The molecular weights of the proteins are indicated in kilodaltons. m, M yristilated; pp, phosphorylated; gp, glycosylated. N, Nuclear; C, cytoplasmic; M, membrane. d DNA bp, DNA-binding proteins; DNAIRNA bp, DNAlRNA-binding proteins; PK, tyrosine protein kinase activity; PK*, serine/threonine protein kinase; GTP bp, GTP-binding protein. emyc’ indicates that it is not identical to v-myc. 0

b

208

KARIN MOELLING

complete intron and at best a few nucleotides (e.g., myc). Therefore, the incorporation of a cellular oncogene into a retroviral genome most likely took place by a genetic interaction between cellular messenger RNA and viral genetic sequences. Since retroviruses are diploid, full-length genomic retroviral progeny RNA transcribed from its DNA provirus and a transcript of the cellular progenitor oncogene could be packaged into one virus particle. Upon reverse transcription in the next infectious cycle, recombination between viral and cellular sequences would take place leading to the acquisition of the oncogene by the virus and loss of viral replicating genes. Since the reverse transcriptase shows a high degree of infidelity, it may have helped to modify the cellular gene toward higher oncogenicity. Except for one case, all viral oncogenes differ from their homologous cellular progenitor genes (for review see Duesberg, 1983). The mechanism of uptake of oncogenes by retroviruses often results in fusion with replicative genes. A frequently encountered situation is the fusion of an oncogene with the 5‘ portion of the gag gene that then becomes the 5’ end of a gag-onc fusion gene. The 5’ end of gag codes for initiator and splice signals which are thus supplied to the gag-onc fusion genes. Only in rare cases are oncogenes inserted into retroviral genomes exactly at the junction of viral genes (e.g., src in RSV and myc in MH2 and OK10, see below). Their expression then is regulated by splice signals and results in subgenomic messages. In a few cases oncogenes are fused to residual en0 sequences (e,g., sis and myb). Three viruses are known so far that have picked up two cellular genes and thus carry two potential oncogenes (MH2, E26, AEV). The total number of oncogenes is unknown. The number in retroviruses appears to be very limited. Twenty-two viral oncogenes are known at present. Some of them have been isolated independently in different viruses. For instance, myc is found in four viral strains originally isolated in Bulgaria (MC29), Germany (CMII), Finland (OKlO), and Great Britain (MH2). It is not very likely that a large reservoir of different oncogenes would give rise to these independent isolates. A second argument supporting that the number of oncogenes is limited comes from the observation that many of them are related. For instance, all the protein kinases and several without enzyme activity show clear homologies (src,f p s , fes, yes, r-os, abl, erbB, fm, sis, raf, ras), suggesting they are derived from the same ancestral sequences. Most surprising is the homology between the oncogenes of a natural avian virus isolate (mil from MH2) with that of a transforming virus created in tissue culture from a nontransforming murine virus (raf, 3611) and the presence of the oncogenes sis and abl in viruses of different animal species (see below). There are, however, other oncogenes known which have never been detected in retroviruses. One such oncogene was identified by transfection of

FUSION PROTEINS IN RETROVIFML TRANSFORMATION

209

DNA from tumor cells into recipient tissue culture cells, the B-lym gene (Goubin et al., 1983). It is possible that retroviruses select for certain oncogenes and therefore lead to a wrong estimate of the total number of oncogenes. Many more oncogenes may be detected independent of transduction by retroviruses, e.g., by transfection of tumor DNA into appropriate recipient cells or by analysis of elevated RNA transcripts in tumor cellswhich seems to be an enormous number (Schutzbank et al., 1982). Fusion of oncogenes to retroviral replicative genes generates protein products that are also fused. These proteins contain viral and oncogene information in a single polypeptide chain. The viral structural portion provides a convenient antigenic marker that allows detection of such oncogene products with antiviral antibodies which are easily available. In contrast, oncogene products are generally poorly antigenic because they closely resemble cellular proteins. Therefore, oncogene products without viral structural portions are much more difficult to detect. Only recently have specific antibodies become available either by overproduction of oncogene products in bacteria by molecular cloning procedures or by synthesis of peptides with sequences derived from known nucleotide sequences of oncogenes. Besides providing a very useful antigenic marker, the viral structural sequences have an essential effect on the oncogenic potential of at least one fusion protein, as has recently been shown for the gag-abl fusion protein of Abelson murine leukemia virus (A-MuLV). The known retroviral fusion proteins will be discussed below emphasizing the potential role of the viral structural protein portions on the transforming proteins. Table I summarizes the oncogenes, their viral origin, their gene products, and some characteristic properties. Details are explained in the text. Some recent reviews will be useful for further details (Bishop, 1983; Bister, 1983; Enrietto and Wyke, 1983; Graf and Stehelin, 1982; Muller and Verma, 1984). II. Fusion Proteins of Avian Retroviruses

The avian retroviruses comprise four groups; three consist of the acute avian leukemia viruses such as the myelocytomatosis virus family (MC29), the avian erythroblastosis viruses (AEV), and the avian myeloblastosis viruses (AMV). The fourth group is represented by the avian sarcoma viruses. The gene products of the three groups of acute avian leukemia viruses are shown in Fig. 1. A. MYELOCYTOMATOSISVIRUS MC29 The acute avian leukemia virus MC29 induces a broad spectrum of neoplastic diseases in fowl including myelocytomatosis, sarcomas, and car-

210

KARIN MOELLING

Viruses

Genome Structure

Protein ( 8 )

MC 29 MC29

Agag

my c

Aenv

pllOA~ag-mYc

CMII

Agag

my c

Aenv

pgoAgag-mYc

MH2

Agag

mil

my c

plOOA~ag-mil, p55m~c

OK10

Agag

ApOl

myc Aenv

~

MC29 mutants

'gag

HB-1 revertant

'gag

AEV-ES4

Agag

AEV-H

Agag

Amyc v / c myc

~

gag-Apol-myc 7 p200 6 ~ ~ ~ , . p55myC

Aenv Aenv

AEV

erbA

erbB

Apol

erbB

p75Agag-erbA

68erbB

* P

~ ~ 7 6 'p7ZerbB ~ ~ .

AMV AMV

Agag

pol

my b

pr76gag, pr180gag-po1

E26

Agag

myb

ets

p135A~a8-myb-ets

FIG. 1. Schematic diagram of the genornes of acute avian leukemia viruses and their protein products. Partially deleted genes are indicated by A. p, Protein, pr, precursor.

cinomas of the kidney and liver (Beard, 1980; Grafand Beug, 1978; Ivanov et al., 1964; Mladenov et al., 1967). MC29 codes for an oncogene Agag-myc consisting of a deleted Agag and a transforming region rnyc and additional sequences corresponding to a deleted Aenv region. The Agag-myc gene product is a fusion protein consisting of 450 amino acids derived from the amino terminus of gag, followed by 422 amino acids that are specific for the myc region. Thus, the gag region contains the entire p19 and p10 but only the first 211 amino acids of p27 followed by the myc sequence. The mycspecific portion corresponds to a molecule of 49,000 MW which is smaller than the equivalent myc gene since a termination codon is located within the rnyc gene, 300 bases upstream of its 3' end (Reddy et al., 1983). The open reading frame of the gag-myc gene codes for a polypeptide of 875 amino acids with a molecular weight of 96,000. This is in close agreement with the observed size of 110,000 MW for the fusion protein synthesized in MC29infected quail nonproducer cell lines, designated as p l l 0 Agag--myc (Bister et al., 1977). The difference in molecular weight is probably due to posttranslational modifications such as phosphorylation. Mapping of the tryptic phosphopeptides of pllOAgag--myc has revealed that the myc domain of the pro-

FUSION PROTEINS IN RETROVIRAL TRANSFORMATION

211

tein is heavily phosphorylated on serine and threonine (Ramsay et al., 1982b). Examination of the amino acid sequence of the transforming viral myc protein reveals that the carboxy terminus is highly hydrophilic, containing a large number of glutamic acid, arginine (13%), and lysine residues (10% of the carboxy terminal 150 amino acids). These numbers have strong similarities to the amino acid composition of certain histones. The basic portion of the pl10Agag-*71yc may account for its ability to bind to DNA (Donner et al., 1982). In contrast, the amino terminus of the protein is more hydrophobic. The two regions are joined to a large number of proline residues, indicating that these two regions of the polypeptide are linked by a highly flexible region. This structure is reminiscent of the hinge region present in immunoglobulin molecules at the junction of the variable and the constant regions. It is interesting to speculate that these two biochemically distinct domains-which are derived from two different exons of the cellular myc gene-may have different functions (Reddy et al., 1983)(see Fig. 2 below).

1 . MC29 Deletion Mutants Three partial transformation-defective deletion mutants of MC29 have been isolated which, even though they still transform fibroblasts in uitro, are no longer tumorigenic in the bird and do not transform macrophages in uitro (Ramsay et al., 1980; Enrietto et al., 1982). The mutants, designated as QIOA, QlOC, and Q I O H , code for fusion proteins p100, p95, and p9OAgUgAtnyc,respectively. The deletions of the mutants map around the center of the myc-specific region corresponding to the junction site of the two exons (Ramsay and Hayman, 1982). A characteristic ClaI cleavage site is lost in all three mutants, (Enrietto and Hayman, 1982);the deletions extend about 200 to 600 nucleotides in the 3’ direction (Bister et al., 1982). Two of the three major sites of phosphorylation apparently reside within the region deleted in the mutants (Ramsay et al., 1982a).

2 . Properties of

p 1 l O A g a g - m ~ in ~

Vitro

The pl10Ag0g-7n~c fusion protein has been characterized recently. It is not a glycoprotein and does not phosphorylate immunoglobulins in uitro, therefore it does not appear to be a protein kinase (unpublished observation). The pllOAgug-m~cprotein has been analyzed by anti-gag antibodies and by monoclonal antibodies against p19 (Greiser-Wilke et al., 1981), since anti-myc antibodies have only become available very recently (see below). The pll0Agag-m~~ is located in the nucleus of transformed fibroblasts as was shown by indirect immunofluorescence (Donner et al., 1982; Abrams et al., 1982) Subcellular fractionation analysis indicated that 60%of the p l lWgna-

212

KARIN MOELLING

wc protein was located in the nucleus of fibroblasts as well as bone marrow cells, which are the actual target cells in uiuo (Donner et al., 1982, 1983; Abrams et al., 1982). Presence of some p l l W @ ~ m yinc other cellular compartments, e.g., in the plasma membrane, has not yet been shown but may be possible. pllOAgag-myc rapidly migrates to the nucleus, within 1or 2 hr, as tested in pulse-chase experiments (Bunte et a[., 1982). A small fraction of about 16% of pl1OAgag-vlyc was found associated with the chromatin (Bunte et al., 1982). Cell cycle-dependent fluorescence analysis using nitrous acid to block mitosis indicated diffusion of the myc gene product into the cytoplasm during mitosis (Greiser-Wilke et al., unpublished observation). Treatment of cells by detergent and DNase resulted in some release of the p l l 0 protein. Subsequent application of RNase and high salt still retained myc protein in the residual nuclear matrix (Moelling et al., 1984a). Association of the myc protein with the nuclear matrix has been observed by Eisenman and collaborators (Hann et al., 1983). Recently O'Malley and collaborators showed that expression of genes is not sufficiently characterized by DNase I sensitivity of chromosomal doxains but that actively expressed genes are associated with the nuclear matrix (Ciejek et al., 1983). Further analyses of the myc gene product, in particular of the roles of the two functional domains, are required. Some properties of the viral pllOASag-nzYchave been studied in uitro. The protein was purified by immunoahity column chromatography using the immunoglobulin fraction of the monoclonal antibodies against p19 which was coupled to Sepharose CL-4B by covalent linkage using suberimidate. Efficiency of purification was about 4000-fold. The purified protein binds to double-stranded normal cellular DNA in uitro (Donner et al., 1982). DNAprotein interaction appears to be of biological significance for the oncogenic potential of the myc gene product, since the fusion proteins isolated from cells transformed with the three transformation-defective deletion mutants mentioned above exhibit reduced abilities to bind to DNA in uitro (Donner et al., 1983). Since the deletions of the mutants map around the ClaI site, this domain is expected to be involved in DNA binding and oncogenicity for bone marrow cells in the animal. Whether phosphorylation of this domain of the myc gene product plays any role is not yet known (Ramsay et al., 1982a). It is also unclear as yet why all three deletion mutants are still capable of transforming fibroblasts in vitro. We have attempted to attribute enzymatic functions to the nayc gene product in the hope of understanding its role during transformation. The purified protein is not a protein kinase, a GTPbinding protein, or an ATP-binding protein, It does not methylate DNA or inhibit DNA-methylase activities (Bunte et al., unpublished observations). It is not known whether the myc protein has any ATPase activity. Attempts

FUSION PROTEINS IN RETROVIRAL TRANSFORMATION

213

to identify a specificity of the myc protein for certain DNA sequences, e.g., regulatory genes such as LTR, 72 base pair (bp) repeats or the origin of Simian Virus 40 (SV40), have failed so far (Moelling et al., 1984b). These negative results may, however, be attributable to the limited amount of protein and artificial in vitro conditions. 3. Speculation on Functions of pllOAgag--myc

The pll@gag-"yC is expected to be involved in gene regulation. Computer analysis of the myc-specific sequences and search for homologies with other known genes suggested sequence homology of the central two-thirds of myc with the EIA gene of adenovirus (Ralston and Bishop, 1983). No homology was detected by computer search with the A repressor or the cro protein of A (Lipsick et al., 1984). The ElA gene has recently been shown to exert functions similar to the myc gene in multistep carcinogenesis. The ElA gene, as well as the myc gene, can transform primary rat fibroblasts if transfected in concert with other oncogenes, such as the EJ-Ha-ras gene (Ruley, 1983; Land et al., 1983). The ElA gene of adenovirus codes for a gene product that is required for the expression of other early adenovirus genes, E2 and E3, which lack enhancer sequences (Schrier et al., 1983). The ElA gene product substitutes for enhancers and regulates gene expression as a transcriptional activator. The notion that myc may function in the same way comes from preliminary experiments reported in Cold Spring Harbor (see Newmark, 1983). In transfection experiments ElA and myc genes can be used interchangeably and turn on the expression of the dihydrofolate reductase (DHFR) gene linked to the promoter of a heat-shock gene. Thus, both myc and ElA appear to be able to turn on gene expression (see Newmark, 1983). The E l A protein per se is not a DNA-binding protein; only in conjunction with other cellular proteins may it bind to DNA. It is also closely associated with the nuclear matrix (Feldman and Nevins, 1983). Using purified pllOAgag--myc protein, preliminary evidence showed that it inhibits transcription in a system consisting of an initiation-transcription complex from HeLa cells and DNA coding for adenovirus major late message (Davison et al., 1983). Presence of p l l O A ~ " ~ "did ~ " not affect a deletion upstream of the promoter. It is unclear whether these effects are due to interaction of the myc protein with DNA or the initiation complex (Moelling et al., 1984a). Inhibition of transcription has also recently been observed in another system. The tumorigenic strain of adenovirus, Ad12, inhibits expression of class I major histocompatibility genes thus allowing the infected tumor cell to escape immune surveillance (Schrier et al., 1983). Decreased synthesis of tropomyosin which is observed in quail cells trans-

214

KARIN MOELLING

formed by MC29, Schmidt-Ruppin D virus, and also in chemically transformed cells, can be interpreted as another example for inhibition of gene transcription by the effect of oncogenes (Hendricks and Weintraub, 1981). A speculative model of the myc gene product is shown in Fig. 2. The amino and carboxy terminal portions, corresponding to exons 2 and 3, are indicated as two independent functional domains which are separated by the ClaI site. The carboxy terminal domain is involved in DNA binding. Whether the amino terminal part is matrix-associated is highly speculative. Since DNA binding of the myc protein is reduced in the transformationdefective mutants, it is suggested that this function plays a role in tumorigenicity in the bird, whereas fibroblast transformation is independent of this property and therefore is tentatively assigned to the amino terminal domain.

4 . Specijc Antibodies against myc Proteins Recently myc-specific sera have become available. Antibodies from tumor-bearing animals had low titers and were of limited use (unpublished observation). However, since the entire nucleotide sequence of the viral and cellular myc genes became available (Alitalo et al., 1983a; Reddy et al., 1983; Watt et al., 1983), synthetic peptides were prepared according to predicted amino acid sequences. Nine and 12 carboxy terminal amino acids containing peptides of the viral myc and a peptide of 17 amino acids corresponding to the ClaI site proved to be useful immunogens (Hann et al., 1983; Moelling et al., 1984b; Bunte et al., 1984). The 12 carboxy terminal amino acids of the human c-myc were also used for antibody production and recognize the myc protein in human cells as well as the viral myc protein (Giallongo et al., 1983). Even though the 12 carboxy terminal amino acids of the viral and human myc only differ in two amino acids, the antibody against the viral myc

v-myc protein

DNA binding region

nuclear matrix associated (7)

&coo-

hydrophilic, histone-like

NH; transformation of fibroblasts

oncogenicity in the bird

FIG. 2. Schematic drawing of the structure of the myc protein. For explanations see text.

FUSION PROTEINS IN HETROVIRAL TRANSFORMATION

215

peptide does not recognize the human myc protein (Moelling et al., unpublished observation). An alternative approach for anti-myc antibody production was taken by molecular cloning of the myc gene in bacteria. Three groups succeeded so far in obtaining antibodies using this approach. All of them expressed myc as fusion proteins. A small portion of the amino terminal part of myc was fused with part of the viral src gene product (Alitalo et al., 1983b). According to unpublished observations, expression of the complete myc gene proved diEcult (Bunte et al., 1984), whereas the carboxy terminal half was expressed to high levels by two groups (Lautenberger et al., 1983; Bunte et al., 1984). Antibodies against the bacterially expressed proteins are more potent than the ones against the synthetic peptides and they cross-react with the human cellular myc protein (Moelling et al., unpublished observation). It will have to be seen how useful the bacterially expressed myc proteins are for functional analyses, since they are truncated, unphosphorylated, and in some cases insoluble. The viral and cellular myc proteins identified with these sera have been described with molecular weights ranging from 49,000 to 62,000. The predicted size of the molecule is 49,000. A molecule of this size was detected in human cells such as HeLa and Burkitt’s lymphoma cells (Giallongo et al., 1983). The viral myc proteins in avian cells were described in the literature with 62,000 (Hann et al., 1983), 57,000 (Alitalo et al., 1983b), and 55,000 MW (Moelling et al., 1984b; Bunte et al., 1984). The differences between the molecular weights may be due to different markers or gel systems. However, we have directly compared the viral and human cellular myc gene products and observed that the human cellular myc protein is larger (62,000 to 64,000 MW) than the viral myc protein, which we defined to be 55,000 (Benter et al., 1984). The myc protein of MH2 is a doublet, whereas the protein from OK10 migrates only as a single band. Both of them are phosphorylated (Hann et al., 1983; Moelling et al., 1984b), which may give rise to microheterogeneity. The normal cellular chicken myc protein differs from the myc protein that is activated indirectly in chicken lymphomas by promoter insertion (Hayward et d., 1981)in that only the one from lymphomas appears to be phosphorylated (Hann et al., 1983). Some indirect support that phosphorylation of myc may be relevant for transformation comes from the study of the partially defective MC29 mutants which lack phosphopeptides (Ramsay et al., 1982a). The p55V-myC behaves very similarly to the pllOAgag-pnyc.Both proteins are predominantly located in the nucleus of fibroblasts (Hann et id., 1983; Moelling et al., 1984b). The viral p55v-myChas recently been purified by immunoaffinity column chromatography with antipeptide antibodies. Its DNA binding site was located in the carboxy-terminal half of the myc pro-

216

KARIN MOELLING

tein, since antibodies against this domain inhibit DNA-protein interaction in oitro (Bunte et al., 1984). B. MC2g-RELATED VIRUSES CMII

AND

OK10

myc-related sequences have been identified in three MC29-related viruses, the myelocytomatosis virus CMII; Okerbloms isolate OKlO and the Mill-Hill virus No. 2, MH2. While CMII and OKlO induce similar pathogenic effects in the animal such as myelocytomatosis and endotheliomas, MH2 shows a somewhat altered spectrum: It causes carcinomas of liver and kidney (Graf and Beug, 1978). All three viral strains transform fibroblasts and macrophages in uitro (Beug et al., 1982). The myc sequences responsible for the oncogenicity of these viruses are expressed either in the form of a gag-myc fusion protein or as myc proteins devoid of viral structural proteins. CMII codes for a protein p90Agag--myc, which is closely related to the pllOAgag-irlycof MC29 and presumably lacks some gag sequences that give rise to the smaller size of the molecule. p9OAgarmyc behaves indistinguishably from pllOAgag-'nVc in all known properties in oiuo and in oitro (Bunte et al., 1982). The OKlO virus is an exceptional case in that it codes for the very large fusion protein p200Agag-Apo2-myc, in addition to the p55V-myC nonlinked to gag. Even though both of the myc gene products are coded for by identical sequences and therefore must be identical, they both reside in different cellular compartments. The majority (80%) of p200 is cytoplasmic (Hann et al., 1983; Bunte et al., 1984), in contrast to p55v-myC,which is predominantly located in the nucleus (70%). After purification the p200 protein binds not only to double-stranded DNA in oitro, but also to single-stranded RNA. Neither of the other myc proteins nor pr7f3'"g share this latter property. However, the RNA-dependent DNA polymerase behaves analogously and binds to DNA as well as RNA in oitro under similar conditions (Bunte et al., 1984). The p200Agag-A1J01--mycis the only example known among the avian viruses in which properties of the oncogene appear to be dramatically altered by the fused gaglpol portion. It is unknown at present whether the p200 protein contributes to the oncogenicity of OK10, whether it cooperates in some way with p55"-"Yc, or whether it is an irrelvant molecule. The other interesting oncogene that is influenced by its gag portion is the abl gene. In this case, the gag region is essential for transformation in lymphocytes (see below). Figure 3 shows a model on the mechanism of transformation by nuclear antigens. The bifunctional domains of rnyc are schematically indicated. The effect on the myc gene product is schematically shown to enhance or inhibit

FUSION PROTEINS IN RETROVIRAL TRANSFORMATION

nuclear matrix

’

217

a or

mRNA

chromatin

FIG. 3. Hypothetical model for the transformation mechanism of the three known nuclear oncogene products: nyc, rnyb, andfos. The effect of the interaction of the transforming proteins with the DNA may result in transcriptional activation (indicated by thick arrow) or inhibition (arrow with bar).

gene transcription. Three oncogene products have been identified as nuclear: myc, myb, and fos. C. HB-1 AND MH2 In the previous sections viral myc gene products either linked or unliked to viral structural proteins have been discussed. There are two more isolates (HB-1 and MH2) that are also myc-containing viruses, both of which, however, exhibit distinct properties. HB-1 is a derivative of MC29, which was recovered from the MC29 deletion mutant QlOH by passage through tissue culture cells (Ramsay et al., 1982a). This process resulted in recovery of cellular myc-specific sequences into the deleted viral myc gene and restored the transforming activity of HB-1. All known myc-specific properties are recovered in HB-1 such as restriction endonuclease cleavage sites, e. g., the characteristic ClaI site, oligonucleotides, and tryptic peptides including the phosphopeptides absent from the mutant (Ramsay et al., 1982a,b; Enrietto and Hayman, 1982). However, none of these methods is sensitive enough to detect minor changes in the myc gene of HB-1. Sequencing is required. The viral and cellular myc genes differ in seven amino acids, and it is not known which proportion of each of them was recombined.

218

KARIN MOELLINC

The fusion protein coded for by HB-1, p108Agug-m~c,is very similar in size to the wild-type pllO protein. The quail nonproducer cell line transformed by HB-1 exhibits a high level of expression of the p108 which is eight times higher than that of the respective proteins in mutant and wild-type quail cell lines. The same difference in expression of pllO and p108 was also observed in newly infected bone marrow cells (T. Bunte and K. Moelling, unpublished observation). The p108 was purified and tested for its proteinDNA interaction. Surprisingly, the purified protein did not show wild-type DNA-binding properties in uitro but rather resembled closely the reduced DNA-binding efficiency of the mutant protein (T. Bunte and K. Moelling, unpublished observation). It is not known whether the oncogenic potential of HB-1 does not involve interaction with DNA or whether the elevated level of expression of the protein compensates for the reduced binding ability of the protein molecules. HB-1 seems to have arisen by two recombinational events since it differs from MC29 not only in the myc but also in its LTR and gag regions (Ramsay et al., 1982a). Restriction analyses indicated that an EcoRI cleavage site is lost in the LTR and a PstI site is lost in gag. It is likely that recombination of this region took place with the helper virus, the ringneck pheasant virus (RNPV), which is known for its efficient promoter (Ramsay et al., 1982a; Bister et al., 1983). HB-1 is exceptional in its pathogenic spectrum. It does not cause myelocytomatosis or renal hepatic tumors in viuo, like MC29, rather it gives rise to rapid lymphomas (Ramsay et al., 1982a; Enrietto et al., 1984). B-cell lymphomas in birds are presumably mediated by viral promoter or enhancer sequences of the helper virus, not by oncogenes. Integration of viral regulatory elements takes place in the vicinity of the cellular myc gene. This leads to an increase in cellular myc gene transcripts and results in lymphomas after long latency periods (Hayward et al., 1981).HB-1 does not induce lymphomas by this mechanism (Enrietto et al., 1984). Tumor induction was independent of the bursa which otherwise would have been involved. It rather induces rapid lymphomas which appear within weeks instead of months and can be of B-cell as well as of T-cell origin. Whether the different pathogenicity of HB-1 reflects differences in the myc region or is due to gagspecific sequences or to the viral LTR is unknown. The oncogenic properties of HB-1 appear to arise by the superposition of two transformation mechanisms, involving an oncogene as well as the LTR. According to a recent analysis, the LTR of a murine retrovirus determines its leukemogenicity (Lenz et al., 1984). Therefore, the LTR may have more influence on the pathogenicity than the avian viral gag sequences. The genetic background of the chicken flock Brown Leghorns may also be relevant. MH2 also differs from the other known myc-containing viral strains. It

FUSION PROTEINS IN RETROVIRAL TRANSFORMATION

219

causes a different pathogenic spectrum and can be defined as a carcinoma virus affecting the liver and kidney of the bird (Grafand Beug, 1978). Recently MH2 was found to be much more oncogenic than MC29 in uiuo and it exhibits a higher efficiency of transformation in quail fibroblasts in uitro (Linial, 1982). This suggests a more distant relationship between the two viruses. MH2 codes for a gag-linked fusion protein, p100, which is not mycrelated. plOO is absent from certain transformed cells which are still very tumorigenic (Linial, 1982). These cells express the myc gene product from a subgenomic message (Pachl et al., 1983). In addition to the myc oncogene, MH2 codes for a second oncogene, designated mil or mht, which is fused to gag and which has only been identified as such very recently (Jansen et al., 1983; Kan et al., 1983). The proteins plOOAgag-mdzand p55v-myChave been purified and characterized. p55V-myC is located in the nucleus of transformed quail fibroblasts and binds to DNA in uitro (Bunte et al., 1984). pl00 is a cytoplasmic protein that binds to poly(A)-containing single-stranded RNA and DNA in uitro (Bunte et al., 1983). This property has never been observed for any other oncogene product including the cytoplasmic gag-erbA from AEV (see below). In the case of p200Agag-Apo1-mycof OK10, its ability to bind to RNA and DNA is thought to be due to presence of pol sequences. Recently, a close sequence homology of mil with the oncogene rafof a murine virus, 3611, has been detected (Jansen et al., 1984). Both oncogenes are distantly related to src (Sutrave et al., 1984). This relationship links together two totally different viral isolates from different animal species supporting the notion that there is only a limited number of oncogenes carried by retroviruses. The individual contributions of the two oncogenes of MH2 (mil and myc) to transformation will have to be sorted out by in uitro mutagenesis and subsequent in uiuo analysis. Recently the plOOgag-mizand p75sag-l"fgene products have been purified by immunoaffinity column chromatography. Both proteins exhibit protein kinase activities in autophosphorylation reactions as well as with exogenous target proteins such as histone H5 or actin. These protein kinases differ, however, from the known tyrosine-protein kinases of the src family, since they phosphorylate serine or threonine but not tyrosine. Independent of their phosphoamino acid specificities all three kinases phosphorylate lipids (see Fig. 4) (Moelling et al., manuscript submitted).

D. AVIANERYTHROBLASTOSIS VIRUS Two independent isolates of the avian erythroblastosis virus (AEV) exist: AEV-ES4 (see Grafand Beug, 1978)and the recent isolate from Japan, AEVH (Hihara et al., 1983). Both of them cause erythroblastosis and fibrosar-

220

KARIN MOELLING

1 \

FIG. 4. Model for the role of oncogene products in the regulation of protein kinase C. src, ros, fps, millraf phosphorylate phosphatidylinositol (PI) to phosphatidyl-mono- and -diphosphates (PIP, PIP2) thus activating diacylglycerol (DG) as second messenger and protein kinase C. Formation of phosphatidic acid (PA)from DG can also be affected by oncogene products (dotted lines). Activation of protein kinase C is also possible by tumor promoters (TPA). Protein kinase C regenerates cellular responses.

coma in vivo. The AEV-ES4 comprises two inserted genes that are expressed separately (gag-erbA and erbB) whereas AEV-H carries essentially only the erbB gene. Deletion of the erbA sequences by in vitro mutagenesis reduces but does not abolish transformation. A complete erbB gene is needed for erythroblast transformation, whereas only partial erbB functions are required for fibroblast transformation. A mutant of AEV which codes for truncated erbA and erbB products will transform fibroblasts but not erythroid cells (Frykberg et al., 1983). However, in the case of AEV-ES4, presence of the erbA gene renders the virus more tumorigenic and therefore it must play some role in the development of the malignant phenotype. Therefore the role of erbA in the transformation process is not clear, but it enhances the effect of erbB. The two erb genes are expressed separately; the erbA gene is a fusion protein (p75*garerM) and erbB is a glycoprotein of 68,000 to 78,000 MW (gp74erbB) that is encoded by a subgenomic message. Its unglycosylated precursor has a molecular weight of 62,000 (Privalsky et al., 1983; Hayman et al., 1983). The gag-erbA protein is predominantly present in the cytoplasm (Donner et al., 1983; Bunte et al., 1982; Hayman et al., 1983; Pachl et al.,

FUSION PROTEINS IN RETROVIRAL TRANSFORMATION

221

1983). No known functions could be attributed to the purified p75ag-erbA (Donner et al., 1983). The gag-erbA protein is included in Fig. 4 since it is cytoplasmic. It does not, however, bind to RNA in uitro. The erbB protein has been shown to be a membrane glycoprotein (Hayman et al., 1983; Hihara et al., 1983). In the case of AEV-ES4, mutants that are temperaturesensitive for transformation have been analyzed, and it has been possible to correlate the synthesis of the plasma membrane protein gp74 with transformation. At nonpermissive temperature, gp68 is not processed into gp74 and no cell surface immunofluorescence with erbB-specific antiserum is seen (Hayman, 1984).Interestingly, nucleic acid sequence analysis of molecularly cloned erbB genes from both AEV-ES4 and AEV-H have revealed that they contain a region of homology with the sequence associated with the tyrosinespecific protein kinase activity found in the sarcoma-inducing viruses, i.e., the Rous sarcoma virus src gene and the Y73 yes gene (Hayman and Beng, 1984). Sequence analysis has further confirmed indirect evidence on a relationship between erbB and a growth factor receptor which was based on receptor-associated tyrosine kinase activities (Cohen et al., 1980; Ek et al., 1982; Kasuga et al., 1983). Sequence analysis indicated that the erbB gene is a truncated epidermal growth factor (EGF) receptor which lacks the external EGF-binding domain but retains the transmembrane domain and a domain involved in stimulating cell proliferation (Downward et al., 1984; Ullrich et al., 1984). Absence of a hormone-binding site could result in a signal-independent constant firing of the receptor and uncontrolled growth. Even though no protein kinase has yet been found that is associated with the erbB oncogene product, it will most likely be a tyrosine kinase, like src or the EGF receptor. How erbB interacts with erbA and what the role of erbA may be is still unclear. The normal cellular homologous c-erbB gene is very highly expressed in erythroblastosis which is induced by avian leukemia virus through promoter insertion (Hayward et al., 1981).It is associated with an insertional mutation of the cellular erbB locus (Fung et al., 1983).Analysis of these leukemia cells has shown that the c-erbB gene products can be found in sizes ranging from 60,000 to 90,000 MW. All of them appear to be plasma membrane glycoproteins (for reference, see Hayman, 1984). It is not known whether the c-erbB gene is mutated in any way in these leukemic cells. Therefore there is no direct evidence yet as to whether the normal c-erbB gene is transforming.

E. MYELOBLASTOSIS VIRUSAND ERYTHROBLASTOSIS VIRUS There are only two avian oncogenic viruses that cause acute leukemias yet do not transform chicken fibroblasts in culture: the avian myeloblastosis

222

KARIN MOELLING

virus (AMV) which causes myeloblastosis exclusively in chickens, and E26, which induces erythroblastosis and a low level of concomitant myeloblastosis in chickens (Ivanov et al., 1964; Moscovici, 1975; Sotirov, 1981; Radke et al., 1982). Surviving animals sometimes succumb much later with liver lymphomas or nephroblastomas caused by the presence of nontransforming associated viruses (MAV 1,2) (Moscovici and Gazzolo, 1982; Moscovici and Moscovici, 1983). Both AMV and E26 viruses are replication-defective and share a sequence termed myb (also known as amu) which is unrelated to viral replicating genes and is therefore considered as the transforming oncogene of these viruses (Duesberg et al., 1980). While AMV has only the myb gene, E26 contains an additional sequence termed ets (for E twenty-six) related to normal cellular sequences (Nunn et al., 1983; Leprince et al., 1983). The acquired cellular sequences in AMV have replaced the 3' terminal portion of the reverse transcriptase and most of the gene coding for the retroviral envelope protein enu. The messenger RNA for the myb protein is a spliced subgenomic message and the protein probably contains a short gag amino terminus, consisting of 6 amino acids, in addition to carboxy terminal 11envspecific amino acids (Lipsick et al., 1984). The AMV-transforming protein is a "minifusion" protein, which contains too little gag and enu information to allow its recognition through the fused proteins. To produce myb-specific antibodies, a portion of the coding region of v-myb virus was cloned for expression in a prokaryotic host. The protein produced was a fusion protein of 37,000 MW fused to some human growth hormone sequences. Antibodies against this protein precipitated a viral myb protein of 45,000 (Klempnauer et al., 1983). The identical myb region has been successfully cloned for myb expression as a fusion protein with MS2 polymerase. Antibodies produced by this protein also precipitate a p48"-"yb protein (Moelling, et aZ., in press). Alternatively, three myb-specific peptides have been produced that specifically precipitate a p48"-"yb protein from leukemic myeloblasts (Boyle et al., 1983). The p48"-"Yb is not glycosylated or phosphorylated and does not appear to act as a protein kinase in vitro. The same sera recognized a 110,000-MW protein in embryonic hematopoietic tissue, which expresses high levels of myb-specific mRNA. This protein is absent from nonhematopoietic cells as well as AMV-transformed of the p48"-"yb leukemic myeloblastosis. It is the cellular homolog pl10c-7a~b protein. In contrast to p48'-"yb, p l 10c-tnybis strongly phosphorylated (Lipsick et al., 1984). Klempnauer detected with his antibodies a normal cellular myb gene product of 75,000 MW (Klempnauer et al., 1983). Whether this represents an alternatively spliced c-myb protein needs to be seen (Lipsick et al., 1984).

