Advances in Cancer Research Volume 32

ADVANCESINCANCER RESEARCH VOLUME 32

Contributors to This Volume Ettore Appella

Lloyd W. Law

William M. Baird

G. David McCoy

Paul H. Black

Thomas G. O’Brien

Leonard A. Cohen

Bandaru S.Reddy

Leila Diamond

Michael J. Rogers

Peter Hill

John H. Weisburger Ernst L. Wynder

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 32-1 980 ACADEMIC PRESS A Subsidiary of Harcourt Brace Jovanovich, Publishers

w

New York

London Toronto

Sydney San Francisco

COPYRIGHT @ 1980, 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.

111 Fifth Avenue, New York, New York 10003

United Kingdom Edition published b y ACADEMIC PRESS, INC. (LONDON) LTD. 24/28 Oval Road, London N W l 7 D X

LIBRARY OF

CONGRESS CATALOG CARD

NUMBER:52-13360

ISBN 0-12-006632-7 PRINTED IN THE UNITED STATES OF AMERICA 80 81 82 83

9 8 7 6 5 4 3 2 1

CONTENTS CONTRIBUTORS TO VOLUME 32 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

vii

Tumor Promoters and the Mechanism of Tumor Promotion LEILADIAMOND, THOMAS G. O'BRIEN,A N D WILLIAM M. BAIRD I . Introduction 111.

IV. V. VI. VII.

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

ogenesis . . . . . . . . . . . . . . . . . . Chemistry of Tumor Promoters.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Experimental Models of Tumor Promotion.. . . . Effects of Tumor Promoters on Cells in Culture .......................... Speculation on the Biochemical Mechanisms of Tumor Promotion.. . . . . . . . . Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . .............................

I 4 6 16

32 54 62 63

Shedding from the Cell Surface of Normal and Cancer Cells PAULH . BLACK I. 11. 111. IV. V. VI . VII. VIII.

IX. X. XI. XII.

Introduction ........................... Membrane S .................................................. Synthesis and Intracellular Translocation of Released Proteins . . . . . . . . . . . . . Shedding and Membrane Protein Turnover . . . . . . . . . . . . . . . . . Shedding and the Activated State . . . . Shedding and Activated Specific Cells ......... Mechanism of Shedding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Consequences of Shedding from the Cancer Cell Surface Shedding and Chronic Viral Disease ........... . . . . . . . . . . . . . . . . . . . . . . . . . Activation and Surface Proteases: Chemotaxis and Rheumatoid Arthritis.. . . Prevention of Shedding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . ...

76 78

85

125 144

172 174 178 179 180

Tumor Antigens on Neoplasms Induced by Chemical Carcinogens and by DNA- and RNA-Containing Viruses: Properties of the Solubilized Antigens LLOYDW. LAW,MICHAEL J. ROGERS,A N D ETTOREAPPELLA I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. Soluble Antigens from Chemically Induced Tumors.. ..................... V

201 203

vi

CONTENTS

111. TATA of Neoplasms Induced by the DNA Oncogenic Viruses, SV40 and Polyoma Virus ........................................................ IV. TATA of Neoplasms (Leukemias) Induced by RNA Tumor Viruses . . . . . . . . V. Soluble Antigens and Immune Deviations of the H o s t . . . . . . . . . . . . . . . . . . . . . VI. Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

216 219 225 229 232

Nutrition and Its Relationship to Cancer BANDARU S. REDDY,LEONARD A. COHEN,G. DAVIDMcCoy, PETERHILL,JOHNH. WEISBURGER, A N D ERNSTL . WYNDER 1. 11. 111. IV. V. VI. VII. VIII.

Introduction .......................................................... Dietary Factors and Cancer of the Large Bowel . . . . . . . . . . . . . . . . . . . . . . . . . . Dietary Factors and Cancer of the Stomach.. . Dietary Factors and Cancer of the Upper A Dietary Factors and Cancer of the Pancreas Dietary Factors and Cancer of the Breast.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Dietary Factors and Cancer of the Prostate Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

238 24 I 27 I 282 29 I 295 324 329 332

SUBJECT INDEX. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347 CONTENTSOF PREVIOUS VOLUMES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353

CONTRIBUTORS TO VOLUME 32

Numbers in parentheses indicate the pages on which the authors’ contributions begin.

ETTORE APPELLA, Laboratory of Cell Biology, National Cancer Institute, Bethesda, Maryland 20205 (201) WILLIAM M. BAIRD,The Wistar Institute of Anatomy and Biology, Philadelphia, Pennsylvania 19/04 ( I ) PAUL H . BLACK,Hubert H . Humphrey Cancer Research Center and Department of Microbiology, Boston University School of Medicine, Boston, Massachusetts 021 18 (75) LEONARDA . COHEN,Naylor Dana Institute f o r Disease Prevention, American Health Foundation, Valhalla, N e w York 10595 (237) LEILADIAMOND, The Wistar Institute of Anatomy and Biology, Philadelphia, Pennsylvania 19/04 ( 1 ) PETER H I L L , NLiyIor Dana InJtitute f o r Diseuse Prei’ention, American Heulth Foundtition, Vulhallu, N e w York 10595 (237) LLOYDW. LAW,Laboratory of Cell Biology, National Cancer Institute, Bethesda, Maryland 20205 (20 I ) G. DAVIDM c C o y , Naylor Dana Institute f o r Disease Prevention, American Heulth Foundation, Valhalla, N e w York 10595 (237) THOMASG. O’BRIEN,The Wistar Institute of Anatomy and Biology, Philadelphia, Pennsylvaniu 19/04 ( 1 ) BANDARUS. REDDY,Naylor Dana Institute f o r Diseuse Prevention, American Health Foundation, Valhalla, N e w York 10595 (237) MICHAELJ . ROGERS,Laboratory of Cell Biology, National Cancer Institute, Bethesda, Maryland 20205 (201) JOHN H . WEISBURGER, Naylor Dana Institute f o r Disease Prevention, American Health Foundation, Valhalla, N e w York 10595 (237) ERNST L. WYNDER,Naylor Dana Institute f o r Disease Prevention, American Health Foundation, Valhalla, N e w York 10595 (237)

This Page Intentionally Left Blank

ADVANCESINCANCER RESEARCH VOLUME 32

This Page Intentionally Left Blank

TUMOR PROMOTERS AND THE MECHANISM OF TUMOR PROMOTION

Leila Diamond, Thomas G. O’Brien, and William M. Baird The Wistar Institute of Anatomy and Biology. Philadelphia. Pennsylvania

I.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

..

11. Concept of Two-Stage Carcinogenesis 111. Chemistry of Tumor Promoters . . . . . . . . .

IV.

V.

VI.

VII.

..

A. Identification of Promoting Components in Croton Oil . . . . . . . . . . . . . . . . . B. Phorbol Diester Tumor Promoters . . , , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Cellular Interaction and Metabolism of Phorbol Esters . . . . . . . . . . . . . . . . . D. Tumor Promoters from Other Plants . . . . . . , . . . . . . . . . . . . . . . . . . . . . . . . . . E. Other Mouse Skin Tumor Promoters . . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F. Tumor Promoters in Tissues Other than S k in .. . . . Experimental Models of Tumor Prom A. Mouse Skin Model . . . . . . . . . . . . . ........................... B. Other Models of Two-Stage Carci . . . . . . . .. Effects of Tumor Promoters o n Cells in A. I n Vitro Models of Two-Stage Carcinogenesis . . . . . . . . . . . . . . . . . . . . . . . . . B. Effects of Phorbol Diester Tumor Promoters on Cell Morphology . . . . . . . C. Biochemical Effects of Phorbol Diester Tumor Promoters . . . . . . . . . D. Other Biological Effects of Phorbol Diester Tumor Promoters . . . . . . . . . . E. Effects of Other Tumor Promoters . . . . . . . . . . . . . . . . . . . . . . . . . . Speculation on the Biochemical Mechanisms of Tumor Promotion . . . . . . . . . . . A. Membrane Interactions . . . . , , . . . . . . . . . . . . . . . . . . . . . . . B. Altered Gene Expression ....................... ................... ........... C. Altered Cell Differentiation . . . . . . . . . . . . . D. Relation between Tumor Promotion and Viral Carcinogenesis . . . . . . . . . . . ................ Perspectives . . . . . . . . . . . . . . . . . . . . . . .

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

1 4 6 6 7 12 13 14 16 16 16 28 32 33 35 38 45 53 54 56 56 59 60 62 63

I. Introduction Current epidemiologic studies are providing evidence that cancer in humans can result from an interaction of multiple factors. One of the most striking examples of this is the 10-fold higher lung cancer mortality among asbestos workers who were cigarette smokers than among asbestos workers who were not smokers (Selikoff and Hammond, 1975). There is also epidemiologic evidence that there is a reversible stage in the induction of human cancer. This is suggested, for example, by the fact that the risk of lung cancer is decreased in exsmokers compared to the I ADVANCES I N CANCER RESEARCH, VOL. 32

Copyright @ 1980 by Academic Press. Inc. All rights of reproduction in any form reserved ISBN 0-12-006632-7

2

LEILA DIAMOND

et al.

risk in those who continue to smoke, and that the risk decreases with the number of years the exsmoker has not smoked (Wynder and Stellman, 1977; Doll, 1978). Multifactorial and multistage model systems of cancer induction in experimental animals were first described over 40 years ago (reviewed in Boutwell, 1964; Van Duuren, 1969). In one of these systems, it was shown that tumor induction can be divided into two distinct treatment stages, initiation and promotion, and that the promotion stage is reversible (reviewed in Boutwell, 1964). Compounds with tumor-promoting activity are a class of cocarcinogens that are not themselves mutagenic or carcinogenic but that cause the formation of tumors when applied repeatedly after an irreversible event has been "initiated" by a carcinogen (reviewed in Boutwell, 1964). Because promotion requires repeated treatment with a promoter and is a reversible process, an understanding of the process is fundamental to developing procedures and strategies for controlling the incidence of cancer in humans. Conclusive evidence for two-stage carcinogenesis was obtained first in mouse skin (Mottram, 1944), and this has been the primary model for investigating the mechanism of tumor promotion. Originally, one of the main advantages of this model was the fact that croton oil was an extremely potent promoter of tumor formation in mouse skin (reviewed in Boutwell, 1964). With the isolation and characterization of the active principles in croton oil and the synthesis of structurally related phorbol diesters (reviewed in Hecker, 1978), a series of compounds with a range of tumor-promoting activity became available for studying the mechanism of promotion in this model. Recently, models of two-stage carcinogenesis have been developed in other species and tissues and with other promoting compounds, for example, phenobarbital is a promoter of liver carcinogenesis in rats initiated with 2-acetylaminofluorene (Peraino et al., 1971, 1973, 1978). The recent profusion of reports (reviewed in Weinstein et ul., 1977; Diamond et al., 1978a) on the effects of phorbol diester tumor promoters on various types of cells in culture show that they produce a variety of biological and biochemical changes; in this respect, these compounds resemble agents such as hormones, retinoids, and transforming viruses that also have multiple effects on target tissues. Although many of the effects induced by phorbol diesters in cell culture reproduce effects seen in mouse skin, there is as yet no in vitro model completely analogous to tumor promotion in that tissue. However, two-stage transformation in vitro, analogous with respect to timing and dosage to two-stage carcinogenesis in v i w , has been described in a mouse fibroblast cell line (Mondal and Heidelberger, 1976; Mondal et al., 1976); this and other well-

TUMOR PROMOTERS

3

defined cell culture systems will be useful in determining not only which of the multitude of effects induced by tumor promoters are specifically related to promotion, but also how these compounds exert promoting activity in vivo and in vitro. This article primarily covers studies on the mechanism of tumor promotion in mouse skin by croton oil and the phorbol diesters and on the effects of these compounds on cell cultures of various types. Most studies of tumor promotion have used these compounds. However, other types of promoters may differ from the phorbol diesters in their primary biochemical sites and modes of action, and the results of studies on the mechanism of action of phorbol diesters may be specific to these promoters. This article describes recent findings in the field of tumor promoters through mid- 1979 and provides sources in the literature useful as background. For an historical perspective of the concept of two-stage carcinogenesis and tumor promotion, particularly in mouse skin, the early reviews of Berenblum (1954, 19691, Setala (1956, 1960), Saffiotti and Shubik (1963), Salaman and Roe (1964), Boutwell (1964), and Van Duuren (1969) can be consulted. Several recent reviews by these and other authors discuss both general and specific aspects of tumor promotion; several contain extensive surveys of the literature (Stenback er al., 1974; Scribner and Suss, 1978). Stenback et al. (1974) review many of the biological aspects of tumor promotion and cocarcinogenesis in skin. Van Duuren (1976), Hecker (1971), and Hecker and Schmidt (1974) discuss the tumor-promoting and cocarcinogenic agents themselves, in particular the nature and structureactivity relationships of the phorbol esters. Hecker (1978) has also recently reviewed the structures and biological activity of other plant diterpene esters. Boutwell (1974) published a review that is concerned primarily with the biochemical mechanisms of tumor promotion. A more general discussion of the concept of two-stage carcinogenesis and of the events that may be critical in promotion is in the review of Scribner and Suss (1978). The effects of tumor promoters and, in particular, of the diterpene esters, on various types of cells in culture were reviewed by Weinstein et ul. (1977) and by Diamond et al. (1978a). These authors discuss the significance of recent observations that these compounds induce the protease plasminogen activator and inhibit terminal differentiation of cells in culture. There are several excellent, relatively up-to-date surveys of the literature on tumor promotion. The proceedings (Slaga et al., 1978) of a symposium on Mechanisms of Tumor Promotion and Cocarcinogenesis

4

LEILA DIAMOND et

al.

held in 1977 contain detailed descriptions of newer findings in the field and some excellent reviews and literature surveys of specific aspects of the subject (as, e.g., the chapter by Hecker). The International Cancer Research Data Bank published an Oncology Overview on Tumor Promotion and Cocarcinogenesis containing 200 abstracts of selected articles published in the period 1970-1977 (Helmes, 1978); they also publish a monthly “Cancergram” on Modification of Carcinogenesis. II. Concept of Two-Stage Carcinogenesis

Yamagiwa and Ichikawa (1918) were the first to demonstrate that skin tumors could be induced in experimental animals; they did so with repeated application of coal tar condensate to rabbit ears. A few years later, Deelman (1927) found that a similar scheme for the induction of skin tumors in mice could be divided into two stages by first repeatedly applying carcinogenic tars to the skin and then wounding the skin. Twort and Twort (1939) described a two-stage procedure of inducing tumors using chemicals for both stages, that is, multiple applications of benzo(a)pyrene followed by multiple applications of oleic acid. Berenblum (1941) was the first to show that croton oil, obtained from the seeds of Croton tigfium L . , and a resin from this oil were extremely potent cocarcinogens when applied to mouse skin alternately with benzo(a)pyrene. The protocol presently used in most two-stage carcinogenesis studies was first described by Mottram (1944) who elicited skin tumors by treating the backs of mice with a single, subcarcinogenic dose of an initiator, benzo(a)pyrene, followed by repeated applications of a promoter, croton oil. Initiation requires only a single application of the carcinogen and is thus a relatively rapid process (Berenblum, 1954; Colburn and Boutwell, 1966). It produces no apparent morphological alterations in the epidermis (Berenblum and Shubik, 1949; Scribner and Suss, 1978). In contrast, promotion is a slow process requiring repeated exposure to the promoter over a latent period of several weeks (Berenblum and Shubik, 1947; Boutwell, 1964, 1974; Stenback et a f . , 1974). The initiation event can probably persist unexpressed in the epidermal cells for as long as a year or for the lifetime of the animal (Berenblum and Shubik, 1949; reviewed in Van Duuren, 1969), even though the epidermis is a continually renewing tissue. The fact that initiation has introduced some change into the epidermis is demonstrated when subsequent promoter treatment results in tumor formation (Berenblum and Shubik, 1949; see Scribner and Suss, 1978). If the order of application of

TUMOR PROMOTERS

5

initiator and promoter are reversed, tumor formation does not occur (Berenblum and Haran, 1955). Two-stage carcinogenesis in which initiation and promotion are clearly separate events mediated by different agents is distinct from experimental protocols in which a single, high dose of a carcinogen is sufficient to elicit tumor formation without subsequent promoter treatment (Terracini et d . , 1960). With the two-stage system of mouse skin tumorigenesis, the first tumors to be observed are papillomas. After continued promoter treatment, malignant tumors appear on the skin: it is not clear if they develop from papillomas (Burns et u / . , 1978). Tumor promotion has been divided by Boutwell (1964) into two steps: first, the "conversion" of an initiated cell into a dormant or latent tumor cell and, second, the "propagation" of that cell. He made the distinction between these two steps on the basis of experiments (Boutwell, 1964) in which it was shown that after several weeks of treatment of initiated skin with croton oil, treatments with the hyperplasiogenic agent, turpentine, could be substituted and tumors would still develop; turpentine without prior promoter treatment was ineffective. Thus, he concluded that hyperplasiogenic agents such as turpentine are effective only in the propagation step and not the conversion step of promotion. Although more direct evidence that promotion actually involves two distinct steps is required, these terms can be useful when discussing what appear to be specific events in the promotion process. There is good evidence that promotion is a reversible stage in carcinogenesis, but it is not clear whether all, or only some, steps are reversible. Increasing the intervals between promoter application beyond the optimal time, while keeping the individual and total (cumulative) doses of promoter constant, greatly reduces or eliminates the promoting effect in mouse skin (Boutwell, 1964). This implies that the increase in time between treatments allows reversion of a critical event(s); this may be the conversion step. That promotion may be reversible even after benign tumors have formed is suggested by the fact that discontinuation of promoter treatment can result in regression of many of these tumors (reviewed in Stenback et u/., 1974). However, promotion can be carried to a point at which formation of malignant, nonregressing tumors is inevitable; if the promoter treatments are discontinued at that time, even before such tumors are visible, a significant number of carcinomas will develop, although fewer than if the treatments had been continued (Burns et NI., 1978). The concept of tumor progression has been distinguished from tumor promotion (Boutwell, 1964), although experimentally absolute distinction between the two is difficult. The term progression usually refers to all

6

LEILA DIAMOND et af.

the multiple steps involved in the transition of an initiated cell into a malignant neoplasm without regard to the method of tumor induction (see Foulds, 1969; Farber, 1976; Nowell, 1976), whereas promotion is an operational term derived from two-stage models of tumor induction.

Ill. Chemistry of Tumor Promoters

A. IDENTIFICATION

OF

PROMOTING COMPONENTS I N CROTONOIL

Initial studies of the biologically active components of croton oil were concerned with isolating its vesicant, purgative, and toxic components. Bohm (1919, using methanol extraction, prepared croton resin containing the active components of croton oil. Bohm and Flaschentrager (1930) isolated a crystalline product, phorbol (Fig. l ) , that was not toxic unless acetylated (Bohm et al., 1935), suggesting that esters of phorbol were the biologically active components of croton oil. However, the studies of croton resin by Spies (1935) suggested that the presence of some free hydroxyl groups was required for its toxicity. The discoveries of Berenblum (1941) that croton oil was a cocarcinogen and that croton

FIG. 1. Structures of phorbol-12,13-diesters (left) and 4-a-phorbol-12,13-diesters(right) as described by Hecker (1978). Phorbol, an alcohol with the carbon skeleton tigliane (Hecker, 1978). is the parent alcohol of the tumor-promoting phorbol- 12,13-diesters found in croton oil (reviewed in Hecker, 1967, 1971; Hecker and Schmidt 1974). The A series of esters has a long-chain fatty acid on C-12 (R') and a short-chain fatty acid on C-13 (R'); the B series of esters has a long-chain fatty acid on C-13 (R') and a short-chain on C-12 (R'). The isolation and identification of these compounds are described in Hecker er ul. (1964a.b); Van Duuren and Orris (1965): Clarke and Hecker (1%5a,b); Hecker and Bresch (1965); Hecker and Schairer (1967). The work has been reviewed by Hecker (1967, 1971) and Hecker and Schmidt (1974).

TUMOR PROMOTERS

7

resin was more active than croton oil suggested that the cocarcinogenic components might also be esters of phorbol. Using countercurrent distribution techniques, the research groups of Hecker (reviewed in Hecker, 1968, 1971) and Van Duuren (reviewed in Van Duuren and Orris, 1965; Van Duuren, 1969) isolated pure tumorpromoting compounds from croton oil and identified their structures as 12,13-diesters of phorbol (Table I). Phorbol itself is inactive in mouse skin (Hecker, 1971; Baird and Boutwell, 1971; Baird er a/., 1972; Slaga et a/., 1976). The two major groups of esters differ in the position of the long-chain ester, which is either on C-12 (A group) or on C-13(B group). All contain one acid of carbon chain length 8 to 16 and one of 2 to 6; all are potent tumor promoters. Of all of the phorbol diesters, 12-0-tetradecanoylphorbol- 13-acetate (TPA)' (Hecker's Fraction Al, Van Duuren's Fraction C), has the highest tumor-promoting activity in mouse skin (Van Duuren and Orris, 1965: Van Duuren, 1969; Hecker, 1968, 1971). It has also been reported to show weak carcinogenic activity after prolonged treatment of the skin of hairless mice (Iversen and Iversen, 1979).

B. PHORBOL DIESTERTUMOR PROMOTERS Even before the chemical structure of the phorbol molecule was established (Hoppe et a/., 1967; Pettersen el ( I / . , 1967), methods were developed for synthesizing both the A and B groups of phorbol diesters from phorbol prepared by hydrolysis of croton oil (reviewed in Hecker, 1971). This made possible the synthesis of symmetrical phorbol diesters with a wide range of ester chain lengths: the role of ester size in the induction of biological activity could now be assessed. The symmetrical diesters may be synthesized from phorbol in only two steps (Thielmann and Hecker, 1969), and a number have been used for studies of the effect of chain length on promoting activity (Van Duuren and Orris, 1965: Van Duuren, 1969; Hecker, 1968, 1971, 1978; Thielmann and Hecker, 1969; Baird and Boutwell, 1971; Baird et u / . , 1972; Scribner and Boutwell, 1972; Slaga er al., 1976) (Table 11). The activity of the diesters increases with increasing chain length up to 8 carbons but, beginning with the 12carbon length, activity falls off with increasing chain length. In both the A numbering system for the phorbol diesters different from the system used here is that of Chemical Abstracts and the International Union of Pure and Applied Chemistry (see Van Duuren ef d.,1978a, 1979). This latter system designates the 12- and 13- positions as 9- and 9a-, respectively, and TPA as phorboI-Pmyristate-9a-acetate, abbreviated in the literature as phorbol myristate acetate or PMA. The designation P M A with the numbering system used in this chapter can also be found in the literature.

8

LEILA DIAMOND

et al.

TABLE 1 PHORBOL DIESTERS PRESENT IN CROTON OILa Number

12-Position Tetradecanoic Decanoic Dodecanoic Hexadecanoic 2-Methylbutyric 2-Methylbutyric Tiglic (rrcrns-2-methyl-2butenoic) Acetic 2-Methylbutyric Tiglic Acetic

13-Position

Molecular weight

Acetic Acetic Acetic Acetic Dodecanoic Dccanoic Decanoic

616.8 560.7 588.8 644.9 630.9 602.8 600.8

Dodecanoic Octanoic Octanoic Decanoic

588.8 574.8 572.7 560.7

The structure of phorbol and positions of the esters are shown in Fig. I . The references describing the isolation and identification of the diesters are cited in the legend to Fig. I ; the work is reviewed in Hecker (1967, 1971) and Hecker and Schmidt (1974).

symmetrical and unsymmetrical diesters, the compounds with a combined chain length of 14 to 20 carbons usually have the highest biological activity (Thielmann and Hecker, 1969). Although the A series of diesters have the long-chain acid on the 12position, monoesters with long-chain acids on the 12-position have low biological activity whereas those with long-chain acids on the 13 have high biological activity (Thielmann and Hecker, 1969). Thus, the biological activity requires esterification with either a long-chain fatty acid on the 13-position or a combination of two fatty acids with a combined length of 14 to 20 carbons on the 12- and 13-positions. Interestingly paradoxical is the aromatic ester, phorbol dibenzoate: at the usual frequency of application, twice weekly, this agent has very weak promoting properties; but at a frequency of four applications per week, it becomes a moderately potent promoter (Baird and Boutwell, 1971; Baird et al., 1972). Further information about which structural features of the phorbol esters are responsible for tumor-promoting activity has been obtained through studies of compounds with alterations in other portions of the phorbol nucleus (Table 111). Some of these occur naturally and others result from synthetic alterations of the phorbol ring (reviewed in Hecker, 1978). The presence of the free allylic hydroxyl group at C-20 is essential for high biological activity. Croton oil contains a series of phorbol- 12,13-

9

TUMOR PROMOTERS

diesters that are also esterified at C-20 (Hecker, 1971, 1978). These have low promoting activity unless the C-20 ester is cleaved to yield an active tumor promoter of the 12,13-diester series. Hecker (1978) has referred to these as "cryptic promoters", although there is no evidence that they are cleaved to active 12,13-diesters in mouse skin. Other alterations at C-20 also reduce the biological activity of the phorbol ester, for example, TABLE I1 TUMOR-PROMOTING ACTIVITY OF MONOESTERS A N D 12.13-SYMMETRICAL DIESTERS OF PHORBOL ~~

Compound Phorhol

12-Decanoate 12-Tetradecanoate 13-Decanoate 13-Tetradecanoate 12.13-Diacetate

~

Promoting activity"

Molecular weight

-b

364.4

-

518.7 574.8 518.7 574.8 448.5

-

++ + -b

+ +

476.6 504.6

12,I3-Dihexanoate 12.13-Diheptanoate 12,13-Dioctanoate

++

+ +' +++

560.7 588.8 616.8

12.13-Didecanoate

+++

672.9

++ ++

729. I 785.2

++

572.7

12,13-Dipropanoate 12.13-Dihutyrate

12.13-Didodecanoate 12.13-Ditetradecanoate

12,13-Dihenzoate

Reference Hecker (1971); Baird and Boutwell (1971); Baird ef c d . (1972); Slaga ef ( I / . (1976) Thielmann and Hecker (1969) Thielmann and Hecker (1969) Thielmann and Hecker (1969) Thielmann and Hecker (1969) Baird and Boutwell (1971); Baird ef ( I / . (1972); Slaga e / a / . ( 1976) Baird (1971) Thielmann and Hecker (1969); Scrihner and Boutwell (1972) Thielmann and Hecker (1969) Hecker (1978) Thielmann and Hecker (1969): S a g a et a / . (1976) Thielmann and Hecker (1969); Baird and Boutwell (1971); Baird er a / . (1972) Thielmann and Hecker (1969) Thielmann and Hecker (1969); Scribner and Boutwell (1972); Slaga e / (/I. (1976) Hecker (1968, 1971, 1978); Baird and Boutwell (1971); Baird ef d . (1972)

On the scale used, TPA would he ++++. The promoting activities are the averages of those reported in several strains of mice; results in individual strains are given in the appropriate references. Has weak promoting activity in a sensitive strain of mice (Baird and Boutwell, 1971; Baird ef u / . , 1972). Reported by Hecker (1978) to have promoting activity similar to that of phorhol-12,13dihutyrate.

10

LEILA DIAMOND

et al.

TABLE 111 EFFECTON BIOLOGICAL ACTIVITYOF MODIFICATION OF PHORBOL Activity" Modification Alteration at C-20 Acetate of TPA Tetradecanoate of TPA Te tradecanoate Decanoate Deox y-PDDb Methyl ether of PDD Aldehyde of TPA

Irritant

Promoting

+++ ++

+++ +

-

-

-

-

-

N.D. N.D.

-

+

++

Acid of PDD Alteration of ring junction 4-wPDD

-

-

-

-

4-0-Methyl TPA Other alterations 6,7-Dihydro TPA 6a.7a-Oxide of TPA 6&7@Oxide of TPA 3-Hydroxy-TPA

+

*

N.D.'

++ +++ +++

a

++

++ +++

On the scale used, TPA would be

* PDD, phorbol- 12.13-didecanoate.

++

Reference Hecker (1978) Hecker (1978) Thielmann and Hecker (1969) Thielmann and Hecker (1969) Thielmann and Hecker ( 1969) Thielmann and Hecker (1969) Hecker (1978); Van Duuren ef ul. (1979) Hecker ( 1978) Hecker (1971, 1978); Van Duuren et a / . ( 1979) Hecker (1978) Segal et ul. (1978) Hecker (1978) Hecker (1978) Segal et a / . (1975, 1978); Hecker (1978)

++++ for both activities.

N.D.. not done.

oxidation to either an aldehyde or acid, loss of the hydroxy group, or formation of an ether linkage (Thielmann and Hecker, 1969; Hecker, 1978) (Table 111). Another feature of the phorbol molecule critical for high promoting activity is the steric configuration at the ring junction of C-4 and C-10. In phorbol itself, the C- 10-hydrogen and C-4-hydroxyl are trans, whereas in the epimer 4-a-phorbo1, they are cis (Hecker, 1978). As can be seen in Fig. 1, this difference in the ring junction greatly alters the overall configuration of the molecule. The esters of the a epimer, such as 4-aphorbol- 12,13-didecanoate, are inactive both as tumor promoters and as irritants (Hecker, 1978; Van Duuren et al., 1979). Loss of the 4-hydroxyl group does not affect biological activity, for 4-deoxyphorbol diesters are biologically active whereas 4-deoxy-4a-phorbol diesters are not (Hecker, 1978). Other features of this ring junction are also important: methylation

TUMOR PROMOTERS

11

of the 4-position gives 4-0-methyl-TPA, an inactive analog (Hecker, 1978), whereas the 4-deoxyphorbol diesters have biological activity. More information about the structure-activity relationships in the phorbol molecule has been obtained from diterpene esters in other plants. For example, such deoxy derivatives of phorbol esters as 12-deoxyphorbol and 4-deoxyphorbol are formed in some plants; such oxidized phorbol are formed in derivatives as 12-deoxy-5/3-hydroxyphorbol-6a,7a-oxide others (reviewed in Hecker, 1978). Oxidized forms such as 6,7-epoxides are also formed as autooxidation products of phorbol esters or by metabolism of the phorbol esters in mouse skin (Schmidt and Hecker, 1975; Hecker, 1978). In general, therefore, a potent promoter must have both a highly lipophilic portion of the molecule and a hydrophilic portion. Alterations that do not drastically alter this relationship as, for example, oxidation of the hydrophilic portion (e.g., 6,7-epoxide) or loss of part of the ester functions (e.g., 12-deoxyphorbol esters), do not greatly decrease the biological activity of the molecule (Hecker, 1978). In contrast, major alterations in the hydrophilic or lipophilic nature of the molecule as, for example, the loss of hydroxyls in the hydrophilic portion (C-20 esters or 4-0-methyl ethers), the lack of a lipophilic portion (such as phorbol itself or short-chain 12,13-diesters) or a total change in the configuration of the molecule (4a-epimers), result in large or complete losses of biological activity (Hecker, 1978). Within this general requirement for both hydrophilic and lipophilic portions, other factors affect biological activity. Monoesters with long chains at the 12-position are inactive, whereas those with long chains at the 13-position are active (Thielmann and Hecker, 1969); this may be because a free C-13-hydroxyl has the effect of increasing the size of the hydrophilic portion of the molecule, whereas a free C- 12-hydroxyl, separated from the hydrophilic portion of the molecule by the C- 13-ester and the two methyl groups at C-17, creates two hydrophilic portions. The presence of a small-chain ester on C-13, as in TPA, would eliminate this second polar region and restore the lipophilic-hydrophilic balance of the molecule. In addition, the absolute promoting activity of a compound is also dependent upon other features of molecular structure. Changes in stereochemical configurations that do not greatly alter the polarity of the molecule but induce changes in its stereochemistry (such as reduction of the 3-ketone to a hydroxyl) or its rigidity (such as reduction of the 6,7bond) decrease but do not abolish biological activity (Segal er al., 1978). Similarly, diesters of very long chain length have decreased biological

12

LEILA DIAMOND

et al.

activity (Thielmann and Hecker, 1969; Scribner and Boutwell, 1972), although the ester portion of the molecule is probably as lipophilic as that of the more active symmetrical diesters. C. CELLULAR INTERACTIONA N D METABOLISM OF PHORBOL ESTERS Because of the highly lipophilic portion of the phorbol diester molecule that is required for promoting activity, the compounds are easily sequestered in cellular membranes. This makes it difficult to determine if the biological changes they induce result from interaction with specific receptors or regions on cell membranes. Studies on membrane interactions of phorbol diesters are discussed in Sections IV,A,2 and V,C,l. With the exception of TPA, very Little is known about the metabolic fate of phorbol esters. Because the phorbol esters produce such rapid effects, it might be expected that they do not require metabolic activation and that the parent compounds are the actual promoting species. Early attempts to prove this in mouse skin were inconclusive because of the rapid clearance of the [3H]TPA from the skin and the resulting low recovery (Van Duuren, 1969). However, neither this nor other more recent studies (Kreibich et al., 1974; Segal et al., 1975; Berry et al., 1977; Hecker, 1978) detected any metabolism of [3H]TPA during the first 24 hours after application to the skin. TPA added to homogenates of either mouse skin or liver is degraded, primarily to the monoester 12tetradecanoylphorbol (Berry et al., 1978). More information about the metabolic pathways of phorbol esters has been obtained in cell culture systems. Kreibich et al. (1974) found that extensive metabolism of [3H]TPAoccurred within 12 hours in cultures of mouse L cells. O’Brien and Diamond (1978a) showed that [3H]TPA is rapidly metabolized by hamster embryo fibroblasts in culture and that the major metabolite is the monoester phorbol- 13-acetate. Very little metabolism of TPA occurs in cultures of human diploid fibroblasts (O’Brien and Diamond, 1978a,b) or HeLa cells (Kreibich et al., 1974). The metabolism of phorbol esters that are not radioactive can be measured by an indirect bioassay that is based on the ability of tumorpromoting compounds to induce the enzyme ornithine decarboxylase in cell culture (O’Brien and Diamond, 1978b). It was shown using this procedure that phorbol- 12,13-dibenzoate is rapidly inactivated by hamster embryo fibroblasts but that phorbol-12,13-didecanoateis not; chromatographic data have confirmed that this diester is not metabolized by hamster cells (T. G. O’Brien, unpublished). It may be that some cells in

13

T U M O R PROMOTERS

culture have an enzyme that can specifically hydrolyze the ester bond at C - 12 on some, but not all, phorbol diesters. The information now available on the metabolism of TPA indicates that metabolism is not required for biological activity and, in fact, leads to loss of activity. More data are needed before it can be determined whether this is a general rule for all phorbol diester tumor promoters. Knowledge about the extent and rate of metabolism of these compounds is important not only for theoretical reasons but also for establishing treatment schedules, particularly in long-term experiments in vivo or in vifro designed to maintain effective concentrations of promoter in a tissue or cell.

D. TUMORPROMOTERS FROM OTHERPLANTS Plants other than Croton tiglium contain different types of diterpene esters with tumor-promoting activity in mouse skin. Hecker (1978) has classified these by the nature of the carbon skeleton of the terpene molecule into three classes: ( I ) esters of tigliane including the phorbol diesters of C . tiglium described in Section III,B (see Fig. 1 and Table I ) and esters from other plants with different configurations of hydroxyl groups on tigliane: ( 2 ) esters of daphnane, including esters of resiniferonol (see Fig. 2 ) or modified resiniferonol molecules such as mezerein, gnidia esters, and pimelea esters; and (3) esters of ingenane, including esters of ingenol (see Fig. 2 ) from Euphorhiri lrithyris and similar esters from other Euphorhici. Hecker (1978) has reviewed the structures of these esters in detail and described their irritant and tumor-promoting activities. The esterified alcohols with high promoting activity all have a hydrophobic portion somewhat similar to phorbol (Figs. I and 2 ) and a lipophilic

a __

OR

on

OH

CH20H

FIG.2. Structures of resiniferonol esters (left) and ingenol esters (right) as described by Hecker (1978). These are the parent alcohols of two series of plant diterpene tumor promoters: those with a daphnane carbon skeleton (left) and those with an ingenane skeleton (right) (Hecker, 1978).

LEILA DIAMOND et

14

al.

portion. The free alcohols from these compounds and esters in which a short chain such as acetate has been substituted for the long-chain ester have little biological activity (Hecker, 1978). Thus, although promoting activity of diterpene esters requires the presence of lipophilic-hydrophilic regions, the differences in promoting activity between the esters are determined by more subtle differences in the structural features of the molecule.

E. OTHERMOUSESKINTUMOR PROMOTERS

In addition to the diterpene esters, other classes of chemicals have tumor-promoting activity in mouse skin (Table IV). Unsaturated fatty acids such as oleic acid are weak promoters at high doses (Twort and Twort, 1939; Gwynn and Salaman, 1953; Holsti, 1959). Methyl esters of several unsaturated fatty acids are somewhat more potent promoters than

TABLE IV COMPOUNDS OTHER THANDITERPENE ESTERSWITH TUMOR-PROMOTING ACTIVITYI N MOUSESKIN Acids and esters Fatty acids Fatty acid methyl esters Surface-active agents Phenols Anthralin

Twort and Twort (1939); Gwynn and Salaman (1953); Holsti (1959) Arffman and Glavind (1971) Setila (1956, 1960) Bock and Burns (1963); Van Duuren et a / . (1978b)

Other phenols

Nonaromatic hydrocarbons Limonene and citrus oils Undecane and other linear alkanes Miscellaneous Cigarette smoke condensate

Weakly acidic fraction of condensate lodoacetic acid

l-Fluoro-2,4-dinitrobenzene Flame retardants

Boutwell and Bosch (1959); Van Duuren (1969); Van Duuren and Goldschmidt (1976) Roe and Peirce (1960) Sick (1966) Gellhorn (1958); Van Duuren et a / . (1971); Van Duuren and Goldschmidt (1976); Hoffmann and Wynder (1976) Wynder and Hoffmann (1961); Bock er a / . (1971); Hecht et (I/. (1975, 1978) Gwynn and Salaman (1953) Bock er a / . (1%9) Loewengart (1977)

TUMOR PROMOTERS

15

the free fatty acids (Arffmann and Glavind, 1971); Rohrschneider and Boutwell (1973) have noted the structural similarity between these fatty acid esters and the phorbol diester molecules. However, all these fatty acids and their methyl esters require several thousand-fold greater doses for promotion than TPA. Another class of tumor promoters are the surface-active agents such as Spans, esters of fatty acids and hexital anhydrides, and Tweens, polyoxyethylene ethers of the Spans (Setala, 1956, 1960). The promoting activity of these agents is very low compared to that of phorbol diesters. Various phenolic compounds including phenol itself have promoting activity in mouse skin (Boutwell and Bosch, 1959; Van Duuren, 1969; Van Duuren and Goldschmidt, 1976). Anthralin (1,8-dihydroxy-9-anthrone) is the compound with the highest promoting activity outside the plant terpene ester series (Bock and Burns, 1963; Van Duuren et al., 1978b) and is frequently used experimentally as a nonphorbol ester-type promoter. It is much less active than the phorbol esters, but much more active than fatty acids and their esters. The presence of free hydroxyls on a portion of the anthralin molecule is required for the promoting activity: esterification of the 1- and 8-hydroxyl groups, either with acetate or myristate, abolishes the promoting activity of the anthralin molecule, whereas formation of an acetate or myristate ester on C-10 (on the opposite side of the molecule) does not alter its promoting activity (Van Duuren et al., 1978b). Some nonaromatic hydrocarbons also have weak promoting activity in mouse skin (Roe and Peirce, 1960; Sice, 1966). These include both cyclic hydrocarbons such as limonene and straight chain alkanes such as tetradecane. Cigarette smoke condensates have carcinogenic, and tumor-promoting activity in mouse skin (Gellhorn, 1958; Van Duuren, et al., 1971; Hoffman and Wynder, 1976). Efforts have been made to fractionate smoke condensate and to identify the components responsible for its promoting activity. The major portion of the promoting activity is associated with the weakly acidic fraction of the condensate (Wynder and Hoffman, 1961; Bock et al., 1971; Hecht et al., 1975). A large number of components have been identified in these fractions including both phenolic compounds and fatty acids, and several of these have promoting activity (Hecht et a l . , 1975, 1978). Other compounds that are weak promoters in mouse skin include two chemicals that react with proteins, iodoacetic acid (Gwynn and Salaman, 1953)and I-fluoro-2,4-dinitrobenzene(Bock et al., 1969), and some flame retardants (Loewengart, 1977).

16

LEILA DIAMOND

et al.

F. TUMOR PROMOTERS I N TISSUES OTHERTHAN SKIN Examples of compounds capable of promoting tumor formation in tissues other than skin are given in Table V. Phorbol, the parent alcohol of the phorbol ester series, is inactive as a promoter on mouse skin, but when administered intraperitoneally , acts as a promoter of leukemia and of tumors of the lung, liver, and mammary glands in initiated rodents (Berenblum and Lonai, 1970; Armuth and Berenblum, 1972, 1974, 1977). When applied topically, TPA induces tumors of other organs, as well as skin in mice (Goerttler and Loehrke, 1976a,b; 1977); when administered intragastrically, it is a promoter in the forestomach of mice (Goerttler et al., 1979a). Croton oil in the drinking water is a promoter of gastric carcinomas in the glandular stomach of rats (Matsukura et. al., 1979). An oxygenated fatty acid that is a skin tumor promoter when applied topically (Arffmann and Glavind, 1971) acts as a promoter for tumors of the forestomach when fed in the diet (Kler et al., 1975). A cyclopropenoid fatty acid promotes liver tumors in aflatoxin B,-initiated rainbow trout (Lee et al., 1971). Phenobarbital is a potent promoter of liver tumors in rats initiated with 2-acetylaminofluorene or other hepatocarcinogcns (Peraino et al., 1971, 1973, 1978; Weisburger e f al., 1975; Kitagawa and Sugano, 1978). Two dietary sweeteners, saccharin and cyclamate, are promoters of bladder tumor induction in rats initiated with N-methyl-N-nitrosourea (Hicks et al., 1973, 1975, 1978; Hicks and Chowaniec, 1977). A number of bile acids promote formation of colon tumors in rats initiated with Nmethyl-N’-nitro-N-nitrosoguanidine(Narisawa et ul., 1974; Reddy et ul., 1976, 1977, 1978). Other compounds with promoting activity in some systems include butylated hydroxytoluene (BHT) (Witschi and Lock, 1978; Peraino et al., 1977, 19781, dichlorodiphenyltrichoroethane(DDT) (Peraino et al., 1975, 1978) and polychlorinated biphenyls (PCBs) (Nishizumi, 1976; Kimura er al., 1976). It is important to note that compounds that are promoters in one system may be inactive in another; phorbol (Hecker, 1971; Baird et al., 1972; Slaga et al., 1976), deoxycholic acid (Glauert and Bennink, 19781, and phenobarbital (Grube et al., 1975), for example, are inactive as promoters in mouse skin, although they are promoters in other tissues (see Table V). IV. Experimental Models of Tumor Promotion

A. MOUSESKINMODEL The mouse skin model of two-stage carcinogenesis has been the model of choice for studying tumor promotion for a number of reasons. In

TUMOR PROMOTERS

17

addition to the availability of a series of promoters with a range of tumorpromoting activity in the skin, other advantages of the model are that the tumors have a relatively short latent period, that tumors are visible so that the animals need not be sacrificed to score tumors, and that the number of tumors that develop on the skin reflects the relative potency of the compound to be evaluated. Research on the mechanisms of two-stage carcinogenesis in mouse skin has been concentrated on the epidermis as the target issue of the promoter, as well as of the initiator. There have been few studies on the effects of promoters on the dermis and how these compounds might influence tumor formation by affecting the physiology or structure of the epidermis. It is possible, however, that the dermal layer is a target of promoter action in mouse skin (not simply because of the leukocytic infiltration that follows promoter treatment) and plays a role in the promotion of skin tumors through some type of cell-cell interaction. In view of the complexity of tumor formation and the paucity of information, the possibility has to be considered that in the skin, as well as in other organ sites, tumor formation involves both direct and indirect effects of the promoter on initiated cells. Numerous biological changes have been observed in mouse skin treated with tumor promoters (reviewed in Boutwell, 1974; Stenback et al., 1974; Scribner and Suss 1978). Before pure phorbol diesters were available, these studies usually used croton oil as the promoting agent and, because of its complex chemical composition, results were often difficult to relate unambiguously to tumor promotion. Once the active principles of croton oil were identified, isolated, and synthesized (see Sections III,A and B), it became possible to study the consequences in mouse skin of exposure to pure tumor promoters of high potency. 1. Biological Effects of Phorhol Diester lumor Promoters

The acute biological effects of treatment of mouse skin with phorbol diester tumor promoters are profound and, despite years of study, the relationship of some of these phenomena to tumor promotion per se is not clear [Scribner and Suss (1978) have an excellent discussion of this problem]. A number of biological effects of promoter treatment, however, have been well established. First, the histological changes that follow promoter treatment appear to be the same whether or not the skin has been initiated; these are described in detail in Stenback et al. (1974) and Scribner and Suss (1978). Within a few hours after application of an effective dose of a promoter, such local tissue reactions as the edema and erythema characteristic of inflammation and irritation are evident