FUSION PROTEINS I N RETROVIRAL TRANSFORMATION

223

p48"-"yb is a nuclear antigen in transformed myeloblasts-in contrast to pllOc-"yb which resides in the cytoplasm of embryonic lymphocytes. It is unlikely that the nuclear location of p48"-"yb is due to the viral coding sequences gag and env since both of them are nonnuclear antigens. According to recent analysis p48v-mybis a DNA-binding protein in vitro (Moelling et al., in press). By comparison of its nucleotide sequence with that of other known DNA-binding proteins, a homology was seen between an internal region of myb and the helix-turn-helix region common to many phage and bacterial regulatory proteins including Acro, A repressor, and E . coli lac repressor products. Therefore, the myb gene product has been included into the transformation model shown in Fig. 3. Interestingly, the human promyelocytic human leukemic cell line H60, which expresses high levels of the human cellular myc gene, also expresses the cellular myb gene, both of which are suppressed when the cell differentiates (Westin et al., 1982). This suggests a role for the myc aad myb genes in maintenance of the transforrued phenotype. pllOC-"yb is induced in T lymphocytes if stimulated by a mitogen (Lipsick et al., 1984), which may indicate that c-myb expression is a general property of rapidly dividing hematopoietic cells-whether transformed or not. Furthermore, recent studies of the Abelson murine leukemia virus (A-MuLV)-induced murine B-cell neoplasms have revealed that some of the tumors contain rearranged and transcribed c-myb loci. These neoplasms do not require the continued presence of the A-MuLV for maintenance of transformation (Mushinski et al., 1983). These findings suggest a possible role for the c-myb gene product also in neoplasms not induced by AMV (Lipsick et al., 1984). E26 unlike AMV, has the additional capacity to induce erythroblastosis in viuo and to transform erythroblastosis (Radtke et al., 1982; Moscovici et al., 1983). Recently, a new oncogene (ets)in addition to the myb oncogene has been identified in E26 (Nunn et al., 1983; LePrince et al., 1983). E26 codes for one large fusion protein, p135Agag--myb--ets and no smaller proteins since no subgenomic RNA was detected. In so far E26 differs from AEV and MH2, which also code for two oncogenes, but express them independently (see above). In E26 the ets oncogene may potentiate the transforming ability of the E26 myb protein, as E26-transformed myeloblasts seem to be more tightly blocked in their differentiation than AMV-transformed myeloblasts. Furthermore, the two oncogenes may allow E26 to transform cells already committed into both the myeloid or erythroid lineages (Radtke et al., 1982), or into uncommitted precursor cells still able to differentiate into these lineages (Moscovici et al., 1983). The p135Agag-7nyb--etswas analyzed in E26-transformed quail fibroblasts as well as bone marrow cells, from which nonproducer clones were selected.

224

KARIN MOELLING

Cell fractionation analysis of transformed fibroblasts indicated that 80% of the p135Agag--myh-ets is located in the nucleus of these cells. Furthermore, the protein was purified by immunoafhity column chromatography from bone marrow cells and proved to be a DNA-binding protein in vitro. DNAbinding in vitro was thermolabile if the protein was purified from temperature-sensitive mutants (Moelling et al., in press). Whether this property is attributable to the ets- or myb-specific portion is unknown at present.

F. SARCOMAVIRUSES The avian sarcoma viruses comprise four subgroups which are classified on the basis of the relatedness of their oncogenes: the Rous sarcoma virus (RSV) subgroup; the Fujinami sarcoma virus (FSV) subgroup consisting of FSV, PRCII, and U R l ; the Yamaguchi Y73 subgroup, which includes Y73 and Esh virus; and the Rochester UR2 subgroup (for review, see Bister and Duesberg, 1982). Except for the RSV subgroup, which codes for the oncogene src, all the other subgroups code for gag-linked oncogenes that are srcrelated. The transforming proteins of the various isolates are: pp6OsrC(RSV), fp~ ppl5OAgag-fpS (URl), pp130/l4OAgag-fp" (FSV), ~ p l 0 5 ~ g ~ g - (PRCII), ppgoAgag-ws (Y73), pp8OAgag-Yes (Esh), and ~ p 6 8 ~ g ~ g - "(Urn). " The structure and function of the sarcoma-specific oncogenes have been studied most extensively with pp60"'". Since progress in understanding the role of pp6OSrc during transformation has been slow during the last year, many review articles are still valid and up to date (see, e.g., Enrietto and Wyke, 1983).Therefore, only some characteristic properties of the sarcomaspecific tyrosine kinases will be summarized here. All of the avian defective sarcoma virus gag-onc fusion proteins are phosphorylated in vivo on both serine and tyrosine, and they have associated protein kinase activities that seem to be directed toward the fusion proteins themselves as well as toward immunoglobulin and cell acceptor proteins. Studies on ts transformation mutants of FSV showed that the tyrosine phosphorylation of pp130/ 140Agag-fp" was thermolabile, thus implicating the kinase activity associated with this protein in cell transformation. ts mutants of PRCII have also been isolated (Neil et al., 1981). Results on the localization of pp6OSrc and other sarcoma-specific v-onc fusion proteins in transformed cells have varied; some reports describe localization at the nuclear envelope around centrioles or in focal adhesion plaques while the bulk of evidence supports its association with the plasma membrane (Rohrschneider, 1979, 1980; Willingham et al., 1979; Shriver and Rohrschneider, 1981). The membrane association seems to occur at the cytoplasmic face, and it has been postulated that the amino terminal portion may anchor the molecule in the membrane which contains phospholipids

FUSION PROTEINS IN RETROVIRAL TRANSFORMATION

225

(Sefton et al., 1982). In the case of pl3OAgag-fPs, lipids have not been detected, a property which may correlate with the more cytoplasmic location of this molecule (Feldman et al., 1983). The intracellular locations of the protein kinases may be important in placing the enzymes in proximity to potential target proteins, whose phosphorylation determines the tumorigenicity of the infected cell. Even though all the sarcoma-specificproteins exhibit protein kinases, it is not convincingly proven yet that the protein kinase activity is necessary and sufficient for transformation. Experiments that link defects in transformation with defects in protein kinase activity do not exclude the possibility that other functionally important properties reside in this domain of the molecule and are af€ected by the defects as well. There are a few peculiar observations supporting this possibility, e.g., RSV-transformed vole cells may revert to normal morphology without reduction of protein kinase activity (Lau et al., 1980). Among the oncogene products described here, there are several src-related ones which, however, do not exhibit detectable protein kinase activities (e.g., erbB,

fm) *

Identification of the cellular targets of tyrosine kinases proved difficult, even though tyrosine phosphorylation is an unusual activity and the level of tyrosine phosphorylation of total cell protein is 6- to 10-fold higher in transformed cells. About 30 P-Tyr-containing proteins have been identified in transformed cells compared to 2 in nontransformed cells (Martinez et al., 1982). So far only a few of these have been investigated further including proteins of 130, 51, 43, 36 and 28 kDa most of which have unphosphorylated or poorly phosphorylated precursors in normal cells (for review, see Hynes, 1980, 1982). In addition, filamin and vimentin have low amounts of phosphotyrosine in transformed cells, the significance of which is unclear. The two proteins best studied and considered likely to be important targets for pp6OSrc are the 36K and the 130K proteins. Cells transformed by FSV, PRCII, and Y73 all show similar tyrosine phosphorylation of the 36K protein (Radke and Martin, 1979; Erikson and Erikson, 1980; Radke et al., 1980). In contrast to 36K whose function and precise cellular location are unknown, the 130K protein has been identified as vinculin, which is located in part in adhesion plaques. It is found close to the plasma membrane and is thought to link actin cables to proteins in the membrane. In RSV-transformed cells the level of phosphotyrosine increases 10-fold (Geiger, 1979; Burridge and Feramisco, 1980; Sefton et al., 1981). Vinculin is also modified in Y73-transformed cells with increased tyrosine phosphorylation-but not in PRCII- and FSV-transformed cells. Two proteins are found associated with pp6OSrc in precipitates from transformed cells of 89K and 50K. The complex of 89K-pp6OSrc-50K is a transport complex in which pp60"" phosphorylation at tyrosine is controlled while the protein is moved to the cell

226

KARIN MOELLING

membrane (Courtneidge and Bishop, 1983). Preliminary evidence indicates that the oncogene products of the other sarcoma viruses also exist as complexes with 89K and 50K. The 43K protein becomes phosphorylated in tyrosine in RSV-transformed cells as well as in cells treated with the tumor promotor TPA (Gilmore and Martin, 1983). Recently glycolytic enzymes have been implicated as targets of pp6OSrc. Three glycolytic enzymes, enolase, phosphoglycerate mutase, and lactate dehydrogenase, are phosphorylated at tyrosine in cells transformed by RSV. However, only 5% of the enolase and phosphoglycerate mutase become phosphorylated upon transformation (Cooper et al. , 1983). Since these two enzymes do not play a key regulatory role for glycolysis, it is unclear at present what the effect of this phosphorylation on glycolysis may be like. Part of the difficulty in identifying the physiological target molecule(s) of the tyrosine kinases resides in the promiscuity of the enzymes in in uitro reactions. Almost any molecule in proximity to the kinases becomes phosphorylated in uitro. To prove the role of a target molecule in transformation would require its isolation and phosphorylation in uitro and subsequent analysis of its altered functions. For example, vinculin phosphorylated in uitro needs to be tested for its effect on actin polymerization, and glycolytic enzymes after in uitro phosphorylation have to be analyzed for their effect on glycolysis. We have recently succeeded in purifying the enzymatically active protein kinase pl3OAgug-fPs, which allows efficient in vitro phosphorylation. The enzyme is immobilized on immunobeads and can therefore be removed from the reaction mixture after phosphorylation. This enzyme should prove useful in studying the properties of phosphorylated vinculin, glycolytic enzymes, or 36K (P. Donner and K. Moelling, unpublished observations). The search for potential protein kinase target molecules has acquired a new impact by two reports in which not protein molecules but phospholipids are considered to interact with the protein kinases. Phosphatidylinositol (PI) can be phosphorylated in uitro by the oncogene products src and ros to phosphatidylinositol-mono- or -diphosphates (PIP, PIP,) which activate protein kinase C. This mechanism mimics what happens during activation of hormone receptors (Sugimoto et al., 1984; Macara et al., 1984). Tumorpromoting phorbol esters stimulate cell division through a similar signal transmission (Nishizuka, 1984; for review see J. L. Marx, 1984). We have recently confirmed phosphorylation of PI to PIP and PIP, by the purifiedfps kinase as well as the serine-threonine-specific millrufkinase (Moelling et al., manuscript submitted). A model on the effect of oncogene products on phospholipids is presented in Fig. 4. These data give a third example besides sis and erbB for the linkage between oncogenes and growth factor activity.

FUSION PROTEINS IN RETROVIRAL TRANSFORMATION

227

Ill. Fusion Proteins of Mammalian Viruses

A. ABELSONMURINELEUKEMIA VIRUS Abelson murine leukemia virus (A-MuLV) is an acute oncogenic retrovirus that induces lymphosarcoma in mice and transforms fibroblasts and pre-B lymphocytes in culture. The genome of A-MuLV is a hybrid of Moloney leukemia virus (MoLV) and the v-abl gene, the expression of which is controlled by the promoter/enhancer of MoLV. Moloney leukemia virus has a strong promoter, so that viral RNA produced from an integrated viral genome constitutes about l%of the total poly(A)-containing RNA of the cells in transformed fibroblasts or pre-B lymphocytes (Wang, 1983). This is about 15 times higher than expression of the cellular homologous gene c-abl. It is not known, however, whether increased expression of the viral abl sequence contributes to the oncogenicity of v-abl. The v-abl is the cDNA copy of at least 10 exons of c-abl which are correctly spliced. However, some 5' c-abl sequences are deleted in v-abl and it contains one base mutation. Thus v-abl differs from c-abl by both deletions and point mutations. v-abl codes for a fusion protein in which the first 30,000 MW fraction is derived from the MLV gag gene and the remaining 130,000 MW fraction from v-abl. The v-abl specific portion has a tyrosine-specific protein kinase activity. About two-thirds of the normal protein can be deleted and still retain protein kinase activity. Even though v-abl and v-src do not crosshybridize, they share amino acid homologies of about 50% in a domain that can be viewed as a tyrosine kinase domain and that is also shared by other tyrosine kinases such as v-fps, v-yes, and v-fes (Shibuya and Hanafusa, 1982; Kitamura et al., 1982; Hampe et al., 1982). The kinase domains are connected to other different sequences in each of the five proteins. The p160Agug-u62fusion protein is the only known example among all the fused onc gene products whose gag protein plays an essential role for oncogenicity of the protein. While the gag protein can be eliminated without any effect on the ability of A-MuLV to transform fibroblasts, the gag-truncated protein failed to induce transformation of bone marrow cells in culture. The gag- A-MuLV proteins showed lower autophosphorylation than their gag+ counterparts, although the cells transformed by gag- virus had a normal level of protein-linked phosphotyrosine indicating that gag deletion only affected phosphorylation of the molecule itself but not of other target proteins (Prywes et al., 1983).The role of gag sequences for lymphoid cell transformation can only be speculated. It has recently been shown that the A-MuLV protein like other transforming proteins, src and ras, has covalently bound lipid (Sefton et al., 1982). The exact nature of the lipid has not

228

KARIN MOELLING

been reported, but the finding of myristate at the N-terminus of the murine p15-gag protein (Henderson et al., 1983) suggests that the lipid may be attached to gag sequences. The role of lipids has been studied for pp6prC of Rous sarcoma virus. Lipid is bound to the amino terminus of pp60"" and appears to be important for membrane association and tumorigenicity (Krueger et al., 1982; Garber et al., 1983), but not for transformation of fibroblasts. A modified amino terminus of pp6WrC resulted in reduced tumorigenicity in chickens. These data suggest a role of the amino terminus of ppW" which may direct it to the plasma membrane and place it in proximity to target proteins whose phosphorylation determines the tumorigenicity of the infected cell.

VIRUSES B. MURINEOSTEOSARCOMA Two murine osteosarcoma viruses (MSVs) have been isolated. One of them, the FBJ-MSV, originated from a spontaneous osteosarcoma in a mouse. Inoculation of the virus into newborn mice induces bone neoplasia. They do not show invasive growth and do not give rise to metastases. FBJMSV also transforms fibroblasts in uitro. The other isolate (a relative of FBJMSV) is the FBR-MSV isolate, which originated from a radiation-induced osteosarcoma of the mouse. Both viruses code for the oncogene v-fos, which 9 , in the case in the case of FBJ-MSV is a protein of 55,000 MW ( ~ 5 5 ~ - fand of FBR-MSV is fused to gag, giving rise to a p75gag-foS fusion protein (for review see Miiller and Verma, 1984). The p55"-foSis located in the nucleus of transformed cells as is its normal cellular homolog p55"-foS, which differs in 5 amino acids of 332 total in the amino terminal part of the v-fos, while the 48 carboxy terminal amino acids are encoded in a different frame in v-fos. The different carboxy termini do not seem to influence the nuclear location of the proteins (Miller et al., 1984). Further studies of the proteins are required to determine their function and potential similarities to the nuclear proteins myc and myb (see Fig. 3).

C. 3611 FIBROSARCOMA VIRUS A new acutely transforming type C retrovirus was isolated from mice inoculated with a virus stock obtained by iododeoxyuridine induction of chemically (methylcholanthrene) transformed mouse cells. This virus has never been found in nature. It was designated 3611-MSV and transforms embryo fibroblasts and epithelial cells in culture and induces fibrosarcomas in uiuo. The virus encodes a 90,000-MW polyprotein (p90Agag-'"f) which is a glycoprotein, and a posttranslational cleavage product (p75Agag-ra4, both of which contain the murine p15, p12, and part of the p27 gag sequences. p7SngaCraj is slightly phosphorylated in serine and unglycosylated. Neither

FUSION PROTEINS IN RETROVIRAL TRANSFORMATION

229

of the two proteins is phosphorylated in tyrosine. The mouse cell line C3H/MCA-5 from which 3611-MSV was derived does not have elevated levels of phosphotyrosine (Rapp et al., 1983). When the ruf oncogene was isolated, no relationship to other known oncogenes was detectable. But recently the oncogene millmht from the avian virus MH2 (see above) was characterized and found to be closely related to raf (Jansen et al., 1984; Sutrave et al., 1984). Thus, the oncogene of a naturally occurring chicken virus is homologous to the oncogene of a murine virus that was produced under laboratory conditions. This again suggests that the number of oncogenes is limited. Other examples of the presence of one oncogene in viruses from different species include the murine abl and the monkey sis oncogenes in feline sarcoma viruses (see below). The p75Agag--raf has been recently purified by immunoaffinity column chromatography and was found to be a protein kinase which phosphorylates serine and threonine but not tyrosine in uitro. The protein kinase phosphorylates itself and exogenous substrates such as actin or histone H5. It is very similar to the avian plOOgag-miz protein kinase from MH2 (see above). Like tyrosine-specific protein kinases, this protein kinase also phosphorylates lipids (Moelling et al., manuscript submitted) (see Fig. 4). SARCOMAVIRUSES D. FELINE

The feline leukemia virus (FeLV) is horizontally transmitted in domestic cats and is a natural etiological agent of leukemia (Hardy, 1980). Feline leukemia virus recombines at a relatively high frequency with normal cellular cat genes resulting in the formation of five different feline sarcoma virus strains. The most extensively studied of the FeSV isolates include the Snyder-Theilen (ST) and Gardner-Arnstein (GA) strains, both of which contain the viral oncogene v-fes, and the Susan McDonough (SM) strain, which contains the viral oncogene v-fm (Snyder and Theilen, 1969; Gardner et al., 1970; McDonough et al., 1971). Recently a new oncogene (v-fgr) was identified in the Garner-Rasheed FeSV strain (Rasheed et al., 1982; Naharro et al., 1983). Two FeSV strains were identified recently, the ParodiIrgens (PI) and the Hardy-Zuckerman-2 (HZ-2) FeSVs (Irgens et al., 1973), which carry other oncogenes, such as v-sis and v-abl as fusion proteins, p75Agag--sis and p95Agacabz,respectively (Besmer et al., 1983a,b). These oncogenes are known from retroviruses of monkeys and mice. Nucleic acid hybridization-which generally failed to detect sequence homology between different oncogenes-detected sequence homology between v-fes and v-fps, the oncogene of the Fujinami and PRCII strains of avian sarcoma virus (Shibuya et al., 1980). Also the products encoded by v-fes and v-fps were found to be biochemically, antigenically, and func-

230

KARIN MOELLING

tionally related (Beemon, 1981; Barbacid et al., 1981). v-fes and v-fps showed 70% overall nucleotide homology suggesting that they were derived from cognate c-onc loci of mammals and birds (Sherr et al., 1984). The v-feslv-fps gene products each exhibit an associated tyrosine-specific protein kinase activity (van de Ven et al., 1980a; Barbacid et al., 1980a,b; Reynolds et al., 1980; Feldman et al., 1980). In this respect v-feslv-fps genes encode enzymes functionally very similar to those of other viral oncogenes including v-src (Collett and Erikson, 1978; Levinson et al., 1978; Sefton et al., 1980), v-abl (Witte et al., 1980), v-yes (Kawai et al., 1980), v-ros (Feldman et al., 1982), and v-fgr (Rasheed et al., 1982). The proteins coded for by each of these genes induce phosphorylation of heterologous substrates in transformed cells, leading to a marked increase in the total level of phosphotyrosine. In addition, each of the viral transforming proteins is itself phosphorylated in tyrosine, generally at a preferred site (Patschinsky et al., 1982). It is interesting that other viral oncogenes including v-mos (van Beveren et al., 1981) and erbB (see above) 'are not thought to encode kinases even though they exhibit homology to members of the tyrosine kinase gene family. These data suggest that portions of all of these genes have descended from a single ancestral sequence and have now evolved to be functionally distinct. The possibility that v-fm, v - m s , and v-erbB specify tyrosine kinases, which have not yet been detected, cannot be rules out. Alternatively, other transforming functions could reside in these molecules which do not involve tyrosine phosphorylation. The related coding regions of the various kinase v-onc genes include the putative catalytic sites of the enzymes as well as the sites for nucleotide triphosphate binding (Barker and Dayhoff, 1982). The divergence of other regions of these genes could possibly 'af€ect their differential expression in cells, their subcellular localization, or the catalytic properties of the enzymes. All of the tyrosine kinases encoding genes may be members of a more extended gene family encoding tyrosine kinases which are associated with cell surface receptors for the extracellular growth factors EGF, PDGF, and insulin (Ushiro and Cohen, 1980; Ek et al., 1982; Kasuga et al., 1982). The association of tyrosine-specific protein kinase activities with these receptors suggests that the mechanism by which some v-onc gene products transform cells could involve subversion of hormonally regulated pathways controlling cell proliferation. In contrast to v-fes, the v-fm gene of SM-FeSV is a glycosylated polyprotein, g~l80g~g-f'". It is posttranscriptionally cleaved near the gag-fms junction to yield two distinct polypeptide products. These include a protein of about 55,000 kDa which contains only gag-coded antigenic determinants and a glycoprotein of 120,000 kDa (gp120fm) (Barbacid et al., 1980b; Ruscetti et

FUSION PROTEINS IN RETROVIRAL TRANSFORMATION

231

al., 1980; Sherr et al., 1980a). All glycosylation occurs within the v-fmscoded portion of g ~ l 8 0 g ~ g - + The ~ . mechanism and site(s) of proteolytic cleavage are not known. Part of the gp120fms appears to be subsequently modified, possibly by further glycosylation to generate gp140fm, which together with the gpl2ofms represents the major forms of transforming glycoprotein in SM-FeSV transformants. The v # m glycoprotein was localized in the juxtanuclear region of the cell cytoplasm (Anderson et al., 1982) where it codistributes with the intermediate filament proteins keratin and vimentin. Also cell fractionation analyses showed that a portion of the v-fm glycoprotein has an intermediate filament-associating domain (Sherr et al., 1984). The v-fm glycoprotein can serve as substrate for tyrosine phosphorylation in immune complex kinase assays, but it is poorly phosphorylated in the cell and lacks an intrinsic kinase activity (Sherr et al., 1980b). Surprisingly, the nucleotide sequencing of the v-fm gene predicted a product with marked homology to the tyrosine-specific protein kinases (Hampe et al., 1983). The regions of homology include the site of tyrosine phosphorylation but do not include an upstream lysine residue that has been implicated in ATP binding (Barker and Dayhoff, 1982). Like pp60Src, the kinase domain of the v-fm glycoprotein appears to be oriented at the cytoplasmic surface of the membrane (Willingham et al., 1979). Recently, the v-onc sequences of the two FeSV strains PI-FeSV and HZ-2 FeSV have been found to be homologous with the v-sis sequences of the simian sarcoma virus and the Abelson murine leukemia virus (Besmer et al., 1983a,b). While the A-MuLV induces tumors of the hematopoietic system, predominantly of the B-lymphocyte lineage, and transforms fibroblasts in culture (see above), HZ-2 is able to transform nonhematopoietic cells in uiuo. The HZ-2 FeSV gene product p98Agag-ab2-enois a protein kinase which phosphorylates itself as well as the heavy chain of immunoglobulin (Besmer et al., 1983b). Further details on this protein are not yet known. The PI-FeSV codes for a p76Agag--sis fusion protein in contrast to SSV, where the sis oncogene is expressed from a subgenomic message and fused to short N-terminal enu leader sequences, p28Aenu-SiS(Besmer et al., 1983a; Robbins et al., 1982). Thus the sis protein appears to be expressed once fused to gag and in the other case fused to env sequences. As far as the feline and the simian sarcoma virus gene products have been characterized, no differences have been observed in uioo or in uitro (Besmer et aZ., 1983a).

E. SIMIANSARCOMA VIRUS The oncogene of simian sarcoma virus (SSV), v-sis, codes for a protein of 28,000 MW (Robbins et al., 1982). The coding sequences for the onc gene

232

KARIN M O E L L I N G

product initiate from the amino terminus of the env gene (Devare et al., 1983). The protein should therefore be designated p28Aen"-sis. The v-sis gene product is not a phosphoprotein, and it does not possess any detectable protein kinase activity (Robbins et al., 1982). These properties distinguish it from the majority of other retroviral onc proteins. The oncogene of SSV recently turned out to be one of the most interesting ones, since it gave a very important clue to the potential physiological role of oncogenes in general (for summary, see Weiss, 1983). The sequence of v-sis is almost identical with the sequence coding for one peptide of the plateletderived growth factor, PDGF (Waterfeld et al., 1983; Doolittle et al., 1983). Ninety contiguous amino acids of PDGF are virtually identical to a large domain in p28Ae"u-sis. Purified PDGF consists of several cleaved peptides linked by disulfide bonds, therefore the identification of PDGF peptides and precursors is not quite clear yet. Platelet-derived growth factor is released from a granules of platelets during blood clotting, and it is the major polypeptide growth factor found in serum. Fibroblastic and neuroglial cells are specifically sensitive to the mitogenic action of PDGF, as they express PDGF receptors. Like the E GF receptor, the PDGF receptor exhibits tyrosine-specific kinase activity on binding the growth factor (Ek et al., 1982). Tyrosine phosphorylation is the characteristic property of the src-related oncogenes. If the oncogenes src and abl encoded a growth factor themselves and both factor and receptor were produced by the identical cell, then autocrine stimulation of proliferation would ensue (Sporn and Todaro, 1980). None of the oncogene products have convincingly been shown to be secretory proteins as yet. As p28Aen-sis contains enu sequences, it is therefore possible that these env sequences allow membrane insertion and lead to secretion of the transforming protein. Thus, cells having PDGF receptors and the capability of responding to the mitogenic effect of PDGF become constitutively activated. Alternatively, cells infected by SSV may express the mitogenic signal only intracellularly and by-pass the need for receptor interactions at the cell surface. It remains to be seen whether p28Aenv-SiS is a growth factor, whether it is secreted, and if it binds to PDGF receptor to induce a mitogenic response. Many cells transformed by oncogenic viruses release mitogenic sarcoma growth factors, resulting in autocrine growth stimulation, which, however, appears to be indirectly induced by the oncogenes in the transformed cells (De Larco and Todaro, 1978; Kaplan et al., 1982). Two other transforming genes besides v-sis have been shown to share at least a partial homology with mitogenic peptides: polyoma virus middle T antigen and gastrin, which can act as a mitogen (Baldwin et al., 1982), and furthermore, the B-lym gene of chicken bursa1 lymphomas. This gene was identified by DNA transfection of NIH/3T3 cells. It encodes a small poly-

FUSION PROTEINS IN RETROVIRAL TRANSFORMATION

233

peptide that shows partial homology to the amino terminus of transferrin (Goubin et al., 1983).The apparent identity of sis with PDGF is nonetheless the first example of an oncogene representing a previously defined growth factor. These results are summarized in a transformation model presented in Fig. 5. Various oncogenes are indicated at various sites in the model. IV. Discussion on the Effect of Viral Structural Proteins on Transformation

Three oncornaviral structural proteins-gag, pol, and enu-exist, and deleted portions of each of them are found fused to transformation-specific portions to give polyproteins. Most frequently the gag region is hsed to oncogenes. There is only one case in the avian virus system in which the identical oncogene is expressed unlinked to structural proteins as well as

3 nucleus

cytoplasm

FIG.5 . Hypothetical model for the transformation mechanisms of the src-related transforming proteins. Some of them are protein kinases (src,fps, abl, etc.), whereas others are not (erbB,fms, sis). Two of the proteins are glycosylated (fm.erbB). B-lym is not a viral oncogene (see Section I). Autocrine stimulation of a tumor cell by oncogene products has not yet been proven. The effect of the protein kinases on potential target proteins is indicated by arrows starting from src. PI indicates phosphatidylinositol.

234

KARIN MOELLING

linked, the v-myc and the Agag-Apol-myc of OK10. In this particular case the effect of the fused proteins affects the subcellular location of myc. Probably due to pol sequences the polyprotein is more cytoplasmic. Since at present it is unknown which of the two proteins plays a biological role or whether they function in concert, nothing can be concluded about the potential effect of fused proteins on the transforming proteins. In other cases such as Agag-myc in MC29 and v-myc in MH2, an effect of Agag on subcellular location, in vitro binding, and other known properties has not been detectable. In the case of src and Agag-fps there may be an effect attributable to Agag in that it may prevent attachment of lipids to the amino terminus of fps, which is observed with src, and thus result in the more cytoplasmic location of Agag-fps. The most clear-cut observation of an effect of Agag on an oncogene product was made with Agag-abl. This is the only fusion protein known where the gag portion is required for transformation, possibly due to its myristilation. The role of Aenv in the case of sis can only be speculated; it may allow attachment to membranes or even secretion of the polyprotein. Since none of the known viral oncogenes is identical to the normal cellular homologous genes but always differs by at least point mutations, deletions, and frame shifts, it cannot be excluded that fusion to structural proteins contributes further differences. The effect of Agag on the abl transforming protein is particularly surprising, since the actual protein kinase activity is located on a small portion of the molecule, whereas a large part appears dispensible. Therefore, the fused structural proteins in some cases affect the location of the molecules more than their actual functions. Since the oncogenic spectra of the various oncogene products of some viruses are rather broad, it may also be possible that the structural proteins have more influence in certain target cells or sites of inoculations than others. Fusion to gag sequences always involves the direct linkage of the oncogene to the viral LTR. This region is assumed to play a role in activation of cellular oncogenes. In the case of HB-1 an elevated level of myc gene expression can be observed by comparison with MC29 and is thought to be due to an exchange of the LTR. The consequence of the elevated rnyc expression on transformation is not quite clear, but may exist, since HB-1 has a different pathogenic effect in the animal. The LTR may have other effects; the most important hypothesis about its role at this point concerns its tissue specificity which is very important in the various oncogenic effects of retroviruses.

REFERENCES Abrams, H. D., Rohrschneider, L. R., and Eisenman, R. N. (1982). Cell 29, 427-439. Alitalo, K . , Bishop, J. M . , Smith, D. H., Chen, E. Y . , Colby, W. W., and Levinson, A. D. (1983a). Proc. Not!. Acad. Sci. U.S.A. 80, 100-104.

FUSION PROTEINS IN RETROVIRAL TRANSFORMATION

235

Alitalo, K., Ramsay, G., Bishop, J. M., Ohlsson-PfeiH‘er, S., Colby, W. W., and Levinson, A. D. (1983b). Nature (London) 306, 274-277. Anderson, S. J., Furth, M., Wolff, L., Ruscetti, S. K., and Sherr, C. J. (1982).J . Virol. 44,696702. Anderson, S. J.. Gonda, M., and Sherr, C. J . (1983). Submitted. Baldwin, G . S., Burgess, A. W., and Kemp, B. E. (1982). Biochem. Biophys. Res. Cornrnun. 109, 656-663. Barbacid, M., Beemon, K., and Devare, S. G. (1980a).Proc. Natl. Acad. Sci. U.S.A. 77,51585162. Barbacid, M., Lauver, A. V., and Devare, S. G. (1980b).J. Virol. 33, 196-207. Barbacid, M., Breitman, M. L., Lauver, A. V., Long, L. K., and Vogt, P. K. (1981). Virology 110, 411-419. Barker, W. C., and Dayhoff, M. 0. (1982). Proc. Natl. Acad. Sci. U.S.A. 79, 2836-2839. Beard, J. W. (1980). In “Viral Oncology” (G. Klein, ed.), pp. 55-87. Raven, New York. Beemon, K. (1981). Cell 24, 145-153. Benter, T., Beimling, P., Pf&, E., Sander, T., Heimann, B., and Moelling, K. (1984). In “Modern Trends in Human Leukemia VI” (R. Neth, R. C. Gallo, M .F. Greaves, M. A. S. Moore, K. Winkler, eds.). Springer-Verlag Berlin and New York, in press. Besmer, P., Snyder, H. W., Murphy, J. R., Hardy, W. D., and Parodi, A. (1983a).J . Virol. 46, 606-613. Besmer, P., Hardy, W. D., Zuckerman, E., Lederman, L., and Snyder, H. W. (1983b).Nature (London) 303, 825-827. Beug, H., and Hayman, M. J. (1984). Cell36, 963-972. Beug, H., Hayman, M. J., and Graf, T. (1982). E M B O J . 9, 1069-1073. Bishop, J. M. (1983). Annu. Reu. Biochem. 52, 301-354. Bister, K. (1983). In “Mechanisms of Viral Leukaemogenesis, Annual Reviews Leukaemia Research Fund” 0. V. Dacie, J. M. Goldman, and J. 0. Jarrett, eds.), Vol. 1. Churchill Livingstone, Edinburgh. Bister, K., and Duesberg, K. (1982). In “Advances in Viral Oncology” (G. Klein, ed.), Vol. 1, pp. 3-41. Raven, New York. Bister, K., Hayman, M. J., and Vogt, P. K. (1977). Virology 82, 431-448. Bister, K., Ramsay, G . M., and Hayman, M. J. (1982).J . Virol. 41, 754-766. Bister, K., Jansen, H. W., Graf, T., Enrietto, P. J., and Hayman, M. J. (1983).J . Virol. 46, 337-346. Boyle, W. J . , Lipsick, J. S., Reddy, E. P., and Baluda M. A. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 2834-2838. Bunte, T., Greiser-Wilke, I., Donner, P., and Moelling, K. (1982). E M B O J. 1, 919-927. Bunte, T., Greiser-Wilke, I., and Moelling, K. (1983). E M B O J . 2, 1087-1092. Bunte, T., Donner, P., Pf&, E., Reis, B., Greiser-Wilke, I., Schaller, H., and Moelling, K. (1984). E M B O J. 3, 1919-1924. Burridge, K., and Feramisco, J. R. (1980). Cell 19, 587-595. Ciejek, E. M., Tsai, M.-J., and O’Malley, B. W. (1983). Nature (London) 306, 607-609. Cohen, S . , Carpenter, G . , and King, L. E. (1980).J . B i d . Chem. 255, 4834-4842. Collet, M. S . , and Erikson, R. L. (1978). Proc. Natl. Acad. Sci. U.S.A. 75, 2020-2024. Cooper, J. A., Nachum, A. K., Schwartz, R. J., and Hunter, T. (1983). Nature (London) 302, 218-223. Courtneidge, S. A , , and Bishop, J. M. (1983). Proc. Nutl. Acad. Sci. U.S.A. 79, 7117-7121. Davison, B. L., Egly, J.-M., Mulvihill, E. R., and Chambon, P. (1983). Nature (London) 304, 680-683. DeFeo-Jones, D., Scolnick, E . M., Koller, R., and Dhav, R. (1983).Nature (London)306,707709.

236

KARIN MOELLING

Cooper, J. A., Nachum, A. K., Schwartz, R. J., and Hunter, T. (1983). Nature (London) 302, 218-223. Courtneidge, S. A., and Bishop, J. M. (1983). Proc. Natl. Acad. Sci. U.S.A. 79, 7117-7121. Curran, T., Miller, A. D., Zokas, L., and Verma, J. M. (1984). Cell (in press). Davison, B. L., Egly, J.-M., Mulvihill, E. R., and Chambon, P. (1983). Nature (London) 304, 680-683. DeFeo-Jones, D., Scolnick, E. M., Koller, R., and Dhav, R. (1983).Nature (London)306, 707709. De Larco, J. E., and Todaro, G . J. (1978). Proc. Natl. Acad. Sci. U.S.A. 75, 4001-4007. Devare, S. G . , Reddy, E. P., Law, J. D., Robbins, K. C., and Aaronson, S. A. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 731-735. Donner, P., Greiser-Wilke, I., and Moelling, K. (1982). Nature (London) 296, 262-266. Donner, P., Bunte, T., Greiser-Wilke, I., and Moelling, K. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 2861-2865. Doolittle, R. F. (1981). Science 214, 149-159. Doolittle, R. F., Hunkapiller, M. W., Hood, L. E., Devare, S. G., Rohbins, K. C., Aaronson, S . A., and Antoniades, H. N. (1983). Science 221, 275-277. Downward, J., Yarden, Y., Mayes, E., Scarce, G . , Totty, N., Stockwell, P., Ullrich, A., Schlessinger, J., and Waterfield, M. D. (1984). Nature (London) 307, 521-527. Duesberg, P. H. (1983). Nature (London) 304, 219-226. Duesberg, P. H., Bister, K., and Moscovici, C. (1980). Proc. Natl. Acad. Sci. U.S.A. 77,51205124. Ek, B., Westermark, B., Wasteson, A., and Heldin, C.-H. (1982). Nature (London) 295, 419420. Enrietto, P. J., and Hayman, M. J. (1982).J . Virol. 44, 711-715. Enrietto, P. J., and Wyke, J. A. (1983). Ado. Cancer Res. 39, 269-309. Enrietto, P. J., Hayman, M. J., Ramsay, G. M., Wyke, J. A., and Payne, L. N. (1982). Virology 124, 164-172. Enrietto, P., Payne, L. N., and Hayman, M. J. (1984). Cell (in press). Erikson, E., and Erikson, R. L. (1980). Cell 21, 829-836. Feldman, L. T., and Nevins, J. R. (1983). Cell Biol. 3, 829-838. Feldman, R. A., Hanafusa, T., and Hanafusa, H. (1980). Cell 22, 757-765. Feldman, R. A., Wang, L. H., Hanafusa, H., and Balduzzi, P. C. (1982).J . Virol. 42,228-236. Feldman, R. A., Wang, E., and Hanafusa, H. (1983).J . Virol. 45, 782-791. Frykberg, L., Palmieri, S., Beug, H., Graf, T., Hayman, M. J., and Vennstrom, B. (1983). Cell 32, 227-238. Fung, Y. K. T., Lewis, W. G., Kung, H.-J., and Crittenden, L. B. (1983). Cell 33, 357-368. Gallwitz, D., Donath, C., and Sander, C. (1983). Nature (London) 306, 704-707. Garber, E. A., Krieger, J. G., Hanafusa, H., and Goldberg, A. R. (1983).Nature (London) 302, 161-163. Gardner, M. B., Rongey, R. W., Arnstein, P., Estes, J. D., Sarma, P., Huebner, R. J . , and Rickard, C. J. (1970). Nature (London) 226, 807-809. Geiger, B. (1979). Cell 18, 193-205. Giallongo, A., Appella, E., Riccardi, R., Rovera, G., and Croce, C. M. (1983). Science 222, 430-432. Gilmore, T., and Martin, G . S. (1983). Nature (London) 306, 487-490. Goubin, G . , Goldman, D. S . , Luce, J., Neiman, P. E., and Cooper, G. M. (1983). Nature (London) 302, 114-119. Graf, T., and Beug, H. (1978). Biochim. Biophys. Acta 516, 269-299. Graf, T., and Stehelin, D. (1982). Biochim. Biophys. Acta 651, 245-271.

FUSION PROTEINS IN RETROVIRAL TRANSFORMATION

237

Greiser-Wilke, I., Owada, M. K., and Moelling, K. (1981). J . Virol. 39, 325-329. Hampe, A., Laprevotte, I., Galibert, F., and Sherr, C. J. (1982). Cell 30, 775-785. Hampe, A. M., Gobet, M., Even, J., Sherr, C. J . , and Galibert, F. (1983).J . Virol. 45, 466471. Hann, S. R., Ahrams, H., Rohrschneider, L. R., and Eisenman, R. N. (1983).Ceu34,789-798. Hardy, W. D., Jr. (1980). I n “Feline Leukemia Virus” (W. J. Hardy, Jr., M. Essex, and A. J. McClelland, eds.), pp. 79-120. Elsevier, Amsterdam. Hayman, M. J., and Beng, H. (1984). Nature (London) 309, 460-462. Hayman, M. J . , Ramsay, G., Savin, K., Kitchener, G., Graf, T., and Beug, H. (1983).Cell 32, 579-588. Hayward, W. S . , Neel, B. C., and Astrin, S. M. (1981). Nature (London) 290, 475-480. Henderson, L. E., Krutzsch, H. C., and Oroszlan, S. (1983).Proc. Natl. Acad. Sci. U.S.A.80, 339-343. Hendricks, M., and Weintraub, H. (1981). Proc. Natl. Acad. Sci. U.S.A. 78, 5633-5637. Hihara, H., Yamamoto, H., Shimohira, H., Arai, K., and Shimizu, T. (1983).J . Natl. Cancer Inst. 70, 891-898. Hynes, R. 0. (1980). Cell 21, 601-602. Hynes, R. (1982). Cell 28, 437-438. Irgens, K., Wyers, M., Moraillon, A., Parodi, A., and Fortuny, V. (1973).C.R. Acad. Sci. Paris 26, 1783-1786. Ivanov, K., Mladenov, Z., Nedyalkov, S., Todorov, T. G., and Yakimov, M. (1964).Bull. Inst. Pathol. Comp. Anim. Acad. Bulg. S c i . 10, 5-38. Jansen, H. W., Ruckert, B., L u n , R., and Bister, K. (1983). E M B O J . 2, 1969-1975. Jansen, H. W., Lurz, R., Bister, K., Bonner, T. I., Mark, G. E.,and Rapp, U. (1984).Nature (London) 307, 281-284. Kan, N. C., Floridellis, C. S., Garon, C. F., Duesberg, P. H., and Papas, T. S. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 6566-6570. Kaplan, P. L., Anderson, M., and Ozanne, B. (1982). Proc. Natl. Acad. S c i . U.S.A. 79, 485489. Kasuga, M., Zick, Y., Blithe, D. L., Crettaz, M., and Kahn, R. (1982). Nature (London) 298, 667-669. Kasuga, M., Fujita-Yamaguchi, Y., Blithe, D. L., and Kahn, C. R. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 2137-2141. Kawai, S., Yoshida, M., Segawa, K., Sugiyama, H., Ishizaki, R., and Toyoshima, K. (1980). Proc. Natl. Acad. S c i . U.S.A. 77, 6199-6203. Kitamura, N . , Kitamura, A., Toyoshima, K., Hirayama, Y., and Yoshida, M. (1982). Nature (London) 297, 205-208. Klempnauer, K.-H., Ramsay, G., Bishop, J. M., Moscovici, M. G., Moscovici, C., McCrath, J. P., and Levinson, A. D. (1983). Cell 33, 345-355. Krueger, J. G., Garber, E. A., Goldberg, A. R., and Hanafusa, H. (1982). Cell 28, 889-896. Land, H., Parada, L. F., and Weinberg, R. A. (1983). Nature (London) 304, 596-602. Lau, A. F., Krzyzek, R. A., Brugge, J. S . , Collett, M. S . , Erikson, R. L., and Faras, A. J. (1980). Cold Spring Harbor Symp. Quant. Biol. 44, 1057-1064. Lautenherger, J. A., Court, D., and Papas, T. S. (1983). Gene 23, 75-84. Lenz, J . , Celander, D., Cowther, R. L., Patrarca, R., Perkins, D. W., and Haseltine, W. A. (1984). Nature (London) 308, 467-470. LePrince, D., Gegonne, A., Coll, J., de Taisne, C., Schneeberger, A., Lagrou, C., and Stehelin, D. (1983). Nature (London) 306, 395-397. Levinson, A. D., Oppermann, H., Levintow, L., Varmus, H. E., and Bishop, J. M. (1978).Cell 15, 561-572.

238

KARIN MOELLING

Linial, M. (1982). Virology 119, 382-391. Lipsick, J. S., Boyle, W. J., Lampert, M. A,, and Baluda, M. A. (1984).CancerCells2,143-151. Macara, I. G., Marinetti, G. V., and Balduzzi, P. C. (1984). Proc. Natl. Acad. Sci. U.S.A. 81, 2728-2732. McDonough, S. K., Larsen, S., Brodey, R. S., Stock, W. D., and Hardy, W. D., Jr. (1971). Cancer Res. 31, 953-956. Martinez, R., Nakamura, K. D., and Weber, M. J. (1982). Mol. Cell B i d . 2, 653-665. Maw, J. L. (1984). Science 224, 271-274. Miller, A. D., Curran, T., and Verma, J. M. (1984). Cell 36, 51-60. Mladenov, Z., Heine, U., Beard, D., and Beard, J. W. (1967).J . Natl. Cancer Inst. 38, 251285. Moelling, K., Benter, T., Bunte, T., Pfalf, E., Deppert, W., Egly, J. M., and Miyamoto, N. B. (1984a).Curr. Top. Microbiol. Immunol. 113, 198-207. Moelling, K., Bunte, T., Greiser-Wilke, I., Donner, P., and Pfalf, E. (1984b). Cancer Cells 2, 173-181. Moelling, K., Heimann, B., Beimling, P., Rapp, U. R., and Sander, T., manuscript submitted. Moscovici, C. (1975). Curr. Top. Microhiol. Immunol. 71, 79-101. Moscovici, C., and Gazzolo, L. (1982).I n “Advances in Viral Oncology” (G. Klein, ed.), pp. 83106. Raven, New York. Moscovici, M. G., and Moscovici, C. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 1421-1425. Moscovici, M. G . , Jurdic, P., Samarut, J., Gazzola, L., Mura, C. V., and Moscovici, C. (1983). Virology 129, 65-78. Miiller, R., and Verma, J. M. (1984). Curr. Top. Microhiol. Immunol. (in pres). Mushinski, J. F., Poker, M., Bauer, S. R., and Reddy, E. P. (1983). Science 220, 795-797. Naharro, G., Tronick, S. R., Rasheed, S., Gardner, M. B., Aaronson, S. A., and Robbins, K. C. (1983).J. Virol. 47, 611-619. Neil, J. C., Ghysdael, J., Vogt, P. K., and Smart, J. E. (1981).Nature (London) 291, 675-677. Newmark, P. (1983). Nature (London) 305, 407-471. Nishizuka, Y. (1984). Nature (London) 308, 693-698. Nunn, M. G., Seeburg, P. H., Moscovici, C., and Duesberg, P. H. (1983). Nature (London) 306, 391-395. Pachl, C., Biegalke, B., and Linial, M. (1983).J. Virol. 45, 133-139. Patschinsky, T., Hunter, T., Esch, F. S., Cooper, J . A., and Sefton, B. M. (1982).Proc. Natl. Acad. Sci. U.S.A. 79, 973-977. Privalsky, M. L., Sealy, L., Bishop, J. M., McGrath, J. P., and Levinson, A. 1). (1983).Cell32, 1257-1267. Prywes, R., Foulkes, J. G., Rosenberg, N., and Baltimore, D. (1983). Cell 34, 569-579. Radke, K., and Martin, G. S. (1979). Proc. Natl. Acad. Sci. U.S.A. 76, 5212-5216. Radke, K., Gilmore, T., and Martin, G. S. (1980). Cell 21, 821-828. Radke, K., Beug, H., Kornfeld, S., and Graf, T. (1982). Cell 31, 643-653. Ralston, R . , and Bishop, J. M. (1983). Nature (London) 306, 803-806 Ramsay, G. M., and Hayman, M. J. (1982).J . Virol. 41, 745-753. Ramsay, G . M., Graf, T., and Hayman, M. J. (1980). Nature (London) 288, 170-172. Ramsay, G . M . , Enrietto, P. J., Graf, T., and Hayman, M. J. (1982a). Proc. Natl. Acad. Sci. U.S.A. 79, 6885-6889. Ramsay, G . M., Hayman, M. J., and Bister, K. (1982b). EMBOJ. 9, 1111-1116. Rapp, U. R . , Reynolds, F. H., and Stephenson, J. R. (1983).J . Virol. 45, 914-924. Rasheed, S., Barbacid, M . , Aaronson, S. A., and Gardner, M. B. (1982). Virology 117, 238244. Reddy, E. P., Reynolds, R. K., Watson, D. K., Schultz, R. A., Lautenberger, J., and Papas, T. S. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 2500-2504.

FUSION PROTEINS IN RETROVIRAL TRANSFORMATION

239

Reynolds, F. H., Jr., Van de Ven, W. J., and Stephenson, J. R. (1980).J . B i d . Chem. 255, 11040-11047. Robbins, K. C., Devare, S. G., Reddy, E. P., and Aaronson, S. A. (1982). Science 218, 11311133. Rohrschneider, L. R. (1979). Cell 16, 11-24. Rohrschneider, L. R. (1980). Proc. Natl. Acad. Sci. U.S.A. 77, 3514-3518. Ruley, H. E. (1983). Nature (London) 304, 602-606. Ruscetti, S. K., Turek, L. P., and Sherr, C. J. (1980).1. Virol. 35, 259-264. Schrier, P. J., Bernards, R., Vaessen, R. T. M . J., Houweling, A., and van der Eb, A. J. (1983). Nature (London) 305, 771-775 Schutzbank, T., Robinson, R., Ore, M., and Levine, A. J. (1982). Cell 30, 481-490. Sefton, B. M., Hunter, T., Beemon, K., and Eckhart, W. (1980). Cell 20, 806-816. Sefton, B. M., Hunter, T., Ball, E. H., and Singer, S. J. (1981). Cell 24, 165-174. Sefton, B. M . , Trowbridge, I. S., Cooper, J. A,, and Scholnick, E. M. (1982). Cell 31,465-474. Sherr, C. J., Donner, L., Fedele, L. A., Turek, L., Even, J., and Ruscetti, S. K. (1980a). In “Feline Leukemia Virus” (W. D. Hardy, Jr., M. Essex, and A. J. McClelland, eds.), pp. 293-307. Elsevier, Amsterdam. Sherr, C. J.. Fedele, L. A., Oskarsson, M., Maizel, J., and Vande Woude, G. (1980b).J . Virol. 34,200-212. Sherr, C. J . . Anderson, S. J., Rettenmeier, C. W., and Roussel, M. F. (1984). Cancer Cells 2, 329-338. Shibuya, M., and Hanafusa, H. (1982). Cell 30, 787-795. Shibuya, M., Hanafusa, T., Hanafusa, H., and Stephenson, J. R. (1980). Proc. Natl. Acad. Sci. U.S.A. 77, 6536-6540. Shilo, B. Z., and Weinberg, R. A. (1981). Proc. Natl. Acad. Sci. U.S.A. 78, 6789-6792. Shriver, K . , and Rohrschneider, L. R. (1981).J . Cell B i d . 89, 525-535. Snyder, S. P., and Theilen, G . H. (1969). Nature (London) 221, 1074-1075. Sotirov, N. (1981).1.Natl. Cancer Znst. 66, 1143-1147. Sporn, M. B., and Todaro; G. J. (1980). New Engl. J . Med. 303, 878-880. Sugimoto, Y., Whitman, M., Cantley, L. C., and Erikson, R. L. (1984). Proc. Natl. Acad. Sci. U.S.A. 81, 2117-2121. Sutrave, P., Bonner, T. I., Rapp, U. R., Jansen, H. W., Patschinsky, T., and Bister, K. (1984). Nature (London) 309, 85-88. Ullrich, A,, Coussens, L., Hayflick, J. S., Dull, T. J., Gray, A., Tam, A. W., Lee, J., Yarden, Y . , Libermann, T. A , , Schlessinger, J.. Downward, J., Mayes, E. L. V., Whittle, N., Waterfield, M. D., and Seeburg, P. H. (1984). Nature (London) 309, 418-425. Ushiro, H., and Cohen, S. J. (1980).J . Biol. Chem. 255, 8363-8365. Van Beveren, C . , Galleshaw, J. A., Jones, V., Berns, A. J. M., Doolittle, R. F., Donoghue, D. J., and Verma, I. M. (1981). Nature (London) 289, 258-262. Van de Ven, W. J. M., Reynolds, F. H., Nalewaik, R. P., and Stephenson, J. R. (1980).1.Virol. 35, 165-175. Watt, R . , Stanton, L. W., Marcu, K. B., Gallo, R. C., Croce, C. M., and Rovera, G . (1983). Nature (London) 303, 725-728. Wang, J. Y. J. (1983). Nature (London) 304, 400. Waterfield, M. D., Scarce, G. T., Whittle, N., Stroobant, P., Johnsson, A,, Wasteson, A,, Westermark, B., Heldin, C. H . , Huang, J. S., and Deuel, T. F. (1983). Nature (London) 304, 35-39. Weiss, R. (1983). Nature (London) 304, 12. Westin, E. H . , Gallo, R. C., Arya, S . K., Souza, L. M . , Baluda, M. A,, Aaronson, S. A., and Wong-Staal, F. (1982). Proc. Natl. Acad. Sci. U.S.A. 79, 2194-2198. Willingham, M. C., Jay, G., and Pastan, I. (1979). Cell 18, 125-134.

This Page Intentionally Left Blank

APPLICATION OF MIGRATION INHIBITION TECHNIQUES IN TUMOR IMMUNOLOGY Robert Szigeti Department of Tumor Biology, Karolinska Institutet, Stockholm, Sweden

I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241 11. The Basic Phenomenon of Migration Inhibition. . . . . . . 242 111. Assay Systems and Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243 A. Assay Systems ..... 243 B. Antigens . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244 C. Cell Preparation for the Assay 245 D. Migration Inhibition Techniques 247 E. Statistical Methods of Evaluating MMI and LMI ........................ 249 IV. Applications in Animal Tumors ........................................... 250 ................................... V. Applications in Human Tumors. . . . 253 A. Migration Inhibition Studies on General Immunocompetence . . . . . . . . . . . . . 254 B. Tumor-Specific Reactions. ............................................ 254 VI. Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295 References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296

.

I. Introduction

Leukocyte and macrophage migration inhibition (LMI and MMI, respectively) are in vitro correlates of delayed hypersensitivity (Bloom and Bennett, 1966; SZborg and Bendixen, 1967; Bloom et al., 1969; Halliday and Webb, 1969; Astor et al., 1973; Rocklin, 1975) and serve to measure the cellmediated immune response of sensitized hosts to the sensitizing antigen. Both assays are based upon the phenomenon that sensitized lymphocytes, upon encountering the corresponding antigen, release chemical mediators of immunity termed lymphokines. Leukocyte migration inhibitory factor (LIF) and macrophage migration inhibitory factor (MIF) are well-defined proteins (Rocklin et al., 1980).They act on polymorphonuclear (PMN) granulocytes and macrophages (M+), respectively, and inhibit their random movement. Migration inhibition studies are clinically valuable and are used to evaluate cellular immunity in a wide variety of experimental and clinical situations. Virus, tumor and tissue antigens that should not ordinarily be applied in vivo can be used in the migration assays. This article will consider the use of the LMI and MMI assay systems in experimental tumor immunology. Furthermore, we will summarize the various findings in migration inhibition studies performed to assess the cell24 1 ADVANCES IN CANCER RESEARCH. VOL. 43

Copyright 0 1985 by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12-006643-2

242

ROBERT SZIGETI

mediated immune reactivity of cancer patients in general and against tumorassociated antigens (TAAs) in particular. II. The Basic Phenomenon of Migration Inhibition

Production and release of lymphokines are among the earliest events of lymphocyte activation. Lymphokines are produced in minute quantities and have significant biological effects. Increasing evidence suggests that these mediators play important roles in v i m and provide a network among various cell types participating in the immune defense system. The characteristics of lymphokines (and monokines) are reviewed in detail by Rocklin et al. (1980). Here we summarize only the main features of MIF and LIF as protein molecules. The existence of MIF was postulated by Svejcar and Johanovsky (1963) and demonstrated by David and David (1972). MIF in all species studied so far is a heat-stable protein. Complete purification of MIF has not yet been achieved. Its molecular weight is smaller than that of immunoglobulin (2543K and 65K in the guinea pig, 48-67K in the mouse, and 23-55K in the human). It may contain disulfide bridges. MIF activity can be destroyed by proteolytic enzymes and in some species it seems to be a glycoprotein in that it is sensitive to neuraminidase. The target of MIF is the M+. MIF interacts with M+s at the cell surface and MIF activity can be removed from supernatants of activated lymphocytes by incubation with M+s. MIF activity can be blocked by a-l-fucose. This sugar is believed to be part of the M+ receptor for MIF. The mode of migration inhibitory action of MIF is not clear. It seems to activate M+s and increase their microtubular density. The distinction between MIF and LIF was first demonstrated by Rocklin (1974), who characterized the latter lymphokine (Rocklin, 1975). Its molecular weight in the human was shown to be 68K. Recently, LIF produced by the cell line REH-1 proved to be a 58K molecular weight protein (Meshulam et al., 1982). Human LIF is a heat-stable protein having a serineesterase property. It is resistant 'to neuraminidase treatment and contains disulfide linkages. LIF acts selectively on PMN granulocytes. The mode of action may be very similar to that of MIF on M+s, but this is not known. Both lymphokines are produced in the early G , phase (Prystovski et al., 1975; Bendtzen et al., 1975) independent of cell proliferation (G6rski et al., 1975). The lymphocyte subpopulation(s) responsible for MIF and LIF production has not yet been clearly defined. At first it was thought that this function was exclusive to T cells. However, it was later found that both T and B

MIGRATION INHIBITION IN T UMOR IMMUNOLOGY

243

lymphocytes are capable of producing MIF and LIF (Yoshida et al., 1973; Rocklin et ul., 1974; Wahl et al., 1974; Chess et al., 1975; Rasenen et al., 1978a,b; Masucci et al., 1984a). T lymphocytes can be activated by mitogens (PHA, Con A) or by specific antigens. Antigen-induced production of MIF-LIF by T cells has been found to be M+ dependent, whereas mitogens can act independently of M+s (Nelson and Leu, 1975; Shevach et al., 1975; Petersen and Bendtzen, 1982; Bendtzen, 1983; Szigeti et al., 1984a). B cells may be activated for MIF-LIF production not only by polyclonal activators (PWM, EBV) but also by stimulation of the Fc, Ig, or C3 receptor both in the guinea pig and in man (Sandberg et al., 1975; Berman and Weigle, 1977; Neville and Lischner, 1977, 1981, 1982), although Rasanen and Arvilommi (1977) have not found LIF production after C3 and Ig receptor stimulation when tested with human unfractionated lymphocytes. The M+ requirement for B cells to produce MIF-LIF seems to be less stringent than that for T lymphocytes (Wahl et al., 1975; Twomey et al., 1977; Masucci et al., 1984a; Szigeti et al., 1984a). 111. Assay Systems and Techniques

A. ASSAYSYSTEMS 1. Direct (One-Step) Method In the one-step method, both lymphocytes and indicator cells (M+s or PMNs) are present in the same culture. When the antigen (mitogen) is added, the lymphocytes become activated and produce MIF-LIF that then acts on the indicator cells. The cells may be placed in capillary tubes, agarose plates, or droplets (see Section II1,D) and incubated with the activating agent for 18-24 hr. The area of migration is then measured. Advantages of the direct assay are that this test (1)is technically easier to perform than the two-step method, (2) requires fewer cells (buffy coat or peritoneal exudate cells) and no separation-fractionation procedure, (3)is less time-consuming, and (4) gives results within 18-24 hr. Negative results, however, may be difficult to interpret, because the lack of migration inhibition may result either from the impairment of the indicator cells, or from the absence of lymphocytes sensitized to the antigens used in the test. Furthermore, the lymphocytes present in the culture may be unable to produce lymphokines. The latter possibility can be excluded with the simultaneous tests of antigen and mitogens (e.g., PHA) (Morison, 1975; Hahn et al., 1976; Szigeti et al., 1980a) that should activate lymphocytes for production of mediators in a nonspecific way if they are able to do it at all. The cause of unresponsiveness

244

ROBERT SZIGETI

can be significant in immunodeficiencies and immunosuppression. The direct test can be offered therefore as a first screening assay. 2 . lndirect (Two-step) Method This assay consists of two steps. First, separated lymphocytes are cultured with the antigen (mitogen) without the indicator PMNs or M+s. The second step involves the assay of the supernatants derived from the cultured lymphocytes using unrelated indicator cells. The MIF content of the supernatants can be determined using human, murine, or guinea pig M+s, and the LIF content using PMNs. The two-step assay makes it possible to test the lymphokine source on fractionated lymphocyte subsets (Rocklin et al., 1974; Chess et al., 1975; Kapadia et al., 1982; Robbins et al., 1982; Szigeti et al., 1984a) if normal indicator cells are used. Frozen lymphocytes can be used as well as fresh ones (Szigeti et al., 1981b). The disadvantages of the indirect test are that it requires a longer time, more sterile conditions, and a larger amount of blood. Interestingly, most human LMI studies with tumor antigens have been performed with the direct test (McCoy, 1979). B. ANTIGENS Any nontoxic material can be applied in the assay system. In tests of tumor-associated antigens crude extracts are mostly used. These contain, of course, not only the presumed TAA, but various other antigens (organ-associated antigens, HLA antigens, etc.). Therefore, it is preferable to perform simultaneous tests with “mock” preparations from the same organ (e.g., normal vs tumorous renal tissue). Furthermore, crude preparations may contain substances that cause nonspecific migration inhibition (Wolberg, 1971; Kadish et al., 1976). Whether migration inhibition is indeed due to a MIF-LIF effect or to nonspecific damage of the migrating cells can be determined by a viability test or by applying anti-MIF (Yoshida et al., 1975; Geczy et al., 1975) or anti-LIF (Bendtzen, 1977) antisera to these lymphokines. However, such antisera are available only in limited quantities and their use has not been widespread. Interspecies differences may cause migration inhibition, e.g., in the ratmouse system (Szigeti et al., 1982a and unpublished observations). In the human, alloantigens do not seem to complicate the test. Reactions of unsensitized lymphocytes to normal HLA antigens do not mask TAA-induced LMI. Indirect evidence supports this fact (McCoy, 1979). Thus, the technique can be applied for testing allogeneic tumor antigens. Various methods have been used for antigen preparation from tumor tissues. Crude saline or buffered homogenates (Andersen et al., 1969), 3 M

MIGRATION INHIBITION IN TUMOR IMMUNOLOGY

245

KCl extract (McCoy et al., 1974), perchloric acid extracts (Guillou and Giles, 1973), cell sonicates (Wolberg, 1971; Szigeti et al., 1974a), cryostat sections (Black et al., 1974a), and formalin-fixed whole tumor cells (Ross et al., 1973) have been employed. Further purification of the crude extracts has been attempted using Sephadex-column-separated preparations (Cochran et al., 1976a; Kadish et al., 1976) or the basic protein component of the tumors (Light et al., 1975). In virus-related animal and human tumors, whole viruses or virus-determined semipurified or homogenously purified antigens have been used, e.g., mouse mammary tumor virus (MMTV) or purified gp52/55 glycoprotein of MMTV (Black et al., 1976; McCoy et al., 1978) with leukocytes of patients with breast cancer and partially purified Epstein-Barr virus (EBV)determined nuclear antigen (EBNA) in the human EBV system (Szigeti et al., 1981a; 1982b). Migration inhibition tests are applicable even following the purification process of antigens that have no detectable antibodies (Lawniczak et al., 1983). Extracts of tumor tissue culture cells have also been used. Such antigenic extracts were prepared from cells derived from in uitro growing lymphoid (Braun et al., 1972; Szigeti et al., 1980b), melanoma (Boddie et al., 1975a), breast and lung carcinoma (McCoy et al., 1976a, 1977a), as well as Ewing’s sarcoma (McCoy et al., 1977b) cell lines. These provide more standardized TAA preparations that might be more valuable for long-term tests. Most investigators select one or two specific concentrations of the antigen in their studies and dose-response curves are not given. Since antigen effectiveness is highly dependent on the concentration used and since almost any compound can cause nonspecific (toxic) migration inhibition at a sufficiently high concentration (Andersen et aZ., 1970; Kjaer, 1975), tests with several concentrations of the antigen are needed and dose-response curves must be carefully analyzed. Based on these results one or more proper concentration(s) can be selected for the tests. Statistical analysis of the results should consider several concentrations of the antigen (see Section 111,E). FOR C. CELLPREPARATION

THE

ASSAY

1 . Direct Test For the one-step (direct) assay both peritoneal exudate cells (PECs) and bu@ coat leukocytes can be used. These heterogeneous cell populations contain both the lymphokine producing lymphocytes and the indicator (migrating) M+s andlor PMNs. PECs can be obtained from experimental animals (guinea pig, mouse, rat)

246

ROBERT SZIGETI

with ip injection of inducer materials (sterile mineral oil, thioglycolate). The PECs obtained 3-4 days after injection can be directly used in the test. For human one-step LMI studies the bufi coat cell population of the blood can be used. Most investigators use a 1 g sedimentation of total blood (with plasmagel or dextran). This can be followed by lysis of the residual red blood cells with ammonium chloride or distilled water treatment. The cells are then washed. Although it has been described that some leukocyte populations that were unresponsive after two to three washings became responsive upon further washes (Kjaer, 1976b), most studies have been performed with leukocytes washed three times. LMI studies with chicken (Nagaraja et al., 1982), rabbit (Ansfield et al., 1980), or (Azadegan et al., 1981), and pig (den Hollander et al., 1981) buffy coat cells have also been reported. The direct test can also be performed using separated mononuclear leukocytes and adding them to unrelated indicator (migrating) cells (Bull et al., 1973). These indicator cells can be either M+s (derived from guinea pig or mouse) or human or chicken PMNs (Bellina and Salerno, 1981).Interspecies differences do not seem to be relevant, since both MIF and LIF cross the species barrier (see Section 11). 2. Indirect Test

The first step of the indirect migration inhibition assay is the preparation of lymphokine-containing culture supernatants. This is performed almost exclusively by using Ficoll-Hypaque-separated mononuclear cells. M+ depletion is usually neglected in most studies. [ M+-free lymphocyte populations can be abolished in antigen-induced MIF-LIF production (see Section I])]. The total mononuclear population can be further separated on the basis of E-EA rosetting (Szigeti et al., 1984a; Masucci et al., 1984a), T-m and T-g rosetting (Kapadia et al., 1982), or active rosetting of T cells (Hobbins et al., 1982). The lymphocytes are then incubated in antigen (mitogen)-containing medium (test cultures) or medium alone (control cultures). Both macro- and microcultures can be set up. Mitogens are usually used for “pulse” induction, and after 1-2 hr incubation are washed out to avoid any direct effect on the indicator cells. In any case, it is preferable to test the direct effect of the mitogens-antigens on the indicator cells at the proper concentrations. Altcrnatively, one part of the control cultures can be reconstituted by the antigen (mitogen) at the end of the “pulsing” or culture period (mitogens, antigens, respectively). Then, supernatants are harvested and used in the migration inhibition assay with or without dialysis.

MIGRATION INHIBITION IN T U M O R IMMUNOLOGY

247

The second step is the preparation of indicator cells. Guinea pig and mouse PECs can be used without any further manipulation except extensive washing. Human PMNs can be obtained with dextran sedimentation and Ficoll-Hypaque gradient centrifugation from total heparinized blood. The pellet contains considerable red blood cells that can be lysed (see above). The residual granulocyte population is applied in the test.

D. MIGRATIONINHIBITION TECHNIQUES

The migration inhibition phenomenon was first observed on spleen fragments of immunized animals (Rich and Lewis, 1932). Since then, various MMI and LMI assay systems have been developed (Pekarek and Krejci, 1974; Bendixen et al., 1976). Three of these techniques have been widely used: the capillary tube, agarose plate, and agarose microdroplet methods. 1. Capillary Tube ( C T ) Technique The technique was in common use when Sflborg and Bendixen (1967) applied it for humans. This technique, micromodified by Federlin et al (1971) and Maini et al. (1973), has been described in detail in the review of Bendixen et al. (1976). Five percent dextran solution is added to heparinized blood (ratio 1:5). After mixing slowly the tubes are warmed for 45-60 min at 37°C. Then the leukocyte-rich plasma is harvested, mixed with serum-free medium at the original volume, and centrifuged. The cell pellets are resuspended in the same volume of 10%horse serum-containing medium and washed (routinely three times). Aliquots of the obtained suspension containing 3 x lo6 cells are transferred to 2 0 - 4 glass capillary tubes (internal diameter 0.6 mm) (Drummond Hemocaps, Drummond Scientific Supplies, Broomall, Pennsylvania). The capillary tubes are sealed by melting one end in a flame, centrifuged for 10 min at 900 g, and cut at the cell-fluid interphase. Immediately after cutting, two tubes are placed in opposite directions in a 0.5-ml culture chamber (Sterilin Migration Chamber, Sterilin Ltd., Middlesex, England) and fixed to the chamber by silicone wax. Antigen-containing or control medium is filled into the chambers to the brim and covered with a microscope coverslip, retained by silicone wax. After 24 hr at 37°C the migration areas are studied with a projection microscope or photographic magnifier and measured. The average migration areas of test cultures ( M , ) and those of control cultures (M,) give the migration index (MI = M J M , ) or the percentage of migration inhibition [%MI = (l-MJM,) x 1001. An MI of less than 1.0 indicates migration inhibition and should be evaluated after statistical

248

ROBERT SZIGETI

analysis (see Section 111,E). The CT technique is the most frequently used migration inhibition technique.