TABLE V COMPOUNDS WITH TUMOR-PROMOTING ACTIVITYI N TISSUEOTHER THANS K I N _ _ _ _ _ _ _ _ _ ~ ~ ~

~~~

Promoter Phorbol and Phorbol Esters Phorbol

W

12-0-Tetradecanoylphorbol-13acetate (TPA)

Croton oil Fatty Acids Methyl- ltoxonital-trans- 10octadecenoate Methyl sterculate Phenobarbital

Method of application'

Initiatolb

Animal Reference' Mouse Mouse Rat

2 3

AAF DMBA

Mouse Mouse

4 5

DMBA DMBA MNNG

Transplacental Skin, lung, liver Intragastric Forestomach Diet Glandular stomach

Mouse Mouse Rat

6, 7 8

Stomach tube

Forestomach

Mouse

9

Liver Liver Liver

Trout Rat Rat

10 11-14 15

Bladder

Rat

16-19

Bladder

Rat

17-19

DMBA DMN DMBA

ip injection Topical Topical Intragastric Diet

Diet Diet Diet

Tissue

Leukemia Lung, liver Mammary gland, leukemia Transplacental Liver T o mothers Skin, other organs

ip injection ip injection ip injection

Diet

~~

Method of application"

NQO

Topical sc injection Stomach tube

Aflatoxin B1 Diet AAF, DAB, 3'-Me-DAB Diet sc injection DENA

Sweeteners Saccharin

Diet

MNU

Cyclamate

Diet

MNU

Instilled in bladder Instilled in bladder

1

30

Bile Acids Taurodeoxycholic acid Lithocholic acid Sodium deoxycholate Sodium cholate Sodium chenodeoxycholate Other Chemicals Butylated hydroxytoluene (BHT)

Intrarectal Intrarectal Intrarectal Intrarectal Intrarectal

MNNG MNNG MNNG MNNG MNNG

Intrarectal Intrarectal Intrarectal Intrarectal Intrarectal

Colon Colon Colon Colon Colon

Rat Rat Rat Rat Rat

20,21 20.21 21.22 21, 23 21, 23

Diet ip injection Diet

AAF Urethane AAF

Diet ip injection Diet

Liver Lung adenoma Liver

Rat Mice Rat

13, 24 25 13, 26

Diet Injection Diet

Liver Liver Bladder

Rat Rat Rat

27 28 29

Dichlorodiphenyltrichloroethane (DDT) Polychlorinated biphenyls (PCBs) Diet Diet DL-Tryptophan Diet

-

0

3'-Me-DAB DENA FANFT

The abbreviations used are: ip, intraperitoneal; sc, subcutaneous. The abbreviations used are: DMBA, 7,12-dimethylbenz(a)anthracene; DMN, dimethylnitrosamine; AAF, 2-acetylaminofluorene; NQO, 4-nitroquinoline I-oxide; DAB, p-dimethylaminoazobenzene;3'-Me-DAB, 3'-methyl-4-dimethylaminoazobenzene;DENA, diethylnitrosamine; MNU, methylnitrosourea; MNNG, N-methyl-N'-nitro-N-nitrosoguanidine; FANFT, N-[4-(5-nitro-2-furyI)-2-thiazolyl]formamide. The references cited are: ( I ) Berenblum and Lonai (1970); (2) Armuth and Berenblum (1972); (3) Armuth and Berenblum (1974); (4) Armuth and Berenblum (1977); (5) Goerttler and Loehrke (1976a): (6) Goerttler and Loehrke (1976b); (7) Goerttler and Loehrke (1977): (8) Goerttler et a/. (1979); (9) Kler et a / . (1975); (10) Lee et a / . (1971); ( 1 I ) Peraino el ( I / . (1971); (12) Peraino et a/. (1973); (13) Peraino et a / . (1978): (14) Kitagawa and Sugano (1978); (15) Weisburger et a/. (1975); (16) Hicks et a / . (1973); (17) Hicks et d.(1975); (18) Hicks and Chowaniec (1977); (19) Hicks et a/. (1978): (20) Narisawa et a / . (1974): ( 2 I ) Reddy el a / . (1978); (22) Reddy et ( I / . (1976); (23) Reddy e? a / . (1977); (24) Peraino et ( I / . (1977); (25) Witschi and Lock (1978): (26) Peraino et a / . (1975); (27) Kimura et a / . (1976): (28) Nishizumi (1976); (29) Cohen et a / . (1979): (30) Matsukura et a / . (1979).

20

LEILA DIAMOND

et al.

and, by 24 hours, leukocytes have infiltrated the dermis (Frei and Stephens, 1968). A sustained stimulation of mitotic activity in the basal cell layer of the epidermis starts within a day or so of treatment and continues for several days, resulting in an increase in the number of nucleated cell layers, from the normal 1-2, to 3-4 or more (Raick, 1973a). The stimulation of proliferation induced by promoter treatment is followed 24-72 hours later by a phase of increased keratinization of the upper layers of the epidermis (Bach and Goerttler, 1971; Raick, 1973a; Balmain, 1976). These responses to the promoter, including the marked hyperplasia, all gradually subside, and within two weeks after treatment, the epidermis has regained its normal appearance (Raick, 1973b). Repeated promoter treatment at regular intervals (every 3-4 days) prevents this “reversion,” and the skin appears to be in a state of constant irritation and hyperplasia (Dammert, 1961). If the skin had been previously initiated, tumors would begin to appear as early as 6 weeks after the start of twice weekly promoter treatment (Boutwell, 1964). The tumors first observed are almost exclusively papillomas; much longer treatment, typically 20 weeks or more, is needed before malignant lesions such as basal cell and squamous cell carcinomas are observed (Boutwell, 1964; Stenback et al., 1974). For a number of reasons, including the shorter latent period and the ease of scoring, experiments in which tumor formation in response to a two-stage protocol is measured frequently use the appearance of papillomas as the endpoint. Although hyperplasia similar to that induced by promoting agents can, of course, be induced in epidermis by weak promoters and inflammatory agents, skin treated with such agents differs histologically from that induced by promoting agents (Raick, 1974). The increase in nucleated cell layers in the epidermis after TPA treatment occurs before an increase in the mitotic index (Raick, 1973a), whereas treatment with the hyperplasiogenic but weakly promoting compound ethylphenylpropiolate first produces an increase in mitotic activity and then an increase in the number of nucleated cell layers (Raick, 1974). There are also ultrastructural differences between skin treated with TPA and with weakly promoting, hyperplasiogenic agents (Raick and Burdzy, 1973). Some of the morphological changes induced by TPA in mouse epidermis resemble changes induced by promoters in cells in culture (see Section V,B). For example, a decrease in cell size and an increase in intracellular spaces have been observed after TPA treatment of epidermis in vivo (Raick, 1973a) and of epidermal cell cultures (Yuspa et al., 1976a). The complex set of tissue reactions to the promoter (erythema, edema, inflammation, irritation, and hyperplasia) creates a high “background” of cellular and molecular changes in the skin, only some of which may

TUMOR PROMOTERS

21

be involved in the promotion process. These same reactions can be elicited by other inflammatory hyperplasiogenic agents that are not promoters. Consequently, it is extremely difficult to uncover those critical changes that promoters induce that distinguish them from irritants in general. Attempts to do so have been focussed in recent years on the effects promoters induce at the cellular and molecular levels: these areas of research are the major emphasis of this review. 2. Biochemical Effects of Phorhol Diester lurnor Promoters Some description of the cells involved is basic to an understanding of the biochemical effects seen after promoter treatment. First, the number of benign and malignant tumors that develop in an initiation/promotion experiment is relatively small (on the order of 50 or less under optimal conditions), so it is likely that the observed biochemical effects of promoter treatment on initiated skin are essentially the responses of normal cells or, at least, the responses of cells not destined to become tumors. It is conceivable that the biochemical response to promoters of initiated cells in the epidermis differs from that of normal cells, but because of the overwhelming number of normal cells, this difference cannot be detected. Second, promotion is a long-term process, and biochemical changes induced by, for example, the tenth promoter application to the skin may very well differ, either qualitatively or quantitatively, from the changes induced by the first application. Few studies have addressed this problem. Third, some investigators have used mouse skin as the experimental tissue, whereas others have used epidermis. If the goal is to uncover critical biochemical events in the cells destined to become tumor cells, it is important to minimize the amount of nontarget tissue in the preparation. This is particularly important in mouse skin for two reasons: (1) Although the dermis contributes much more to the total weight of the skin than the epidermis, it is relatively acellular, so that results obtained with whole skin are not strictly comparable with those from more homogeneous preparations. (2) Because tumor promoters are excellent inflammatory agents, infiltration of leukocytes into the dermis may affect the results of experiments on whole skin, particularly those in which radioactive isotopes are used or in which measurements are made at later times (>24 hours) after promoter treatment when infiltration has occurred. The recent development of relatively simple procedures for isolating epidermis free of dermis (Marrs and Voorhees, 1971; Mufson et al., 1977a) has given impetus to biochemical studies of mouse epidermis. With these points in mind, the studies on the biochemical effects of tumor promoters on mouse skin will be reviewed briefly. In view of the

22

LEILA DIAMOND et

al.

enormous biological effect of promoter treatment on skin, it is not surprising that numerous biochemical changes have been reported. There are probably many others that have not yet been described. Possibly only a few of these changes are essential for effective promotion, and several different experimental approaches have been used to try to identify the nature of these critical biochemical events. a. Effects on Macromolecular Synthesis. i . Nucleic acid synthesis. The rates of RNA and DNA synthesis in mouse skin or epidermis are increased by a single application of promoter (Paul and Hecker, 1969; Hennings and Boutwell, 1970; Baird et a l . , 1971); phorbol diesters with strong promoting activity stimulate incorporation of nucleic acid precursors into mouse epidermis that is sustained over several days, whereas the stimulation caused by weak promoters lasts less than a day (Baird et al., 1971). An initial inhibition of the incorporation of [3H]thymidine into skin or epidermal DNA is also seen (Paul and Hecker, 1969; Baird et al., 1971; Raick, 1973a). Hennings and Boutwell (1970) and Baird et al. (1972) used methylated albumin Kieselguhr chromatography and polyacrylamide gel electrophoresis, respectively, to detect the species of RNA synthesized at early times after treatment of skin with promoters and irritants. Promoters stimulated precursor incorporation largely into ribosomal and transfer RNA species, whereas irritants stimulated the incorporation of precursors into transfer RNA to a greater extent than into ribosomal RNA (Baird et al., 1972). ii. Protein synthesis. Boutwell and colleagues (see Boutwell, 1974) have proposed that altered gene expression is fundamental to the action of tumor promoters. This would mean that promoters change the pattern of protein synthesis either qualitatively or quantitatively. Promoters do induce extensive hyperplasia in mouse skin and, consequently, the rate of total protein synthesis is increased (Baird et al., 1971). It is not clear, however, whether the pattern of synthesis after promoter treatment differs qualitatively from that produced by hyperplasiogenic but weakly promoting agents. Scribner and Boutwell (1972) described two "new" protein bands seen on polyacrylamide gel electrophoretograms of the soluble protein from mouse skin treated for 3.5 hours with TPA; these were absent or greatly reduced in electrophoretograms of the soluble protein from untreated skin. They also were not found in skin treated with either a high dose of acetic acid, a poor promoter, or initiating doses of carcinogens. Balmain (1976) has also described the appearance in TPAtreated mouse epidermis of two "new" proteins that appear to be different from those described by Scribner and Boutwell (1972). They are found in newborn epidermis, but not in untreated adult epidermis; at

TUMOR PROMOTERS

23

least one is involved in epidermal keratinization for it is similar to the "histidine-rich protein(s)" in keratohyalin granules (Hoober and Bernstein, 1966; Sibrack et al., 1974). Whether any of these "new" proteins are specifically involved in promotion or simply reflect the wave of keratinization in response to TPA is unknown. b. Effects on Specific Proteins. Changes in the activity or synthesis of specific proteins with known functions have also been found in promoter-treated epidermis. Many of these changes, such as the increased synthesis and modification of histones (Raineri et al., 1973, 1978), are probably a part of the general program of events associated with the stimulation of cell proliferation. There is evidence, however, for specific changes that cannot be ascribed simply to enhanced cell division or h yperplasia. i . Ornithine decarboxylase. The induction by promoters of the enzyme ornithine decarboxylase (ODC), the rate-limiting enzyme in polyamine biosynthesis, was reported first by O'Brien et al. (1975a). A single application of croton oil or TPA induces more than a 200-fold increase in the activity of this enzyme in the epidermis. A very sharp peak of activity at 5-6 hours is followed by a return to the very low basal level by 24 hours. Subsequent to the increased ODC activity, the levels of putrescine and spermidine are elevated (O'Brien, 1976). The extent of induction of ODC by promoters correlates well with the promoting activity of different doses of TPA and with the promoting activity of other phorbol diester and nonphorbol tumor promoters (O'Brien et al., 1975b). There is relatively little induction of ODC by nonpromoting, hyperplasiogenic agents (O'Brien et al., 1975b), suggesting that the induction of this enzyme by promoters is relevant to the mechanism of promotion. Another enzyme involved in polyamine synthesis, S-adenosyl-L-methionine decarboxylase, is also induced by TPA and other promoters, but to a much lesser extent than ODC (O'Brien et N / . , 1975a). This enzyme is induced by hyperplasiogenic agents as well as by promoters, indicating that the induction of these two polyamine-biosynthetic enzymes can be regulated independently. O'Brien et al. (1975b) and O'Brien (1976) also found that skin tumors produced by a two-stage protocol have permanently elevated levels of ODC in contrast to the extremely low levels in untreated epidermis and to the transient increase in ODC that follows promoter treatment. O'Brien and Boutwell (O'Brien et a/., 1975b; O'Brien, 1976; Boutwell, 1977)have proposed a theory about the nature of initiation and promotion that takes into account their observations on ODC activity in promoter-treated epidermis and in tumors; the theory, which has stimulated considerable interest, is discussed in Section VI,B. Recent studies have confirmed

24

LEILA DIAMOND

et al.

and extended some of their findings. For example, it has been shown that ODC induction by TPA is reduced greatly when mice are kept on a diet deficient in vitamin B6 (Murray and Froscio, 1977), an essential cofactor for ODC. Another observation is that skin wounding, a promoting stimulus, induces ODC, whereas skin massage does not induce the enzyme or promote (Clark-Lewis and Murray, 1978). ODC is also induced by TPA in primary cultures of newborn mouse epidermis (Yuspa et al., 1976b) (see Section V,C,2). ii. Proteases. The suggestion that proteases may play a role in the mechanism of tumor promotion (Troll et al., 1975, 1978) came from initiation/promotion experiments in mouse ear skin that showed a delay in the appearance of tumors when such synthetic protease inhibitors as the chloromethyl ketones of tosyl lysine (TLCK) and tosyl phenylalanine (TPCK) were applied to the skin after applications of croton oil (Troll et al., 1970). However, it is difficult to interpret these data, for the inhibitors are not specific for proteases and probably attack other protein species in the epidermis. In other experiments in mouse dorsal skin (Hozumi et al., 1972), it was found that protease activity increases after TPA treatment, although at the time the measurements were made (24 hours after TPA) infiltrating leukocytes may have accounted for much of the increased activity, as Troll et al. (1970) had originally suggested. iii. Histiduse. A decrease in epidermal histidase activity has been observed after promoter treatment (Colburn et al., 1975). This enzyme is normally present in adult epidermis, but little activity is found in fetal tissue or skin tumors (Baden et al., 1968). The decreased activity in response to promoters may be related to the stimulation of keratinization, as this enzyme may help to regulate the amount of histidine available for the synthesis of “histidine-rich protein” (Hoober and Bernstein, 1966; Sibrack et al., 1974). i v . Chalones. The concept that natural, tissue-specific substances called chalones regulate the balance between cell loss and cell division in differentiating tissue (Bullough and Laurence, 1964) is perhaps best illustrated in epidermis. It has been proposed (Marks, 1971) that the epidermis contains both a G , chalone that inhibits the GI + S transition and a G2 chalone that controls progression from G , + mitosis. The stimulation of epidermal DNA synthesis by TPA is not subject to inhibition by epidermal GI chalone for about two days and then the tissue regains its sensitivity to the chalone (Krieg et al., 1974). Marks (1976) and Marks et al. (1978) have suggested that a disturbance of this homeostatic mechanism may be involved in tumor promotion, as discussed in Section V1,C. c . Effects on Small Molecules. i. Cyclic nucleotides. The effects of

TUMOR PROMOTERS

25

tumor promoters on mouse skin are similar in several respects to the effects of hormones on their target tissue. An early event in some hormone-stimulated systems is a change in cyclic nucleotide levels (see Robison et a/., 1971). Cyclic nucleotides also are important regulators of epidermal physiology (Voorhees et a/., 1974). Several investigators, therefore, have studied the effects of promoters on various aspects of cyclic nucleotide metabolism. Grimm and Marks (1974) found that after the application of TPA there was an early rise in epidermal cAMP levels, with a subsequent decline to below control levels at 9-24 hours. Mufson et a / . (1977b) found no change in the level of cAMP during the first 24 hours after TPA treatment; Belman et a / . (1978) reported that TPA had no effect on epidermal cAMP levels at any time up to 72 hours after treatment. Despite these conflicting results (probably due to the use of different methods), a consistent finding from several laboratories is that promoters render the skin incapable of responding to p-adrenergic agonists (Marks and Grimm, 1972; Grimm and Marks, 1974; Verma and Murray, 1974; Mufson et a/., 1977b). Such catecholamines as isoproterenol and epinephrine usually induce a rapid rise in epidermal cAMP that is mediated by p-adrenergic receptors (Marks and Grimm, 1972); this effect is blocked by prior treatment of the skin with TPA or other promoters. The p-adrenergic block appears not to be specific for tumor promotion for acetic acid has a similar effect (Verma and Murray, 1974). Cyclic nucleotides probably do not play a role in TPA-mediated induction of ODC (Mufson et a/., 1977~1,but the question of whether they act as a “signal” or “second messenger” for any of the other biochemical effects of promoters is still open. ii. Prostaglandins. High doses of indomethacin, a potent inhibitor of a key enzyme in prostaglandin synthesis (Robinson and Vane, 1974), inhibit the induction of ODC by TPA (Verma et a / . , 1977); this inhibition is counteracted by treatment of the skin with prostaglandins El and E2, but not prostaglandins of the F series. High doses of indomethacin also block TPA-induced DNA synthesis (Furstenberger and Marks, 1978), and this also is prevented by administering prostaglandin E2simultaneously with the TPA. These studies suggest that prostaglandins may mediate some of the short-term biochemical effects of TPA in mouse epidermis. However, Furstenberger and Marks (1979) have shown that TPA does not compete with prostaglandin El for its membrane receptor in isolated rat adipose cells, suggesting that the phorbol diesters are not prostaglandin agonists as Smythies et a f . (1975) had proposed. That prostaglandins may play a role in the mechanism of promotion is supported by the finding that they act as cocarcinogens in mouse skin (Lupulescu, 1978), but is perhaps refuted by indomethacin’s only slight in-

26

LEILA DIAMOND et

al.

hibitory effect on tumor promotion by TPA (Viaje e f al., 1977). Although the levels of prostaglandins in promoter-treated mouse skin have not been reported, in human skin the levels are increased in the involved areas of psoriasis (Hammarstrom et al., 1975), a disease characterized by increased epidermal cell proliferation. d . Eflects on Cell Membranes. A number of studies have attempted to determine whether the primary or predominant effect of tumor promoters is mediated through a specific interaction with the plasma membrane, whether the promoters have more affinity for some specific regions or entities in the cellular membranes than for others, and what the nature and function of these hypothetical sites might be. For example, the studies described earlier showed that there was a decreased responsiveness of promoter-treated skin to p-adrenergic agonists and chalones. The induction of ODC by TPA is largely prevented by prior administration of colchicine (O’Brien et al., 1976); this has been interpreted to mean that the membrane cytoskeletal elements responsible for signal transmission from membrane to nucleus had been disrupted by the colchicine. Because phorbol diester tumor promoters have highly lipophilic portions of the molecule, they would be expected to be associated more with the membranous fractions of cells and tissues than with the cytosolic compartment. Helmes et al. (1974) found that in mouse skin treated with tritiated phorbol- 12,13-didecanoate, the radioactivity in subcellular fractions paralleled the membrane content of the fractions, an indication of the difficulties in determining specific interactions of these compounds with membrane components. There is other indirect evidence for the interaction of phorbol diester tumor promoters with cell membranes: (1) The incorporation of [3H]choline (Suss et al., 1971) and 32Pi(Rohrschneider et al., 1972; Balmain and Hecker, 1974) into membrane phospholipids is enhanced by promoters [although some of the observed increase may be due to changes in pool size (Balmain and Hecker, 1974)]. (2) Purified rat liver microsomal membranes exposed to phorbol diesters show a change in buoyant density and an altered affinity for DNA (Kubinski et al., 1973). (3) Exposure to TPA decreases the native fluorescence of rat liver plasma membranes, suggesting an altered conformation of proteins in the lipid bilayer (Van Duuren et d.,1976). (4) The fluorescence polarization of the probe 1,6-diphenyl-1,3,5-hexatriene decreases after treatment of lymphoblastoid or rat embryo cells with promoters but not with nonpromoters (Fisher et al., 1979; Castagna e f a l . , 1979a). Fisher et al. (1979) propose that the lipid moieties of the cell membrane are more “fluid” in the presence of the promoters.

TUMOR PROMOTERS

27

Studies on the interaction of tumor promoters with the plasma membrane of intact cells in culture are described in Section V,C,l.

3 . Modifiers of Tumor Promotion in Mouse Skin Experimental modification of a biological process can lead to increased understanding of the basic mechanism(s) involved. This section briefly reviews studies on the modifying effects on tumor promotion of specific chemicals; it does not review studies on other types of modifiers of tumor formation such as caloric restriction (see Tannenbaum, 1959; Clayson, 1975). Tumor promotion is inhibited by cortisone (Ghadially and Green, 1954) and such synthetic glucocorticoids as dexamethasone (Belman and Troll, 1972) and fluocinolone acetonide (Schwarz et al., 1977). Topical application of dexamethasone completely suppresses TPA-mediated promotion (Scribner and Slaga, 1973). Fluocinolone acetonide is as effective as dexamethasone at lower dose levels and also blocks promoter-induced hyperplasia and DNA synthesis (Schwarz et al., 1977). These and other studies (Slaga et al., 1975) have shown that steroids that are effective inhibitors of promotion also inhibit promoter-induced cell proliferation, hyperplasia, and inflammation. This may, in fact, explain why they are antipromoters, since cell proliferation is a sine qua non for promotion. Another class of antipromoters are the retinoids, which probably act by a mechanism different from that of the steroids. Although both prophylactic and therapeutic effects on skin cancer have been ascribed to the retinoids (Bollag, 1972, 1974), their mechanism of action is not known. With a series of natural and synthetic retinoids, there is a striking correlation between the ability to inhibit ODC-induction by TPA in mouse skin and an inhibitory effect on promotion, as measured by the formation of skin papillomas (Verma e f al., 1978, 1979). The active retinoids are effective at doses that cause no local or systemic toxicity and have little apparent effect on TPA-induced hyperplasia. Verma et ul. (1978, 1979) have suggested that the inhibitory effect of retinoids on promoter-induced ODC activity might be exploited to screen synthetic retinoids for potential antipromoting activity in uivo. Retinoids that inhibit this activity in mouse skin also inhibit some biological effects of promoters in cell culture (Kensler and Mueller, 1978; Wertz et ul., 1979). Another use of modifiers in the study of promotion is to inhibit a specific, or at most, a limited number of, cell function(s1 and to determine the short-term (biochemical) and long-term (tumorigenesis) effects. The

28

LEILA DIAMOND

et al.

results of such studies, however, are sometimes difficult to interpret. As described above, synthetic protease inhibitors such as TLCK and TPCK have antipromoting activity (Troll et al., 1970). The natural protease inhibitor, leupeptin, also inhibits tumor promotion in mouse skin (Hozumi et al., 1972). However, clear-cut evidence that tumor promoters stimulate protease activity in mouse epidermis is still lacking. Nevertheless, these modifiers can be useful for studying the mechanism of promotion as long as it is recognized that they probably not only inhibit proteases but affect the functions of some other proteins as well. The effect of the prostaglandin biosynthesis inhibitor, indomethacin, on the action of promoters in mouse skin is discussed in Section IV,A,c. The fact that at relatively low doses it is only a weak inhibitor of promotion (Viaje et d.,1977) suggests that its inhibition (at high doses) of the biochemical effects of promoters (Verma et d . , 1977; Furstenberger and Marks, 1978) may be mediated by effects on cell functions other than prostaglandin synthesis. Many other agents have been tested for antipromoting activity, including butyric acid (Belman and Troll, 1974), theophylline (Belman and Troll, 1974), polyinosinic-polycytidylic acid (Elgjo and DegrC, 1973), and Bacillus Calmette-Guerin (BCG) vaccine (Schinitsky et a / . , 1973). The most effective inhibitors of promotion so far have been retinoids and the antiinflammatory steroids. Future studies on the relationship between structure and activity of these two classes of compounds and on their mode of action can lead to increased understanding of the mechanism of promotion.

B. OTHERMODELSOF TWO-STAGECARCINOGENESIS Several model systems that also demonstrate the two distinct stages of carcinogenesis, initiation and promotion, have been described within the past few years (see Table V). Some may be too recent for the identification of the most effective initiators or promoters and for unequivocal evidence that the promoting compound is acting as a pure promoter and not only as a cocarcinogen. Nevertheless, these model systems and other studies with phorbol and TPA (Table V) show that the phenomenon of two-stage carcinogenesis as described in mouse skin is not limited to that tissue, but can be extended to many other species and tissues and to promoting compounds of varied chemical structure. The large number and diversity of experimental models subject to "promotion" also strongly suggest that environmentally induced human cancer involves similar stages in its development.

TUMOR PROMOTERS

29

1. Skin

Mice are the only species in which two-stage carcinogenesis in skin with phorbol diesters as the promoting agents has been clearly demonstrated (Shubik, 1950; reviewed in Van Duuren, 1969). However, no other species has been as thoroughly evaluated for susceptibility to this protocol and differences in the structure of the epidermis or in the ability to metabolize initiators or promoters may account for the failure to reproduce this system in other species. Yuspa e f a / . (1979) are developing an in vivo model of carcinogenesis for human skin in which neonatal foreskins are grafted to athymic nude mice. The grafts can be maintained morphologically intact for about 6 months and respond to topical application of TPA with hyperplasia. Recently, Goerttler et a / . (1979b) reported two-stage tumorigenesis of dermal melanocytes in the dorsal skin of Syrian hamsters initiated with 7,12-dimethylbenz(a)anthracene administered intragastrically and promoted topically with TPA. The dorsal epidermis did not show morphological alterations after TPA treatment nor did tumors develop in the epidermis. The benign melanomas that developed underwent malignant transformation when transplanted to normal hosts. 2. LiveiRecent studies suggest that the development of experimental liver cancer in rats proceeds through several stages, two of which may be analogous to initiation and promotion in mouse skin [see Pitot (1979) for a recent review and Table V]. The first evidence for this was presented by Peraino et a / . (1971 , 1973) who reported a promoting effect of dietary phenobarbital on liver tumorigenesis in rats previously fed the carcinogen 2-acetylaminofluorene. The enhancing effect of phenobarbital is most clearly demonstrated when the prior regimen of 2-acetylaminofluorene is limited to only a few weeks and itself results in only a few tumors. Friedrich-Freksa et a/. (1969) were the first to describe the appearance in the liver of small foci of enzyme-deficient cells after carcinogen treatment. These and other investigators (Farber, 1976; Cameron et a/., 1976; Pitot e f a / . , 1978) have studied the biochemical and biological properties of these foci and such other early lesions as hyperbasophilic foci, hyperplastic foci, and hyperplastic nodules. Pitot et a / . (1978) used several “markers” (deficiency of glucose-6-phosphatase and cannicular ATPase, presence of y-glutamyl transpeptidase) to follow the development of these enzyme-altered islands during hepatocarcinogenesis. They have found that phenobarbital itself does not induce their formation in the liver

30

LEILA DIAMOND et

al.

but that, when fed to rats after they are treated with 2-acetylaminofluorene or other liver carcinogens, it does enhance the number and size of these foci and inhibit their phenotypic reversion. Taken together with the findings of Peraino et al. (1971, 1973, 1978) and Kitagawa and Sugano (19781, these data indicate that phenobarbital can promote carcinogenaltered cell populations into autonomous neoplasms. Little is known about the mechanism of the promoting action of phenobarbital. It does not appear to involve hepatotoxicity (see Peraino et id., 1978). Phenobarbital is not a carcinogen or mutagen (Peraino et al., 1975; McCann and Ames, 1976). It is an inducer of several drug- and carcinogen-metabolizing enzymes (see Schulte-Hermann, 1974);this does not account for the promoting effect, however, since tumor formation is inhibited rather than enhanced when phenobarbital is administered at the same time as the carcinogen (Peraino et al., 1971). Furthermore, the promoting action of phenobarbital can be demonstrated even when several weeks of control diet intervene between carcinogen and phenobarbital treatment (Peraino et al., 1973). Farber and his colleagues (Solt and Farber, 1976; Solt et al., 1977; Farber and Soh, 1978) have described another experimental protocol for studying the stages of hepatocarcinogenesis. It is based on the observation that putative preneoplastic hepatocyte populations are resistant to the cytotoxic effects of carcinogens (Cameron et al., 1976). The development of “carcinogen (diethy1nitrosamine)-initiated hepatocytes” is selectively stimulated by exposure to a toxic growth inhibitor, (e.g., a low level of 2-acetylaminofluorene in the diet) followed by partial hepatectomy. The hypothesis is that under these conditions normal liver cells activate the 2-acetylaminofluorene to a metabolite that inhibits their growth, whereas “altered” or “initiated” cells are resistant, perhaps because they have a decreased ability for activation, and proliferate rapidly to form foci and hyperplastic nodules. The resistant cells can be stimulated to undergo at least 15 cycles of cell proliferation without expressing evidence of autonomous growth; their eventual autonomous growth is a late phenomenon in their evolution to cancer. This is an excellent model for studying initiation, because foci of “altered” cells can be induced relatively rapidly with several classes of hepatocarcinogen. However, only a small fraction of foci develop into tumors, even with the enhancing stimulus of a cytotoxic environment and partial hepatectomy (Soh et al., 1977). It may be that the majority of these foci would be sensitive to the additional enhancing action of promoting agents such as phenobarbital; this would then be an even more valuable model for studying the mechanism of promotion in liver.

TUMOR PROMOTERS

31

Many investigators (Bannasch, 1976; Soh et a/., 1977; Watanabe and Williams, 1978; Pitot et a / . , 1978) have asked whether the enzyme-altered islands induced by a carcinogen alone or by a carcinogen and promoter (phenobarbital) are the immediate progeny of initiated hepatocytes and precursors of hepatocellular carcinomas. Pitot et al. (1978) have shown that the islands, like hepatocellular carcinomas, are phenotypically heterogeneous with respect to several marker enzymes and presumably to other markers as well. However, the biochemical phenotype of the carcinomas is maintained through several hundred transplant generations (Pitot et u/., 1978), suggesting that it may be fixed in initiated hepatocytes soon after initiation. Thus, it may be possible to identify initiated cell populations in the liver and determine the biological and biochemical characteristics that distinguish those altered foci that can be “promoted“ to carcinomas from those that cannot. Various selective procedures for growth of “premalignant” cells such as in vivo + in vivo (Laishes and Farber, 1978) and in vivo + in vitro (Borenfreund et al., 1977; Laishes e f a / ., 1978)transfer of “carcinogen-initiated” hepatocytes can be combined with biochemical techniques to study both the characteristics of the various cell types at different stages in the development of hepatomas and factors such as promotion that mediate development from one stage to the next.

3. Bladder Hicks et u/. (1973)provided the first evidence for the multistage nature of experimental bladder cancer in rats. They showed that a single instillation into the bladder of the nitrosamide, methylnitrosourea, produces few tumors unless the rats are subsequently maintained on a diet containing either of the artificial sweeteners, sodium saccharin or sodium cylcamate, in which case there is a high incidence of bladder tumors (Hicks et a / . , 1975, 1978; Hicks and Chowaniec, 1977). Cohen et al. (1979) showed that high doses of dietary DL-tryptophan also may act as a “promoter” of urinary bladder carcinogenesis in rats that have been (FANFT) as the first fed N-[4-(5-nitro-2-furyl)-2-thioazolyl]formamide initiator. Control experiments to rigorously demonstrate that tumor promotion analogous to the mouse skin model does occur in this tissue (e.g.., exposure to promoter first and then to initiator) have not yet been done. However, the general features of the model indicate that qualitatively different stages occur in the development of the bladder tumors and that different chemicals can mediate each stage.

32

LEILA DIAMOND et

ul.

V. Effects of Tumor Promoters on Cells in Culture

Cell and organ culture systems offer to the study of chemical carcinogenesis and tumor promotion the advantages and disadvantages they offer to the study of other biological phenomena at the cellular and molecular levels. Ideally, in vitro model systems for studying a particular phenomenon will reproduce the biological and biochemical responses that occur in vivo. A clear-cut demonstration of two-stage transformation, analogous with respect to timing and dosage to two-stage carcinogenesis in vivo, has been described with a cell clone derived from a pool of mixed embryo fibroblasts (Mondal and Heidelberger, 1976; Mondal et al., 1976). However, other cell culture systems are needed that are more analogous to in vivo tumor promotion models. As discussed below, studies to improve techniques for culturing such epithelial tissues as epidermis and liver are proceeding rapidly and cell culture systems that closely approximate the tissues and conditions for two-stage carcinogenesis in vivo probably will be developed soon. One of the problems in developing in vifro systems for the study of the mechanism of tumor promotion is that very precise experimental conditions are required to demonstrate the actual promotion (conversion) of initiated cells. This effect of tumor-promoting compounds must be distinguishable from the effects of the compounds on the progression or expression of “premalignant” cells already destined to become malignant, effects that may not be unique to promoters. As described below, the phorbol diester tumor promoters have effects on cells that may very well enhance the expression of transformation but may not be involved in the specific phenomenon of conversion (Boutwell, 1964). Most in vitro studies with tumor promoters have used the phorbol diesters, and in particular, the most potent promoter in the series, TPA. Very few of these studies, however, have used a two-stage transformation system or even cell types in which two-stage carcinogenesis and tumor promotion have been demonstrated in vivo. Instead, the effects induced by phorbol diesters have been studied in such diverse cell types as chick embryo fibroblasts and Friend erythroleukemia cells. Such studies have led to the realization that these compounds are powerful tools for studying not only promotion but also such biological phenomena as cell proliferation and differentiation. Although they also have led to hypotheses for the mechanism of tumor promotion, the testing of these hypothesis requires the development of epithelial cell systems in which two-stage transformation can be induced at will. Such model systems are also needed to help define the structural requirements for the promoting ac-

T U M O R PROMOTERS

33

tivity of plant esters and to identify other types of compounds with promoting activity. This section is divided into the specific biochemical and biological effects of phorbol diester tumor promoters; however, these effects are often overlapping events that have simply been investigated separately. In general, when a series of phorbol diesters has been tested for a specific effect in lvitro, a positive correlation has been found between promoting activity in mouse skin and the ability of a compound to induce that effect in vitro (see, e.g., Yuspa et a / . , 1976b; Rovera et al., 1977).

A. I n Vitro MODELSOF TWO-STAGE CARCINOGENESIS Sivak and Van Duuren (1967) were the first to look at the effects of a tumor-promoting compound on cells in culture. They showed that in a mixed population of contact-inhibited 3T3 (Swiss) mouse fibroblasts and spontaneously- or SV40-transformed cells, outgrowth of the transformed cells was enhanced by a purified fraction of croton resin. With contactinhibited 3T3 cells in which a few transformed foci appear spontaneously, the number of foci was not increased by exposure to 7,12-dimethylbenz(a)anthracene alone or prior to exposure to the resin; that is, this was not an in iiitro initiation/promotion transformation system. An extension of this observation that tumor promoters enhance phenotypic expression of transformation is the report of Fisher et a / . (1978) that TPA added after infection enhances 2-3 fold the number of transformed foci in rat embryo cell cultures infected with adenovirus type 5. TPA does not affect the cloning efficiency of adenovirus-transformed cells grown either alone or with an excess of rat embryo cells, indicating that the enhancement of adenovirus transformation by TPA is probably due to its ability to facilitate expression of the transformed state. The first description of what appeared to be two-stage malignant transformation of cells in culture was that of Lasne et al. (1974). They reported acceleration of the time required for transformation of rat embryo fibroblasts when they were treated first with benzo(c1)pyrene and then, beginning at passage 16, were subcultured in TPA-containing medium. In these experiments, and in a later, more extensive study (1977), these authors found that transformation occurred eventually in most experimental groups, including those treated with carcinogen alone, TPA alone, or untreated. However, it occurred earliest in those groups treated with both an initiator and a promoter, leading them to conclude that they had reproduced two-stage carcinogenesis in tissue culture.

34

LEILA DIAMOND

et a / .

More convincing evidence that two-stage transformation in vitro can be obtained that is analogous with respect to timing and dosage to two-stage carcinogenesis in vivo was presented by Mondal and Heidelberger (1976) and Mondal et a / . (1976). In cultures of a clone of the C3H mouse fibroblast cell line C3H/10T% (Reznikoff e f a / . , 1973) treated (initiated) with nontransforming doses of carcinogenic hydrocarbons or with UV irradiation, continuous treatment with phorbol diester tumor promoters beginning 2-4 days later produces transformation. The promotion of transformation is not due to a simple stimulation of cell division by the promoters (Mondal et N / . , 1976). Saccharin also enhances transformation of these cells when they are pretreated with a nontransforming dose of 3-methylcholanthrene but not with UV irradiation (Mondal et a / . , 1978). X-ray-induced transformation of C3HIlOTM cells is also enhanced by TPA (Kennedy et a / . , 1978) but not as much as is UV-irradiation- or carcinogen-induced transformation. The promoter acts most effectively on cells treated with doses of X-irradiation that by themselves yield low levels of transformation, and is as effective when added to the cultures immediately after irradiation as when added 48 or 96 hours later. The protease inhibitor, antipain, suppresses both X-ray-induced transformation of C3H/10T% cells and its enhancement by TPA (Kennedy et ul., 1978; Kennedy and Little, 1978, 1979). Several groups are attempting to develop two-stage carcinogenesis models in epithelial cell systems, but as yet no such model is available. The tissues being used include rat trachea (Steele et a / . , 1978a), mouse submandibular gland (Knowles, 1979), and mouse epidermis (Colburn et a / ., 1978a), the latter being analogous to the classic initiation/promotion in vivo model. Procedures for the preparation of primary cultures of newborn mouse epidermis have been developed by two groups (Yuspa and Harris, 1974; Fusenig and Worst, 1974, 1975). The primary cells differentiate in r~itrowith the production of keratin proteins that are identical to those of the stratum corneum of mouse skin (Yuspa and Harris, 1974: Steinert and Yuspa, 1978). Cells derived from cultures of BALB/c 3T3 mouse epidermis have been malignantly transformed after treatment with N-methyl-N’-nitro-N-nitrosoguanidine (MNNG)followed by repeated TPA treatment; the transformation process required 3-5 months (Colburn et N / . , 1978a). This study did not establish, however, that treatment with TPA was required for transformation of the MNNG-initiated cells. Several mouse epidermal cell lines have been developed that are non-tumorigenic at early passages but subsequently undergo a TPA-dependent conversion to the neoplastic phenotype (Colburn, et UI., 1978, 1979). Promoter-sensitive cell lines

TUMOR PROMOTERS

35

such as these will be useful for studying sequential events during transformation. An organ culturekell culture system for studying transformation of rat tracheal epithelium has been developed by Nettesheim and co-workers (Marchok et a/., 1977; Steele et a/., 1977). Although permanent epithelial cell lines could not be established from solvent-treated control explants, they were established from tracheas exposed to TPA (Steele et a/., 1978a,b). Tracheal explants were treated every 6 days for 1 hour to stimulate DNA synthesis and, after several weeks, the epithelial cells were allowed to grow out on a substrate. After 5-6 weeks, the primary cultures showed altered morphology and grew more rapidly than control cells. The cell lines derived from these cultures had growth and morphological characteristics similar to those derived from carcinogen-treated tracheal explants, but even one year after TPA exposure, they still did not grow in soft agarose or produce tumors in immunosuppressed recipients (Steele et a / . , 1978b). Epithelial cultures derived from tracheal explants exposed to MNNG underwent subsequent malignant transformation with a dose-dependent response to MNNG (Steele et a/., 1978a). With explants exposed to low concentrations of MNNG followed by TPA, the promoter increased the frequency of subsequent transformation and decreased the time at which it could first be observed (Steele and Marchok, 1979). Thus, this system has the potential to develop into a model for two-stage carcinogenesis. In addition, it will be interesting to determine how TPA enhances the ability of tracheal epithelium to become established as permanent, nonmalignant cell lines in vitvo. Other epithelial cell culture systems that might eventually be developed into two-stage carcinogenesis models include hamster (Sisskin and Barrett, 1979)and human (Rheinwald and Green, 1975: Sun and Green, 1976) epidermis; rat urinary bladder (Hashimoto, 1978); mouse (Lin et al., 1976; Tonelli et a / . , 1979) and rat (Richards and Nandi, 1978) mammary gland, and rat hepatocytes (Michalopoulos and Pitot, 1975; Williams, 1976: Laishes et al., 1978), perhaps in some cases after initiation or initiation and promotion in vivo.

B. EFFECTSOF PHORBOL DIESTERTUMOR PROMOTERS O N CELL MORPHOLOGY The potent promoter TPA can have striking, reversible effects on the morphology and size of cells growing in monolayer culture. These effects are most apparent in sparse, growing cultures, although similar changes

36

LEILA DIAMOND

et al.

may occur in confluent cultures treated with TPA [e.g., human or hamster fibroblasts (L. Diamond and T. G. O’Brien, unpublished)]. Within a few hours after growing cultures of human diploid fibroblasts or chick embryo fibroblasts (CEF) are exposed to low concentrations of TPA, the cells become elongated, the number of narrow cytoplasmic processes increases, and the nuclear and cytoplasmic membranes are more distinct than in control cultures (Diamond et al., 1974; Driedger and Blumberg, 1977).TPA-treated CEF assume a criss-crossed, disoriented arrangement similar to but distinguishable from that induced by transformation with Rous sarcoma virus (RSV) (Driedger and Blumberg, 1977; Wilson and Reich, 1979). They lose the ordered actin-containing cytoskeletal structures found in untreated normal cells; the actin pattern, as visualized by indirect immunofluorescence with antiactin antibodies, is diffuse and resembles that of RSV-transformed cells (Rifkin et al., 1979). The mean cell volumes of TPA-treated human and chick fibroblasts are 25-30% less than those of control cells (Diamond et al., 1974; Driedger and Blumberg, 1977). Driedger and Blumberg (1977) determined that the protein per cell of TPA-treated CEF is also one-third less than that of control cells and concluded that the TPA-induced decrease in cell volume is due to changes in the overall composition of the cell. Similar changes in cellular morphology, refractility , and volume occur in 3T3 mouse fibroblasts treated with TPA (Sivak e f al., 1969; Diamond et al., 1974; Sivak, 1977). The promoter also alters the morphology of mouse epidermal cells in culture, with the cells shifting from a clearly epithelial to a fibroblastic appearance with wider intercellular spaces (Yuspa et al., 1976a); these changes are similar to those seen in mouse skin treated with TPA (Raick, 1973a). The promoter has little effect on the morphology of CEF from subdermal connective tissue but induces striking morphological changes in chick chondroblasts from vertebral cartilage (Lowe et al., 1978). The polygonal, epithelioid chondroblasts are transformed into fibroblastic cells with long, fine overlapping processes, some as long as 50 pm. These effects of TPA on cell morphology and volume are not the result of an overall generalized toxicity, since in low-density cultures of some cells [e.g., CEF (Driedger and Blumberg, 1977)], the growth rate in TPAcontaining medium may be no different than in control medium. In others [e.g., 3T3 mouse fibroblasts and WI-38 human fibroblasts (Diamond et al., 1974)], TPA may cause an inhibition of cell division for up to 48 hours, after which the growth rate in TPA-containing medium is similar to that of untreated cultures; this occurs even with cells such as human fibroblasts, which do not deplete the TPA in the medium by metabolic degradation (O’Brien and Diamond, 1978a,b). These effects appear to be

TUMOR PROMOTERS

37

due instead to a direct interaction of the compound with the plasma membrane; this induces an immediate "shock" or paralysis of the cells from which they gradually recover even in the continued presence of the compound. Similar morphological changes occur in cells treated with other tumor-promoting phorbol diesters (L. Diamond and T. G. O'Brien, unpublished); it is not clear if they are a prelude to some or all of the other effects the promoters induce in cells, effects that may be apparent only much later. It may be possible to answer this question in cell systems in which such clearly membrane-associated phenomena as the induction of adherence can be dissociated from phenomena that clearly require altered gene expression. Growing cultures treated with TPA may eventually achieve higher cell densities than control cultures (Diamond et a / . , 1974; Lowe et ril., 1978). Diamond et ril. (1974) have suggested that the TPA-induced early decrease in the surface area covered by each cell may enhance final densities by delaying the onset of contact- or density-dependent growth inhibition in such cells as human and 3T3 mouse fibroblasts that are sensitive to such controls. TPA-treated BALB/c 3T3 cells do, however, retain their density-dependent growth controls. Although the first exposure of confluent cultures to TPA results in the morphological changes described above and an approximate doubling of cell number, subsequent treatments with TPA induce less striking morphological alterations and no additional cycles of cell division (Sivak, 1972, 1977; Diamond et al., 1977). The early observation of Sivak and Van Duuren (1967) that a croton oil derivative can enhance the outgrowth of transformed Swiss 3T3 cells in a mixed population containing an excess of contact-inhibited fibroblasts suggests that transformed cells may not be sensitive to a "block" in untransformed cells that prevents repeated stimulation of cell division by TPA. In addition to the rapid morphological changes described above, CEF infected with RSV show later TPA-induced morphological changes that resemble those induced by transformation with this virus. In CEF infected with a temperature-sensitive RSV mutant (ts68) and maintained at the temperature permissive for transformation (37°C) or infected with wild-type RSV, treatment with TPA accentuates such morphological alterations induced by virus transformation as the spherical cell shape, multilayering, and disorganization of the monolayer (Wilson and Reich, 1979; Quigley, 1979). Under these conditions, the cells either detach or form large multicellular aggregates and networks of these aggregates, probably because of the high levels of plasminogen activator induced by the combination of TPA and virus transformation (Wilson and Reich, 1979; Quigley, 1979). These morphological alterations are inhibited by

38

LEILA DIAMOND

et ul.

several protease inhibitors, whereas the early cellular elongation that follows TPA treatment of these and many other cells is not prevented by protease inhibitors (Quigley, 1979). The morphological changes induced by TPA in CEF infected with the ts68 RSV mutant and maintained at the nonpermissive temperature (41°C) are similar to, but more pronounced than, those seen in uninfected CEF treated with TPA (Wilson and Reich, 1979), indicating that accentuation of the transformed phenotype by TPA in RSV-infected CEF requires expression of the virus genome.

C. BIOCHEMICAL EFFECTSOF PHORBOL DIESTERTUMOR PROMOTERS 1. Interactions with the Plasma Membrane

One of the first reports of TPA’s effects on the membranes of intact cells in culture described its stimulation of incorporation of [3H]choline into the phospholipids of HeLa cells (Suss et a / . , 1971). The effect was rapid and occurred with very low concentrations of TPA. The promoter also stimulates incorporation of [3H]choline into the phospholipids of bovine lymphocytes (Wertz and Mueller, 1978). TPA stimulates uptake of 86Rb+,a K+ analog, and of 32Piby mouse (Swiss) 3T3 cells (Moroney et nl., 1978). The stimulation of uptake is ouabain-sensitive, suggesting that the (Na+,K+)-ATPase is the target of the TPA. TPA can have marked effects on glucose transport and metabolism. It stimulates uptake of the glucose analog 2-deoxyglucose in CEF and rodent and human cell cultures (Driedger and Blumberg, 1977; Lee and Weinstein, 1979). In human polymorphonuclear leukocytes, TPA mimics the effect of phagocytosis and greatly stimulates oxidation of glucose via the hexose monophosphate shunt (Repine et al., 1974; DeChatelet et ul., 1976). It also stimulates the decarboxylation of 2-deoxyglucose in phagocytic leukocytes (Zabos et ul., 1978). In BALB/c 3T3 preadipose cells, TPA increases uptake of 2-deoxyglucose (T. G. O’Brien, unpublished) and stimulates lactate production from glucose (O’Brien et ul., 1979a). The stimulation of glucose metabolism by TPA can complicate the interpretation of sugar uptake measurements, particularly with such substrates as 2-deoxyglucose which may be further metabolized beyond the initial phosphorylated species. TPA and other phorbol diester tumor promoters can stimulate prostaglandin synthesis in cells in culture. The most responsive of the cells studied by Levine and Hassid (1977)is the canine kidney cell line MDCK

TUMOR PROMOTERS

39

in which chemical carcinogens also stimulate prostaglandin production (Levine, 1977). TPA also increases prostaglandin synthesis in normal and transformed mouse fibroblasts, human fibroblasts, and bovine endothelial cells (L. Levine, L. Diamond, and T. G . O’Brien, unpublished), and in mouse peritoneal macrophages in vitro (Brune et a / . , 1978). Thus, this may be a general response of cells to promoter treatment. Ohuchi and Levine (1978a) have shown that TPA stimulates synthesis of prostaglandin in MDCK cells by causing deacylation of membrane phospholipids. This increases the availability of free arachidonic acid, which can be converted to endoperoxides by a membrane-bound cyclooxygenase; the endoperoxides are then converted enzymatically to prostaglandins, thromboxanes, or prostacyclin. Ohuchi and Levine (1978a) suggest that the high efficiency of arachidonate conversion to prostaglandins in the presence of TPA reflects an ability of the promoter to bring together spatially the deacylating and endoperoxide-generating systems, resulting in a coupled reaction. Retinoids also stimulate deacylation of cellular lipids and prostaglandin production of MDCK cells; the combined stimulatory effects of TPA and a trimethylmethoxyphenyl analog of retinoic acid are synergistic (Levine and Ohuchi, 1978). TPA stimulates release into the culture medium of free arachidonic acid and its metabolites from [3H]arachidonic acid-labeled MDCK cells, but not the release of radioactivity from [14C]linoleicacid-labeled cells (Ohuchi and Levine, 1978b). Since linoleic acid probably occupies the same C-2 position as arachidonate in membrane phosphoglycerides, TPA may be increasing specifically the activity of a phospholipase A2 that has a greater affinity for arachidonate than for linoleic acid ester bonds. Other possible sites for interaction of tumor promoters with cell membranes are hormone receptors. Promoters may have an effect on these receptors that alters the cellular response to the hormones. One suggestion is that they may “take over” the receptors for some endogenous hormone, perhaps through structural resemblance to that hormone (Rohrschneider and Boutwell, 1973; Wilson and Huffman, 1976; Weinstein et ul., 1977). Lee and Weinstein (1978) reported that TPA and other tumor promoters inhibit the binding of the polypeptide hormone, epidermal growth factor (EGF), to the surface of HeLa cells by reducing the number of binding sites available. On the other hand, Brown et ul. (1979) and Shoyab er d.( 1979)reported that TPA decreases the binding of EGF by decreasing the affinity of cellular receptors for EGF without changing the total number of available binding sites. Brown er ul. (1979)found that in 3T3 mouse cells there are different mechanisms for the inhibition of binding of 1251-labeledEGF by TPA and native EGF, which suggest that the promoter does not bind directly to EGF receptors. Shoyab er ul.

40

LEILA DIAMOND et

al.

(1979) also found that TPA did not affect either the number or affinity of other membrane receptors for their specific ligands, as for example, concanavalin A, multiplication stimulating activity, and insulin. Thus, decreased affinity of membrane receptors for their ligands is not a general property of TPA-treated cells. At concentrations that inhibit neurite formation of nerve growth factorstimulated chick ganglia, TPA does not affect the amount of lZ5I-labeled nerve growth factor bound to specific sites on dissociated dorsal root ganglion cells (Ishii, 1978). In chick embryo muscle cells, TPA acts like RSV to cause both an increase in the rate of degradation of acetylcholine receptors and a decrease in the rate of receptor synthesis, thereby greatly reducing the steady-state level of surface receptors (Miskin et al., 1978a).TPA induces these effects in post mitotic myotubes, an example of the fact that cells that are unable to divide can respond to the promoter with a biological effect (see also Miskin et al., 1978b; Toyama et al., 1979). The effects of TPA on hormone binding may reflect a specific interaction with certain membrane receptors or, alternatively, may indicate that such receptor-ligand binding is very sensitive to a general membrane perturbation induced by TPA. Different receptors may be more or less affected, depending on their sensitivity to alterations in their lipid microenvironment. 2. Effects on Specific Proteins a . Ornithine Decurboxylase. Treatment of 24-hour-old primary cultures of newborn mouse epidermal cells with TPA (1.6 x lo-’ M ) induces a 5- 10-fold increase in the activity of the enzyme ornithine decarboxylase (ODC), with peak induction at about 9 hours (Yuspa et a / . , 1976b). DNA synthesis is stimulated in the treated cultures after 48-72 hours. Fluocinolone acetonide potentiates the increase in ODC activity but inhibits the stimulation of DNA synthesis (Lichti et al., 1977). In these epidermal cell cultures, promoting but not nonpromoting phorbol diesters induce ODC and stimulate DNA synthesis and, in this respect, the cultures resemble adult mouse epidermis in vivo (see Section IV,A,2). Other types of promoters such as anthralin and carcinogens such as 7,12-dimethylbenz(a)anthracene have little effect on ODC activity in these cultures, however, in contrast to their effects in vivo (Lichti et al., 1978). Phorbol diester tumor promoters also induce ODC in hamster embryo cell cultures, with the extent of induction generally being greater in transformed than in normal cells (O’Brien and Diamond, 1977). A striking difference between TPA’s effects on the two types of cells is that in

TUMOR PROMOTERS

41

transformed cells, TPA potentiates induction of ODC by fresh serumcontaining medium, whereas in normal cells, the effects of the two inducers are additive. In hamster cells, the induction of ODC by TPA is not followed by an increase in DNA synthesis (O’Brien and Diamond, 1977), in contrast to what occurs following ODC induction in many cells in vitro and in mouse skin in v i ~ w . h. LEZS Protein (Fihronectin). Tumor promoters induce transient effects on specific cellular proteins such as the major surface protein LETS (large-external-transformation-sensitive) (Vaheri et crl., 1978) and the serine protease plasminogen activator (PA); these effects are similar to the permanent effects induced by transforming viruses. The addition of TPA to exponentially growing cultures of CEF leads to a decrease in the amount of LETS protein on the cell surface, with the total amount remaining reduced by 90% in 3 days (Blumberg et a / . , 1976). The treated cells regain LETS protein within 3 days of trypsinization and replating in medium without TPA. c. Plasminogen Activator. TPA is an inducer of PA in a number of cell types, including CEF, HeLa cells,and HTC rat hepatoma cells (Wigler and Weinstein, 1976). The extracellular PA activity in cultures of vascular endothelium derived from rabbit vena cava is also increased by TPA (Loskutoff and Edgington, 1977). In cultured mouse macrophages, TPA acts in a manner similar to that of the lectin concanavalin A to stimulate synthesis and secretion of PA; the total fibrinolytic activity and the fraction of secreting cells is dramatically increased in cultures of macrophages from unprimed and endotoxin-primed mice, whereas macrophages from thioglycollate-primed mice show a high basal level of activity that is increased only slightly by TPA (Vassalli et al., 1977). In CEF, PA is induced by concentrations of TPA as low as 5 X lo-’ M ;requires de n o w RNA and protein synthesis; is observed as early as 3 hours after exposure to TPA: and persists as long as TPA remains in the medium but, after its removal, induced levels gradually return to control levels (Wigler and Weinstein, 1976; Weinstein et al., 1977). In chick embryo fibroblasts and myogenic cells transformed by RSV, TPA markedly increases the already elevated levels of PA (Weinstein et LII., 1977; Goldfarb and Quigley, 1978; Miskin et al., 1978b; Wilson and Reich, 1978, 1979). Transformation lowers the threshold concentration for response to the action of TPA and a clearly synergistic effect between transformation and TPA on PA synthesis can be demonstrated (Wilson and Reich, 1979). Quigley (1979) has shown that PA itself is the serine protease responsible for the “supertransformed” morphology of TPAtreated, RSV-infected CEF; it can act catalytically on an unknown cellular protein substrate that is not plasminogen and alter cellular behavior.

42

LEILA DIAMOND et a / .