2. Agarose Plate Test (AP) Clausen's original method (1971) has also been detailed in the same review (Bendixen et al., 1976). Briefly, a 2% agarose solution in distilled water is heated in boiling water. After sol formation, the agarose is cooled to 47°C in a water bath. For the preparation of the 100 ml agarose culture medium, 50 ml of 2% agarose solution is added to 31 ml of sterile distilled water, 9 ml of 10-fold concentrated TC-199 medium, 10 ml horse serum, 200 pl penicillinstreptomycin in sterilized water (TC-penicillin/streptomycin) and 540 p1 of 10% sodium bicarbonate. Thus the final agarose medium contains 1% agarose, 10% serum, and antibiotics. The 47°C agarose medium is transferred to plastic petri dishes (48 x 8.5 mm, 6 ml medium each). After the gel has solidified, the dishes are incubated at 37°C in a humidified CO, atmosphere. Six to nine wells of2.3 mm diameter are punched in the gel and 9-pl aliquots of migrating cells mixed with antigens (test supernatants) or with medium (control supernatants) from a suspension of 15 X lo6 cells in 90 p1 volume are placed in the wells. After 18-24 hr of incubation and fixation for 1 hr with 5% glutaraldehyde, the agarose -ayer can be removed and migration areas can be measured. 3. Agarose Microdroplet Assay (AD) This modification of the migration inhibition assay has been developed by Harrington and Stastny (1973) and Harrington (1974) in the guinea pig and mouse, respectively, and applied to humans by Hahn et al. (1976) and McCoy et al. (1977~).In our laboratory we use this technique (Szigeti et al., 1980a, 1981a), as follows: 0.4% agarose solution is heated to boiling, then placed in a 37°C water bath for several minutes. The same volume of twofold concentrated RPMI-1640 medium, containing 20%horse serum and antibiotics, is added to the liquefied agarose giving a final medium containing 0.2% agarose and 10% horse serum. Washed bufl) coat cells (20 x 106) from heparinized, dextran-sedimented blood is added to 135 pl of this nutrient agarose medium. After vigorous mixing, 3 X 2 pl droplets of the suspension are placed into each migration chamber (Sterilin, as above). When the droplets have solidified, the chambers are filled with the proper (antigen-containing or antigen-free) medium, covered, and kept at 37°C in a humidified CO, atmosphere for 18-24 hr. The migration areas are projected and the diameters of the inner (droplet = d)and outer (cell migration from the droplet = D)circles are measured. The area of the migration (A)is calculated according to the formula of Weese et al. (1978a):A = (.rr/4)(D,D2-d,d,) in a computer

MIGRATION INHIBITION I N TUMOR IMMUNOLOGY

249

program. The AD technique can also be used for indirect assay (Hahn et al., 1976; Suslov et al., 1980).

4 . Other Technical Variations Standardization of the biological methods mentioned above has been attempted using 51Cr labeled indicator cells (Noronha-Blob and Huang, 1980). Another experimental approach takes advantage of the esterase activity of LIF (see Section 11), using radiolabeled amino acid ester substrates. The method is highly sensitive and results with mitogen-induced LMI are reproducible but do not correlate well with antigen-induced LMI (Rocklin and Rosenthal, 1977; Rocklin and Urbano, 1978; Bendtzen and Rocklin, 1980).

E. STATISTICALMETHODSOF EVALUATING MMI

AND

LMI

Statistical evaluation is the most critical point in migration inhibition studies, especially in the early diagnosis of cancer or in longitudinal tests. The ideal statistical method should point out the boundary between “negative” and “positive” and minimize the incidence of “false” reactions. In spite of its importance, a significant number of investigators evaluate their results with a practical, but arbitrary cut-off point for significant inhibition. This point is most often MI = 0.8 or %MI = 20. In those studies in which statistical analysis has been performed, the Student’s t test is the most commonly used and it is generally permissible. With fewer than 6-8 simultaneous determinations, however, a nonparametric test (Wilcoxon’s)is preferred (Bendixen et al., 1976). These analyses should be performed at each concentration of the extract (antigen) used in the test. Few statistical models have been developed that might allow the use of a test to determine whether a given MMI-LMI value reflects a truly positive response. Andersen et al(1970), Kjaer (1975), and Kadish et al. (1976) used “normal nonpathological range” for this purpose, assuming a normal distribution pattern of the data. The model is based on the mean +SD of the MIS of normal control donors with antigenic extracts. The 2 SD cut off borders the normal range, and each MI out of this range is to be considered as pathological. In our own studies with EBV antigen-related LMI we use this model, and it correlates well with serological data and other immunological tests (Szigeti et al., 1980b, 1981a-c). McCoy et al. (197613)adopted other cut-off values. They consider an MI as positive if it is lower than the lower tenth percentile of the normal control donor observation. This diminishes the incidence of “false positives.” In this model, however, 10% of the normal donors will always have a “pathological” MI value.

250

ROBERT SZIGETI

It should be emphasized that all of these analyses are based on the reactivity of normal donors. Furthermore, the technical variation of repeated tests on the same donor can be in the range of 7-10% (Andersen et al., 1970). Therefore, data obtained with separate, repeated tests may be more accurate as parameters of cell-mediated immunity. This relatively detailed description of the practical performance of the assays is provided here to allow the reader to evaluate MMI-LMI tests in tumor immunology in a critical manner. IV. Applications in Animal Tumors

The inhibition of migration of peritoneal exudate cells (PECs) in sensitized experimental animals was first proposed as a model for the detection of delayed hypersensitivity by George and Vaughan (1962).This technique was then applied to a variety of experimental tumor systems. Cellular immune reactions to spontaneous, chemically, and virally induced tumor-associated or tumor-specific transplantation antigens (TAA,TSTA, respectively) were studied in various animal tumor systems. The pioneering experiments of Bloom et al. (1969) in guinea pigs (GPs) revealed that soluble antigens of three different chemically induced tumors inhibited the migration of PECs from sensitized animals and no cross-reaction was found. These observations have been confirmed by Halliday and Webb (1969) in mice and extended with the experimental proof that in uiuo hypersensitivity measured with the footpad reaction was correlated with MMI. The same results were found by Kronman et al. (1969)in G P hepatoma. Churchill et al. (1972) noted that DMBA-induced hepatoma cells treated by X irradiation, trypsinization, or storage in liquid nitrogen after freezing retained their capacity to stimulate MIF production by sensitized lymphocytes in GPs. Similarly, a 3 A4 KCI extract of intact or mitomycin-treated tumor cells was found to induce specific MIF release in sensitized GPs (Littman et al., 1973). Tumor immunity was lower in animals with tumors than those immunized against the tumor, and the reaction was augmented and appeared earlier after immunization with a mixture of tumor cells and bacillus Calmette-Gukrin (BCG). The purification of TAA in chemically induced GP sarcoma was attempted by Suter et al. (1972) and resulted in two active fractions. These studies, however, have not been continued. Similar studies have also been performed in the murine methylocholanthrene (MCA)-induced tumor system: the migration of PECs was specifically inhibited by corresponding tumor extracts (Halliday and Webb, 1969; Halliday, 1971; Vaage et al., 1972; Cerilli and Smith, 1972) and the reaction was abrogated by serum from tumor-bearing mice (Halliday, 1971). Depression of fully developed immunity to MCA tumors and spontaneous

MIGRATION INHIBITION IN TUMOR IMMUNOLOGY

251

mammary carcinoma by excess tumor antigen was not, however, reflected in alteration of MMI response (Vaage et al., 1972). In later studies with the direct MMI test, Padarathshingh et al. (1978) detected as low as picogram quantities of NP-40 solubilized MCA sarcoma membranes in sensitized mice, and with the indirect assay they demonstrated that the reaction was indeed mediated by MIF. The physicochemical properties of the membrane antigen, however, have not yet been clarified. Further experiments revealed that MCA tumor-bearing mice responded with MMI even to human myelin basic protein (Yong and Halliday, 1982) and MMI appeared to be more sensitive than E rosette augmentation or leukocyte adherence inhibition. It is noteworthy, that in both studies, the agarose microdroplet assay (Harrington and Stastny, 1973) was performed. In addition to investigating cell-mediated immunity to chemically induced tumors, the suitability of MMI for the study of immune response to virally induced tumors and the activity of such tumor extracts as antigens in MMI have been studied. Halliday (1971, 1972) demonstrated that PECs from mice bearing primary Moloney virus-induced tumors (“progressor” mice) and those from mice in which these tumors had spontaneously regressed (“regressor” mice) are distinguishable. The corresponding soluble tumor antigen induced MMI with “regressor,” but not with “progressor” PECs. Sera from the two types of mice were also different: “progressor” serum blocked the MMI found with “regressor” PEC. “Regressor” serum failed to have this property. The Moloney murine sarcoma virus (MSV)-MMI system was also used to clarify some general events in (experimental tumor) immunology, particularly studies on the regulatory role of M+s in cell-mediated reactions against tumors (Herberman et al., 1980; Holden et al., 1980). Basically from the works of Landolfo et al. (1977a, 1977b, 1978), it was found that viable and metabolically active M+s were required to interact with soluble (3 M KC1) tumor antigen and the M+-lymphocyte interaction was controlled by the ZA subregion of the H-2 complex. When intact tumor cells were used as antigens, M+s and H-2 compatibility were not required for MIF production. These findings suggested that two different mechanisms, mediated by two separate subpopulations of immune lymphocytes, might be responsible for production of MIF. Further experiments revealed that M+s could mediate negative or inhibitory effects on production of MIF and other lymphokines and on the response of M+s to MIF. The suppression was not antigen specific and the effector cells were heterogeneous with several subpopulations evident in regressing tumors but only one population in progressing tumors. A significant difference was found between immune spleen- and tumor-infiltrating lymphocytes inasmuch as the latter did not produce MIF, possibly due to the suppressive effect of M+s (Herberman et al., 1980;

252

ROBERT SZIGETI

Holden et al., 1980). These observations were extended in a recent study of Sarlo and Mortensen (1983) with murine sarcoma virus (MSV)-immune T cells: MIF was formed only by Lyt-l+ cells and tumor-infiltrating M+s elicited a leukocyte adherence enhancing factor from Lyt-2 cells. Tumorrelated MMI reactions in EL-4 lymphoma (Harrington, 1977) and suppression of MMI by spleen cells in P388 murine lymphocytic leukemia (Nori and Gothoskar, 1983) have also been reported. The nature and properties of the antigen(s) responsible for MMI reactions are not known. Preliminary studies in our laboratory revealed that only extracts containing Moloney virus-determined cell surface antigen (MCSA) and specific MMI inductive activity, whereas gp70 and p30 virally coded proteins had no such effect (Szigeti, Fenyo, and Asjo, unpublished data). These findings suggest that this defense mechanism is directed against cellular components determined by the virus but not against viral proteins. Tumor-specific migration inhibition has also been found in animals with SV40 induced tumors. In their early study, Pekarek et al. (1968) used spleen cell migration in hamsters with the SV40 system. McCoy et al. (1977d), using the direct agarose microdroplet technique, could detect MMI activity with cell-free crude membranes, and with papain-solubilized and N P-40 detergent-solubilized membrane extracts at concentrations as low as 250 pg protein/ml in SV4O-induced murine sarcoma. Specific MMI has also been reported in adenovirus 12-induced murine tumors (Rees and Potter, 1973). As mentioned above, Halliday (1971, 1972) reported the blocking effect of sera from tumor-bearing mice on TAA-induced MMI. Other studies confirmed these findings. Poupon and Lespinats (1972) and Poupon et al. (1974) found that such sera directly induced MMI without addition of extraneous antigen, indicating the presence of free, circulating, tumor-specific antigen. These experiments were carried out in BALB/c mouse plasmacytoma. The nature of the antigen, however, remained unclear. A characteristic T cell-dependent immune surveillance system is represented by the resistance of mice against the outgrowth of polyomavirus (PV)transformed tumor cells. The effector mechanism has not been dissected in detail (Allison, 1980), nor has the target antigen(s) been identified so far, in spite of the discovery of PV-induced TSTA more than 20 years ago (Sjogren et al., 1961). One obvious reason for the latter is the lack of detectable antibodies against TSTA and the lack of appropriate short-term tests that would allow its detection. We have applied the MMI assay to study CMI responses to antigens of PV-transformed cells. M+s from PV-immunized mice, but not from control mice, showed significant migration inhibition when exposed to extracts of mouse and rat PV-transformed cells. Extracts of non-PV-induced tumors had no such effect (Szigeti et al., 1982a). The nature of the target antigen and its relationship to known T antigens are unknown, +

MIGRATION INHIBITION IN TUMOR IMMUNOLOGY

253

although our preliminary experiments showed that NP-40-solubilized membrane fractions from PV-transformed cells had the same virus-specific effect, even at 2-10 pg protein/ml (Szigeti et al., 1984d). In a recent study, we examined the MMI response of mice immunized with mutant PV or with cells transformed/infected by them. The MMI reaction revealed individual differences. In several cases mice immunized with a mutant virus responded preferentially or exclusively to extracts of cells transformed or infected with the corresponding mutant. Moreover, in the MMI test, mutant viruses were usually less immunogenic than the corresponding transformed/infected cells (Szigeti et a l . , 1984b). These findings suggested that abnormal protein products of the middle-T mutants alone or in association with the cell membrane could act as the main immunogen in the MMI test. Alternatively, a cellcoded component would be responsible for the viral mutation-associated distinctness of the MMI. Migration inhibition studies have also been performed in tumor-bearing rats. Steiner and Watne (1970) were able to show MMI response in rats in which PECs were incubated with autologous tumor extracts. Interestingly, spleen cells from normal rats produced MIF upon culture in uitro with puroymycin-treated rat cell lines expressing leukemia viruses or endogenous C-type viruses. In addition, lymphoid cells from the blood, peritoneal exudate, and lymph nodes but not thymus also were able to produce MIF in response to tumor cells. The MIF production of spleen cells was M+ dependent (Sharma et al., 1979). Similarly, MMI response was demonstrated in rat Zajdela hepatoma (Fahlbusch et al., 1981; Fahlbusch and Zschiesche, 1981). A unique application of LMI was the detection of cell-mediated immune reactivity of Sinclair melanoma-bearing swine to 3 M KCl extracts of swine and human melanoma (Hook et al., 1983). LMI response was found in 70% of melanoma-bearing compared to 10% or less of normal swine. Normal fetal swine tissue was not active in the test. Swine and human melanoma extracts were equally capable of inducing LMI both in melanoma swine and in human melanoma patients suggesting that these tumors might share a common TAA. V. Applications in Human Tumors

The assessment of cellular immunity in cancer patients is of considerable interest in two different respects: (1)in determining whether the presence of the tumor is associated with depressed (altered) cell-mediated immune competence and (2) in determining whether the neoplasms have tumor-associated antigens which can induce immune reactions and which might be involved in the defense mechanism of the host. Migration inhibition assays have been applied to answer both questions. Studies on general immu-

254

ROBERT SZIGETI

nocompetence will be briefly discussed in this section and we will focus in particular on tumor-specific migration inhibition studies.

A. MIGRATIONINHIBITION STUDIESON GENERAL IMMUNOCOMPETENCE The same recall antigens (tuberculin, candidin, mumps, streptokinase/streptodornase)that are ordinarily used to elicit delayed cutaneous hypersensitivity reactions can also be applied in migration inhibition tests. Generally, a decreased cellular hypersensitivity to these antigens has been observed, especially in advanced metastatic malignant disease (Alth et al., 1973;Tautz et al., 1974; Kadish et al., 1976; Walden et al., 1976; Yamamoto, 1980; Wolf et al., 1980; Pettersson et al., 1982; Abai and Fekete, 1982). Investigations with recall antigens, however, require a confirmed presensitization state. Therefore, the results of such studies are difficult to interpret in infants and young children, and in populations for which certain vaccinations (e.g., BCG) are not obligatory or general. The application of mitogens (mostly PHA, Morison, 1974) in the LMI test (as the measure of the ability of lymphocytes for LIF production) can eliminate this problem, since no presensitization is required. PHA-induced LMI assays have revealed decreased reactivity in patients with breast cancer (Wolberg, 1971, 1974; Pacheco-Rupil et al., 1982), Hodgkin’s disease (Golding et al., 1977; Szigeti et al., 1982c), and immunodeficiencies with increased susceptibility to neoplasms, such as Down’s syndrome (Hahn et al., 1976; Szigeti et al., 1980a) and ataxia telangiectasia (Masucci et al., 198413). On the other hand, LMI studies in cancer patients with Con A did not show decreased reactivity (Lidereau et al., 1983). Although Abai and Fekete (1982) claimed that PHA was found to inhibit directly the migration of indicator cells by agglutinating them and others found this mitogen to be highly cytotoxic (Wolberg, 1971), in our experience 1 pg/ml purified phytohemagglutinin (PHA-P, Burroughs Wellcome, Ltd., England) failed to have any direct cytotoxic or leukoagglutinating effect (Szigeti et al., 1980a,c). Moreover, direct PHA-LMI assay can be recommended as a first screening test for immunodeficiencies (Szigeti et al., 1980a), followed by selective measuring of LIF production of lymphocytes with indirect assay (Szigeti et al., 1981b; Masucci et al., 1984a).

B. TUMOR-SPECIFIC REACTIONS I . Breast Cancer ( B C ) Breast cancer was the first human tumor tested with LMI. Andersen et al. (1969) found in their preliminary study that a significant proportion of pa-

MIGRATION INHIBITION IN TUMOR IMMUNOLOGY

255

tients with BC demonstrated cellular hypersensitivity against extracts of autologous tumor tissue. Since then a large number of LMI studies in BC has been reported. Table I gives the results of many of these studies and reflects the advantages and disadvantages of the assay system (see Section 111). Andersen et al. (1969, 1970)first showed that leukocyte migration in some BC patients was inhibited by their autologous tumor extracts, whereas normal breast tissue extracts had no such effect. These findings were confirmed by others (Segall et al., 1972; Churchill and Rocklin, 1973). The rate of reactivity was found to be 30-50% among BC patients in these studies. The same figures for LMI responses have been reported in subsequent studies in which good reactivity of BC patients was found even to allogeneic BC tumor extracts (Cochran et al., 1972a; Jones and Turnbull, 1974; McCoy et al., 1974; Black et al., 1974a). The results, however, were difficult to compare because various antigen preparations and statistical (or arbitrary) evaluations were used. Standardization of the antigenic substance was attempted by McCoy et al. (1976a), who found reactivity of the majority of BC patients against a 3 M KCI extract of MCF-7, a BC-derived cell line. Although only a small proportion of normal control donors were responders, patients with benign breast diseases (BBD) also had a higher frequency of reactivity (Cochran et al., 1974; McCoy et al., 1974, 1976a; Kadish et al., 1976; Cannon et al., 1978a, 1982). These findings indicate that patients with both neoplastic and benign breast disease might become sensitized to “organ-specific,” breast tissue antigen(s). Furthermore, cross-reactions with other, histologically dissimilar tumor extracts (similar to those from patients with laryngeal cancer and from patients with tumors of the digestive tract) suggested that the reaction was directed toward a “group” of TAAs shared by several tumors or a “common epithelial antigen” (Mantovani et al., 1981). The issue is still far from being clear, since the TAA responsible for the migration inhibition response has not been identified or purified. Kadish et al. (1976) claimed they had found a high molecular weight “cancer-specific” and a low molecular weight “organ-specific” antigenic fraction active in LMI, when BC tissue extracts were separated on Sephadex G-200 columns. This finding is similar to those described for delayed cutaneous hypersensitivity reactions (Hollinshead et nl., 1974a; Weese et a l . , 1978b). Further purification, however, was not performed. Some investigators reported LMI responses when BC patients leukocytes were exposed to fetal antigens (Albrecht et al., 1978; Matzku et al., 1979; Pasternak et al., 1983). The biological significance of these findings is not yet clear. An intriguing issue of migration inhibition studies in BC is the observation

TABLE I TUMOR-SPECIFIC LMI STUDIES IN BREAST CANCER

Reference

lo u1 Q,

Leukocyte donors.

Andersen et al. (1969)

8 BC

Andersen et al. (1970)

22 BC, 9 BBD, 9 NCD

Wolberg and Goelzer (1970)

5 BC

Wolberg (1971)

7 BC (of 21 cancer patients)

Cochran et al. (1972a)

45 BC, 45 MM, 95 NCD

Antigenb Saline extract of BC tissue (100 wg/ml) Saline extract of BC tissue (100 wdml)

Cell suspension or ultrasonicate of BC tissue Cell suspension or ultrasonia t e of BC tissue

Supernatant of tumor homogenate (12.5-100 wgW

TechniqueC

Positive reaction (MI)d

Ratio of positive reactions

D,CT

c0.80

5/8 BC

D,CT

c0.78

8/22 BC

D,CT

nge (calculated with t test)

5/5 BC

D,CT

ng (calculated with U test)

6/7 BC (15/21 all cancer patients) to autologous tumor extracts

D,CT

ng (calculated with U test

28/45 (12145 prior to operation) BC (6/8 BC in autologous test), 37/46 MM (12/46 preoperative) 18/64 NCD

Note

Extracts from normal breast tissue did not induce LMI. BC tissue-induced LMI was lost within the first month after operation and/or X-irradiation theraPY Antigen concentrations and tests on NCD are not indicated Antigen concentrations and tests on NCD are not indicated. Tumor extracts inhibited migration of both autologous and allogeneic leukocytes The degree of LMI was reduced after operation until day 6-22

Segall et al. (1972)

Churchill and Rocklin (1973)

Anderson et al. (1973)

w

2

Cochran ct al. (1973a)

Ross et aE. (1973)

13 BC (of 57 various cancer patients) 7 BC, 9 c c , 10 NCD

11 BC with and 12 BC without autograft, 16 NCD 73 BC, 55 MM, 162 NCD

39 BC, 27 MM, NCD (ng)

Hypotonic extracts of tumor tissues

D,CT


33/57 cancer patients (8113 BC)

CEA, tumor cells mixed with autologous lymphocytes Tumor cell suspension

1,CT (GP M44

ng (probably

417 to tumor cells, 019 to CEA

D,CT (tumor cell mixed with donor leukocytes) D,CT

ng (calculated witht test)

ng (calculated with U test)

38/73 BC, 33/55 MM to allogeneic tumor extract, 9/63 and 19/99 NCD to BC,MM extracts, resp.

D,CT

ng

43/66tests to

Supernatant of tumor homogenates (25-200 CLg/ml)

Formalin-fixed tumor cells

<0.80)

Cross-reactions were observed in the MM group. No correlation with the clinical stage. Tests on NCD are not indicated

LMI of autograft patients was significantly higher than that of nonautograft patients in the presence of allogeneic tumor cells and than LMI of NCD LMI in patients was infrequent on contact with extracts from histologically dissimilar tumors and nonmalignant breast tissue. There was a striking decline in the frequency of LMI in patients with metastases

histologically similar, 9/39 to dissimilar tumor cells (continued)

TABLE I (Continued)

Reference

Leukocyte donors.

Antigenb

Positive reaction (MI)d

Ratio of positive reactions

D,CT

<0.75 (calculated with t test)

12/14 BC, 25/41 all donors)

Saline extract of tumor homogenates

D,CT

ng (at 5% level with U test)

54% BC, 15% NCD

34 BC

lo00 g supernatant and 3000 g sediment of tumor homogenates

D,CT

<0.79

16/34 in autologous, 16/30 in allogeneic tests

14 BC, 9 BBD, NCD, other cancers

3 M KCI extract

D,CT

<0.80

20126 BC, 3/25 NCD, 2/10 BBD, 219

14 BC of 41 other cancer patients

Ultrasonicates of tumor tissue (1-5

Cochran et nl. (1974)

138 BC, 96 BBD, 28 other cancer, 37 nonmalignant disease, 26 NCD

Jones and Turnbull (1974)

McCoy et al. (1974)

Wolberg (1974)

TechniqueC

mglml)

of BC, BBD, and normal breast tissue

other cancer patients (no. of tests)

Note Tests on NCD are not reported. Patients with tumors that induced LMI had poor prognosis. Only few BC preparations inhibited leukocyte migration Patients with localized tumor showed LMI more frequently than those with advanced disease. BBD extracts infrequently caused LMI in BC patients. No axrelations were found with clinical features No difference was found between tests performed in day 7 and weeks 8-12 after operation. Autologous-allogeneic cross-reactions were interpreted as indicating viral etiology Allogeneic reactions suggested that ditferent BCs had common antigen

Black et al. (1974a)

BC, BBD, NCD (numbers are not given) 124 BC, BBD, NCD (numbers are not given)

Cryostat sections of BC tissue

D,CT


Cryostat sections of BC tissue and R 111 mouse milk samples

D,CT

<0.75

Akiyoshi et al. (1975)

80 various cancer patients

Saline extracts of tumor tissues (mostly 1 mglml)

I,CT (GP M4s)

<0.75

Ross et d.(1975)

81 BC, 69 MM, 61 NCD and other cancer 105 BC

Formalin-&xed tumor cells

D,CT


55/81 BC, 38/69 MM, 5% NCD

loo0 g supernatant and 3oOO g sediment of BC homogenate, (50-200 w/mU

D,CT, with human bufFy coat or mononuclear cells with GP M+s


At day 7:47% of BC to autologous, 40% to allogeneic BC extract; 23% at 2 months, 19% at 6 months

Black et al. (1974b)

8

Jones and Turnbull (1975)

36% BC to autologous (invasive) BC tissue, 0/12 to normal tissue 31% BC to FUII milk, 33%BC to homologous in situ BC tissue 29% BC to autologous invasive, 16% to homologous invasive BC tissue 26% of total group, 9/22 BC to autologous tumor extract

Stage-response relationship to autologous BC tissue: in situ 1,II. Cancer-free patients showed 2 years persistence of LMI response Responsiveness to BC tissue correlated with cross-reactions to RIII milk. Antigenicity of BC was a reflection of a component that was similar to that found in murine mammary tumors

Relative frequency of MIF production varied inversely with progress of the diseases with an expressed postoperative decline N o cross-reactions were found with MM extracts or extracts from histologically dissimilar tumors Only postoperative tests! Reactions to BBD tissue extracts were rare

(continued)

TABLE I (Continued)

Reference

Leukocyte donors"

Antigenb

Techniquec

Positive reaction (MI)d

Ratio of positive

reactions and 34% at 1 year after surgery to autologous extract 14/40 BC, 11/26 BBD

Roberts and Bathgate (1975)

40 BC, 26 BBD

Saline extracts of BC tissues (50-150 wdml)

D,CT

<0.80

Black et ol. (1976)

198 BC

Cryostat sections of BC, RIII milk gp55 of MuMTV Supernatant of tumor homogenates andlor formalin-fixed tumor cells

D,CT

<0.80

39/77 to autologous, 65/198 to RIII milk, RIII gp55

D,CT

ng

8/10 BC to autologous, 55/104 BC to dlogeneic tumors, 241105 NCD to tumors 8/23 cancer patients' sera induced LMI

D,CT

<0.95 (loth percentile of normal Val-

27/32 BC, 1/24 NCD, 7/16 other cancers

Cochran et al. (1976~) 59 BC, 54 MM, 60 BBD, and 51 NCD

McCoy et al. (1976a)

32 BC, 24 NCD, 16 other cancers

3 M KCl extract of MCF-7 cell line (2.5 50 PdmU

ues)

Note

Tests were performed to autologous BC extracts in the 3rd week after mastectomy. Poor correlation was found with skin tests In situ BC tissues were more regularly immunogenic. Responsiveness to RIIII milk and MuMTV were correlated to gp55 reaction LMI failed to show ditference between cancer patients with early and advanced disease

MCF-7 cells possessed a TAA to which most BC patients were sensitized. The cell line provides

standardization of antigen for longitudinal BC-LMI studies Jones (1976)

Jones et

al. (1976)

BC, BBD, BCD, other cancer patients (numbers are not clearly indicated) 107 BC, 58 others

@BD, NCD, Rieche et al. (1976)

various cancers) 58 BC,

BBD

lo00 g supernatant and 3OOO g sediment of BC homogenates (100200 pg/ml)

D,CT and 1,CT

lo00 g superna-

D,CT

tant and 3OOO g sediment of BC homogenates (100200 pglml) BC homogenate (100 d m l )

<0.80

11/24 in D, 10124 in I tests

c0.80

55/107 BC at day 7, 28/106 at month 2, 31/79 at year

(with human

PMN granulocytes)

one after oper-

D , a

<0.70

(number is not given)

Brandes and Goldenberg (1976)

33 BC, 129 NCD

Saline BC extracts (mostly at 100-250 &ml)

ng

C0.80 (andlor t test)

ation 52/58 BC to autologous and allogeneic extracts, 1-40 postoperative day. 22.7% of NCD 23/23 BC in remission, 6/10 in relapse, intermittently. 126/129 NCD intermittently

Autologous postoperative tests. At 2 months LMI was frequently found in patients with axillary lymph node involvement

BC sera abrogated LMI. Positive reaction was associated with progressive disease. Neither BBD nor normal tissue extracts induced LMI in BC patients Repeated tests. Highly overlapping LMI reactivity between BC and NCD

(continued)

TABLE I (Continued)

Reference

Leukocyte donorsa

Springer et al. (1976)

39 BC, 10 BBD, 27 NCD

Kadish et al. (1976)

30 BC, 6 BBD, 9 NCD, 7 other cancers

Dean et al. (1977)

34 BC (2 female and 2 male BC tested with LMI) BC, BBD, NCD (numbers are not given)

TechniqueC

Positive reaction (MI)d

Ratio of positive reactions

ThomsenFriedenreich (T), antigen (500-4000 CLdml) 3 M KCI extracts of BC

D,AP


38% BC, 1/10 BBD, 1/27 NCD

Postoperative tests

D,m

C0.78

315 BC to autologous, 11/20 to allogeneic extracts, 3/22 NCD

3 M KCI extract of BC and MCF-7 cells (50 and 750

D,CT

ng

100% BC

Cross-reactivity was found between BC and BBD extracts. Fractionation on Sephadex G-200 column resulted in a high MW “cancer”- and a low MW “organ”-specificfraction. Further purification was not performed BC in males and females possessed common TAA. One third of the BC patients showed lymphoproliferative response with the extracts

D,CT

<0.84

38/72 BC, 4/26 other cancer, 16/34 BBD, and 6/61 NCD to BC extract, 741104 BC to MCF-7

Antigenb

%

w

Cannon et al. (1978a)

CL~W 3 M KCI extracts from BC, BBD, and MCF-7 cells (50-750 &ml)

Note

Pre- and postoperative tests. The incidence of reactivity in BC and BBD was similar. LMI was not useful in discriminating malignant and benign breast lesions

Cannon et al. (197%)

250 tests in BC and MM, 50 tests in other cancer patients, 113 tests in NCD 60 BC, 26 BBD, 15 other cancer patients, 16 NCD

3 M KCI extracts of BC and MCF-7 (50-750 pglml)

D,CT

CO.84 (BC)
MuMTV (1-500

D,CT


Mantovani et al. (1978)

47 BC, 14 BBD, 56 NCD

Formalin-bed BC cells

D,CT


38/47 BC, 10/14 BBD

Albrecht et al. (1978)

17 various cancers, 18 NCD

3 M KCI extracts of BC and fetal tissues

D,AP


11/17 cancer patients (314 BC)

McCoy et al. (1978)

&ml,

(MuMTV), <0.92 (gp52), <0.91 (MCF-7)

MuMTV gp52 (0.5 wlml), 3 M KCI extract of MCF-7 (25-50 Pdm4

21-87% of BC and MM to homologous tumor extracts, 8-14% NCD, 11-20% of other cancers with various extracts 13/15 BC, 315 BBD, 1/16 NCD, 0/9 other cancers to MuMTV; 39/60 BC, 10126 BBD, 1/15 NCD, 3/15 other cancers to

Variability might he due to the patient’s reactivity and the immunogenicity of the extracts. Primary BC and local MM tissues were more immunogenic than metastatic lesions

Studies with gp69/71 of Rauscher MuLV and Kirsten MSV gave high incidence of reactivity in BC, BBD and NCD. There was also found LMI reaction to fetuin and transfenin

gpsz Patients with metastases were more frequently unresponsive. No correlation was found with histopathology and type of the treatment

(continued)

TABLE I (Continued)

Reference

E

Leukocyte donorsa

Fukuda et al. (1979)

96 BC, 22 BBD, 10 head and neck cancer, 40 NCD

Wanebo et d.(1979)

96 BC, 32 BBD, 67 NCD

Matzku et d.(1979)

9 BC, 29 CC, 41 RC, 11 LC, 22 GC, 34 benign disease, 36 NCD

Fukuda et al. (1980)

94 BC, 32 BBD, 20 head and neck cancer, 67 NCD

Antigenb Cryostat section, Saline and 3 M KCl extract of BC, MCF-7 and 2 other cell lines MuMTV, Mason-Phr monkey virus (MPMV), MCF-7 3 M KCl extract of tumor and fetal tissue, fetal organs

MuMTV, MuLV, MPMV (5-30 pglml), saline extract of MCF-7 cells (70 &ml)

Positive reaction (MI)d

Ratio of positive reactions

D,CT

Various (at 5th- loth percentile of NCD)

D,CT

Various (at 5th-10th percentile of NCD)

D,CT

(0.80 (and U

20162 BC to autologous, 42/96to allogeneic BC cryostats, 36/72 to MCF-7 extract BC: 49% to MuMTV, 50% to MCF-7; 75% in simultaneous tests 71-93% to organ-related tumor extracts 048% to nonorgan-related tumor extracts 70% to organrelated, 16% to non-organrelated fetal extracts 46/94 BC to MuMTV, 75% BC to MuMTV and MCF-7 simultaneously

Techniquec

test)

D,CT

Lowest loth percentile of NCD

Note

High response in BC to allogeneic BC antigens, low response in NCD and BBD were found

No relation was found to age, menstrual status, estrogen receptors, and stage of the disease The patient’s leukocytes

were sensitized to crossreactive fetal determinants of organ-related specificities. BC patients showed pronounced reactivity to fetal skin extracts

Springer et a1. (1980)

Mantovani et al.

(1981)

Akiyoshi et al. (1981)

2!