Studies with a temperature-sensitive RSV mutant (ts68) show that the enhancement of PA synthesis by TPA requires continuous expression of the transforming gene src (Weinstein et ul., 1977, 1978a,b; Wilson and Reich, 1979) and that independent transcriptional events are required to mediate the individual responses to the virus and to TPA (Wilson and Reich, 1979). Elevated levels of CAMP inhibit enzyme induction by TPA or transformation in chick fibroblasts (Wilson and Reich, 1978, 1979) but not myogenic cells (Miskin et al., 1978b), suggesting that the mechanism of induction does not involve changes in cyclic nucleotide concentrations. Retinoic acid, which is also an inducer of PA production in CEF, does not affect PA production in normal or ts68-transformed CEF maximally stimulated by TPA (Wilson and Reich, 1978). However, with suboptimal concentrations of TPA, retinoic acid and TPA act synergistically to stimulate PA production, with an effect similar to that of either TPA or retinoic acid with virus transformation (Wilson and Reich, 1978). All three PA inducers require RNA synthesis for the reversal of their effect that occurs either spontaneously, after removal of the inducer, or after shifting to the nonpermissive temperature (Rifkin et a / ., 1975; Weinstein et al., 1977; Wilson and Reich, 1978). d . Membrane Glycopeptides. The biochemical profile of cell surface glycopeptides is altered in CEF transformed by RSV (Warren et a / . , 1972). This alteration is more pronounced when RSV-transformed cells are treated with TPA (Weinstein et a/., 1977, 1978a). 3. Effects on Replica five DNA Synthesis The effects of phorbol diester tumor promoters on the DNA synthetic phase of the cell cycle of cells in culture are complex. TPA has been considered to be mitogenic by some, whereas others have suggested that the observed stimulation of DNA synthesis is due to a partial synchronization of the cells that follows an initial inhibition of DNA synthesis by TPA (see discussion in Peterson et a / . , 1977). Both conclusions probably have some truth, since it appears that the cell type, such culture conditions as the presence of serum or specific serum factors, and the stage of the cell cycle all play a role in determining the effects of these compounds on DNA synthesis. A transient inhibition of thymidine incorporation after TPA treatment has been observed in monolayer cultures of HeLa cells (Suss et a / . , 1972), C3HIIOTM cells (Peterson et a/., 1977), and primary cultures of mouse epidermal cells (Yuspa et al., 1976a). The relative inhibition of thymidine incorporation appears to be greater in low-density cultures than in high, although this difference in sensitivity to TPA is not seen in

TUMOR PROMOTERS

43

suspension cultures of different densities (Kinzel et a/., 1973). The inhibition of thymidine incorporation by TPA usually results only in a temporary delay of growth and may be followed by considerable increase in the rate of DNA synthesis (Colburn et a/., 1978a; Fusenig and Samsel, 1978). In mouse epidermal cell cultures, the TPA-induced stimulation of thymidine incorporation is not simply an overshooting, synchronization effect after initial inhibition: there is an actual increase in the proliferative pool of the cell population (Fusenig and Samsel, 1978). TPA also has a transient inhibitory effect on proliferation of mouse skin fibroblasts in culture, but it is not followed by as great a stimulation of DNA synthesis as occurs with epidermal cells (Fusenig and Samsel, 1978). In some cells growing in complete serum-containing medium, TPA may have a stimulatory effect on DNA synthesis, but only after the cell growth rate slows down as the medium becomes depleted, that is, when the growth factors present in the medium are inadequate to maximally stimulate growth (Diamond et a/., 1974: Driedger and Blumberg, 1977; Frantz et a/., 1979). TPA does stimulate DNA synthesis in stationary cultures of human fibroblasts (O'Brien et ( I / . , 1979b) and BALB/c 3T3, C3H/lOTM, and AKR-2B mouse fibroblasts (Sivak, 1977; Moses ef a/., 1978; O'Brien et a/., 1979b). In responsive cell strains such as these, TPA also usually enhances thymidine incorporation stimulated by fresh serum-containing medium. It does not stimulate DNA synthesis in quiescent cultures of hamster embryo fibroblasts (O'Brien and Diamond, 1977) or of some transformed cell lines derived from the C3HIIOTM and AKR-2B cell lines (Moses et a/., 1978). Under certain conditions, such as in the presence of serum-free medium, TPA alone may not stimulate DNA synthesis in quiescent cultures, but may enhance stimulation when the cells are induced to cycle by exposure to fresh serum-containing medium (Boynton et a / . , 1976). The role of specific serum factors in this response has been investigated in several recent studies. For stimulation of quiescent cultures of Swiss or BALB/c 3T3 mouse cells by TPA, serum can be replaced by any of the purified growth factors, EGF, insulin, or fibroblast growth factor (Dicker and Rozengurt, 1978; Frantz et a / . , 1979). With BALB/c 3T3 cells, serum can also be replaced with either of the defined serum growth fractions: platelet-derived growth factor or platelet-poor plasma (Frantz et a / . , 1979). When combined with any of these factors, TPA acts synergistically to stimulate thymidine incorporation and cell division (Frantz rt a / . , 1979). Thus, it can replace each of the three classes of hormonal growth factors, that is, platelet-derived or fibroblast growth factor, the insulinlike somatomedins and EGF, as long as at least one member of another

44

LEILA DIAMOND

et a/.

class is present. It cannot, however, replace the requirement for at least one growth factor during a cell cycle, and in this respect differs from SV40. TPA stimulates DNA synthesis in lymphocyte populations from several animals species, as first shown by Mueller and Kajiwara (1965) and Whitfield et u / . (1973). Stimulation of thymidine incorporation into mouse spleen cells by TPA is enhanced in cells pretreated with concanavalin A (Suss and Schuster, 1974). TPA is mitogenic for a subpopulation of human T-lymphocytes different from those responsive to phytohemagglutin and having relatively high affinity for sheep erythrocytes (Estensen et a/., 1974; Touraine et a/., 1977). In bovine lymphocyte populations, TPA acts synergistically with concanavalin A or phytohemagglutinin to stimulate DNA synthesis, affecting cells that would not otherwise respond to the lectins (Mastro and Mueller, 1974). The concurrent addition of retinoic acid selectively blocks the comitogenic action of TPA without altering the basic response of the lymphocytes to phytohemagglutinin (Kensler and Mueller, 1978). Retinoic acid does not affect the TPA-induced increases in overall rates of RNA and protein synthesis, indicating that it probably does not compete with TPA for a phorbol ester receptor or binding site on the lymphocyte membrane, but selectively inhibits an early event in the TPA-mediated transition of the lymphocytes from Go + S. In contrast to its comitogenic effect on bovine lymphocytes stimulated with lectins, TPA inhibits the induction of DNA synthesis in cultures of lymphocytes undergoing the mixed-lymphocyte reaction (Mastro and Mueller, 1978). It does not affect cells already undergoing DNA replication, suggesting that it probably acts on some early step that is necessary for the triggering of cell replication. In some human cell lines derived from either normal lymphocytes or lymphoblastic leukemia cells, TPA induces cell clumping, adherence to the substrate, and transient inhibition of growth (Castagna et al., 1979b). 4. Effects on DNA Repuir Synthesis

Inhibition of DNA repair was proposed by Gaudin et a / . (1971, 1972) as the general mechanism of action of tumor promoters, after finding that the promoting compounds they tested inhibited DNA repair synthesis in UV-irradiated human lymphocytes. However, there are a number of reports that tumor promoters do not specifically inhibit DNA repair synthesis. TPA does not inhibit repair in human amnion cells after /ow doses of either UV irradiation or N-acetoxy-2-acetylaminofluorene

TUMOR PROMOTERS

45

(Trosko et a/., 1975). It does not inhibit the “post-replication repair” that occurs in growing V79-4 Chinese hamster cells in response to U V irradiation (Trosko et ul., 1975). Finally, in several cell types, tumor promoters inhibit normal DNA-replicative synthesis as much as, or more than, they inhibit repair replication (Poirier et nl., 1975; Cleaver and Painter, 1975; Langenbach and Kuszynski, 1975).

D. OTHERBIOLOGICAL EFFECTSOF PHORBOL DIESTERTUMOR PROMOTERS 1. Effects on l’ermincil Cell Dflerentiution

Much of the recent interest in tumor promoters was stimulated by the finding that these compounds inhibit terminal differentiation of normal and malignant cells in culture. This finding was in agreement with earlier reports that, in mouse skin, tumor promoters affect differentiation both morphologically and biochemically (see Section IV,A), and suggested possible mechanisms of tumor promotion based on the effects of promoters on terminal differentiation in \,itro, In addition, the finding that phorbol diester tumor promoters modulated cell differentiation introduced to the study of differentiation a new class of compounds with many chemical and biological properties distinct from those of compounds previously used. (1. Inhibition of Cell Dgferentitition. i . Rodent cells. The inhibitory effects of tumor promoters on terminal differentiation of Friend murine erythroleukemia cells (FELC) (Patuleia and Friend, 1967) in culture were first reported by Rovera et r i l . (1977) and Yamasaki et [ I / . (1977). In clones of FELC characterized by the presence of a high percentage (4070% benzidine-positive) of cells undergoing erythroid differentiation in the absence of inducing drugs (Rovera and Bonaiuto, 1976; Rovera and Surrey, 1977), there is a positive correlation between the known tumorpromoting activity of a particular phorbol diester in mouse skin and the ability of the compound to inhibit spontaneous differentiation of these cells (Rovera et nl., 1977). The inhibition is reversed by removal of the promotor from the culture medium, even after continuous treatment with the promoter and consequent inhibition of differentiation for a period as long as 6 months (Rovera et ctl., 1977: Diamond et ill., 1978b). An interesting theoretical and practical point is that the cells treated with TPA for a prolonged period retain their potential for a high frequency of spontaneous differentiation, unlike control cells, which gradually revert

46

LEILA D I A M O N D et a / .

to a population with a low percentage of differentiating cells (Rovera and Surrey, 1977). TPA is apparently able to stabilize the differentiation program in FELC clones such as these. In other clones of FELC characterized by a low percentage of spontaneously differentiating cells, tumor promoters inhibit induced erythroid differentiation, with the extent of inhibition depending on the type of inducer and the particular clone (Yamasaki et a/., 1977, Diamond et N/., 1978b: Fibach et d., 1978). Clones of FELC inducible by hexamethylene bisacetamide can be maintained for prolonged periods in the presence of the inducer and TPA; the cells continue to divide, but will differentiate when released from the TPA block (Yamasaki et ul., 1979a; Fibach et a/., 1979). Because the accumulation of globin mRNA and the synthesis of globin chains were reduced in the presence of TPA, Rovera et a / . (1977) and Yamasaki et t i / . (1977) concluded that tumor promoters inhibit spontaneous and induced differentiation of FELC by restricting expression of a product of differentiation, that is, hemoglobin. Fibach et a/., (1979) have shown that promoters also suppress expression of such other erythroid characteristics as synthesis of spectrin, heme synthetic enzymes, and heme. Subclones that vary in their response to the inhibitory effects of TPA on induced erythroid differentiation have been isolated from a subline of FELC (Fibach et mi., 1978). In some of these clones, TPA causes more than 90% inhibition of differentiation, whereas in others, differentiation M or by is not inhibited either by concentrations of TPA as high as other structurally related tumor promoters. However, TPA-resistant clones are not resistant to dexamethasone (Fibach et a / . , 1978), another type of inhibitor of induced differentiation of FELC (Lo et a/., 1978). In TPA-sensitive, but not in resistant clones, the promoters also enhance cell adherence to the surface of the culture vessel (Yamasaki et a/., 1979b). This adherence is probably not a prerequisite for TPA inhibition of differentiation, since inhibition also occurs when these clones are grown in bacterial flasks in which they do not adhere. Furthermore, in other clones of FELC, for example, clone 745 isolated by Friend et a / . (1971) and the spontaneously differentiating clones of Rovera and Bonaiuto (1976), tumor promoters do not induce adherence, although they do inhibit differentiation (Rovera et d . , 1977; Diamond et al., 1978b). Another rodent cell system in which phorbol diester tumor promoters inhibit terminal differentiation is mouse preadipose cells (Diamond et a/., 1977). The preadipose BALB/c 3T3 (clone A31T) cells used in these studies differentiate to mature adipocytes in a manner similar to that reported for clones of Swiss 3T3 cells (Green and Kehinde, 1974; Green,

TUMOR PROMOTERS

47

1978); when held as confluent monolayers, they gradually acquire many of the biochemical characteristics of adipocytes in vivo (Diamond et al., 1977; O’Brien et a/., 1979a).The inhibition by tumor promoters of either spontaneous or insulin-stimulated adipose conversion of BALB/c 3T3 cells is reversed when the compounds are removed from the medium (Diamond et al., 1977; O’Brien et al., 1979a). However, the cells sometimes spontaneously overcome the inhibitory effects of the promoters, even when treated frequently with fresh solutions of promoters. This is in contrast to FELC, which do not overcome the inhibitory effects of the promoters on either spontaneous or induced differentiation (Rovera et al., 1977; Yamasaki et ul., 1979a). Treatment with TPA markedly enhances lactate production from glucose by these preadipose cells after 4-7 days (O’Brien el al., 1979a). The concentration dependence for the TPA-mediated stimulation of lactate production is similar to that for its inhibitory effect on adipose conversion, suggesting that the ability of TPA to interfere with the normal pattern of glucose metabolism may be important in the inhibitory effect of TPA on triglyceride accumulation in these cells. The phorbol diesters that inhibit adipose conversion of BALB/c 3T3 cells also stimulate cell division (Diamond e f al., 1977),but it is difficult to determine if inhibition of differentiation in this system is dependent on continuous cell cycling. The antiviral glycoprotein, interferon, also inhibits adipose conversion of these cells (Cioe et al., 1979). In cultures treated with low doses of interferon and ineffective concentrations of TPA (1.6 x low9M 1, the two agents act synergistically to block differentiation. Several lines of evidence suggest that they differ in the mechanism by which they inhibit adipose conversion. In cultures of the mouse C1300 neuroblastoma cell line (Augusti-Tocco and Sato, 1969), tumor promoters reversibly inhibit, during a 3-hour period, the spontaneous neurite formation that occurs in response to serum deprivation (Ishii et al., 1978). TPA also blocks, over a 2-day time course, neurite outgrowth induced in these cells by prostaglandin El, papaverine, and bromodeoxyuridine (Ishii er al., 1978). In primary cultures of newborn mouse epidermis, TPA inhibits the appearance of fibrillar structures and the development of a multilayered cellular organization (Yuspa et ul., 1976a; Fusenig and Samsel, 1978). The block in this keratinization-like differentiation is reversed after 2 days. An interesting recent observation on the effects of tumor promoters on cell differentiation is that epithelial cell lines established from rat trachea treated with TPA in vitro show no evidence of cilia while in culture, but when transplanted to stripped trachea in v i w are able to

48

LEILA DIAMOND et a / .

repopulate them with highly differentiated, ciliated cells (Steele et a / ., 1978b). The authors suggest that TPA may cause prolonged cellular dedifferentiation in vitro, which can be reversed in vivo by association with normal mesenchymal tissue and serum components. ii. Chick embryo cells. The first clear demonstration that a tumor promoter can inhibit the differentiation program of cells in culture was made by Cohen et al. (1977) using chick embryo muscle. They showed that TPA is mitogenic for presumptive myoblasts and inhibits their conversion to definitive myoblasts synthesizing muscle-specific myosin heavy and light chains. TPA also blocks the fusion of postmitotic myoblasts to form myotubes, presumably by acting directly on the cell surface (Cohen et al., 1977). When added to cultures in which postmitotic multinucleated myotubes have already formed, TPA reversibly inhibits accumulation of myofibrils and enhances the density of the longitudinallyoriented 10-nm filaments (Toyama et al., 1979). This is a clear example of the fact that TPA can affect the differentiation process of cells not traversing the cell cycle (see also Miskin et a/., 1978a,b). The type of mitogenic effect induced in myogenic cells by TPA had been observed earlier in myogenic cells that had either incorporated bromodeoxyuridine into their DNA (Bischoff and Holtzer, 1970) or been transformed with RSV (Holtzer et a/., 1975). In the latter case, evidence was obtained using a temperature-sensitive RSV mutant that suppression of the cells' phenotypic program is dependent on the continuous function of the virus transformation gene. With respect to the effect of TPA on myogenesis, Cohen et al. (1977) pointed out that the behavior of TPAsuppressed myogenic cells is remarkably similar to that of temperaturesensitive RSV-transformed and bromodeoxyuridine-suppressed myogenic cells. They suggested that these dissimilar agents, that is, TPA, bromodeoxyuridine, and the transforming RSV gene, may exert their common effect on myogenesis by forcing presumptive myoblasts to undergo proliferative cell cycles, thereby preventing activation of the normal synthetic program. Chick chondroblasts are a cell system similar to chick myoblasts in that the effects of TPA on differentiation (Pacifici and Holtzer, 1977; Lowe et ul., 1978)resemble those of bromodeoxyuridine and temperaturesensitive RSV-transformation (Pacifici et ul., 1977). The earliest effect of TPA on chick chondroblasts prepared from vertebral cartilages is enhanced attachment to the substrate, although cell-cell adhesion is greatly reduced. The normally sessile, polygonal chondroblasts are transformed by TPA into motile, multilayered fibroblasts with long, pseudopodial processes. In the presence of TPA, the chondroblasts no longer synthesize two of their unique proteins, the type IV sulfated proteoglycan and

TUMOR PROMOTERS

49

a glycosylated protein of apparent MW 180,000. The reversibility of the effect of TPA upon its removal is gradually lost after 4 days and is irreversible after 12 days in TPA. TPA also induces morphological changes in CEF from subdermal connective tissue, but is less mitogenic for these cells than for chondroblasts and, in contrast to its effects on differentiation of chondroblasts, does not block synthesis of the type 111 sulfated proteoglycan characteristic of chick fibroblasts (Lowe et ul., 1978). The differential response to TPA of these two cell types with their close lineal relationship is of considerable theoretical interest from the point of view of cell differentiation and the molecular target of growth factors. It also demonstrates how the activity of an exogenous molecule may depend largely on the phenotypic activities of the responding cell (see also Rovera et al., 1979a,b). TPA does inhibit synthesis of another differentiated property of CEF, collagen (Delclos and Blumberg, 1979). An initial decrease in collagen synthesis is apparent after 24 hours, but 5 days are required for maximal reduction to 20% of controls. Collagen synthesis is also decreased in CEF transformed by RSV (Levinson et ul., 1975; Kamine and Rubin, 1977) and treatment of such cells with TPA further decreases synthesis to 50% of transformed controls (Delclos and Blumberg, 1979). In cultured embryonic chick sensory and sympathetic ganglia, tumor promoters transiently and noncompetitively inhibit nerve growth factorprovoked neurite outgrowth without causing retraction of already established neurites (Ishii, 1978). TPA appears to enhance cell viability and after 24 hours in its presence, ganglia slowly overcome the block in neurite outgrowth. This escape from the inhibitory effects of the promoters is also sometimes seen in preadipose 3T3 mouse cells treated with TPA (Diamond et NI., 1977), but occurs much sooner in the ganglia than in the preadipose cells. h. Stimulution of Cell Differentiation. Although the first reports of the effects of phorbol diester tumor promoters on terminal differentiation of cells in culture indicated that these compounds were specifically inhihitors of cell differentiation, a series of recent papers indicates that in some target cells they stimulate differentiation. Miao et al. (1978) found that TPA induces differentiation in two clones of Rauscher virus-transformed murine erythroid cells. Maximum induction (60% benzidine-positive cells) requires a concentration of M TPA, but lower concentrations also have some stimulatory effect. The commitment to differentiate apparently occurs within a few hours after exposure to TPA, in contrast to a period of 48 hours or longer that is required for commitment following exposure to another type of inducer, dimethyl sulfoxide. A line of human promyelocytic leukemia cells, HL-60, can also be

50

LEILA DIAMOND

et u/.

induced to differentiate into mature cells by TPA and related phorbol diesters (Huberman and Callaham, 1979; Rovera et a / . , 1979a,b). This cell line, which grows in suspension, was established in culture by Collins et a / . (1977). Most of the cells in the population are promyelocytes, approximately 20% are more mature myeloid cells; treatment with dimethyl sulfoxide, dimethyl formamide, or butyric acid induces differentiation along the myeloid series (Collins et a / . , 1978). The first apparent effects of the tumor promoters on HL-60 cells are the induction of adherence and the cessation of DNA synthesis and cell division: these effects are irreversible and independent of the continued presence of the promoters (Huberman and Callaham, 1979; Rovera et a/., 1979a,b). The differentiation of HL-60 cells after TPA treatment was described by Huberman and Callaham (1979) as an increase in the percentage of myelocytes, metamyelocytes, and other mature myeloid cells in the population and an increase in the percentage of cells phagocytizing yeast. Rovera et a / . (1979a,b), however, found that in many respects the TPAtreated cells resemble macrophages rather than granulocytes. They concluded from an extensive study of the morphological, biochemical, and functional characteristics of the adherent cells that TPA-induced differentiation of HL-60 cells is along the monocytic rather than the myeloid series. Among the features that distinguish TPA-treated HL-60 cells from untreated and dimethyl sulfoxide-treated cells are retention of the nucleolus, resistance to trypsinization, ability to phagocytize sensitized sheep erythrocytes, and high levels of a monocyte-specific enzyme, non1971) in all cells. These authors have specific acid esterase (Yam et d., concluded that TPA and other phorbol diester tumor promoters can not only induce differentiation but can induce it along a pathway apparently different from that normally followed. They have also pointed out that the induction of differentiation in HL-60 cells by tumor promoters differs from induction by other types of inducers in this and other cell systems by the rapidity with which the proliferative capacity is lost and by the fact that the entire cell population is affected, with no nonresponsive, proliferating cells remaining. Both Huberman and Callaham (1979) and Rovera et u / . (1979a)have proposed that compounds such as these, which can inhibit cell growth and induce differentiation so effectively, merit investigation as antitumor drugs with this particular mechanism of action. Another cell system in which tumor promoters stimulate cell differentiation is the human melanoma line, HO (Giovanella et al., 1976). Treatment with TPA results in inhibition of cell growth, stimulation of melanin synthesis, and formation of dendrite-like structures characteristic of normal melanocytes (Huberman et a / ., 1979). Glucosamine incorporation into the GM3membrane ganglioside is increased 8-10-fold and

TUMOR PROMOTERS

51

decreased 50% into the GMlganglioside, compared to control cultures. The TPA-induced changes in the glycolipid composition of the cell membranes may play a role in the formation of the dendrite-like structures. 2 . Effects on Blood Elements

The fact that TPA affects terminal differentiation of leukemic cells is perhaps not surprising in view of its striking effects on the biochemistry and ultrastructure of normal blood elements. It is a potent stimulator of platelet aggregation, causing swelling and labilization of platelet granules (Zucker et a / . , 1974; White et ul., 1974). It also markedly affects erythrocytes, converting them into stomatocytes and stomatospherocytes containing multiple invaginations on the concave membrane faces (White and Repine, 1978). Phorbol diester tumor promoters affect phagocytic cells in a manner similar to particulates such as bacteria, and TPA has been used as a tool for detecting functional abnormalities in the neutrophils and alveolar macrophages of individuals with specific diseases (Repine et al., 1974; Hoidal et al. , 1978). In normal human polymorphonuclear leukocytes, TPA causes concentration-dependent increases in oxygen utilization, hexose monophosphate shunt activity, nitroblue tetrazolium dye reduction, and superoxide release (Repine er al., 1974; DeChatelet er al., 1976; Johnston et al., 1976; Goldstein, 1978). It induces rapid development of surface membrane-derived vacuoles and disappearance of specific granules (White and Estensen, 1974). This appears to be a selective labilization, since azurophilic lysosomes do not participate in vacuole formation, and contrasts with the effects of TPA on HL-60 promyelocytic leukemia cells in which azurophilic granules disappear (Rovera et al., 1978b). In human alveolar macrophages, TPA induces biochemical changes similar to those seen in polymorphonuclear leukocytes, but the increase in vacuolization is not as striking (Hoidal et a / . , 1978). The promoter generates formation of superoxide anion in several human and rodent granulocytehacrophage cell lines, including HL-60, as well as in mature granulocytes (Greenberger et a / . , 1978). It induces granulocytehacrophage “colony-stimulating activity” in a murine monocyte line (PU51.8), acting on these cells in the same way as other macrophage activators that inhibit cell growth (Ralph et a / . , 1977). It would be interesting to know if TPA-treated HL-60 cells have macrophage “colony-stimulating activity.” The effects of TPA and other promoters on lymphocytes are discussed in Section V,C,3.

52

LEILA DIAMOND et al.

3. Effects on Mutation Frequencies and Sister Chromatid Exchanges TPA may affect the recovery of spontaneous and induced mutations in V79 Chinese hamster lung cells. The recovery of 6-thioguanine- and ouabain-resistant colonies is slightly increased (52-fold) by treatment of UV-irradiated cells with TPA and, to a lesser extent, by treatment with phorbol, neither of which is itself mutagenic for these cells (Trosko et ul., 1977). TPA is most effective at increasing the mutation frequency when present after the mutation expression time and the completion of most DNA repair. TPA and phorbol also increase slightly the frequency of ouabain-resistant mutations induced by the chemicals MNNG and methylazoxymethanol acetate (Lankas et a l . , 1977). Kinsella et al. (1978) and Kinsella and Radman (1978) observed that in V79 cells TPA, but not the nonpromoter 4-O-methyl-TPA, increases approximately 2-fold the frequency of sister chromatid exchanges (SCE) at concentrations of TPA that induce no increase in chromosome aberrations; however, over 100 SCE per metaphase are induced by TPA in about 25% of the cells. Induction of SCE by TPA is inhibited by the steroid fluocinolone acetonide and the protease inhibitors antipain and leupeptin. The authors considered the frequency of SCE to be a cytological indicator of cellular recombinational activity and suggested that the induction of SCE by TPA might result from the induction of cellular DNA recombination enzymes by the promoter. Nagasawa and Little (1979) also found a slight elevation in the frequency of SCE in C3H/lOT% cells treated with a concentration of TPA (0.1 pg/ml) that did not affect cell viability and that does not transform these cells (Mondal and Heidelberger, 1976; Kennedy et ul., 1978). TPA also enhances direct X-ray-induced SCE but has no effect on the frequency of SCE in irradiated cells maintained in stationary growth for several hours (recovery-induced SCE) (Nagasawa and Little, 1979). Protease inhibitors suppress both TPA-induced and recovery-induced SCE but not direct X-ray-induced SCE. In contrast to these observations, Loveday and Latt (1979) did not detect any increase due to TPA in either the immediate or delayed induction of SCE by 5-bromodeoxyuridine or mitomycin C in V79 or Chinese hamster ovary (CHO) cells. 4. Induction of Herpesvirus Synthesis

TPA and other diterpene derivatives induce synthesis of Epstein-Barr virus (EBV) antigens and particles in some lymphoblastoid cell lines with persisting EBV DNA (zur Hausen er al., 1978, 1979). Spontaneous induction of viral antigen and particle synthesis usually occurs in these cell

T U M O R PROMOTERS

53

lines at a low rate. Treatment with TPA for 10 days increases the recovery of viral DNA 20-40-fold. The viral DNA from TPA-treated cultures has the same restriction endonuclease cleavage pattern as viral DNA from uninduced cells. In some cell lines carrying the EBV genome, tumor promoters induce an abortive infection, with increased synthesis of early antigen but not viral capsid antigen. Yamamoto et ril. (1979) observed that EBV induction by tumor promoters and the pyrimidine analog iododeoxyuridine was inhibited by cisretinoic acid but that the retinoid did not affect viral antigen synthesis induced by superinfection. They suggest that inducing chemicals may activate persisting EBV genomes by a common retinoic acid-sensitive mediator, whereas a different pathway leads to synthesis of early antigens after spontaneous induction or superinfection.

5 . Effects on Virrts-induced Cell lransformrition TPA enhances EBV-induced transformation of human leukocytes, as detected by its ability to enhance colony formation in soft agar by infected cells (Yamamoto and zur Hausen, 1979). There is also evidence that under certain conditions TPA may enhance slightly transformation of Chinese hamster lung cells by simian virus 40 mutants defective in the synthesis of 20K t-antigen (Martin et ul., 1979a,b).

E. EFFECTSOF OTHERTUMOR PROMOTERS As with the phorbol diester tumor promoters, macrocyclic plant diterpene esters from the families Euphorbiaceae and Thymelaeaceae (see Section III,D) show some correlation between tumor-promoting activity in iivo and the ability to induce some specific effect in vitro. For example, the daphnane derivative, mezerein, induces plasminogen activator in CEF (Wigler et d . , 1978), stimulates DNA synthesis in phytohemagglutinin-treated bovine lymphocytes (Kensler and Mueller, 1978), inhibits differentiation of murine FELC and neuroblastoma cells (Fibach et d., 1978; Ishii et ul., 1978), and stimulates adherence and differentiation of HL-60 human leukemia cells (Rovera et a/., 1979a). However, comparative studies of the activity of tumor-promoting compounds in viva and in vitro are just beginning and there have already been some exceptions to the correlation between the two. For example, within a series of diterpene derivatives there is some correlation between promoting activity and the capacity to induce herpesvirus synthesis in lymphoblastoid cell lines (zur Hausen et ul., 1979). However, two nonpromoting diter-

54

LEILA D I A M O N D et a / .

pene derivatives, one of the tigliane type and one of the daphnane type, also induce viral antigen synthesis, whereas the promoter anthralin, which is a phenolic compound and not a diterpene, does not. I t is not unusual to find that a single diterpene ester such as TPA or mezerein has many different effects on cells; promoters that induce a specific effect in one cell system in vitro usually induce other effects in the other systems in which they have been tested. This generalization does not, however, extend to other promoters that are unrelated in structure to the diterpene esters and are generally weaker promoters (see Sections III,E and F). Many of these compounds are so toxic for cells in culture that it is difficult to determine whether they induce specific biological or biochemical effects. A few promoters were tested for the ability to enhance outgrowth of SV40-transformed clones in mixed culture (Sivak and Van Duuren, 1970); some tobacco leaf extracts were active, whereas Tween 80 and phenol were not. In another study (Driedger and Blumberg, 1978), oleic acid caused a 20% decrease in LETS protein in CEF, and anthralin, lauric acid, and limonene caused an increase in 2-deoxyglucose transport. However, none of these compounds affected LETS protein levels or 2-deoxyglucose transport to the extent that phorbol diesters did (Blumberg et a/., 1976; Driedger and Blumberg, 1979). Anthralin, cantharidin, Tween 80, and a tobacco leaf extract did not induce plasminogen activator in CEF or HeLa cells (Wigler et al., 1978). Again in contrast to the phorbol diester promoters, anthralin did not induce adherence and differentiation of HL-60 cells (Rovera et d., 1979a) nor induce herpesvirus synthesis in lymphoblastoid cell lines with persisting virus genomes (zur Hausen et a / . , 1979). On the other hand, high concentrations of saccharin, a compound that may have promoting or carcinogenic activity in rat bladder (Hicks ef ul., 1973, 1975, 1978; Hicks and Chowaniec, 1977; Reuber, 1978), enhances transformation of C3HIlOTM cells initiated by 3-methylcholanthrene but not UV irradiation (Mondal et a/., 1978).

VI. Speculation on the Biochemical Mechanisms of Tumor Promotion

Speculation on the biochemical mechanism of promotion, particularly the promoting action of the phorbol diesters, must involve consideration of both the mechanism by which these compounds induce their myriad effects on cells and identification of those effects that are specifically related to promotion. (In this discussion, instead of citing all the refer-

TUMOR PROMOTERS

55

ences in which the work is described, the reader is referred to the appropriate section in the chapter). Irritation was proposed as a mechanism of tumor induction as early as 1761 when John Hill reported six cases of “polypusses” (nasal carcinomas) in snuff users (see Redmond, 1970). Early theories of the mechanism of promotion were based on the fact that promoters induced inflammation and hyperplasia, but there is now substantial evidence that promotion involves more than simply this. Many hyperplasiogenic irritants have little or no promoting activity (see Boutwell, 1974; Raick, 1974; Scribner and Suss, 1978) and, within the diterpene series of tumor promoters, inflammatory activity is not directly correlated with promoting activity (see Van Duuren, 1976; Hecker, 1978). However, all potent tumor-promoting phorbol diesters are irritants (see Van Duuren, 1976; Hecker, 1978) and the induction of inflammation and hyperplasia is probably an essential part of the promotion process; antiinflammatory steroids that inhibit inflammation also inhibit promotion (Ghadially and Green, 1954; Belman and Troll, 1972; Schwarz et ul., 1977). Boutwell’s experiments (Boutwell, 1964) indicate that the induction of inflammation and hyperplasia by such nonpromoting irritants as turpentine and cantharidin may be related to the propagation stage of promotion, whereas only promoters can accomplish the conversion stage that precedes propagation. Any theory of the mechanism of promotion must take into account this specific activity of promoters. It must also be consistent with two requirements for two-stage carcinogenesis: first, an irreversible initiation step resulting from a single application of a low dose of carcinogen and, second, the need for repeated treatments with the promoter, with reversion of the promotion process if treatment stops. Most of the recent hypotheses about the mechanism of tumor promotion are based on the biochemical and biological changes induced in cells, both in vivo and in vifro, by the phorbol diester series of promoters and structurally related plant esters. Too little is known about the effects of other classes of promoters to speculate on whether all promoters act through the same basic mechanism by, for example, activating a common set of critical changes. It may also be that, with their different architectures and responses to exogenous chemicals, each tissue responds differently to promoting agents. In addition, although recent studies on the mechanism of tumor promotion in mouse skin have considered the epidermis as the target tissue of both promoter and initiator, promoters might perhaps alter the interactions between different cell types in the skin and affect the epidermis indirectly. Critical experiments to help resolve this question might be done in epidermal cell culture models of two-stage carcinogenesis.

56

LEILA DIAMOND

et al.

A. MEMBRANE INTERACTIONS

An important unanswered question about tumor promoters is that of the identity of the cellular “target(s)” with which they interact to trigger the observed changes in cellular behavior (see Sections IV,A,2,d and V,C,l). It has been suggested (Rohrschneider and Boutwell, 1973; Wilson and Huffman, 1976; Weinstein et al., 1977, 1978b) that promoters may act by usurping the action of some endogenous hormone or growth factor, perhaps by binding to its cellular receptor. Promoters have been shown to inhibit the binding of some growth-promoting substances to their receptors, but more studies are needed to determine whether this is a general membrane effect involving most receptor species or is confined to a unique population(s) of receptors. The synergistic effect of TPA on the stimulation of DNA synthesis and cell division in 3T3 mouse cells treated with various growth-promoting substances (Dicker and Rozengurt, 1978; Frantz et al., 1979) suggests that under certain conditions promoters may enhance the binding of some hormones to their receptors, perhaps by uncovering “masked” receptors, and, thereby, increase the biological response to that hormone. Because phorbol diester tumor promoters are readily sequestered in cellular membranes, they might easily pass into the cell and be available for interaction with cytosolic or nuclear components (i.e., receptors). There is as yet no evidence that specific interaction with receptors such as those with which either peptide or steroid hormones interact is required for induction of biological effects by phorbol diesters. The promoters may instead produce a generalized membrane perturbation, as, for example, an increased fluidity of critical components, that could lead directly to subsequent changes within the cell or alter the accessibility of the cell surface to other growth-controlling substances.

B. ALTEREDGENEEXPRESSION One hypothesis for promoter action that has engendered considerable discussion is based on the alteration of gene expression by promoters. According to this hypothesis, promoters cause increased expression of a gene@)that has been altered by initiation; this results in the conversion of the initiated cell into a cell capable, after further promoter treatment, of multiplying and progressing to a tumor. If the theory, which has been reviewed in detail by Boutwell (1974), is correct, the crucial question becomes: The expression of which altered gene(s) leads to tumor formation?

TUMOR PROMOTERS

57

O’Brien and Boutwell (O’Brien et NI., 1975b; O‘Brien, 1976; Boutwell, 1977) proposed a theory of promoter action that accords with both the hypothesis of somatic mutation for initiation and the hypothesis of specific gene activation for promotion. They suggested that since promoterinduced epidermal ornithine decarboxylase (ODC) activity has such a short half-life (17 minutes) and there is a transient “spike” of activity after each promoter treatment (see Section IV,A,2,b), there must exist a specific system for degrading or inactivating the enzyme. Since the normal levels of ODC are very low, such degradative machinery would be needed only after the large induction of the enzyme that follows promoter treatment. If the initiator-induced ”lesion” were a defect in the gene(s) responsible for this degradation system, then promoter treatment of the initiated cell would result in a permanently elevated ODC level, as well as high intracellular polyamine concentrations. This would confer on initiated cells a selective advantage over the surrounding normal cells that have only transient elevation of ODC and polyamine levels in response to promoter treatment. An important aspect of this theory is that the critical alterations in cancer cells would be in the function of regulatory, and not structural, genes. The key aspect of promoter action would be the activation of a gene(s) that can take advantage of the “lesion” in the regulatory gene that the initiator had previously introduced. Thus, the theory is compatible not only with a mechanism based on the genes controlling ODC activity but also with a mechanism involving alterations in other important metabolic functions. The determination of the specific genes involved, however, is hindered by the difficulties in identifying initiated cells so that they can be studied. Several investigators (Troll ef NI., 1975, 1978; Weinstein et al., 1978a; Kinsella et NI., 1978: Kinsella and Radman, 1978) have suggested that cellular proteases may play a role in promotion by affecting gene expression through proteolytic cleavage of protein repressors. Troll (1975, 1978) has proposed a molecular model of two-stage carcinogenesis with a role for proteases that is based on analogy with the SOS DNA repair system in E . coli (Witkin, 1976). According to the theory, subcarcinogenic doses of carcinogen would damage the DNA slightly, but the lesions would remain “invisible“ until a promoter, by stimulating cell division and inducing proteases, caused the lesion to be noticed during DNA synthesis. This would lead to derepression of an SOS error-prone DNA repair system. Promoters, by inducing proteases, would derepress genes having repressors that are sensitive to hydrolysis by the protease(s). This theory is compatible with reports that TPA induces proteases (see Sections IV,A,2 and V,C,2) and that promotion in vivo and in v i m can be blocked

58

LEILA DIAMOND

et a / .

by protease inhibitors (see Sections IV,A,3 and V,A). However, there is as yet no clear evidence that animal cells have an inducible, error-prone DNA repair system or that, in mouse skin, promoters induce proteases in epidermal cells. Troll et a / . (1978) have discussed the indirect evidence that these events could occur. Kinsella et a / . (1978) and Kinsella and Radman (1978) have proposed yet another hypothesis in which tumor promoters act by inducing expression of the activity of specific genes, those involved in genetic recombination. They propose that initiated cells contain specific recessive, autosomal, somatic mutations and that promoters cause the expression of these mutations by inducing cellular DNA recombination enzymes and, as a consequence, an aberrant mitotic segregation event. Such an event would convert the initiated heterozygous cell into a homozygous or hemizygous cell. If one accepts the assumption of these authors that sister chromatid exchanges are cytological indicators of cellular recombination, then their hypothesis is consistent with the reports that TPA induces sister chromatid exchanges and that protease inhibitors suppress this induction and, in the same cells, suppress TPA-enhanced transformation (see Sections V,A and V,D,3). Kinsella et a / . (1978) also found that TPA induces segregation of two recessive traits, 6-thioguanine-resistanceand growth "on" agar, from doubly heterozygous cell hybrids. As they have pointed out, however, their hypothesis that promoters act by inducing mitotic segregation can apply only to an irreversible step(s) in promotion. As described in Sections V,B and V,C, many of the morphological and biochemical effects of tumor promoters on cells in culture resemble those induced by transformation, particularly transformation induced by infection with Rous sarcoma virus (RSV) (reviewed also in Weinstein et a/., 1977: Diamond et a / . , 1978a). An important difference, however, is that the addition of tumor promoters to soft agar medium usually does not confer the ability for anchorage-independent growth on cells that have not already acquired that property. The mimicry of transformation by tumor promoters has led Weinstein ef a / . (1978a) to propose that the role of the initiator in two-stage carcinogenesis may be to alter the regulation of the genes involved in the tumor cell phenotype so that once the regulatory system is turned on by a promoter, it tends to remain on autonomously. This is a concept of promoter-induced gene activation not unlike that proposed by O'Brien and Boutwell. Recently, Soprano and Baserga (1979)found that treatment with TPA reactivates silent ribosomal R N A genes in human-mouse hybrid cells that normally express the ribosomal R N A genes of only one species. There is also a synergistic interaction between promoters and RSV that results in greater expression of the cellular characteristics associated

TUMOR PROMOTERS

59

with the transformed phenotype than that induced by either agent alone (see Sections V,B and V,C,2). In chemically transformed hamster embryo cells, tumor promoters have a synergistic effect on serum induction of ODC rather than the additive effect that occurs in normal cells (O’Brien and Diamond, 1977). The fact that normal and transformed cells respond differently to promoter treatment suggests that there may also be differences between the responses of normal and initiated cells, and that this enhanced expression of specific genes in promoter-treated initiated cells may give them the required selective advantage.

C. ALTEREDCELLDIFFERENTIATION

A number of investigators have suggested that tumor promotion in mouse skin involves some alteration in normal cell differentiation such as dedifferentiation or redifferentiation (e.g., Berenblum, 1954; Raick, 1974; Marks, 1976). In fact, as early as 1954 Berenblum discussed a theory of promotion based on “inhibited maturation.” Marks (Marks, 1976: Marks et al., 1978) has suggested that the effect of promoters may be to prevent the maturation of epidermal stem cells, including the “initiated” cells, by promoting their dedifferentiation to a “pluripotent” population no longer committed to give rise specifically to interfollicular epidermis. According to this concept, the differentiation program “cancer” would only occur in the course of subsequent redifferentiation and a general reprogramming of the genetic read-out. Marks also proposes that it is this dedifferentiated population that is insusceptible to chalone control in response to TPA (see Section IV,A,2,b) and undergoes “hyperplastic transformation.” Such hyperplasia occurring in an initiated population redifferentiated to undergo the program “cancer” would lead to the promotion of tumor growth. Inactivation of the chalone mechanism, which is also produced by such nonpromoting treatments as acetic acid and mechanical manipulation (Krieg et al., 1974), would be necessary but not sufficient for promotion. That tumor promoters can inhibit or stimulate terminal differentiation, of cells in culture has been shown in many systems (see Section V,D,I), but there is no evidence that these effects include an alteration in commitment to a particular program of differentiation or an induction of dedifferentiation and reprogramming. Lowe et al. (1978) have pointed out that the effect of TPA on the differentiation of several different types of embryonic chick cells is similar to the effect of bromodeoxyuridine and RSV in that it can block expression of the terminal phenotype, with the cells and their progeny remaining committed to their respective li-

60

LEILA DIAMOND

e f a/.

neages. This inhibitory effect of promoters in the absence of dedifferentiation to a primitive stem cell is contrary to the redifferentiation theory of promotion envisioned by Marks. Rovera et af. (1979a,b) report that promoters can induce a leukemic cell line to differentiate along a pathway different from that normally followed; however, the two types of differentiated cells are assumed to be progeny of the same bipotent (not pluripotent) stem cells. Rovera et a / . (1977)and Diamond et a/.(1978a) suggested a mechanism for promotion based on the observation that tumor promoters inhibit terminal cell differentiation in vitro. They proposed that in a self-renewing tissue such as skin, transient blocking of differentiation that favors proliferation of the stem cell population could lead to amplification of cell populations in which the potential for malignancy has been initiated by a carcinogen but is not yet expressed. By being able to transiently shift the cell population from a commitment to differentiate (and become end cells) to a commitment to proliferate (and increase the stem cell population), tumor promoters could give initiated cells in the population additional time to acquire or express the malignant phenotype. A similar hypothesis of promotion, also based on the inhibition of differentiation in vitro, has been discussed by Weinstein et al. (1979). These hypotheses make two assumptions about promotion in mouse skin: first, that epidermal differentiation is inhibited by promoters and, second, that promoters act directly and not indirectly on initiated cells. These assumptions are important in considering these hypotheses in view of the recent findings that tumor promoters can stimulate terminal differentiation of some cells in culture (see Section V,D,l,b). One possibility is that the particular effects of promoters on differentiation that are relevant to promotion are not effects on the specific target cell but are instead, or in addition, effects on other types of cells in the tissue involved. In any case, there is ample evidence from both in vivo and in vifro studies that some type of altered cell differentiation in response to promoter treatment may be a mechanism by which tumor promoters help to convert initiated cells into proliferating tumor cells.

D. RELATION BETWEEN TUMOR PROMOTION A N D VIRAL CARCINOGENESIS The observation that tumor promoters produce reversible phenotypic changes resembling virus-induced transformation has suggested possible similarities between the mechanisms of viral and chemical carcinogenesis, as well as possible interactions between carcinogens, promoters, and

TUMOR PROMOTERS

61

transforming viruses in inducing tumors. Such speculations are particularly appropriate in view of recent data indicating that, in many species, genetically transmitted viral and transforming genes are part of the normal genetic make-up and can be activated by such agents as chemical carcinogens and hormones (see Aaronson and Stephenson, 1976). Also relevant perhaps is the recent observation that tumor promoters induce herpesvirus synthesis in lymphoblastoid cell lines with integrated viral genomes (see Section V,D,4). Todaro et ul. ( 1978) proposed a general model of transformation based to some extent on the observation that sarcoma growth factor induces phenotypic transformation in a clone of rat fibroblasts and that this transformation can be blocked by retinoids, as can many of the effects induced by TPA. Sarcoma growth factor is structurally similar to epidermal growth factor and is produced by murine sarcoma virus-transformed cells that lack available receptors for epidermal growth factor (DeLarco and Todaro, 1978). According to the model of Todaro et a / . (1978), tumor viruses and chemical carcinogens act by inducing cells to produce normally repressed or inactive growth-promoting factors. These factors may be exogenous (as TPA) or endogenous (as sarcoma growth factor) for given cells and would act as proximal effectors of transformation. In twostep carcinogenesis, initiation would involve gene damage that facilitates the production of normally repressed growth factors and/or their cellular receptors. Initiated cells might be producing some, but not enough, of their growth factors to cause phenotypic transformation, but additional exogenous promoters would allow the critical concentration of growth factors for transfOrmation to be reached. Bissell et cil. (1979) have speculated on the possibility that the product of the RSV transforming gene src itself may be a promoter rather than the initiator of transformation. They propose that viral carcinogenesis may occur in two stages: an “initiation” step caused by expression of a part of the viral genome other than src and a promotion step caused by the activation of the src gene. The scr gene product could be enhanced or replaced by such promoting agents as TPA. Among other things, the model predicts that infection with transformation-defective viruses that have a deletion within the src gene should also produce a transformed phenocopy upon the addition of a promoter: this has not yet been demonstrated. Pastan and Willingham (1978), in considering the mechanism of action of the src gene product, state “One is struck by the variety of effects that develop upon expression of a single gene and this has led us to wonder how a single gene product can have its diverse effects on cell growth, cell shape, and the synthesis of specific proteins.” One can

62

LEILA DIAMOND et

al.

substitute for “expression of a single gene (product)” in this statement “treatment with a single chemical” and wonder how treatment with that chemical, TPA, can also cause permanent expression of the transformed phenotype in an initiated cell. It may be that the mechanisms by which various agents such as RSV, growth-promoting hormones, and tumor promoters induce their diverse, multiple effects are similar, but that the effect(s) responsible for promoting action and perhaps the mechanism by which that effect is induced are unique to promoters. Until more is known about the biological consequences, in both normal and initiated cells, of such promoter-induced effects as altered regulation of polyamine biosynthesis or mimicry of the transformed phenotype, it will be difficult to identify the biochemical events that are unique to, and required for, promotion. VII. Perspectives

As this review indicates, impressive progress has been made in the field of tumor promotion, particularly in the last few years. Ten years ago research in this field was largely restricted to one system, mouse skin, and to one poorly defined tumor promoter, croton oil. Now the models of two-stage carcinogenesis being used in the study of the mechanism of tumor promotion are numerous and diverse. Much of the progress has resulted from the isolation and identification of the phorbol diesters as the active principles of croton oil and from the observation that these compounds have a multiplicity of intriguing effects on cells in culture. Impetus to research on tumor promotion has also come from the discovery that other types of chemicals can promote tumor formation in tissues other than skin. New systems amenable to initiation and promotion protocols, and new classes of promoters, will undoubtedly be discovered soon. The demonstration of two distinct stages in experimentally induced cancer in animals suggests that there are two stages in environmentally induced cancer in humans. Most human cancer probably results from a complex interaction of carcinogens, cocarcinogens, and tumor promoters. Prevention of human cancer requires the identification and, wherever possible, the elimination of these agents. Because promotion is a reversible process that requires repeated exposure over a period of time, in contrast to the rapid, irreversible process of initiation by carcinogens, manipulation of promotion would seem to be the best method of cancer prevention. Fundamental to designing methods of controlling cancer, therefore, is identification of promoters and an understanding of their

T U M O R PROMOTERS

63

biochemical mechanism(s) of action. As more is learned about these mechanisms, a battery of short-term tests can be developed t o screen for compounds with promoting activity. The identification of promoters in the environment, and the elucidation of their mechanisms of action, could lead to prophylactic measures such as the development of drugs specifically designed to interfere with the promotion mechanism and prevent tumor formation. If research in the field of tumor promotion continues to make the progress it has in the past decade, there is little doubt that answers will soon be found to the many unresolved questions concerning the biochemical mechanism(s) of tumor promotion, the identification of tumorpromoting compounds in the environment, and the role of tumor promotion in the etiology of cancer in humans. ACKNOWLEDGMENTS The authors' research is supported by grants CA 08936, CA 19948, CA 21778, and CA 10815 from the National Cancer Institute, DHEW, and grant E S 01664 from the National Institute for Environmental Health Sciences, DHEW. They thank Dr. Andrew Sivak for his critical reading of the manuscript.

REFERENCES Aaronson, S. A., and Stephenson, J . R. (1976). Biochim. Biophys. Acta 458, 323-354. Arffmann, E., and Glavind, J. (1971). Experientia 27, 1465-1466. Armuth, V . , and Berenblum, I . (1972). Cancer Res. 32, 2259-2262. Armuth, V . , and Berenblum. I. (1974). Cancer Res. 34, 2704-2707. Armuth, V . , and Berenblum, I. (1977). Inr. J . Cancer 20, 292-295. Augusti-Tocco, G., and Sato, G . (1969). Proc. Natl. Acad. Sci. U . S . A . 64, 31 1-315. Bach, H . , and Goerttler, K. (1971). Virchows Arch. (Zellparhol.)8, 1%-205. Baden, H . P., Sviokla, S., Mittler, B., and Pathak, M. A. (1968). Cancer Res. 28, 14631468. Baird, W. M. (1971). "Studies on the Mechanism of Action of Phorbol Esters in Skin Tumor Formation." Ph.D. dissertation, University of Wisconsin, Madison, Wisconsin. Baird, W. M., and Boutwell, R . K . (1971). Cancer Res. 31, 1074-1079. Baird, W. M . , Sedgwick, J . A., and Boutwell, R. K. (1971). Cancer Res. 31, 1434-1439. Baird, W. M., Melera, P . W., and Boutwell, R. K. (1972). Cuncer Res. 32, 781-788. Balmain, A. (1976). J. Invest. Dermatol. 61, 246-253. Balmain, A , , and Hecker, E. (1974). Biochim. Biophys. Actu 362,457-468. Bannasch, P. (1976). Cancer Res. 36, 2555-2562. Belman, S., and Troll, W. (1972). Cuncer Res. 32,450-454. Belman, S., and Troll, W. (1974). Cuncer Res. 34, 3446-3455. Belman, S., Troll, W., and Garte, S. J . (1978). Cuncer Res. 38, 2978-2982. Berenblum, I . (1941). Cmncer Res. 1, 807-814.

64

LEILA DIAMOND

et al.

Berenblum, I. (1954). A d v . Cancer Res. 2 , 129-175. Berenblum, I. (1969). Prog. Exp. Tumor Res. 11, 21-30. Berenblum, I., and Haran, N. (1955). Br. J. Cancer 9, 268-271. Berenblum, I., and Lonai, V. (1970). Cancer Res. 30, 2744-2748. Berenblum, I . , and Shubik, P. (1947). Br. J. Cancer 1, 379-382. Berenblum, I . , and Shubik, P. (1949). Br. J. Cancer 3, 384-386. Berry, D. L., Lieber, M. R., Fischer, S . M., and Slaga, T. J . (1977). Cancer Lett. 3, 125132. Berry, D. L., Bracken, W. M. Fischer, S. M., Viaje, A., and Slaga, T. J. (1978). Cancer Res. 38, 2301-2306. Bischoff, R., and Holtzer, H. (1970). J. Cell Biol. 44, 134-150. Bissell, M. J., Hatie, C., and Calvin, M. (1979). Proc. Nail. Acad. Sci. U . S . A . 76, 348352. Blumberg, P. M., Driedger, P. E., and Rossow, P. W. (1976). Nature (London) 264, 446447. Bock, F. G., and Burns, R. (1963). J. Nail. Cancer Inst. 30, 393-398. Bock, F. G., Fjelde, A,, Fox, H. W., and Klein, E. (1969). Cancer Res. 29, 179-182. Bock, F. G., Swain, A. P.. and Stedman, R. L. (1971). J. Natl. Cancer Inst. 47, 429-436. Bohm, R. (1915). Arch. Exp. Pathol. Pharrnacol. 79, 138-153. Bohm, R., and Flaschentrager, B. (1930). Arch. Exp. Puthol. Pharmucol. 157, 115-1 16. Bohm, R., Flaschentrager, B., and Lendle, L. (1935). Arch. Exp. Pnthol. Pharmacol. 177, 212-220. Bollag, W. (1972). Eur. J . Cancer 8, 689-693. Bollag, W. (1974). Eur. J . Cancer 10, 731-737. Borenfreund, E., Higgins, P. J., and Bendich, A. (1977). Cancer Lett. 3, 145-150. Boutwell, R. K. (1964). Prog. Exp. Tumor Res. 4, 207-250. Boutwell, R. K. (1974). CRC Crit. Rev. Toxicol. 2 , 419-443. Boutwell, R. K . (1977). I n "Origins of Human Cancer" (H. H. Hiatt, J . D. Watson and J . A . Winsten, eds.), pp. 773-783. Cold Spring Harbor Laboratory, New York. Boutwell, R. K., and Bosch, D. K. (1959). Cancer Res. 19, 413-424. Boynton, A. L., Whitfield. J. F., and Isaacs, R. J. (1976). J. Cell. Physiol. 87, 25-32. Brown, K . D., Dicker, P., and Rozengurt, E. (1979). Biochem. Biophys. Res. Commun. 86, 1037-1043. Brune, K., Kalin, H., Schmidt, R., and Hecker, E. (1978). Cancer Lett. 4, 333-342. Bullough, W. S . , and Laurence, E. B. (1964). Exp. Cell Res. 33, 176-194. Burns, F. J . , Vanderlaan, M., Snyder, E., and Albert, R. E. (1978). In "Carcinogenesis, Vol. 11, Mechanisms of Tumor Promotion and Cocarcinogenesis" (T. J. Slaga, A. Sivak and R. K. Boutwell, eds.), pp. 91-96. Raven, New York. Cameron R., Sweeney, G. D., Jones, K., Lee, G., and Farber, E. (1976). Cancer Res. 36, 3888-3893. Castagna, M., Rochette-Egly, C., Rosenfeld, C., and Mishal, Z. (1979a). FEES Lett. 100, 62-66. Castagna, M., Rochette-Egly, C., and Rosenfeld, C. (1979b). Cancer L e t f . 6, 227-234. Cioe, L . , O'Brien, T. G., and Diamond, L . (1979). J. Cell B i d . Int. Rep., in press. Clarke, E., and Hecker, E . (1965a). 2.Krebsforsch. 67, 192-204. Clarke, E., and Hecker, E. (1965b). Naturwissenschaften 15, 446-447. Clark-Lewis, I . , and Murray, A. W. (1978). Cancer Res. 38, 494-497. Clayson, D. B. (1975). Cancer Res. 35, 3292-3300. Cleaver, J. E., and Painter, R. B. (1975). Cancer Res. 35, 1773-1778.

TUMOR PROMOTERS

65

Cohen, R., Pacifici, M., Rubinstein, N., Biehl, J., and Holtzer, H. (1977). Nuture (London) 266, 538-540. Cohen, S. M., Arai, M., Jacobs, J. B., and Friedell, G. H. (1979). Cuncer Res. 39, 12071217. Colburn, N. H., and Boutwell, R. K. (1966). Cuncer Res. 26, 1701-1706. Colburn, N. H., Lau, S., and Head, R. (1975). Crincer Res. 35, 3154-3159. Colburn, N. H., Vorder Bruegge, W. F., Bates, J. R., Gray, R. H., Rossen, J. D., Kelsey, W. H., and Shimada, T. (1978a). Cancer Res. 38, 624-634. Colburn, N. H., Vorder Bruegge, W. F., Bates, J., and Yuspa, S. H . (1978b). In "Carcinogenesis, Vol. 11. Mechanisms of Tumor Promotion and Cocarcinogenesis" (T. J . Slaga, A. Sivak, and R . K. Boutwell, eds.), pp. 257-271. Raven, New York. Colburn, N. H., Former, B. F., Nelson, K. A., and Yuspa, S. H. (1979). Nature (London) 281, 589-591. Collins, S. J., Gallo, R. C., and Gallagher, R. E. (1977). Nature (London) 270, 347-349. Collins, S. J., Ruscetti, F. W., Gallagher, R. E., and Gallo, R. C. (1978). Proc. N u l l . Acud. Sci. U . S . A . 75, 2458-2462. Dammert, K. (1961). Actu Pothol. Microbiol. Scond. 53, 33-49. DeChatelet, L. R., Shirley, P. S., and Johnston, Jr., R. B. (1976). Blood 47, 545-554. Deelman, H. T. (1927). Br. Med. J . 1, 872. DeLarco, J . E., and Todaro, G. J. (1978). Proc. N u t / . Acud. Sci. U . S . A . 75, 4001-4005. Delclos, K. B., and Blumberg, P. M. (1979). Cancer Res. 39, 1667-1672. Diamond, L., O'Brien, S., Donaldson, C., and Shimizu, Y. (1974). Int. J. Cuncer 13,721730. Diamond, L. O'Brien, T. G., and Rovera, G. (1977). Nutitre (London) 269, 247-249. Diamond, L., O'Brien, T . G., and Rovera, G. (1978a). Life Sci. 23, 1979-1988. Diamond, L., O'Brien, T. G., and Rovera, G. (1978b). In "Carcinogenesis, Vol. 11, Mechanisms of Tumor Promotion and Cocarcinogenesis" (T. J. Slaga, A. Sivak and R. K. Boutwell, eds.), pp. 335-341, Raven, New York. Dicker, P., and Rozengurt, E. (1978). Nutitre (London) 276, 723-726. Doll, R. (1978). Cuncer Res. 38, 3573-3583. Driedger, P. E., and Blumberg, P. M. (1977). Cuncer Res. 37, 3257-3265. Driedger, P. E., and Blumberg, P. M. (1978). Int. J. Cancer 22, 63-69. Driedger, P. E., and Blumberg, P. M. (1979). Cancer Res. 39, 714-719. Elgjo, K., and Degre, M. (1973). J . N u l l . Cuncer Inst. 51, 171-177. Estensen, R. D., Hadden, J. W., Hadden, E . M., Touraine, F., Touraine, J. L., Haddox, M. K., and Goldberg, N. D. (1974). I n "Control of Proliferation in Animal Cells" (B. Clarkson and R. Baserga, eds.), pp. 627-634. Cold Spring Harbor Laboratory, New York. Farber, E. (1976). In "Liver Cell Cancer" (H. M. Cameron, D. A. Linsell and G. P. Warwick, eds.), pp. 243-277. Elsevier, Amsterdam and New York. Farber, E., and Solt, D. (1978). In "Carcinogenesis, Vol. 11, Mechanisms of Tumor Promotion and Cocarcinogenesis" (T. J. Slaga, A. Sivak and R. K . Boutwell. eds.), pp. 443-448. Raven, New York. Fibach, E., Yamasaki, H., Weinstein, I. B., Marks, P. A., and Rifkind, R. A. (1978). Cnncer R e s . 38, 3685-3688. Fibach, E., Gambari, R., Shaw, P. A,, Maniatis, G., Reuben, R. C., Sassa, S., Rifkind, R. A,, and Marks, P. A. (1979). Proc. N u t / . Accid. Sci. U . S . A . 76, 1906-1910. Fisher, P. 0 . . Weinstein, I. B., Eisenberg, D., and Ginsberg, H . S. (1978). Proc. Nurl. Acnd. Sci. U . S . A . 75, 2311-2314.

66

LEILA DIAMOND

et al.

Fisher, P. B., Flamm, M., Schachter, D., and Weinstein, I. B. (1979). Biochem. Biophys. Res. Commun. 86, 1063-1068. Foulds, L. (1969). "Neoplastic Development," Vol. I. Academic Press, New York. Frantz, C. N., Stiles, C. D., and Scher, C. D. (1979). J . Cell. Physiol., 100, 413-424. Frei, J. V . , and Stephens, P. (1968). Br. J. Cancer 22, 83-92. Friedrich-Freksa, H., Gossner, W., and Borner, P. (1969). 2. Krebsforsch. 72, 226-239. Friend, C., Scher, W., and Holland, J. G., and Sato, T. (1971). Proc. Nail. Acad. Sci. U.S.A. 68, 378-382. Fiirstenberger, G., and Marks, F. (1978). Biochem. Biophys. Res. Commun. 84, 1103- 1 1 1 1 . Fiirstenberger, G., and Marks, F. (1979). Cancer Leii. 6, 73-77. Fusenig, N. E., and Samsel, W. (1978). In "Carcinogenesis, Vol. 11, Mechanisms of Tumor Promotion and Cocarcinogenesis" (T. J . Slaga, A. Sivak, and R. K. Boutwell, eds.), pp. 203-220. Raven, New York. Fusenig, N. E., and Worst, P. K. M. (1974). J . Invest. Dermatol. 62, 187-193. Fusenig, N. E., and Worst, P. K. M. (1975). Exp. Cell Res. 93, 443-457. Gaudin, D., Gregg, R. S., and Yielding, K . L. (1971). Biochem. Biophys. Res. Commun. 45,630-636. Gaudin, D., Gregg, R. S., and Yielding, K. L. (1972). Biochem. Biophys. Res. Commun. 48, 945-949. Gellhorn, A. (1958). Cancer Res. 18, 510-517. Ghadially, F. N., and Green, H. N. (1954). Br. J. Cancer 8, 291-295. Giovanella, B. C., Stehlin, J. S. , Santamaria, C., Yim, S. O., Morgan, A. C., Williams, Jr., L. J., Leibovitz, A., Fialkow, P. J., and Mumford, D. M. (1976). J. Nail. Cancer Inst. 56, 1131-1142. Glauert, H . P., and Bennink, M. R. (1978). Res. Commun. Chem. Paihol. Pharmacol. 22, 609-6 12. Goerttler, K., and Loehrke, H. (1976a). Virchows Arch. Pathol. Anat. Hisiol. 370,97-102. Goerttler, K., and Loehrke, H. (1976b). Virchows Arch. Paihol. Anat. Hisiol. 372, 29-38. Goerttler, K., and Loehrke, H. (1977). Virchows Arch. Paihol. Anat. Histol. 376, 117-132. Goerttler, K., Loehrke, H., Schweizer, J., and Hesse, B. (1979a). Cancer Res. 39, 12931297. Goerttler, K., Loehrke, H., Schweizer, J . , and Hesse, B. (1979b). Cancer Res., 40, 155161. Goldfarb, R. H., and Quigley, J . P. (1978). Cancer Res. 38, 4601-4609. Goldstein, I. M. (1978). I n "Carcinogenesis, Vol. 11, Mechanisms of Tumor Promotion and Cocarcinogenesis" (T. J. Slaga, A. Sivak and R. K. Boutwell, eds.), pp. 389-400. Raven, New York. Green, H. (1978). In "10th Miami Symp. on Differentiation and Development" (F. Ahmad, ed.), pp. 13-36. Academic Press, New York. Green, H., and Kehinde, 0. (1974). Cell 1, 113-1 16. Greenberger, J . S . , Newburger, P. E., Karpas, A., and Moloney, W. C. (1978). Cancer Res. 38, 3340-3348. Grimm, W., and Marks, F . (1974). Cancer Res. 34, 3128-3134. Grube, D. D., Peraino, C., and Fry, R. J. M. (1975). J. Invesr. Dermaiol. 64, 258-262. Gwynn, R. H., and Salaman, M . H. (1953). Br. J . Cancer 7, 482-489. Hammarstrom, S . , Hamberg, M., Samuelsson, B., Duell, E. A., Stawiski, M., and Voorhees, J. J . (1975). Proc. Nail. Acad. Sci. U.S.A. 72, 5130-5134. Hashimoto, Y. (1978). I n "Carcinogenesis, Vol. 11, Mechanisms of Tumor Promotion and Cocarcinogenesis" (T. J. Slaga, A. Sivak and R. K. Boutwell, eds.), pp. 533-543. Raven, New York.

TUMOR PROMOTERS

67

Hecht, S. S., Thorne, R. L., Maronpot, R. R., and Hoffmann, D. (1975). J. Natl. Cancer Insr. 55, 1329-1336. Hecht, S. S ., Carmella, S. , and Hoffmann, D. (1978). J . Anal. Toxicol. 2, 56-59. Hecker, E. (1967). Naturwissenschafren 54, 282-284. Hecker, E. (1968). Cancer Res. 28, 2338-2349. Hecker, E. (1971). In “Methods in Cancer Research” (H. Busch, ed.), Vol. VI, pp. 439484. Academic Press, New York. Hecker, E. (1978). I n “Carcinogenesis, Vol. 11, Mechanisms of Tumor Promotion and Cocarcinogenesis” (T. J. Slaga, A. Sivak and R. K . Boutwell, eds.), pp. 11-48. Raven, New York. Hecker, E., and Bresch, H. (1965). 2. Narurforsch. 2Ob, 216-226. Hecker, E., and Schairer, H. U. (1967). 2. Krebsforsch. 70, 1-12. Hecker, E., and Schmidt, R. (1974). I n “Progress in the Chemistry of Organic Natural Products” (W. Herz, H. Grisebach and G. W. Kirby, eds.), pp. 377-467. SpringerVerlag, Berlin and New York. Hecker, E., Bresch, H., and v. Szczepanski, Ch. (1964a). Angew. Chem. Inr. Ed. 3, 225226.

Hecker, E., Kubinyi, H., and Bresch, H. (1964b). Angew. Chem. Int. Ed. 3, 747-748. Helmes, C. T. (Ed.) (1978). “Oncology Overview: Selected Abstracts on Tumor Promotion and Cocarcinogenesis.” International Cancer Research Data Bank, National Cancer Institute, Bethesda, Maryland. Helmes, C. T., Hillesund, T., and Boutwell, R. K. (1974). Cancer Res. 34, 1360-1365. Hennings, H., and Boutwell, R. K. (1970). Cancer Res. 30, 312-320. Hicks, R. M., and Chowaniec, J. (1977). Cancer Res. 37, 2943-2949. Hicks, R. M., Wakefield, J. St. J., and Chowaniec, J. (1973). Nature (London) 243, 347349.

Hicks, R. M., Wakefield, J . St. J., and Chowaniec, J. (1975). Chem. Biol. Interact. 11, 225-233.

Hicks, R. M., Chowaniec, J . , and Wakefield, J. St. J. (1978). I n “Carcinogenesis, Vol. 11, Mechanisms of Tumor Promotion and Cocarcinogenesis” (T. J. Slaga, A. Sivak and R. K. Boutwell, eds.), pp. 475-489. Raven, New York. Hoffmann, D., and Wynder, E. L. (1976). I n ”Chemical Carcinogens” (C. E. Searle. ed.), pp. 324-365. American Chemical Society, Washington, D.C. Hoidal, J. R., Repine, J. E., Beall, G. D., Rasp, Jr., F. L., and White, J. G. (1978). A m . J . Parhol. 91, 469-482. Holsti, P. (1959). Acra Pathol. Microbiol. Scand. 46, 51-58. Holtzer, H., Biehl, J., Yeoh, G., Meganathan, R., and Kaji, A. (1975). Proc. Null. Acad. Sci. U . S . A . 72,4051-4055. Hoober, J. K., and Bernstein, I. A. (1966). Proc. Natl. Acad. Sci. U . S . A . 56, 594-601. Hoppe, W., Brandl, F., Strell, I., Rohrl, M., Gassmann, I., Hecker, E., Bartsch, H., Kreibich, G., and v. Szczepanski, Ch. (1967). Angew. Chem. I n t . Ed. 6 , 809-810. Hozumi, M., Ogawa, M., Sugimura, T., Takeuchi, T., and Umezawa, H. (1972). Cancer Res. 32, 1725-1728. Huberman, E., and Callaham, M. F. (1979). Proc. Null. Acad. Sci. U . S . A . 76, 1293-1297. Huberman, E., Heckman, C., and Langenbach, R. (1979). Cancer Res. 39, 2618-2624. Ishii, D. N. (1978). Cuncer Res. 38, 3886-3893. Ishii, D. N., Fibach, E., Yamasaki, H., and Weinstein, I. B. (1978). Science 200,556-559. Iversen, U. M., and Iversen, 0. H. (1979). Virchows Arch. B Cell Pathol. 30, 33-42. Johnston, Jr., R. B., Lehmeyer, J. E., and Guthrie, L. A. (1976). J . Exp. Med. 143, 15511556.

68

LEILA DIAMOND et a / .

Kamine, J., and Rubin, H. (1977). J . Cell. Physiol. 92, 1-12. Kennedy, A. R., and Little, J. B. (1978). Nature (London) 276, 825-826. Kennedy, A. R., and Little, J. B. (1980). I n “Radiation Biology in Cancer Research” (R. E. Meyn and H. R. Withers, eds.), pp. 295-307. Raven, New York. Kennedy, A. R., Mondal, S., Heidelberger, C., and Little, J. B. (1978). Cancer Res. 38, 439-443. Kensler, T. W., and Mueller, G. C. (1978). Cancer Res. 38, 771-775. Kimura, N . T., Kanematsu, T., and Baba, T. (1976). Z. Krebsforsch. 87, 257-266. Kinsella, A. R., and Radman, M. (1978);. Proc. Natl. Acad. Sci. U . S . A . 75, 6149-6153. Kinsella, A., Mousset, S.. Szpirer, C., and Radman, M. (1978). I n “DNA Repair Mechanisms” (P. C. Hanawalt, E. C. Friedberg and C. F. Fox, eds.), pp. 733-738. Academic Press, New York. Kinzel, V., Schmid, E., Suss, R., and Kreibich, G. (1973). Biochem. Pharmacol. 22, 30913097. Kitagawa, T., and Sugano, H. (1978). Cann 69, 679-687. Kler, H. W., Glavind, J., and Arffmann, E. (1975). Acta Pathol. Microhiol. Scund. 83, 550-558.

Knowles, M. A. (1979). J. Nutl. Cancer Inst. 62, 349-352. Kreibich, G . , Suss, R., and Kinzel, V. (1974). 2. Krebsforsch. 81, 135-149. Krieg, L., Kiihlmann, I., and Marks, F. (1974). Cancer Res. 34, 3135-3146. Kubinski, H., Strangstalien, M. A., Baird, W. M., and Boutwell, R. K. (1973). Cancer Res. 33, 3 103-3107. Laishes, B. A., and Farber, E. (1978). J. Natl. Cancer I n s t . 61, 507-512. Laishes, B. A., Roberts, E., and Farber, E. (1978). Inr. J . Cancer 21, 186-193. Langenbach, R., and Kuszynski, C. (1975). J. Natl. Cancer I n s t . 55, 801-802. Lankas, Jr., G. R., Baxter, C. S., and Christian, R. T. (1977). Mutation Res. 45, 153-156. Lasne, C., Gentil, A., and Chouroulinkov, I. (1974). Nature (London) 247, 490-491. Lasne, C., Gentil, A,, and Chouroulinkov, I. (1977). Br. J . Cancer 35, 722-729. Lee, D. J . , Wales, J. H., and Sinnhuber, R. 0. (1971). Cancer Res. 31,960-963. Lee, L.-S., and Weinstein, I. B. (1978). Science 202, 313-315. Lee, L.-S., and Weinstein, I. B. (1979). J. Cell. Physiol. 99, 451-460. Levine, L. (1977). Nature (London) 268, 447-448. Levine, L., and Hassid, A. (1977). Biochem. Biophys. Res. Commun. 79, 477-484. Levine, L., and Ohuchi, K. (1978). Nature (London) 276, 274-275. Levinson, W., Bhatnagar, R. S., and Liu, T.-Z. (1975). J . Natl. Cancer I n s t . 55, 807-810. Lichti, U . , Slaga, T. J., Ben, T., Patterson, E., Hennings, H., and Yuspa, S. H., (1977). Proc. Natl. Acad. Sci. U . S . A . 74,3908-3912. Lichti, U.,Yuspa, S. H., and Hennings, H. (1978). I n “Carcinogenesis, Vol. 11, Mechanisms of Tumor Promotion and Cocarcinogenesis” (T. J . Slaga, A . Sivak and R. K. Boutwell, eds.), pp. 221-232. Raven, New York. Lin, F. K., Banerjee, M. R., and Crump, L. R. (1976). Cancer Res. 36, 1607-1614. Lo, S.-C., Aft, R., Ross, J., and Mueller, G. C. (1978). Cell 15, 447-453. Loewengart, G. (1977). J. Toxicol. Environ. Health 2, 539-546. Loskutoff, D. J . , and Edgington, T. S. (1977). Proc. Natl. Acad. Sci. U . S . A . 74, 39033907. Loveday, K. S., and Latt, S. A. (1979). Mutation Res. 67,343-348. Lowe, M. E., Pacifici, M., and Holtzer, H. (1978). Cancer Res. 38, 2350-2356. Lupulescu, A. (1978). Nature (London)272, 634-636. Marchok, A. C., Rhoton, J. C., Griesemer, R. A., and Nettesheim, P. (1977). Cancer Res. 37. 1811-1821.

TUMOR PROMOTERS

69

Marks, F. (1971). Z. Physiol. Chem. 353, 1273-1274. Marks, F. (1976). Cancer Res. 36, 2636-2643. Marks, F., and Grimm, W. (1972). Nature (London) New Biol. 240, 178-179. Marks, F., Bertsch, S.. Grimm, W., and Schweizer, J . (1978). In "Carcinogenesis, Vol. 11, Mechanisms of Tumor Promotion and Cocarcinogenesis" (T. J . Slaga, A. Sivak and R. K. Boutwell, eds.), pp. 97-116. Raven, New York. Marrs, J . M., and Voorhees, J . J . (1971). J . Invest. Dermutol. 56, 174-181. Martin, R. G., Setlow, V. P., and Edwards, C. A . F. (197%). J . Virol. 31, 596-607. Martin, R. G., Setlow, V. P., Edwards, C. A. F., and Vembu, D. (1979b). Cell 17, 635643. Mastro, A. M., and Mueller, G. C. (1974). Exp. Cell Res. 88, 40-46. Mastro, A. M., and Mueller, G. C. (1978). Biochim. Biophys. Actu 517, 246-254. Matsukura, N., Kawachi, T., Sano, T., Sasajima, K., and Sugimura, T. (1979). J. Cuncer Res. Clin. Oncol. 93, 323-327. McCann, J . , and Ames, B. N. (1976). Proc. Nrztl. Acad. Sci. U . S . A . 73, 950-954. Miao, R. M., Fieldsteel, A. H., and Fodge, D. W. (1978). Nature (London) 274, 271-272. Michalopoulos, G., and Pitot, H . C. (1975). Exp. Cell Res. 94, 70-78. Miskin, R., Easton, T . G., Maelicke, A , , and Reich, E. (1978a). Cell 15, 1287-1300. Miskin, R., Easton, T. G., and Reich, E. (1978b). Cell 15, 1301-1312. Mondal, S ., and Heidelberger, C. (1976). Nirrure (London) 260, 710-71 1. Mondal, S . , Brankow, D. W., and Heidelberger, C. (1976). Cancer Res. 36, 2254-2260. Mondal, S . , Brankow, D. W., and Heidelberger, C. (1978). Science 201, 1141-1142. Moroney, J . , Smith, A., Tomei, L. D., and Wenner, C. E. (1978).J. Cell. Physiol. 95, 287294. Moses, H. L., Proper, J. A., Volkenant, M. E., Wells, D. J., and Getz, M. J . (1978). Cuncer Res. 38, 2807-2812. Mottram, J. C. (1944). J . Parhol. Bucteriol. 56, 181-187. Mueller, G. C., and Kajiwara, K. (1965).In "Developmental and Metabolic Control Mechanisms and Neoplasia," pp. 452-474. Williams & Wilkens, Baltimore, Maryland. Mufson, R. A,, De Young, L. M., and Boutwell, R. K. (1977a). J . Invest. Dermatol. 69, 547-550. Mufson, R. A , , Simsiman, R. C., and Boutwell, R. K. (197713). Cuncer Rrs. 37, 665-669. Mufson, R. A,, and Astrup, E. G., Simsiman, R. C., and Boutwell, R. K. ( 1 9 7 7 ~ )Proc. . N a t l . Acad. Sci. U . S . A . 74, 657-661. Murray, A . W., and Froscio, M. (1977). Biochem. Biophys. Res. Commun. 77, 693-699. Nagasawa, H., and Little, J . B. (1979). Proc. N u l l . Acad. Sci. U . S . A . 76, 1943-1947. Narisawa, T., Magadia, N. E., Weisburger, J . H . , and Wynder, E. L. (1974). J . Nutl. Cuncer Inst. 53, 1093- 1097. Nishizumi, M. (1976). Cancer Lett. 2, 1 1 - 16. Nowell, P. C. (1976). Science 194, 23-28. O'Brien, T. G . (1976). Cancer Res. 36, 2644-2653. O'Brien, T. G . , and Diamond, L. (1977). Cancer Res. 37, 3895-3900. O'Brien, T. G., and Diamond, L. (1978a). Cancer Res. 38, 2562-2566. O'Brien, T. G . , and Diamond, L. (1978b). Cuncer Res. 38, 2567-2572. O'Brien, T. G., Simsiman, R. C., and Boutwell, R. K. (1975a). Cancer Res. 35, 1662-1670. O'Brien, T. G., Simsiman, R. C., and Boutwell, R. K. (1975b). Cancer Res. 35,2426-2433. O'Brien, T. G., Simsiman, R. C., and Boutwell, R. K. (1976). Cancer Res. 36, 3766-3770. O'Brien, T. G . , Saladik, D., and Diamond, L. (1979a). Biochem. Biophys. Res. Commun. 88, 103- 110. O'Brien, T . G., Lewis, M. A., and Diamond, L . (1979b). Canrer Res. 39, 4477-4480.

70

LEILA DIAMOND et

ul.

Ohuchi, K., and Levine, L . (1978a). J. Biol. Chem. 253, 4783-4790. Ohuchi, K . , and Levine, L. (1978b). Prosrag. Med. 1,421-431. Pacifici, M., and Holtzer, H . (1977). A m . J. Anut. 150, 207-212. Pacifici, M., Boettiger, D., Roby, K., and Holtzer, H. (1977). Cell 11, 891-899. Pastan, I . , and Willingham, M. (1978). Nature (London) 274, 645-650. Patuleia, M . C., and Friend, C. (1967). Cancer Res. 27, 726-730. Paul, D., and Hecker, E. (1969). Z. Krehsforsch. 73, 149-163. Peraino, C., Fry, R. J. M., and Staffeldt, E. (1971). Cancer Res. 31, 1506-1512. Peraino, C . , Fry, R. J. M., Staffeldt, E., and Kisieleski, W. E. (1973). Cancer Res. 33, 2701-2705. Peraino, C., Fry, R. J. M., Staffeldt, E., and Christopher, J. P. (1975). Cancer Res. 35, 2884-2890. Peraino, C., Fry, R. J. M., Staffeldt, E., and Christopher, J . P. (1977). Food Cosmet. loxicol. 15, 93-96. Peraino, C., Fry, R. J. M., and Grube, D. D. (1978). In "Carcinogenesis, Vol. 11, Mechanisms of Tumor Promotion and Cocarcinogenesis" (T. J. Slaga, A. Sivak and R. K. Boutwell, eds.). pp. 421-432. Raven, New York. Peterson, A . R., Mondal, S. , Brankow, D. W., Thon, W., and Heidelberger, C. (1977). Cancer Res. 37, 3223-3227. Pettersen, R. C., Ferguson, G., Crombie, L., Games, M. L., and Pointer, D. J . (1967). Chem. Commun. 716-717. Pitot, H. C. (1979). Annu. Rev. Med. 30, 25-39. Pitot, H. C., Barsness, L., Goldsworthy, T., and Kitagawa, T . (1978). Nature (London) 271, 456-458. Poirier, M. C., DeCicco, B. T., and Lieberman, M. W. (1975). Cancer Res. 35, 1392-1397. Quigley, J . P. (1979). Cell 17, 131-141. Raick, A . N. (1973a). Cancer Res. 33, 269-286. Raick, A. N . (1973b). Cancer Res. 33, 10%-1103. Raick, A. N. (1974). Cancer Res. 34, 920-926. Raick, A. N., and Burdzy, K. (1973). Cuncer Res. 33, 2221-2230. Raineri, R., Simsiman, R. C., and Boutwell, R. K. (1973). Cancer Res. 33, 134-139. Raineri, R . , Simsiman, R. C., and Boutwell, R. K. (1978). Cancer Lett. 5, 277-284. Ralph, P., Broxmeyer, H. E., and Nakoinz, I. (1977). J. Exp. Med. 146, 611-616. Reddy, B. S . , Narasawa, T . , Weisburger, J. H., and Wynder, E. L. (1976). J. Nutl. Cancer Inst. 56, 441-442. Reddy, B. S . , Watanabe, K., Weisburger, J. H., and Wynder, E. L. (1977). Cancer Res. 37, 3238-3242. Reddy, B. S., Weisburger, J . H., and Wynder, E. L. (1978). In "Carcinogenesis, Vol. 11, Mechanisms of Tumor Promotion and Cocarcinogenesis" (T. J . Slaga, A. Sivak and R. K. Boutwell, eds.), pp. 453-464. Raven, New York. Redmond, Jr., D. E. (1970). N . Engl. J . Med. 282, 18-23. Repine, J . E., White, J. G., Clawson, C. C., and Holmes, B. M. (1974). J . Lab. Clin. Med. 83, 91 1-920. Reuber, M. D. (1978). Environ. Health Perspect. 25, 173-200. Reznikoff, C. A., Brankow, D. W., and Heidelberger, C. (1973). Cancer Res. 33, 32313238. Rheinwald, J. G., and Green, H. (1975). Cell 6, 331-344. Richards, J . , and Nandi, S. (1978). Proc. Natl. Acad. Sci. U . S . A . 75, 3836-3840. Rifkin, D. B., Crowe, R. M., and Pollack, R. (1979). Cell 18, 361-368. Rifkin, D. B., Beal, L. P., and Reich, E. (1975). In "Proteases and Biological Control" (E.

TUMOR PROMOTERS

71

Reich, D. Rifkin and E. Shaw, eds.), pp. 841-847. Cold Spring Harbor Laboratory, New York. Robinson, H. J., and Vane, J. R. (Eds.) (1974). “Prostaglandin Synthetase InhibitorsTheir Effects on Physiological Functions and Pathological States.” Raven, New York. Robison, G. A,, Butcher, R. W., and Sutherland, E. W. (1971). “Cyclic AMP.” Academic Press, New York. Roe, F. J. C., and Peirce, W. E. H. (1960). J . Natl. Cancer Inst. 24, 1389-1403. Rohrschneider, L . R., and Boutwell, R. K. (1973). Nature (London) New Biol. 243, 212213. Rohrschneider, L. R., O’Brien, D. H., and Boutwell, R. K. (1972). Biochim. Biophys. Acta 280, 57-70. Rovera, G., and Bonaiuto, J. (1976). Cancer Res. 36, 4057-4061. Rovera, G., and Surrey, S. (1977). Cancer Res. 37,4211-4219. Rovera, G., O’Brien, T. G., and Diamond, L. (1977). Proc. Natl. Acad. Sci. U . S . A . 74, 289-2898, Rovera, G., O‘Brien, T. G., and Diamond, L. (197%). Science 204,868-870. Rovera, G., Santoli, D., and Damsky, C. (1979b). Proc. Natl. Acad. Sci. U . S . A . 76,27792783. Safiotti, U . , and Shubik, P. (1963). Natl. Cancer Inst. Monogr. 10, 489-507. Salaman, M. H., and Roe, F. J . C. (1964). Br. Med. Bull. 20, 139-144. Schinitsky, M. R., Hyman, L. R., Blazkovec, A. A,, and Burkholder, P. M. (1973). Cancer Res. 33, 659-663. Schmidt, R., and Hecker, E. (1975). Cancer Res. 35, 1375-1377. Schulte-Hermann, R. (1974). CRC Crit. Rev. Toxicol. 3, 97-158. Schwarz, J. A., Viaje, A., Slaga, T. J., Yuspa, S. H., Hennings, H., and Lichti, U. (1977). Chem. Biol. Interact. 17, 331-347. Scribner, J. D., and Boutwell, R. K. (1972). Eur. J. Cancer 8, 617-621. Scribner, J . D., and Slaga, T. J . (1973). Cancer Res. 33, 542-546. Scribner, J. D., and Suss, R. (1978). I n “International Review of Experimental Pathology” (G. W. Richter and M. A. Epstein, eds.), Vol. XVIII, pp. 137-198. Academic Press, New York. Segal, A., Van Duuren, B. L., and Mate, U. (1975). Cancer Res. 35, 2154-2159. Segal, A., Van Duuren, B. L., Mate, U . , Solomon, J. J., Seidman, I., Smith, A., and Melchionne, S. (1978). Cancer Res. 38, 921-925. Selikoff, I. J., and Hammond, E. C. (1975). I n “Persons at High Risk of Cancer: An Approach to Cancer Etiology and Control” (J. F. Fraumeni, Jr., ed.), pp. 467-483. Academic Press, New York. Setala, H. (1956). Acta Pathol. Microbiol. Scand. Suppl. 115, 7-91. Setala, K. (1960). Prog. Exp. Tumor Res. 1, 225-278. Shoyab, M., De Larco, J . E., and Todaro, G. J . (1979). Nature (London) 279, 387-391. Shubik, P. (1950). Cancer Res. 10, 13-17. Sibrack, L. A., Gray, R. H., and Bernstein, I. A. (1974). J. Invesr. Dermatol. 62, 394-405. Sice, J. (1966). Toxicol. Appl. Pharmacol. 9, 70-74. Sisskin, E . E., and Barrett, J . C. (1979). Proc. A m . Assoc. Cancer Res. 20, 197. Sivak, A. (1972). J. Cell. Physiol. 80, 167- 173. Sivak, A. (1977). In Vitro 13, 337-343. Sivak, A., and Van Duuren, B. L. (1967). Science 157, 1443-1444. Sivak, A,, and Van Duuren, B. L. (1970). J . Natl. Cancer I n s t . 44, 1091-1097. Sivak, A., Ray, F., and Van Duuren, B. L. (1%9). Cancer Res. 29,624-630. Slaga, T. J., Thompson, S., and Smuckler, E. A. (1975). J. Narl. Cancer I n s / . 54, 931-936.

72

LEILA DIAMOND

et al.

Slaga, T. J., Scribner, J. D., Thompson, S., and Viaje, A. (1976). J. Nail. Cancer Inst. 57, 1145- 1149. Slaga, T. J., Sivak, A., and Boutwell, R. K. (Eds.) (1978). “Carcinogenesis, Vol. 11, Mechanisms of Tumor Promotion and Cocarcinogenesis.” Raven, New York. Smythies, J. R., Benington, F., and Morin, R. D. (1975). Psychoendocrinol. 1, 123-130. Solt, D., and Farber, E. (1976). Nature (London) 263, 701-703. Solt, D. B., Medline, A.. and Farber, E. (1977). Am. J . Pathol. 88, 595-618. Soprano, K. J . , and Baserga, R. (1979). Proc. Natl. Acad. Sci. U.S.A., in press. Spies, J. R. (1935). J . Am. Chem. Soc. 57, 182-184. Steele, V. E., and Marchok, A. C. (1979). Proc. Am. Assoc. Cancer Res. 20, 136. Steele, V. E., Marchok, A. C., and Nettesheim, P. (1977). Int. J . Cancer 20, 234-238. Steele, V. E., Marchok, A. C., and Nettesheim, P. (1978a). In “Carcinogenesis, Vol. 11, Mechanisms of Tumor Promotion and Cocarcinogenesis” (T. J. Slaga, A. Sivak and R. K. Boutwell, eds.), pp. 289-300. Raven, New York. Steele, V. E., Marchok, A. C., and Nettesheim, P. (1978b). Cancer Res. 38, 3563-3565. Steinert, P., and Yuspa, S. H. (1978). Science 200, 1491-1493. Stenback, F., Garcia, H., and Shubik, P. (1974). In “The Physiopathology of Cancer, Vol. I, Biology and Biochemistry,” pp. 155-225. Karger, Basel. Sun, T.-T., and Green, H. (1976). Cell 9, 51 1-521. Suss, R., and Schuster, A. (1974). Experientia 30, 81. Suss, R., Kinzel, V . , and Kreibich, G. (1971). Experientia 27, 46-47. Suss, R., Kreibich, G., and Kinzel, V. (1972). Eur. J . Cancer 8, 299-304. Tannenbaum, A. (1959). In “The Physiopathology of Cancer,” 2nd Ed., pp. 517-562. Hoeber-Harper, New York. Terracini, B., Shubik, P.. and Della Porta, G. (1960). Cancer Res. 20, 1538-1542. Thielmann, H. W., and Hecker, E. (1969). In “Fortschritte der Krebsforschung” (C. G. Schmidt and 0. Wetter, eds.), Vol. VII, pp. 171-179. Schattauer, Stuttgart, New York. Todaro, G. J., De Larco. J . E., and Sporn, M. B. (1978). Nature (London) 276, 272-274. Tonelli, Q. J., Custer, R. P., and Sorof, S. (1979). Cancer Res. 39, 1784-1792. Touraine, J.-L., Hadden, J. W., Touraine, F., Hadden, E. M.,Estensen, R., and Good, R. A. (1977). J . Exp. Med. 145, 460-465. Toyama, Y., West, C. M., and Holtzer, H . (1979). Am. J . Anat., 56, 131-137. Troll, W., Klassen, A., and Janoff, A. (1970). Science 169, 1211-1213. Troll, W., Rossman, T., Katz, J., Levitz, M., and Sugimura, T. (1975). In “Proteases and Biological Control” (E. Reich, D. B. Rifkin and E. Shaw, eds.), pp. 977-987. Cold Spring Harbor Laboratory, New York. Troll, W., Meyn, M. S., and Rossman, T. G. (1978). In “Carcinogenesis, Vol. 11, Mechanisms of Tumor Promotion and Cocarcinogenesis” (T. J. Slaga, A Sivak and R. K. Boutwell, eds.), pp. 301-312. Raven, New York. Trosko, J . E., Yager, Jr., J. D., Bowden, G. T., and Butcher, F. R. (1975). Chem. Biol. Interact. 11, 191-205. Trosko, J. E., Chang, C.-C., Yotti, L. P., and Chu, E. H. Y. (1977). Cancer Res. 37, 188193. Twort, S . M., and Twort. C. C. (1939). Am. J . Cancer 35, 80-85. Vaheri, A., Ruoslahti, E.. and Mosher, D. F. (Eds.) (1978). “Fibroblast Surface Protein.” Annals N.Y. Acad. Sci., Vol. 312, New York Academy of Sciences, New York. Van Duuren, B. L. (1969). Prog. Exp. Tumor Res. 11, 31-68. Van Duuren, B. L. (1976). In “Chemical Carcinogens” (C. E. Searle, ed.), pp. 24-51. American Chemical Society, Washington, D.C. Van Duuren, B. L., and Goldschmidt, B. M. (1976). J . Nutl. Cancer Inst. 56, 1237-1242.

T U M O R PROMOTERS

73

Van Duuren, B. L., and Orris, L . (1965). Cancer Res. 25, 1871-1875. Van Duuren, B. L., Sivak, A., Katz, C., and Melchionne, S. (1971). J . Natl. Cancer I n s t . 47, 235-240. Van Duuren, B. L., Banerjee, S., and Witz, G. (1976). Chem. Biol. Interact. 15, 233-246. Van Duuren, B. L., Witz, G., and Goldschmidt, B. M. (1978a). In “Carcinogenesis, Vol. 11.. Mechanisms of Tumor Promotion and Cocarcinogenesis” (T. J. Slaga, A. Sivak and R. K. Boutwell, eds.), pp. 491-507. Raven, New York. Van Duuren, B. L., Segal, A,, Tseng, S.-S., Rusch, G. M., Loewengart, G., Mate, U., Roth, D., Smith, A., Melchionne, S., and Seidman, I. (1978b). J . Med. Chem. 21, 2631. Van Duuren, B. L., Tseng, S.-S., Segal, A., Smith, A. C., Melchionne, S., and Seidman, I . (1979). Cancer Res. 39, 2644-2646. Vassalli, J.-D., Hamilton, J., and Reich, E. (1977). Cell 11, 695-705. Verma, A. K., and Murray, A. W. (1974). Cuncer Res. 34, 3408-3413. Verma, A. K . , Rice, H . M., and Boutwell, R. K. (1977). Biochem. Biophys. Res. Commun. 79, 1160- 1166. Verma, A. K., Rice, H . M., Shapas, B. G., and Boutwell, R. K. (1978). Cancer Res. 38, 793-801. Verma, A. K., Shapas, B. G., Rice, H . M., and Boutwell, R. K. (1979). Cancer Res. 39, 4 19-425. Viaje, A,, Slaga, T . J., Wigler, M., and Weinstein, I. B. (1977). Cancer Res. 37, 15301536. Voorhees, J . J., Duell, E. A., Stawiski, M., and Harrell, E. R. (1974). A d v . Cyclic Nucleotide Res. 4, 117-162. Warren, L., Critchley, D., and Macpherson, I. (1972). Nulure (London) 235, 275-278. Watanabe, K., and Williams, G. M. (1978). J . Nut/. Cancer Inst. 61, 1311-1314. Weinstein, I. B., Wigler, M . , and Pietropaolo, C. (1977). I n “The Origins of Human Cancer” (H. H. Hiatt, J . D. Watson and J . A. Winsten, eds.), pp. 751-772. Cold Spring Harbor Laboratory, New York. Weinstein, I. B., Wigler, M., Fisher, P. B.. Sisskin, E., and Pietropaolo, C. (1978a). I n “Carcinogenesis, Vol. 11. Mechanisms of Tumor Promotion and Cocarcinogenesis” (T. J . Slaga, A. Sivak and R. K. Boutwell, eds.), pp. 313-333. Raven, N e w York. Weinstein, I. B.. Wigler, M., Yamasaki, H., Lee, L.-S., Fisher, P. B., and Mufson. A. (1978b). In “Biological Markers of Neoplasia: Basic and Applied Aspects“ (R. W. Ruddon. ed.), pp. 451-471. Elsevier, Amsterdam and New York. Weinstein, I. B., Yamasaki, H . , Wigler, M., Lee, L.-S., Fisher, P. B., Jeffrey, A., and Grunberger, D. (1979). In “Carcinogens: Identification and Mechanisms of Action” (A. C. G r i f i n and C. R. Shaw, eds.), pp. 399-418. Raven, New York. Weisburger, J. H . , Madison, R. M., Ward, J . M., Viguera, C., and Weisburger, E. K. (1975). J . Null. Cancer Inst. 54, 1185- 1188. Wertz, P. W., and Mueller, G. C. (1978). Cancer Res. 38, 2900-2904. Wertz, P. W., Kensler, T. W., Mueller, G. C., Verma, A. K., and Boutwell, R. K. (1979). Nature (London) 277, 227-229. White, J. G . , and Estensen, R. D. (1974). A m . J . Pathol. 75, 45-60. White, J . G., and Repine, J . E. (1978). A m . J . Pathol. 91, 571-580. White, J . G., Rao, G. H . R., and Estensen, R. D. (1974). A m . J . Pathol. 75, 301-314. Whitfield, J . F., MacManus, J . P., and Gillan, D. J . (1973). J . Cell. Physiol. 82, 151-156. Wigler, M., and Weinstein, I . B. (1976). Nature (London) 259, 232-233. Wigler, M., DeFeo, D., and Weinstein, I . B. (1978). Cancer Res. 38, 1434-1437. Williams, G. M. (1976). Methods Cell B i d . 31, 357-364.

74

LEILA DIAMOND

et al.

Wilson, E. L., and Reich, E. (1978). Cell 15, 385-392. Wilson, E. L., and Reich, E. (1979). Cancer Res. 39, 1579-1586. Wilson, S. R., and Huffman, J. C. (1976). Experientia 32, 1489-1490. Witkin, E. M. (1976). Bacreriol. Rev. 40, 869-907. Witschi, H., and Lock, S. (1978). In “Carcinogenesis, Vol. 11, Mechanisms of Tumor Promotion and Cocarcinogenesis” (T. J. Slaga, A. Sivak and R. K. Boutwell, eds.), pp. 465-474. Raven, New York. Wynder, E. L., and Hoffmann, D. (1961). Cancer 14, 1306-1315. Wynder, E. L., and Stellman, S . D. (1977). Cancer Res. 37, 4608-4622. Yam, L. T., Li, C. Y., and Crosby, W. H. (1971). Am. J. Clin. Pathol. 55, 283-290. Yamagiwa, K., and Ichikawa, K. (1918). J. Cancer Res. 3 , 1-29. Yamamoto, N., and zur Hausen, H. (1979). Nature (London) 280, 244-245. Yamamoto, N., Bister, K., and zur Hausen, H. (1979). Nature (London) 278, 553-554. Yamasaki, H., Fibach, E., Nudel, U., Weinstein, I. B., Rifkind, R. A., and Marks, P. A. (1977). Proc. Nail. Acad. Sci. U . S . A . 74, 3451-3455. Yamasaki, H., Fibach, E., Weinstein, I. B., Nudel, U., Rifkind, R. A., and Marks, P. A. (1979a). In “Oncogenic Viruses and Host Cell Genes” (Y. Ikawa and T . Odaka, eds.), pp. 365-376. Academic Press, New York. Yamasaki, H., Weinstein, I. B., Fibach, E., Rifkind, R. A., and Marks, P. A. (1979b). Cancer Res. 39, 1989-1994. Yuspa, S . H., and Harris, C. C. (1974). Exp. Cell Res. 86, 95-105. Yuspa, S . H., Ben, T., Patterson, E., Michael, D., Elgjo, K., and Hennings, H. (1976a). Cancer Res. 36, 4062-4068. Yuspa, S. H., Lichti, U., Ben, T. Patterson, E., Hennings, H., Slaga, T. J., Colburn, N., and Kelsey, W. (1976b). Nature (London) 262, 402-404. Yuspa, S . H., Viguera, C., and Nims, R. (1979). Cancer Lett. 6, 301-310. Zabos, P., Kyner, D., Mendelsohn, N., Schreiber, C., Waxman, S., Christman, J., and Acs, G. (1978). Proc. Natl. Acad. Sci. U . S . A . 75, 5422-5426. Zucker, M. B., Troll, W., and Belman, S. (1974). J . Cell. Biol. 60, 325-336. zur Hausen, H., O’Neill, F. J., Freese, U. K., and Hecker, E. (1978). Nature (London) 272, 373-375. zur Hausen, H., Bornkamm, G. W., Schmidt, R., and Hecker, E. (1979). Proc. Natl. Acad. Sci. U . S . A . 76, 782-785.

SHEDDING FROM THE CELL SURFACE OF NORMAL AND CANCER CELLS*

Paul H. Black Hubert H. Humphrey Cancer Research Center and Department of Microbiology, Boston University School of Medicine. Boston, Massachusetts

I . Introduction . . . . . . . . . . . . . . . . . . . . . . 11. Membrane Structure

111.

1v.

V.

VI.

VII.

VIII.

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

Endoskeletal and Exoskeletal Protein Factors Maintaining Membrane ................................................. Structure Synthesis M a r Translocation of Released Proteins . . . . . . . . . . . . . A. Secretion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Exocytosis and Membrane Fusion.. . . . . . . . . . . . . . . . . . . . . . . , . . . . . . . . . C. Stimulus-Secretion Coupling D. Energy Requirements an E. Synthesis of Membrane ( F. Vesicular Stomatitis Virus Shedding and Membrane Prote A. Characteristics of Sheddin ................................ B. Plasma Membrane Protein Shedding and the Activated State A. Cell Activation ................................................... B. Shedding from Growing Normal and Cancer Cells C. Shedding from Mitogen-Stimulated Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Biochemical Events of Cell Activation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Shedding and Activated Specific Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Ovulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Eggs-Fertilization C. Embryogenesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Photoreceptor Cells-Rods . . . . E. Shedding from Cells F. Shedding from Virus Mechanism of Shedding A. Proteins and Glycop B. Membrane Structures: Microvilli, Vesi Consequences of Shedding from the Cancer A. Cancer Cell Surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Cryofibnnogenemia C. Shedding, Coagulation, and Fibnnolysis in Cancer . . . . . . . . . . . . . . . . . . . D. Invasion and Metastases .......................................... E. Shedding and Tumor Immunity . . . . ...... F. Glycosyl Transferases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

76 78 79 85 85 88 92 93 96 97 99 99 101 104 104 105

116 121 125 125 126 127 127 130 140 144 144 147 148 149 151 153 156 161 168

* Publication number 2 of the Hubert H. Humphrey Cancer Research Center. 75 Coovrieht @3 1980 bv Academic Press, Inc.

76

PAUL H. BLACK

IX. Shedding and Chronic Viral Disease . . ........................... X. Activation and Surface Proteas A . Proteases and Chemotaxis B. Rheumatoid Arthritis.. . . . . . . . . . . ................... XI. Prevention of Shedding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XII. Summary and Conclusions.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References ............................................................

172

177 178 179 180

I. Introduction

This article will be concerned with a meagerly studied aspect of cell membrane turnover, namely, the release of macromolecules from the surface of viable cells. Proteins, glycoproteins, lipoproteias, glycolipids, and glycosaminoglycans, either alone or in combination with other cell membrane constituents, may be released. This phenomenon has been referred to by different names: desquamation (Kraemer and Tobey, 1972); exfoliation (Koch and Smith, 1978); sloughing; or shedding (Kapeller et al., 1973). The term shedding will be utilized in this article. Shedding implies the release, in soluble or particulate form, of cell surface constituents without affecting cell viability and should be distinguished from secretion, which implies the release of the soluble contents of secretory vesicles by exocytosis. Shedding occurs normally and, as such, plays an important role in membrane turnover. It also appears to play a role in the following phenomena: determination of certain cell-to-cell interactions during morphogenesis and embryological development; the production of factors mediating various immunological interactions between cells; the release of certain cell membrane proteases; the production of growth factors from cells; and the modulation of certain cell surface receptors. The phenomenon of shedding is especially important in cancer cells where release of cell surface components such as proteases and adhesion molecules occurs continually. Shedding from neoplastic cells may therefore be an important determinant of the cancer cell phenotype and of the pathophysiology of cancer. In this article, consideration will be given to the following: the cell surface of eukaryotic cells and the factors maintaining its stability; what is known about the synthesis, transport, and insertion of cell surface proteins and glycoproteins; the phenomenon of shedding and its role in plasma membrane turnover; shedding during normal cell growth, its role in causing the phenotype of the mitotic cell, and its relationship to cell

CELL SURFACE SHEDDING

77

activation. Shedding from eggs, embryonic cells, retinal rods, lymphocytes, macrophages, and virus-infected cells will also be described. Finally, shedding from the cancer cell, as an extreme example of the activated cell, and the relationship between shedding and the cancer cell phenotype will be considered. Some possible consequences of shedding from the cancer cell will be explored with respect to the following: separation, infiltration, and metastases; abnormal hematological events such as enhanced coagulation and/or fibrinolysis and cryofibrinogenemia; immunological hyporesponsiveness and the role of various blocking and suppressor factors in its causation; and the appearance in the serum of certain enzymes such as glycosyl transferases that may be used for diagnostic purposes. Throughout the article, evidence will be marshalled that indicates the coupling of synthesis of plasma membrane proteins with membrane insertion and shedding; this occurs in both activated normal cells and cancer cells. Since there is evidence that certain proteases are components of the cell surface, their presence during activation may play a role in shedding and in the movement of normal cells through tissues (which occurs during embryogenesis), in chemotactic movement, and wound regeneration. However, proteolytic activity, either present on the cell surface or shed from cells activated by an unknown stimulus, may be important in the pathophysiology of certain disease processes. This will be considered in relation to a disease such as rheumatoid arthritis. Many of the studies that will be reviewed have been carried out with cells from different species cultivated in vim) or in animals. In addition, studies utilizing transformed cells' will be frequently mentioned. Differences between species certainly exist, and it is always hazardous to extrapolate from one species to another. Whenever possible, studies utilizing human cells will be quoted. Moreover, results may be affected by conditions of tissue culture, such as passage number of cells, mycoplasma contamination, or the physiological state of the cells. Notwithstanding these differences, an attempt has been made to summarize and synthesize the results and to present an overview of the basis for and significance of cell surface shedding. Of necessity, several complicated and controversial subjects could not be examined in detail, and whenever possible, the reader will be referred to appropriate reviews of the subject.

* Transformed cells are cells rendered neoplastic in vitro by viruses, chemicals, or various forms of radiation. In general, they resemble the tumor cells induced in animals by these agents and have been extensively studied as in vitro models of carcinogenesis.

78

P A U L H. BLACK

II. Membrane Structure

There is general agreement that the plasma membrane is a lipid bilayer oriented with polar regions outside and nonpolar regions inside. In the interior, hydrophobic portion of the membrane, hydrocarbon chains, which are held together by London-van der Waals interactions, can achieve considerable lateral motion; thus a more fluid state exists in the interior of the bilayer leaflet. The hydrophilic head groups of the plasma membrane phospholipids interact with nonlipid material, largely proteins, by ionic bonding and perhaps other noncovalent interactions (see Nicolson et al., 1977 for review). There has been much speculation as to whether the lipid bilayer is continuous and as to the nature of the interrelationship between membrane protein and membrane lipid. For the purposes of this article, it is convenient to view the membrane as a mosaic of lipid and protein (Singer and Nicolson, 1972; Singer, 1974a). According to this model, “integral” membrane proteins are globular and amphipathic, i .e., their three-dimensional structures are segmented into hydrophilic and hydrophobic regions, with the hydrophilic regions (containing most of the ionic and highly polar amino acid residues of the protein) protruding from the membrane, either on the external or the cytoplasmic surfaces, and with the hydrophobic regions embedded within the nonpolar interior of the lipid bilayer (Singer, 1974a). An integral protein might span the membrane and protrude from both surfaces if it had two hydrophilic ends separated by a hydrophobic middle segment. Thus, proteins are thought to be enmeshed in a discontinuous lipid bilayer that forms the matrix of the mosaic, and, since the membrane lipid is generally fluid rather than crystalline under physiological conditions, the mosaic is a dynamic one with its components able to undergo lateral diffusion in the plane of the membrane (Singer, 1977; Stein, 1972). In addition to the integral membrane proteins, another class of proteins, designated “peripheral” proteins, are also associated with the membrane. Such proteins are weakly bound to the surfaces of membranes and do not appear to interact with the membrane lipids, in contrast to the integral membrane proteins that are bound more strongly to the membrane and exhibit functionally important interactions with membrane lipids. The peripheral proteins can associate with integral membrane proteins, and perhaps also glycolipids, by ionic interactions and are removed from the cell membrane by mild procedures such as exposure to an environment of high ionic strength or to metal ion-chelating agents. Whereas, removal of integral proteins requires the use of more extreme measures such as

CELL SURFACE SHEDDING

79

treatment with detergents, organic solvents, or chaotropic agents (Singer, 1974a; Nicolson et al., 1977). A region of the plasma membrane external to the trilaminar membrane of most vertebrate cells is composed of a morphologically detectable cell coat or glycocalyx and is the location of many peripheral membrane proteins (Luft, 1971). The precise composition and molecular organization of the cell coat is not completely understood and may be quite variable, but its basic composition is known. It is primarily composed of small, branching carbohydrate chains that vary in length from one to twenty sugars and are covalently linked to protein or lipid. In addition to these glycoproteins and glycolipids, high-molecular-weight linear carbohydrate molecules called glycosaminoglycans are also present; these include hyaluronic acid, chondroitin sulfate, and heparan sulfate (Vogel and Kelley, 1977; Rittenhouse et a/., 1978). The glycocalyx is thought to be important in various aspects of cell-to-cell interaction such as cell recognition and cell adhesion. The glycocalyx may be rather labile, particularly in malignant cells where it may be rapidly and spontaneously released into the medium while the cells remain viable (Kim et al., 1975; Rittenhouse et a/., 1978). The plasma membrane can, therefore, be viewed as a two-dimensional solution of a mosaic of integral membrane proteins embedded in a lipid bilayer with peripheral proteins bound loosely to either surface. Such an arrangement has two important implications. First, it permits the membrane to be organized asymmetrically with certain components localized predominantly on the outer or inner half of the membrane. Thus, glycolipids and glycoproteins are localized exclusively on the external surface of the membrane where they function as receptors for antibodies, viruses, lectins, hormones, etc. Second, components can diffuse laterally, permitting rapid and reversible changes in the topography of specific surface components. Some proteins diffuse freely; however, the existence of constraining elements in the cytoplasm prevents some components from free movement and completely random distribution. The cell, therefore, can exert control over the distribution of certain components by restraining movement and can therefore maintain specific topographic arrange1977). ments or patterns (Nicolson et d., ENDOSKELETAL A N D EXOSKELETAL PROTEIN FACTORS MAINTAINING MEMBRANE STRUCTURE

The plasma membrane gains structural stability by interactions with proteins inside the cell and from side-to-side interactions of

80

PAUL H. BLACK

(glyco)proteins2 on the external surface of the cell; these have been designated the endoskeleton and exoskeleton, respectively (see Black et a/., 1975, and Roblin et al., 1975b, for reviews). 1. Endoskeleton

The endoskeletal or cytoskeletal system is composed of tubules and filaments that maintain cell structure and affect various movments of cells, organelles, and components of the plasma membrane. Microtubules are polymeric complexes of the protein tubulin arranged into long tubular structures with a diameter of approximately 250 A (see Olmsted and Borisy, 1973, for review). Microtubules are thought to play important roles in maintaining the overall shape and stability of the cell, in intracellular movement of organelles, and, perhaps, in the movement of certain cell surface components. The evidence that the degree of microtubule polymerization influences the stability of plasma membrane lectin3 binding sites and their movement when exposed to exogenous lectin is mostly indirect and is based on the observation that depolymerization of the microtubules (with colchicine or at 4°C) results in patch and cap formation of certain cell surface receptors (e.g., Con A receptors) in a number of different cell types (Edelman et al., 1973; Toh and Hard, 1977, Williams et al., 1977). Recently, direct evidence that microtubules, perhaps in concert with microfilaments, play a role in receptor movement has been obtained (Gabbiani et ul., 1977) (see following). The microfilament system is composed of actin-containing filaments that are 50-70 or 100 A in diameter (Goldman and Follett, 1969; McNutt et al., 1971, 1973; Perdue, 1973, see Hitchcock, 1977, for review); the former are frequently aggregated into bundles in the peripheral cytoplasm of confluent cells and are abundant in regions of cell-to-cell apposition or at areas of contact with the dish, where they form dense plaques and associate with the plasma membrane (McNutt et al., 1971, 1973; Perdue, 1973; Revel et al., 1974). Fluorescent antibody studies have revealed that myosin, as well as actin, is present in the microfilament bundles, and that large bundles traverse normal fibroblast cells of a number of species,

* See footnote 4, Section IV,B. Lectins such as concanavalin A (Con A) are generally derived from plants and are divalent or multivalent carbohydrate-binding proteins that cross-link specific carbohydrate components of the cell surface. Such cross-linking may result in patch and cap formation in an individual cell or in cell-to-cell agglutination when a critical number of intercellular bridges have been achieved by a multivalent lectin.

CELL SURFACE SHEDDING

81

including humans (Lazarides and Weber, 1974). The presence of microfilament bundles in the anterior expansions of cells, in microvilli, and in ruffled membranes suggest that they play a role in cell motility (McNutt et al., 1973; Hitchcock, 1977). Thus, it is likely that nonmuscle cells have an actin-linked, contractile system consisting of subunits that can be reversibly polymerized or depolymerized within the cell. Such a system could supply the mechanical force required for intracellular translocation of organelles and for the movement of cell surface receptors. However, there is no convincing evidence at present that a troponin complex, similar to that found in skeletal muscle, exists in nonmuscle cells (Hitchcock, 1977). Some component of the contractile apparatus would have to be anchored to a membrane to explain the movements that take place and, in fact, morphological evidence indicates that such attachment is present in many types of cells (Hitchcock, 1977). Attachment of microfilaments may be direct or by bridges and a-actinin has been demonstrated at the attachment sites (Mooseker and Tilney, 1975). a-Actinin is the binding or anchoring protein in the Z-line of muscle, which provides a precedent for this type of association of contractile proteins. Furthermore, a-actinin has been found to be a component of certain secretory vesicles, further suggesting that it may link microfilaments to membranes and thereby provide contractile force during intracellular translocation (Jockusch et ul., 1977). The exact nature, however, of the association between actin and the plasma membrane is not known. Much evidence has accumulated indicating that the endoskeletal elements, microfilaments and microtubules, control the movement and distribution of various receptors on the cell surface (see Nicolson, 1976a,b, for review). Both microfilaments and microtubules have been thought to stabilize or anchor the membrane and to restrict the mobility of certain cell surface receptors. In addition to providing structural stability to the membrane, evidence indicates that actin may actually cause the movement of cell surface receptors. For example, surface antigens of lymphocytes move into patches and caps in relation to actin filaments (Bray, 1978). Furthermore, several studies utilizing various fluorescent techniques have demonstrated that the region of a lymphocyte immediately beneath an antigen cap contains an accumulation of actin (Toh and Hard, 1977; Bourguignon and Singer, 1977), myosin, and tubulin (Schreiner et ul., 1977; Gabbiani et a l . , 1977);this is also true of patching (Bourguignon and Singer, 1977). Precisely how the ligand receptor becomes associated with actin is not known. Several studies indicate that surface antigens, lectin receptor sites, antibodies, etc., become associated with microfilaments when they

82

PAUL H . BLACK

cluster or aggregate in response to a multivalent ligand (Ash and Singer, 1976; Ash et al., 1977; Bourguignon and Singer, 1977; Flanagen and Koch, 1978). Other studies, however, indicate that Con A receptors are linked to actin filaments normally by an actin subset bound to the plasma membrane, which normally limits the movement of Con A receptors (Toh and Hard, 1977; Williams et al., 1977). Conversion of this subset of bridging actin filaments to aggregates may be necessary for clustering and capping, and binding to Con A in some way accomplishes this. Thus, although aggregates of actin may remain associated with the membrane, dissociation from an anchoring microfilament system appears to be necessary for receptor movement. In general, treatments that act to depolymerize microfilaments (e.g., exposure to cytochalasin B) enhance clustering of receptors. Also, transformed or actively dividing normal cells, which have a poorly developed microfilament system, show enhanced clustering in response to a lectin (Toh and Hard, 1977). Moreover, exposure to various proteolytic enzymes, which causes the dissolution of microfilaments and formation of aggregates, results in an increase in the lateral mobility of Con A sites. These observations indicate that intact microfilaments may restrain receptor mobility. The role of microtubules in directing receptor movement is not entirely understood. Recent studies indicate that tubulin cocaps with actin when the surface Ig of splenic lymphocytes is capped with antibody against this receptor (Gabbiani et al., 1977). This suggests that microtubules and microfilaments are linked to each other or, possibly, to the same plasma membrane component. Close physical proximity of microtubules and microfilaments has been documented (Goldman and Follett, 1969; Goldman, 1971; Perdue, 1973), and microtubules have been shown to be linked to cellular membranes (Brinkley et al., 1975). Thus far, however, the mechanisms by which they might interact are not known. Moreover, there is no universal agreement that microtubules play an active role in movement of surface receptors (Bourguignon and Singer, 1977). In summary, it is clear that the microfilaments are attached to the plasma membrane, either directly or by a subset of microfilaments, and that they may be anchored by a-actinin. Whether certain receptors are associated with actin or become associated during receptor distribution is not known. It is also clear that surface receptors are associated with aggregated actin as they patch and cap in response to their various ligands and that they most likely are moved by these contractile proteins. It also seems likely that dissociation from the microfilament bundles may be a necessary prerequisite for receptor movement and that this may be accomplished by depolymerization of either an anchoring subset or the microfilament bundles to aggregates. The fact that normal growing cells,

CELL SURFACE SHEDDING

83

transformed cells, and cells treated with proteolytic enzymes or cytochalasin B, have poorly developed microfilament systems and relatively mobile receptor sites argues that intact microfilaments restrain receptor movement. It is of interest that the plasma membrane of Chinese hamster ovary cells, which has a transformed phenotype and a poorly developed microfilament system, has many knobs or blebs that are associated with aggregated actin deposits and that “oscillate” continuously. Exposure of these cells to CAMP, or agents that increase CAMP levels, results in reversion to a normal phenotype containing a well developed microfilament system that is associated with a more tranquil cell membrane (Puck, 1977). 2 . Exoskeleton The factors that regulate the polymerization of the microfilament system are largely unknown, but calcium and the cyclic nucleotides are likely to be of importance. These, however, are apparently influenced by another factor affecting the extent of microfilament polymerization, at least for cells observed in vitro; this factor is the degree of cell spreading. Thus, cells going from a round (or suspended) to a more flattened morphology quickly develop an organized system of microfilaments, whereas the converse occurs when cells round up and become detached; in the latter situation, the primary event is usually a disintegration of the microfilament system (see Rees et d.,1977, for review). There is evidence that the organization of surface proteins, glycoproteins, and glycosaminoglycans to each other and/or to one another provides for the development of the microfilament system, possibly by direct attachment to transmembrane proteins. It is of interest that the distributions of a glycocalyx component, fibronectin (see later), and microfilaments are similar, and both are lost or return in parallel upon rounding or flattening of a cell, respectively (Wells and Mallucci, 1978). Exoskeletal elements provide for the adhesion of cells to each other, and/or to the intercellular matrix, and/or to the substrate on which the cells are growing, i.e., the “stick” of the peripheral proteins. Conditions enhancing a more intact microfilament system (by means of their association with the plasma membrane) would further enhance the adhesion of cells, i.e., the “grip” of the microfilament system (Rees et al., 1977). If lateral associations of peripheral proteins are important in maintaining structural stability of membranes, their presence should limit movement of surface receptors and their absence should promote such lateral movement. A number of studies with transformed cells (where certain peripheral membrane proteins are lost spontaneously) or trypsinized cells

84

PAUL H . BLACK

(where similar proteins are released by proteolysis) indicate that this is the case. For example, addition of exogenous glycosaminoglycans such as dextran sulfate causes transformed cells to become more spread and reduces their saturation density and their agglutinability by Con A (Goto et al., 1973). Addition of another glycocalyx component, fibronectin, to transformed cells causes the same effects (Yamada et al., 1977; Willingham et d., 1977). Normal cells, treated with certain proteases, lose as much as 80% of their glycosaminoglycan (Roblin et al., 1975a) (as well as certain other peripheral proteins) generally round up, lose their microfilament system, and agglutinate very readily with Con A (Roblin et al., 1975b, for review). We have defined the term exoskeleton to include the normal lateral associations of surface proteins, glycoproteins, and glycosaminoglycans, which provide for cell adhesion and thereby the extent of cell spreading; such associations result in stabilization of cell membrane receptor sites (Black et al., 1975; Roblin et al., 1975b). Whether the latter effect can occur directly, or is an indirect effect of microfilament polymerization secondary to adhesion and cell spreading, is not known. The importance of the exoskeleton will be stressed repeatedly, as will the consequences of its loss by shedding.

3. Calcium Metabolism and Membranes Calcium (Ca), one of many plasma membrane constituents, is important in maintaining the integrity of plasma membrane structure, and a brief review will be given of the complex pathway of Ca metabolism. The concentration of Ca in the cytoplasm of a resting eukaryotic cell M ) is maintained at 104-fold below the concentration outside the cell ( lop3M ) , and various homeostatic mechanisms maintain cytosol Ca at low concentrations. Intracellular calcium is distributed asymmetrically between the principal sequestering sites from which it is rapidly exchangeable: the mitochondria, endoplasmic reticulum, and the plasma membrane (Moore and Pastan, 1977; see Rasmussen and Goodman, 1977, and Mikkelsen, 1978, for reviews). A complex set of factors, including the energy charge of the cells, the Na+ distribution across the plasma membrane, the pH and phosphate concentration extracellularly are responsible for maintaining the asymmetric distribution of Ca. Its concentration in the cytosol is achieved by pump-leak systems at both the plasma membrane and the inner mitochondria1 membrane. Ca influx into the cell is a passive process, probably by carrier-mediated facilitated diffusion and involves at least two separate channels: a relatively specific Ca channel independent of membrane potential, which is probably uti-

CELL SURFACE SHEDDING

85

lized in certain hormone-receptor interactions not involving depolarization, and a second channel, which is a potential-dependent Ca permeability channel and which is presumably utilized when certain cells (particularly those activated by acetylcholine) undergo depolarization of the plasma membrane. Interaction with a specific receptor in the former case or depolarization in the latter case, results in an increase in permeability of the Ca channels with a subsequent rise in intracellular Ca concentration (Rasmussen, 1977; Rasmussen and Goodman, 1977). Efflux involves active transport, i.e., the specific Ca-activated ATPase or Ca pump (Borle, 1973). Another energy-dependent mechanism regulating cellular Ca metabolism at the cell membrane involves a Na+-Ca2+ exchange (extracellular Na' exchanges with intracellular Ca2+)and depends on the Na,K-activated ATPase or Na pump (Rasmussen and Goodman, 1977). Ca also has a compacting and hardening effect on the cell membrane, presumably by several types of interaction (Manery, 1966): ( I ) it may form salt linkages with dissociated anionic groups of phospholipids, sialic acid, proteins, and anionic glycosaminoglycans, and these bridges may play a role in the adhesion of the glycocalyx (or portions thereof) to the cell membrane (Deman et a / . , 1971; Bulkin and Hauser, 1973); ( 2 ) it plays a role in packing of phospholipids, limiting their motational freedom (Sauerheber and Lordon, 1975); (3) it probably forms stable chelates with ATP, a characteristic of the alkaline earth metals (Manery. 1966); and (4) it may promote cross-linking between membrane proteins or between membrane proteins and peripheral cytoskeletal proteins by activation of transglutaminase, a Ca-dependent enzyme that cross-links proteins (An1977). Ca also is an important determinant of both the state derson et d., of polymerization of cytoskeletal elements and their association with the plasma membrane (Poste et a / . , 1975). Thus, Ca, and possibly ATP as well, are important factors in maintaining the structural rigidity and permeability of the plasma membrane.

Ill. Synthesis and lntracellular Translocation of Released Proteins

A. SECRETION Relatively little is known concerning the synthesis of integral plasma membrane (g1yco)protein.sand its regulation, in comparison to secretory proteins. Much current evidence, however, indicates that plasma membrane proteins utilize the same pathway as secretory proteins and that similar events in their processing are operative. Therefore, the process

86

PAUL H. BLACK

of synthesis and release of secretory proteins will be considered in some detail. The guinea pig pancreatic exocrine cell has been studied most extensively (see Palade, 1975, and Jamieson and Palade, 1977, for reviews). The secretory apparatus of this cell consists of a series of compartments segregated from the rest of the cell by a membrane that is generally impermeable. During secretion, the products are transferred directly from one compartment to the next; the flow is vectorial, essentially irreversible (i.e., unidirectional), and is characterized by highly specific membranemembrane interactions (see Fig. 1). Concomitant with transfer of soluble contents, there is also a flow of membrane from one compartment to another. The process of secretion has been divided into six stages: synthesis, segregation, intracellular transport, concentration, intracellular storage, and discharge (Palade, 1975).

FIG.1. Diagrammatic representation of typical exocrine cell indicating intracellular compartments utilized in the processing of released proteins: RER, rough-surfaced endoplasmic reticulum; TR, transitional elements of the RER; GV, Golgi transporting vesicles; GC, Golgi cisternae; CV, condensing vacuoles; ZG, zymogen granules. The numbers indicate steps of the secretory process described in the text. (Adapted from Fig. I , Jamieson and Palade, 1977.)

CELL SURFACE SHEDDING

87

Proteins destined for export are synthesized on polysomes attached to the rough endoplasmic reticulum (RER); they pass through this membrane and appear in the cisternal space of the RER. There has been much interest in the mode of attachment of the protein synthesizing complex to the RER, and a model has been proposed that suggests a mechanism for this association (see Blobel, 1977, for review). This model is called the “signal hypothesis” and is supported by much-experimental evidence. Blobel and associates postulate that all mRNA’s for secretory proteins contain information (a sequence of codons referred to as “signal codons”) localized on the 3’ side of the AUG initiation codon that is translated into a unique amino-terminal sequence (the “signal sequence” or peptide, which is composed of 15-30 hydrophobic amino acids) of the nascent polypeptide chain (Blobel and Sabatini, 1971; Blobel and Dobberstein, 1975). When the signal sequence of the nascent polypeptide chain emerges from within the large ribosomal subunit, it is thought to initiate attachment of protein, ribosome, and mRNA to the membrane of the RER, possibly by a hydrophobic interaction between the signal sequence and the RER membrane. In addition, attachment may occur by the association of several ribosome receptor proteins with receptor proteins in the RER in such a way that a tunnel in the large ribosomal subunit is linked directly to a tunnel in the membrane (Sabatini and Blobel, 1970). It had been known that ribosomes bind to the membrane through the larger of the two ribosomal subunits (Sabatini et al., 1966). The growing peptide is then vectorially discharged through the tunnel in the RER into the intraluminal compartment of the RER, so that synthesis and secretion form a continuous process. The prefix (i.e., presecretory protein) is then removed posttranslationally by a “signal peptidase,“ which is associated with the RER membrane in the case of parathyroid hormone (Habener, 1977), prolactin, and growth hormones (Jackson and Blobel, 1977). From the cisternal space, the secretory proteins move, most likely by diffusion, to the transitional elements of the RER (Fig. 1). Part of the transitional elements (toward the cis side of the Golgi elements) are smooth surfaced, and small vesicles bud off, presumably by a fusionfission mechanism, to become the Golgi vesicles. These will then fuse with the condensing vacuole or with the dilated cisternae of the Golgi stacks. Vesicles that fuse with the condensing vacuole may also be derived from the dilated cisternae of the Golgi stacks. In the condensing vacuole, the secretory proteins are concentrated, with the result that the large condensing vacuoles are converted into mature secretion granules (zymogen granules). The secretion granules are stored in the apical region of the cell between the Golgi stacks and acinar lumen. The last step in

88

P A U L H . BLACK

secretion involves the fusion of the secretory granule with the plasma membrane.

B. EXOCYTOSIS A N D MEMBRANE FUSION Because the fusion reaction will be referred to repeatedly in this article, the process will be considered in some detail (see Poste, 1972, and Poste and Allison, 1973, for reviews). Although the precise mechanism of fusion is not completely understood, ultrastructural studies of various secretory cells have revealed several steps in the fusion process: approximation; aggregation of intramembranous particles (IMP, i.e., large integral membrane proteins); actual fusion of the membranes; and restoration of the lipid bilayer structure (Lawson et al., 1977) (see Fig. 2). Thus, large secretory vesicles become apposed to the plasma membrane, which A