30 BC, 84 BBD, 25 NCD 110 BC, 48 cancer of the digestive tract, 60 laryngeal cancer, 93 NCD

T antigen (10 pg110 pl)

D,AP

c0.90

16/30BC, 11/84 BBD, 1/25

Formalin-fixed tumor cells

D,CT

<0.80

82% BC to homologous, 7295% to lung

23 GC, 16 BC, 10

3 M KCI extract

CRC

Cannon et al. (1982)

BC, BBD, other benign gynecologic diseases (only test numbers are given)

NCD

of autologous tumors

MuMTV, 25 pdmk MuMTVgp52, %-1 pglml; 3 M KCl extract of MCF-7 cells (so CLglml)

I, MMI, CT (GP MW

<0.75

D,CT 1,(with autologous PMN granulocytes)

Lower loth percentile of NCD values, preoperative benign gynecological diseases (<0.86-

and laryngeal tumor cells, resp. 22% NCD to various tumor cells (66% NCD to laryngeal tumor cells) 12/49patients

5-35% BC in various tests with no differences between BC and NCD

Skin test with T antigen showed higher sensitivity and specificity than LMI LMI of cancer patients was not tissue specific and a p peared to be directed toward a “group” of antigens shared by several tumors. Is it a “common epithelial antigen”?

No correlation was found between lymphocyte proliferation and LMI. Only postoperative tests were carried out (7-21 days after surgery) Blind preoperative tests. Patients with hyperplastic BBD had higher reactivity to gp52 than those without hyperplasia. MuMTV-induced LMI did not appear to be valuable

C0.93) (continued)

TABLE I (Continued) ~

Reference Bemhard et al. (1983)

Leukocyte donorso

171 BC, 16 BBD, 66 NCD

Antigenb Saline extract of MCF-7 cells (20 w g W , MuMTV

Techniquec D,CT

Positive reaction (MI)d

Ratio of positive reactions

Lowest 10th percentile established in NCD

17-23% BC to

<0.65

9/30 BC

the various antigens

(2.5-20

pg/ml), T antigen (5-20

wLgW

Lidereau d al. (1983)

30 BC, 9 BBD, 30 NCD

fuMTV (5-25 -50 p.g/ml)

D. D

Note

T antigen-induced LMI mainly in patients with advanced disease. Patients with other cancers did not react with T antigen. MCF-7 induced response was low after surgery and returned to preoperative level at 6 months. Tests with various antigens suggested the presence of cross-reacting antigen(s) N o significant LMI response in controls and BBD

0 BC, Breast cancer;BBD, benign breast disease; NCD, normal control donors; MM, malignant melanoma; CC, colon cancer; RC, rectal cancer; CRC, colorectal cancer; LC, lung cancer; GC, gastric cancer. b Where available, protein concentrations are indicated. CEA, Carcinoembryonic antigen; MuMTV, murine mammary tumor virus; MuLV, rnurine leukemia virus; MSV, murine sarcoma virus. c D, Direct; I, indirect; CT,capillary tube technique; AP, agarose plate technique; AD, agarose microdroplet technique; GP, guinea pig; M+, macrophages. MI, Migration index (see Section 111). ng, Not given.

MIGRATION INHIBITION IN TUMOR IMMUNOLOGY

267

by Black et al. (197413, 1976) that around 30% of BC patients react in LMI assays using mouse milk from strain that is a highly susceptible to murine mammary tumor virus (MMTV), as well as to a gp52/55 protein of MMTV, but not in LMI assays using milk from a strain with low susceptibility to MMTV. Although these data have been confirmed by others (McCoy et al., 1978; Wanebo et al., 1979; Fukuda et al., 1980; Cannon et al., 1982, Bernhard et al., 1983) the significance of these reactions also remains unclear, since some of the BC patients also reacted in LMI studies using the gp69/71 protein of Rauscher murine leukemia virus and/or Kirsten murine sarcoma virus, which do not cross-react serologically with MMTV gp52 (McCoy et al., 1978). Furthermore, the reaction appeared to depend on the hyperplastic rather than the malignant nature of the patient’s tumor since those with hyperplastic BBD showed higher reactivity to gp52 than those without hyperplasia (Cannon et al., 1982). Thus, the possible etiological relationship remains to be determined in spite of correlative findings with other techniques (Mesa-Tejada et a l . , 1978; Lloyd et al., 1983). This subject is reviewed by Moore et al. (1983). Recently, LMI reactions to Thomsen-Friedenreich (T) antigen (a precursor of the major antigens of the MN human blood group system) have been observed in BC patients and less frequently in BBD patients, whereas normal control women did not show such reactivity (Springer et al., 1976, 1980; Bernhard et al., 1983). The nature and characterization of antigen(s) participating in the induction of migration inhibition responses in BC patients are not yet fully understood. The difference between BC patients and normal donors in their degree of LMI reaction to cancer tissue extracts reported in the earliest studies stimulated investigators to develop a technique which would be useful in the clinical diagnosis of BC. The ideal diagnostic method should be specific, i.e., reactivity against the tumor extract(s) by normal donors and patients with other cancers or benign diseases should be minimal or negative. In addition, at least a considerable (major) proportion of BC patients must show the reaction. In BC-LMI studies, however, the overall percentage of responder patients was reported to be between 30 and 50%, as mentioned above, and the highest responder BC patient population was 84.4% (McCoy et al., 1976a) using the 3 M KCI extract of MCF-7 cell line as antigen. Further studies by the same investigators did not confirm these data when performed on peroperative patients (the optimal period for diagnosis with this noninvasive technique) (Cannon et al., 1982). The ratio of responders in the BC population could be increased up to 75% with simultaneous use of antigens (Wanebo et al., 1979; Fukuda et al., 1980). The incidence of “false positive” reactions in normal donors and in patients with benign lesions of the breast could not be eliminated (Cannon et

268

ROBERT SZIGETI

al., 1978b, 1982). Thus, LMI studies cannot distinguish malignant and benign pathological changes of the breast in a reliable way. Another possible application of this technique was to attempt to correlate the results of migration inhibition with clinical stage and to assess the prognosis of BC patients; however such studies gave conflicting results. Despite the early study of Cochran et al. (1972a) reporting decreased LMI reactions in BC patients when tests were performed postoperatively vs preoperatively, most subsequent studies were performed after surgery. Furthermore, reactivity may also be influenced by the therapy and the general immunocompetence (see Section V,A) of the patient. These parameters are frequently not indicated in the studies. Several workers reported that with advancing (metastasizing) disease there was a reduction in migration inhibition reactivity (Cochran et al., 1973a, 1974; Black et al., 1974a; Akiyoshi et al., 1975; Mantovani et al., 1978). Others did not observe this correlation (McCoy et al., 1974; Rieche et al., 1976). Jones (1976) and Jones et al. (1976) reported that LMI actually increased in BC patients with metastases. Similarly, Bernhard et al. (1983) found a higher incidence of LMI response to T antigen in patients with advanced disease. Some groups did not find any correlation between LMI and the clinical stage or prognosis (Segall et al., 1972; Wanebo et al., 1979). Since the reaction is determined by the immunocompetence of the patient, by many clinical variables, and even by the antigenicity of the tumor extract used in the test (Black et al., 1976; Cannon et al., 1978b), the relationship between migration inhibition and the clinical stage, tumor burden, and prognosis of the patient is still not clear. These data show that migration inhibition studies in BC are still of research interest but have not yet reached clinical diagnostic significance.

2 . Malignant Melanoma ( M M ) It has been known for approximately 15 years that MM is immunogenic in tumor-bearing patients (Morton et al., 1968)and that it induces both humora1 and cell-mediated immune responses. The latter have been demonstrated by lymphocyte cytotoxicity, blastogenesis (Cochran, 1978), and LMI studies (Cochran et al., 1972b). During the last decade, LMI studies in MM (Table 11)demonstrated that not only leukocytes of MM patients, but also those of others with benign melanocytic lesions are more frequently reactive than those of normal control donors (Cochran, 1978). In an early study MacKie et al. (1972) described complete reactivity in relatively few patients in preoperative LMI tests. Later on, however, various studies reported that 42-80% of MM patients recognized antigens of autologous and allogeneic MM tumors presented as extracts or formalin-

MIGRATION INHIBITION IN TUMOR IMMUNOLOGY

269

fixed cells with migration inhibition (Table 11). Responses of normal individuals and patients with benign nevus were relatively rare. MM patients’ leukocytes, however, were also reactive in LMI to nevus extracts and showed response more frequently when exposed to extracts derived from normal skin or that containing atypical melanocytes (Wilson et al., 1979). Furthermore, first trimester fetal extracts also selectively inhibited leukocyte migration of MM patients, suggesting the existence of fetal antigen(s) that cross-react with those obtainable from melanocytic tumors (Cochran et al., 1982). Reaction of M M patients to extracts from tumors of other origins as well as response of patients with other tumors to melanoma extracts were rarely observed (Cochran et al., 1972a, 1973a; Falk et al., 1973; Ross et al., 1973; Boddie et al., 1975a; Herberman et al., 1975; McCoy et al., 1975; Cannon et al., 1978b). The incidence of such cross-reactivity was always lower than that of reactivity of MM patients to M M extracts, but was higher than that found in normal control donors (NCDs). As is the case with BC (see Section V,B,l), the nature of the responsible antigen(s) in MM remains unclear so far, although a soluble MM-associated antigen has been isolated and it has been tested with other techniques, such as complement fixation (Grimm et al., 1976) and elicitation of delayed type hypersensitivity (DTH) skin reaction (Hollinshead et al., 1974b; Herberman et al., 1975; Holmes et al., 1975; Roth et d.,1976). Interestingly, these partially purified proteins have not been applied in the LMI assay. Since the LMI assay does not distinguish benign nevus and MM, it probably would not be useful for differential diagnosis. The assay, however, might serve a role in monitoring the course of the disease. Conflicting data have been reported on the correlation between LMI and the clinical status of MM patients. Cochran et al. (1972b, 1973b, 1975) found that LMI occurred most often in patients who had a localized tumor, whereas patients with disseminated disease responded less frequently. A well-controlled study of Morrison et al. (1979)from the same group reported that 65% of patients in Stage I1 showed LMI whereas only 20% of those in Stage I (without detectable tumor) showed this response. Others found no correlation between LMI and the stage of the disease (McCoy et al., 1975; Herberman et al., 1975; Boddie et al., 1975a, 1976; Cochran et al., 1976b). Earlier applications of the assay for monitoring therapy for MM (including the addition of BCG or transfer factor) revealed contradictory experiences (Spitler et aZ., 1973; Falk et al., 1973; Cochran et al., 1976b). Recent data on the diagnostic value of LMI, and successful monitoring of the clinical course in ocular melanoma using this technique (Manor et al., 1978; Delpre et al., 1983) are promising, however. Because of the small number of such investigations, this issue needs further analysis.

TABLE I1 TUMOR-SPECIFIC LMI STUDIESIN MALIGNANT MELANOMA"

Reference MacKie et al. (1972)

Leukocyte donors

12 MM

Cochran et al. (197%) Cochran et al. (1972b) 22 MM, 12 skin cancer, 16

Antigen

Technique

Supernatant of MM homogenates

D,CT

Fluid from a cystic melanoma

Positive reaction (MI)

Ratio of positive reactions

nge (at 5% level with U test)

10/10MM pre-

D,CT

ng (at 5% level with U test)

11/22 MM, 3/5 skin cancer, 2/16 NCD

Fluid from a cystic melanoma

D,CT

ng (at 5% level with U test)

10/16with local MM, 116 with

Saline extract of MM tissue

I,AP

operatively, 11/12MM postopera tively

NCD Cochran et 01. (1973b) 22 MM, 5 skin cancer, 16 NCD

Spitler (1973);Spitler

et al. (1973)

51 MM, 18 NCD

ng

metastatic MM, 3/5skin cancer, 2/16 NCD 6/11disseminated MM, 3/17primary MM, 3/14localized-spread MM, 1/18 NCD

Note In the immediate postoperative period, LMI was lost and returned between 4 and 22 postoperative days See Table I LMI occurred most often in patients who had no evidence of recurrence or whose tumor did not extend beyond the primary site Fractionation of the cystic fluid on Sephadex G-200 column resulted in an active fraction with l00K MW

No LMI to MM tissue extract was observed in healthy black donors. Response occurred late in the disease course. No serum-blocking effect was found. In one MM treated with transfer factor, a good relationship was observed with the clinical course

Cochran et al. (1973a) Ross et a!. (1973)

N

; I

See Table I

43/66 tests to histologically similar, 9/39 to dissimilar tumor cells; 24/104 tests on NCD 24/31 MM, 25/25 gastrointestinal cancer (serum-free tests)

27 MM, 39 BC, NCD

Formalin-fixed tumor cells

D,CT

ng

Falk et 01. (1973)

4 5 MM and gastrointestinal cancer, ? NCD

Tumor membranes (preparation is not given in detail)

D,CT

<0.W

Tautz et al. (1974)

2 MM, ? NCD (both black and white do-

MM tumor cell suspension

D,AP

<0.80

819 healthy African blacks, 112 MM, 1/8 white NCD

Char et al. (1975)

11 ocular

3 M KCI extract of MM tissue

D,CT

c0.80

10/11 choroidal melanoma, 1/11 other malignancies, 3/21 NCD

LMI was induced only when tumor membranes derived from histologically similar tumors. Autologous serum abrogated LMI, allogeneic sera decreased the response. BCG therapy did not influence the reaction

nors)

melanoma, 11 other (systemic) malignancies, 21 NCD

The antigenicity of the three tumor extracts applied was different

(continued)

TABLE I1 (Continued)

Reference McCoy et al. (1975)

Leukocyte donors

64 MM, 43 NCD,

Antigen

3 M KCl extract

Technique D,CT

Positive reaction (MI)

C0.80

of MM tissue

? other cancers

tQ l w -

Ross et al. (1975)

Cochran et al. (1975)

69 MM, 81 BC, 61 NCD and other cancers 81 MM, 8 NCD, 29 other malignant, 28 nonmalignant disease

Formalin-fixed tumor cells

D,CT

<0.76-<0.83

Formalin-fixed MM cells, supernatants of tumor homogenates

D,CT

(depending on tumors, at 5% level with U test) ng

Ratio of positive reactions

571134 tests in MM, 3/50 NCD, 4/27 other cancers to MM extract, 2/26 MM to extracts from other tumors 38/69 MM in autologous, 6/10 in allogeneic tests, 55/81 BC, 5% NCD 43/81 MM, 10/65 other disease

Note Benign nevus tissue did not induce LMI. Stage I, patients reacted slightly more frequently, otherwise all stages had the same incidence of response. Decreased reac' tivity was seen 0-14 days after surgery Mouse melanoma cells infrequently caused LMI

LMI was most frequent in patients with early, localized tumor and lost for 5-12 days after surgery. Some correlation was observed between positive LMI and favorable prognosis. Disappearance of LMI preceded the development of metastases within 6-8 weeks

Boddie et al. (1975a)

36 MM, 23 other cancers, 28 NCD

3 M KCI extract of M14 MM cell line

D,AP


Herberman et al. (1975)

61 MM, 40 NCD, 170 other cancers

3 M KCl extract of MM tissue

D.CT

C0.80

Formalin-hed MM cells

D,CT

ng (calculated with U test)

19/36MM, 4/23 other cancers, 4/28NCD to MM extract (7132MM to non-MM extracts) 45/96 tests on MM, 3/36 NCD (1115of black donors), 3/17other cancers

m

a

Cochran et ol. (1976b) 89 MM, 13 BC, 2 osteosarcoma, 43 NCD

Normal muscle tissue extract did not induce LMI. Patients in stages I and 111 reacted with roughly equal frequency

Skin tests on MM patients with the same extract revealed 29/43positive reactions and appeared to be more sensitive. LMI decreased between 11 and 30 days after surgery. No correlation was found with the clinical status

Sera from MM (38/89),other cancers (4/15), NCD (3/43) caused LMI in autologous system; sera from MM (31171). NCD (3131) increased MM cells-induced LMI (continued)

TABLE I1 (Continued)

Reference Boddie et al. (1976)

t9

2

Char(1977)

Spitler et al. (1977)

Leukocyte donors

Antigen

Technique

Positive reaction (MI)

36 MM, 36 LC, 27 sarcoma, 26 other cancers, 50 NCD

3 M KCI extract of MM tissue and cell lines (181-454 dml)

D,AP

12 ocular MM, 12 other ocular tumors or ocular metastases 86 MM, 13 household contacts of MM, 100 NCD

3 M KCI extract from a metastatic choroidal melanoma

D,CT

<0.82

Saline extract of MM tissue

1,AP (with GP M+s) D,AP

c0.80

<0.70 (calculated with t test)

Ratio of positive reactions

Note

19/36 MM, 23/36 LC, 18/27 sarcoma to histologically similar tumor extracts 4/28 NCD to MM extracts 8/12 MM, 0/12 other diseases

No correlation was found with clinical stage

12/31 MM, 8/11 household contacts, 3/13 NCD in 1,AP 10/50 MM, 5/7 contacts, 20% of NCD in D.AP

The frequency of LMI responses in MM was intermediate between contacts and NCD. Poor correlation was found with clinical stage. Skin tests demonstrated normal cellmediated immunity in MM

Cannon et al. (1978b) Manor et a!. (1978)

Wilson et al. (1979)

Morrison et al. (1979)

8 choroidal MM, 10 choroidal nevus, 5 other, systemic malignancy, 10 NCD 57 MM, 52 benign nevus, 98 NCD

72 MM, 36 NCD, 29 nonmalignant disease, 10 other cancer

Saline extract of melanoma from enucleated eyes

I (with GP

Supernatants of tissue homogenates

D,CT

At 5% level with U test

35/57 MM, 9/24 nevus, 10/64 NCD to MM extract; 16/50 MM, 5/52 nevus, 14/98 NCD to nevus extracts

Formalin-fixed MM cells

I,CT

At 5% level with U test

65% MM in St.11, 58% St.1, when MM was present, 29% St.111, 20% St.1 without detectable MM

<0.80

7/8 ocular melanoma, 0/5 other cancers, 0/10 nevus

Md?4

See Table I Four patients who were tested before enucleation had already positive MIF test. There was no response to iris extract from the same eyes

MM patients more frequently showed LMI with nevus extracts than nevus patients and NCD. Extracts of skin containing atypical melanocytes selectively caused LMI in MM patients. The same was found with normal skin extracts

(continued)

TABLE I1 (Continued)

Reference Cochran et d.(1982)

Delpre d ~ l (1983) .

Leukocyte donors

Antigen

Technique

97 MM, 194 pregnant women, 123 NCD

Supernatants of homogenates or formalinfixed MM cells, saline supernatants from whole fetus or fetal tissues

D,CT

63-year-old man with choroidal melanoma

Epitheloid and spindle-type extract of choroidal melanoma (not detailed)

I (with GP M+) (not detailed)

See Table I footnotes a, c , d, e .

Positive reaction (MI) At 5% level with U test

<0.80

Ratio of positive

reactions 56197 MM, 39/194 pregnants, 231123 NCD to MM extracts; 511169 MM, 31/197 NCD, 271142 pregnants to 1st trimester fetal preparations

Note 1 s t trimester fetal prepara-

tions selectively caused LMI in MM. 2nd trimester preparations showed no selective LMI activity. No evidence was found that pregnant woman were immunized to fetal or MM antigens

Initially, when only the ocular result was found, MMI was positive and remained so for 2 years, even when hepatic difFusion developed. Six months later the test became negative and the patient’s condition went to an abrupt decline

MIGRATION INHIBITION IN TUMOR IMMUNOLOGY

277

3. Gastric and Colorectal Cancer (GC, C R C ) Migration inhibition studies in GC and CRC (Table 111) generally reveal greater sensitivity and specificity compared to those found in BC and MM. Lurie et al. (1975) reported 100% positives with this assay in 12 CRC patients, while none of 17 NCD reacted. Sirianni et al. (1981) also observed 100% positive reactions in 10 patients. In larger patient samples, however, the ordinary ratio of pathological LMI reactions varied between 30 and 90% (Table 111). Usually, migration inhibition assays were performed with allogeneic tumor extracts, although one group of investigators did not find sensitization only in the autologous system (Elias and Elias, 197513). In one study an extract of CRC cell line HT 29 gave more positive reactions than tumor extracts (Burtin et al., 1977). The use of an antigenic (extract) panel instead of a single tumor extract together with the evaluation that a 3/5 positive MI was considered as a pathological reaction (Zoller et al., 1977a-c) significantly improved both the sensitivity and the specificity of the assay. Moreover, the combination of LMI with other immunological techniques (Lurie et al., 1975; Emmrich et al., 1978; Matzku et al., 1978; Zoller et al., 1983) proved to be even more useful, even in clinical practice. Response of NCDs to GC or CRC tissues was found to be infrequent in these sutdies. GC and CRC patients, however, frequently showed LMI (in most studies) when their leukocytes were exposed to extracts from normal rnucosa obtained from the stomach or colon (Gouillou and Giles, 1973; Gouillou et al., 1975; Zoller et al., 197713; Emmrich et al., 1978). Recent studies reported significantly lower figures on positive reactions of GC and CRC patients to extracts from normal mucosa (Iwaki et al., 1979; Sirianni et al., 1981; Akiyoshi et al., 1982). When cross-reactivity was tested, 25% or more positive reactions were found between GC and CRC (Zoller et al., 1977b; Kojima et al., 1981a,b), whereas cross-reactions with extracts from other cancers were rarely observed (Burtin et al., 1977; Zoller et al., 1977b). Among noncancer patients with diseases of the stomach, those with atrophic gastritis and intestinalization (both known to be precancer states) were found to respond most often to GC extract with LMI (Zoller et al., 1977a). Patients with Crohn’s disease or ulcerative colitis did not appear to show as high reactivity to CRC extracts as CRC patients (Matzku et al., 1978; Zoller et al., 1977b), although patients with both diseases were reported to show LMI with colon antigens (Bendixen, 1969; Richens et al., 1974; Eckhardt et al., 1976). The data do not indicate whether the response detected with migration inhibition is directed to tumor-specific or organ-specific antigen(s). Findings on the role of fetal antigens are contradictory: the Heidelberg group found

TABLE 111 TUMOR-SPECIFIC LMI STUD~ES I N GASTRIC AND COLORECTAL CANCER“ _

Reference Bull et al. (1973)

GuiUou and Giles (1973)

Rocklin and Churchill (1973) McIllmurray et a/. (1974)

Leukocyte donors

_

_

~

Antigen

27 CRC, 52 cancerfree patients

Tumor homogenates

22 CRC, 12 NCD

Perchloric acid extract of CRC and normal intestinal rnumsa (50 &ml)

25 CRC, 25 nonmalignant colon disease

Saline homogenate of CRC and normal mucosa (1-2 mg/ml)

~

Technique

D,CT (mixed mononuclear cell population) D,CT

Positive reaction (MI)

Note

<0.80

24/27 CRC, 0152 other patients

The observed data did not reveal, whether the reaction had specificity for CRC or for a “common tumor antigen”

<0.79-<0.83 (calculated

15/22 CRC, 15/15 NCD to CRC extract, 8/22 CRC to normal mucosa

Only six CRC patients showed LMI when tests were performed in autologous sera. “Normal” colonic mucosa was obtained from the same CRC patients at a site at least 10 cm away from the tumor See Table I

ng

A higher degree of migration inhibition was observed with leukocytes obtained from CRC patients. LMI was more marked in response to CRC extracts than to normal colonic rnucosa in the same patients

with mean22 SD

of NCD)

D,CT

Ratio of positive reactions

nge (calculated with t test)

Akiyoshi et al. (1975)

80 various cancer patients (50 GC, 6 CRC) 14 CRC, 10 benign gastrointestinal disease

Saline extract of tumor tissues (0.5-1.28 mglml)

1,MMI (GP

Perchloric acid extract of CRC and normal mucosa (50 )IgW

1,CT; D,CT

<0.68-<0.83 (calculated with mean 22SDof the controls)

Armitstead and Gowland (1975)

9 CRC, 6 NCD

Perchloric acid extract of CRC tissue (10 &ml)

D,CT


Straus et al. (1975)

4 CRC, 4 pancreatic cancer, 14 nonmalignant colon disease, 5 NCD

CEA (50 and 250 pg/ml)

D,CT

(0.80

Guillou et al. (1975)

<0.75

M W

25% of total group (10/50 GC, 216 CRC) to autologous extract 18/25 lymph node cells, 11/14 PBL to CRC extracts, 6/25 lymph node cells to normal mucosa

5/9 CRC with bufFy coat, 717 CRC with mononuclears, 019 CRC with granulocytes Only 1 patient with pancreatic cancer

See Table I

Preoperative tests. No significant differences were found among peripheral blood lymphocytes, draining and other lymph node cells in the reaction. Sensitization of local nodes occurred early in the disease and might be lost when the lymph nodes became involved with metastasis. Reactions in autologous serum were less frequent Preoperative tests. Mononuclear cells showed increased sensitivity to CRC antigen. Autologous serum interfered directly with granulocyte migration

(continued)

TABLE I11 (Continued)

Reference House et al. (1975)

Leukocyte donors

31 CRC, 81 NCD

~

Positive reaction (MI)

Ratio of positive reactions

Antigen

Technique

Saline extract of CRC tissue and normal mucosa

D,CT

ng (calculated with t test)

Not detailed

Lurie et al. (1975)

16 CRC, 3 benign colon disease, ? NCD

Tumor homogenates

D,CT

<0.80

12/12 CRC prior to surgery, 0/16 CRC after surgery (2/44 month), 0117 NCD

Elias and Elias (1975a)

35 patients with varous cancers (27 CRCand GC) ng

Autologous tumor cells and/or tumor extracts (100 pg/ml)

D,CT

<0.80

18/20 with localized tumor, 11/15 with metastasis

Autologous and homologous CRC extracts and cells

D,CT

ng (calculated with t test)

ng

8

Elias and Elias (1975b)

Note

LMI in patients in the 1st2nd week after surgery did not differ from those of NCD, but later they showed significant reactivity to the tumor antigen. Patients with localized and metastatic disease reacted equally Simultaneously, skin tests to recall antigens and serum CEA-level determination were performed. The postoperative negativity of LMI occurred both in potentially cured and noncured patients Fresh, frozen-thawed (dead) tumor cells and tumor cell extracts gave the same reaction as antigens. Preoperative and follow-up studies were performed Leukocytes of patients were sensitized to autologous hut not to homologous tumor cells and extracts

Ehas et al. (1977)

Burtin et al. (1977)

62 CRC

68 CRC, 40 other cancers, 31 NCD

Tumor cell homogenates (100 &ml)

Saline extract from CRC homogenate and HT-29 cultured cells (20-600

<0.80 (calcu-

lated with multiple comparison statistics)

D,CT

63% in St.B,

82% in St.C, 30% in St.D in Dukes classification (initial reactions)

<0.80 (and 10th percentile of control values)

30168 CRC, 3/31

co.80 (calcu-

39/43 GC, 1/32 NCD, 8/76 benign gastric disease

NCD, 8/40 other cancers to HT-29 extract

wW)

N a, r

Zijller et al. (1977a)

43 GC, 135 other cancers, 94 benign diseases, 32 NCD

A panel of 3 M KCl extracts from GC tissue (0.52.5 pglml)

D,CT

lated with mean22 SD of NCD values)

Patients with localized or regional disease were initially more sensitized to their tumors when compared to those with metastasis. The reactivity, however, has been lost in the progressive course of the disease HT-29 extract more frequently induced LMI than tumor extracts. Among patients with other cancers, those with gynecological cancers were rarely positive, whereas those with cancer of the respiratoryldigestive system showed LMI in 6/15 cases. The antigen(s) involved showed a broad tissue specificity Considerable sensitization was found in patients with atrophic gastritis and intestinalization. Using a panel of extracts both sensitivity and specificity was improved. Positive reactions persisted for 3 months after surgery (continued)

TABLE 111 (Continued)

Reference Ziiller et al. (197%)

Ziiller et al. (1977~)

N a ! N

Imai et al. (1977)

Emmrich et al. (1978)

Leukocyte donors 59 CRC, 27 benign colorectal disease, 37 NCD 74 GC, benign gastric diseases, NCD (numbers are "8)

20 GC, 12 other cancers, 15 NCD 3 GC, 3 CRC, 6 NCD

Positive reaction (MI)

Ratio of positive reactions

D,CT

<0.80 ( C ~ C U lated with mean22 SD of NCD val-

55/59 CRC, 2/27 benign diseases, 0137 NCD

D,CT

<0.79 (based on mean52 SD of NCD values)

32/35 GC preoperatively, 24/43 atrophic gastritis

3 M KCI extract of GC tissues (500 CLgW

D,AP

<0.83 (based on mean22

CEA, 3 M KCI extracts from tumors

D,AP

4/10 GC in autologous, 10126 GC in allogeneic tests 6/15 cancers to CEA, 11/15 to histologically similar, 9/15 to dissimilar type

Antigen A panel of 3 M

KCI extracts from CRC tissue (0.52.5 pglml) A panel of 3 M KCl extract from GC tissue and normal gastric mucosa

Technique

ues)

SD) <0.79

Note 13% CRC reacted with normal colonic mucosa, obtained from a megacolon. Reactivity persisted for 2 months after surgery 2 months after surgery most patients' LMI reaction was in the normal range and reappeared in patients with local recurrence or metastasis. Patients with benign disease when showed positive LMI with GC extracts, also gave LMI with normal gastric mucosa extract No relation was found with the clinical stages

Matzku et al. (1978)

Zaller et al. (1979)

N

179 CRC, 51 benign colon disease, 101 other cancer, 40 NCD 24 GC, 43 CRC, 54 other cancers, 19 benign diseases, 10 NCD

3 M KCI extract of CRC tissue

D,CT

(0.80

158/179 CRC, 8/51 benign diseases, 0140 NCD, 5/101 other cancer

3 M KCI extract from tumors and fetal tissues

D,CT

C0.80 and at

82% GC, 68% CRC, 50% other cancer, 2% NCD to tumor extracts; 59% GC and CRC, 5% NCD to fetal extracts

22 GC, 14 CRC, 6 LC, 28 NCD, 21 benign diseases

3 M KCl extract of tumor tissues (22 CLg/mc

D,AP

5% level with U test

w x-

Matzku et al. (1979) Iwaki et al. (1979) ’


12/40 GC, 7/14 lung cancer, 10/21 CRC, 2/68 NCD, 0/67 benign disease to tumor extracts; 2/75 cancer patients to normal gastric mucosa (test numbers)

LMI was independent of the clinical stage. Simultaneous CEA level determinations resulted in a CEA-, LMI+ pattern in primary tumors and CEA+,LMI+ with advanced disease Gestation age of the fetus did not influence the antigenicity of the extracts applied in LMI. Multiparous donors bearing tumors might respond more strongly or frequently

See Table I The various tumor extracts differed in antigenicity. Leukocytes from patients with scirrhous and poorly differentiated tumors were more strongly sensitized than those from patients with medullary or well-differentiated cancers

(continued)

TABLE 111 (Continued) Leukocyte donors

Antigen

Matzku et al. (1980)

CRC and lung cancers (not detailed)

3 M KCI extract of tumor, normal and fetal tissues

D,CT

<0.80

11-82% to vanous extracts

Burtin and Chavanel (1980)

42 CRC, 40 NCD, 17 benign disease

Saline extract of fetal colon

D,CT

Calculated with t test or <0.78 at 10th percentile level of NCD

Significant LMI was not observed at any antigen concentration. Migration enhancement was frequently found

10 CRC, 32 other cancer, 10 NCD, 16 benign colon disease

3 M KCI extract of CRC and normal colon tissue

D,CT

c0.98 (arbitrarily)

10110 CRC, 1/10 NCD, 6/32 other cancer, 4/14 BC

Reference

Akiyoshi et al. (1981) Sirianni et d.(1981)

Technique

(50 P.8-2

mg/ml)

Positive reaction (MI)

Ratio of positive reactions

Note CEA content of tumor extracts did not influence LMI reaction of the patients. Furthermore,CEA plasma level did not affect LMI reactivity Leukocytes were only weakly or not at all sensitized to fetal colon antigen, which by contrast provoked stronger reactions in patients with nonmalignant intestinal diseases See Table I No reaction was observed to normal colon tissue

Kojima et al. (1981a)

79 GC, 20 CRC, 20 benign gastrointestinal disease, 10 NCD

3 M KCI extract of GC and CRC tissues (250 P g w

D,AP

(0.77 (based on mean+2 SD of NCD values)

Kojima et al. (198lb)

50 CRC, 15

3 M KCI extract of GC and CRC tissues (250iJ,g/ml)

D,AP

<0.82 (based on mean22 SD of NCD values)

GC, 10 NCD, 15 benign gastrointestinal disease

Mantovani et al. (1981) Akiyoshi et al. (1981)

GC: 48%to single, 79% to a panel of GC extracts; 25% of benign disease and CRC to GC tissue extracts 2/30 CRC to normal mucosa, 48% CRC, 16% GC, 9% benign, 2% NCD to single, 80% CRC, 13% GC, to a panel of CRC extracts, %% CRC to GC extracts

GC extracts had a wide range of cross-reactivity when compared to CRC extracts

Marked depression of LMI was observed from 10 to 14 days after curative resection. In patients without recurrence, LMI was found to remain in the normal range during 2-4 weeks after surgery. Positive reaction reappeared in patients with local recurrence

See Table I 86 GC, 31 other cancer, 42 NCD

3 M KCI extract of GC and CRC tissues (250 pglml)

I,MMI,CT

<0.75 (based on mean22 SD of NCD values)

12/52 GC to autologous, 4/47 GC to allogeneic GC tissue, 2/35 GC to allogeneic normal mucosa

Postoperative tests. Indirect MMI appeared to show low sensitivity and high specificity

(continued)

TABLE 111 (Continued)

Reference

Leukocyte donors

Antigen

Technique

Positive reaction (MI)

Ratio of positive reactions

Note

_ _

=Her el a!. (1983)

Elias et d.(1983)

a

A total of 1120 CRC, among them 320 CRC were tested three times or more

A panel of 3 M

46 CRC

Autologous CRC extracts

See Table I footnotes a, c, d, e.