FIG.2. (A) Schematic representation of interactions of the secretory granule membrane with the plasma membrane during exocytosis: (a) approximation of zymogen granule to plasma membrane at cell apex; (b) fusion of granule membrane with plasma membrane; (c) fission of fused area with release of soluble contents of vesicle. (B) Schematic representation of the interactions of a vesicle containing a transmembrane protein synthesized by the RER with the plasma membrane: (a) approximation; (b) fusion; (c) fission. Note reversal of polarity after fusion with plasma membrane (see text).

CELL SURFACE SHEDDING

89

bulges as it accomodates the curvature of the vesicles. The fusion areas of the plasma and granule membranes are almost devoid of IMP, which are aggregated around the bulges of the membranes. Once apposition has occurred, there is an elimination of layers as fusion progresses, until rupture (fission) of the membrane occurs with release of the vesicle contents to the extracellular space. This sequence of events has also been observed in virro during fusion of chromaffin granules (isolated from the hypophysis) to each other or to sheets of plasma membrane in the presence of Ca (Gratzl and Dohl, 1976; Schober et al., 1977). There is an almost universal requirement for Ca and energy during fusion. One of the probable actions of Ca is to bind to the anionic surfaces of the secretory vesicle and the plasma membrane and thereby neutralize the electronegative charge; with the resultant reduction of electrical repulsion between vesicle and the plasma membrane, approximation of the two membranes would be facilitated (Hall and Simon, 1976). A distance of 510 A is essential for fusion to occur (Poste, 1972). Since membrane fusion appears to involve interactions of membrane lipids, there must be movement of the IMP away from the area of fusion (Poste and Allison, 1973); this is evident from electron microscopic (EM) examination of spontaneous (Lawson et nl., 1977), chemical (Zakai et al., 1976, 1977), or viral-induced fusions (Okada et al., 1974; Fuchs et a / . , 1978; Loyter et a / . , 1977). Restricting the mobility of surface glycoproteins with doses of a lectin, which immobilizes receptor redistribution, prevents fusion (Toyama et d., 1978). The precise mechanism for IMP clustering is not known. Much evidence, however, indicates that reactions of Ca and/or protein kinases are involved. IMP aggregation in human RBC ghosts, as determined by freeze-fracture EM studies, can be induced by increasing intracellular Ca concentration (Hart et al., 1976; Elgsaeter et al., 1976; Vos ef al., 1976; Volsky and Loyter, 1977). Ca is known to induce clustering of acid phospholipids on the internal side of the plasma membrane (phase separations, see later) and this can lead to aggregation of IMP (Grant and McConnell, 1974). Ca may also induce the contraction of the spectrinactin complex in RBC, which can, in turn, result in aggregation of the IMP (Nicolson and Painter, 1973; Elgsaeter et nl., 1976). Microfilaments, which also contract with increased cytosol Ca concentration, may also be responsible for IMP aggregation (see later). It is known that cytochalasin B inhibits Sendai virus-induced cell fusion (Maeda et al., 1977) and the fusion occurring intracellularly after endocytosis in macrophages (Raz and Goldman, 1976). Some of the effects of Ca on contractile elements may be due to its effect in inducing various phosphorylation reactions, either directly by activating protein kinases or indirectly by its

90

PAUL H . BLACK

effects on cyclic nucleotide metabolism. Thus, phosphorylation of spectrin has been shown to induce the polymerization of actin in vitro, which results in a cross-linked complex of spectrin and actin (Pinder et al., 1977). There is evidence that phosphorylation of membrane proteins affects the fusion reaction, possibly by affecting IMP aggregation; however, a precise relationship between phosphorylation/dephosphorylation and aggregation/disaggregation is not yet clear (see Greengard, 1978, for review). A number of studies indicate enhanced secretion with phosphorylation. For example, phosphorylation of two platelet proteins and platelet granule release occurs when the cytosol Ca concentration is increased by exposure to an ionophore, collagen, or thrombin (Haslam and Lynham, 1977). Moreover, Ca-dependent phosphorylation of two synaptic vesicle proteins appears to be involved in the fusion reaction during neurotransmitter release (DeLorenzo and Freedman, 1977, 1978). It is not certain from these studies, however, whether Ca stimulates a protein kinase directly or acts by changing cyclic nucleotide levels (Schulman and Greengard, 1978). Dephosphorylation reactions also have been shown to affect IMP aggregation. For example, Sendai virus induce a specific and rapid dephosphorylation of human RBC ghosts concomitant with IMP aggregation (Loyter et al., 1977). The precise mechanism for such dephosphorylation is not clear. Phosphoprotein phosphatases may be involved; dephosphorylation may also occur by loss of intracellular and/or membrane ATP. A correlation has been observed between ATP loss, dephosphorylation of proteins, and IMP aggregation in human RBC ghosts (Gazitt et al., 1976). Thus, clustering of IMP occurs spontaneously in ATPdepleted RBC ghosts (Gazitt et al., 1976). Moreover, chemical-induced membrane fusion occurs only in ATP-depleted ghosts (Zakai e f al., 1977) and addition of ATP, presumably resulting in the pumping out of Ca, prevents fusion from occurring (Volsky and Loyter, 1977). It is of interest that similar events may be operative in transformed cells; Con A has been shown to induce redistribution and clustering of binding sites in cells with a low ATP content. Increasing cellular ATP concentration inhibited movement of Con A binding sites (Vlodovsky et al., 1973). With ATP depletion, rapid loss of 32Pi-labeledproteins from the RBC membranes occurs, indicating a rapid turnover of covalently bound phosphorus (Gazitt et al., 1976). ATP depletion can also result in loss of association of integral membrane proteins with cytoskeletal elements, which can lead to structural rearrangements of IMP. Thus, in aged RBC, which have low ATP, spontaneous vesiculation may occur, and the vesicles, which are deficient in IMP, contain no associated spectrin (Lutz et

CELL SURFACE SHEDDING

91

al., 1977). These studies suggest that phosphorylation-dephosphorylation reactions cause changes in the charge and conformation of proteins that may affect their interaction with other proteins or with neighboring lipids. Actual fusion occurs in areas of the apposed membranes composed mainly of protein-free, lipid bilayers. Most studies indicate that some disorder or perturbation of the lipid bilayer is necessary for fusion to occur. The introduction of a variety of compounds called fusogens (e.g., certain esters, fatty acids, alcohols, and vitamin A), which are generally lipid-soluble and which, when inserted into the membrane, exert a fluidizing or disordering effect, further indicates that some perturbation of lipid structure is important in fusion (Lucy, 1970; Kosower et al., 1975, 1977). There is evidence from nuclear magnetic resonance (NMR) studies that lipid phase changes occur during fusion induced in RBC ghost membranes by the fusogens, oleic acid and glycerol monooleate (Cullis and Hope, 1978). Thus, a variable portion of the phospholipids undergo a reorganization, i.e., phase transition from a bilayer to a globular phase (inverted micelle). This is strongly dependent on the presence of Ca and may be related to the well known capacity of Ca to cause lateral phase separations of charged lipid species in mixed bilayer systems (Cullis and Hope, 1978). As stated above, proteins are segregated away from such areas. Phase transitions from a lamellar (bimolecular leaflet) to a globular (micelle or inverted micelle) phase in the apposed membranes would promote disorder, and fusion would be expected to occur in these areas. The mechanism whereby Ca induces phase changes that promote fusion has been studied in mixed phospholipid bilayer systems in vitro (Breisblatt and Ohki, 1975, 1976; Duzgunes and Ohki, 1977; Papahadjopoulos et al., 1974, 1977). Ca has been shown to cause the fusion of certain phospholipid vesicles, especially those containing phosphatidyl serine (PS), which has a strong negatively charged polar group associated with the phosphorus atom and will bind to Ca ions; this causes segregation of individual lipids and formation of separate domains. By NMR techniques, it has been demonstrated that such an interaction causes a water exclusion effect, i.e., there is a displacement of water of hydration; this would make the complex hydrophobic (Hauser et al., 1975; Breisblatt and Ohki, 1976). The hydrophobicity results in a higher free energy state, which in turn would result in a high aggregation rate among PS. spherical membranes. In addition, the asymmetric distribution of Ca (Ca2+ concentration higher outside the vesicle) would cause packing and would decrease the area per molecule of PS; this also increases the free energy and the permeability to water of the highly curved vesicle configuration (Hall and Simon, 1976). Inverted micelles would form at the waterhydrophobic interface. These micelles are also unstable and, together

92

PAUL H . BLACK

with the high free energy state, would tend to favor fusion of the two apposed membranes. Stabilization would occur after fusion with the establishment of the bilayer structure. With fusion and exocytosis, the secretory vesicle membrane is incorporated into the plasma membrane and has been shown to be endocytized and reutilized (Herzog and Farquhar, 1977; Farquhar, 1978). Since PS is a component of t h e inner aspect of the plasma membrane (Zwaal, 1975) as well as the outer surface of the secretory vesicle, the in vitro studies utilizing membranes composed, in part, of this phospholipid are likely to be relevant to in vivo fusions. In this connection, it is of interest that PS is a selective enhancer of anaphylactic histamine release from mast cells induced by antigen, dextran, or Con A (Goth et af., 1971), or Sendai virus (Sugiyama, 1977). In addition to the potentiation by PS, these stimuli are dependent on Ca and temperature, suggesting that C a : PS interactions, which promote fusion in mixed bilayer systems, are also operative in the mast cell.

C. STIMULUS-SECRETION COUPLING With some membrane stimuli (e.g., cross-linking by a ligand) the permeability of the Ca channel(s) increases and cytosol [Ca"] rises. It is likely that some of this increment is due to the removal of Ca from the plasma membrane and possibly other sequestered sites. It was suggested by Woodin and Wieneke (1964) that ATP and Ca were removed from the plasma membrane of stimulated polymorphonuclear leukocytes (PMN's) and that such removal would produce decreased structural rigidity and cross-linking locally; these changes would render the membrane susceptible to fusion. It was further suggested that such removal was initiated by hydrolysis of membrane-bound ATP by an ATPase that was present in the leukocyte granule membrane and that had the properties of the Ca,Mg-activated enzyme (Poste, 1972). Moreover, high concentrations of ATP and Ca were found to inhibit the fusion of leukocyte granules with the plasma membrane and to act independently (Woodin and Wieneke, 1964); this has also been found in many other secretory systems (Poste, 1972). If ATPase is involved, inhibition of this enzyme should prevent fusion. Indeed, a variety of inhibitors of ATPase (e.g., high [Caz+])(Poste and Allison, 1973) prevent secretion, presumably by inhibiting fusion of the secretory granule with the plasma membrane. The inhibition of secretion by omission of Mg from the external medium might also be interpreted as inhibition of an ATPase wiith an obligatory Mg

CELL SURFACE SHEDDING

93

cofactor requirement (Poste and Allison, 1973). Displacement of Ca from membranes would therefore result in the Ca-dissociated state in which the membrane is less rigid, more deformable, more permeable to cations, and more fusion susceptible. In summary, Ca is the coupling factor between stimulus and secretion and is likely to be involved in each step of the fusion reaction. An increase in cytosol Ca concentration may result in approximation of vesicle to the plasma membrane and may enhance IMP aggregation directly by affecting a contractile system (which might move the IMP) and/ or possibly indirectly by its effect on phosphorylation reactions of membrane or cytoskeletal proteins. Aggregation of IMP may also occur by dephosphorylation reactions, possibly due to loss of membrane-associated ATP. Cytoplasmic Ca also appears to be essential in causing phase changes of the phospholipids, presumably by its interaction with PS, that result in lipid-lipid fusion. The events of exocytosis in stimulus-secretion coupling have been considered in some detail since it involves an activated cell. The similarity of the events occurring at the plasma membrane of these cells with those occurring in growing or mitogen-stimulated normal and neoplastic cells will be considered later.

D. ENERGYREQUIREMENTS A N D INTRACELLULAR TRANSLOCATION D U R I N G SECRETION The secretory process is dependent on energy; inhibition of ATP production leads to an abrupt arrest of intracellular transport (Palade, 1975). The most proximal site for energy requirement is localilized at the level of the transitional elements of the RER; in the absence of ATP generation, the secretory proteins remain in the RER. Only with resumption of ATP synthesis is transport to the condensing vacuole resumed. The precise nature of this energy-requiring reaction is not known, although an energygenerating system is known to be required for membrane fusion during exocytosis. Energy is also likely to be required ( I ) during membrane fusion and fission as Golgi vesicles are generated from the transitional elements, and (2) when Golgi vesicles fuse with condensing vacuoles or with Golgi cisternae. Energy-dependent processes may also be involved in the movement of secretory granules. There is much conflicting evidence about the relative roles of endoskeletal elements in intracellular transport, and both microtubules and microfilamentg have been implicated in this process. In certain cells,

94

PAUL H . BLACK

microtubules have been seen to be associated with secretory elements (Bed et al., 1973), and tubulin has been shown to associate with pituitary granule membranes in vitvo (Sherline et al., 1977). Much of the evidence for the involvement of microtubules in secretion is derived from the inhibitory effects of colchicine and the vinca alkaloids on secretion of certain products. Thus, exposure to colchicine or vinblastine has been found to inhibit release of the following: catecholamines (Schneider et al., 1977); parathyroid hormone (Chertow et al., 1975); amylase (Butcher and Goldman, 1972); procollagen (Ehrlich et al., 1974); insulin, in response to glucose stimulation; and mucin, in response to carbamylcholine stimulation (Zurier et al., 1973; Jamieson and Palade, 1971; Malaise-Lagge et al., 1971; Rossignol et al., 1972); prolactin; and growth hormone (Sherline et al., 1977). In other studies, however, colchicine stimulates release of various secretory products such as collagenase (Harris and Krane, 1971), plasminogen activator (Hull et al., 1977), or other neutral proteases (Gordon and Werb, 1976), glucagon (Edwards and Howell, 1973), and corticosteroids (Temple et al., 1972). The implication from the inhibition studies is that colchicine depolymerizes microtubules that are involved in secretion; such depolymerization has been documented in certain studies. However, one must interpret these results with caution, since colchicine has other effects that may play a role in inhibition of secretion. For example, colchicine may bind to membranes (Stradler and Franke, 1974), cause membrane expansion (Seeman et al., 1973), inhibit a number of transport systems (Wilson et al., 1974), and impair the mobility of cell surface receptors (Berlin, 1975); and colchicine or vinblastine may inhibit fusion (Palade, 1975). Recently, colchicine and lumicolchicine have been found to cause the aggregation of IMP in the plasma membrane of transformed cells (where IMP are randomly distributed) (Furcht and Scott, 1975). Since lumicolchicine does not cause the disruption of microtubules, its membrane effect is presumably independent of any effect on microtubules. All these studies indicate the colchicine can have a direct effect on the cell membrane and may affect the fusion-fission step in exocytosis. Nevertheless, it appears likely that microtubules, either alone or with microfilaments, may play some role in the intracellular movement of some secretory products. There is more evidence that the microfilament system is involved in intracellular movement of secretory products. Structurally, the microfilament system is associated with the junctional elements of the RER (Palade, 1975). In oxyntic cells, which have two interconvertible membrane systems in the secretory pole, actin is associated with both systems and is thought to be involved in the movement of vesicles from the

CELL SURFACE SHEDDING

95

precursor membrane pool to the plasma membrane upon stimulation (Vial and Garrido, 1976). Actin and myosin (Berl et al., 1973), as well as tropomyosin (Blitz and Fine, 1974), are present in synaptosomes. Frog retinal pigment granules are associated with and move along microfilaments upon stimulation with light (Murray and Dubin, 1974). Moreover, a microfilament cell web thought to be involved in insulin secretion has been described in pancreatic beta cells (VanObberghen et al., 1973). In contradistinction to microtubule disruptive drugs, microfilament disruptive drugs generally enhance secretion. Thus, exposure of cells t o cytochalasin B results in enhanced secretion of the following: parathyroid hormone (Chertow et al., 1975); insulin (VanObberghen et al., 1973; Orci et al., 1972; Lacy et al., 1973); and amylase (Butcher and Goldman, 1972). In addition, lysosomal enzyme release (Koza et al., 1975; Tou and Stjernholm, 1975; Zurier et al., 1973) and exocytosis in polymorphonuclear cells (Goldstein et al., 1977) are also enhanced by cytochalasin B treatment. Dosage is apparently an important factor, since low doses of cytochalasin B may act as a secretogogue in pancreatic acinar cells, whereas higher doses may prevent secretion (Bauduin et al., 1975). Also, the type of microfilament affected by cytochalasin B may be a factor, since the 50-70 A filaments are more reactive than the 100 A filaments (Bauduin et al., 1975). Cytochalasin B may also inhibit secretion in certain cases, for example, histamine release from mast cells mediated by antigen (Orr et al., 1972). Like colchicine , cytochalasin B has other cellular effects, particularly on the cell membrane; this may be a factor in its total effect. It binds to the plasma membrane and may change the physical properties of proteins (Carter, 1967; Spudich and Lin, 1972); this may alter membrane permeability (Estensen and Plagemann, 1972; Tou and Stjernholm, 1975). Also, at least three different transport systems (glucose, uridine, and thymidine) may be affected by cytochalasin B (Zurier et al., 1973). Although experiments utilizing microfilament disruptive agents must be interpreted with caution, the many structural studies demonstrating the association of contractile elements with secretory vesicle walls suggest that a contractile mechanism that regulates intracellular translocation of certain secretory products is operative and may also act to constrain translocation (Zurier et al., 1973; Tou and Stjernholm, 1975). Such constraint of movement is likely to be mediated by a subpopulation of microfilaments in a manner similar to the restraint imposed on cell surface receptors by an intact microfilament system. The effect of cytochalasin B might be to remove the restraints, resulting in exaggerated, unregulated intracellular movement of secretory products and leading to enhanced fusion and secretion.

96

PAUL H . BLACK

E. SYNTHESIS OF MEMBRANE (GLYCO)PROTEINS The secretory pathway is presumably involved in the intracellular translocation of peripheral membrane and glycocalyx proteins, but little is known concerning the mechanism. That integral membrane proteins might utilize the secretory pathway was first suggested by Palade (1969), and an increased amount of evidence that supports this hypothesis has accumulated. Such a model of biosynthesis would explain the asymmetric distribution of membrane proteins (See Rothman and Lenard, 1977, for review). According to this model, transmembrane proteins would be synthesized by the signal mechanism on the RER by membrane-bound ribosomes, like secretory proteins. However, a portion of the nascent chain (presumably hydrophobic) may become immobilized in the membrane during translation and may have the N-terminal end protruding into the cisternal space and the C-terminal end protruding into the cytosol (transmembrane proteins, see Fig. 3). Proteins predominantly localized on the luminal aspect may also be immobilized in this way; in this instance, the amino terminus would be in the cisternal space and the carboxyl terminus would be embedded in the lipid bilayer. Alternatively, luminal proteins may be secreted into the cisternal space by the signal mechanism and, with subsequent folding, could generate a hydrophobic domain that inserts into the lipid bilayer. Proteins localized to the cytoplasmic side of the membrane might be synthesized on free ribosomes; changes in conformation leading to a hydrophobic domain might then cause the insertion into the lipid bilayer. Thereafter, it is thought that the membrane with its inserted protein(s) may move along the secretory

CYTOPLASM

FIG.3. Hypothetical scheme of three principal modes of association of membrane protein with the lipid bilayer of the RER: (1) proteins predominantly located on the luminal aspect of the RER (ectoproteins); (2) transmembrane proteins; (3) proteins predominantly located to the cytoplasmic side of the RER (endoproteins). (Reprinted from Fig. 3, Blobel, 1977, with permission of the publisher and author.)

CELL SURFACE SHEDDING

97

pathway by fusion with other membranes in the Golgi complex and finally with the plasma membrane. Such movement has been termed "membrane flow," and this concept implies that the biogenesis of cellular endomembrane is facilitated by physical transfer of membrane material from one cell compartment to another (Franke el al., 1971; see Morre, 1977, and Cook, 1977, for reviews). Sugars are added to proteins as the latter move to various compartments; core sugars such as mannose and N-acetylglucosamine are added in the RER, whereas more terminal sugars such as galactose, fucose, and sialic acid are added in the Golgi complex or in the plasma membrane by a group of membrane-bound glycosyl transferases (Ronzio and Mohrlok, 1977). After fusion with the plasma membrane, a reversal of polarity occurs such that the portion of protein in the luminal aspect of the vesicle is then localized on the external surface of the membrane (ectoproteins), whereas proteins on the outside of the vesicle, exposed to the cytoplasm, are located on the inner aspect of the membrane (endoproteins, see Fig. 2). The mechanism for maintenance of polarity during the more proximal fusions is not understood. Although little is known about these fusions, similar mechanisms are presumably operative in different types of fusion (Poste, 1972; Poste and Allison, 1973). Thus, Ca induces aggregation of IMP in isolated Golgi membranes from rat liver (Gratzl et al., 1977) and also induces fusion of Golgi vesicles in v im (Judah and Quinn, 1978).

F. VESICULAR STOMATITIS VIRUS(VSV) SYNTHESIS An example of membrane biosynthesis that corroborates much of the aforementioned model has recently been delineated by utilizing the G glycoprotein of VSV that is present as a transmembrane protein on the plasma membrane of infected cells. The virus almost certainly utilizes host cell pathways for its synthesis, glycosylation, and processing since its genome encodes only five proteins: all structural proteins of the virus. These are the G glycoprotein and the M protein, which are both associated with the cell membrane during biogenesis, and three nonmembrane proteins, which are associated with the 40s viral RNA (Morrison and Lodish, 1975) (see Fig. 4). The envelope of VSV is composed of a trilaminar lipid membrane that contains external spikes made up of a single glycoprotein (G) and surrounds an RN A-containing nucleocapsid; the matrix protein M forms a bridge between the glycoprotein and the nucleocapsid. The G glycoprotein is synthesized on membrane-bound ribosomes that are bound to the RER membrane by nascent polypeptide chains (Morrison and Lodish, 1975). It is then processed and in 30 min-

98

PAUL H . BLACK

FIG.4. Schematic diagram illustrating the pathways of maturation of the major structural prbteins of VSV. (Note, of the 3 nonmembrane proteins associated with the viral RNA, only the N protein is depicted). The G glycoprotein is partially glycosylated in the RER ( G , )and becomes fully glycosylated in the smooth ER (G.J. (See text). (Modified from Fig. I , Katz et al., 1977.)

Utes arrives at the plasma membrane where it will be incorporated into mature virions. Details of the process of synthesis and insertion have been investigated by utilizing a cell-free system for synthesis of VSV G glycoprotein; translation of viral specific mRN A by wheat germ ribosomes occurs in the presence of RER vesicles from dog pancreas (Rothman and Lodish, 1977). The resulting polypeptide is incorporated into and spans the membrane asymmetrically with the amino terminus inside the RER vesicle, whereas the carboxyl terminus protrudes into the cytoplasmic space (Rothman and Lodish, 1977; Katz et al., 1977; Toneguzzo and Ghosh, 1978). Protein synthesis and insertion are tightly coupled; therefore, the protein can cross the membrane only during synthesis and not thereafter (Katz et al., 1977; Rothman and Lodish, 1977). Pulse chase experiments reveal the movement of the G glycoprotein from the rough to the smooth and then to the plasma membrane fractions (Knipe et al., 1977a,b). Glycosylation occurs in two stages: the protein is core glycosylated (N-acetylglucosamine and mannose) in the RER and sialylated in the smooth endoplasmic reticulum (SER) (Rothman and Lodish, 1977; Knipe et al., 1977b).

CELL SURFACE SHEDDING

99

The M protein is also tightly associated with the membrane and protrudes on the cytoplasmic side (Morrison and McQuain, 1978). It presumably acts as a bridge binding the G and nucleocapsid proteins. Virus assembly consists of aggregation of the G glycoprotein around the M protein and binding of the nucleocapsid in the cytoplasm to the M protein exposed to the cytosol. Budding then occurs or is concomitant with virus assembly. Shedding of the G glycoprotein also occurs and will be discussed later.

IV. Shedding and Membrane Protein Turnover

A. CHARACTERISTICS OF SHEDDING If a molecule is characterized as being shed and not secreted, it must be established that the molecule released was located on the cell surface. The major external labeling technique used to investigate shedding of cell surface molecules is iodination of proteins and glycoproteins catalyzed by lactoperoxidase (Phillips and Morrison, 1971); iodination is limited to the exposed proteins and glycoproteins on the cell surface. A second widely used technique is metabolic labeling, whereby precursor compounds such as glucosamine are used to label primarily cell surface glycoproteins. Most of the labeled, acid-precipitable, carbohydrate-containing molecules are accessible to protease cleavage and are presumably located on the external cell surface (Onodera and Sheinin, 1970); this further establishes that the shed material was synthesized and released from the cell, rather than being adsorbed from the serum present in the culture medium. Further, characterization of a molecule as shed precludes that cell surface materials are liberated as a consequence of cell death. In a number of studies, no decrease in cell number or cell viability during the course of the experiment was demonstrated (Doljanski and Kapeller, 1976). The absence of soluble cytosol proteins such as lactic dehydrogenase among the materials shed indicates that neither leakage from the cell nor cell autolysis occurred (Doljanski and Kapeller, 1976). The spontaneous release of isotopically labeled cell surface proteins and glycoproteins in vitro has been demonstrated with a number of cell types (Kornfeld and Ginsburg, 1966; Kraemer, 1967; Cone et al., 1971; Huang et af., 1974; Rahman er af., 1977). Such release is an active metabolic process requiring respiration, protein synthesis, and energy,

100

PAUL H. BLACK

and can be inhibited by low temperatures (Cone er a/., 1971; Doljanski and Kapeller, 1976). Release follows first order kinetics; this indicates that molecules are shed randomly (i.e., "old" and "newly" synthesized molecules mix and are released irrespective of their age). The amount of shed materials increases linearly with time; this indicates the coupling of synthesis with shedding and the existence of an intracellular pool (Cone er al., 1971; Siekevitz, 1972). Shedding has been found to be selective in that some labeled proteins are released, whereas others are not (Cone er al., 1971; Huang er a/., 1974). In studies to determine the characteristics of shedding, a comparison was made between glucosamine-labeled macromolecules spontaneously released from fibroblasts and those released by treatment with proteases such as trypsin (trypsinate). Separation by column chromatography and gel electrophoresis revealed remarkably similar profiles, indicating identity of both size and charge (Kapeller et al., 1973; Truding and Morell, 1977; Rahman et al., 1977). Studies of the kinetics of release and turnover rates of these macromolecules revealed that release was biphasic; the two populations of molecules, presumably representing different glycoproteins, had half-lives of approximately 1-24 hours and 2-3 days. These conclusions were based on the biphasic loss of radioactivity from the trypsinate. The labeled macromolecules appeared in the medium and similar kinetics were observed. The identity of the kinetics of release of glycoprotein both spontaneously and by trypsin treatment, the identity of size and charge properties by fractionation procedures, and the dependence of shedding on protein synthesis establish that shedding of at least a portion of membrane macromolecules is one of the normal physiological consequences of surface membrane protein turnover (Kapeller et al., 1973; Kraemer and Tobey, 1972). The mechanism of shedding is not known; current models will be discussed later. The identity of the trypsinate with the spontaneously shed molecules suggests that proteolysis may be involved; this is consistent with the temperature dependence of shedding and further implies that an active metabolic or enzymatic process is operative. However, shedding of different macromolecules may involve different mechanisms. Utilizing appropriate isotopes, shed cell surface molecules have been found to be a heterogeneous population of lipids, glycolipids, proteoglycans, proteins, and glycoproteins that may vary in size from individual molecules to large vesicles composed of several membrane components (Doljanski and Kapeller, 1976). Most studies, however, have been concerned with proteins and glycoproteins, and the plasma membrane turnover of these molecules will now be considered.

CELL SURFACE SHEDDING

101

B. PLASMA MEMBRANE PROTEIN TURNOVER The plasma membrane is a dynamic organelle in that there is continual synthesis and degradation of membrane (glyc~)proteins,~ as well as lipids. There is considerable controversy concerning the mechanism of plasma membrane protein turnover. Two models have been proposed: the heterogeneous turnover model (in which proteins turn over at different rates-) and the unit turnover model (in which proteins turn over at the same rate) (see Tweto and Doyle, 1977. for review). The studies of protein turnover in rat liver (endoplasmic reticulum and plasma membrane) by Schimke and associates provide the basis for the heterogeneous model of membrane turnover and reveal that: ( 1 ) there is continual synthesis and degradation of membrane proteins; (2) the degradation process follows first-order kinetics and is therefore random; and (3) the rates of turnover are heterogeneous, i.e., proteins, especially the glycoproteins, are degraded with different half-lives, with larger molecules being degraded faster (see Schimke, 1975, for review). These studies have led to a model that postulates that proteins leave the membrane individually and at different rates. As discussed above, shedding is a process by which individual proteins might be lost from the membrane. Another mechanism for loss of individual membrane proteins is by release into the cytoplasm (with subsequent degradation or reentry into the plasma membrane) (Schimke, 1975). Although it would be difficult, on thermodynamic grounds, to imagine the existence for any considerable length of time of integral membrane proteins in the cytosol (because of their hydrophobicity) (Tweto and Doyle, 1977), there is evidence that some type of molecular exchange of integral proteins may take place between the cytosol and both Golgi (Autuori et ul., 1975) and plasma membranes (Thompson et al., 1978). Other studies have also indicated the continuous and random turnover of plasma membrane proteins, but have not found such, if any, heterogeneity in the rates of protein degradation. Instead, such studies reveal that the half-lives of a large number of different plasma membrane proteins in thymocytes (Schmidt-Ullrich et al., 1974), fibroblasts (Roberts and Yuan, 1974, 1975; Hubbard and Cohn, 1975a,b), and hepatoma cells Several aspects of the turnover of membrane proteins and glycoproteins are similar. Statements applying to both will be designated by use of the term (glyco)proteins; when either a radioactive sugar or amino acid is employed to label glycoproteins or proteins, respectively, or when reference is made to more specific studies, the designation will be more specific. This general review, however, will be concerned with both glycoprotein and protein turnover, and there will be considerable intermingling of both in the discussion.

102

PAUL H. BLACK

(Tweto and Doyle, 1976) are very similar or the same. Such studies indicate that the plasma membrane must be degraded as a unit and thus, form the basis of the unit turnover model. The major mechanism of plasma membrane protein turnover in this model is internalization and degradation (Tweto and Doyle, 1977). There is much evidence showing that discrete areas of membrane are internalized and degraded and that much of the degraded material is reutilized as, for example, the secretory vesicle membrane in exocytozing cells. An extreme example of this is evident in studies of endocytozing cells where, during pinocytosis or phagocytosis, the equivalent of the entire plasma membrane may be internalized within 2 hours (Steinman et al., 1974). It is likely that internalization of membrane is a major mechanism of protein loss from the plasma membrane of cells other than secretory or pinocytosing cells, and, at least for the areas being internalized, would result in more homogeneous turnover rates (Tweto and Doyle, 1977). Whether individual proteins are internalized is not known; however, it seems likely in the case of capping and internalization of receptor sites by specific ligands (Edelman, 1976). Proteins are likely to be degraded by both mechanisms, i.e., heterogeneously and as a unit. Abundant evidence shows that the plasma membrane is not homogeneous throughout (Nicolson, 1976a) and that specialized areas exist. Therefore, it is possible that internalization of such distinct functional areas or "units" occurs at different rates. It is also clear that individual (g1yco)proteins have different turnover rates and are selectively released to the medium with different kinetics; this would contribute further to heterogeneity of protein turnover. Whether individual proteins are endocytozed or exchanged with cytosol proteins is not definitely known, although it is likely that both occur. The fact that cell surface proteins are catabolized by different mechanisms further supports the heterogeneity model of protein turnover. Moreover, the type of cell involved and the extent to which different cells internalize or shed their cell surface macromolecules, may also contribute to heterogeneity of protein turnover. In this regard, it is of interest that in rapidly pinocytozing cells, where membrane turnover occurs predominantly by internalization, 13- 17% of the original isotopic iodine label was found as acid precipitable counts in the culture fluid by 24-47 hours, suggesting that shedding had also occurred (Hubbard and Cohn, 1975b). Another factor that influences protein turnover in general is the physiological state of the cell. In nongrowing or nutritionally deprived eukaryotic cells from many species, there is increased protein breakdown (see Goldberg et al., 1974, for review; Hendil, 1977; Tanaka and Ichihara, 1977; Castor, 1977, Otsuka and Moskowitz, 1978). Such loss, together

CELL SURFACE SHEDDING

103

with the diminished rate of protein synthesis in these cells (Tanaka and Ichihara, 1978), accounts for the lack of protein accumulation (Bradley, 1977; Castor, 1977). In contrast, intracellular proteins are relatively stable during growth (Hendil, 1977). Moreover, the rate of protein synthesis in growing cells may be as much as three times that in resting cultures (Meedel and Levine, 1978). In general, an inverse correlation between the rate of protein degradation and the growth rate has been found in mouse, hamster, and human fibroblasts (Otsuka and Moskowitz, 1978). In contradistinction to general cellular protein turnover, there is no general agreement concerning the turnover of plasma membrane proteins during growth and quiescence of eukaryotic cells. Various turnover studies have indicated that proteins and/or glycoproteins of the plasma membrane are not being degraded during growth (Warren and Glick, 1968; Hughes et al., 1972), or are being degraded at the same rate as in nongrowing cells (Kaplan and Moskowitz, 1975a; Hubbard and Cohn, 1975b; Tweto and Doyle, 1976.). Various factors may be responsible for these results: (1) measurement of specific activity dilution may be too insensitive to detect turnover during growth; (2) increased reutilization of labeled precursors during growth may make for an apparently longer halflife of proteins in growing cells (Warren and Glick, 1978; Hughes et al., 1972); (3) utilization of pinocytozing cells may indicate little or no difference in turnover (Hubbard and Cohn, 1975b) since pinocytosis, which accounts for the major portion of protein loss, occurs to the same extent in growing and nongrowing cells and is thus independent of the growth state (Tweto and Doyle, 1977); and (4) comparisons of turnover rates during different growth states were made in cells of malignant origin Tweto and Doyle, 1976). Various factors may be responsible for these cells differ from that of normal cells (Hendil, 1977). In an extensive study of the turnover of plasma membrane (g1yco)proteins in growth-inhibited cells, specific activity dilution of labeled components in isolated membranes was examined; measures were taken to avoid reutilization of isotopes (Kaplan and Moskowitz, 1975a). These studies revealed that growing cells synthesized membrane proteins and carbohydrates at approximately four times the rate of nongrowing cells, whereas the rate of membrane protein degradation was approximately the same in both growth states. Such findings would be expected in view of the requirement that new cell membrane must be synthesized in preparation for or during cell division. These studies measured turnover of total labeled membrane proteins and individual components that were turning over more rapidly than the total rate might not be reflected, especially if they were a minor part. However, when individual (g1yco)proteins were examined, utilizing a double isotope technique,

104

PAUL H . BLACK

there was marked heterogeneity in the rates of membrane component turnover (Kaplan and Moskowitz, 1975b). This was most marked in rapidly growing cells. When exponentially growing cells became growth inhibited, less heterogeneity and decreased rates of turnover of membrane components occurred. The membrane components from exponentially growing cells exhibiting the highest rate of turnover were glycoproteins; this was thought to be due to their location on the exterior surface of the cell. Indeed, one membrane glycoprotein from growing cultures had a very rapid rate of turnover (half-life, 2 hours); this component could be found in the medium, suggesting that it was shed (Kaplan and Moskowitz, 1975b). Other studies have also indicated rapid rates of turnover of membrane glycoproteins (Baumann and Doyle, 1978). In summary, general cellular protein degradation in steady-state eukaryotic cells appears to occur to a greater extent in nongrowing cells. With respect to plasma membrane (glyco)proteins, recent studies indicate that there is an increased turnover in growing cells and that rapidly turning over (g1yco)proteins may be shed into the medium. These studies suggest that shedding of cell surface macromolecules occurs to a greater extent in rapidly dividing cells. Shedding and membrane protein turnover in general have similar characteristics: first-order kinetics of loss; dependence on time, temperature, and protein synthesis; heterogeneity of turnover with, possibly, the fastest turnover occurring with high-molecular-weight components. Shedding, however, if it occurs to a greater extent in growing cells, would be different from the turnover of internalized plasma membrane proteins (this appears to be independent of growth state) or turnover of intracellular proteins, which are relatively more stable in the growing than in the nongrowing state. V. Shedding and the Activated State

A. CELLACTIVATION Cell activation occurs during stimulus-secretion coupling and results in discharge of the soluble contents of secretory granules. Shedding occurs during another type of cell activation associated with various types of growing cells (normal, transformed, and mitogen-stimulated cells). The molecules released by such shedding will be considered after a brief discussion of some aspects of growth control. Normal cells in tissue culture grow exponentially until they become confluent, i.e., form a monolayer. Growth and movement then cease, and the cells are “density-” or “growth-inhibited”. They are arrested in the G I (or Go)phase of the cell cycle (see Pardee et ul., 1978 for review).

CELL SURFACE SHEDDING

105

Further growth of nonconfluent dividing cells can also be inhibited by deprivation of serum. Thus, two types of growth inhibition exist: ( I ) density-dependent and (2) serum-dependent. Resting or serum-deprived cells in the G I (or Go) phase can be stimulated to grow by reculturing, exposure to fresh serum, various proteases, or certain growth-promoting factors (also called mitogens) such as fibroblast growth factor (FGF), epidermal growth factor (EGF), or insulin (Gospodarowicz and Moran, 1976). The cells will then grow rapidly and undergo a series of metabolic events (called the positive pleiotypic response), which is characterized by increased rates of uptake of various nutrients into the cell, changes in the intracellular levels of cyclic nucleotides, and increases in the rates of RNA, DNA, and protein synthesis (Hershko e f al., 1971). This presumably reflects the sequential unfolding of an entire program of events from GI to cell division. Growing normal and rnitogen-stimulated cells undergo the positive pleiotypic response in a temporally regulated manner. Transformed or cancer cells have altered growth characteristics, which lead to a loss of density-dependent inhibition of growth and movement, and grow to high densities in vifro.They also are, in general, less sensitive than untransformed cells to the growth limitation imposed by lack of serum, and they generally do not undergo the changes characteristic of normal cells when deprived of serum; such normal changes include diminished RNA and protein synthesis, decreased uptake of certain nucleic acid precursors, and increased protein breakdown (i.e., the negative pleiotype response). Reexposure of previously deprived transformed cells to serum has little to no activating effect on the processes under pleiotypic control (Hershko e f al., 1971). Thus, the transformed cell resembles a cell under positive pleiotypic control and can also be considered to be an activated cell; however, it is relatively independent of the growth stimulants normally required. Thus, similarities exist between growing normal, mitogen-stimulated, and transformed cells.

B. SHEDDING FROM GROWING NORMALA N D CANCER CELLS Much of the interest in the past several years in studying molecules released from cells has arisen from the fact that such release from cancer cells occurs quite rapidly. Often these molecules (such as proteases) are said to be "secreted" when, in fact, there is evidence for their association with the plasma membrane: this implies shedding. In addition, studies comparing normal cells with cancer cells have often failed to control for the growth state of the former cells. In this section, the release of proteins from growing and nongrowing normal (or untransformed) and cancer

106

PAUL H . BLACK

(transformed), cells will be discussed. In addition, the possible role of shedding in causing the phenotypic changes associated with both the cycling normal cell and the cancer cell will be discussed. Two types of proteins will be examined: cell surface proteases (in particular, plasminogen activator) and a large-molecular-weight glycoprotein component of the cell coat, fibronectin.

I . Proteases Although it had been known since 1925 that cancer cells could lyse a fibrin clot (Fischer, 1925), the features of the fibrinolytic system in transformed cells have been delineated only recently by Reich and his group. Their early studies and other studies revealed that cells transformed by a variety of viruses and chemicals, as well as cells from spontaneously arising cancers, contained and released into the medium a serine protease, plasminogen activator (PA), which was not present in untransformed cells. This protease converts the serum zymogen plasminogen to plasmin, which hydrolyzes fibrin (Unkeless et al., 1973, 1974a; Ossowski et al., 1973, 1974; Quigley et al., 1974). Generally, assays for plasminogen activator in either fluids or cell lysates are carried out with lZ5I-labeled fibrin as the substrate in the presence of plasminogen; the extent of fibrinolysis is determined by the amount of released 1251-labeledfibrinopeptides (see Christman et al., 1977, for review). Early studies of the fibrinolytic activity of untransformed cells emphasized the low or background levels of PA when cells were assayed soon after plating (Unkeless et al., 1973; Ossowski et al., 1973). Subsequently, Chou et al. in our laboratory found that the fibrinolytic activity of untransformed mouse cells varies greatly during the cell growth curve (Chou et al., 1974b). An increase in cell-associated activity, with a considerable increase in released PA, occurs in actively growing untransformed mouse cells. However, the release of PA decreases when the cells become confluent, indicating a density-dependent control of release. In contrast, transformed cells, which continue to grow, continue to release PA (Chou et al., 1974b, 1977a). Cell lysate activities remain at approximately the same level (Chou et al., 1977a) or increase for a period of time (Loskutoff and Paul, 1978) in confluent, untransformed mouse cells. In subsequent studies of cells from a number of other species, cell-associated and released PA were also shown to increase as the cells become confluent. Evidence was presented that this was due to increased synthesis of PA (which, in turn, was dependent on RNA and protein synthesis and energy) rather than to changes in cellular inhibitor levels or to adsorption of released molecules to the cell (Rohrlich and Rifkin, 1977). These studies

CELL SURFACE SHEDDING

107

indicate that an increase in synthesis and release of PA occurs in growing normal cells and that release decreases upon cell confluence. These findings, together with the turnover studies presented, further indicate that shedding is associated with cell growth. The plasma membrane association of PA has been found in a number of cell types from different species. The increase in PA activity in cell homogenates, which occurs after addition of the detergent Triton X- 100 but not after treatments that would normally solubilize the contents of vesicles, suggests that the enzyme is not contained within a vesicle, but rather exists in a membrane-bound form (Unkeless et al:, 1974a; Quigley, 1976; Jaken and Black, 1979a). In addition, PA has been shown to be associated with plasma membrane-enriched fractions of virus-transformed cells from a number of species (Christman et al., 1975). Further purification of plasma membranes from chick, hamster, and mouse cells by differential and sucrose density gradient centrifugations reveal that approximately 80% of the cellular PA activity is contained in a plasma membrane-enriched fraction (Quigley, 1976; Jaken and Black, 1979a). Furthermore, PA copurifies with the plasma membrane marker enzymes, 5’-nucleotidase and Na,K-activated ATPase (Quigley , 1976). There is also evidence for the plasma membrane location of PA in human basophils (Dvorak et al., 1978) and pig kidney cells (Paul et af., 1978). The requirement for detergent treatment to solubilize the enzyme suggests that it may be an integral membrane protein, but this is not known for certain (Quigley, 1976; Jaken and Black, 1979a). All available evidence, however, indicates the plasma membrane association of PA; its release, therefore, is likely to be by shedding rather than by secretion. When the intracellular distribution of PA activity in growing and nongrowing mouse cells is compared, approximately 70% of PA activity is found to be associated with the plasma membrane-enriched fraction of the growing cells, whereas only 20-25% of PA activity is associated with a similar fraction in the nongrowing cells; in nongrowing cells, the remainder of the activity is associated with a heavier membrane fraction, presumably the RER (Jaken and Black, 1979a). This suggests that PA molecules move from heavy (RER) to lighter (plasma membrane-enriched) fractions during growth; this is analogous to the movement of the G glycoprotein of VSV. Such movement of PA is associated with its shedding, since there is a correlation between shedding of PA and its predominantly plasma membrane location (Jaken and Black, 1979a). Thus, the evidence at present suggests that (1) PA is generally membraneassociated in the cell; (2) in confluent cells, the majority of PA is associated with a heavy membrane (RER) fraction; and (3) in tumor and growing normal cells in which continued synthesis occurs, the major

108

P A U L H. BLACK

portion of PA activity is associated with plasma membrane-enriched fractions; these observations indicate movement of PA inactivated cells. Molecular weight determinations of partially purified PA from normal and transformed cells have revealed some differences between species (see Roblin et a / . , 1975b, for review). In general, there is a major component of approximately 50,000 daltons (d) and two minor components of approximately 75,000 d and 35,000 d; all have PA activity. Both the 50,000 d and 35,000 d, but not the 75,000 d, species of human PA crossreact with an antiserum prepared against human urokinase, suggesting that the lower-molecular-weight forms of PA are related (Roblin, 1978; Astedt and Holmberg, 1978; Vetterlein et a l . , 1979; but see Aoki, 1974). Whether the 75,000 d form represents a seperate gene product is not known. In quiescent mouse cells, the predominant form of PA is a 75,000 d species that is associated with a heavy membrane fraction (as determined by sucrose gradients). In growing and transformed mouse cells, however, nearly all the PA activity is present as the 50,000 d form that is associated with a plasma membrane-enriched fraction (Jaken and Black, 1979b). Following stimulation of quiescent cells with phorbol myristate acetate (PMA), calcium, or serum, there is an induction of the 50,000 d species, an increase in plasma membrane-associated activity, and release of PA (Jaken and Black, 1979b). This indicates that with activation and movement of PA to the plasma membrane, the 50,000 d species of PA is generated, but whether a precursor-product relationship exists between the 75,000 d and 50,000 d species of PA is not known. A relationship between cell surface-associated, non-plasminogen-dependent protease activity and cell growth rate has been reported; the labeled proteins, casein or hemoglobin, were used as substrates (Hatcher ef a / . , 1976). Although the number of proteases involved is not known, cells with the shortest doubling time contained the highest levels of surface protease activity, whereas nondividing cells (e.g., peripheral lymphocytes or cells with doubling times of greater than 3 days) had little to no surface protease activity. In other studies utilizing labeled casein as substrate, surface protease activity was found in growing normal smooth muscle, fibroblast, and endothelial cells. The surface activity was accompanied by released activity in some studies (Tokes and Sorgente, 1976) and disappeared when growth stopped (Hatcher et a / . , 1976, 1977; Tokes and Sorgente, 1976). In malignant cells (e.g., human melanoma or transformed mouse cells), proteolytic activity and the rate of cell proliferation remained high (Hatcher et al., 1977). Moreover, higher cell surface and released proteolytic activities were. observed after treatment of normal

CELL SURFACE SHEDDING

109

endothelial cells with a carcinogen or after spontaneous transformation (Tokes and Sorgente, 1976). From these studies, one may conclude that cell surface proteolytic activity (both plasminogen-dependent and independent) increases with growth and that proteolytic activity from growing cells may be shed into the medium; such activity is diminished or absent from nongrowing cells. Growing cells, then, resemble cancer cells in having increased surface proteolytic activity that may be shed. 2 . Proteases and the Mitotic Cycle

The fact that proteases appear on the plasma membrane in actively cycling cells suggests that such an association occurs at a particular phase@) of the cell cycle. Comparatively little information is available about cell surface proteolytic activity during the cell cycle. However, in synchronized cells, an increase in proteolytic activity (utilizing labeled acetylated casein or hemoglobin as substrates) has been found during or just before the M period in rabbit fibroblasts, mouse skin tumor, and human melanoma cells (Hatcher et ul., 1976). The surface levels decreased following the M phase. Recently, PA has been found to become associated with the cell surface dixring the G2-M periods of the cell cycle; it is also released at this time (Aggeler et al., 1978). Release of a serine protease, active at pH 7.8, has been observed from murine leukemia cells during the M period, but it is not clear that it was associated with the cell surface (Bosmann, 1974a). Thus, there is some evidence that proteases become associated with the cell surface during the G2-M phases of the cell cycle. 3 . Nature o f Proteases in Cancer Cells

There has been much speculation concerning a unique association between increased cellular PA activity and cancer. Generally, a good correlation has been found between elevated cellular and/or shed PA levels and malignant properties (e.g., invasiveness and/or metastases) in tumor and transformed cells of animals (Laug et al., 1975), and in various human cancers (Nagy et oil., 1977; Bigbee and Jensen, 1978); in some studies, the most malignant rat tumors were found to have the highest PA levels (Pollack et al., 1974). However, such a correlation does not always exist (Jones et d.,1975, 1976; Rifkin and Pollack, 1977; Wolf and Goldberg, 1978), and some transformed cell clones may have elevated cellular levels of PA but release little (Jones et al., 1975). Cells, cultivated

110

P A U L H . BLACK

in vitro or derived from tumors induced with these clones, actively release the enzyme, suggesting that release may more readily occur with progression of a tumor (Jones ef al., 1975). Numerous studies, however, indicate that proteases other than PA (for example, plasminogen-independent proteases active at either neutral or acid pH when assayed on substrates such as acetylated hemoglobin or casein) may be elevated in cancer cells (Bosmann, 1972; Schnebli, 1972) or may be present on the cancer cell surface. Such activity is absent from normal cells (Spataro et al., 1976). Moreover, plasminogen-independent fibrinolysin activity may be present in transformed cell cultures (Chen and Buchanan, 1975) and in human tumor cells (Nagy et al., 1977). Increases in PA, therefore, do not invariably occur with neoplastic transformation, and other proteolytic enzymes may be elevated intracellularly, may be present on the tumor cell surface, and may be shed into the medium in v i m . 4 . Fibronecrin An almost ubiquitous change that occurs in transformed or tumor cells is the loss of a large-molecular-weight glycoprotein that is a differentiated product of fibroblasts, myoblasts, and some epithelial cells and that is normally present on the peripheral portions of the cell membrane, i.e., the glycocalyx and/or in the intercellular matrix (see Hynes, 1974; Yamada and Olden, 1978; and Vaheri and Mosher, 1978, for reviews). It has been called by several different names: LETS (large, external, transformation sensitive) (Hynes, 1973, 1974); galactoprotein a (Gahmberg and Hakomori, 1973); SFA (surface fibroblast antigen) (Vaheri and Ruoslahti, 1974); Zeta protein (Wickus et al., 1974); and Li (Hogg, 1974). It will be referred to as fibronectin. It is present on the cell surface mainly as a dimer of MW 440,000 and consists of two equally sized polypeptide chains (MW 220,000) that are linked by a disulfide bond (see Fig. 5 ) ; however, larger aggregates do exist (Hynes and Destree, 1977). It comprises 1-3% of the cellular protein of normal cells and is the major species labeled when untransformed cells are iodinated by the lactoperoxidase technique (Yamada et al., 1977; Yamada et al., 1978). By irnmunofluorescent staining of confluent fibroblast cells in culture, it appears as a reticular, fibrillar network that is most concentrated at areas of cell-to-cell contact and on the ventral aspect of cells in contact with the substrate (Birdwell et al., 1978). It is also frequently associated with membrane processes and surface ridges. Transformed cells lack fibronectin or have markedly diminished levels present mostly at areas of cell-to-cell contact (Hynes, 1973). Metabolic labeling with radioactive amino acids also fails to label this protein on the cell surface of transformed cells, indicating its absence, rather than its masking. However, the protein is synthesized in transformed cells; the extent of synthesis varies in different species (Olden and Yamada,

CELL SURFACE SHEDDING

-ASN-

-2

111

CORE

TERMINAL SlALlC OR FUCOSE

FIG.5. Current model of fibronectin structure. Note elongated shape and location of intersubunit disulfide bonds and adjacent sites, which are highly sensitive to the proteases plasmin and trypsin. The oligosacchandes consist of a core of N-acetylglucosamine and mannose linked to asparagine residues with distal galactose residues that may be linked to terminal sialic acid or fucose residues. (Adapted from Fig. 2, Yamada and Olden, 1978.)

1977). In human cells, there is little to no inhibition of synthesis (Vaheri and Ruoslahti, 1975a). Most studies indicate a more rapid turnover of fibronectin and the shedding of a greater proportion of the synthesized fibronectin in transformed and tumor cells than in normal cells (Pearlstein and Waterfield, 1974; Vaheri and Ruoslahti, 1975a,b; Rieber ef al., 1975; Critchley ef al., 1976; Olden and Yamada, 1977). The protein can be recovered as the dimer in the supernatant fluids of cultures, indicating that it is shed from the cell surface more or less intact (Critchley et al., 1976; Baum et al., 1977; Olden and Yamada, 1977). The amount of fibronectin (determined by surface iodination) is diminished in growing cells (Hynes and Bye, 1974; Rieber et al., 1975). Immunoscanning EM and membrane immunofluorescence of synchronized cells reveal that fibronectin is present on cells in the G I , S, and early Gz periods of the cell cycle but not at the late G , or M periods (Hynes and Bye, 1974; Wartiovaara et al., 1974; Critchley et a l . , 1976; Stenman et al., 1977). During mitosis when cells round up, fibronectin is localized to the strands attaching the cells to the substrate (dish) or may be seen around the periphery of dividing cells, suggesting that it is shed from mitotic cells (Stenman et al., 1977). There is also a network of fibronectin that may appear on the cell substrate adjacent to cells or remaining after cells have migrated (Wartiovaara et a l . , 1974; Culp, 1976, 1977). Fibronectin binds to collagen (Engvall and Ruoslahti, 1977) and

112

PAUL H . BLACK

has been shown to play a role in cell adhesion; release of fibronectin or its absence may result, therefore, in diminished cell-to-substrate, cell-tointercellular matrix and, perhaps, cell-to-cell adhesion (see following). Although fibronectin has been the most intensively studied protein released from the cell surface, other cell surface (g1yco)proteins are also absent from the surface of various transformed cells; whether these are shed is not clear at present (Bussel and Robinson, 1973; Stone et al., 1974; Robbins et al., 1974; Wartiovaara et al., 1974). There is evidence that tnransformed cells retain a smaller proportion of their glycosaminoglycans and that these molecules are also shed (Roblin et al., 1975a; Voyles and Moskowitz, 1976; Dunham and Hynes, 1978). Shedding of glycosaminoglycans is also increased during growth (Klagsburn, 1976). Moreover, the cell surface glycosaminoglycan heparan sulfate is “desquamated” or shed from the cell during the G2-M period of the cell cycle, and it is possible that a fibronectin-heparan sulfate complex is shed (Kraemer and Tobey, 1972).5Both of these glycocalyx components are also associated with the substrate-attached material left behind when cells move on a dish or are removed from the dish with chelating agents (Culp and Black, 1972b; Roblin et al., 1975a; Culp, 1976; Dunham and Hynes, 1978). The interaction of fibronectin with the glycosaminoglycans and collagen would confer structural stability to the cell membrane and, perhaps, would account for the inability to patch or cap fibronectin of attached cells by antibody raised against purified cell surface fibronectin (Schlessinger et al., 1977).

5 . Shedding and the Mitotic Cycle Whether the increase in proteolytic activity on the cell surface during G,-M is the cause of shedding of fibronectin and glycosaminoglycans is not known; the relationship between proteolysis and shedding will be considered later. One might speculate on the relationship between shedding during the cell cycle and the phenotype of mitotic cells. The loss of fibronectin (and possibly heparan sulfate as well) is likely to be responsible for the rounding that occurs during the G2-M or premitotic phase of the cell cycle; rounding, in turn, may result in depolymerization of microfilaments (Willingham et al., 1977), which is a feature of mitotic cells (see earlier). The increase in lateral mobility of lectin receptors that also occurs at this period of the cell cycle (Smets and Deley, 1974) presumably results from these changes, and it is likely that the random Recent evidence suggests that fibronectin has a binding site for heparan sulfate and that cell-substrate adhesion may be mediated principally by a fibronectin-heparan sulfate complex (Culp er a / . , 1978 Rollins and Culp, 1979).

CELL SURFACE SHEDDING

113

orientation of the IMP during mitosis (Furcht and Scott, 19741, which are normally clustered at other stages of the cell cycle, may also be the result of membrane destabilization. Rapidly growing normal cells have more microvilli than stationary cells (Allen and Iype, 1976). When such cells are synchronized, the microvilli are observed during mitosis (Willingham, et a / . , 1977; Collard and Temmink, 1976). Cells in mitosis may also have other types of membrane projections such as cytoplasmic blebs and ruffles, further indicating loss of membrane stabilizing factors (Porter et a/., 1973a; Lundgren and Roos, 1976; Willingham et a / . , 1977). Indeed, exposure of certain normal confluent cells to trypsin, which removes certain glycocalyx components, results in the appearance of microvilli (Willinghamet al., 1977), indicating that structural changes in the plasma membrane may result from loss of glycocalyx and/or the events that follow. The studies of the mitotic cycle suggest that shedding is a factor regulating the phenotype of the mitotic cell and may, indeed, be the initial event causing the phenotypic change. The loss of cell surface material is likely to be responsible for the increased fusion potential of cells in the M phase of the cell cycle (Stradler and Adelberg, 1972). Rapidly growing cells, which are shedding, also have a high fusion rate. Moreover, treatment with trypsin of cells at any phase of the cell cycle confers on them the fusion potential of mitotic cells (Stadler and Adelberg, 1972), further suggesting that loss of cell surface material in the M phase favors cell-to-cell fusion. 6. Shedding and the Cancer Cell Phenotype

Transformed cells are characterized by increased synthesis and release of PA and/or other proteases, shedding of fibronectin and glycosaminoglycans, a poorly organized microfilament system, decreased adhesion, increased lateral mobility of certain cell surface receptors, an increased tendency to fuse, and increased transport of certain molecules such as glucose. In addition, numerous studies of transformed or tumor cells of epithelial (Allen and Iype, 1976; Karasaki et a / . , 1977) or fibroblastic (Vorbrodt and Koprowski, 1969; Willingham et al., 1977; Glaser et a / . , 1977) origin, especially by scanning electron microscopy, reveal the presence of a more active cell surface containing many microvilli, blebs (Glaser et a / . , 1977; Borek and Fenoglio, 1976; Karasaki et al., 1977), filopodia, and ruffles (Vorbrodt and Koprowski, 1969; Karasaki et al., 1977). Not all transformed cell lines, however, have increased microvilli (Porter el al., 1973b; Collard and Temmick, 1976). Nearly all of the changes described above in transformed cells also occur in normal mitotic cells and, in many respects, the phenotypes of

114

PAUL H. BLACK

mitotic and transformed cells are similar. One may, therefore, speculate that shedding of exoskeletal elements may be the main determinant of the transformed phenotype. Indeed, adding fibronectin to transformed cells results in enhanced adhesion and increased spreading, a reversion to a normal morphology, restoration of contact inhibition of movement, and polymerization of microfilaments (Yamada et al., 1977; Willingham, 1977). The similarities in phenotype between normal mitotic and cancer cells suggests that the expression of the cancer cell phenotype follows from a programmed series of events, i.e., the positive pleiotypic response that may be directed by cycles of DNA synthesis (Lundgren and Roos, 1976). It follows, therefore, that the expression of the cancer cell phenotype is an effect of some transformation event that results in dysregulated DNA synthesis. 7. Structures Shed from Cancer Cells

In addition to individual molecules, certain cancer cells contain and shed various surface projections. Thus, in certain cases of human reticulum cell sarcoma and malignant histiocytosis, various cell surface structures such as blebs, blisters, and long worm-like processes identified by scanning E M may be shed (Skinnider and Ghadially, 1977). Blebs (see Figs. 6 and 7) are surface protrusions containing cytoplasm and resemble the blebs present during mitosis. [The term zeiosis has been utilized to describe bleb formation from the cytoplasm in tissue culture cells at mitosis (Belkin and Hardy, 1961).] Blisters (see Fig. 8) contain clear fluid and appear to develop spontaneously from a single membrane-bound vacuole that acquires a second membrane as it is discharged from the cell surface. It is interesting that depolymerization of the microfilaments by cytochalasin B can induce blebbing. Formation and release of a large number of blisters from leukemic lymphoblasts can be induced by exposure to vinca alkaloids (Krishan and Frei, 1975); the same phenomenon could not be induced in normal fibroblasts growing in culture. These findings indicate that formation of such structures may be related to abnormalities of the cytoskeletal system. From cells of various human (Petitou ef a l . , 1978) and mouse leukemias (Calafat ef al., 1976; VanBlitterswijk et nl., 1977), vesicles having plasma membrane marker enzymes are shed and have been recovered from the blood and pleural fluid. It is of interest that the shed vesicles had a more rigid membrane when compared to the plasma membrane of leukemic cells; they contained a higher cholesterol :phospholipid ratio and had an increased microviscosity . A greater concentration of lectin receptors was also present on such vesicles compared to the cell membrane of leukemic cells, sug-

CELL SURFACE SHEDDING

115

FIG.6. Scanning electron micrograph of peripheral blood of a 60-year-old male with malignant histiocytosis (histiocytic medullary reticulosis) showing leukemic histiomonocytic leukemic cells. Note spherical and hemispherical surface blebs (arrow), and small wormlike process (arrowhead). ~ 7 3 5 0 (Reprinted . from Fig. 6, Skinnider and Ghadially, 1977, with permission of the publisher and author.)

gesting that the shedding was not of random portion of the cell membrane and may have arisen from a more rigid part. Such vesicles could not be identified in normal sera or in sera of patients in remission (Petitou et a / . , 1978). Shedding of portions of the long microvillous processes of hairy cell leukemia cells may also occur (Burns et a / . , 1977). A number of different structures may be shed from mouse tumor cells. Certain mouse ascites tumor cells growing in suspension may shed all or nearly all of the cell coat or glycocalyx rapidly and spontaneously without loss of viability; the intact glycocalyx may be recovered from the ascites fluid (Rittenhouse et al., 1978). Other mouse tumors, such as embryonal carcinoma cells growing as aggregates in suspension, may shed round disc-like structures essentially composed of two large-molecular-weight glycoproteins, one of which has the properties of fibronectin (Chung et al., 1977). Shedding of microvilli, which cover the surface of mouse mastocytoma cells and which contain H-2 antigens, has been documented (Koch and Smith, 1978). H-2 antigens have been shown by immunological means to be concentrated on microvilli (dePetris, 1978).

116

PAUL H. BLACK

FIG. 7. Transmission electron micrograph of peripheral blood of a 91-year-old patient with malignant histiocytosis. Appearance suggests the pinching off and discharge of blebs. x30,800. (Reprinted from Fig. 13, Skinnider and Ghadially, 1977, with permission of the publisher and author.)

Although only a few reports that characterize structures shed from cancer cells are available, it is of interest that all of the aforementioned reports concern leukemic cells, or circulating neoplastic lymphoreticular cells, or cells growing in suspension. Thus, rounded or unanchored cells are apparently more prone to shed discrete portions of the cell membrane rather than individual proteins or glycoproteins. In this connection, it has been shown that tumor cells grown in the ascites form contain very labile surface coats and that they readily release large amounts of glycoproteins into the culture medium (Molnar et al., 1965; Dorval et al., 1976). The factors involved in such release will be considered later.

c. SHEDDING FROM MITOGEN-STIMULATED CELLS Exposure of cells to certain mitogens results in changes characteristic of the positive pleiotypic response, and eventually results in cell division.

CELL SURFACE SHEDDING

117

1. Tumor Promoters

Tumor promoters are agents that facilitate tumor growth after initiation with a carcinogen, according to the two-stage experimental carcinogenesis system described by Berenblum (see Berenblum, 1975, for review). One of the most effective tumor promoters is croton oil, and the active principles have been identified as certain phorbol esters, among which phorbol myristate acetate (PMA) is the most active. Treatment with PMA at very low concentrations (lo-*- 10-lo M ) increases both cell-associated and shed PA in chick, human (Wigler and Weinstein, 1976), and rabbit cells (Loskutoff and Edgington, 1977). The increase in intracellular activity occurs as early as 3 hours after exposure and precedes the increase in PA release. The necessity for new gene expression is likely, since both actinomycin D (which inhibits transcription) and cycloheximide (which inhibits translation) prevent PA induction (Wigler and Weinstein, 1976;

FIG.8. Transmission electron micrograph of peripheral blood of patient described in legend of Fig. 7. Appearance suggests the pinching off of a double-membrane-bound blister and its impending detachment from a leukemic cell. X26,OOO. (Reprinted from Fig. 9, Skinnider and Ghadially, 1977, with permission of the publisher and author.)

118

PAUL H. BLACK

Loskutoff and Edington, 1977). The effect is transient and disappears after removal of the promoter. PMA also induces the production and shedding of fibronectin from chick embryo fibroblasts (Driedger and Blumberg, 1977). In general, the activity of phorbol esters as inducers of PA and fibronectin parallels their tumor-promoting potential (Viage et a/., 1977; Wigler et al., 1978). In addition to causing the positive pleiotypic response in normal cells, tumor promoters also induce other phenotypic changes characteristic of transformed cells (Weinstein and Troll, 1977; Fisher et al., 1978). Thus, changes in morphology, cell size, and growth characteristics, as well as increases in glucose transport, an elevation of cGMP, and a decrease in CAMP content (Belman and Troll, 1974), may all occur in treated cells (Driedger and Blumberg, 1977; Yamasaki et a/., 1977; Loskutoff and Edgington, 1977; Weinstein and Troll, 1977). However, these changes are transient and revert with removal of the promoter, whereas reversion of transformed cells occurs rarely and only under unusual circumstances (Culp et al., 1971; Culp and Black, 1972a). The mechanism by which tumor promoters produce positive pleiotypic effects is unknown, as is the mechanism of tumor promotion in general. There is evidence that tumor promoters may act by increasing the sensitivity of the cell to Ca (Boynton et a/., 1975). Low Ca induces a GI-S block, and DNA synthesis may be enhanced by overcoming this block. It is of interest that PMA is not mitogenic in the absence of Ca (Boynton et a/., 1975). However, little is known about the effects of tumor promoters on cell Ca metabolism. Promoters also induce ornithine decarboxylase in mouse skin; ornithine decarboxylase is a regulatory enzyme involved in the synthesis of the polyamine putrescine (Weinstein and Troll, 1977). The induction of polyamine biosynthetic enzymes may be the regulatory event stimulating DNA synthesis (Peterson et al., 1977). Tumor promoters may also prevent terminal differentiation of muscle cells (Cohen et al., 1977),fat cells (Diamond et al., 1977),neuroblastoma, and erythroleukemia (Yamasaki et al., 1977) cells. Promoters also affect the plasma membrane. PMA is lipid soluble and induces changes in ion movements (e.g., increasing the influx of 32Piand K) and increases Na,Kactivated ATPase and 5’-nucleotidase activities (Moroney et a / ., 1978). The increase in 32Piinflux would tend to raise the intracellular phosphate concentration; this may be an essential factor in cell activation (see later). Perhaps the phosphate influx is responsible for the degranulating effect of PMA on human leukocytes or its activating effect on platelets, including aggregation and the release reaction (White and Estensen, 1974; Estensen et a/., 1974). It is of interest that in such activated cells, cGMP accumulates (White and Estensen, 1974). Increases in Ca and/or cGMP

CELL SURFACE SHEDDING

119

occur in several types of activated cells and may be the second messenger(s) mediating the positive pleiotypic effects of PMA (see following). 2. Calcium

The mitogenic action of Ca has been established by a number of investigators (Whitney and Sutherland, 1972a; Dulbecco and Elkington, 1975; Boynton et al., 1975). Exposure of confluent mouse cells to supranormal (4.3 m M ) Ca results in marked increases in both cell-associated and released PA (Chou et al., 1977b; Chou and Black, 1979). The increase in cell-associated PA commences at approximately 4-8 hours and precedes the increase in released PA that appears at approximately 10-12 hours; the latter reaches levels 14 times those of control cells exposed to a normal Ca concentration of 1.8 mM. Like the stimulation by the phorbol esters, RNA and protein synthesis, and therefore, new gene expression, are required, as is continuous energy production. It is of interest that strontium (Sr), in a Ca-deficient medium, also stimulates PA production and release to approximately the same extent as Ca, but other divalent cations (Ba, Mn, Mg) do not (Chou and Black, 1979); Sr can substitute for Ca in certain secretory systems (Douglas, 1974). Although both phorbol esters and Ca are mitogenic, the induction and release of PA preceded DNA synthesis, which is therefore not necessary for these events. 3 . Growth Factors-Epidermal

Growth Factor

A number of growth factors, which are mitogenic, have been identified (Gospodarowicz and Moran, 1976). Recently, epidermal growth factor (EGF) at very low concentrations (10-8-10-10 M ) has been found to induce increases in both cell-associated and shed PA in HeLa cells. These increases were dependent on new synthesis, had a time course similar to PMA-induced increases, and were restricted to cells of only certain species (Lee and Weinstein, 1978). EGF also causes an increase in cell surface fibronectin in serum-depleted mouse cells, but no data indicating mitogenesis or shedding was given (Chen et al., 1977). It is of interest that EGF can act as a promoter of mouse skin carcinogenesis (Lee and Weinstein, 1978). 4. Thrombin

Thrombin is mitogenic for certain tissue culture cells (Blumberg and Robbins, 1975; Teng and Chen, 1975). Bovine thrombin stimulates both the production and release of fibronectin in cultures of human and chick fibroblasts. Both cell surface (iodine-labeled) and newly synthesized

120

P A U L H . BLACK

(metabolically labeled) fibronectin molecules were shed after 20 hours of exposure to thrombin (Mosher and Vaheri, 1978).

5 . Coupling of Synthesis and Shedding Although the nature of the coupling between protein synthesis and degradation (which must exist in order to explain the steady state concentration of specific proteins) is not understood, there is evidence indicating that the magnitude of the rates of protein turnover is governed solely by their rates of synthesis and not, as implied, by their rates of degradation (Siekevitz, 1972). The studies of shedding from growing normal and cancer cells and from mitogen-activated cells in which protein synthesis is actively occurring support the hypothesis that a coupling mechanism exists between protein synthesis and shedding; shedding apparently occurs in response to excess or continuous synthesis of PA. In general, an increase in intracellular concentration of PA precedes an increase in shedding. This is most apparent in growing cells and cells stimulated by mitogens. Conversely, inhibition of protein synthesis, either at the transcriptional or translational levels, prevents PA shedding. Moreover, inhibition of energy metabolism, which is required for protein synthesis, also prevents shedding. Thus, it is likely that increased synthesis of a protein such as PA must precede and is coupled t o its shedding, although the mechanism of linkage between specific protein synthesis and shedding is entirely unknown. From studies of Ca stimulation of PA production and release in mouse cells (in which optimal stimulation results in shed PA levels approximately 14-fold higher than nonstimulated cells) intracellular levels never exceed 2-3 times the control levels (Chou and Black, 1979); this is true also for PMA-stimulated mouse cells of the same strain (Jaken and Black, 1979b). It therefore appears that intracellular pool size may be regulated within a narrow range and that enhanced synthesis of a plasma membrane protein such as PA, by continuous exposure to a mitogen, may result in continuous shedding. In mitogen-stimulated cells, therefore, as in growing cells, PA and probably other cell membrane proteins are continually synthesized and transported to the cell membrane for eventual insertion; this sequence presumably involves processing and modification of the vesicle or granule containing the membrane protein by the elements of the secretory system as described. It is of interest that exposure of confluent mouse cells [in which the majority of PA is associated with a heavy membrane fraction (presumably RER)], to PMA results in movement of a large proportion of the PA to a lighter plasma membrane-enriched fraction. This is accompanied by an induction of the 50,000 d species of PA, which is eventually

CELL SURFACE SHEDDING

121

shed (Jaken and Black, 1979b). Such movement of PA, therefore, together with the fusion reactions that are presumably involved, can be considered part of the positive pleiotypic response to PMA. 6 . Comparison of Shedding with Secretion

Shedding of a plasma membrane protein such as PA occurs from certain activated (i.e., growing and mitogen-stimulated) cells, whereas secretion by exocytosis occurs from cells activated by various stimuli. Both involve the secretory pathway. Cellular products destined for secretion are contained within secretory granules or vesicles where they are soluble, whereas molecules to be shed are more tightly membranebound, presumably since their synthesis. Both exocytosis and shedding [the latter in certain experimental systems (Chou et al., 1977b; Chou and Black, 1979)], require Ca, may utilize Sr but not Mg in place of Ca, and are inhibited by an excess of Ca or Mg (Douglas, 1974; Rubin, 1974). Both require continuous energy production. Mitogen-induced shedding of PA requires both RNA and protein synthesis and commences in hours (greater than four with some mitogens and up to 10-12 for Ca-induced stimulation), whereas exocytosis of secretory products does not require protein synthesis (Jamieson and Palade, 1971) and occurs soon after exposure to the stimulus (Douglas, 1974; Rubin, 1974; Palade, 1975; Chou and Black, 1979).

D. BIOCHEMICAL EVENTSOF CELLACTIVATION

The molecular events responsible for cell activation are not completely understood at present. Rasmussen has pointed out a nearly universal involvement of Ca; cell activation generally leads to a rise in the concentration of cytosol Ca (see Rasmussen, 1977; Rasmussen and Goodman, 1977; Mikkelsen, 1978, for reviews). Ca is the coupling factor or second messenger in a number of phenomena that involve cell stimulation, e.g., stimulus-secretion coupling in exocrine or endocrine glands (see earlier). Ca is also involved in the stimulation of all forms of muscle irrespective of whether the contraction of the muscle is initiated by hormonal or neural means, i.e., stimulus-contraction coupling. This is accompanied by another type of coupling that also involves Ca: stimulus-metabolism coupling. In this latter coupling, changes in Ca ion concentration, in addition to excitation-contraction coupling, concomitantly couples metabolism to excitation. Thus, the appropriate metabolic events (e.g., glycogenolysis) are activated together with the contractile events in order

122

PAUL H. BLACK

to supply metabolic energy as ATP to the working tissue (Rasmussen and Goodman, 1977). Whether Ca is involved in the activation involving growth or mitogen stimulation is speculative and will be considered later.

I . Interaction of Second Messengers The cyclic nucleotides, cGMP and CAMP, are also involved with Ca in the coupling reactions (mentioned in the preceding section) in several different ways: they may act to regulate cell Ca metabolism; they may act as comessengers, often in sequence with Ca; or they may act as cell messengers whose concentrations are regulated by changes in cell Ca concentration (Whitfield et al., 1976; Rasmussen, 1977). Thus, though the final coupling factor is Ca, its effect may be modulated by the actions of cyclic nucleotides or intracellular Ca transport systems. No one sequence of events appears to be responsible for all activation, and although the second messengers are nearly universal and interrelated in the activation of particular cell types by specific extracellular messengers, a number of particular variations exist. In activation of some cells, Ca and cGMP are stimulatory, whereas in others, Ca and cAMP are apparently involved. For example, stimulation of PMN leukocytes by zymosan-coated particles results in lysosomal enzyme release, which is accompanied by an increase in uptake of Ca and an increase in cGMP. The increase in cGMP is Ca dependent and presumably results from Ca stimulation of soluble guanyl cyclase. Similar relationships between Ca and cGMP occur in activated smooth muscle, the mast cell, and circulating basophil upon interaction of antigen with its IgE receptor, and during platelet activation by collagen, thrombin, ADP, etc. (Rasmussen, 1977). In all these systems, and possibly during growth as well (see following), a rise in cytosol Ca is accompanied by an increase in cGMP. Whether cGMP mobilizes additional Ca from intracellular sequestering sites is not known. Also, in these cell systems, cAMP or agents that raise intracellular cAMP levels block or prevent the effect. In other systems (for example, catecholamine stimulation of the rat salivary gland and insulin secretion from the pancreatic beta cell), stimulation results in an increase in both cAMP and Ca in the cytoplasm (Rasmussen, 1977). In this instance, both act as sequential messengers in the control of the cellular secretory response, and each regulates the concentration of the other; an increase in cAMP causes an increase in Ca probably by mobilizing it from intracellular sequestering sites, whereas an increase in Ca then causes a decrease in CAMP, presumably by inhibiting adenyl cyclase and possibly by activating phosphodiesterase (Rasmussen and Goodman, 1977).

CELL SURFACE SHEDDING

123

2. Ca, cGMP, and Growth Whether Ca is involved in growth and/or the coupling between a mitogenic stimulus and DNA synthesis is not known. Nevertheless, a number of studies indicate that increases in cytoplasmic Ca and/or cyclic nucleotides are associated with growth. In several recent studies, evidence was presented that indicates that decreased levels of Ca are present in the plasma membranes of growing normal and transformed cells with a concomitant elevation of cytosol Ca. Mouse fibroblasts labeled with 45Ca were exposed to the Ca chelating compound EGTA and the removable, labeled Ca was considered cell surface Ca. Low levels were present during growth, but cell surface Ca increased progressively with increasing cell density. Quiescent or nongrowing normal cells had approximately seven times the level of cell surface Ca of growing cells. Transformed cells had levels similar to growing cells (Tupper and Zorgniotti, 1977). In subsequent studies utilizing 45Ca exchange, two cell compartments were identified: one was a rapidly exchanging (cell surface) compartment and the other was a more slowly exchanging (cytosol) compartment. Serum stimulation of quiescent cells resulted in a rapid loss of cell surface Ca, an 8-fold increase in Ca influx, and a decrease in Ca efflux (Tupper et al., 1978). Other studies indicate that the Ca concentration in tissue culture medium is decreased during growth, whereas the concentration remains unchanged in the medium of quiescent cells (Dulbecco and Elkington, 1975). Moreover, a diminution of medium Ca occurs in response to cell stimulation with serum and growth factors (Dulbecco and Elkington, 1975). There is, therefore, some evidence that in growing, transformed, and mitogenstimulated fibroblasts, increases in Ca influx occur that, together with decreases in cell surface Ca, result in an elevation of cytosol [Ca]. Loss of plasma membrane Ca, i.e., the Ca dissociated state (Poste and Allison, 1973), would enhance membrane fusion required for certain events of the positive pleiotypic response discussed earlier. Elevation of cytosol Ca also occurs in mitogen-activated lymphocytes; this, in turn, causes an elevation of cGMP (Whitfield et al., 1974, 1976). An increased uptake of 45Caoccurs within minutes after exposure to a mitogen (Whitney and Sutherland, 1972b); this was found to be due to an alteration in the binding affinity of the Ca carrier rather than to an increased number of carrier molecules (Whitney and Sutherland, 1973). The molecular events whereby Ca and/or the cyclic nucleotides may act to stimulate growth and in which cyclic nucleotides are actually involved are poorly understood.

124

P A U L H. BLACK

There is, however, an impressive body of evidence indicating that cGMP may act as a positive signal for the pleiotypic or mitogenic events, whereas cAMP may act as a negative signal. Transformed cells or rapidly growing tumor cells have elevated levels of cGMP (Goldberg et al., 1974; Shoji et al., 1977) and generally lower levels of cAMP (Rudland et al., 1974). Moreover, growing normal fibroblast cells have elevated levels of cGMP and decreased cAMP levels, resulting in a lower ratio of CAMP/ cGMP than confluent normal cells (Rudland et al., 1974); similar events occur with growing and nongrowing lymphocytes (Watson, 1975). The ratio generally increases as normal cells become quiescent; however, the ratio does not change during different growth states of transformed cells (Rudland et al., 1974). Moreover, addition of a high-molecular-weight glycoprotein, which restores contact inhibition of growth to malignant hamster melanocytes, is followed by an increase in the cAMP/cGMP ratio (Knecht and Lipkin, 1977). Fetal liver and lung cells, like transformed cells, also have lower cAMP/cGMP ratios (Shoji et al., 1977). In situations where quiescent cells are stimulated with serum, cGMP increases within minutes (9- 1 1-fold); insulin and FGF (both mitogens) also increase cGMP while lowering cAMP (Seifert and Rudland, 1974a). Adlop4M ) to quiescent 3T3 cells induces DNA ministration of cGMP ( synthesis (Seifert and Rudland, 1974a). A similar mitogenic effect from cGMP occurs in lymphocytes; this effect can be prevented by agents that elevate cAMP (Seifert and Rudland, 1974a; Johnson er al., 1977). Utilizing synchronized cells, Seifert and Rudland have found that cGMP is elevated only when cells pass through the G, stage or when they are stimulated to grow (from the Go phase) (Seifert and Rudland, 1974b). Based on these data, they postulate that cGMP acts as a positive signal. Normal cells have decreased cAMP levels during the M phase (Burger er al., 1972); however, cAMP levels may rise or fall during the cell cycle (Rasmussen and Goodman, 1977). The chemical carcinogens, nitrosourea (Derubertis and Craven, 1977a) and 4-nitroquinoline 1-oxide (Derubertis and Craven, 1977b), which promote growth, both stimulate guanyl cyclase and cause an increase in cGMP that is independent of Ca. Stimulation of the soluble enzyme was the major effect, but some stimulation of particulate activity also occurred with nitrosourea. The tumor promoter PMA also has been reported to increase the levels of cGMP very quickly, again indicating an association of growth promoting substances with elevation of cGMP (Estensen er al., 1974). It is of interest that an aqueous extract of the balsam pear inhibited guanyl cyclase activity, lowered cGMP, and inhibited growth of undifferentiated rat prostatic carcinoma cells that had high levels of cGMP prior to treatment (Claflin et al., 1978).

CELL SURFACE SHEDDING

125

Other studies indicate that cAMP may exert negative control. Unique cytoplasmic phosphoproteins, phosphorylated mainly by CAMP-dependent protein kinases, are associated with cell growth arrest (Kletzien et al., 1977). It is of interest that a separate phosphoprotein (phosphorylated by a CAMP-independent protein kinase) was identified in growing cells, and cAMP inhibited the phosphorylation of this protein. These studies suggest that cAMP may influence cell growth in a negative way and, further, that certain phosphoproteins may function as regulatory proteins to maintain the G I block (Kletzien, 1977). The cause(s) of the alterations in cyclic nucleotide levels that apparently occur in cancer cells is entirely unknown; presumably these fluctuations are a reflection of altered Ca and/or cyclase and/or phosphodiesterase activities (Mikkelsen, 1978). In conclusion, although little is known about growth regulation in normal cells and the dysregulation that occurs in cancer, at least preliminary evidence suggests that elevations of cytosol Ca and/or cGMP are associated with growth. The question arises as to whether these changes are related to shedding. VI. Shedding and Activated Specific Cells

Shedding may occur in a variety of cells as a consequence of cell activation. In this section, we will consider hormone-activated ovarian granulosa cells, fertilized eggs, certain embryonic cells, visual cells of the retina, cells of the immune system, and virus-infected cells. Only a few such studies in a limited number of species have been carried out.

A. OVULATION PA production and release are related to events surrounding rupture of the egg from the follicle at the time of ovulation. In studies of rat ovary cells in vitro, PA is detectable only in follicles destined to ovulate and increasing amounts of PA are found in granulosa cells as ovulation approaches (Beers et al., 1975). Inactive cells can be stimulated to produce PA by exposure to follicle stimulating hormone (FSH); less stimulation occurred with luteinizing hormone (LH) (Strickland and Beers, 1976). PA is synthesized de novo, since it is inhibited by inhibitors of RNA and protein, but not of DNA, synthesis. During the activation process, there is an increase in cellular PA, as well as shedding of PA into the follicle fluid. Time-course experiments indicate that release of PA peaked hours after exposure to hormone [FSH (6- 10) and LH (15-

126

PAUL H . BLACK

2 5 ) ] . Since plasminogen is present in the follicle fluid and plasmin has been shown to weaken follicle wall strips in vitro, it was postulated that PA is responsible for the disruption of the follicle wall at the time of rupture of the egg at ovulation (Strickland and Beers, 1976).

B. EGGS-FERTILIZATION Several studies, utilizing fish eggs, have revealed that activation occurs upon fertilization. For instance, fertilization of sea urchin eggs results in a marked elevation of intracellular calcium (Steinhardt et al., 1977; Ridgway et al., 1977). A rapid (within minutes) mobilization and discharge of cortical granules then occurs and results in elevation of the fertilization membrane (i.e., separation of the vitelline membrane from the egg surface) (Schuel et al., 1973, 1976). Removal or hydrolysis of the sperm receptor sites occurs upon elevation of the fertilization membrane, thus avoiding polyspermy (penetration of the egg by supernumery sperm). A trypsin-like serine protease has been localized to the cortical granules in unfertilized eggs. Following fertilization, a portion of the protease becomes relocated, a fraction of which is released into the sea water (Fodor et al., 1975). Whether another portion actually becomes associated with the surface of the egg and whether it is shed and/or secreted has not been determined; however, the protease is thought to be responsible for both elevation of and removal of sperm receptor sites from the fertilization membrane (Schuel et al., 1973). Certain serine protease inhibitors (soybean trypsin inhibitor, antipain, and leupeptin) inhibit these reactions, and polyspermy occurs (Schuel et al., 1976). It is of interest that a similar mechanism appears to be operative in mammalian eggs, where a trypsinlike protease is thought to change the zona pellucida that prevents polyspermy (Schuel et d., 1973). Other studies, in which the uninfected egg membrane is labeled with lZ5I, have revealed that 15-25% of surface labeled material (the majority being a glycoprotein of MW 150,000) is released from the surface after fertilization or parthenogenetic activation (Johnson and Epel, 1975). This release occurs with or shortly after cortical release. The released material (after dialysis and concentration) causes a decrease in protein synthesis to prefertilization levels when added to an activated egg preparation that is actively synthesizing protein. It is, therefore, thought to be responsible for the low metabolic state of unfertilized eggs, and its removal apparently results in the release of suppression of the egg (Johnson and Eppel, 1975). Thus, activation of eggs is associated with the shedding or modification of cell surface components. It is likely that these events are associated

CELL SURFACE SHEDDING

127

with the movement of cellular proteases to the egg surface, but this has not been conclusively proven. C. EMBRYOGENESIS

PA is produced by the mouse blastocyst. By microsurgical dissection of tissues, PA was found to be associated with certain types of cells and, furthermore, in certain cells, it was temporally related to various functions (Sherman et al., 1976; Strickland et al., 1976). Pure trophoblast cultures release PA during a limited period of the life cycle (6-10 days after fertilization); this time corresponds to the invasion of the uterine wall by trophoblast cells. A second cell type, the parietal endoderm cell, releases PA continuously. This cell type arises at the junction of the inner cell mass and the blastocoel cavity. These cells then multiply and migrate outside the yolk sac cavity and along the trophectoderm until they form a continuous layer apposed to the trophectoderm, where they produce and adhere to a thick basement membrane containing collagen (Reichert's membrane). As the embryo grows, with expansion of the trophoblast layer to accomodate the embryo, the parietal endodermal cells also proliferate. During these events of embryogenesis, autolysis and/or modification of tissues must be induced by both the invading trophoblastic cells and the migrating parietal endodermal cell with its need to separate from Reichert's membrane. Thus, PA may be responsible for certain interactions between a cell and the surrounding matrix turnover (Strickland and Beers, 1976; Strickland et al., 1976). It is also apparent that in certain embryonic cells such as the trophoblast, cell surface or shed PA activity is controlled temporally, as in mitotic cells. D. PHOTORECEPTOR CELLS-RODS Visual cells of the vertebrate eye maintain their integrity for life by a continuous replacement of their cellular constituents. The photoreceptor cells, rods and cones, contain their light sensitive pigments in a stack of closed discs derived from and contained within the plasma membrane that makes up the outer segment of the cell (see Fig. 9) (Young, 1973; O'Day and Young, 1978). These discs are continuously synthesized and shed. Much more is known about shedding of discs from rod outer segments (ROS) than from cones and only shedding of ROS will be discussed. A rhythmic daily shedding of ROS occurs in most vertebrate species (O'Day and Young, 1978). During experimental studies in fish,

128

P A U L H . BLACK

FIG.9. Schematic diagram of a visual cell. The outer segment consists of a stack of discshaped membranes that contain visual pigment molecules. A connecting cilium joins the outer and inner segments. The ellipsoid portion of the inner segment is filled with mitochondria. The myoid portion is the region where most of the cell's protein, carbohydrate, and, possibly, phospholipid are produced. The Golgi complex is also located in the myoid. (Reprinted from Fig. 1, Young, 1973, with permission of the publisher and author.)

amphibians, birds, and mammals, the shedding of ROS occurs soon after function, and since rods are used during the dark, shedding occurs early after an exposure to light. Optimal ROS length is maintained, however, since new discs are synthesized (Young, 1973). Evidence for coupling between synthesis and shedding has been obtained. By utilizing [3H]leucine and autoradiographic techniques in amphibians, the rate of addition to the ROS base can be determined (Hollyfield et al., 1977). In cyclic light, the rate at which new discs are added is greater during the light portion of the cycle than during darkness. The fact that shedding of ROS discs occurs predominantly after light exposure (at which time disc

CELL SURFACE SHEDDING

129

addition begins) suggests that the two processes are coupled (Hollyfield et al., 1977; Besharse e f al., 1977). Moreover, in cyclic light, an increase in disc addition that occurs with elevation of temperature is accompanied by a proportionate increase i n ROS shedding (Hollyfield et al., 1977). The shed ROS are ingested by the pigmented epithelial cells or by amoeboid phagocytes, which degrade them (Young, 1973). It has been estimated that each pigmented epithelial cell engulfs and destroys approximately 2000-4000 ROS discs daily (Young, 1971). Failure to carry out this function is thought to be an important factor in the pathogenesis of certain retinal dystrophies such as retinitis pigmentosa and certain types of hereditary retinal dystrophies of the rat (Young, 1973). There is evidence that Ca and possibly cGMP are involved in the transduction process of vertebrate rod photoreceptors (Hagins and Yoshikami, 1974). Calcium is sequestered in the membrane discs of the ROS (Fishman et al., 1977) and is released to the cytoplasm subsequent to a light-induced conformational change of rhodopsin, which is also present in the ROS discs (Zucker and Nolte, 1978). The released Ca diffuses to the outer membrane and decreases the membrane conductance for Na+, thus generating an increase in membrane potential (Hagins and Yoshikami, 1974; Brown and Flaming, 1978). During darkness, a steady inward current of Na+ enters the plasma membrane covering the ROS and depolarizes the cell membrane potential; this dark current is reduced by light. The hyperpolarizing effect of light is thought to be the mechanism by which the photochemical apparatus in the outer segments communicates its state to the presynaptic membrane at the junctions between receptor cells and secondary retinal neurons. Light also simultaneously activates a ROS membrane enzyme cGMP phosphodiesterase; this leads to a fall in cGMP (Lolley et al., 1977; Wheeler and Bitensky, 1977). As Ca increases, it inhibits the activity of the phosphodiesterrase, leading to an increase in cGMP, which then stimulates the uptake of Ca by the disc (Rasmussen and Goodman, 1977). Thus, both Ca and cGMP are involved in photoreceptor activation; since shedding of ROS occurs subsequent to light stimulation, one might speculate that shedding is related to either or both of these changes, i.e., the loss of Ca from ROS membranes and changes in cGMP levels. Evidence that cGMP may be involved in ROS turnover has been derived from certain hereditary retinal dystrophies of mice and rats. Inherited defects in the species of cGMP phosphodiesterase have been found in certain strains of mice (C,H/HeJ) and result in disorganization of the ROS and degeneration of the photoreceptor cells in the retina during the second to third week of postnatal life (Farber and Lolley, 1976; Lolley and Farber, 1976a). Normally, phosphodiesterase activity

130

P A U L H . BLACK

is low at birth and increases approximately 6-fold as the photoreceptor cells differentiate and mature; in the diseased retina, the low level of activity persists. The cGMP concentrations are elevated approximately 3-fold. The inherited retinal dystrophy of rats has also been associated with an alteration of the kinetic characteristics of cGMP phosphodiesterase thought to be due to an inhibitor in the diseased retina (Lolley and Farber, 1976b). These findings indicate that elevated levels of cGMP occur secondarily to dysfunction of the ROS membrane phosphodiesterase and may result in disturbances in the synthesis and/or turnover of ROS. E. SHEDDING FROM CELLSOF

THE

IMMUNE SYSTEM

It is clear that cells of the immune system (e.g., T cells, B cells, and macrophages) interact with each other in the generation of an immune response. Such interactions may occur at the cell surface, and much recent evidence indicates that plasma membrane molecules, either cellassociated or soluble, mediate many of these interactions. The surface location of these regulatory molecules, their release by shedding, and the conditions stimulating their release and involving cell activation will be considered. 1. B Lymphocytes The resting B lymphocyte of a number of species has immunoglobulin (Ig) inserted in the plasma membrane, where it exists as an 8s IgM monomer (MW approximately 70,000) and serves as the receptor protein (see Vitetta and Uhr, 1975, and Schreiner et al., 1977, for reviews). Evidence that it is an integral membrane protein is based on the fact that detergent is required for both its solubilization from the membrane and the maintenance of its solubility in an aqueous environment (Melcher et al., 1975), but the precise nature of its membrane association is not known (Singer, 1974b). Synthesis of the 8s IgM monomer occurs in the RER, with subsequent transport through the Golgi apparatus and eventual insertion of a Golgi vesicle into the plasma membrane (Vitetta and Uhr, 1975). However, little is known about the release of 8s IgM from the cell. Shedding of the IgM monomer from mouse spleen cells in vitro has been demonstrated. The shed material is recovered quantitatively from the medium; the monomer is apparently bound noncovalently to a small amount of plasma membrane lipid (Vitetta and Uhr, 1972). These studies suggest a fairly

CELL SURFACE SHEDDING

131

rapid turnover. However, later studies revealed heterogeneity of the B cell population with a biphasic loss of IgM from the cell surface; a rapid loss occurs from a population of large, stimulated B cells in 1-3 hours, whereas a slower loss from small spleen cells occurs in 10-30 hours (Andersson et al., 1974; Melchers and Cone, 1975). Although this former population was not extensively characterized, stimulated B cells apparently have a high turnover of surface IgM. Upon release, the monomeric IgM, which is insoluble in an aqueous environment, becomes soluble in serum. Stimulation of both mouse and human B cells with anti-Ig antibody results in patching, capping, and subsequent endocytosis and removal of surface IgM and is followed by regeneration of surface IgM (Loor et al., 1972; Vitetta and Uhr, 1975). Moreover, prolonged stimulation of B cells by the continuous presence of anti-Ig serum results in continuous synthesis and expression of surface Ig in B cells. Indeed, the continuous expression of surface IgM is required for continuous stimulation (Weiner et al., 1978). Thus, synthesis of IgM monomer is apparently linked to insertion that is part of the overall pattern of membrane biogenesis; stimulation of B cells, however, results in the preferential increase in the rate of IgM synthesis over that of other proteins (Vitetta and Uhr, 1975). Stimulation of B cells with antigen results in marked changes in B cell physiology and leads to differentiation and cell division, a marked increase in endoplasmic reticulum, increased synthesis and secretion of Ig, a change in the class of Ig secreted, and a high turnover of surface IgM. The Ig secreted from stimulated B cells is 19s IgM or other Ig classes. The 19s IgM is not radioiodinated when cell surface proteins are labeled with 1251, indicating a different mechanism of release (presumably involving secretion) from that of surface IgM, which is shed (at least in vitro) (Unanue and Schreiner, 1977). Whether shedding of the 8s monomer occurs in vivo is not known.

2 . Immune Response: la Antigens Most of the knowledge of the factors involved in immune interactions has been generated utilizing inbred strains of mice. Within the major histocompatibility region (H-2) of the mouse, there is a group of genes that regulates the immune response (Ir region). A portion of this region (I region) is composed of genes that code for antigens distinguishable (by alloantisera) on the surface of certain cell types involved in immunological reactions. These antigens are designated Ia antigens and are primarily expressed on B cells, but may also be present on both T cells (especially if stimulated) and macrophages. There is an intimate relationship at the

132

PAUL H . BLACK

cell surface (or in supernatant fluid) between Ia antigens and the molecules that mediate Ir gene function (T-B cell cooperation, T cell suppression, stimulation in the mixed lymphocyte reaction, and graft vs host (GVH) reactions). In mice, the I region has been separated by serological means into five subregions (see Fig. lo) (Shreffler et al., 1976; van Rood et al., 1977). In several species, the Ia antigens are glycoproteins that exist in a complex containing two moieties of approximate molecular weights of 33,000 and 27,000 (Cullen et al., 1976). They are differentiation antigens in that they are present on both nonactivated and activated B cells and are generally lost upon differentiation to plasma cells (Halper et al., 1978), whereas they are acquired upon activation to T cells (Woody, 1977). 3 . T Cells: Helper and Suppressor Factors

In a number of species, augmentation or suppression of the functional capacity of T and/or B cells is mediated by helper and suppressor T cell subpopulations, respectively. Soluble helper (Feldman, 1972) and suppressor (Tada et al., 1975) factors, derived from T cells, may also mediate these effects. The factors studied in mice are Ir gene products that carry la determinants (Armerding et al., 1974; Munro and Taussig, 1975). They mediate T cell-B cell or T cell-T cell interactions with cell-bound acceptor sites that are also Ir gene products and that adsorb and interact with the factors (Taussig and Finch, 1977). The factors may be antigen specific (Taussig, 1974; Tada et al., 1975) or nonspecific (Armerding and Katz, 1974; Schimpl and Wecker, 1972). Helper factors are involved in the regulation of the proliferation and differentiation of antibody forming cells in response to T-dependent antigens (Armerding and Katz, 1974), in the proliferation of cytotoxic T cells (Plate, 1976), and perhaps in the continued growth in culture of normal and antigen-specific T cells (Gillis et al., 1978). T cell helper activity may derive from the capacity of helper factors to activate T or pre-T cells in the B population, since helper activity is frequently related to proliferative activity (Koopman et al., 1977; Hirano et al., 1977). Both helper factors and helper cells in mice are characterized by the presence of antigenic determinants coded for by the IA subregion of the H-2 complex (McDougal et al., 1977). Recently, Tada et al. have demonstrated

K

A

B

I J

E

c s

FIG.10. Schematic diagram of genes coding for histocompatibility complex of the mouse.

CELL SURFACE SHEDDING

133

the presence of two distinct populations of helper T cells on the basis of their differential adsorption to nylon wool columns; these two populations can act independently or synergistically in enhancing the secondary antibody response (Tada et al., 1978). One class is presumably coded for by determinants of the I, subregion, whereas the other has determinants coded for by the I J subregion. The latter subregion also codes for suppressor determinants (see later); thus, closely linked, but separate and perhaps complementary, cell surface structures are likely to be involved in the interactions of surface molecules (Tada et af., 1978). Suppressor cells produce factors that inhibit antibody responses (Tada et al., 1975; Taniguchi et al., 1976) and contact sensitivity reactions (Greene et al., 1977b),and enhance tumor growth by production of tumor specific factors that suppress the immune rejection of the tumor (Fujimot0 et al., 1976). Suppressor T cells and specific suppressor factors of mice have certain similar biological and immunochemical properties in common; both are characterized by the presence of determinants coded for by the I, subregion of the H-2 complex. The cell surface location of suppressor determinants has been corroborated recently by experiments carried out in tumor-bearing mice. Alloantisera directed against IJ determinants abolishes the suppressor activity present in the mice, with subsequent tumor destruction. The effect was prevented by absorption of the antisera with cells bearing I,J determinants (Greene et al., 1977a). Conversely, animals inoculated with a soluble suppressor T cell factor obtained from cultures of spleen cells from tumorous animals augmented tumor growth in both syngeneic and allogeneic animals (Treves et al., 1976). It is apparent, therefore, that helper and suppressor factors are derived from the cell surface and their release might more appropriately be termed shedding. In general, shedding occurs to a greater extent from T cells than from B cells (Jones, 1973) and, as in other cell types, is dependent upon or is augmented by cell activation. Spontaneous shedding is apparently due to lymphocyte activation and/or cell division. For example, a spontaneous, rapid, temperature-dependent release of T cell receptors for alloantigens occurs at regular (8 hour) intervals from mouse T cells in culture; this release was thought to be related to the division cycle of the cells (Ramseier, 1974, 1975). When stimulated with antigen, the T cells shed considerably greater quantities of receptors than nonstimulated cells (Ramsier, 1974, 1975). Shedding of Ia antigens in mice also occurs, since such antigens can be detected in normal mouse sera (Parish and McKenzie, 1977). However, stimulation with mitogens and antigens, both in vivo and in vitro, results in an enhanced shedding of Ia antigens (up to 125-fold); these antigens are thought to be derived from T cells and are

134

PAUL H . BLACK

associated with helper and/or suppressor factors (Parish and McKenzie, 1977). In general, release of T cell helper or suppressor factors requires activation of the T cell by mitogen, antigen, or a mixed lymphocyte reaction (McDougal et al., 1977; Gillis et al., 1978); it also may require the presence of macrophages (Herman et al., 1977). The evidence that T cells acquire Ia antigens upon activation (Woody, 1977) suggests that synthesis and/or membrane insertion of molecules bearing these determinants occurs with activation. 4. Nature of the Released Material

Although helper and suppressor factors are associated with la determinants [the antigenic nature of which is thought to reside in a carbohydrate moiety (Parish and McKenzie, 1977)], little is known about the precise nature of the released material. There is some evidence that membrane complexes containing gangliosides may mediate helper and/or suppressor activity in mice. Gangliosides are sialic acid-containing glycosphingolipids that, like most glycolipids, are selectively located on the cell membrane (see Fishman and Brady, 1976, for review). Thus, a lowmolecular-weight antigenic component (containing helper activity and thought to be a ganglioside) was dissociated from a large Ia antigencontaining lipoprotein complex present in the serum of antigen-stimulated mice (Parish and McKenzie, 1977). In other studies, introduction of a ganglioside (GM,) (isolated from the suppressor substance in suppressor medium and incorporated into cholesterol-lecithin liposomes) into B cells in spleen cell cultures resulted in a depression of the antibody response. The effect was specific for B cells and could be neutralized with antiGMI antibodies (Miller and Esselman, 1975). These studies suggest that helper and/or suppressor modulatory activity is associated with a cell surface product that is probably a membrane complex of lipid and protein containing the modulatory glycolipid (Esselman and Miller, 1977; Freimuth et al., 1978). 5 . Human la-Like Antigens

In humans, alloantigens, which are preferentially expressed on B lymphocytes, are linked to the major histocompatibility complex (HLA), and are involved in mixed lymphocyte stimulation and other reactions, may be the human counterpart of the mouse Ir genes; such antigens have been referred to as Ia or Ia-like antigens (Halper et al., 1978). The analogy is strengthened by the fact that a bimolecular complex on human B lymphoblastoid cell lines (detected with various heteroantisera) is com-

CELL SURFACE SHEDDING

135

posed of two moieties of molecular weights of 28,000 and 37,000; this is similar to the complex found in other species (Ikeman er al., 1978; Halper et al., 1978). At least two distinct genetic loci for human B cell Ia-like alloantigens have been described in genetic studies utilizing families in which one member is an HLA recombinant; one is associated with the HLA-D locus and a second is associated with the HLA-A locus (Mann er al., 1976). Recently, it has been demonstrated that helper factors, derived from supernatant fluids of stimulated mouse T cells, adsorb to human peripheral blood lymphocytes (Taussig and Finch, 1978). By using HLA-recombinant families, the existence of at least two acceptor loci, separable by crossing over, were demonstrated; it was postulated that these acceptor sites are products of the Ir genes of man and that the extent of adsorption may be the determinant of the extent of immune responsiveness (Taussig and Finch, 1978). Whether human helper and/or suppressor factors are similar in chemical composition to such factors of mice is not known. Recently, a human helper factor was partially purified and was found to be a protein of approximately 80,000 d (Rutenberg et al., 1979). 6. Release of Other Lymphocyte Cell Surface Molecules from Normal and Lymphoma Cells

Both human and mouse T cells release Con A, phytohemagglutinin (PHA), and sheep red blood cell (SRBC) receptors spontaneously in culture (Jones, 1973; Sarmay et al., 1978) or after stimulation (Jones, 1973; Owen and Fanger, 1975); stimulation of T cells may also result in the release of Fc receptors (Fridman et al., 1974). Human peripheral blood lymphocytes that are incubated may release Fc and C3B receptors, as well as BTmicroglobulin and histocompatibility antigens (Sarmay et al., 1978) [release of the latter two components may account for their presence in human sera and urine (Reisfeld et al., 1976)l. Shedding of receptors from tumor cells of the lymphoreticular system may be accelerated. Fc receptors for both IgG and IgM are shed spontaneously from both B and T human lymphoblastoid cell lines (Molenaar er al., 1977), and increased shedding of Fc and C3 receptors may occur from lymphocytes of patients with chronic lymphatic leukemia (Sarmay et al., 1978). 7. Macrophages

Although few data are available, there is some evidence (1) that certain "secreted" macrophage products that play a role in modulating immune

136

PAUL H . BLACK

cell function are actually cell surface molecules that are shed; (2) that release of such molecules occurs with activation; and (3) that certain cell surface molecules may act as growth factors for macrophages and/or other cells. 7a. The Activated Macrophage There is much controversy and confusion concerning the precise meaning and nature of the activated macrophage (see discussions by North, 1978; Karnovsky and Lazdins, 1978; Cohn, 1978). As originally introduced by Mackanass (1964), the term “activated” described the adaptive changes that occurred in macrophages that enabled them to express enhanced antimicrobial resistance. Such cells showed a pronounced ruffling of the plasma membrane, increased adhesion to and spreading on the substrate, an enhanced phagocytic capacity, and an increase in lysosomes and endocytotic vesicles. In addition, activated macrophages may release soluble products including neutral proteases such as plasminogen activator, a specific elastase, and a specific collagenase (Gordon, 1976; Werb, 1978). Macrophages may be activated metabolically by the presence of a tumor, or by certain lymphocyte-derived products (North, 1978). Karnovsky and Lazdins (1978) have suggested that activated macrophages be considered in three categories: macrophages “specifically activated” by immunologic means; macrophages nonspecifically “elicited” by a variety of agents inoculated into the peritoneal cavity (such as casein, peptone, or thioglycollate); or cultured peritoneal macrophages exposed in vitro to products of stimulated lymphocytes and referred to as “specifically conditioned in vitro.’ * 7b. Immunological Interactions There is considerable evidence that Ir genes exist in macrophages from a number of species and that their products, the Ia or Ia-like antigens, play an important role in macrophage-lymphocyte interactions. Most studies indicate that one of the first reactions in generating an immune response is the association of specific antigen with macrophage Ir gene products. The cell surface complex may then bind to sensitized T cells, inducing T cell proliferation and an immunologic reaction (such as delayed hypersensitivity) (Lipesky and Rosenthal, 1975; Lopez er al., 1977), or perhaps inducing the shedding of soluble helper or suppressor factors that, in turn, interact with other T and/or B cells. Sensitized T cells may recognize antigen only in association with specific Ir gene products, since in certain sensitization reactions carried out in vitro, the antigen specific

CELL SURFACE SHEDDING

137

proliferative response was inhibited by antibody directed against the macrophage Ia antigen used for the original immunization (Schwartz et a / . , 1976; Thomas et a l . , 1977). Thus, T cells may exhibit histocompatibility restriction with respect to the immunizing macrophage (Thomas et a / . , 1977). The macrophage Ir gene products may provide a receptor function that is important in orienting or focusing antigen for T lymphocyte recognition (Rosenthal et al., 1977). Macrophage Ir genes, associated with antigen, may also modulate antibody induction in B cells (Howie and Feldman, 1978); similar results, with respect to histocompatibility restriction, have been obtained in assays for antibody induction (Pierce et al., 1976). Macrophage function may be served by soluble factors. One such factor of approximately 55,000 d consists of macrophage Ia antigen complexed to a fragment of immunogen and could replace macrophages in induction of helper T cells (Erb et al., 1976). It is likely that such complexes are shed from the macrophage cell surface upon activation, but the precise relationship between activation and shedding of soluble macrophage immunologic mediators is not yet clear. Shedding of glucosamine-labeled membrane components from activated macrophages occurs (Schroit et a / . , 1973). Presumably, such components were derived from the cell surface, since activation of labeled macrophages that had been treated with antimacrophage serum led to shedding of labeled immune complexes (Schroit et m l . , 1973).

7c. Activation cind Release of P A The release of PA from macrophages is thought to be an index of activation (North, 1978) since it occurs with the three types of activation mentioned: immunological (Nogueira et al., 1977); elicited (Unkeless et af., 1974b; Werb, 1978; Vassalli and Reich, 1977); lympholine-induced (Nogueira et a/., 1977); or by various combinations (Klimetzek and Sorg, 1977). The release of PA has generally been referred to as "secretion," but several characteristics of the release suggest that the enzyme may be shed from the cell surface. 1 . With activation, a large increase in both cell-associated and released PA occurs with time; this indicates that new synthesis is required (Nogueira et al., 1977; Hamilton et al., 1976). The prolonged time course for release suggests that shedding rather than secretion is involved. Thus, maximal release may not occur until 2-3 days (Nogueira et al., 1977; Hamilton et al., 1976), and continuous production and release from thioglycolate-elicited macrophages may persist for 4 days (Unkeless et af.,

138

PAUL H. BLACK

1974b), or for as long as 9 days in endotoxin-elicited macrophages that have phagocytized latex particles (Gordon et al., 1974). 2. The release of PA was unaccompanied by the secretion of lysozyme and intracellular acid hydrolases in several studies, indicating a dissociation between PA release and lysosomal enzyme secretion (Gordon et al., 1974; Hamilton et al., 1976; Vassalli et al., 1976). 3. In macrophages elicted by intraperitoneal thioglycolate, endotoxin (Vassalli et al., 1976), or asbestos and exposed in vivo to [3H]thymidine during the 4-day interval between stimulation and harvest (Hamilton et al., 1976), nearly all of the macrophages had incorporated [3H]thymidine into their DNA, suggesting that this population consisted of young, recently arrived, and replicating macrophages. Such macrophages were actively releasing PA. In other studies, macrophages produced very high levels of PA, and this was closely related to their ability to proliferate in the presence of colony stimulating factor (Gordon, 1978) (see later). These studies suggest that, as in other cells, a relationship between growth and shedding may exist in macrophages. 4. Mitogens such as PMA and Con A induce both production and release of PA from macrophages (Vassalli et a l . , 1977). Although large amounts of PA were released, the intracellular levels were not increased by more than 2-fold, suggesting a rapid turnover of intracellular PA and a coupling of synthesis with release, as in other stimulated cells releasing PA. 5 . Intracellular PA is tightly membrane-bound. This is supported by the fact that it is solubilized only by detergents and not by treatments that would solubilize the contents of secretory granules or loosely associated membrane proteins (Solomon et al., in preparation). 6. Recent studies indicate that macrophages activated by periodate treatment or exposure to Con A had increased surface proteolytic activity; this was indicated by their ability to hydrolyze an 1251-labeledcasein substrate covalently coupled to latex beads that were in contact with the cells (Tokes et al., 1978). Some proteolytic activity also occurred with beads not in contact with the cells; this suggests the presence of released activity, but there is no conclusive evidence that this soluble activity represents the same protease(s) as that on the cell surface (Tokes et al., 1978). The studies suggest, however, that cell surface proteolytic activity appears with macrophage activation. Shedding from an activated macrophage would have to occur in a cell that is actively ingesting material, and endocytosis is the major mechanism of membrane protein turnover in phagocytic cells (Hubbard and Cohn, 1975b; Edelson and Erbs, 1978). Nevertheless, as discussed earlier, protein loss from the cell membrane may occur by both mechanisms

CELL SURFACE SHEDDING

139