D,CT

KC1 extracts of CRC tissue

ng

Calculated with U test

ng

ng

Simul~neousdeterminations of CEA plasma level were performed. Patients with persistent absence of the disease showed either persistent negativity in the tests (50%) or conversion in one of the two parameters. Persistent positive profile was rarely found. Patients with recurrence or metastases showed either positive profiles or conversion to positivity and only 1%was negative in both tests. Because of its technical complexity, LMI has not been recommended for postsurgical monitoring Various LMI patterns were found during different immunotherapeutical procedures

MIGRATION INHIBITION IN TUMOR IMMUNOLOGY

287

that more than half of GC and CRC patients were sensitized to fetal antigenic determinants (Zoller et al., 1979; Matzku et al., 1980), whereas Burtin and Chavanel (1980) claimed that leukocytes from CRC patients were only weakly or not at all sensitized to fetal colon antigen, which, by contrast, elicited a stronger reaction in patients with nonmalignant intestinal disease. It seems to be obvious, however, that the carcinoembryonic antigen (CEA) by itself either cannot induce LMI in CRC (Churchill and Rocklin, 1973; Straus et al., 1975; Matzku et al., 1980) or does it very rarely (Emmrich et al., 1978). Furthermore, the antigenicity of the tumor extracts is not influenced by their CEA content and the plasma level of CEA does not affect LMI reactivity of the patients (Matzku et al., 1978, 1980): actually, the two patterns are rather complementary, providing a good tool for monitoring the patients (Lurie et al., 1975; Zoller et al., 1983). The correlation between LMI and clinical status revealed that patients with localized tumors exhibited greater reactivity in LMI than patients with metastasis and, in follow-up studies, their initial reactivity was lost during the progressive course of the disease (Elias et al., 1977); on the contrary, more often monitoring of the patients revealed another pattern: after surgery a marked depression of the initial reactivity was seen (Lurie et al., 1975; Zoller et al., 1977~).Subsequently, in patients without recurrence, the reactivity remained in the normal range and in those developing recurrence or metastasis a positive reaction reappeared (Zoller et al., 1977c; Kojima et al., 198lb). In a very recent, well-controlled postsurgical follow-up study, Zoller et al. (1983) reported that the simultaneous use of LMI and plasma CEAlevel determinations provided clinically useful, disease-related patterns: patients with absence of the disease after surgery showed persistent negativity or conversion of one or both parameters, whereas those with recurrences or metastases revealed either a positive profile or conversion to positivity. Thus, it appears that in GC and CRC the LMI assay, especially when combined with other tests, is a useful tool in immunological diagnosis and monitoring. 4 . Lung Cancer (LC)

A high frequency (60-100%) of response in migration inhibition assays has been observed with leukocytes from LC patients using various techniques (Boddie et al., 1974, 197513, 1976; McCoy et al., 1977a; Zoller et al., 1980; Suslov et al., 1981). The relatively lower incidence of LMI reactions in the study of Vose et al. (1977) may be due to the unusually short cultivation period (2-3 hr) performed in these experiments. The reactivity reported in most studies was not influenced by the histological type of the tumors of LC patients whose leukocytes were tested or

288

KOBEKT SZIGETI

those used as antigen source (McCoy et al., 1977a; Zoller et al., 1980; Suslov et al., 1981). The specificity of the assay in LC appears to be high, since reactivity of NCDs, patients with nonpulmonary cancers, or patients with nonmalignant diseases of the lung was infrequently found. Some weak cross-reactivity was observed with LC antigens and patients with BC or Ewing’s sarcoma (McCoy et al., 1977a). On the contrary, 72% of BC patients in a recent study of Mantovani et al. (1981) reacted to formalin-fixed LC cells, suggesting the existence of a “common epithelial antigen” (see Section V,B, 1).Reactivity of LC patients with tumor extracts from other organs rarely occurred (Zoller et al., 1980). Since neoplasms of nonpulmonary origin frequently give metastases in the lung, a crucial question is raised related to the reaction of such patients. Zoller et al. (1980) found that patients with pulmonary metastases from various nonpulmonary tumors frequently showed reactivity to 3 M KC1 extracts from LC. These patients, however, gave a similar incidence of LMI response to normal lung tissue extracts. An intriguing application of the assay showed that NCDs in contact with LC patients or materials responded as frequently in indirect LMI studies as LC patients when their leukocytes were exposed to 3 M KCI extracts of a LC pleural effusion (7661), whereas noncontact NCDs were nonresponders. No correlation was found in this study between smoking history and LMI reactivity (Suslov et al., 1980).These data emphasize the importance of proper selection of control donors in future studies, even in those tumor systems where such epidemiological studies have not yet been performed. Another unique application of the LMI assay has been done by Cerni (1977), who found that 55%of LC patients showed a specific LMI reaction to formalinized E-14 cells (an established cell line from LC) and three different hybrids which had only a few human chromosomes left over. There was no reactivity to the other parent, hamster fibroblast cells. Generally, there was no correlation (when tested) between LMI reactivity and the clinical stage of the disease (McCoy et al., 1977a; Suslov et al., 1981). In a large-scale immunological monitoring of LC patients, Oldham et al. (1976) were unable to find major changes in immunological parameters during chemo- and/or immunotherapy. Earlier, Boddie et aZ(1975b) found that patients tested within 3 weeks of surgery might show no LMI, but subsequently develop it 1 month later. Others did not observe any decrease in the reaction due to possible immunodepression from surgery and they found no changes in the 10- to 30-day postoperative period, but 1-12 months later the incidence of reactivity did decrease (Suslov et al., 1981).Nevertheless, data on the application of LMI tests in clinical practice in LC patients are few, conflicting, and not satisfactory for drawing any conclusions.

MIGRATION INHIBITION IN TUMOR IMMUNOLOGY

289

As with other tumor systems, the antigen that is responsible for the migration inhibition in LC patients is completely unknown. In this regard, a preliminary attempt to isolate an LC-associated antigen from immune complexes derived from a pleural effusion of LC has been performed (Lawniczak et al., 1983). The antigenic activity was monitored with direct LMI assay and suggested an antigenic protein of about 150,000 MW.

5. Renal Cancer (RC) LMI studies in hypernephroma are based mainly on the early studies of Kjaer (1974a,b; 1975, 1976a,b), reviewed by Kjaer and Bendixen (1976). Apart from careful attention to some technical details of the capillary tube technique (see Section V,C) and comparing it with the agarose plate method (Kjaer and Sorensen, 1976), he found that patients with hypernephroma responded with LMI when their leukocytes were exposed to autologous as well as allogeneic tumor and fetal kidney tissue. Moreover, he demonstrated that the reactivity was almost exclusively confined to the group of patients without distant metastasis and he found a highly significant correlation between LMI response and survival. Further investigations (Kjaer and Christensen, 1977) showed that the general capacity for developing cell-mediated immune reactions in hypernephroma was generally not impaired, but the tumors had different levels of antigenicity: in allogeneic combinations small, localized tumors usually had a high antigenicity, whereas those that were more dispersed or showed early metastases had a low antigenicity. Other investigators found that preoperative local irradiation of the patients increased the incidence of LMI response (Pappas and Schwarze, 1974) and in uitro irradiation of the hypernephroma resulted in a higher antigenicity of the tumor tissue (Schwarze et al., 1979). Purification of the antigen from renal cell carcinoma has been attempted by Wright et al. (1977) with preparative gradient acrylamide gel electrophoresis. The antigenic activity of the fractions was monitored by direct capillary tube LMI technique. A high-molecular-weight fraction showed activity only in RC patients, whereas a low-molecular-weight fraction was associated with responses in both RC patients and normal controls. Further physicochemical characterization of the putative antigen(s) has not been performed. The same authors demonstrated that 71% of RC patients that had a localized tumor and were clinically free of disease 1year postnephrectomy lost their LMI reactivity. On the contrary, patients with metastases were LMI positive unless they received irradiation or hormone therapy. Thus, it was suggested that LMI correlated with the presence of active disease (Wright et al., 1978). These findings do not appear to support those of Kjaer (1976b). To

290

ROBERT SZIGETI

our knowledge, however, no recent studies have been carried out to settle this issue. 6. Carcinomas of the Cervix and Ovary

Early LMI studies in squamous carcinoma of the cervix demonstrated high sensitivity and specificity with both autologous and allogeneic tumor extracts (Goldstein et al., 1971; Chen et al., i975). No inhibition of leukocyte migration was found for other gynecologic tumors challenged by cervical cancer antigen (Chiang et al., 1976). Either no correlation could be shown between LMI test, clinical status of the patient, and prognosis (Chen et nl., 1975), or the LMI was positively correlated with the clinical stage in invasive cancer, i.e., LMI sensitivity increased during progression (Chiang et al., 1976). On the contrary, later findings of Schmeisser and Kademann (1980) indicated a lower incidence of tumor-directed LMI in advanced disease states. In a well-controlled study of Rivera et al. (1979) there was a significantly greater degree of LMI by cervical cancer antigen with cells of patients with invasive compared to patients with in situ carcinoma. In patients with invasive carcinoma, however, no relationship could be observed between the degree of LMI and stage of the disease. Furthermore, more vigorous reactions were seen in patients under the age of 50. Surgical removal of the cancer tissue might result in disappearance of the initially positive reaction, whereas radiation therapy had no specific effect (Yamamoto, 1980). The antigenicity of the tumors used as antigen source did not appear to be influenced by the histological type (Rivera et al., 1979; Yamamoto, 1980). Partial purification of the putative antigen using a Sephadex G-200 column resulted in two active out of a total of four fractions and increased both the specificity and sensitivity of the test (Rivera et al., 1979). The small number of diagnostic studies indicated that the tumor might have to reach a critical size or level of invasiveness before it could elicit a detectable LMI response (Rivera et al., 1979). Follow-up of a few patients suggested, however, that the test could be applied to monitor the therapeutic effect (Schmeisser and Kademann, 1980; Yamamoto, 1980). LMI studies in ovarian carcinoma also indicated the existence of tumordirected cellular hypersensitivity (Chen et al., 1973; Faiferman et al., 1977). A high incidence of LMI response was detected when patients’ leukocytes were incubated with autologous or allogeneic KCl extracts. Other tumor extracts (saline, deoxycholate, and perchloric acid) were less effective. The nature of the KCl antigen in ovarian carcinoma remains to be clarified, It is distinct from CEA and seems to have a high molecular weight on Sephadex column fractionation (Faiferman et al., 1977). No correlation was observed in these studies between the histological grade or clinical stage of ovarian carcinoma and the degree of LMI.

MIGRATION INHIBITION IN TUMOR IMMUNOLOGY

29 1

7 . Prostatic and Bladder Cancer Immunological reactions in prostate cancer were recently reviewed by Ablin and Bhatti (1981). Application of migration inhibition assays is rare in this tumor (Ablin, 1975; Ablin et al., 1975, 1976, 1978; Evans and Bowen, 1977). In these studies a considerable number of patients were found to be reactive to tissue extracts from prostate cancer when compared with leukocytes from healthy age-matched donors or patients with benign hypertrophy of the prostate. There appeared to be little or no influence of tumor stage on the LMI test, although the number of patients tested in these studies was rather small. In a recent study of Wright et al. (1981), tests in a relatively greater number of patients with prostate cancer revealed higher sensitivity and specificity, but no correlation with the stage of the disease. In spite of the high figures of positivity reported in this study, the authors suggested that the assay in its present form was not capable of detecting prostate cancer patients in the general population and its discriminating function was purely a statistical one. LMI studies in patients with transitional cell carcinoma of the bladder also demonstrated tumor-directed reactions. Lavergne et al. (1979) used the agarose microdroplet assay and showed that the positive reaction with KCl extract of various bladder tumor cell lines strongly correlated with the presence of the tumor. Schelhammer et al. (1979) found that the LMI reactivity in the presence of extract from T-24, an established bladder carcinoma cell line, was the same regardless of the stage of the bladder tumor. On the contrary, Meleady et al. (1980) reported that patients without a clinically evident tumor at the time of testing showed slightly more frequent reactivity than those with evident tumor, while patients with the heaviest tumor burden showed the lowest reactivity. Moreover, these latter investigators found LMI reactions only when a 3 M KCI extract of fresh tumor specimens were used as antigens. No specific reaction was observed with glutaraldehydefixed tumor cells and even the 3 M KCI extract of the T-24 cell line was ineffective. Thus, LMI findings in bladder carcinoma are strongly contradictory at the present.

8. Other Nonlymphphoid Solid Tumors Both George et al. (1975) and Rocklin et al. (1977) reported migration inhibition in response to tumor antigen in a group of patients with familial medullary thyroid carcinoma. While the former authors could not observe such response in family members who did not have cancers, the latter group found that 6 of 12 family members gave an LMI response to thyroid tumor extract. In the study of Amino et al. (1975),a few patients with various forms

292

ROBERT SZICETI

of thyroid cancer showed significant response to tumor extracts but not to homogenates from Grave’s thyroid gland. These findings were extended later on by the same group (Aoki and DeGroot, 1982): they observed 50% positive reactions in metastatic thyroid cancer patients to 3 M KCl extracts from tumors without cross-reactions with noncancer antigens. Fractionation by column chromatography resulted in a fraction of around 45,000 MW which selectively retained the TAA activity. Tumor-specific LMI reactions have also been found in hepatoma (Lee et al., 1977), Ewing’s sarcoma (McCoy et al., 1977b), neuroblastoma (GrahamPole et al., 1976), acoustic neuroma (Rasmussen et al., 1982), oral squamous carcinoma (Airoldi et al., 1982), and, at surprisingly high ratios, in patients with laryngeal carcinoma (Mantovani et al., 1981; Cortesina et al., 1982). Similarly, some patients with oral leukoplakia, a premalignant lesion, showed tissue-specific cellular hypersensitivity measured with LMI (RoedPetersen et al., 1973; Roed-Petersen, 1982).

9. Leukemias and Lymphomas Hilberg et al. (1973) studied the ability of two acute lymphocytic leukemia (ALL) patients and their families to produce MIF in response to leukemic blast and remission cell homogenates. One of the patients and both mothers responded in MMI. We found that a majority of children with ALL in remission showed leukemia-specific LMI with allogeneic ALL blast extracts (Szigeti et al., 1974a). Similarly, a group of patients with Down’s syndrome, known to have increased susceptibility to leukemia, responded with LMI when their leukocytes were exposed to extracts from ALL or acute myelogerous leukemia (AML) blasts (Szigeti et al., 1974b). Leukemia-specific LMI reactions were rarely observed in the family members of ALL patients and contact persons (Szigeti et al., 1976). In the study of Sidi et al. (1981), 8 of 27 patients with chronic neutropenia reacted with MIF production when their lymphocytes encountered AML blasts. The positivity of the MIF test was considered by the authors as a warning sign of the preleukemic state among such patients. The LMI assay was used to investigate specific immune responses of patients with CML (Gothoskar et al., 1976; Gangal et al., 1977, 1979). A significant number of patients revealed LMI reaction with blast extracts and the results suggested the sharing of antigens between myeloid leukemic cells, myeloid blasts, and lymphoid blasts. These interesting results need to be enlarged upon and confirmed before conclusions can be made about their significance.

10. Human, Tumor-Related Viruses LMI studies on the relationship between human, breast cancer and MMTV were already mentioned in Section V,B,1. Apart from sporadic in-

MIGRATION INHIBITION IN TUMOR IMMUNOLOGY

293

vestigations of migration inhibition with potentially oncogenic viruses (e.g., hepatitis B, Gupta et al., 1980)or viruses that obviously cause premalignant lesions (e.g., wart virus, Lee and Eisinger, 1976), most studies have focused on detection of immune reactions against Epstein-Barr virus (EBV) or EBVrelated antigens, although the oncogenic potential of herpesviruses type 1 and type 2 (reviewed by Nahmias and Norrild, 1980)and human cytomegalovirus (reviewed by Geder, 1980)is well known. In this article we only refer to LMI findings in the EBV system. Human EBV is the causative agent of infectious mononucleosis (IM) and is closely associated with two human neoplasms, African Burkitt’s lymphoma (BL) and nasopharyngeal carcinoma (NPC), and it may be involved in their etiology (reviewed by de The, 1980). The difference between self-limiting lymphoproliferation in IM and unlimited cell proliferation in the EBV-related cancers may be partly due to the host’s immune response. This is reflected by the bulk of immunological studies in this field (reviewed by Rickinson and Moss, 1983). Cell-mediated immunity to EBV measured by LMI has been studied by several authors both in healthy persons and in patients with IM (Fimmel et al., 1974; Lai et al., 1974, 1975, 1977a,b; Periman et al., 1977; Wainwright et al., 1979; Lai and Alpers, 1980) as well as in NPC (Levine et al., 1981).All authors found that healthy EBV-seropositive (SP) donors showed LMI in the presence of the extract from the EBV-genome-carrying P3HR-1 or Raji cells. No response was observed however in the acute phase of IM, and LMI appeared only after convalescence. NPC patients were shown to have EBVrelated LMI responses. These results are, however, difficult to interpret, especially concerning virus specificity, because of the small number of control (i.e., EBV seronegative, SN) subjects, and the frequent lack of tests with mock antigens (extracts from corresponding EBV-negative cells). For these reasons and because of the undefined protein concentration of extracts used in some studies and the surprisingly high protein concentrations used in others, nonspecific effects cannot be excluded (see Section 111,B). Furthermore, it is not yet clear which of the EBV-determined antigens is responsible for the effect. We developed an EBV-specific LMI assay and made a critical comparison between extracts from EBV-negative lymphoma lines and EBV-genomecarrying sublines derived from the same cell lines. We found that leukocytes of SP individuals responded to extracts from EBV-carrying cells in LMI tests, while extracts from EBV-negative cells elicited no reaction. SN persons did not respond to any of the extracts (Szigeti et al., 1980b, 1981a). Analysis of the EBV-specific LMI tests showed that at least three different groups of antigens could elicit the reaction: the nuclear antigen EBNA (Szigeti et al., 1981a, 1982b), a membrane component of EBV-transformed virus nonproducer cells (Szigeti et al., 1981c), and the antigen complex

294

ROBERT SZIGETI

associated with lytic infection (early antigen, EA; virus capsid antigen, VCA) (Szigeti et al., 1982~).All healthy SP donors showed EBNA-induced LMI, which was T-cell mediated and required the presence of macrophages or interleukin 1 (Szigeti et al., 1984a). LMI could be induced in SP donors not only with extracts derived from cells growing in vitro, but even those from EBV-genome-containing BL and NPC biopsies, while EBV-negative tumor extracts gave no difference in LMI between SP and SN donors. Thus, it seemed that EBV-LMI assay might be useful for detecting EBV-genomes in tissue and tumor extracts (Szigeti et al., 1983). Recently we extended our observations and found that LMI induced by solubilized membranes in healthy SP donors could be blocked with sera from patients with BL or NPC. The blocking activity of the sera could be removed by absorption with EBVcarrying but not EBV-negative cells, suggesting that the antigen involved was present on the cell surface and was associated with EBV transformation of the cells (Szigeti et al., 1984~).The identification and purification of this antigen should be subjected to further analysis. In contrast to healthy SP donors, both cell extracts containing only EBNA and those containing partially purified EBNA were ineffective in patients in the acute phase of IM. These gave, however, significant LMI to extracts containing EA-VCA in addition. This “mononucleosis-type” (EBNA-negative, EA-VCA-positive) LMI pattern was found even in chronic mononucleosis, X-linked lymphoproliferation, and in a group of patients with Hodgkin’s lymphoma, ataxia telangiectasia, and elevated antibody titers to EBV antigens (Szigeti et al., 1982~).Patients with this type of LMI response also showed certain impairment in other immunological tests (Masucci et al., 1984b). Similar LMI studies in BL and NPC have not yet been performed. 1 1 . Effect of Cancer Patients’ Sera on Migration Inhibition

Some reports document the presence of MIF- or LIF-like activity in human sera. Cohen et al. (1974) demonstrated that migration inhibitory activity was recovered from the serum of 81% of patients with chronic lymphocytic leukemia (CLL), Hodgkin’s disease, lymphoma, or myeloma. In the material of Benjamin et al. (1979), MIF activity was detected in the sera of 50% of CLL patients. LMI by the patients’ sera was found in various solid tumors (Cochran et al., 1976b; Bice et al., 1976; Botto et al., 1977; Abai and Fekete, 1982). The preliminary physicochemical characterization of the migration inhibitory serum component showed properties similar to LIF (Cohen et al., 1974). These data, however, have not since been confirmed. Circulating antigens, as in murine tumors (Poupon et al., 1974), or spontaneous lymphokine production as a sign of in uiuo activation of lymphocytes, as it occurs in infectious mononucleosis (Palit et al., 1978) cannot be excluded.

MIGRATION INHIBITION IN TUMOR IMMUNOLOGY

295

Generally, migration inhibitory activity of sera showed no relationship to the clinical stage of the disease. Botto et aZ. (1977), however, found that disappearance of this activity in the serum preceded skin anergy. Thus, such an LMI assay might be useful as an in uitro test for prognosis. On the other hand, this serum migration inhibitory effect showed a negative correlation with recall- or tumor-antigen-induced LMI, skin reactions, and other immunological parameters. Such a blocking effect was described by a number of investigators (Gouillou and Giles, 1973; Kjaer, 1974a, 1976a); Favila et al., 1978). The nature of the blocking factor is unknown. VI. Concluding Remarks

1. MMI and LMI phenomena are in vitro correlates of delayed hypersensitivity, and migration inhibition assays are used to evaluate cellular immunity to tissue, tumor, and virus antigens. Any nontoxic immunogenic material can be applied as antigen in these assays at proper concentrations, which must be determined in preliminary, pilot experiments. 2. Technically the assays are relatively simple and can be evaluated after 18-96 hr, depending on whether the direct or indirect technique was applied. 3. One obvious advantage is that the assays can be used in allogeneic combinations of cells, since during the short incubation period no apparent sensitization and reactivity to histocompatibility antigens can be found. Another advantage is that tests can be performed with cell extracts and cell fractions (whole living cells are not required). 4. The vast majority of animal and human tumors studied with MMI and LMI appeared to contain tumor-determined (TAA,TSTA) antigen, usually not present in normal adult tissue. Tumors of the same histological type can share common TAA/TSTA, but those of different histological type reveal no or weak cross-reaction. 5. The tumor-deterwined antigens are always immunogenic in autochthonous/autologous and usually in allogeneic tumor-bearing animaldpatients. Reactivity of normal healthy controls/donors is rare or absent. The assays (especially the agarose microdroplet assay) are sensitive and can detect tumor-determined antigens at very low concentrations. 6. The assays are available for follow-up studies of antigen purification particularly in cases when no detectable antibodies can be found. In spite of many attempts with the assay, however, no identification or purification of TAA/TSTA has been accomplished so far. 7. The lack of standardization in antigen preparation, and the use of various techniques as well as different methods of statistical evaluation led to inconsistency in the results obtained among various laboratories.

296

ROBERT SZIGETI

8. Therefore, these tests cannot be used for immunodiagnosis in human oncology at present. Although the application of these tests in monitoring general cell-mediated immune reactions of cancer patients (especially in colorectal cancer) provided promising results, no uniform pattern has yet been obtained in regard to tumor-specific LMI reactivity of patients as a function of stage of disease or as a predictive prognostic sign. Thus, at present the findings with these assays are still only of research interest. REFERENCES Abai, A. M.. and Fekete, B. (1982). Haematologia 15, 111-125. Ablin, R. J. (1975). Oncology 31, 177-202. Ablin, R. J., and Bhatti, R. A. (1981). In “Prostatic Cancer” (R. J. Ablin, ed.), pp. 183-204. Dekker, New York. Ablin, R. J., Bush, I. M., Bruns, G. R., and Guinan, P. 1).(1975). Clin. Res. 23, 485A. Ablin, R. J., Marrow, C., Guinan, P. D . , Bruns, G. R., Sadoughi, N., John, T., and Bush, I. M. (1976). Urol. Znt. 31, 444-458. Ablin, R. F., Bruns, G. R., Sadoughi, N., Bush, I. M., John, T., and Guinan, P. D. (1978). Urology XI, 289-292. Airoldi, M., Mastromatteo, V., Negri, L., Fazio, M., and Gandolfo, S. (1982). Boll. Soc. Ztal. Biol. Sper. 58, 1213-1220. Akiyoshi, T., Hata, A., and Tsuji, H. (1975). Gann 66, 167-174. Akiyoshi, T., Miyazaki, S., Koba, F., Kawaguchi, M., and Tsuji, H. (1981). Gann 72,988-991. Akiyoshi, T., Koba, F., Kawaguchi, M., Miyazaki, S., and Tsuji, H. (1982).J . Natl. Cancer Znst. 69, 345-347. Albrecht, S . , Thuy, L. T., Pasternak, G . , Schlott, B., Reinhofer, J., Lunow, E., Loewe, G . , and Gutz, H . J. (1978). Acta B i d . Med. Germ. 37, 1747-1757. Allison, A. C. (1980). In “Viral Oncology” (G. Klein, ed.), pp. 481-487. Raven, New York. Alth, G., Denck, H., Fischer, M., Karrer, K., Kokron, O., Korizek, E., Miksche, M., Ogris, E., Reider, C.. Titscher, R., and Wrba, H. (1973). Cancer Chemother. Rep. 4, 275-277. Amino, N., Pysher, T., Cohen, E. P., and DeGroot, L. J. (1975). Cancer 36, 963-973. Andersen, V., Bendixen, G., and Schipldt, T. (1969). Acta Med. Scand. 186, 101-103. Andersen, V., Bjerrum, 0.. Bendixen, C . , Schipldt, T.,and Dissing, I. (1970).Znt. J . Cancer 5, 357-363. Anderson, J. M., Kelly, F., Wood, S. E., Rodger, K. D., and Freshney, R. I. (1973). Br. J . Cancer 28 (Suppl.), 83-96. Ansfield, M. J., Shalaby, M. R . , and Kaltreider, H. B. (1980).J . Zmmunol. Methods 34, 43-48. Aoki, N., and DeGroot, L. J . (1982). Acta Endocrinol. 99, 56-63. Armitstead, G. R., and Gowland, G. (1975). Gut 16, 968-972. Astor, S. H . , Spitler, L. E., Frick, 0. L., and Fudenberg, H. H. (1973). J . Zmmunol. 110, 1174- 1179. Azadegan, A , , Kaneene, J. M . B., Muscoplat, C. C., and Johnson, D. W. (1981). Am. J . Vet. Res. 42, 122-125. Bellina, L., and Salerno, A. (1981).J . Immunol. Methods 43, 277-281. Bendixen, G. (1969). Gut 10, 631-636. Bendixen, G., Bendtzen, K., Clausen, J. E., Kjaer, M., and Splborg, M. (1976). Scand. J . Zmmund. 5 (SUPPI.),175-184. Bendtzen, K. (1977). Cell. Zmmunol. 29, 382-393. Bendtzen, K. (1983). Allergy 38, 219-226.

MIGRATION INHIBITION IN TUMOR IMMUNOLOGY

297

Bendtzen, K.,and Rocklin, R. E. (1980).J. Zmmunol. 125, 1775-1781. Bendtzen, K., Andersen, V., and Bendixen, G . (1975). Acta Alkrgol. 30, 133-149. Benjamin, D., Aderka, D., Livni, E., Joshua, H., Shaklai, M., and Pinkhas, J. (1979).J. Natl. Cancer Znst. 63, 1175-1179. Berman, M. A., and Weigle, W. 0. (1977).J . E x p . Med. 146, 241-256. Bernhard, M. I., Wanebo, H. J., Helm, J., Pace, R. C., and Kaiser, D. L. (1983). Cancer Res. 43, 132-137. Bice, D. E., Gruwell, D., and Salvaggio, J. (1976). J . Reticuloendothel. Soc. 19, 281-289. Black, M. M . , Leis, H. P., Shore, B., and Zachrau, R. E. (1974a). Cancer 33, 952-958. Black, M . M., Moore, D. H., Shore, B., Zachrau, R. E., and Leis, H. P. (1974b). Cancer Res. 34, 1054-1060. Black, M. M., Zachrau, R. E., Dion, A. S., Shore, B., Fine, D. L., Leis, H. P., and Williams, C. J. (1976). Cancer Res. 36, 4137-4142. Bloom, B. R., and Bennett, B. (1966). Science 153, 80-82. Bloom, B. R., Bennett, B., Oettgen, H. F., McLean, E. P., and Old, L. J. (1969). Proc. Natl. Acad. Sci. U . S . A . 64, 1176-1180. Boddie, A. W., Holmes, E. C., and Roth, J. A. (1974). Surg. Forum 25, 109-110. Boddie, A. W., Urist, M. M., Chee, D. O., Holmes, E. C., and Morton, D. L. (1975a).Znt. J. Cancer 16, 1035-1041. Boddie, A. W., Holmes, E. C., Roth, J. A., and Morton, D. L. (1975b).Znt. J . Cancer 15,823829. Boddie, A. W., Urist, M. M., Chee, D. O., Holmes, E. C., and Morton, D. L. (1976). Znt. J. Cancer 18, 161-167. Botto, H. G . , Gauthier, H., Pouillart, P., Bruley-Rosset, M., and Huguinenin, P. (1977). Eur. J . Cancer 13, 329-332. Brandes, L. J., and Goldenberg, G. J. (1976). Cancer Res. 36, 3707-3710. Braun, M., Sen, L., Bachmann, A. E., and Paulovski, A. (1972). Blood 39, 368-376. Bull, D. M., Leibach, J. R., Williams, M. A., and Helms, R. A. (1973). Science 181, 957-959. Burtin, P., and Chavanel, G. (1980). Br. J . Cancer 42, 345-348. Burtin, P., Chany, E., Beaulieu, R., Maunoury, M. T., Chavanel, G . , and Sabine, M. C. (1977). Cancer 39, 2378-2385. Cannon, G. B., McCoy, J. L., Jerome, L. J., Reddick, R., Alford, C., Tinley, V., and Herberman, R. B. (1978a).J . Natl. Cancer Znst. 61, 1181-1186. Cannon, G . B., McCoy, J. L., Connot, R. J., Jerome, L., Keys, L., Dean, J. H., and Herberman, R. B. (1978b).J . Natl. Cancer Znst. 60, 969-978. Cannon, G. B., Barsky, S. H., Alford, T. C., Jerome, L. J., Tinley, V., McCoy, J. L., and Dean, J. H. (1982).J . Natl. Cancer Znst. 68, 935-943. Cerilli, J., and Smith, M. C. (1972). Surg. Gynecol. Obstet. 134, 739-745. Cerni, C. (1977). Oncology 34, 216-218. Char, D. H. (1977). Znoest. Ophthalmol. Visual Sci. 16, 176-179. Char, D. H., Jerome, L., McCoy, J. L., and Herberman, R. B. (1975).Am. J . Opthalmol. 79, 812-816. Chen, S. S., Koffler, D., and Cohen, C. J. (1975). Am. J. Obstet. Cynecol. 121, 91-95. Chen, S. Y., Koffler, D., and Cohen, C. J. (1973). Am. J. Obstet. Cynecol. 115, 467-470. Chess, L., Rocklin, R. E., MacDermott, R. P., David, J. R., and Schlossman, S. F. (1975).J. Zmmunol. 115, 315-317. Chiang, W. T., Wei, P. Y., and Alexander, E. R. (1976).Am. J. Ohstet. Cynecol. 126, 116-121. Churchill, W. H., and Rocklin, R. E. (1973). Natl. Cancer Znst. Monogr. 37, 135-137. Churchill, W. H., Zbar, B., Belli, J . A., and David, J. R. (1972).J . Natl. Cancer Inst. 48, 541549. Clausen, J. E. (1971). Acta Allergol. 26, 56-80.

298

ROBERT SZIGETI

Cochran, A. J. (1978). “Man, Cancer and Immunity.” Academic Press, New York. Cochran, A. J., Spilg, W. G . , MacKie, R. M., and Thomas, C. E. (1972a).Br. Med. J. 4,67-70. Cochran, A. J.. Jehn, U. W., and Gothoskar, B. P. (1972b). Lancet 1, 1340-1341. Cochran, A. J., MacKie, R . M., Thomas, C. E., Grant, R . M., Cameron-Mowat, D. E., and Spilg, W. G. (1973a). Br. J . Cancer 28 (Suppl.), 77-82. Cochran, A. J., Jehn, U. W., and Gothoskar, B. P. (1973b). Pigment Cell 1, 360-365. Cochran, A. J., Grant, R. M., Spilg, W. G . , MacKie, R. M . , Ross, C. E., Hoyle, D. E., and Russell, J. M. (1974). Znt. J. Cancer 14, 19-25. Cochran, A. J., Ross, C. E . , MacKie, R. M., Grant, R . M., and Hoyle, D. E. (1975). Behring Znst. Mitt. 56, 125-130. Cochran, A. J., MacKie, R. M., Ross, C. E . , Ogg, L. J., and Jackson, A. M . (1976a). I n “Clinical Tumor Immunology” 0. Wybran and M. Staquet, eds.), pp. 69-76. Pergamon, Oxford. Cochran, A. J . , MacKie, R. M., Ross, C. E., Ogg, L. J., and Jackson, A. M. (1976b). Znt. J. Cancer 18, 274-281. Cochran, A. J., MacKie, R. M., Grant, R . M., Ross, C. E., Connell, M. D., Sandilands, G., Whaley, K., Hoyle, D. E., and Jackson, A. M. (1976~).Znt. J. Cancer 18, 298-309. Cochran, A. J., Todd, G . , Hart, D.M., Morrisdn, L. J.. and MacKie, R . M. (1982). Cancer Zmmunol. Zmmunother. 14, 78-81. Cohen, S., Fisher, B., Yoshida, T., and Bettigole, R. E. (1974).N . Engl. J. Med. 290,882-886, Cortesina, G., Cavallo, G. P., Beatrice, F., Sartoris, A., Bussi, M., Morra, B., DiFortunato, V., Poggio, E., and Rendine, S. (1982). Tumori 68, 39-46. David, J. R . , and David, R . R. (1972). Prog. Allergy 16, 300-449. Dean, J. H . , McCoy, J. L., Cannon, G . B., Leonard, C. M., Perlin, E., Kreutner, A,, Oldham, R. K., and Herberman, R. B. (1977).J . Natl. Cancer Inst. 58, 549-555. Delpre, G., Pitlik, S., Halevy, J., Perry, G., Avidor, I., Kadish, U., and Rosenfeld, J. (1983). Arch. Znt. Med. 143, 363-365. Eckhardt, R., Heinisch, M., and Meyer zum Buschenfelde, K. H. (1976). Scand. J . Gastroenterol. 11, 49-54. Elias, E. G., and Elias, L. L. (1975a). Cancer 36, 1393-1398. Elias, E. G . , and Elias, L. L. (1975b). Surg. Cynec. Obstetr. 141, 715-718. Elias, E. G., Elias, L. L., Didolkar, M. S., and Hebel, J. R. (1977). Cancer 40, 687-692. Elias, E. G., Didolkar, M. S., Reed, W. P., and Buda, B. S. (1983). Proc. Am. Soc. Clin. Oncol. 2, C-203 (Abstr.). Emmrich, F., Irmscher, J., Kotzsch, M., Miiller, M., and Ambrosius, H. (1978). Br. J. Cancer 37, 461-465. Evans, C. M., and Bowen, J. G . (1977). Proc. R . Soc. M e d . 70, 417-420. Fahlbusch, B., and Zschiesche, W. (1981). Acta Biol. M e d . Gem. 40, 793-800. Fahlbusch, B . , Zschiesche, W., and Schumann, I. (1981).Acta B i d . Med. Germ. 40,785-792. Faiferman, I . , Gleicher, N., Cohen, C. J., and Koffler, D. (1977).J. Natl. Cancer Znst. 59, 1593- 1597. Falk, R. E., Mann, P., and Langer, B. (1973). Arch. Surg. 107, 261-265. Favila, J., Jimenez, L., Castro, M. E., Garcia-Garcia, G., Gutierrez, A., and Pulido, 1. (1978). 1. Natl. Cancer Inst. 60, 1279-1283. Federlin, K . , Maini, R. N., Russell, A. S., and Dumonde, D. C. (1971).J. Clin. Pathol. 24, 533-536. Firnmel, P. J., Lai, P. K., Keast, D., Alpers, M. P., and Mackay-Scollay, E. M. (1974). Znt. J. Cancer 13, 599-605. Fukuda, M., Wanebo, H. J., Tsuei, L., and Sarkar, N. H. (1979).J . Natl. Cancer Znst. 63, 1161-1169.