in such cells. This is supported by the data of Hubbard and Cohn (1975b), which indicate that, although the major portion of iodine-labeled cell surface molecules is endocytozed, 13- 17% of acid-insoluble, labeled material appeared in the medium during 24-47 hours of incubation. 7d. Release of Growth Factors Growth factors such as macrophage growth factor (MGF; assayed by stimulation of the proliferation of peritoneal macrophages) (Bradley and Metcalf, 1966) and colony stimulating factor (CSF; assayed by stimulating the differentiation of individual progenitor cells from bone marrow, spleen, and blood to colonies of granulocytes and/or macrophages) (Pluznik and Sachs, 1%5), are present i n both mouse and human tissues. MGF is found on the cell surface of mouse fibroblasts during the S and early M periods of the cell cycle and is spontaneously shed and lost from the cell surface during the M phase of the cell cycle (Cifone and Defendi, 1974). Shed MGF, purified from mouse fibroblast supernatant media, is a glycoprotein with a molecular weight of approximately 70,000 and is composed of two disulfide-bonded 35,000 MW subunits (Stanley and Heard, 1977). It is identical to mouse CSF, and has similar physicochemical properties and is immunologically cross-reactive with the CSF from human urine (Stanley et al., 1976). CSF can also be released from mouse peritoneal macrophages upon stimulation by antigen or endotoxin. The release elicited by endotoxin occurs in 3 hours and is related to the concentration of endotoxin, suggesting that activation is necessary (Eaves and Bruce, 1974). Proliferation of macrophages is dependent on the presence of CSF (Lin and Stewart, 1974; Gordon, 1978; Stanley et al., 1978). Thus, a growth factor, associated with the surface of at least some cells, is released from activated macrophages, and in turn stimulates the proliferation of macrophages. Several other helper factors are released from macrophages. A mitogenic factor (MW 10-15,000) for T, as well as B, cells can also provide helper activity for B cells, and is found in the culture fluids of macrophages activated by a T cell factor (Unanue et a / . , 1976; Unanue and Kiely, 1977). A number of helper factors for T cells released from adherent human peripheral blood cells thought to be macrophages, were found to copurify and to have identical properties (Koopman et al., 1978; Gery et al., 1972). T cells also augment production of this helper factor(s). Other helper and/or growth factors elaborated from macrophages have been described (Salmon and Hamburger, 1978; Hoffman et al., 1979). One may raise the question as to whether certain of these helper factors and/or CSF are components of the macrophage cell surface and are shed

140

PAUL H . BLACK

in a manner similar to the shedding of certain lymphocyte-derived proliferative or helper factors.

F. SHEDDING FROM VIRUS-INFECTED CELLS The utilization of host cell pathways in the biosynthesis of the envelope (G) glycoprotein of the lytic VSV has been considered earlier. This also occurs with other lytic and nonlytic budding viruses. The replication of the latter viruses is dependent on an activated cell state and shedding of individual virus plasma membrane (g1yco)proteins may occur. Because these aspects of virus infection have many analogies with host plasma membrane (g1yco)protein synthesis and shedding, and since the mechanism(s) underlying release of cellular and viral (g1yco)proteins may be similar, certain factors regulating nonlytic virus replication and viral shedding will be considered. 1 . Oncoviruses: Structure and Replication

Oncoviruses are a large group of RNA viruses that are transmitted horizontally as infectious virus particles or may be derived from proviral genetic information integrated into the genome of normal host cells and transmitted vertically in a large number of species, including chickens, mice, cats, baboons, and possibly humans (see Hirsch and Black, 1974, for review). These viruses contain a nucleocapsid portion [composed of RNA and, generally, four structural (capsid) proteins] and an envelope (composed of one or more glycoproteins). Oncoviruses replicate in either normal or transformed cells without causing cell death. Oncoviruses synthesize their structural proteins as large distinct precursor polypeptides, which subsequently undergo proteolytic cleavage and processing to generate the virion structural proteins (see Shapiro and August, 1976, for review). Although the structure and replication of oncoviruses from different species are generally similar, several differences do exist; this discussion will deal with mouse and avian oncoviruses. The precursor of the envelope glycoprotein of murine oncoviruses is made on membrane-bound ribosomes in the RER and then moves to the SER (Witte et al., 1977) in a manner similar to the VSV G glycoprotein (Morrison and Lodish, 1975). It is probably not cleaved until it reaches the plasma membrane. As it is cleaved, it becomes exposed at the outer surface; the inner aspect also becomes available for interaction with the capsid precursor protein (Witte et al., 1977).

CELL SURFACE SHEDDING

141

The capsid precursor protein of both avian and murine oncoviruses is probably synthesized on soluble ribosomes (Diggelmann et al., 1976) and is found in the cytosol and in the plasma membrane (Hayman, 1978). It is organized on mature viral buds containing viral nucleoprotein prior to cleavage (Witte and Baltimore, 1978). Cleavage may be associated with the budding process (Hayman, 1978), or the capsid precursor may appear in the mature viral bud (Hunter et al., 1976) with cleavage occurring after incorporation into the virion (Schochetman et al., 1978). Little is known about the enzymes effecting cleavage or the regulatory mechanisms involved. It is of interest that the capsid precursor protein of both avian (Eisenman and Vogt, 1978) and certain murine oncoviruses (Racevskis and Sarkar, 1978) is phosphorylated prior to cleavage, suggesting that phosphorylation may play a role in the subsequent cleavage; several of the intermediate cleavage products are also phosphorylated (Racevskis and Sarkar, 1978). The proteases that effect the cleavages have been postulated to be host proteases (Shapiro and August, 1976; Reynolds et al., 1977); however, virus-coded proteases may also be involved. Recent analyses of the amino and carboxyl terminal amino acid sequences of capsid proteins derived from the capsid precursor protein of certain murine oncoviruses has revealed the presence of tyrosylproline and phenylalanylproline bonds that are resistant to most of the known proteases (Oroszlan et al., 1978). This suggests the involvement of a novel protease, perhaps coded by the virus genome. In this regard, it is of interest that a unique and specific proteolytic factor has been identified in purified mouse oncovirus preparations that is capable of cleaving the capsid precursor protein to its constituent polypeptides (Yoshinaka and Luftig, 1977a,b; Luftig and Yoshinaka, 1978). Further study is required in order to determine whether the proteolytic factor(s) is a host cell or virus coded enzyme@). 2. Viral Replication and Cell Growth A number of studies have indicated that murine oncovirus-infected cells release virus only during periods of active growth and cease virus production in the Go state of proliferative arrest (Panem and Kirsten, 1973; Paskind et al., 1975; Sherton et al., 1976; Evans et al., 1977). Virus production and release resumes after the cell cycle is reinstituted. Several major events, therefore, during viral growth are apparently related to specific phases of the cell cycle. This is further substantiated by studies, utilizing synchronized permissive cells, that show that virus is released during mitosis from a number of cell lines infected with murine oncoviruses (Panem and Schauf, 1973; Paskind et al., 1975; Schauf and Panem,

142

PAUL H . BLACK

1976). Studies with synchronized cells have further established that synthesis occurs in three distinct periods of the cell cycle; this synthetic pattern is consistent with the time required for processing of precursors, appearance of viral antigens at the cell surface, and budding during mitosis. Maximal expression of viral antigens, as determined by fluorescence microscopy, occurred in the S and G2phases about 1-2 hours prior to release (Naso and Brown, 1977). Thus, viral protein synthesis and its insertion into the membrane occur at a specific time in the cell cycle that is similar to the biosynthesis of certain host plasma membrane proteins, as discussed earlier. It is of interest that endogenous viral genetic information may be induced spontaneously in growing cells since many more oncovirus particles are seen in intestinal crypt cells that actively proliferate than in fully differentiated villous cells of the jejunum of normal mice (Croker et al., 1977). 3. Shedding in Oncovirus-Infected Cells

The envelope glycoprotein, in soluble form, of certain muri oncoviruses has been detected in culture fluids of infected cells maintained in vitro (Bolognesi et al., 1975) and in the serum of many mouse strains, even in the absence of infectious virus (Strand and August, 1976; Kennel, 1977). This glycoprotein is presumably shed from the cell surface; endogenous viral genes coding for the various oncovirus proteins may be expressed noncoordinately, which explains the presence of this viral product in the absence of infectious virus (Strand and August, 1976; Kennel, 1977; Schochetman et al., 1978). The capsid precursor protein can also be found in the culture fluid of infected cells (Herberman et al., 1973) or in mouse sera in the presence or absence of infectious virus (Ledbetter and Nowinski, 1977; Snyder et al., 1977; Schochetman et al., 1978). The latter product, found in the serum of leukemic mice, was formerly thought to be a tumor antigen derived from the host cell surface (Gross cell surface antigen). Shedding of the major envelope glycoprotein of mouse mammary tumor virus (MMTV) into sera of mice bearing MMTV-induced breast tumors occurs (Ritzi et al., 1976, 1977; Arthur et al., 1978). Moreover, such shedding into the milk of certain strains of lactating female mice has been described; the latter shedding may be found in the presence or absence of whole virus and suggests that both synthesis and shedding of the glycoprotein occur from stimulated (i.e., lactating) cells (Ritzi et al., 1976). In mice bearing MMTV-induced breast tumors, there is a good correlation between serum envelope glycoprotein levels and tumor size (Ritzi et al., 1976); removal of tumor causes a decrease and recurrence

CELL SURFACE SHEDDING

143

of tumor causes an increase in serum envelope glycoprotein levels. Thus, monitoring serum viral envelope glycoprotein levels is both of diagnostic and prognostic significance (Ritzi et al., 1977), as it is with shed host cell products (see following). 4. Shedding in VSV-Infected Cells

In addition to released virus, shedding of the G glycoprotein from the cell surface occurs during infection with VSV. Approximately one of six molecules of G is shed, and newly synthesized molecules are more likely to be shed, suggesting that shedding may be a consequence of an initial unstable association between G and M protein possibly resulting from an inability of G to recognize M (Little and Huang, 1977). Indeed, more shedding occurs with virus mutants that are defective in their M proteins (Little and Huang, 1977). However, proteolysis may also be involved, since the shed product is smaller and the inhibitor phenylmethylsulfonylfluoride (PMSF) decreases shedding (Little and Huang, 1978). It has been suggested that the failure to organize and aggregate G glycoprotein, presumably by an intact M protein, might result in an increased lability for proteolysis (Little and Huang, 1978).

5 . Proteolysis at the Cell Surface: Lytic Viruses Lytic enveloped viruses may also require proteolysis of viral membrane precursor proteins during replication. Generally, host cell proteases are thought to be involved. The envelope glycoprotein of Sindbis virus is cleaved on the cell membrane (Smith and Brown, 1977), as is the hemagglutinin precursor protein of influenza virus (Shied and Choppin, 1975; Klenk et al., 1977). Two precursor glycoproteins, one carrying neuraminidase and hemagglutinin activities and the other containing the fusion glycoprotein, are cleaved on the membrane of cells infected with certain strains of Newcastle disease and Sendai viruses (Nagai and Klenk, 1977). 6. Summary

The studies of nonlytic enveloped viruses indicate that (1) cellular permissiveness for the synthesis and processing of viral structural proteins is determined by the stage of the cell cycle and, therefore, cell growth is required; (2) virus budding occurs at the M phase; and (3) shedding of viral membrane (g1yco)proteins occurs and is also associated with cell growth (Evans et al., 1977). These factors are also important determinants of host plasma membrane protein synthesis and shedding. Although the mechanism of shedding is not known, cleavage of viral

144

PAUL H . BLACK

precursor membrane proteins is known to occur at the cell surface, and preliminary evidence suggests the involvement of proteases in virus shedding. The possibility that a poorly organized or aggregated viral glycoprotein (G of VSV) may render it more susceptible to proteolysis raises the question as to whether similar factors are operative in the cancer cell. Moreover, the fact that M phase is the cell cycle phase where oncoviruses bud (a process that requires fusion as the bud pinches off) further suggests that fusion and release of membrane components may more readily occur in the absence of anchoring restraints of the cell membrane. VII. Mechanism of Shedding

Little is known of the events that cause shedding in normal or cancer cells. Possible mechanisms of shedding of individual (glyco)proteins, as well as of relatively large structures, will be considered.

A . PROTEINS A N D GLYCOPROTEINS: PROTEOLYSIS AND SHEDDING Since activated and transformed cells have increased cell surface and released protease activities, proteolysis at the cell surface is one possible mechanism of shedding of cell surface proteins or glycoproteins. This is supported by experiments, which showed that molecules released by protease treatment are similar to those spontaneously shed and that certain molecules such as fibronectin are extremely sensitive to proteolytic cleavage. It has also been shown that exposure of certain normal chick and human fibroblasts to certain proteases alters many of their properties to those of the transformed phenotype and may even stimulate cell growth (Blumberg and Robbins, 1975; Zetter et al., 1976). All of these results suggest that an increase in cell surface proteolytic activity is responsible for shedding and certain features of the transformed phenotype (Burger, 1970; Sefton and Rubin, 1970; Vaheri et al., 1973; Roblin et al., 1975b). There are a number of examples of proteolysis at the cell surface. The cleavage of certain viral precursor proteins of enveloped viruses already has been mentioned. Cell surface proteases may generate cell surface enzymatic activity by limited proteolysis (see Neurath and Walsh, 1976, for review; Richert and Ryan, 1977). Receptors on certain cells (e.g., neutral retinal cells) during development may be generated by limited proteolysis; such cleavage products may be released from the cell and are recovered in the culture fluid (Rutishauser et al., 1976). There is also

CELL SURFACE SHEDDING

145

evidence that immunoglobulins on the surface of tumor cells may be cleaved (Ran et a / . , 1975; Cotropia et al., 1977; Yefenof et a/., 1978). One obvious approach to investigation of this question would be to utilize various protease inhibitors in an attempt to decrease proteolytic activity and determine whether growth inhibition and/or reversion of the transformed to a more normal phenotype occurs (See Roblin et a / . , 1975b, for review). In early experiments, growth inhibition of virus-transformed cells was achieved by several protease inhibitors, particularly the synthetic chloromethyl ketones, N-tosyl-L-phenylalanylchloromethylketone (TPCK) and N-tosyl-L-lysylchloromethyl ketone (TLCK) (which are specific inhibitors for chymotrypsin and trypsin, respectively); the effect was thought to be selective for transformed cells (Schnebli and Burger, 1972). However, Chou et al. demonstrated that these compounds, which are potent alkylating agents, inhibited cellular protein synthesis and that the effects were both secondary to toxicity and not selective for transformed cells (Chou et a / . , 1974a,c). Furthermore, the inhibitor-treated cells were not arrested in Go-G, as originally suggested (Schnebli and Haemmerli, 1974; Collard and Smets, 1974). The generation of plasmin (fibrinolytic) activity by PA was thought to be responsible for several features of transformed cells (Ossowski et a / . , 1973, 1974). However, inhibition of fibrinolytic activity by eaminocaproic acid (EACA), a known inhibitor of fibrinolysis, did not restore density-dependent growth inhibition to transformed cells (Chou et a/., 1974b). Moreover, transformed cells exposed to macromolecular inhibitors (for example, soybean or pancreatic trypsin inhibitors and ovomucoid, which are active against the plasmin system and less toxic than the chloromethyl ketones) did not change in morphology or growth characteristics; nor did fibronectin reappear on the cell surface of treated transformed cells (Hynes et d.,1975; Hynes and Pearlstein, 1976). These experiments suggest that activation of plasminogen is not required for maintenance of the transformed phenotype. The question of whether a correlation exists between generation of fibrinolytic activity and loss of fibrinoectin from transformed cells has been investigated. Intact fibronectin molecules are recovered in the culture fluid bathing transformed cells; molecules with a slightly smaller subunit size and a missing interchain disulfide link (like those obtained after plasmin digestion) are not detected (Vaheri and Mosher, 1978). In studies of transformed cell lines, little correlation existed between presence of fibrinolytic activity and absence of fibronectin at the cell surface (Wolf and Goldberg, 1976; Pearlstein et a / . , 1976; Gallimore et af., 1977). The evidence presented does not indicate that plasmin is responsible for the transformed phenotype or the shedding of fibronectin. Certainly,

146

PAUL H . BLACK

other proteases may be involved, as may plasminogen activator itself. With respect to the latter, Quigley and Goldfarb (1978) have recently reported that the changes in morphology of chick cells induced by PMA are accompanied by a large increase in synthesis and shedding of PA and are prevented by inhibitors of PA (diisopropylfluorophosphate (DFP), leupeptin, antipain, benzamidine) but not by inhibitors of plasmin (trasylol, EACA, and soybean trypsin inhibitor). These results suggest that an arginine-specific, DFP-sensitive serine protease may be responsible for the PMA-induced morphological and cellular alterations; such alterations, as discussed above, may be initiated by shedding. However, the involvement of such a protease in shedding remains to be proven. This is not unlikely, since there is evidence that protease(s) may be involved in virus shedding as previously discussed. Although the studies described suggest the involvement of cell surface PA in causing cellular phenotypic changes, one must be cautious in the interpretation of experiments utilizing protease inhibitors: ( 1) the compound may be toxic (for example, the chloromethyl ketones); (2) little information is available as to whether certain protease inhibitors remain outside or enter the cell; and (3) little conclusive proof exists to show that an external protease inhibitor (alone or bound to some type of inert support that remains outside the cell) is exposed to and actually inhibits certain protease(s) during the experiment. One must, therefore, conclude at present that there is suggestive evidence, but not conclusive proof, for the involvement of cell surface protease(s) in phenotypic conversion and shedding. Further work should clarify this issue and may have important therapeutic implications. It is also possible that release of a (g1yco)protein may require specific intervention of a chemical process (e.g., covalent modification by phosphorylation, glycosylation, or acetylation); such alteration may also alter the susceptibility of a (g1yco)protein to proteolysis. The fact that certain virus precursor proteins are phosphorylated prior to cleavage is of interest in this regard (Eisenman and Vogt, 1978; Racevskis and Sarkar, 1978). The virus studies also indicate that failure to organize or aggregate the G glycoprotein of VSV by the M protein may render it more susceptible to proteolysis (Little and Huang, 1977, 1978). The fact that cell membrane stabilizing factors derived from the exoskeleton and endoskeleton are diminished, together with the randomness of the IMP, in cancer cells suggests that similar events may be operative. Thus, the increased lateral mobility of cell surface (g1yco)proteins in transformed as compared to normal cells suggests that cleavage by a cell surface protease might more readily occur in transformed cells, Indeed, glycosylation reactions occur in individual transformed mouse cells (cis-glycosylation), whereas such

CELL SURFACE SHEDDING

147

reactions occur only on adjacent contact-inhibited mouse cells (transglycosylation); this suggests that more contact between enzyme and acceptor molecules was achieved in the former cells (Roth and White, 1972). B. MEMBRANE STRUCTURES: MICROVILLI , VESICLES,BLEBS The evidence that Ca is lost from the cell membrane during fusion, during growth, and during mitogen stimulation, and the changes iil membrane structure that may occur were discussed earlier. The coupling of synthesis and insertion of plasma membrane proteins with shedding, at least for PA, suggests that events that occur at the plasma membrane and enhance fusion should facilitate shedding. Such conditions, together with changes in membrane structure and its stabilizing elements, are likely to contribute to the release of membrane structures (for example, portions or all of the glycocalyx, microvilli, vesicles, and blebs described in Section V,B) from certain neoplastic cells. It is likely that components of the exoskeleton, cell surface Ca' and/or ATP, and endoskeletal elements may all be involved in microvillus formation. Poste has pointed out that the formation of microvilli is influenced by the amount of cell coat material; cells with extensive cell coats (>35 A) have few microvilli, whereas cells with thinner cell coats (<35 A) have considerably more microvilli (Poste, 1972). This may reflect the membrane stabilizing effect of the exoskeleton. Tumor cells and cells continually passed in tissue culture contain many microvilli and the smallest amount of cell coat material (Poste, 1972). Microvilli can be induced in fibroblasts and lymphocytes by inhibiting cellular energy generation with sodium azide or dinitrophenol (Vlodovsky et al., 1973; dePetris, 1978). The mechanism(s) for microvillus formation under these conditions is not understood. Diminished cellular ATP may favor membrane destabilization and microvillus formation by affecting the state of polymerization of endoskeletal elements; for instance, enhanced Con A agglutinability in transformed cells occurs following ATP depletion (Vlodovsky et al., 1973). However, cell surface ATP content may also be an important determinant of microvillus formation. The fact that cell surface ATPase activity is elevated in several tumor cell types (Stefanovic et af., 1974; Ronquist and Agren, 1975; Cittadini, 1977), especially over microvilli (Epstein and Holt, 1963; Karasaki et al., 1977; Cittadini et al., 1977), suggests that loss of ATP locally may be important in microvillus formation. The presence of numerous microvilli in normal mitotic and cancer cells, as well as in activated lymphocytes (dePetris,

148

PAUL H . BLACK

1978), suggests that events of cell activation favor microvillus formation. Shedding of microvilli and/or other protrusions from the surface of cancer cells presumably results from fusion at the base. A model for the formation of blebs and/or microvesicles that separate from the cell has been studied in red blood cells. Mention has been made of the pronounced rearrangement of IMP that occurs in aged SRBC, presumably due to ATP depletion. Spontaneous vesiculation and release of protein-poor vesicles may then occur (Lutz ef al., 1977). No actin or spectrin is associated with the vesicles and such a dissociation may be important in vesicle formation and release. In several other studies utilizing fowl or human RBC, increases in intracellular Ca have resulted in pinching off of protein-poor vesicles (Allan et af., 1976; Hart ef al., 1976; Zakai et al., 1977) that lacked actin and spectrin (Allan et al., 1976; Zakai et al., 1977). When studied by freeze-fracture EM, the vesicles were seen budding from IMP-free areas (Zakai et al., 1977). It is possible that increases in intracellular Ca caused the contraction of actin and/or spectrin and the lateral movement of proteins, resulting in protein-poor lipid vesicles (Allan et al., 1976; Zakai et al., 1977; Lutz et al., 1977). In these studies, dissociation from protein, especially from actin and spectrin, apparently occurred at the inner surface of the RBC and presumably enhanced microvesicle formation; pinching off of the vesicle is a result of fusion at the base of the vesicle and would be enhanced in the absence of protein. Although the data are scant, one may hypothesize that events occurring in the experimental systems (leading to microvillus, bleb, and vesical formation and release) also occur in certain cancers, especially leukemias and cancers characterized by circulating cells. Such cancer cells may have shed portions or all of their glycocalyx. Their cell membranes may be analogous to the Ca-dissociated state of the fusion reaction (showing a reduction in membrane Ca and/or ATP, possibly due to increased plasma membrane ATPase). The randomness of the IMP found in cancer cells (perhaps due to loss of restraint from the microfilament system) would favor fusion at the base and shedding of structures that form from a destabilized plasma membrane. VIII. Consequences of Shedding from the Cancer Cell Surface

Increased synthesis, incorporation into the plasma membrane, and release of cell surface components by the cancer cell have been discussed earlier. In this section, possible consequences of these events will be

CELL SURFACE SHEDDING

149

considered in an attempt to explain certain aspects of the pathogenesis of cancer and certain phenomena associated with cancer.

A. CANCERCELLSURFACE 1. Shedding and Malignancy

There is evidence from certain animal models that the extent of shedding from the tumor cell surface is an important determinant of the degree of malignancy of a tumor. For example, spontaneously metastasizing rat mammary carcinomas had much less glycocalyx and were less immunogenic than the nonmetastasizing tumors (Kim et al., 1975). Evidence was presented that this was due to increased shedding rather than diminished synthesis and, indeed, circulating tumor antigens were found in the sera (see following). Another mouse tumor provides a striking example of the relationship between shedding and malignancy. Adaptation of mouse mammary adenocarcinoma cells to growth intraperitoneally as an ascites-forming tumor resulted in the isolation of two sublines. One subline (Ha) lost strain specificity and grows in allogeneic hosts, whereas the other (St) retained strain specificity and grows only in syngeneic hosts. A largemolecular-weight glycoprotein, epiglycanin (approximately 450-500,000 d), is present on the surface of Ha, but not St, cells and is shed into the ascites fluid of mice bearing the Ha tumor (Cooper et al., 1974; Miller et al., 1977). The Ha tumor is less immunogenic than the St tumor, presumably because epiglycanin covers the H-2 specific antigens; this masking allows allotransplantation (Cooper et al., 1974; Codington et al., 1978). Thus, in this system, a large glycoprotein is continually synthesized and shed from the cell surface and seems to be responsible for allotransplantability by masking cell surface tumor antigens (Killion et al., 1976); it is also possible that shed material may block host immune mechanisms also (see following) (Nowotny et al., 1974, 1976). Although glycocalyx shedding is apparently correlated with malignancy, the pathogenetic mechanism(s) underlying this relationship is not clear. One component of the glycocalyx, fibronectin, has been studied in some detail. In various animals tumors, fibronectin loss correlates with degree of malignancy. Indeed, cell lines established from the most malignant and invasive tumors, both spontaneous and experimentally induced, were found to have little to no fibronectin (determined by immunofluorescence) (Chen et al., 1976; Gallimore et al., 1977). In these studies,

150

P A U L H . BLACK

an inverse correlation was found between tumor induction by transformed cells and amount of fibronectin present. In fact, loss of fibronectin was thought to be the best parameter of malignancy (Gallimore et a/., 1977). Since fibronectin is an adhesion protein (Yamada and Olden, 1978) and loss of adhesion characterizes cancer cells, we might consider the interactions of fibronectin with other substances and also some possible physiological functions of this molecule.

2 . Importance of Fibronectin Fibronectin (derived either from the cell surface or from media) has been found to be similar in most properties to cold insoluble globulin (CIG), which is a protein in blood normally present in concentrations of 0.3 pg/ml (see Yamada and Olden, 1978, and Vaheri and Mosher, 1978, for reviews). Most available evidence indicates that CIG is the circulating form of fibronectin and presumably arises from normal shedding (Vaheri and Ruoslahti, 1975b; Macarak et a/., 1978). It was known for some time that certain serum factor(s) were required for attachment of cells to plastic substrates, with or without collagen; CIG has now been identified as one such cell adhesion factor (Pearlstein, 1976; Yamada and Olden, 1978). Purified fibronectin and/or CIG can also increase aggregation of dissociated cells in vitro (Yamada et a / . , 1977). Fibronectin and CIG bind to collagen or gelatin (Engvall and Ruoslahti, 1977). Purified chick and human fibronectin bind most avidly to heatdenatured collagen types I, 11, and 111. Less binding to native collagen occurs; of the native collagens, binding to type 111 is the most pronounced (Engvall et ul., 1978). Evidence that fibronectin and collagen interact in vivo has been obtained from immunofluorescent studies of rat fibroblast monolayer cultures in which collagen and/or procollagen were localized on the cell surface in a reticular pattern similar to fibronectin (Bornstein and Ash, 1977). Both components were lost when cells were in suspension culture and during mitosis. It is probable that fibronectin is a cellular binding site for collagen, since collagen will not bind to cells in mitosis, which lack fibronectin (Bornstein and Ash, 1977). Moreover, recent evidence indicates that fibronectin is the collagen receptor on platelet membranes (Bensusan et a / ., 1978). Fibronectin-collagen binding may be important in the cellular interaction with and organization of connective tissue, since fibronectin is especially concentrated at the basal lamina of cells and is abundant in various basement membranes (Linder et af., 1975). It is of interest that fibronectin-mediated attachment of cells to collagen or plastic substrates requires Ca (Yamada and Olden, 1978). This suggests that loss of Ca from the cell membrane under different

CELL SURFACE SHEDDING

15 1

physiological conditions would decrease adhesion, and is consistent with the well known dispersal effect of divalent cation chelating agents on adherent cells in culture. Binding to collagen is important in the process of removing collagenous debris from the blood by the macrophages of the reticuloendothelial system. A factor from serum, the anti-gelatin factor (AGF), was found to bind gelatin and was required for macrophage ingestion of denatured collagen (Wolffetal., 1967). Recently, this factor has been purified, and it has properties identical to CIG (Dessau et al., 1978). Either CIG or AGF can be found on the surface of guinea pig peritoneal exudate macrophages and can be removed by trypsin; the ability to bind denatured collagen by such trypsinized macrophages can be restored by addition of either purified CIG or AGF (Dessau et al., 1978). CIG has also been identified as the opsonic protein from humans, azSB glycoprotein, which is necessary for clearance of gelatin-coated colloidal particles from the blood by the reticuloendothelial system (Blumenstock et al., 1978). Antiserum to either purified CIG or purified opsonic protein will decrease phagocytic uptake by Kupffer cells of the liver (Blumenstock et nl., 1978). Thus, CIG is important in regulating hepatic reticuloendothelial phagocytic activity and possibly host defense. Fibronectin also binds to fibrinogen and fibrin (Ruoslahti and Vaheri, 1975; Engvall et al., 1978; Colvin et al., 1979), and there is evidence indicating that the same site on fibronectin is involved in both fibrinogen and collagen binding (Engvall et nl., 1978). Such binding may play a role in cell migration during wound healing (see following). Fibronectin also binds to heparan sulfate and a fibronectin-heparan sulfate complex may be one of the essential elements mediating cell-to-substrate adhesion (Culp et af., 1978; Rollins and Culp, 1979), and possibly cell-to-cell and cell-to-intercellular matrix adhesion as well.

B. CRYOFIBRINOGENEMIA A precipitate may develop in the cold from citrated, oxalated, or EDTA-treated plasma of certain individuals, and can be dissolved, in part, by rewarming (Stathakis et al., 1978). Such a pathological precipitate has been called cryofibrinogen. Cryofibrinogenemia is not uncommonly associated with thromboembolic conditions (Ball and Goldman, 1976), especially the migratory thrombophlebitis accompanying neoplasia (Trousseau’s syndrome) (Sack et al., 1977) or during disseminated intravascular coagulation (DIC) (Mosesson et al., 1968). It can be produced

152

PAUL H. BLACK

by adding small amounts of thrombin to normal plasma o r to plasma treated with agents enhancing fibrin formation. Both fibrinogen and fibrin had been known for some time to be components of cryofibrinogen. A third component, CIG, first described in the cryoprecipitate of a patient with ovarian cancer and a DIC syndrome, is also essential (Mosesson et al., 1968). Stathakis et al. (1978) have recently examined the role played by each of the three components in cryofibrinogenemia by utilizing fibrinogen- Sepharose columns and measuring the binding of CIG before and after conversion to fibrin. Although CIG binds to fibrinogen in the cold, CIG will not form a cold insoluble precipitate alone or with added fibrinogen; fibrin is required for precipitation. Fibrinogen is also essential since it forms complexes with fibrin; these complexes are required to keep fibrin in a soluble state (Shainoff and Page, 1962) prior t o complexing and precipitation with CIG. CIG is essential, since precipitation does not occur in its absence, even when solutions containing fibrinogen-fibrin complexes are saturated with fibrin. However, the amount of precipitate that forms can be increased by increasing the ratio of fibrin t o fibrinogen. CIG is thought to act as a nucleus with multiple binding sites for the fibrinogen-fibrin complex (Stathakis, 1978). CIG can also be covalently cross-linked to fibrinogen-fibrin complexes by activated factor XI11 (transglutaminase) and CIG-fibrinogen-fibrin cryoprecipitates can be stabilized by activated factor XI11 (Yamada and Olden, 1978). Thus, an important biological role may be served by CIG in modifying the solubility of fibrinogen-fibrin complexes. The increased shedding of cell surface fibronectin, together with the increased coagulation that occurs with certain neoplasms, would be likely to promote cryofibrinogenemia. CIG may also be cross-linked to fibrin during the final stages of blood-clotting by activated factor XI11 (Vaheri and Mosher, 1978). Thus, CIG may be consumed during malignancy, especially in hypercoagulable states such as DIC. This may account for the decreased CIG levels that occur during advanced malignancy in animals and man (Blumenstock et al., 1978). It is of interest that the amount of cryoprecipitate in the patient with ovarian cancer and DIC mentioned above, fell with heparin therapy and rose with cessation of such therapy (Mosesson et al., 1968). Since heparin limits coagulation by its antithrombin action, it is likely that the reduction in cryofibrinogen is due to decreased fibrin formation. It would be of interest to determine plasma CIG levels under these circumstances. Binding to tumor cells may also account for the lower CIG levels in advanced cancer. CIG may also be sequestered locally at sites of entry into the vascular systems of malignant cells or foreign particulate material (Blumenstock et a / ., 1978).

CELL SURFACE SHEDDING

153

C. SHEDDING, COAGULATION, A N D FIBRINOLYSIS IN CANCER 1 . Coagulation and Cancer In this section, aspects of the excessive stimulation of the coagulation and/or fibrinolytic systems that may occur with neoplasia and the role of shedding in these processes will be considered. Normally, coagulation is induced locally at the site of vascular injury (via the extrinsic mechanism) by the release or exposure of thromboplastin or tissue factor (TF), which is a glycoprotein whose activity may be enhanced by associated lipid (Zacharski and McIntyre, 1973a,b). When exposed, TF binds factor VII and generates an enzyme that can directly activate the plasma protein factor X (Nemerson, 1966). A number of other reactions occur and culminate in the conversion of prothrombin to thrombin. Thrombin cleaves fibrinogen to two sets of fibrinopeptides, A and B, and activates factor XIII, which cross-links fibrin by introducing multiple covalent bonds between approximated fibrin chains. The hemostatic mechanism in cancer seems to be poised between hyperactivity of the coagulation system (with widespread fibrin deposition and thrombosis) and excessive stimulation of the fibrinolytic system (this occurs with DIC). Examples of the former are the high incidence of thrombosis in autopsied cancer cases (Sack et al., 1977) and the thromboses that occur with occult tumors of pancreas, lung, stomach, and other neoplasms. Pathologic activation of both the coagulation and fibrinolytic systems, however, results in thrombolytic and hemorrhagic events (DIC). In this syndrome, consumption of platelets, fibrinogen, and other factors occurs during coagulation, while hydrolysis of fibrinogedfibrin with generation of split products results from the fibrinolytic activity. Much evidence indicates that the consumptive process is initiated by the release of some thromboplastin-like material into the circulation, with resulting coagulation; the fibrinolytic activity is thought to be induced secondarily (Straub et al., 1967). Indeed, the syndrome can be induced by the administration of thromboplastin to animals (Harker and Slichter, 1972). TF (thromboplastin) is thought to be sequestered on the surface of a number of cell types [including fibroblasts (Zacharski and McIntyre, 1971, 1972; Maynard et al., 1975), endothelial, and smooth muscle cells] and may also be associated with subendothelial structures such as collagen and basement membrane (Maynard el d . , 1977). It is the “procoagulant activity ,” described in cultures of fibroblast cells (Zacharski et al., 1969; Zacharski and McIntyre, 1971; Green et al., 1971), that was

154

PAUL H . BLACK

later identified as T F (Zacharski and McIntyre, 1972). Early studies indicate that T F activity of human fibroblasts in culture increased with time, and it was thought that attachment and spreading triggered the synthesis of TF. However, total cellular and surface activities have been found to be highest in growing fibroblast and endothelial cells, whereas confluent cells have little or no surface activity; the appearance of T F activity requires RNA and protein synthesis (Zacharski and McIntyre, 1971, 1973a; Maynard et al., 1977). One may interpret these findings in light of the model presented in this review: cell growth evokes new TF synthesis, transport, and insertion into the cell membrane. The cryptic nature of T F in confluent cells indicates that either little T F is present on the cell surface or that it is masked; the former is likely to be the case if T F resembles PA in its metabolism. This, of course, would be desired especially in endothelial cells. If total and cell surface TF activity are increased in growing cells, the same situation might be expected to occur in cancer cells. Indeed, tumor cells are particularly rich in thromboplastic activity (O’Brien ef al., 1968; Sakuragawa et al., 1977). Recent evidence also indicates that TF is shed from tumor cells in v i m (Dvorak et al., 1979), and it is likely that such shedding from tumor cells occurs in vivo; such release of TF is presumably responsible for activation of the extrinsic pathway of clotting and may occur both locally and systematically in cancer. In retrospect, it is likely that T F was the lipid-protein complex associated with a variety of human cancer cells and produced fibrin in and around the tumor nidus (O’Meara, 1958; O’Meara and Thornes, 1961). There is evidence that both animal (Dvorak et a / . , 1979) and human (O’Meara, 1958; O’Meara and Thornes, 1961) tumors may be invested in a fibrin gel. Local fibrin formation (due presumably to shedding of TF) is thought to be important in the growth of both primary and metastatic tumors (Hilgard and Thornes, 1976), since tumor cells may require a fibrin network as a matrix for growth (Laki and Yancy, 1968). The importance of the association of fibrin with embolic tumor emboli in arrest and vascular infiltration during metastatic spread will be considered later. Release of large amounts of T F into the systemic circulation is likely to result in DIC. This is supported by a number of studies in which elevated thromboplastin activity in various solid tumors (Sakuragawa er a l . , 1977) and leukemias (Pollack, 1971) was thought to be responsible for the DIC that accompanied these neoplasms; the mechanism of release of TF, whether by shedding and/or autolysis, is not yet clear, however. It is of interest that normal leukocytes contain little to no TF. However, endotoxin increases the TF content appreciably (Niemetz, 1972); such

CELL SURFACE SHEDDING

155

induction in these and/or other cells with shedding may be operative in the DIC syndrome induced by endotoxin. 2. Fibrinolysis and Cancer

Normally, fibrinolysis is confined locally to a clot. Plasminogen binds to fibrin by lysine binding sites; such sites are also involved in the binding of plasminogen or plasmin to antiplasmin, which is the antiprotease that rapidly neutralizes plasmin in the circulation. Plasminogen activator also binds avidly to fibrin. Thus, PA can convert plasminogen to plasmin, which, while sequestered locally from antiplasmin, can hydrolyze fibrin. Released plasmin can then be neutralized by antiplasmin (Collen et al., 1977; Rakoczi et al., 1978; Wiman and Collen, 1978). In disease processes, plasmin may overwhelm the system and cause a fibrinolytic syndrome. It may be produced experimentally by the selective infusion of urokinase (PA) with subsequent conversion of plasminogen to plasmin and cleavage of fibrinogen (Bell et al., 1968; Harker and Slichter, 1972). This may also occur in the primary fibrinolytic syndrome associated with carcinoma of the prostate, a very rich source of urokinase (Brassine et al., 1976). Usually, however, the fibrinolysis associated with cancer is secondary to increased coagulation, and primary fibrinolytic syndromes are rare (Davidson et al., 1969; Harker and Slichter, 1972). Whether increased shedding of PA by tumor cells other than prostate cells and generation of plasmin locally is important clinically with respect to either fibrinolysis or tumor invasion and metastasis is speculative (see following). It is clear that many human tumors, both of epithelial and connective tissue origin, have a high spontaneous fibrinolytic activity (Davidson et al., 1969; Svanberg and Astedt, 1975). Such activity may prolong bleeding from a locally incised tumor (Davidson et al., 1969). Localized fibrinolysis induced by tumor PA may also be a factor in the hydrolysis of the fibrin meshwork, which is presumably utilized for growth and must eventually be hydrolyzed. Elevated PA activity is especially common in highly vascularized tumors (Todd, 1964; Astrup, 1966). In some studies, the release of fibrin degradation products (FDP) from various human tumor explants cultured on performed clots varies with the vascularity of the tumor (Bjorlin et al., 1972). It is known that endothelial cells in culture have high PA activity (Loskutoff and Edgington, 1977) but the significance of these findings is not understood. The possible implications of the high fibrinolytic activity of endothelial cells in the vascular space and in granulation tissue will be considered later.

156

PAUL H. BLACK

D. INVASION A N D METASTASES The ability of cells to invade normal tissues and metastasize to distant sites characterizes them as malignant cells. If one hopes to interrupt this process, the mechanisms underlying these events must be known. For invasion, and/or metastases to occur, cells must separate from the primary tumor mass. In addition, during the process of metastasis, cells must enter vascular channels, be disseminated throughout the body, become arrested, penetrate the endothelium, move out of the vascular space, and infiltrate and grow in the extravascular space. Some aspects of these events will be considered, particularly those associated with blood-borne metastases. Although the discussion will be focused on single cells or groups of cells separated from the primary tumor, it is well known that solid cancers can invade both blood vessels and lymphatic channels and give rise to metastases (see Poste, 1977; Sugarbaker and Ketcham, 1977; Weiss, 1977; and Roos and Dingeman, 1979, for reviews). The mechanism(s) whereby cells separate from the primary tumor mass is not understood. Increased tumor tissue pressure, the loss of certain cell-to-cell junctions, the known motility of cancer cells, a decrease in cell-to-cell or cell-to-intercellular matrix adhesion, and the increase in cell surface and/or released proteolytic activity may all be factors (Poste, 1977; Sugarbaker and Ketcham, 1977). In keeping with the concepts delineated in this article, the latter two factors will be stressed. The loss of adhesive forces between cancer cells, which facilitates separation, has been known for some time (Coman, 1944). Such loss presumably results from some change in the physicochemical bonds on the surface of adjacent cells or between the cell and intercellular matrix. The decrease or absence of adhesion proteins such as fibronectin or a fibronectin-glycosaminoglycan (e.g., heparan sulfate) complex (Culp et al., 1978; Rollins and Culp, 1979) would influence such interactions and is likely to be an important factor in the loss of adhesion. More insight into the mechanism of separation can be gained by considering the properties of peripheral tumor cells, which separate, infiltrate, and may metastasize. This area of a tumor is usually well vascularized and the cells have a high rate of DNA synthesis compared to the more hypoxic core (Sugarbaker and Ketcham, 1977). It is known that in culture an increased growth rate of both normal and cancer cells is associated with cell detachment from a solid surface, and the fact that cells in mitosis round up and partially detach from the monolayer has been discussed. It is also known that the most actively dividing tumors (and also the most anaplastic) have the greatest tendency to infiltrate and metastasize (Weiss, 1977). Thus, significant correlations have been made

CELL SURFACE SHEDDING

157

between growth “coefficients” of a number of human epithelial cancers and melanomas with the number of circulating tumor cells and metastatic spread (Glucksmann, 1948), or between the rapidity of doubling time and metastases (Weiss, 1977). This may seem inconsistent with the fact that the growth fraction (i.e., the percentage of tumor cells actively synthesizing DNA) becomes lower as a primary tumor enlarges, which is probably due to the fact that an increasing proportion of cells enter the Go or remain in the GI phase of the cell cycle (Schabel, 1975). Such cells can become activated, however. If portions of a large tumor are transplanted in animals, the growth fraction is again elevated in the transplanted tumors and metastases may occur; it is also high in the metastatic foci of animals bearing transplanted and primary tumors. The aforementioned evidence suggests that actively growing cells are associated with separation and metastases. The shedding of cell surface molecules (particularly the adhesion (glyco)proteins), which occurs during activation and cell cycle traverse, is likely to be involved in separation. Whether cell surface proteases, either directly or indirectly, affect the separation process is uncertain. The possible involvement of proteolytic enzymes elaborated by tumor cells (particularly collagenase) in modifying surrounding normal tissues to facilitate invasion and metastases has received much attention (see Strauch, 1972, for review). A number of studies indicate a relationship between malignancy and collagenase activity. For example, clinical aggressiveness of certain human and animal tumors has been correlated with high collagenase activity (Abramson et al., 1975), and degeneration of dermal collagen has been observed at the border of basal cell epitheliomas (Hashimoto et al., 1972; Yamanishi et al., 1972) and at the infiltrating edge of squamous cell carcinomas of the skin (Hashimoto et al., 1973). Villous projections from the advancing tumor border were observed to perforate the basal lamina in the latter study. Moreover, metastatic cells may also have high proteolytic activity. Thus, metastatic tumor cells isolated from the venous effluent draining a tumor were shown to have greater proteolytic activity than primary tumor cells (Liotta et al., 1977); such cells have an increased capacity to solubilize type 1 collagen from human dura and to degrade isolated pulmonary basement membrane tissue. Similar proteolytic activity associated with tumors is presumably responsible for the defects in basement membrane, which have frequently been visualized by EM and are seen adjacent to invading tumor cells (Kellner and Sugar, 1967; Hashimoto et al., 1973). Because of the complexity of basement membrane, more than one hydrolytic enzyme was probably involved. These results, together with reports of the selection of supermetastatic sublines from tumors (Fidler

158

PAUL H . BLACK

and Kripke, 1977), indicate that cells in the tumor population are heterogeneous and that cells that have exposed or are releasing proteases may more successfully infiltrate and metastasize (Bosmann et al., 1973); metabolically active cells at the periphery of tumors are more likely to exhibit these characteristics. Additional evidence for the involvement of collagenase in tumor invasion is derived from studies of cartilage. Collagenase released from human osteosarcomas and mammary carcinomas is inhibited by a lowmolecular-weight protein, that can be isolated from bovine hyaline cartilage and is apparently an integral component of the cartilage matrix (Kuettner et al., 1977). This is of interest since cartilage is rarely invaded by tumor tissue, whereas adjacent bone tissues, whose major organic component is also collagen, is often invaded and replaced by osteosarcoma. This is true for bone metastases from breast cancer as well. Although bone and cartilage contain genetically distinct types of collagen, both are susceptible to collagenase cleavage, although cartilage collagen is somewhat more resistant (Kuettner et al., 1977). Recently, protease inhibitors extracted from bovine nasal cartilage were found to inhibit both surface proteolytic activity and the proliferation of normal and malignant cells (Burk and Hatcher, 1978). The studies with cartilage indicate that the presence of inhibitors of cellular proteases may determine the extent to which cells infiltrate a tissue by affecting their capacity to grow, and that both growth and invasion may be related to surface proteolytic activity. If collagenolysis occurs in the area surrounding an infiltrating tumor, the question arises as to whether the collagenolytic activity is derived from cell surface-bound or released enzyme. There is evidence that collagenase activity is associated with the cell surface of normal, non-confluent, human gingival fibroblasts and acts only when the membrane contacts collagen fibrils (Rose and Robertson, 1977). This suggests that procollagenease and/or collagenase reside on the cell surface where activation may occur under certain circumstances, with subsequent release by shedding. It is of interest that plasmin can convert latent or procollagenase to collagenase, and there is evidence that both procollagenase and PA are present in the same cell (Werb ez af., 1977); plasmin, therefore, may be responsible for generation of collagenase activity. Plasminogen activation might occur on the cell surface or, more likely, in association with fibrin around the tumor cell. Both PA and plasminogen bind tightly to fibrin, and conversion to plasmin with subsequent fibrinolysis or other proteolytic activity could occur in this sequestered locus away from any antiplasmin present, as described earlier. A hypothetical model is presented in Fig. 1 1 . Such a series of events may be operative

CELL SURFACE SHEDDING

T

159

FIBRIN TISSUE FACTOR

+

PLASMINOGEN ACTIMTOR

PLASMINOGEN

#

PL ASMIN

COLLAGENA PROCOL SE LAGENASE

FIG. 1 I . Model for generation of tumor cell collagenolytic activity by plasmin (see text).

in the neoplasms in which marked fibrinolytic activity was detected at both the advancing edge of invading tumors and at metastatic foci (Weiss and Beller, 1969; Peterson et af., 1973). Activation of procollagenase may also be caused by other neutral proteases present in or released from the same cell (Horwitz et af., 1976). However, more must be learned about collagenase release, its activation, and its site of activity. Although the mechanism(s) involved in the infiltration of connective tissue by tumor, and the role of collagenase and/or other proteolytic enzymes in this process are far from settled, some tissue breakdown presumably occurs. In this connection, it is of interest that large amounts of proteins of MW 5,000-8,000 are excreted in the urine of patients with disseminated neoplastic disease, especially during the last 6 months of life; the amino acid profile of the peptides suggest that they are likely to be breakdown products of connective tissue components (Rudman et al., 1969). Tumor cells released into the vascular space are subjected to a number of stresses: mechanical and physicochemical trauma, immunologic factors such as humoral antibody, and possibly nutritional and metabolic stress (Weiss, 1977). Only a small portion of cells survive to form metastases. In one study of mouse melanoma cells, only 0.1% of inoculated cells formed lung metastases, the principal organ where metastases occur (Fidler, 1973). Much of the information about tumor cell arrest has been derived from studies utilizing phase cinemicroscopy of animal blood vessels (particularly the central artery of the rabbit ear) after inoculation of tumor cells (Wood, 1971; Warren, 1973). Generally, single cells do not adhere t o a normal vascular endothelium and are swept into the vessels. Tumor emboli containing clumps of tumor cells and platelets more readily arrest on blood vessel walls (Fidler, 1973). Adherent tumor emboli become enmeshed in a fibrin network. Autoradiographic studies using lZ5Ilabeled fibrinogen have established that fibrin augments the entrapment of adherent tumor cells in a microthrombus (Wood, 1971). It is likely that

160

PAUL H . BLACK

fibrin is associated with the tumor embolus; EM and immunological studies of lung tissues of rats inoculated with tumor and labeled fibrinogen indicate that fibrin is associated with tumor emboli within 30 seconds (monomeric fibrin), while stable or cross-linked fibrin was observed within 5 minutes (Chew and Wallace, 1976). The integrity of the vascular endothelium is also a major factor in tumor cell arrest. If the endothelium is damaged or absent, tumor emboli become firmly embedded in fibrin and progressively pass through the fibrin and blood vessel wall, possibly utilizing fibrin as a matrix (Warren, 1973). There is some suggestion from studies of capillary endothelial physiology that endothelial cells may be released, creating gaps and exposing bare basement membrane (Warren, 1973). Adhesion of tumor emboli to these areas readily occurs, with subsequent transcapillary migration. There is also evidence that tumor cells may modify vascular endothelium or mesothelial cells by causing retraction of cells, exposure of basement membrane, and subsequent tumor cell arrest (Poste, 1977). Such retraction has been induced by thrombin (Shimamoto, 1974) or histamine (Majno, 1969) and results in exposure of tumor emboli to subendothelial structures, thereby enhancing arrest. CIG is known to bind to injured vascular endothelium, or more precisely, to exposed subendothelial components containing collagen. It is possible that its presence might promote the binding of circulating tumor cells, especially those containing fibrin, since CIG binds to fibrin as well (Engvall et al., 1978). These studies indicate the importance of tumor cell clumps containing fibrin and of some compromise in the integrity of the endothelial surface during arrest. It is likely that intact endothelial cells prevent the arrest of most tumor cell emboli by generating plasmin, with subsequent fibrinolysis of the fibrin in the emboli. That endothelium is important in tumor cell arrest is evident from the localizing effect of metastases during acute tissue injury. Thus, disruption of the capillary bed with trauma to spleen, liver, or extremity results in tumor cell localization at these sites (Sugarbaker and Ketcham, 1977). Little is known about the mechanism of transcapillary migration of tumor cell emboli. Some studies with experimental animal tumors indicate that cellular processes from arrested tumor emboli push through and destroy the endothelium at these points; the tumor cells thereby gain access to the extravascular space (Chew et al., 1976). Other studies indicate a transcapillary penetration between endothelial cells in a manner similar to the movement of a PMN across an endothelial surface (Sindelar et al., 1975). The involvement of a fibrin support and proteolytic enzymes in such movement remains to be determined.

CELL SURFACE SHEDDING

161

The importance of coagulation in tumor cell arrest and entrapment has prompted studies investigating the effects of anticoagulation on the incidence of metastases. Although few studies have been carried out, anticoagulation has generally resulted in beneficial effects. Thus, less invasion and spread of carcinoma cells were observed in the arterial circulation of the ear of a rabbit treated with sodium warfarin (Wood, 1971); a decrease in the arrest of 1251-labeledtumor cells with decreased metastases were observed in several species of animals rendered thrombocytopenic or treated with aspirin; and anticoagulation after surgery of patients with osteosarcoma resulted in a decrease in pulmonary metastases (Sugarbaker and Ketcham, 1977). Throughout the cascade of events that occur during metastases, there is continuous selection: certain cells separate, infiltrate, and enter the vascular space; only a fraction of circulating cells successfully adhere and penetrate; and only a proportion acquire a vascular supply and stroma and successfully grow. One may hypothesize that dividing and shedding tumor cells are involved in separation, infiltration, and metastases. Such cells are relatively devoid of adhesion proteins and contain and shed surface proteases. Such proteases, either alone or together with others activated either on the cell surface or in the immediate tumor vicinity, may induce the tissue hydrolysis necessary for invasion and metastasis. Surface and/or released T F activity is likely to promote fibrin formation associated with both the primary tumor and the tumor emboli.

E. SHEDDING A N D TUMOR IMMUNITY Although factors associated with tumor cells are operative in facilitating invasion and metastases, host immune mechanisms are marshalled in an attempt to contain tumor cell growth and spread. These responses and some possible mechanisms whereby tumors evade such control will now be considered. The transformation of normal to malignant cells is accompanied by the appearance of new antigens on the surface of malignant cells in a large number of experimentally induced and spontaneous cancers. The host may respond to these antigens with humoral antibody and specifically immune T cells. In addition, non-T, and non-B cells (the socalled “null” or K cells that have Fc and complement receptors) mediate antibody-dependent cellular cytotoxicity (ADCC); either ihe same cell or a different K (the “natural killer” or NK) cell may also be involved in tumor cell destruction and may play a role in immune surveillance (Pross and Baines, 1976). Macrophages, activated both specifically and nonspecifically , also may destroy tumor cells, but their precise role in tumor

162

PAUL H. BLACK

immunity has not yet been delineated. The nature of the interaction of these factors in mediating tumor immunity has not been definitively established and there may be important differences between tumors induced by different agents, between different species, and between experimental and spontaneous tumors. Notwithstanding these differences, the host must frequently recognize the nonself nature of the antigens and presumably eliminates such cells. However, in many instances, despite the presence of new antigens and an apparent immune response, the tumor effectively escapes the immune constraints of the host, grows, and eventually may kill the host. 1. Tumor Escape Mechanisms

A number of mechanisms by which animal tumor cells escape host immune destruction have been described (see Klein, 1975, for review). Tumor antigens may be masked or covered by another molecule, for example, epiglycanin in the mouse mammary ascites tumor discussed earlier. Another escape mechanism is antigenic modulation, which may result when tumor cells are exposed to antibody specific for cell surface antigens (Old et al., 1968). The antibody causes a redistribution (patching, capping, then endocytosis with or without shedding) and eventual elimination of the tumor antigen from the cell surface; reappearance of antigen does not occur in the presence of antibody, and the cell effectively escapes from the cytotoxic effect of complement in this in vitro model (Stackpole e f al., 1974). Aside from masking or modulation of tumor antigen, a change in or, more rarely, loss of tumor antigen may occur in certain tumors. The former may result from immunoselective pressures that tend to select for cells with “weaker” tumor antigens, whereas a loss of tumor antigen(s) may occur in rapidly growing tumors (Biddison and Palmer, 1977) o r after prolonged cultivation of certain tumors as organ cultures in vifro (Parks, 1975; Jacobs, 1976). After such cultivation, the cells are less immunogenic, since they can be transplanted readily to allogeneic hosts and have an increased tendency to metastasize. Whether such loss is important during metastases in vivo is not known. It is of interest, however, that cells that separate from the primary tumor mass and infiltrate normal tissue to form metastases may have diminished antigenicity (Jacobs and Uphoff, 1974; Sugarbaker and Cohen, 1972; Goldman et al., 1974), and the fact that metastasizing tumors are less antigenic than their nonmetastasizing counterparts has been discussed (Kim et al., 1975). Although the mechanism of such change in or loss of antigenicity is not known, shedding is one mechanism whereby cell surface tumor antigens are lost. Shedding may play a key role in tumor