MIGRATION INHIBITION IN TUMOR IMMUNOLOGY

299

Fukuda, M., Wanebo, H. J., Tsuei, L., Ashikari, R . , and Sarkar, N. H. (1980).J . Natl. Cancer Zmt. 64, 431-437. Gangal, S. G . , Joshi, C. S., Gothoskar, B. P., Gollerkeri, M. P., and Advani, S. H. (1977). Haematologica 62, 469-478. Gangal, S. G., Damle, N . K., Khare, A. G . , and Advani, S. H. (1979). Br. J . Cancer 40, 391396. Geczy, C. L., Friedreich, W., and de Weck, A. L. (1975). Cell. Zmmunol. 19, 65-77. Geder, L. (1980).I n “Oncogenic Herpesviruses” (F. Rapp, ed.), Vol. 11, pp. 47-60. CRC Press, Boca Raton, Florida. George, J. M., Williams, M. A., Almoney, R., and Sizemore, G. (1975).Cancer36, 1653-1661. George, M., and Vaughn, J. M. (1962). Proc. Soc. E x p . Biol. Med. 111, 514-521. Golding, B., Golding, H . , Lomnitzer, R., Jacobson, R., Koornhof, H. J., and Rabson, A. R. (1977). Clin. Zmmunol. Zmmunopathol. 7, 114-122. Goldstein, M. S., Shore, B., and Gusberg, S. B. (1971).Am. J . Obstet. Gynecol. 111, 751-755. Gbrski, A. J., Dupont, B., Hansen, J. A., and Good, R . A. (1975).Proc. Natl. Acad. Sci. U.S.A. 72, 3197-3200. Gothoskar, B. P., D’Silva, H., Gharpure, H., and Advani, S. H . (1976). Eur. J . Cancer 12, 815-821. Guillou, P. J . , and Giles, G. R. (1973). Gut 14, 733-738. Guillou, P. J., Brennan, T. G., and Giles, G. R. (1975). Gut 16, 290-297. Graham-Pole, J., Ross, C. E., Ogg, L. J., and Cochran, A. J. (1976). Lancet 1, 1376-1379. Grimm, E. A., Silver, H. K. B., Roth, J. A., Cbee, D. 0.. Gupta, R. K., and Morton, D. L. (1976). Znt. J . Cancer 17, 559-564. Gupta, B. P., Misra, R. C . , Agarwal, S. K., Saha, K., and Gupta, P. S. (1980). ZndianJ. Med. Res. 72, 891-894. Hahn, T., Levin, S., and Handzel, Z. T. (1976). Clin. Exp. Zmmunol. 24, 476-480. Halliday, W. J. (1971).J. Zmmunol. 106, 855-857. Halliday, W. J. (1972). Cell. Zmmunol. 3, 113-122. Halliday, W. J., and Webb, M. (1969).J . Natl. Cancer Znst. 43, 141-149. Harrington, J. T. (1974). Cell. Zmmunol. 12, 476-480. Harrington, J. T. (1977). Cell. Zmmunol. 30, 261-271. Harrington, J. T., and Stastny, P. (1973).J . Zmmunol. 110, 752-759. Herberman, R. B., Hollinshead, A , , Char, D., Oldham, R., McCoy, J. L., and Cohen, M. (1975). Behring Znst. Mitt. 56, 131-138. Herberman, R. B., Holden, H. T., Djeu, J. Y., Jerrels, T. R . , Varesio, L., Tagliabue, A., White, S. L., Oehler, J. R., and Dean, J. H. (1980). In “Macrophages and Lymphocytes: Nature, Functions and Interactions” (E. Friedman, ed.), pp. 361-379. Plenum, New York. Hilberg, R. W., Balcerzak, S. P., and LoBuglio, A. F. (1973). Cell. Immunol. 7, 152-158. Holden, H. T . , Varesio, L., Taniyama, T., and Puccetti, P. (1980). In “Macrophages and Lymphocytes: Nature, Functions and Interactions” (E. Friedman, ed.), pp. 509-520. Plenum, New York. Hollander, den, F. C., van Lieshut, J. I., and Schuurs, A. H. W. M. (1981). Zmmunopharmacology 3, 161-178. Hollinshead, A. C., Jafhrs, W. J., Alpert, L. K., Harris, J. E., and Herberman, R. B. (1974a). Cancer Res. 34, 2961-2968. Hollinshead, A. C., Herberman, R. B., Jafhrs, W. J., Alpert, L. K., Minton, J. P., and Harris, J. E. (1974b). Cancer 34, 1235-1243. Holmes, E. C . , Roth, J. A., and Morton, D. L. (1975). Surgery 78, 160-164. Hook, R. R., Hamby, C. V., Millikan, L. E., Berkelhammer, J., Stills, H. F., and Oxenhandler, R. W. (1983). Int. J . Cancer 31, 633-637.

300

ROBERT SZIGETI

House, A. K., Wisniewski, S., and Woodings, T. L. (1975). Dis. Colon Rectum 18, 100-106. Imai, K., Hosokawa, Y., Abe, H., Yachi, A., and Wada, T. (1977). Gann 68, 581-587. Iwaki, J., Akagi, M., Sairenji, T., and Hinuma, Y. (1979).J. Natl. Cancer Znst. 63, 5-10. Jones, B. M. (1976). Br. J. Cancer 3, 14-19. Jones, B. M., and Turnbull, A. R. (1974). Br. J. Cancer 29, 337-339. Jones, B. M., and Turnbull, A. R. (1975). Br. J. Cancer 32, 339-344. Jones, B. M., Conolly, C. E., Isaacson, P., Turner, D. T. L., andTurnbull, A. R. (1976). Br. J. Cancer 34, 94-96. Kadish, A. S., Marcus, D. M., and Bloom, B. R. (1976). Znt. J. Cancer 18, 581-586. Kapadia, A., O’Reilly, R. J., Good, R. A., and Gupta, S. (1982). Fed. Proc., Fed. Am. Soc. E x p . B i d . 1365. Kjaer, M. (1974a). Acta Pathol. Microbiol. Scand. Sect. B 82, 294-296. Kjaer, M. (1974b). Eur. J . Cancer 10, 523-531. Kjaer, M . (1975). Eur. J. Cancer 11, 281-289. Kjaer, M. (1976a).Eur. J. Cancer 12, 783-792. Kjaer, M. (1976b). Eur. J. Cancer 12, 889-898. Kjaer, M., and Bendixen, G . (1976). Ann. N.Y. Acad. Sci. 276, 260-271. Kjaer, M., and Christensen, N. (1977). Cancer Zmmunol. Zmmunother. 2, 41-48. Kjaer, M., and Sflrensen, T. B. (1976). Acta Allergol. 31, 141-158. Kojima, 0.. Oh, A,, Kitagawa, N., Uehara, Y.,Nishioka, B., Fujita, Y., and Majima, S. (1981a). Jpn. J. Surg. 11, 413-421. Kojima, 0.. Fujita, Y., Iwai, N., Yamane, T., Kitagawa, N., Nishioka, B., and Majima, S. (1981b).Aust. N. Z. J. Surg. 51, 434-439. Kronman, B. S., Wepsic, H. T., Churchill, W. H., Zbar, B., Borsos, T., and Rapp. H. J. (1969). Science 165, 296-297. Lai, P. K., and Alpers, M. P. (1980). Comp. Zmmun. Microbiol. Infect. Dis. 2, 565-585. Lai, P. K., Mackay-Scollay, E. M., Fimmel, P. J . , Alpers, M. P., and Keast, D. (1974).Nature (London) 252, 608-610. Lai, P. K., Alpers, M. P., and Mackay-Scollay, E. M. (1975).J . Natl. Cancer Inst. 55, 13191322. Lai, P. K., Alpers. M. P., and Mackay-Scollay, E. M. (1977a). Znt. J. Cancer 20, 21-29. h i , P. K., Alpers, M. P., and Mackay-Scollay, E. M. (197%). Infect. Zmmun. 17, 28-35. Landolfo, S.,Herberman, R. B., and Holden, H. T. (1977a).J. Natl. Cancer Inst. 59, 16751683. Landolfo, S., Herberman, R. B., and Holden, H. T. (1977b). J . Zmmunol. 118, 1244-1248. Landolfo, S.,Herberman, R. B., and Holden, H. T. (1978).J. Zmmunol. 121, 695-701. Lavergne, J. A., Lamm, D. L., Redwin, H. M., and Harrington, J. T. (1979). Cancer Res. 39, 1985-1988. Lawniczak, M., Jezewska, E., and Zeromski, J. (1983). Annu. Meet. Znt. Soc. Oncodeo. Biol. Med., I l t h , Sept., Stockholm p. 148 (Abstr.). Lee, A. K. Y., and Eisinger, M. (1976). Clin. E x p . Zmmunol. 26, 419-424. Lee, C., Chon, S., and Lin, T. (1977). Cancer Res. 37, 918-921. Levine, P. H., Pizza, G . , Cannon, G . , Ablashi, D., Armstrong, G., and Vim, D. (1981). I n “Cancer Campaign, Vol. 5, Nasopharyngeal Carinoma” (F. Grundmann, ed.), pp. 137144. Fischer, Stuttgart. Lidereau, R., Ricqles, de, D., and Saracino, R. T. (1983).Ann. Virol. Inst. Pasteur 134E, 97107. Light, P. A., Preece, A. W., and Waldron, H. A. (1975). Clin. Exp. Zmmunol. 22, 279-284. Littman, 8. H . , Meltzer, M. S., Cleveland, R. P., Zbar, B., and Rapp, H. J. (1973).J . Natl. Cancer Inst. 51, 1627-1635.

MIGRATION INHIBITION IN TUMOR IMMUNOLOGY

301

Lloyd, R. V., Rosen, P. P., Sarkar, N. H., Jimenez, D., Kinne, D. W., Menendez-Botet, C., and Schwartz, M. K. (1983). Cancer 51, 654-661. Lurie, B. B., Bull, D. M., Zamchek, N . , Steward, A. M., and Helms, R. A. (1975). J. Natl. Cancer Znst. 54,319-325. MacKie, R. M., Spilg, W. G . S., Thomas, C. E., and Cochran, A. J. (1972). Br. J . D e m t o l . 87, 523-528. McCoy, J. L. (1979). In “Immunodiagnosis of Cancer” (R. B. Herberman and K. R. McIntire, eds.), pp. 979-998. Dekker, New York. McCoy, J. L., Jerome, L. F., Dean, J. H., Cannon, G . B., Alford, T. C., Doering, T., and Herberman, R. B. (1974).J. Natl. Cancer Znst. 53, 11-17. McCoy, J. L., Jerome, L. F., Dean, J. H., Perlin, E., Oldham, 0. K., Char, D. H., Cohen, M. H., Felix, E. L., and Herberman, R. B. (1975). J . Natl. Cancer Znst. 55, 19-23. McCoy, J. L., Jerome, L. F., Anderson, C., Cannon, G . B., Alford, T. C., Connor, R. J., Oldham, R. K., and Herberman, R. B. (1976a). J. Natl. Cancer Znst. 57, 1045-1049. McCoy, J. L., Jerome, L. F., Dean, J. H., Cannon, G . B., Connor, R. J., and Herberman, R. B. (1976b). In “Zn Vitro Methods in Cell Mediated and Tumor Immunity” (B. R. Bloom and J . R. David, eds.), pp. 607-612. Academic Press, New York. McCoy, J. L., Jerome, L. F., Cannon, G. B., Weese, J. L., and Herberman, R. B. (1977a).J. Natl. Cancer Znst. 59, 1423-1417. McCoy, J. L., Jerome, L. F., Cannon, G. B., Pomeroy, T. C., Connor, R. J., Oldham, R. K., Weese, J. L., and Herberman, R. B. (1977b). J. Natl. Cancer Znst. 59, 1119-1125. McCoy, J. L., Dean, J. H., and Herberman, R. B. (1977~). J. Zmmunol. Methods 15, 355-371. McCoy, J. L., Padarathshingh, M . , Dean, J. H., Henriksen, O., Natori, 0.. and Law, L. H. (1977d). J . Zmmunol. 119, 306-310. McCoy, J. L., Dean, J. H., Cannon, G. B., Jerome, L. F., Alford, T. C., Parks, W. P., Gilden, R. V., Oroszlan, S. T., and Herberman, R. B. (1978).J. Natl. Cancer Inst. 60,1259-1267. McIllmurray, M. B., Price, M. R., and Langman, M. J. S. (1974). Br. J. Cancer 29, 305-311. Maini, R. N., Roffe, L. M., Magrath, I. T., and Dumonde, D. C. (1973).Znt. Arch. Allergy 45, 303-321. Manor, L. S., Livni, E., Joshua, H., and Ben-Sira, I. (1978). Znuest. Ophthalmol. Visual Sci. 17, 684-687. Mantovani, G . , Maxia, G., Caporelli, S., Cossu, F., and Sestu, M. (1978). Tumori 64,487-494. Mantovani, G . , Manca, M. A., Cossu, F., Proto, E., Taglieri, G . , Mirigliani, F., and Gaspardini, G . (1981). Tumori 67, 169-175. Masucci, G., Szigeti, R., Stevens, D., Masucci, M. G., Klein, E., Petersen, J., Bendtzen, K., and Klein, G. (1984a). Cell. Immunol. 85, 511-518. Masucci, G., Berkel, I., Masucci, M. G . , Ernberg, I., Szigeti, R., Ersoy, F., Sand, O., Yegin, O., Henle, G., Henle, W., Pearson, G., Aman, P., and Klein, G . (198413).J. Clin. Zmm. (in press). Matzku, S., Zijller, M., and Schulz, U . (1978). Ercerpta Med. 439, 161-163. Matzku, S., Zoller, M., Ikinger, U., and Price, M. R. (1979). Br. J. Cnncer 39, 516-523. Matzku, S., Zoller, M., Schulz, U., Dietze, W., Bartb, H. O., Saeger, H. D., and Price, M. R. (1980). J. Natl. Cancer Znst. 54, 1345-1348. Meleady, D. F., Butler, M. R., Greally, J. F., and Mullins, G . M. (1980). Zr. J . Med. Sci. 149, 391-400. Mesa-Tejada, R., Keydar, I., Ramanarayanan, M., Ohno, T., Fenoglio, C., and Spiegelman, S. (1978). Proc. Natl. Acad. Sci. U.S.A. 75, 1529-1533. Meshulam, D. H., Blair, H. E., Wong, B. L., Charm, S., Minowada, J., and Rocklin, R. E. (1982). Proc. Natl. Acad. Sci. U.S.A. 79, 601-605. Moore, D. H., Moore, I. I., and Moore, C. T. (1983). Ado. Cancer Res. 40, 189-253.

302

ROBERT SZIGETI

Morison, W. L. (1975).J. Clin. Pathol. 27, 113-117. Morrison, L. J., Cochran, A. J . , MacKie, R. M., Ross, C. E., Todd, G . , and Garland, C. G. (1979). Znt. J. Cancer 24, 11-16. Morton, D. L., Malmgren, R. A., Holmes, C. E., and Ketcham, A. S. (1968). Surgery 64, 233240. Nagaraja, K. V., Newman, J. A., and Pomeroy, B. S. (1982). Am. J. Vet. Res. 43, 916-918. Nahmias, A. J., and Norrild, B. (1980). In “Oncogenic Herpesviruses” (F. Rapp, ed.), Vol. 11, pp. 25-45. CRC Press, Boca Raton, Florida. Nelson, R. D., and Leu, R . W. (1975).J. Zmmunol. 114, 606-609. Neville, L. E., and Lischner, H. W. (1977).J. Reticuloendothel. Sac. 22 (Suppl.), 45a. Neville, M. E., and Lischner, H. W. (1981).J. Erp. Med. 154, 1868-1880. Neville, M . E., and Lischner, H. W. (1982).1. Irnmunol. 128, 1063-1069. Nori, M . , and Gothoskar, B. P. (1983). Neoplasmu 30, 287-293. Noronha-Blob, L., and Huang, S. W. (1980). Clin. E x p . Zmmunol. 40, 627-632. Oldham, R. K., Weese, J. L., Herberman, R. B., Perlin, E., Mills, M., Heims, W., Blom, J., Green, D., Reid, J., Bellinger, S., Law, I., McCoy, J. L., Dean, J. H., Cannon, C. B., and Djeu, J. (1976). Znt. J. Cancer 18, 739-749. Pacheco-Rupil, B., Depiante-Depaoli, M., Trevino, E., and Bengoela, 0. (1982). Cell. Mol. Biol. 28, 73-77. Padarathshingh, M. L., Dean, J. H., McCoy, J. L., Lewis, D. D., Northing, J. W., Natori, T., and Law, L. W. (1978).J . Zmmunol. 120, 1981-1985. Palit, J., Bendtzen, K., and Andersen, V. (1978). Clin. E x p . Immunol. 31, 66-71. Pappas, A., and Schwarze, G. (1974). Z. Zmmunitaetsforsch. 148, 142-150. Pasternak, G., Schlott, B., Gryschek, G., Albrecht, S., Reinhofer, J . , Matthes, E., and von Broen, B. (1983). Annu. Meet. Znt. Soc. Oncodeo. Biol. Med., I l t h , Sept., Stockholm p. 144 (Abstr.). Pekirek, J . , and Krejci, J. (1974).J. Zmmunol. Methods 6, 1-22. Pekirek, J . , Svejcar, J., Vonka, V., and Zavadova, H. (1968). 2. Zmmunitaetsforsch. 134, 449461. Periman, P., Levine, P. H., Ablashi, D. V., and Royston, I. (1977).Znt. J. Cancer 20,506-511. Petersen, J., and Bendtzen, K. (1982). Acta Pathol. Microbiol. Scand. Sect. C 90, 229-235. Petterson, T., Welin, M. G . , and Weber, T. H. (1982).J. Clin. Lab. Zmmunol. 8, 107-111. Poupon, M. F., and Lespinats, G. (1972).J . Natl. Cancer Inst. 48, 1297-1301. Poupon, M . F., Lespinats, G., and Kolb, J. P. (1974). J. Natl. Cancer Inst. 52, 1127-1134. Prystovsky, M. B., Sorokin, C. F., Ceglowski, W. S., Hirschorn, K., and Glade, P. (1975). Znt. Arch. Allergy Appl. Zmmunol. 48, 225-229. Rashen, L., and Arvilommi, H. (1977). Zmmunology 32, 843-847. Wanen, L . , Karchumaki, E., and Krohn, K. (1978a). Cell. Immunol. 37, 221-228. Risanen, L., Karhumaki, E., and Arvilommi, H. (1978b).J. Zmmunol. 121, 418-420. Rasmussen, N., Bendtzen, K., Thomsen, J . , and Tos, M. (1982). Acta Otolaryngol. 94, 261265. Rich, A. R., and Lewis, M. R. (1932). Bull. Johns Hopkins Hosp. SO, 115-131. Richens, E. R . , Williams, M. J., Cough, K. R., and Ancill, R. J . (1974). Gut 15, 19-23. Rickinson, A. B . , and Moss, D. J. (1983). Ado. Viral Oncol. 3, 213-238. Rieche, K., Arndt, A., and Pasternak, G. (1976). Znt. J. Cancer 17, 212-218. Rivera, E. S., Hersh, E. M., Bowen, J. M., Barnett, J. W., Wharton, T., and Murphy, S . G . (1979). Cancer 43, 2297-2305. Rohbins, D. S., Belden, E. L., Hoffman, P. M., Fudenberg, H. H., and Strelkauskas, A. J. (1982). Scand. J . Zmmunol. 15, 563-566. Roberts, M. M., and Bathgate, E. M. (1975). Br. J. Surg, 62, 820-822.

MIGRATION INHIBITION IN TUMOR IMMUNOLOGY

303

Rocklin, R. E. (1974).J. Zmmunol. 112, 1461-1465. Rocklin, R. E. (1975).J . Zmmunol. 114, 1161-1165. Rocklin, R. E . , and Rosenthal, A. S. (1977).J . Zmmunol. 119, 249-252. Rocklin, R. E., and Urbano, A. M. (1978).J . ZmtnunoZ. 120, 1409-1414. Rocklin, R. E., MacDermott, R. P., Chess, L., Schlossman, S. F., and David, J. R. (1974).J . E x p . Med. 140, 1303-1316. Rocklin, R. E., Gagel, R., Feldman, Z., and Tashjian, A. H. (1977).N. Engl. 1. Med. 296,835838. Rocklin, R. E,, Bendtzen, K., and Greineder, D. (1980). Ado. Zmmunol. 29, 55-136. Roed-Petersen, B. (1982). Allergy 37, 429-435. Roed-Petersen, B., Roed-Petersen, J., and Bendixen, G. (1973). Int. J . Cancer 12, 66-72. Ross, C. E., Cochran, A. J., Hoyle, D. E., Grant, R. M., and McKie, R. M. (1973). Lancet 1, 1087- 1088. Ross, C. E., Cochran, A. J., Hoyle, D. E., Grant, R. M., and MacKie, R. M. (1975).Clin. E r p . lmmunol. 22, 126-132. Roth, J. A., Holmes, E. C . , Reisfeld, R. A., Slocuni, H. K., and Morton, D. L. (1976). Cancer 37, 104-110. Sandberg, A. L., Wahl, S. M., and Mergenhagen, S. E. (1975). J. Zmmunol. 115, 138-144. Schelhammer, P. F., Wright, G. L., Rosato, F. E., and Faulconer, R. J. (1979).J. Urol. 122, 746-749. Schmeisser, G., and Kademann, J. (1980). Arch. Geschwuelstforsch. 50, 169-172. Schwarze, G., Dietz, R., and Pappas, A. (1979). Eur. J. Cancer 15, 205-210. Segall, A,, Weiler, O., Genin, J., Lacour, J . , and Lacour, F. (1972). Znt. J. Cancer 9,417-425. Sharma, J. M., Herberman, R. B., Djeu, J. Y., and Nunn, M. E. (1979).J . Zmtnunol. 123,222231. Shevach, E. M., Lee, L., and Ben-Sasson, S. Z. (1975). In “Immune Recognition” (A. S. Rosenthal, ed.), pp. 627-643. Academic Press, New York. Sidi, Y . , Livni, E., Shaklai, M., and Pinkhas, J. (1981). Zsr. J . Med. Sci. 17, 1119-1121. Sirianni, M. C., Luzi, G . , Tavarone, C., Papaluca, M., Fiorilli, M., Messinetti, S., and Aiuti, F. (1981). T ~ m ~67, r i 163-167. Sjogren, H. O., Hellstriim, I., and Klein, C. (1961). Cancer Res. 21, 329-337. Sfiborg, M., and Bendixen, G. (1967). Acta Med. Scand. 181, 247-256. Spitler, L. E. (1973). Clin. Res. 21, 654. Spitler, L. E., Wybran, J . , Fudenberg, H. H., Levin, A. S., Lewis, M., and Horn, L. (1973). Clin. Res. 21, 654. Spitler, L. E., Littooy, F. N., and Sagebiel, R. W. (1977). Cancer Zmmunol. Zmmunother. 2, 69-76. Springer, G. F., Murthy, M. S., Desai, P. R., and Scanlon, E. F. (1976).Naturwissenschaften 63, 300. Springer, G. F., Murthy, S. M., Desai, P. R., and Scanlon, E. F. (1980). Cancer 45, 29492954. Steiner, T., and Watne, A. L. (1970). Cancer Res. 30, 2265-2270. Straus, E., Vernace, S., Janowitz, H . , and Paronetto, F. (1975).Proc. Soc. E x p . Riol. Med. 148, 494-497. Suslov, I. M., McCoy, J. L., Cannon, G. B., and Herberman, R. B. (1980).J . Natl. Cancer Inst. 65, 709-713. Suslov, I., McCoy, J. L., and Herberman, R. B. (1981). J . Natl. Cancer Inst. 66, 233-237. Suter, L., Bloom, B. R., Wadsworth, E. M., and Oettgen, H. F. (1972).J . Zmmunol. 109, 766775. Svejcar, J , , and Johanovsky, J. (1963). Proc. Eur. Congr. Allegy, 5th pp. 375-376.

304

ROBERT SZICETI

Szigeti, R., RBvBsz, T., and Gero-Ferencz, E. (1974a). Acta Allergol. 29, 288-290. Szigeti, R., RCvCsz, T., and Schuler, D. (1974b). Acta Paediatr. Acad. Sci. Hung. 15, 31-39. Szigeti, R., RBvBsz, T., and Schuler, D. (1976). In “Medical Genetics” (G. Szab6 and Z. Papp, eds.), pp. 831-836. Excerpta Medica, Amsterdam. Szigeti, R., Bbkay, J., RBvCsz, T., and Schuler, D. (1980a). Allergy 35, 557-562. Szigeti, R., TimAr, L., and RBvBsz, T. (1980b). Allergy 35, 97-103. Szigeti, R . , Masucci, M . G., Masucci, G . , Klein, E., Klein, G., and Berthold, W. (1980~). Nature (London) 288, 594-596. Szigeti, R., Luka, J . , and Klein, G . (1981a). Cell. Immunol. 58, 269-277. Szigeti, R . , Stevens, D., Ernherg, I., Klein, G., Masucci, M. G . , Svedmyr, E., and Klein, E. (198lb). J. Immunol. Methods 46, 369-374. Szigeti, R., Volsky, D. J., Luka, J., and Klein, G. (1981~).J. Immunol. 126, 1676-1679. Szigeti, R . , Dalianis, T., Klein, G., Ito, Y., and Magnusson, G. (1982a). Int. J. Cancer 30, 6974. Szigeti, R., Luka, J . , Sternas, L., and Klein, G . (1982b). Int. J. Cancer 29, 413-416. Szigeti, R., Masucci, M. G., Henle, G . , Henle, W., Purtilo, D. T., and Klein, G. (1982~).Clin. Immunol. Immunopathol. 22, 128-138. Szigeti, R., Saemundsen, A. K., Synnerholm, A. C., Henle, G . , Henle, W., and Klein, G. (1983). Cancer Res. 43, 2273-2276. Szigeti, R., Masucci, G., Ehlin-Henriksson, B., Bendtzen, K., Henle, W., Henle, G . , Klein, G., and Klein, E. (1984a). Cell. Immunol. 83, 136-141. Szigeti, R., Dalianis, T., Ramqvist, T., Ito, Y., and Klein, G. (1984b). Cancer Res. 44, 10771080. Szigeti, R., Sulitzeanu, D., Henle, G . , Henle, W., and Klein, G . (1984~).Proc. Natl. Acad. Sci. U.S.A. 81, 4178-4182. Szigeti, R., Dalianis, T., Ramqvist, T., Sulitzeanu, D., and Klein, G . (1984d). To be published. Tautz, C., Khuen, F., and Ax, W. (1974). 2. Imrnunitaetsforsch. 147, 155-161. The, de, G . (1980). In “Viral Oncology” (G. Klein, ed.), pp. 769-797. Academic Press, New York. Twomey, J. T., Lazar, S., and Rocklin, R. E. (1977). Clin. Immunol. Immunopathol. 8, 396404. Vaage, J., Jones, R . D., and Brown, B. W. (1972). Cancer Res. 32, 680-687. Vose, B. M., Kimber, I., and Moore, M. (1977). J. Natl. Cancer Inst. 58, 483-488. Wahl, S. M., Iverson, G. M . , and Oppenheini, J. J. (1974). J. E x p . Med. 140, 1631-1645. Wahl, S. M., Wilton, J. M., Rosenstreich, D. L., and Oppenheini, J. J. (1975). J. Immnnol. 114, 1296-1301. Wainwright, W. H., Veltri, R. W., and Sprinkle, P. M. (1979). J. Infect. Dis. 140, 22-32. Walden, P. A. M . , Pentycross, C. R . , Leyton, R . , Browne, P., Dent, J., andVernon, P. (1976). Eur. J. Cancer 12, 277-282. Wanebo, H. J., Fukuda, M., Tsuei, L., Sarkar, N . , and Ashikari, H. (1979). Surgery 86, 275283. Weese, J. L., McCoy, J. L., Dean, J. H . , Ortaldo, J. R., Burk, K. R., and Herberman, R. B. (1978a).1. Immunol. Methods 24, 363-370. Weese, J. L., Herherman, R. B., Hollinshead, A. C., Cannon, G . B., Keels, M., and Oldham, R. K. (1978b).J. Natl. Cancer Inst. 60, 255-263. Wilson, N. I., Cochran, A. J . , Ross, C . E., MacKie, R. M., Ogg, L. J., and Todd, G. (1979). Cancer Immunol. Immunother. 6 , 27-32. Wolberg, W. H. (1971). Cancer Res. 31, 798-802. Wolberg, W. H. (1974). Arch. Surg. 109, 211-214. Wolberg, W. H., and Goelzer, M. L. (1971). Nature (London) 229, 632-634.

MIGRATION INHIBITION IN TUMOR IMMUNOLOGY

305

Wolf, G . T., Kerney, S. E., and Chretien, P. B. (1980). Am. J . Surg. 140, 531-537. Wright, G. L., Schelhammer, P. F., and Faulconer, R. J. (1977). Cancer Res. 37,4228-4232. Wright, G. L . , Schelhammer, P. F., Rosato, F. E., and Faulconer, R. J. (1978). Urology 12, 525-531. Wright, G . L., Schelhammer, P. F., Faulconer, R. J., Reid, J. W., Brassil, D. N., and Sieg, S . M. (1981). Prostate 2, 121-131. Yamamoto, H. (1980). Acta Obstet. Gynaecol. Jpn. 12, 1917-1926. Yong, W. K . , and Halliday, W. J. (1982). Br. J. Cancer 45, 754-761. Yoshida, T., Sonozaki, H., and Cohen, S. (1973).J . E r p . Med. 138, 784-797. Yoshida, T., Bigazzi, P. E., and Cohen, S. (1975).J. Zmmunol. 110, 688-691. Zoller, M., Matzku, S., and Schulz, U. (1977a).1. Natl. Cancer Inst. 58, 897-904. Zoller, M., Matzku, S., and Schulz, U. (197%). Cancer Zmmunol. Zmmunother. 2, 257-265. Zoller, M., Matzku, S., Rapp, W., and Schulz, U. (1977~).Digestion 15, 373-384. Zoller, M., Matzku, S., Schulz, U.,and Price, M. R. (1979).J . Natl. Cancer Znst. 63, 285-293. Zoller, M., Matzku, S., and Zeidler, D. (1980). Cancer Zmmunol. Zmmunother, 8, 143-148. Zoller, M., Schulz, U., Dietze, W., Wesch, H., and Matzku, S. (1983).J . Natl. CancerInst. 70, 581-587.

This Page Intentionally Left Blank

INDEX A Acute avian leukemia viruses genome structure, 210 protein products of, 210 Acute lymphocytic leukemia, leukocyte migration inhibition studies, 292 Adenoviruses, large T antigen, complexes with p53 tumor antigen in transformed cells, 115 Adipocytes, angiogenic activity, 186, 190 Angiogenesis capillary endothelial cells and, 33-36 in clinical oncology, 196- 198 induction by nonmalignant cells, 185-187 inhibition by protamine, 35 metastasis and, 33-36 methods for studying chicken embryo chorioallantoic membrane, 177-178 cloned capillary endothelial cells, 178179 corneal micropocket, 176-177 sustained-release polymer implants, 177 as preneoplastic marker, 180-181 prevascular phase, 181-182 promotion by heparin, 35 solid tumor dependence on, 181-183 tumor, see Tumor angiogenesis vascular phase, 181-183 blood supply alteration during, 183-184 Angiogenesis inhibitors from natural sources, 191-196 cartilage-derived factors, 191-192 heparin-cortisone combination, 193-196 hexasaccharide-cortisone combination,, 195-196 protamine, systemic administration, 192-193 therapeutic applications, 198

307

Angiogenic factors, 187-191 from nonneoplastic cells and tissues, 189191 tumor-derived, 188-190 Animal tumors, see also speciflc tumors migration inhibition and, 250-253 transplantable, in metastasis research, 611 selection and testing, 9-11 subpopulation differences, 8-1 1 An tibodies to myc proteins from tumor-bearing animals, 214-216 to p53 tumor antigen, titer in serum from cancer patients, 129 Antibody-dependent cellular cytotoxicity, against canine TVT, 93-94 Antibody response, to canine TVT in uitro, 90-91 Antigens DL-A, on canine TVT cell surface, 87-90, 93, 96-99, 104 Epstein-Barr virus-related, leukocyte migration inhibition studies, 293-294 on human breast carcinomas, heterogeneity, 161-166 large T adenovirus, complex with p53 tumor antigen, 115 simian virus 40, complex with p53 tumor antigen, 114-116 major histocompatibility complex class I absence on canine TVT cell surface, 98-102, 104 on various tumor cells, 100-101 class 11, on canine TVT cells, 89-90, 99, 104 p53, cellular-encoded, tumor, see Tumor antigen p53

308

INDEX

tumor-associated, effect on migration inhibition, 244-245, 295-296 Avian retroviruses oncogene-coded fusion proteins of erythroblastosis viruses, 219-224 HB-1 isolate, 217-218 MC29-related viruses CMII and OK10, 216-217 MH-2 isolate, 217-219 myeloblastosis virus, 221-224 myelocytoniatosis virus MC29, 209216 sarcoma viruses, 224-227

Basement membrane, as main harrier to cancer invasion, 36-40 Benign breast tumors, reaction with monoclonal antibody DF3 to human breast carcinoma, 151-155 Bladder cancer, leukocyte migration inhibition studies, 291 Blood-brain barrier, breakdown in tumors, angiogenesis and, 197 Breast cancer leukocyte migration inhibition studies, 254-268 tumor-specific reactions, 256-267 Breast carcinoma antigenic phenotypes, 161-166 detection by radiolabeled antibodies to, 155-160 enhancement by recombinant interferon aA, 167-169 reaction with monoclonal antibodies to, 143-171 comparison with benign tumor reaction, 151-155 immunoperoxidase assay, 148-155, 162-163 therapy with monoclonal antibodies to, 160-161 future directions, 169-171 transplanted into athymic mouse, 157159 detection by radiolabeled monoclonal antibodies, 157-158

C Cancer invasion barrier structures to, 36-38 enzymatic theory, 41-43 host defence reactions, 45-46 in uitro assay system, 38-40 mechanical pressure theory, 41 migratory theory, 41-44 sequential steps in, 40-44 tissue resistance to, 44-45 Capillaries endothelial cells angiogenesis and, 33-36 cloned for angiogenesis study, 178-179 growth by sequential steps, 178-179 in tumors during vascularization, 183184 Carbohydrates, on tumor cell surface, metastasis formation and, 12-14 Cartilage, angiogenesis inhibitors from, 191192 Cervical carcinoma, leukocyte migration inhibition studies, 290 Chemotherapy canine TVT, 81-82 metastasis drug combination and timing, 59-62 Chorioallantoic membrane, chicken embryo, angiogenesis study, 177-178 Chromosome 11, murine, p53 gene mapping in hamsterhouse hybrid cells, 120, 122-123 Coitus, canine TVT transmission, 103 Colon carcinoma detection by monoclonal antibodies, 167169 increase by recombinant interferon (Y A, 167-169 Colorectal cancer leukocyte migration inhibition studies, 277-287 tumor-specific reactions, 278-286 Corpus luteum, angiogenic activity, 186

D DNA demethylation, tumor progression and, 32-33

309

INDEX

p53 protein-controlling sequences in transformed cells, 133-139 comparison with genomic sequence in L12 cells lacking p53, 133-139 Dog, transmissible venereal tumor, 75-107 Dormant tumors angiogenesis and, 197-198 animal models, mechanisms of, 47-48 human, clinical reports, 46-47 metastasis and, 47

E Endothelial mitogens from nonneoplastic cells and tissues, 189191 tumor-derived, 188-190 Enzymes degradative extracellular matrix alteration in cancer invasion, 36-44 Epstein-Barr virus antigens, leukocyte migration inhibition studies in healthy donors, 293-294 in patients with infectious mononucleosis, 293-294 Erythrocytes, count increase in TVT-bearing dogs, 83 Erythropoietin, serum, increase in TVTbearing dogs, 83-84 Extracellular matrix tumor cell invasion and, 36-44 enzymatic degradation, 36-44

F Fusion proteins avian retroviral, 209-227 gp74erbB of erythroblastosis virus, 22022 1 p48"-qb of myeloblstosis virus, 222223 p55"-myc of MH2 virus, 219 p90Agag-*nycof MC29-related virus CMII, 216 p1Wgag-m" of MH2 virus, 219 pl08Agag-mycof HB-1 virus, 218 pll0apog-myc of myelocytomatosis virus MC29, 210-216

p135Ag**g-myb-ets of E26 erythroblastosis virus, 223-224 p2wgag-ApoCmycof MC29-related virus OK10, 216, 219 pp60src of sarcoma viruses, 224-226, 228 mammalian retroviral, 227-233 gpl2ofms of feline sarcoma virus, 231 p28Aenu-sbof simian sarcoma virus, 232 p55"-fos of murine osteosarcoma virus FBJ, 228 p7Sag-fos of murine osteosarcoma virus FBR, 228 ~ 7 5 ~ g a g - r of a f 3611 fibrosarcoma virus, 229 p7Sag-sb of feline sarcoma virus, 230 p76Agag-sb of feline sarcoma virus, 231 ~95~gQg-abl of feline sarcoma virus, 230 plW@g-abl of AbeIson murine Ieukemia virus, 227

G Gastric cancer leukocyte migration inhibition studies, 277-287 tumor-specific reactions, 278-286 Gene products of acute avian leukemia viruses, 210 myc proteins, antibodies to, 214-216 protein kinase regulation by, model, 220 Genes

P53 chromosomal mapping in hamster/mouse hybrid cells, 120, 122-123 structure in transformed cells, 133-139 Genital organs, transmissible venereal tumor, canine, 75-107 Genome acute avian leukemia viruses, 210 protein products of, 210 changes, tumor progression and, 25-33 Glucoconjugates, sialylation on tumor cell surface, metastasis formation and, 1214

H Heparin -cortisone combination, angiogenesis inhibition, 193-196

310

INDEX

tumor angiogenesis potentiation, 185, 192-193 prevention by protamine, 193 Hepatocytes, interaction with liver-metastasizing tumor cells, 12-14 Hexasaccharide-cortisone combination, angiogenesis inhibition, 195-196 Human tumors, see also specific tumors migration inhibition and, 253-296

I Immune responses to canine TVT, 86-94, 104, 105 in oitro studies, 89-94 antibody-dependent cellular cytotoxicity, 93-94 antibody response, 90-91 mixed lymphocyte culture, 89-90, 99 in oivo studies, 86-87 to metastasis antigen-specific, 51-53 tumor immunological heterogeneity and, 53-55 unspecific, 49-51 natural killer cell activity, 49-50 tumoricidal macrophage activation, 50-51 Immunodeficiencies in cancer patients leukocyte migration inhibition studies, 254 phytohemagglutinin and, 254 Immunoperoxidase assay for human breast tumors antigenic heterogeneity detection, 162163 reactions with monoclonal antibodies to, 148-1553 Immunotherapy for metastasis, new strategies, 56-59 T-cell role, 57-59 Infectious mononucleosis Epstein-Barr virus related antigens leukocyte migration inhibition studies, 293-294 Interferon aA recombinant, increase of monoclonal antibody effectiveness in detection of

human breast carcinoma, 167-169 human colon carcinoma. 167-169

K Karyotype, canine TVT, characteristics, 95-

96 Kidney, fragments, angiogenic activity, 186187

1 LAI, see Leukocyte adherence inhibition Leukocyte adherence inhibition (LAI) canine TVT assay antigen immunogenicity, 88 leukocyte immune reactivity increase, 91-92 Leukocyte migration inhibition (LMI) animal tumor responses to, 253 phytohemagglutinin-induced,in immunodeficiency assay, 254 studies of acute lymphocytic leukemia, 292 bladder cancer, 291 breast cancer, 244-268 tumor-specific reactions, 256-267 cervical carcinoma, 290 Epstein-Barr virus related antigens, 293-294 gastric and colorectal cancer, 277-287 tumor-specific reactions, 278-286 lung cancer, 287-289 malignant melanoma, 268-276 tumor-specific reactions, 270-276 ovarian carcinoma, 290 prostatic cancer, 291 renal cancer, 289-290 thyroid carcinoma, 291-292 Leukocyte migration inhibitory factor assay systems, 243-244 production and properties, 241-243 response to tumor-associated antigens, 244-245 statistical evaluation, 249-250 LMI, see Leukocyte migration inhibition Lung, metastasis from metastasis, 22 Lung cancer leukocyte migration inhibition studies, 287-289

INDEX

311

spontaneous tumor regression, 48-49 unspecific, 49-51 Macrophage migration inhibition (MMI) natural killer cell activity, 49-50 animal tumor responses to, 250-253 tumoricidal macrophage activation, Macrophage migration inhibitory factor 50-51 assay systems, 243-244 immunological heterogeneity, 53-55 production and properties, 241-243 immunoresistant variants, 52 response to tumor-associated antigens, irnmunotherapy, new strategies, 56-59 244-245 oncogenes and, 4-5, 25 statistical evaluation, 249-250 phenomenon complexity, 1-4 Macrophages research at different levels, 4-6 activated, angiogenesis induction, 185 subpopulation interactions, 24-27 antimetastatic activity, 50-51 tumor dormancy and, 47 peritoneal, see Peritoneal macrophages Microenvironment Malignant melanoma signals in tumors leukocyte migration inhibition studies, genomic rearrangements and, 26-32 268-276 phenotypic shifts and, 26-32 tumor-specific reactions, 270-276 Pz-Microglobulin Mammalian retroviruses absence on canine TVT cell surface, 97oncogene-coded fusion proteins of 102 Abelson murine leukemia viruses, 227-228 on various tumor cells, 100-101 feline sarcoma viruses, 229-232 Migration inhibition 3611 fibrosarcoma virus, 228-229 animal tumors and, 250-253 murine osteosarcoma viruses, 228 assay systems, 243-247 simian sarcoma virus, 232-233 basic phenomenon, 241-243 Mast cells cancer patients’ sera effects on, 294-295 heparin release, 185 human tumors and, 253-296 tumor angiogenesis potentiation, 184-185 immunodeficiencies and, 254 Metastasis of leukocytes, see Leukocyte migration angiogenesis and, 33-36, 197 inhibition animal model, 6-11 of macrophages, see Macrophage migrasubpopulation differences, 8-11 tion inhibition cascade theory technique, 241-249 clinical data, 14-18 agarose microdroplet assay, 248-249 major pathways, 15-17 agarose plate test, 248 experimental data capillary tube, 247-248 extravasion of tumor cells, 21 tumor-associated antigen detection, 244host property effects, 22-23 245, 295-296 metastasis from metastasis, 22 Mixed lymphocyte tumor cell culture, antiorgan site predilection, 20-21 gens on canine TVT cells, 89-90, 99 primary tumor effects, 22-23 MMI, see Macrophage migration inhibition sequential steps, 19 Monoclonal antibodies tumor cell inefficiency, 18, 20 antigen-specific response to metastasis, 53 cell surface properties and, 12-14 to human breast carcinomas, 143-171 chemotherapy applications, 143-144 drug combination and timing, 59-62 DF3, reactivity with malignant and beclonal instability, 23-25 nign breast tumors, 151-155 conclusions, 63-64 future directions, 169-171 host immune response to, 48-55 generation and characterization, 144antigen-specific, 51-53 151

312

INDEX

radiolabeled, tumor detection, 155-160 enhancement by recombinant interferon aA, 167-169 tumor antigenic heterogeneity and, 161-166 tumor therapy, 160-161 to p53 tumor antigen, production and properties, 118-120 Mouse athymic bearing human mammary tumors, 157-159 tumor localization by radiolabeled monoclonal antibodies, 157-159 Murine leukemia virus -transformed L12 cell line DNA genomic sequence, 133-139 TPA-induced p53 synthesis, 132-133 tumor antigen p53 absence, 130-132

N Natural killer cells, effect on metastasis, 4950 Neutrophils migration inhibition, 241-243 tumor antigen detection by, 243-249 statistical evaluation, 249-250 Nonneoplastic cells angiogenic factors from 189-191 endothelial mitogens from, 189-191 Nontransformed cells p53 tumor antigen detection, 116-117, 121 subcellular localization, 123-124, 126127 Nuclei, p53 localization in virus-transformed cells, 123-127

0 Oncogenes avian retroviral, 206, 209-227 transforming proteins coded by, 206, see also Fusion proteins, avian retroviral mammalian retroviral, 207, 227-233 transforming proteins coded by, 207, see also Fusion proteins, mammalian retroviral

metastasis formation and, 4-5, 25 retroviral, mechanism of action, 208-209 Ovarian carcinoma, leukocyte migration inhibition studies, 290

P Peritoneal macrophages migration inhibition, 241-243 tumor antigen detection by, 243-249 statistical evaluation, 249-250 Phosphorylation, p53 tumor antigen in complexes with viral large T antigens, 116 Phytohemagglutinin leukocyte migration inhibition by, immunodeficiency assay, 254 Polycythemia, canine TVT-induced, 82-84, 105 Polymers, sustained-release implants, angiogenesis study, 177 Primary tumors invasive capacity, host effects, 22-23 metastasis inhibition, 22 Prostatic cancer, leukocyte migration inhibition study, 291 Protamine angiogenesis inhibition, systemic administration, 192- 193 Protein kinase regulation by oncogene products, model, 220 retroviral fusion protein activities of, 219222, 224-234 in retroviral transformation mechanism, model, 233 Proteins fusion retroviral, see Fusion proteins structural retroviral, cell transformation and, 234-235

R Radiotherapy, canine TVT, 81 Renal cancer, leukocyte migration inhibition studies, 289-290 Retina, angiogenic activity, 187, 190 Retroviruses avian, see Avian retroviruses cell transformation fusion proteins and, 205-235 structural proteins and, 234-235

313

INDEX

mammalian, see Mammalian retroviruses oncogenes, see Oncogenes

RNA messenger (mRNA) p53-specific, structure in transformed cells, 133-135

S Sialic acid, on tumor cell surface, metastasis formation and, 12-14 Simian virus 40 large T antigen complexes with p53 tumor antigen in transformed cells, 114-117

in several mammalian species, 121 as specific marker for, 127-129 subcellular localization, 123- 127 Transmissible venereal tumor (TW), canine cell markers, 80 cell surface, absence of class I of major histocompatibility complex antigen, 98-102, 104 P~-macroglobulin,97-102 cellular transmission, 94-97, 103-104 cytogenetic studies, 94-96 immunological studies, 96-97 characteristics, products, interactions with host organs, scheme, 106-107 clinical diagnosis, 80-81 DL-A antigens on cell surface, 87-90, 93,

96-99, 104

T T cells cytotoxic, antigen-specific response to metastasis, 51-53 lymphokine-producing, angiogenic activity, 186 in metastasis immunotherapy, 57-59 Testis, fragments, angiogenic activity, 186 12-0-Tetradecanoylphorbol-13-acetate (TPA), p53 tumor antigen induction in murine leukemia virus-transformed L12 cells, 132-133 Thyroid carcinoma, leukocyte migration inhibition studies, 291-292 TF'A, see 12-O-Tetradecanoylphorbol-13acetate Transformation neoplastic, angiogenesis as preneoplastic marker, 180-181 retroviral fusion proteins and, 205-235 models, 217, 233 structural protein effects, 234-235 Transformed cells p53 tumor antigen absence in murine leukemia virus-transformed L12 line, 130-132 induction by prolonged tumor promoter treatment, 132-133 complexes with viral-encoded protein,

114-117 functions, 130-133

histology, 78-79 immune responses to, 86-94 in uitro studies, 89-94 in uiuo studies, 86-87 karyotype, characteristics, 95-96 natural occurrence, 75-78 polycythemia induction, 82-84, 105 therapy, 81-82 transmission by coitus, 103 transplan tation intraperitoneal, 85-86 intravenous, 85 subcutaneous, 84-86 ultrastructure, 79-80 Transplantation canine TVT intraperitoneal, 85-86 intravenous, 85 subcutaneous, 84-86 human mammary tumor into athymic mice, detection, 157-159 Tumor angiogenesis angiogenesis inhibitors from natural sources and, 191-196, 198 potentiation by heparin, 185, 192-193 prevention by protamine, 193 mast cells, 184-185 Tumor antigen p53, 113-130 absence in murine leukemia virus transformed L12 line, 130-132 induction by prolonged tumor promoter treatment, 132-133

314

INDEX

immunogenicity, 117- 120 as marker for cancer cells, 127-129 anti-p53 antibody titer, 129 immunofluorescence, 129 immunoprecipitation, 127-128 in nontransformed cells detection, 116-117, 121 subcellular localization, 123-124, 126127 overproduction in various tumors, 113-114 in transformed cells complexes with viral-encoded protein, 114-116 functions, 130-133 genomic DNA and, 133-139 in several mammalian species, 121 subcellular localization, 123-127 Tumor cells angiogenic factors from, 187-190 circulating extravasion in metastasis formation, 21 metastatic inefficiency, 18, 20 organ site predilection for metastasis, 20-21 class I of major histocompatibility complex antigen on surface, 100-101 endothelial mitogens from, 187-190 fusion with host cells, tumor progression and, 30-31 immunological heterogeneity, 53-55 liver-metastasizing, interaction with hepatocytes, 12-14

locomotion, cancer invasion and, 43-44 Pz-microglobulin on surface, 100-101 surface carbohydrates in metastasis formation, 12-14 Tumor immunology migration inhibition and, 241-296 animal tumor detection, 250-253 assay systems, 143-247 basic phenomenon, 241-243 human tumor detection, 253-296 technique, 247-249 Tumors animal, see Animal tumors human, see Human tumors metastatic capacity, heterogeneity, 2325 phenotypic shifts in metastatic properties, 27-30 primary, see Primary tumors progression mechanisms, 25-33 DNA demethylation, 32-33 fusion between tumor and host cells, 30-31 genetic factors, 25-33 microenvironment, 26-32 regression by angiogenesis inhibition, 194 spontaneous, 48-49 solid, angiogenesis-dependent, 181-183 TPA, see 12-O-Tetradecanoylphorbol-13acetate TVT, see Transmissible venereal tumor

CONTENTS OF RECENT VOLUMES Volume 20 Tumor Cell Surfaces: General Alterations Detected by Agglutinins Annette M. C. Rapin and M a x M. Burger Principles of Immunological Tolerance and Immunocyte Receptor Blockade G . 1. V. Nossal The Role of Macrophages in Defense against Neoplastic Disease

Michael H. Levy and E. Frederick Wheelock Epoxides in Polycyclic Aromatic Hydrocarbon Metabolism and Carcinogenesis P . Sims and P . L. Grover Virion and Tumor Cell Antigens of C-Type RNA Tumor Viruses

Heinz Bauer Addendum to “Molecular Repair, Wound Healing, and Carcinogenesis: Tumor Production a Possible Overhealing?”

Sir Alexander Haddow

Selection of Biochemically Variant, in Some Cases Mutant, Mammalian Cells in Culture G . B. Clements The Role of DNA Repair and Somatic Mutation in Carcinogenesis James E. Trosko and Ernest H. Y . Chu SUBJECT INDEX

Volume 22 Renal Carcinogenesis J. M . Hamilton Toxicity of Antineoplastic Agents in Man: Chromosomal Aberrations, Antifertility Effects, Congenital Malformations, and Carcinogenic Potential Susan M . Sieber and Richard H .

Adamson Interrelationships among RNA Tumor Viruses and Host Cells

Raymond V. Gilden

SUBJECT INDEX

Proteolytic Enzymes, Cell Surface Changes, and Viral Transformation

Richard Roblin, lih-Nan Chou, and Paul H. Black

Volume 21 Lung Tumors in Mice: Application to Carcinogenesis Bioassay Michael B . Shimkin and Gary D. Stoner Cell Death in Normal and Malignant Tissues

E. H. Cooper, A . J. Bedford, and T. E. Kenny The Histocompatibility-Linked Immune Response Genes

Immunodepression and Malignancy

Osias Stuttnun SUBJECT I N D EX

Volume 23 The Genetic Aspects of Human Cancer

W . E. Heston

Baruj Benacerraf and David H . Katz Horizontally and Vertically Transmitted Oncornaviruses of Cats

M. Essex Epithelial Cells: Growth in Culture of Normal and Neoplastic Forms

The Structure and Function of Intercellular Junctions in Cancer Ronald S. Weinstein, Frederick B. Merk,

and Ioseph Alroy Genetics of Adenoviruses

Harold S . Ginsberg and C. S . H . Young

Keef A . Rafferty, ]r. 315

316

CONTENTS OF RECENT VOLUMES

Molecular Biology of the Carcinogen, 4Nitroquinoline 1-Oxide

Minako Nagao and Takashi Sugimura Epstein-Barr Virus and Nonhuman Primates: Natural and Experimental Infection A . Frank, W . A . Andiman, and G . Miller Tumor Progression and Homeostasis

Richmond T . Prehn Genetic Transformation of Animal Cells with Viral DNA or RNA Tumor Viruses

Miroslav Hill and Jana Hillova SUBJECT INDEX

Volume 24 The Murine Sarcoma Virus-Induced Tumor: Exception or General Model in Tumor Immunology? J . P. Levy a n d ] . C . Leclerc Organization of the Genomes of Polyoma Virus and SV40 Mike Fried and Beverly E . Grflin &-Microglobulin and the Major Histocompatibility Complex Per A . Peterson, Lars Rask, and Lars

Ostberg Chromosomal Abnormalities and Their Specificity in Human Neoplasms: An Assessment of Recent Observations by Banding Techniques

Joachim Mark Temperature-Sensitive Mutations in Animal Cells

Claudio Basilico Current Concepts of the Biology of Human Cutaneous Malignant Melanoma Wallace H. Clark, J r . , Michael 3. Mastrangelo, Ann M . Ainsworth, David Berd, Robert E . Eellet, and Evelino

A . Bernardino

Malignancy and Transformation: Expression in Somatic Cell Hybrids and Variants Harvey L. Ozer and Krishna K . ]ha Tumor-Bound Immunoglobulins: In Situ Expressions of Humoral Immunity

lsaac P. Witz The A h Locus and the Metabolism of Chemical Carcinogens and Other Foreign Compounds Snorri S . Thorgeirsson and Daniel W .

Nebert Formation and Metabolism of Alkylated Nucleosides: Possible Role in Carcinogenesis by Nitroso Compounds and Alkylating Agents

Anthony E. Pegg Immunosuppression and the Rule of Suppressive Factors in Cancer

lsao Kamo and Nerman Friednian Passive Immunotherapy of Cancer in Animals and Man

Steven A . Rosenberg and William D . Terry SUBJECT INUEX

Volume 26 The Epidemiology of Large-Bowel Cancer

Pelayo Correa and William Hoenszel Interaction between Viral and Genetic Factors in Murine Mammary Cancer J . Hilgers and P. Bentvelzen Inhibitors of Chemical Carcinogenesis

Lee W . Wattenberg Latent Characteristics of Selected Herpesviruses Jack G . Stevens Antitumor Activity of Corynebacteriuin

parvum Luka Milas and Martin T . Scott SUBJECT INUEX

SUBJECT INDEX

Volume 27 Volume 25 Biological Activity of Tumor Virus DNA F. L. Graham

Translational Products of Type-C RNA Tumor Viruses John R . Stephenson, Sushikumar G . Devare, and Fred H . Reynolds, J r .

CONTENTS OF RECENT VOLUMES

Quantitative Theories of Oncogenesis

Alice S. Whittemore Gestational Trophohlastic Disease: Origin of Choriocarcinoma, Invasive Mole and Choriocarcinoma Associated with Hydatidiform Mole, and Some Immunologic Aspects J . 1. Brewer, E . E . Torok, B . D. Kahan, C . R. Stanhope, and B . Halpern The Choice of Animal Tumors for Experimental Studies of Cancer Therapy

Harold B . Hewitt Mass Spectrometry in Cancer Research

John Roboz Marrow Transplantation in the Treatment of Acute Leukemia E . Donna11 Thomas,C . Dean Buckner,

Alexander Fefer, Paul E. Neiman, and Rainer Storb Susceptibility of Human Population Groups to Colon Cancer

Martin Lipkin Natural Cell-Mediated Immunity Ronald B . Herbermon and Howard T .

Holden SUBJECT INDEX

317

Kettmunn, M . Leckrcq, J . Leunen, M . Mammerickx, and D. Poi-tetelle Molecular Mechanisms of Steroid Hormone Action

Stephen J . Higgins and Vlrich Gehring SUBJECT INDEX

Volume 29 Influence of the Major Histocompatibility Complex on T-cell Activation J . F. A. P. Miller Suppressor Cells: Permitters and Promoters of Malignancy?

David Naor Retrodifferentiation and the Fetal Patterns of Gene Expression in Cancer Josd Uriel The Role of Glutathione and Glutathione STransferases in the Metabolism of Chemical Carcinogens and Other Electrophilic Agents

L. F. Chasseaud a-Fetoprotein in Cancer and Fetal Development

Erkki Ruoslahti and Markku Seppiili Mammary Tumor Viruses

Volume 28 Cancer: Somatic-Genetic Considerations F. M . Burnet Tumors Arising in Organ Transplant Recipients

Dan H. Moore, Carole A . Long, Akhil B . Vaidya, Joel B. Sheffwld, Arnold S. Dion, and Etienne Y. Lasfargues Role of Selenium in the Chemopreventioo of Cancer

A. Clark Griffin SUBJECT INDEX

lsrael Penn Structure and Morphogenesis of Type-C Retroviruses Ronald C . Montelaro and Dani P.

Bolognesi BCG in Tumor Immunotherapy Robert W . Baldwin and Malcolm V . Pimm The Biology of Cancer Invasion and Metastasis lsaiuh J . Fidler, Douglas M . Gersten, and lan R. Hart Bovine Leukemia Virus Involvement in Enzootic Bovine Leukosis A. Burny, F . Bex, H . Chantrenne, Y.

Cleuter, D. Dekegel, J . Ghysdael, R.

Volume 30 Acute Phase Reactant Proteins in Cancer E. H . Cooper a n d l o a n Stone Induction of Leukemia in Mice by Irradiation and Radiation Leukemia Virus Variants

Nechama Harana-Ghera and Alpha Peled On the Multiform Relationships between the Tumor and the Host

v. s. Shapot

Role of Hydrazine in Carcinogenesis

Joseph Bald

318

CONTENTS OF RECENT VOLUMES

Experimental Intestinal Cancer Research with Special Reference to Human Pathology

Kazymir M . Pozharisski, Alexei]. Likhaocheo, Valeri F. Klimasheoski, and lacob D. Shaposhnikov The Molecular Biology of Lymphotropic Herpesviruses

Bill S u g h n , Christopher R . Kintner, and Willie Mark

and RNA-Containing Viruses: Properties of the Solubilized Antigens

Uoyd W . Law, Michael]. Rogers, and Ettore Appella Nutrition and Its Relationship to Cancer

Bandoru S . Reddy, Leonard A . Cohen. G . David McCoy, Peter Hill, ]ohn H . Weisburger, and Ernst L. Wynder INDEX

Viral Xenogenization of Intact Tumor Cells

Hiroshi Kobayashi Virus Augmentation of the Antigenicity of Tumor Cell Extracts Faye C. Austin and Charles W . Boone INDEX

Volume 31 The Epidemiology of Leukemia

Michael Alderson The Role of the Major Histocompatibility Gene Complex in Murine Cytotoxic T Cell Responses

Hennann Wagner, Klaus Pfizenmaier. and Martin Rollinghoff The Sequential Analysis of Cancer Development

Enlmanuel Farber and Ross Cameron Genetic Control of Natural Cytotoxicity and Hybrid Resistance

Edward A . Clark and Richard C. Harmon Development of Human Breast Cancer Sefton R . Wellings

Volume 33 The Cultivation of Animal Cells in the Chemostat: Application to the Study of Tumor Cell Multiplication Michael G . Tooey Ectopic Hormone Production Viewed as an Abnormality in Regulation of Gene Expression

Hiroo lmuru The Role of Viruses in Human Tumors

Harald zur Hausen The Oncogenic Function of Maininalian Sarcoma Viruses

Poul Andersson Recent Progress in Research on Esophageal Cancer in China Li Mingxin (Li Min-Nsin), Li Ping, u n d Li

Baorong (Li Pao-lung) Mass Transport in Tumors: Characterization and Applications to Chemotherapy

Rakesh K. lain, ]onas M. Weissbrod, und lames Wei INDEX

INDEX

Volume 34 Volume 32 Tumor Promoters and the Mechanism of Tumor Promotion Leila Diamond, Thomas G . O’Brien, and

William M . Baird Shedding from the Cell Surface of Normal and Cancer Cells

Paul H . Black Tumor Antigens on Neoplasms Induced by Chemical Carcinogens and by DNA-

The Transformation of Cell Growth and Transmogrification of DNA Synthesis by Simian Virus 40

Robert G .Martin Immunologic Mechanisms in UV Radiation Carcinogenesis

Margaret L. Kripke The Tumor Dormant State

E . Federick Wheelock, Kent 1. Weinhold, and Judith Leoich

CONTENTS OF RECENT VOLUMES

Marker Chromosome 14q- in Human Cancer and Leukemia Felix Mitelman Structural Diversity among Retroviral Gene Products: A Molecular Approach to the Study of Biological Function through Structural Variability James W. Gautsch, John H. Elder, Fred C. Jensen, and Richard A. Lerner Teratocarcinomas and Other Neoplasms as Developmental Defects in Gene Expression Beatrice Mintz and Roger A. Fleischman Immune Deficiency Predisposing to Epstein-Barr Virus-Induced Lymphoproliferative Diseases: The X-Linked Lymphoproliferative Syndrome as a Model David T. Purtilo INDEX

Volume 35 Polyoma T Antigens Walter Eckhart Transformation Induced by Herpes Simplex Virus: A Potentially Novel Type of Virus-Cell Interaction Berge Hampar Arachidonic Acid Transformation and Tumor Production Lawrence Zmine The Shope Papilloma-Carcinoma Complex of Rabbits: A Model System of Neoplastic Progression and Spontaneous Regression John W. Kreider and Gerald L. Bartlett Regulation of SV40 Gene Expression Adolf Graessman, Monika Graessmann, and Christian Mueller Polyamines in Mammalian Tumors, Part I Giuseppe Scalabrino and Maria E. Feriolo Criteria for Analyzing Interactions between Biologically Active Agents Morris C. Berenbaum INDEX

319

Volume 36 Polyamines in Mammalian Tumors, Part I1 Giuseppe Scalabrino and Maria E. Ferioli Chromosome Abnormalities in Malignant Hematologic Diseases Janet D. Rowley andJoseph R. Testa Oncogenes of Spontaneous and Chemically Induced Tumors Robert A. Weinberg Relationship of DNA Tertiary and Quaternary Structure of Carcinogenic Processes Philip D. Lipetz, Alan G. Galsky, and Ralph E. Stephens Human B-Cell Neoplasms in Relation to Normal B-Cell Differentiation and Maturation Processes Tore Coda1 and Steinar Funderud Evolution in the Treatment Strategy of Hodgkin’s Disease Gianni Bonadonna and Armando Santoro Epstein-Barr Virus Antigens-A Challenge to Modern Biochemistry David A. Thorley-Lawson, Clark M. Edson, and Kathi Geilinger INDEX

Volume 37 Retroviruses and Cancer Genes J. Michael Bishop Cancer, Genes, and Development: The Drosophila Case Elbabeth Gateff Transformation-Associated Tumor Antigens Arnold J. Levine Pericellular Matrix in Malignant Transformation Kari Alitalo and Antti Vaheri Radiation Oncogenesis in Cell Culture Carmia Borek Mhc Restriction and I r Genes Jan Klein and Zoltan A. Nagy Phenotypic and Cytogenetic Characteristics of Human B-Lymphoid Cell Lines and Their Relevance for the Etiology of Burkitt’s Lymphoma Kenneth Nilsson and George Klein

320

CONTENTS OF RECENT VOLUMES

Translocations Involving Ig Locus-Carrying Chromosomes: A Model for Genetic Transposition in Carcinogenesis

George Klein and Gilbert Lenoir INDEX

Volume 38 The SJL/J Spontaneous Reticulum Cell Sarcoma: New Insights in the Fields of Neoantigens, Host-Tumor Interactions, and Regulation of Tumor Growth

Benjamin Bonavida The Initiation of DNA Excision-Repair

Chromosomes and Cancer in the Mouse: Studies in Tumors, Established Cell Lines, and Cell Hybrids Dorothy A . Miller and Orlando J . Miller Polyomarvirus: An Overview of Its Unique Properties Beoerly E . Griffin and Stephen M.

Dilworth The Pathogenesis of Oncogenic Avian Retroviruses Paula J . Enrietto and John A . Wyke Adjuvant Chemotherapy for Common Solid Tumors

David A . Berstock and Michael Bautn INDEX

George W . Teebor and Krystyna Frenkel Steroid Hormone Receptors in Human Breast Cancer George W.Sledge, Jr. and William L.

Volume 40

McGuire Relation between Steroid Metabolism of the Host and Genesis of Cancers of the Breast, Uterine Cervix, and Endometrium Mitsuo Kodoma and Toshiko Kodania Fundamentals of Chemotherapy of Myeloid Leukemia by Induction of Leukemia Cell Differentiation

Motoo Hozumi The in Vitro Generation of Effector Lymphocytes and Their Employment in Tumor Immunotherapy Eli Kedar and David W . Weiss Cell Surface Glycolipids and Glycoproteins in Malignant Transformation

G . Yogeeswaran INDEX

Volume 39 Neoplastic Development in Airway Epithelium

5-Methylcytosine, Gene Regulation, and Cancer Arthur D. Riggs and Peter A. Jones Immunobiology of Infection with Human Cytomegalovirus

H . Kirchner Genetics of Resistance to Virus-Induced Leukemias

Daniel Meruelo and Richard Bach Breast Carcinoma Etiological Factors Dan H . Moore, Dan H . Moore 11, and

Cathleen T. Moore Treatment of Actue Leukemia-Advance\ i n Chemotherapy, Immunotherapy, and Bone Marrow Transplantation

Costa Gahrton The Forty-Year-Old Mutation Theory of Luria and Delbruck and Its Pertinence to Cancer Chemotherapy

Howard E . Skipper Carcinogenesis and Aging Vlodimir N . Anisimoo INDEX

P. Nettesheim and A . Marchok Concomitant Tumor Immunity and the Resistance to a Second Tumor Challenge

E. Corelik Antigenic Tumor Cell Variants Obtained with Mutagens

Thiery Boon

Volume 41 The Epidemiology of Diet and Cancer Tim Byers and Saxon Graham

CONTENTS OF RECENT VOLUMES

Molecular Aspects of Immunoglobulin Expression by Human B Cell Leukemias and Lymphomas

John Gordon Mouse Mammary Tumor Virus: Transcriptional Control and Involvement in Tumorigenesis Nancy E . Hynes, Bernd Groner, and Rob

Michalides Dominant Susceptibility to Cancer in Man

David Hamden, John Morten, and Terry Featherstone Multiple Myeloma, Waldenstrom’s Macroglobulinemia, and Benign Monoclonal Gammopathy: Characteristics of the B Cell Clone, Immunoregulatory Cell Populations and Clinical Implications

Hdkan Mellstedt, Goran Holm, and Magnus Bjorkholm Idiotype Network Interactions in Tumor Immunity

Hans Schreiber Chromosomal Location of Immunoglobulin Genes: Partial Mapping of These Genes in the Rabbit and Comparison with Ig Genes Carrying Chromosomes of Man and Mouse

Leandro Medrano and Bernard Dutrillauw INDEX

321

Volume 42 Immunological Surveillance of Tumors in the Context of Major Histocompatibility Complex Restriction of T Cell Function

Peter C. Doherty, Barbara B. Knowles, and Peter 1. Wettstein Immunohistological Analysis of Human Lymphoma: Correlation of Histological and Immunological Categories Harald Stein, Karl Lennert, Alfred C . Feller, and Dauid Y . Mason Induced Differentiation of Murine Erythroleukemia Cells: Cellular and Molecular Mechanisms

Richard A . Aifiind, Michael S h e f f e y , and Paul A . Marks Protoneoplasia: The Molecular Biology of Murine Mammary Hyperplasia

Robert D. CardijJ Xiphophorus as an in Vivo Model for Studies on Normal and Defective Control of Oncogenes

Fritz Anders, Manfred Schartl, Angelika Barnekow, and Annerose Anders Contrasuppression: The Second Law of Thymodynamics, Revisited

Douglas R. Green and Richard K. Gershon INDEX

This Page Intentionally Left Blank