CELL SURFACE SHEDDING

163

escape not only by eliminating tumor antigen, but also by the capacity of the shed product to act as or to generate other blocking factor(s). 2 . Blocking Factors: Tumor Antigen Much evidence indicates that at least a portion of the antitumor response is mediated by effector T cells with specific receptors for the tumor antigen and that engagement of the receptor by cellular tumor antigen is necessary for killing (Kuppers and Henney, 1977). Lymphocytes from tumor-bearing animals or humans often kill cells from their respective tumors in vitro (see Golub, 1975, for review). Sera from tumorbearing animals block the killing of tumor cells by autologous lymphocytes in vitro in an immunologically specific manner; this has been thought to represent the in vitro correlate of tumor enhancement (Hellstrom and Hellstrom, 1969, 1970). The blocking activity was first attributed to antibody that, when complexed to cell surface tumor antigen, was thought to block cell-mediated immune ractions (Hellstrom and Hellstrom, 1969). Although there is evidence that free antibody may be present on the surface of tumor cells (Braslawsky et a l . , 1976), current evidence does not support the concept that such antibody mediates immunological enhancement in vivo (see Currie, 1976, for review). Actually, antibody is likely to be protective, since bound antibody is required for ADCC by its interaction with the Fc receptor of K cells. Moreover, antibody has been found to be protective against blood-borne metastatic spread (Procter et a l . , 1973). Most evidence indicates that blocking activity in serum is related either directly to free circulating tumor antigen and/or antigenantibody complexes or indirectly to suppressor factor(s). Neither the precise role played by each of these blocking factors in tumor escape nor their interrelationship(s) are completely clear a t present. Direct blocking will be considered first. Blocking of cellular immune reactions mediated by lymphocytes from humans and animals bearing tumors and autologous tumor cells in vitro has been achieved in many instances by tumor antigen alone; this suggests that tumor antigen may be the active blocking factor (see Price and Baldwin, 1977, for review). Blocking by antigen is of the central type (where tumor antigen and effector lymphocytes block), whereas tumor antigen reacted with target cells does not cause blocking. Blocking centrally by tumor antigen is generally specific and is dependent on the amount of antigen present in the blocking serum (Bonavida, 1974). The tumor antigen in serum originates by shedding from the surface of tumor cells. This assertion is based on the following evidence: (1) labeled cell surface molecules containing antigenic activity are released in vitro (Ben-

164

PAUL H . BLACK

Sasson et al., 1974; Bystryn, 1977); (2) the released material is antigenically similar to molecules cleaved from the cell surface with proteolytic enzymes (Ting and Rogers, 1977; Stuhimiller and Seigler, 1977); and (3) release is unaccompanied by cell lysis (determined by continued cell viability and failure to demonstrate soluble cytosol markers). Shedding from tumor cells may occur rapidly; shedding of approximately half of cell surface human melanoma tumor antigen occurs in 3 hours (Bystryn, 1977). Large amounts of free tumor antigen may be found in serum (Rao and Bonavida, 1977; Lopez and Thomson, 1977) and urine of patients with advanced malignant disease (Lopez and Thomson, 1977). In a number of instances, tumor antigen is complexed to other material and is, therefore, released as a large macromolecular aggregate (Ben-Sasson et al., 1974; Bystryn, 1977; Murray et al., 1978). Further purification of the released cell surface material has resulted in antigenic molecules of relatively lower molecular weight (15,000 MW in the case of human melanoma tumor antigen) (Murray et al., 1978); higher-molecular-weight tumor antigens (of approximately 50-60,000) have been obtained from tumors induced with certain chemical carcinogens (Bowen and Baldwin, 1976), but the degree of purity achieved is not certain. 3. Blocking Factors: Immune Complexes Immune complexes can also block in vitro. The incidence of immune complexes in cancer sera is fairly high, ranging from 50-80%, depending On the type of assay employed (Rossen et al., 1977; Theofilopoulos et al., 1977). Immune complexes in cancer sera may be composed of cell surface material of varying antigenic specificities, e.g., tumor antigen (Loughbridge and Lewis, 1971), fetal antigen (Costanza et al., 1973), or viral antigen (Oldstone et al., 1974). Such antigens, having evoked a humoral immune response in the tumor-bearing hosts, are likely to combine with antibody in the circulation subsequent to their shedding from the tumor cell surface (Thomson et al., 1973b; Theofilopoulos et al., 1977; Rossen et al., 1977). Indeed, blocking factors have been produced artificially by reacting free antitumor antibody with tumor antigen solubilized from whole tumor cells (Baldwin et al., 1973). Immune complexes may also be formed by stripping of cell surface antigen with specific antibody (Calafat et a l . , 1976; Nordquist et al., 1977; Perrin and Oldstone, 1977). Although immune complexes can act as a blocking factor, it is not clear whether blocking occurs centrally or at the target cell level. Binding of immune complexes to the target cell, possibly by the Fc receptor of tumor cells (Targowski et al., 1977) or to tumor antigen (Thomson et al.,

CELL SURFACE SHEDDING

165

1973a; Long et al., 1977), may occur, and target cells may remove blocking factor, which can subsequently be eluted at low pH and shown to contain two components (Oldstone, 1975a). Immune complexes may also facilitate binding of antigen to the effector T lymphocyte by antibodymediated cross-linking processes, resulting in specific blockade and low zone tolerance (Diener and Feldman, 1972). Whether immune complexes act centrally or peripherally may depend on whether they occur in antigen or antibody excess, respectively (Theofilopoulos et a/., 1977). There is also evidence that immune complexes can bind to the Fc receptor of the K cell and block ADCC (MacLennan, 1972; Prather and Lausch, 1976; Sugamura and Smith, 1977). In a recent study, sera from 74% of cancer patients inhibited ADCC; the greatest inhibition was found in sera from patients with metastatic disease (Mikulski et al., 1977). Immune complex binding to macrophages via the Fc receptors has also been demonstrated (Ryan et al., 1975). The presence of blocking factors in serum is related to the size of the tumor and the extent of metastases (Youn et a / . , 1973; Grosser and Thompson, 1975; Pross and Baines, 1976; TheofiIopoulos et al., 1977; Lopez and Thomson, 1977). With a number of human and animal tumors, blocking factors are present with growing tumors (Bennet et all, 1975; Urovitz et a/., 1976) and not in hosts with regressed tumors; cellmediated immune reactivity against autochthonous tumor cells is present in both, however. Moreover, the most rapidly growing tumors in certain experimental animal systems (i.e., those accompanied by metastases) are associated with the largest amount of blocking factor(s); and elevation of blocking factor(s), especially tumor antigen, frequently precedes metastases (Alexander, 1974; Jamasbi et al., 1978). A number of such metastasizing and nonmetastasizing tumors have been studied; the former shed cell surface components more rapidly and to a greater extent, have less glycocalyx, and are less immunogenic than their nonmetastasizing counterparts (Alexander, 1974; Kim et a / . , 1975; Davey et a/., 1976; Jamasbi et al., 1978). In addition, a greater lability of cell surface molecules is found in metastasizing tumors than in nonmetastasizing tumors (i.e., antigenic modulation occurred much more readily in the former than in the latter tumors) (Davey et al., 1976). Excision of tumor has generally been accompanied by a loss of blocking factors (Steele et a / . , 1975) and may occur within 24 hours with certain experimental animal tumors (Bray and Keast, 1975). Furthermore, treatment of patients with chemotherapy has resulted in a loss of serum blocking activity (Sinkovics ef al., 1972; Noonan et a/., 1977; Brown et a/., 1978), whereas recurrence of disease may be accompanied by an increase in blocking activity (Brown ef al., 1978). It is of interest that

166

PAUL H . BLACK

after excision of a tumor or other nonsurgical treatment, serum antibody to tumor antigen increases, suggesting that the tumor was the source of the antigen that complexed with antibody and thereby prevented the levels of free antibody from increasing (Odili and Taylor, 1971). In summary, the amount of circulating tumor antigen and/or immune complexes is related both to the mass of tumor present and to the extent of shedding; the latter is apparently related to the metabolic activity of the tumor. Shed antigen may provide powerful protection of a tumor by effectively neutralizing humoral antibody and sensitized effector T lymphocytes. With generation of immune complexes, abrogation of ADCC and further blocking of host cell-mediated tumor immune mechanisms may occur. That antigen from the tumor is the cause of blocking activity is supported by the fact that removal of tumor generally eliminates blocking activity. The discussion thus far has been concerned with the direct blocking effect(s) of tumor antigen and/or immune complexes in tests of cellmediated immunity in vitro. Whether such blocking by these serum factors occurs in vivo is not known. However, in recent years, suppressor activity which promotes tumor growth, has been found in tumorous hosts. The questions therefore arise as to the relationship between blocking and suppressor activities, and whether tumor antigen may evoke suppressor activity. 4. Suppressor Activity

There is evidence that host suppressor cells, which are found in the spleens or thymuses of tumor-bearing animals, or specific soluble products, which are shed from and can substitute for suppressor cells (Treves et al., 1976; Nelson et al., 1975), play a role in promoting tumor growth by decreasing the effective host antitumor immune response (see above and Broder and Waldmann, 1978, for review). Suppressor cells are a subset of T cells, but non-T cells may also be present in tumor-bearing animals and humans and can nonspecifically suppress a wide spectrum of immunological functions. Certain studies suggest that suppressor cells are evoked in response to antigenic stimulation. With respect to tumor antigen, immunization may stimulate a tumor immune host response against a transplanted tumor; however, a paradoxical effect may result and growth of a transplanted tumor may be enhanced. Thus, repeated exposure of an animal to soluble tumor antigen, tumor extracts, or X-irradiated tumor cells prior to tumor challenge may result in enhanced tumor growth. Suppressor cells specific for the tumor have been demonstrated in such animals (Rao and Bon-

CELL SURFACE SHEDDING

167

avida, 1976; Embleton, 1976; Hellstrom and Hellstrom, 1978). Such a sequence of events apparently occurs during the growth of a transplanted tumor (Manor et al., 1976). Early during tumor growth ( 1 week), splenic T cells protected an animal against transplanted tumor, whereas splenic T cells (or medium of cultures of such cells) from animals bearing tumors for a longer interval (4 weeks) enhanced the growth of a transplanted tumor. Moreover, when tumor antigen, in the form of irradiated tumor cells, is administered to a mouse at the time of extirpation of tumor, there is a decrease in antitumor resistance, whereas animals not receiving such tumor antigen exhibit increased antitumor resistance (Vaage, 1973). There is evidence that tumor antigens enhance tumor growth in vivo by interacting with a radiosensitive population of T cells; suppressor cells are generally sensitive to sublethal doses of X-irradiation (Hellstrom and Hellstrom, 1978). The induction of such suppressor activity by large amounts of free tumor antigen in tumor-bearing animals is presumably responsible for the depressed tumor-specific immunity and results in what might be viewed as a state of immunological tolerance (Embleton, 1976). Although the aforementioned studies suggest that tumor antigen evokes suppressor activity, it is not always clear from these studies whether the specific immunosuppressive effect is due to tumor antigen or to suppressor factor (Embleton, 1976). In this connection, a tumor-specific blocking factor (a glycoprotein of MW approximately 56,000) has been purified from the serum of mice bearing a methylcholanthrene-induced fibrosarcoma (purification utilized affinity chromatographic techniques and an antiserum raised against tumor cells) (Nepon et al., 1976, 1977). Such a factor could represent either tumor antigen or a suppressor molecule. Both are specific, both are diminished after tumor excision, and both can block in specific cytotoxicity tests; however, binding of the molecule by tumor target cells would be more consistent with a suppressor molecule. In addition, the fact that “unblocking” antibodies from the serum of tumor-immunized mice bind to the factor and not to tumor-specific antigen further indicates that the molecule is a suppressor factor (Nepon et al., 1977). Furthermore, if added tumor antigen (in the form of irradiated tumor cells) was given to an irradiated animal (presumably devoid of suppressor cells), no enhancement of growth occurred; enhancement only occurred in the presence of radiosensitive cells from the spleens of tumor-bearing or even normal mice (Hellstrom and Hellstrom, 1978).. From these studies, it appears likely that tumor antigen enhances tumor growth by interacting with a radiosensitive population of T cells, presumably suppressor cells. The role of tumor antigen in inducing suppression is further suggested

168

P A U L H . BLACK

by the relationship that exists between the presence of tumor and suppressor factors, both specific and nonspecific. Thus, a parallelism of tumor growth and the presence of specific suppressor cells has been observed (Poupon e f al., 1976). Moreover, spleen cells from animals with progressively growing tumors contained suppressor cells for lymphocyte mitogenesis and the mixed lymphocyte reaction, whereas spleen cells from regressor animals did not (Glaser et al., 1975). In addition, excision of tumor results in loss of a factor that inhibited the response of spleen cells to PHA (Gillette and Boone, 1975). In summary, both specific and nonspecific immunosuppressive factors are present in cancer patients. Although the in vivo blocking activity of tumor antigen and/or immune complexes has not been definitively established, tumor antigen appears to be the essential factor in eliciting suppressor activity. Whether suppressor activity is evoked to a greater extent by shed antigen or by cell surface antigen is not known. The relationship between shedding of tumor antigens and malignancy, however, suggests that shedding is an essential element in the abrogation of the immune response in cancer and that one approach to reduction of suppressor activity might involve attempts to reduce shedding.

F. GLYCOSYL TRANSFERASES There is much current interest in glycosyl transferase metabolism in normal and cancer cells, especially since the reported elevations in patients with cancer has led to utilization of glycosyl transferase levels as diagnostic tests for cancer. The glycosyl transferases are a group of enzymes that catalyze the sequential addition of individual monosaccharides (for example, galactose, fucose, and sialic acid) to a parent protein or to various lipid molecules during glycoprotein and glycolipid synthesis, respectively. The enzymes are membrane-bound and are primarily associated with the Golgi membranes and vesicles (Schachter er al., 1970). However, recent biochemical (Bernacki, 1974; Lloyd and Cook, 1974; Verbert et al., 1976; Chatterjee e f al., 1976; Lamont et al., 1974) and EM techniques (Porter and Bernacki, 1975) have demonstrated glycosyl transferase activity on the plasma membrane as well. 1. Sialyl Transferase

Several reports indicate elevated levels of sialyl transferase in tumors and/or in the sera of experimental animals or humans with cancer; assays utilized desialylated fetuin and/or desialylated human q-acid glycopro-

CELL SURFACE SHEDDING

169

tein as acceptor (Bosmann and Hall, 1974; Bernacki and Kim, 1977; Jaken and Mason, 1978; Ip and Dao, 1978). In general, there is a correlation between elevation of serum sialyl transferase activity and the rapidity of tumor growth or the extent of tumor involvement (Bosmann et al., 1975; Bernacki and Kim, 1977). For example, animals with certain metastatic rat mammary tumors had higher tumor and serum levels of sialyl transferase than those animals with nonmetastasizing tumors and/ or normal liver, suggesting that the tumor cells were the source of the serum sialyl transferase. A direct correlation between serum sialyl transferase levels and the amount of serum sialoglycoconjugate was also found in these studies; this further suggests that both were derived from the tumor cell surface (presumably by shedding) (Bernacki and Kim, 1977). Evidence that the increase in sialylation occurred on the cell surface was derived from studies of y-glutamyl transferase in mouse mammary tumors. y-Glutamyl transferase is a cell surface enzyme thought to be involved in transport of amino acids (Meister and Tate, 1976); increased sialylation of this enzyme was found and may reflect the marked increase in sialyl transferase activity (7- 13-fold) found in these tumor cells (Jaken and Mason, 1978). Increased sialylation of y-glutamyl tranferase also occurs to a greater extent in fetal tissues than in adult tissues in both animals and humans, and in proliferating intestinal crypt cells as opposed to differentiated intestinal villous cells (Kottgen et al., 1976). The increase in protein sialylation that occurs in these growing tissues also may reflect the observed increase in sialyl transferase activity (Weiser, 1973; Bauer et al., 1977) (see later). It is of interest that the concentration of yglutamyl transferase itself is increased in human hepatoma tissue relative to normal liver levels. It is apparently shed, since it is present in the accompanying ascites fluids (Peters et al., 1977). Increased levels may also occur in chirrotic livers with active cellular regeneration and in the accompanying ascites. Thus, the cell surface enzyme y-glutamyl transferase is also synthesized and shed to a greater extent in cancer and growing normal tissue than in resting cells (Peters et al., 1977). Whether the increased sialic acid content of the enzyme plays a role in the shedding process is uncertain. Elevated levels of glycolipid-bound sialic acid may also occur in mammary tumors and in the sera of both mice and humans bearing such tumors (Kloppel et al., 1977). The altered ganglioside pattern of serum has been found to correspond to that of the tumor (Skipski et ul., 1975). The increase in sialic acid content of such gangliosides is likely to be related to the increase in sialyl transferase activity present in such tumors (Moskal et al., 1974; Kloppel el al., 1977). Elevation of serum ganglioside levels may precede the appearance of a palpable tumor, and following

170

P A U L H . BLACK

excision of tumor, the levels return to normal in both mice and humans (Kloppel et a f . , 1977). These observations, together with their predominantly cell surface location, suggest that shedding is likely t o be a mechanism of release of such gangliosides. 2 . Fucosyl Transferuse At least two fucosyl transferases have been described in human and mouse sera. One enzyme transfers L-fucose from GDP-L-fucose to the terminal galactose of oligosaccharides or desialylated glycoproteins such as feutin (fucosyl transferase A), whereas the other catalyzes the addition of L-fucose to terminal N-acetyl-D-glucosamine o r D-glucose, or to fetuin from which both sialic acid and galactose have been removed (fucosyl transferase B) (Bauer et a f . , 1977; Chatterjee and Kim, 1978). The majority of the activity of both enzymes is associated with plasma membrane-enriched fractions of various rat tumors, suggesting that the plasma membrane may be the locus of greatest activity in tumors. However, both enzymes may not always appear in parallel; a 6-7-fold increase in fucosyl transferase B activity is present in metastatic as compared to nonmetastatic rat mammary carcinomas (Chatkrjee and Kim, 1977, 1978). This is consistent with the finding that elevation of fucosyl transferase B (3-7 fold) has been reported in 85% of all human cancer sera, especially from patients with highly malignant or metastatic disease (Bauer et al., 1977); less marked elevations of fucosyl transferase A occurred. Thus, although both enzymes are present in plasma membraneenriched fractions, fucosyl transferase B is apparently present to a greater extent in, and is shed from, the more malignant tumors. 3. Galactosyl Transferuse Galactosyl transferase activity using either endogenous or exogenous (asialo, agalactofetuin) acceptor has been found to be elevated in transformed cells (Lamont et al., 1977) and in some tumor tissues (Kessell et a f . , 1977; Chatterjee and Kim, 1977). Markedly elevated levels are found in spontaneously metastasizing as opposed to nonmetastasizing rat mammary carcinomas (Chatterjee and Kim, 1977). In this study, most of the activity was associated with plasma membrane-enriched fractions of either type of tumor. However, the specific activity of various cellular fractions was much higher in the metastasizing than in the nonmetastasizing tumors, and the former tumors shed large amounts of enzyme; this suggests a more rapid turnover of this cell surface enzyme in the metastasizing tumor cells (Kim et a f . , 1975; Chatterjee and Kim, 1977).

CELI. SURFACE SHEDDING

171

The metastasizing tumor discussed earlier represents a highly passaged malignant neoplasm that has presumably undergone “tumor progression”. In general, serum galactosyl transferase levels in cancer sera are only slightly higher than in control serum (Podolsky and Weiser, 1975). However, utilizing a discontinuous polyacrylamide gel electrophoresis fractionation technique, Weiser et a/. detected a more slowly moving isozyme of galactosyl transferase (GT II), in 43/58 patients with cancer; isozyme level was highest in patients with widespread disease and was absent in controls (Weiser et a / . , 1976). Podolsky et al. (1977), utilizing an animal model, showed that the isozyme (GT 11) is unique to tumor tissue, that it might appear before evidence of the tumor, that the activity was linearly related to tumor growth, and that the Michaelis constant of the isozyme differed from that of the galactosyl transferase activity present in both tumor and normal serum (GT I) [i.e., K , UDP-gal of GT I1 (1.0 x lop5 M) was half that of GT I (2.0 x lop5 M)]. These results provide strong evidence that the enzyme originated from the tumor. Recently GT I1 activity was demonstrated in 70-80% of patients with carcinomas; the extent of elevation generally was proportional to the extent and progression of disease (Podolsky et a f . , 1978). 4. Summary and Conclusions These studies suggest that elevated levels of glycosyl transferases occur with neoplastic growth and that only certain isozymes may be involved. An increase in activity, however, is associated with growth in general. Growing normal cells have elevated levels of surface galactosyl transferase and a soluble form of that enzyme appears in the medium of growing cells (Lloyd and Cook, 1974; Lamont et al., 1977). When the cells become confluent, cell surface galactosyl transferase activity is lost (Lamont et a / ., 1977). Neither sialyl nor fucosyl transferase activity increased in parallel in this study (Lamont et al., 1977), but the proper substrate may not have been utilized for the latter enzyme (Chatterjee and Kim, 1978). Activity, therefore, is apparently cell cycle dependent. Indeed, increased cell surface galactosyl transferase activity has been found in the M phase of synchronized cells (Bosmann, 1974b; Webb and Roth, 1974). However, spontaneously transformed mouse cells had surface activity throughout the cell cycle (Webb and Roth, 1974). Serum stimulation of cells also results in increased cell surface activity and release of enzyme (Lamont et a/., 1977). Increased galactosyl transferase and the increased sialyl transferase activity mentioned earlier are also associated with the surface of intestinal crypt cells, whereas no activity is present on the microvillus portion of the plasma membrane of intestinal

172

PAUL H. BLACK

villous cells (Weiser, 1973; but see Weiser et al., 1978). The crypt cells are actively dividing cells, since intestinal epithelium undergoes constant renewal, with movement of cells up to the villous structure. In addition, activation of T lymphocytes by mitogens results in the appearance of surface galactosyl transferase activity (Lamont et al., 1974). Thus, with activation and growth, there is apparently an increased synthesis, movement to the cell surface, and shedding of galactosyl transferase. This is also the pattern of other cell surface (glyco)proteins, whose activity is cryptic in resting cells. One may speculate about the significance of the increase in surface glycosyl transferase activity and the increase in glycosylation of glycoproteins and/or glycolipids that may be shed. The increase in sialylation may increase the half life of circulating glycoconjugates shed from the cell surface by decreasing their plasma clearance rates (Lunney and Ashwell, 1976). If such glycoproteins are circulating tumor antigens, tumor escape mechanisms such as blocking and/or induction of suppressor activity may be enhanced. It has also been shown that gangliosides have a relatively longer retention time in plasma than asialolipids (neutral glycolipids) (Barkai and Di Cesare, 1975); gangliosides, therefore, may accumulate in cancer sera. The possible suppressor activity of gangliosides has been mentioned earlier and one may question whether similar gangliosides released from tumor cells may mediate some of the suppressor activity associated with neoplasia. In view of the increased sialyl transferase activity and the increased sialylation of certain cell surface enzymes during growth (Critchley et al., 1976), the questions might be raised as to whether the change in the properties of galactosyl transferase resulting in formation of the GT I1 isozyme is due to increased sialylation, or whether differences in sialic acid content of other glycosyltransferases may occur during growth or neoplasia.

IX. Shedding and Chronic Viral Disease

For several viruses of animals and humans, infection results in virus persistence throughout the host’s life, evoking a host antiviral immune response and immune complex formation (see Oldstone, 1975b, for review). Lymphocytic choriomeningitis virus (LCM) and oncoviruses in mice, and the human hepatitis B virus, in some instances, are examples of such chronic infections. Much of the antiviral immune response is directed at viral cell surface antigens, which may be shed from the cell

CELL SURFACE SHEDDING

173

surface and may also exist as free uncomplexed antigens in the circulation. The reason(s) for the persistence of chronic virus disease is not known, but much evidence indicates some type of impairment of the host's immune response that is analogous, in several respects, to the immunodepression of cancer patients. Chronic human hepatitis B infection will be briefly considered. Lymphocyte responsiveness to hepatitis B, antigen (HB,Ag) may be specifically depressed in patients with acute hepatitis, in patients with chronic persistent hepatitis and carriers, and in patients with chronic active hepatitis (Tong et al., 1975). Activity normally reappears during convalescence from acute hepatitis (Tong et al., 1975). This specific decrease in cell-mediated reactivity may be due to blocking by HB,Ag (which is presumably shed from the surface of hepatitis infected liver cells), by immune complexes or by specific suppressor factor(s). Nonspecific serum inhibitory factors of T cell function, which inhibit spontaneous and induced lymphocyte cytotoxicity and blast transformation, may also be present in patients with acute or persistent hepatitis infections; their presence is also related to the persistence of virus or viral antigen (Wands et al., 1975; Paronetto and Vernace, 1975; Brattig and Berg, 1976). The question again arises, as in the case of tumor immunity, of the relative contributions of specific blocking and/or various serum inhibitory or suppressor factors in the diminution of immune competence in chronic virus infections. The continuous presence of viral antigen is likely to be responsible for the depressed immune response, presumably by inducing suppressor factors. Whether these are more effectively induced by cell surface or shed viral antigen is not known. Perhaps the efficacy of interferon in the treatment of chronic hepatitis is due to its ability to prevent viral shedding (see later); HB,Ag in the sera of such patients may decrease after such therapy (Greenberg et al., 1976). It is possible, however, that the decrease in circulating HB,Ag after interferon treatment may be due to decreased synthesis (Greenberg et a / . , 1976; Purcell et al., 1976; Desmyther ez cil., 1976). It is of interest that infections with budding viruses are the predominant virus infections accompanied by anergy. Thus, acute infection with measles, rubella, mumps, varicella, influenza, and EBV, for example, which are all budding viruses, may be accompanied by transient anergy during acute infection (Kantor, 1975). This raises the question of whether shedding of viral antigen may be associated with a transient anergic state during acute infection and a tolerance-like state during chronic or persistent infection.

174

PAUL H . BLACK

X. Activation and Surface Proteases: Chemotaxis and Rheumatoid Arthritis

A. PROTEASES A N D CHEMOTAXIS The involvement of cell surface and, possibly, shed proteases in the chemotactic response will be considered. Chemotaxis implies the directed movement of a cell in response to a substance in the cell’s environment; the direction is determined by the concentration gradient of the chemoattractant (Keller et af., 1977). The chemotactic factor is assumed to bind to a receptor on the cell surface, and this binding results in a series of metabolic events, culminating in cell movement. Although little is known about these events (see following), evidence utilizing rabbit PMN’s indicates that upon such interaction, a chymotrypsin-like serine protease is activated and is required for chemotactic migration; a parallel inhibition of both chemotaxis and protease activity occurs with certain protease inhibitors (phosphonates) (Ward and Becker, 1968, 1970). This enzyme has been localized to the plasma membrane of PMN’s by fractionation studies; it is thought to be an integral membrane protein (Tsung et al., 1978). Binding of a chemotactic peptide occurs to the greatest extent in a plasma membrane-enriched fraction(s), indicating that the receptor is also localized in this fraction (Tsung et al., 1978). A similar enzyme has been found in human PMN’s (Tsung et al., 1977). Both chemotactic responsiveness and migration of isolated human PMN’s have been found to be sensitive to various inhibitors of serine proteases (a,-antitrypsin, a2-macroglobulin, TPCK, TLCK) (Goetzl, 1975); synthetic inhibitors with a specificity for chymotrypsin-like serine proteases were most effective, indicating that an enzyme similar to that of the rabbit may be important in human PMN chemotaxis (Goetzl, 1975). However, other serine proteases (e.g., PA) may also be involved in chemotaxis of human PMN’s (Kaplan et al., 1973). Recent studies of human PMN leukocytes activated by PMA or Con A correlate the synthesis and release of PA with movement; this observation further suggests an involvement of PA in migration of PMN’s (Granelli-Piperno et af., 1977). There is also evidence that migrating macrophages have cell surface proteolytic activity (Remold and Rosenberg, 1975) that may be essential for their chemotactic response. As with PMN’s, synthesis and release of PA from human macrophages occurs with cell activation and is associated with movement (Vassalli et al., 1977). In addition, exposure of macrophages to a number of protease inhibitors results in an enhanced response to migration inhibitory factor (MIF) (Remold and Rosenberg, 1975; Pies-

CELL SURFACE SHEDDING

175

sens et al., 1977). These studies suggest that migrating macrophage cell surface proteases may alter the receptor for MIF or, possibly, MIF, itself. The evidence presented suggests that interaction of a PMN with a chemotactic factor results in the appearance of surface proteolytic activity. Presumably, protease synthesis, movement, and insertion into the plasma membrane occurs with cell activation. Such a sequence of events would be similar to the activation of PMN’s by PMA or Con A, which results in PA synthesis and release (Granelli-Piperno et al., 1977). The mediators of such activation might be considered. cGMP increases human PMN chemotaxis, and PMA, which causes an increase in cellular cGMP, has a potentiating effect on the PMN response to a bacterial chemotactic factor (Estensen et al., 1973). With respect to human monocytes, increased movement is induced by serotonin and vitamin C, both of which increase cGMP (Gallin et al., 1978). With both PMN’s and monocytes, cAMP diminishes the chemotactic response and antagonizes the effect of elevated cGMP in augmenting chemotaxis (Estensen et al., 1973; Gallin et al., 1978). Thus, it is likely that cell activation and the chemotactic response are modulated by the cyclic nucleotides. It is more difficult to generalize about the role of Ca, since treatment of PMN’s with ionophore A23 187 (which elevates cytosol Ca) results in an increase in cGMP (Smith and Ignarro, 1975), whereas exposure of monocytes to this ionophore results in an elevation of cAMP and diminished chemotaxis (Gallin et al., 1978). Thus, the precise sequence of second messenger interaction resulting in activation has not been established. Furthermore, whether cellular activation results in the conversion of cell membrane proenzymes to their active form [as hypothesized for the chymotrypsin-like esterase of rabbit PMN’s (Becker, 1972)] is not known. Since most of the studies have been carried out with stimulated (i.e., exudate) PMN’s, it would be of interest to determine the subcellular location of proenzymes and protease activity in unstimulated cells. Although cell surface proteases appear to be essential for chemotaxis, the mechanism is not clear. Movement of cells is only possible if unoccupied surface receptors are available for binding to the chemoattractant. Therefore, the cell must synthesize new receptors or reutilize existing ones when saturated by chemoattractant (Dierich et al., 1977). Surface proteases may hydrolyze chemotactic peptides directly (Aswanikumar et al., 1976) or the cell may internalize the peptide-receptor complex (Harris, 1976). It is also possible that shedding of surface chemotactic factors with or without their receptors may occur; examination of the fluid bathing cells undergoing chemotaxis in vitro for these components would help clarify the issue.

176

PAUL H . BLACK

Another example of chemotaxis, accompanied by cell activation and the appearance of surface proteolytic activity, is the migration of endothelial cells into a tumor in response to the chemoattractant, tumor angiogenesis factor (Folkman, 1975). A recent study in the rabbit (utilizing EM and slit lamp stereomicroscopy) visualized the sequence of events accompanying migration of endothelial cells from the limbal vascular plexus to an implanted carcinoma in the cornea 1 mm away (Ausprunk and Folkman, 1977). Morphological evidence of endothelial cell activation occurred during the first day with an increase in number of cell organelles, an increase in cell size, and the appearance of cell surface projections; these changes resembled those seen in regenerating endothelium. Fragmentation of the basal laminar was evident at this time, and within the next 24 hours, cell junctions were loosened, gaps in the vascular endothelium appeared, and the cells commenced to migrate to the tumor. Thus, the events accompanying separation and movement of endothelial cells are apparently associated with activation and proteolysis. The advancing edge of the capillary sprouts actively synthesized DNA from day 2 and peaked at day 8. These cells would encounter the highest concentration of chemoattractant (Ausprunk and Folkman, 1977). In view of their activated state, they might also be the cells most likely to shed chemoattractant with or without receptor. During wound healing, endothelial cells and fibroblasts infiltrate a fibrin clot. Both endothelial cells, as described earlier, and fibroblasts [which also respond to a gradient of chemoattractive substances (Carter, 1965; Harris, 1973)l are presumably activated during such migration. In this case, however, the cells utilize a fibrin scaffolding to which they must adhere and from which they must separate. One possible mechanism of adhesion and release might involve fibronectin, since this is synthesized and rapidly released from activated cells. Perhaps fibronectin’s transient residence at the cell surface provides sufficient adhesion for traction during cell movement, especially since fibronectin has receptors for fibrin. Shedding of fibronectin (and possibly glycosaminoglycans) might then free the cell from this constraining matrix. Such a sequence has been observed in vitro, where fibronectin and glycosaminoglycan have been found on the culture dish in areas from which cells have migrated (the so-called “cellular footprints”) (Culp, 1976). CIG, present in a fibrin clot cross-linked by factor XI11 (Vaheri and Mosher, 1978), may also be involved in the transient adhesion of migrating cells. Hydrolysis of fibrin by any newly synthesized plasmin would be less likely to occur in a cross-linked fibrin clot (Yamada and Olden, 1978). The studies of the chemotactic response of phagocytic and endothelial cells suggest that the events accompanying activation of cells may also

CELL SURFACE SHEDDING

177

occur in these migrating cells and that surface and/or shed proteolytic activity may play an essential role in such regulated movement. Such migrating cells have many analogies with tumor cells. The fact that endothelial cells, like tumor cells, fail to invade cartilage, which contains an anti-protease that inhibits TAF-induced vascular proliferation, further suggests that protease activity is essential for movement of endothelial cells (Burk and Hatcher, 1978). The normal cell movement discussed in this and in previous sections dealing with embryogenesis and tissue remodeling may involve tissue destruction, but the process is orderly and controlled.

B. RHEUMATOID ARTHRITIS Abnormal cell movement and tissue destruction may occur in certain nonmalignant diseases such as rheumatoid arthritis. One may speculate that such movement and destruction are mediated by surface protease activity generated as a consequence of abnormal cell activation. Whatever the cause of rheumatoid arthritis, synovial cells proliferate and form an invasive pannus that grows along cartilage into the joint space (Krane, 1974). The rheumatoid synovial cell would have the characteristics of an activated cell (which has many features in common with a malignant cell) (Harris, 1976). Thus, cell surface proteases (e.g., PA) and/or other molecules (e.g., TF) are likely to be present and may be shed. The process of erosion is thought to start at the pannus-cartilage junction (Krane, 1974); this implies that some activity of the synovial cell is responsible. The destruction of collagen that occurs at this junction suggests that collagenase is activated locally, possibly on the synovial cell surface or after release from the cell. A model of collagenase activation involving plasmin was discussed in Section VII1,D. The sequence of events postulated to occur at the cell surface (possibly involving fibrin) during tumor cell invasion (see Fig. 11) may also occur during the erosion of rheumatoid arthritis. Fibrin may be present in the pannus; it is also seen by EM at the surface of rheumatoid articular cartilage (Kimura et al., 1977). Local fibrin deposition could be generated by cell surface or locally shed TF or by other reactions. The presence of fibrin, however, would facilitate plasmin formation (Wiman and Coleen, 1978). Thus, collagenolytic activity would be generated locally. Destruction of collagen may occur only in this way, since cartilage matrix unassociated with pannus is not destroyed, even in the presence of collagenase(s) in the joint fluid (Harris et af., 1969). Some support for this model can be derived from recent studies, which indicate that both collagenase and a protease that activates it may be

178

PAUL H . BLACK

released from the same cell. Thus, a complete collagenolytic system was generated in cultures of rheumatoid synovial cells by adding plasminogen to the cultures (Werb et al., 1977). Evidence was presented that indicated that the cultured cells produced both PA and latent collagenase and that the plasmin generated activated collagenase. Moreover, rabbit fibroblasts stimulated by certain proteases release both PA and collagenase (Werb and Aggeler, 1978), and latent collagenase from rabbit macrophages may be activated by a neutral protease released by the same cell (Horwitz et al., 1976). Whether collagenase activity is generated in vivo by PA and/ or other proteases from the same cell and whether this occurs as a consequence of rheumatoid synovial cell activation can be determined only by further study. Xi. Prevention of Shedding

Since a relationship exists between the extent of shedding, the amount of circulating tumor antigen, and malignant progression, attempts to decrease shedding might result in an enhancement of tumor immunity. Substances that are known to prevent the shedding of PA should be briefly mentioned in this regard-interferon (Schroder et al., 1978) and certain corticosteroids (Vassalli et al., 1976; Werb, 1978). The antitumor effect of interferon is well known, but the mechanism underlying this effect is not understood. Exposure of cells to interferon increases the expression of cell surface molecules such as H-2 antigens (Lindahl et al., 1973; Lindahl, et al., 1976) and the amount of cell surface viral glycoprotein (Chang et al., 1978). The enhanced expression may be due to the effect of interferon on membrane turnover, i.e., its ability to decrease shedding (Schroder et al., 1978). It is not known, however, whether interferon affects the shedding process directly or indirectly by its ability to inhibit cell proliferation (Strander and Einhorn, 1977; Lindahl-Magnusson et al., 1971). Whatever the mechanism, retention on the cell surface of tumor antigen might augment the host antitumor response. Certain corticosteroid hormones enhance the adhesion of transformed cells and also prevent shedding. These effects might be responsible for the phenotypic reversion of certain cell lines from a transformed to a normal morphology after treatment with hydrocortisone at physiological concentrations (Armelin and Armelin, 1978). Such effects on phenotype, however, may be secondary to the inhibition of DNA synthesis caused by hydrocortisone. Glucocorticoids may also inhibit the growth of normal cells (Dougherty et al., 1973; Fodge and Rubin, 1975). As with interferon,

CELL SURFACE SHEDDING

179

any effects on phenotype and shedding could be indirectly due to the antigrowth effects of these compounds that are evident in certain cell lines. Corticosteroids can prevent tumor promotion by PMA and other promoters; this effect is likely to be related to their antigrowth properties (Viaje et al., 1977). Their antipromoter effects, in general, parallel their antiinflammatory effects. The capacity to inhibit production and release of PA also parallels these actions of corticosteroids (Vassalli et al., 1976; Viaje et al., 1977; Werb, 1978). Since much of the inflammatory response is generated by release of cell products (by secretion and probably by shedding as well), the antiinflammatory response of certain corticosteroids is likely to be secondary to their ability to prevent or limit cell activation (Viaje et al., 1977) and cell movement (Vassalli et al., 1976). XII. Summary and Conclusions

The phenomenon of shedding of cell surface macromolecules, its role in normal membrane (g1yco)protein turnover, and its importance in certain disease processes have been reviewed. With cell activation (during growth or mitogen-stimulation of normal cells) there is an increase in synthesis, processing, and eventual insertion of certain membrane (glyco)proteins; some of these are proteases. The increased cell surface and/or released proteolytic activity appears to be necessary for the cell movement that occurs during embryogenesis, wound healing, and chemotaxis. Continued cell activation, which occurs in certain diseases such as rheumatoid arthritis, may result in abnormal cell movement, and sufficient cell surface proteolytic activity may be generated to cause tissue destruction. Shedding of cell surface (glyco)proteins, including proteases, is coupled to their synthesis and membrane insertion, but whether release of these macromolecules is secondary to proteolysis is not known. Glycocalyx components are apparently shed at a specific phase of the cell cycle, which may be an important determinant of the phenotype of the normal mitotic cell. Soluble factors shed from the surface of activated cells of the immune system are likely to mediate various immune reactions. Shedding is also involved in the turnover of certain cell surface receptors and the outer segments of photoreceptor cells of the retina. Shedding occurs continuously from cancer cells and may be responsible for several features of the cancer cell phenotype. Evidence has been reviewed that indicates that generation of surface proteolytic activity together with release of components of the cell surface may be important

PAUL H. BLACK

180

factors determining both the behavior of the cancer cell and various phenomena that accompany cancer. Thus, cell surface and/or shed proteases are likely to be important in separation, invasion, and metastasis of cancer cells. Release of tissue factor from the surface of tumor cells is likely to promote the excessive coagulation that occurs in some cancers, and released molecules may cause fibrinolytic syndromes, or be the components of abnormal serum complexes that are cryoprecipitable. Released tumor antigen, either alone or complexed with antibody, may cause blocking of cell-mediated immune tumor cell reactions; free tumor antigen may also evoke immunosuppressive factor(s). Cell surface enzymes such as glycosyl transferases, which are utilized as markers for malignancy, appear in the sera of patients with cancer as a result of shedding. ACKNOWLEDGMENTS The author wishes to express his gratitude to his colleagues Drs. I. N. Chou, S. Jaken, and E. Schroder for many helpful discussions; their work forms the basis of several of the hypotheses presented herein. H e is also grateful to Drs. I. N. Chou, W. Carney, and V. Zurawski for reviewing the manuscript and to Ms. Linda Parlee for invaluable help in preparation of manuscript.

REFERENCES Abramson, M., Huang, C.-C., Schilling, R. W., and Salome, R. G. (1975). Ann. Otol. 84, 158.

Aggeler, J., Kapp, L. N., and Werb, Z. (1978). J. CeI/ Eiol. 79, 10a. Alexander, P. (1974). Cancer Res. 34, 2077. Allan, D., Billah, M. M.,Finean, J. B., and Michell, R. H. (1976). Nature (London) 261, 58.

Allen, T. D., and Iype, P. T. (1976). In Vitro 12, 837. Anderson, D. R., Davis, J. L., and Carraway, K. L. (1977). J. B i d . Chem. 252, 6617. Anderson, J., Lafleur, L., and Melchers, F. (1974). Eur. J . Immunol. 4, 170. Aoki, N. (1974). J. Eiochem. 75, 731. Armelin, M. C. S., and Armelin, H. A. (1978). Proc. Narl. Acad. Sci. U.S.A. 75, 2805. Armerding, D., and Katz, D. H. (1974). J . Exp. Med. 140, 19. Armerding, D., Sachs, D. H., and Katz, D. H. (1974). J . Exp. Med. 140, 1717. Arthur, L. O., Bauer, R. F., Orme, L. S., and Fine, D. L. (1978). Virology 87, 266. Ash, J. F., and Singer, S. J. (1976). Proc. Natl. Acad. Sci. U.S.A. 73, 4575. Ash, J . F., Louvard, D., and Singer, S. J. (1977). Proc. Natl. Acad. Sci. U.S.A. 74, 5584. Astedt, B., and Holmberg, L. (1978). I n “Progress in Chemical Fibrinolysis and Thrombolysis” (J. F. Davidson, R. M. Rowan, M. M. Samama and P. C. Desnoyers, eds.), Vol. 3 p. 555. Raven, New York. Astrup, T . (1966). Fed. Proc. 25,42.

CELL SURFACE SHEDDING

181

Aswanikumar, S., Schiffmann, E., Corcoran, B. A., and Wahl, S. M. (1976). Proc. Natl. Accid. Sci. U . S . A . 73, 2439. Ausprunk, D. H., and Folkman, J. (1977). Microvas. Res. 14, 53. Autuori. F., Svensson, H.. and Dallner, G. (1975). J. Cell B i d . 67, 687. Baldwin, R. W., Price, M. R., and Robins, P. A. (1973). Br. J. Cancer 28 (Suppl. I), 37. Ball, G. V., and Goldman, L. N. (1976). Ann. Int. Med. 85, 464. Barkai, A., and DiCesare, J. L. (1975). Biochim. Biophys. Acta 398, 287. Bauduin, H., Stock, C., Vincent, D., and Grenier, J. F. (1975). J . Cell B i d . 66, 165. Bauer, Ch., Kottgen, E., and Reutter, W. (1977). Biochem. Biophys. Res. Commun. 76, 488. Baum, B. J., McDonald, J. A., and Crystal, R. G. (1977). Biochem. Biophys. Res. Commun. 79, 8. Baumann, H., and Doyle, D. (1978). J. Biol. Chem. 253, 4408. Becker, E. L. (1972). J. Exp. Med. 135, 376. Beers, W. H., Strickland, S . . and Reich, E. (1975). Cell 6, 387. Belkin, M., and Hardy, W. G. (1961).J . Biophys. Biochem. 9, 733. Bell, W. R., Pitney, W. R., and Goodwin, J. F. (1968). Lancet 1, 490. Belman, S., and Troll, W. (1974). Cancer Res. 34, 3446. Bennett, B. T., Debelak-Fehir, K. M., and Epstein, R. B. (1975). Cancer Res. 35, 2942. Ben-Sasson, Z . , Weiss, D. W., and Doljanski, F. (1974). J. Natl. Cancer Inst. 52, 405. Bensusan, H. B., Koh, T. L., Henry, K. G., Murray, B. A , , and Culp, L. A. (1978). Proc. Natl. Acud. Sci. U . S . A . 75, 5864. Berenblum, I. (1975).In “Cancer” (F. F. Becker, ed.). Vol. I , p. 323. Plenum, New York. Berl, S., Puszkin, S., and Nicklas, W. J . (1973). Science 17, 441. Berlin, R. D. (1975). Ann. N. Y . Acad. Sci. 253, 445. Bernacki, R. J. (1974). J. Cell. Physiol. 83, 457. Bernacki, R. J., and Kim, U. (1977). Science 195, 577. Besharse, J. C., Hollyfield, J. G., and Rayborn, M. E. (1977). J . Cell B i d . 75, 507. Biddison, W. E., and Palmer, J . C. (1977). Proc. Null. Acrid. Sci. U . S . A . 74, 329. Bigbee, M . L., and Jensen. R. J. (1978). Biochim. Biophys. Actci 540, 285. Birdwell, C. R., GoSpodarowicz. D., and Nicolson, G. L. (1978). Proc. Natl. Acad. Sci. U . S . A . 75, 3273. Bjorlin, G., Pandolfi, M., and Astedt, B. (1972). Experientia 28, 833. Black, P. H., Chou, I. N., and Roblin, R. 0. (1975). In ”Proceedings of the XIth International Cancer Congress, Vol. I , Cell Biology and Tumor Immunology” (P. Bucalossi, U. Veronesi and N. Cascinelli, eds.), p. 119. Excerpta Medica, Amsterdam. Blitz, A . L., and Fine, R. E. (1974). Proc. N u t / . Acad. Sci. U . S . A . 71, 4472. Blobel, G. (1977). In “International Cell Biology 1976-1977” (B. R. Brinkley and K. P. Porter, eds.), p. 318. The Rockefeller UNv. Press, New York. Blobel, G., and Dobberstein, B. (1975). J. Cell B i d . 67, 835. Blobel, G., and Sabatini, D. D. (1971). I n “Biomembrane” (L. A. Manson, ed.), Vol. 2, p. 193. Plenum, New York. Blumberg, P. M., and Robbins, P. W. (1975). Cell 6, 137. Blumenstock, F. A , , Saba, T. M., Weber, P., and Laffin, R. (1978). 1. B i d . Chem. 253, 4287. Bolognesi, D. P., and Langlois, A. J . (1975). Virology 68, 550. Bonavida, B. (1974).J. Immunol. 112, 926. Borek, C . , and Fenoglio, C. M. (1976). Cirncer Res. 36, 1325. Borle, A. B. (1973). Fed. Proc. 32, 1944. Bornstein, P., and Ash, J . F. (1977). Proc. Null. A m d . Sci. U . S . A . 74, 2480.

182

PAUL H. BLACK

Bosmann, H. B. (1972). Biochim. Biophys. Acia 264, 339. Bosmann, H. B. (l974a). Nature (London) 249, 144. Bosmann, H . B. (l974b). Biochim. Biophys. Acia 339, 438. Bosmann, H . B., and Hall, T. C. (1974). Proc. Nail. Acad. Sci. U.S.A. 71, 1833. Bosmann, H. B., Bieber, G. F., Brown, A. E., Case, K. R.,Gersten, D. M., Kimmerer, T. W., and Lione, A. (1973). Nature (London) 246, 487. Bosmann, H. B., Spataro, A. C., and Myers, M. W. (1975). Res. Commun. Chem. Paihol. Pharmacol. 12,499. Bourguignon, L. Y.,and Singer, S. J. (1977). Proc. Nail. Acad. Sci. U . S . A . 74, 5031. Bowen, J . C., and Baldwin, R. W. (1976). Transplanration 21, 213. Boynton, A. L., Whitfield, J. F., and Isaacs, R. J. (1975). J. Cell. Physiol. 87, 25. Bradley, M. 0. (1977). J . B i d . Chem. 252, 5310. Bradley, T. F., and Metcalf, D. (1966). Ausi. J . Exp. Biol. Med. Sci. 44, 287. Braslawsky, G. R., Yaackubowicz, M., Frensdorff, A., and Witz, I. P. (1976).J . Immunol. 116, 1571. Brassine, C., Coune, A., Nijs, M., and Tagnon, H. J. (1976). Thromb. Res. 8, 803. Brattig, N . , and Berg, P. A. (1976). Clin. Exp. Immunol. 25, 49. Bray, A. E., and Keast, D. (1975). Br. J . Cancer 31, 170. Bray, D. (1978). Nafure (London) 273, 265. Breisblatt, W., and Ohki, S. (1975). J. Membrane Biol. 23, 385. Breisblatt, W . , and Ohki, S. (1976). J. Membrane Biol. 29, 127. Brinkley, B. R . , Fuller, G. M., and Highfield, D. P. (1975). Proc. Nail. Acad. Sci. U . S . A . 72, 498 I . Broder, S . , and Waldmann, T. A. (1978). N . Engl. J . Med. 299, 1281. Brown, C. A., Hall, C. L., Long, J. C., Carey, K., Weitzman, S. A., and Aisenberg, A. C. (1978). Am. J. Med. 64, 289. Brown, K. T., and Flaming, D. G. (1978). Proc. Nail. Acad. Sci. U.S.A. 75, 1587. Bulkin, R. J . , and Hauser, R. (1973). Biochim. Biophys. Acia 326, 289. Burger, M. M. (1970). Nature (London) 227, 170. Burger, M. M., Bombik, B. M., Breckenridge, B. M., and Sheppard, J. R. (1972). Nafure (London) New Biol. 239, 161. Burk, P. G., and Hatcher, V. B. (1978). Clin. Res. 26, 567A. Burns, G. F., Baker, C. R., Cawley, J. C., and Hayhoe, F. G. J. (1977). J . Immunol. 119, 1279. Bussel, R. N., and Robinson, W. S. (1973). J. Virol. 12, 320. Butcher, F. R., and Goldman, R. H. (1972). Biochem. Biophys. Res. Commun. 48, 23. Bystryn, J.-C. (1977). J. Nail. Cancerlnsi. 59, 325. Calafat, J . , Hilgers, J., VanBlitterswijk, W. J., Verbeet, M.,and Hageman, P. C. (1976). J . Nail. Cancer Ins;. 56, 1019. Carter, S. B. (1965). Nature (London) 208, 1183. Carter, S. B. (1967). Nature (London) 213, 261. Castor, L. N. (1977). J. Cell. Physiol. 92, 457. Chang, E. H., Jay, F. T., and Friedman, R. M. (1978). Proc. Nail. Acad. Sci. U.S.A. 75, 1859. Chatterjee, S . K., and Kim, U. (1977). J. Nail. Cancer Insi. 58, 273. Chatterjee, S. K., and Kim, U . (1978). J. Nafl. Cancer I n s i . 61, 151. Chatterjee, S. K., Kim, U., and Bielat, K. (1976). Br. J. Cancer 33, 15. Chen, L. B., and Buchanan, J. M. (1975). Proc. Nail. Acad. Sci. U.S.A. 72, 1132. Chen, L. B., Gallimore, P. H., and McDougall, J. K. (1976). Proc. Nail. Acad. Sci. U.S.A. 73, 3570.

CELL SURFACE SHEDDING

183

Chen, L. B., Gudor, R. C., Sun, T.-T., Chen, A. B., and Mosesson, M. W. (1977). Science 197, 776. Chertow, B. S., Buschmann, R. J., and Henderson, W. J. (1975). Lab. Invest. 32, 190. Chew, E. C., and Wallace, A. C. (1976). Cancer Res. 36, 1904. Chew, E. C., Josephson, R. L., and Wallace, A. R. (1976). I n “Fundamental Aspects of Metastasis” (L. Weiss, ed.), p. 121. Elsevier, Amsterdam. Chou, I. N., and Black, P. H. (1979). J . Cell. Physiol. 100, 547. Chou, I. N., Black, P. H., and Roblin, R. (1974a). Proc. Natl. Acad. Sci. U.S.A. 74, 1748. Chou, I. N., Black, P. H., and Roblin, R. 0. (1974b). Nature (London) 250, 739. Chou, I. N., Black, P. H., and Roblin, R. (1974~).I n “Control of Proliferation in Animal Cells” (B. Clarkson and R. Baserga, eds.), p. 339. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York. Chou, 1. N., ODonnell, S. P., Black, P. H., and Roblin, R. 0. (1977a). J . Cell. Physiol. 91, 31. and Black, P. H. (l977b). J. B i d . Chem. 252, 6256. Chou, I. N., Roblin, R. 0.. Christman, J. K., Acs, G., Silagi, S., and Silverstein, S. C. (1975). I n “Proteases and Biological Control“ (E. Reich, D. Rifkin and E . Shaw, eds.), p. 827. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York. Christman, J. K., Silverstein, S. C., and Acs, G. (1977). I n “Research Monographs in Cell and Tissue Physiology” (A. J. Barret, ed.), p. 90. Elsevier, Amsterdam. Chung, A. E., Freeman, I. L., and Braginski, J. E. (1977). Biochem. Biophys. Res. Commun. 79, 859. Cifone, M., and Defendi, V. (1974). Nature (London) 252, 151. Cittadini, A., Bossi, D., Rosi, G., Wolf, F., and Terranova, T . (1977). Biochim. Biophys. Acta 469, 345. Claflin, A. J., Vesely, D. L., Hudson, J. L., Bagwell, C. B., Lehotay, D. C., Lo, T. M., Fletcher, M. A., Block, N. L., and Levey, G. S. (1978). Proc. Natl. Acad. Sci. U . S . A . 75, 989. Codington, J. F., Klein, G., Cooper, A. G., Lee, N., Brown, M. C., and Jeanloz, R. W. (1978). J. Nail. Cancer Inst. 60, 81 1. Cohen, R., Pacifici, M.,Rubinstein, N., Biehl, J., and Holtzer, H. (1977). Nature (London) 266, 538. Cohn, Z. A. (1978). J. Immunol. 121, 813. Collard, J. G., and Smets, L . A. (1974). xp. Cell Res. 86, 75. Collard, J. G., and Temmink, J. H. M. (1976). J . Cell B i d . 68, 101. Collen, D., Billiau, A., Edy, T., and DeSomer, P. (1977). Biochim. Biophys. Acta 499, 194. Colvin, R. B., Gardner, P. I., Roblin, R. O., Verderberg, E. L., and Mosesson, M. W. (1979). Lab. Invest. 41, 464. Coman, D. R. (1944). Cancer Res. 4,625. Cone, R. E., Marchalonis, J . J., and Rolley, R. T. (1971). J. Exp. Med. 134, 1373. Cook, G. M. W. (1977). I n “Cell Surface Reviews” (G. Poste and G. L. Nicolson, eds.), Vol. 4, p. 85. Elsevier, Amsterdam. Cooper, A. G., Codington, J. F., and Brown, M. C. (1974). Proc. Natl. Acad. Sci. U . S . A . 71, 1224. Costanza, M. E., Pinn, V., Schwartz, R. S., and Nathanson, L. (1973). N . Engl. J . Med. 289, 520. Cotropia, J . P., Gutterman, J. U.,Hersh, E. M., and Mavligit, G. M. (1977). I n t . J . Cancer 20, 520. Critchley, D. R., Wyke, J. A.. and Hynes, R. 0. (1976). Biochim. Eiophys. Acra 436, 335.

184

PAUL H . BLACK

Croker, B. P., McConahey, P. J., Murphy, E. D., and Dixon, F. J. (1977).J . Natl. Cancer Inst. 59, 199. Cullen, S. E., Freed, J. H., and Nathenson, S. G. (1976). Transplant. Rev. 30, 236. Cullis, P. R., and Hope, M. J. (1978). Nature (London) 271, 672. Culp, L. A. (1976). J. Supramol. Struct. 5, 239. Culp, L . A. (1977). Biochemistry 15, 4094. Culp, L. A., and Black, P. H . (1972a).J. Virol. 9, 611. Culp, L. A., and Black, P. H. (1972b). Biochemistry 11, 2161. Culp, L. A . , Grimes, W. J., and Black, P. H. (1971). J. Cell Biol. 50, 682. Culp, L. A., Rollins, R. J., Buniel, J., and Hitri, S. (1978). J . Cell Biol. 79, 788. Currie, G. (1976). Biochim. Biophys. Acra 458, 135. Davey, G. C., Currie, G. A., and Alexander, P. (1976). Br. J . Cancer 33, 9. Davidson, J. F., McNicol, G. P., Frank, G. L.,Anderson, T. J., and Douglas, A. S. (1969). Br. Med. J . 1, 88. DeLorenzo, R. J., and Freedman, S. D. (1977). Biochem. Biophys. Res. Commun. 77, 1036. DeLorenzo, R. J., and Freedman, S. D. (1978). Biochem. Biophys. Res. Commun. 80, 183. Deman, J., VanVaerenbergh, P. M., and Joos, P. (1971). Eur. J . Cancer 7, 317. dePetris, S. (1978). Nature (London) 272, 66. DeRubertis, F. R., and Craven, P. A. (1977a). Biochim. Biophys. Acta 499, 337. DeRubertis, F. R . , and Craven, P. A. (1977b). J . Natl. Cancer Inst. 59, 1741. Dessau, J., Jilek, F., Adelmann, B. C., and Hormann, H. (1978). Biochim. Biophys. Actu 533, 227. Desmyter, J., DeGroote, J., Desmet, V. J., Billau, A., Ray, M. D., Bradburne, A. F., Edy, V. G., and DeSomer, P. (1976). Luncet 2, 645. Diamond, L., O’Brien, T. G., and Rovera, G. (1977). Nature (London) 269, 247. Diener, E., and Feldman, M. (1972). Transplant. Rev. 8, 76. Dierich, M. P., Wilhelmi, D., and Till, G. (1977). Nature (London) 270, 351. Diggelmann, H., Eisenman, R., Zucco, F., and Gallis, B. (1976). Experientiu 32, 791. Doljanski, F., and Kapeller, M. (1976). J. Theor. B i d . 62, 253. Dorval, G., Witz,.I. P., Klein, E., and Wigzell, H. (1976). Inr. J. Cuncer 17, 109. Dougherty, T. F., Stevens, W., and Schneebeli, G. L. (1973). Rec. Proy. Horm. Res. 29, 287. Douglas, W. W. (1974). Biochem. SOC. Symp. 39, 1. Dnedger, P. E., and Blumberg, P. M. (1977). Cancer Res. 37, 3257. Dulbecco, R., and Elkington, J. (1975). Proc. Natl. Acad. Sci. U . S . A . 72, 1584. Dunham, J. S ., and Hynes, R. 0. (1978). Biochim. Biophys. Acta 506, 242. Duzgunes, J., and Ohki, S. (1977). Biochim. Biophys. Acta 467, 301. Dvorak, H . F., Orenstein, N. S., Rypyse, J., Colvin, R. B., and Dvorak, A. (1978). J. Immunol. 120, 766. Dvorak, H. F., Orenstein, N. S., Carvalho, A. C., Churchill, W. H.,Dvorak, A. M., Galli, S. J., Feder, J., Bitzer, A . M., Rypyse, J . , and Giovinco, P. (1979).J . Immunol. 122, 166. Eaves, A. C., and Bruce, W. R. (1974). Cell Tissue Kinet. 7, 19. Edelman, G. M. (1976). Science 192, 218. Edelman, G. M., Yahara, I., and Wang, J. L. (1973). Proc. Narl. Acad. Sci. U . S . A . 70, 1442. Edelson, P. J., and Erbs, C. (1978). J. Immunol. 120, 1532. Edwards, J. C., and Howell, S. L. (1973). FEES Lett. 30, 89. Ehrlich, H . P., Ross, R., and Bornstein, P. (1974). J . Cell Biol. 62, 390.

CELL SURFACE SHEDDING

185

Eisenman, R. N., and Vogt, V. M. (1978). Biochim. Biophys. Acta 473, 187. Elgsaeter, A., Shotton, D. M., and Branton, D. (1976). Biochim. Biophys. Acra 426, 101. Embleton, M. J . (1976). f n t . J . Cancer 18, 622. Engvall, E., and Ruoslahti, E. (1977). f n t . J. Cancer 20, I . Engvall, E., Ruoslahti, E., and Miller, E. J . (1978). J. Exp. Med. 147, 1584. Epstein, M. A . , and Hoit, S. J. (1963). J . Cell Biol. 19, 325. Erb, P., Feldman, M.,and Hogg, N. (1976). Eur. J. fmmunol. 6, 365. Esselman, W. J . , and Miller, H. C. (1977). J. fmmunol. 119, 1994. Estensen, R. D., and Plagemann, P. G. W. (1972). Proc. Natl. Acad. Sci. U . S . A . 69, 1430. Estensen, R. D., Hill, H. R., Quie, P. G., Hogan, N., and Goldberg, N. D. (1973). Nature (London) 245,458. Estensen, R. D., White, J. G., and Holmes, B. (1974). Nature (London) 248, 347. Evans, L. H., Dresler, S., and Kabat, D. (1977). J. Virol. 24, 865. Farber, D. B., and Lolley, R. N. (1976). J. Cyclic Nucl. Res. 2, 139. Farquhar, M. G. (1978). J . Cell Biol. 77, R35. Feldmann, M. (1972). J . Exp. Med. 136, 737. Fidler, I . J . (1973). Eur. J . Cancer 9, 223. Fidler, I. J., and Kripke, M. L. (1977). Science 197, 893. Fischer, A . (1925). Arch. Entw. Mech. Arg. 104, 210. Fisher, P. B., Weinstein, I. B., Eisenberg, D., and Ginsberg, H. S. (1978). Proc. Natl. Acad. Sci. U . S . A . 75, 231 1. Fishman, M. L., Oberc, M. A., Hess, H. H., and Engel, W. K. (1977). Exp. Eye Res. 24, 341. Fishman, P. H., and Brady, R. 0. (1976). Science 194, 906. Flanagan, J . , and Koch, G. L. E. (1978). Nature (London) 273, 278. Fodge, D. W., and Rubin, H . R. (1975). Narure (London) 257,804. Fodor, E. J . B., Ako, H., Walsh, K. A. (1975). Biochemistry 14, 4923. Folkman, J . (1975). In “Cancer” (F. F. Becker, ed.), VoL 3, p. 355. Plenum, New York. Franke, W. W., Monre, D. J . , Deumling, B., Cheetham, R. D., Kartenbeck, J., Jarasch, E. D., and Zentgraf, H. W. (1971). Z. Naturforsch. 26b, 1031. Freimuth, W., Esselman, W. J., and Miller, H. C. (1978).J. fmmunol. 120, 1651. Fridman, W. H., Nelson, R. A., Jr., and Liabeuf, A. (1974). J . fmmunol. 113, 1008. Fuchs, P., Spiegelstein, M.,Haimsohn, M., Gitelman, J., and Kohn, A. (1978). J. Cell. Physiol. 95, 223. Fujimoto, S., Greene, M. I., and Sehon, A. H . (1976). J . Immunol. 116, 791. Furcht, L. T., and Scott, R. E. (1974). Exp. CeN Res. 88, 311. Furcht, L. T., and Scott, R. E. (1975). Exp. Cell Res. 96, 271. Gabbiani, G., Chaponnier, C., Zume, A., and VassaUi, P. (1977). Nature (London) 269, 697. Gahmberg, C. G., and Hakomuri, S.-I. (1973). Proc. N u / / .Acad. Sci. U . S . A . 70, 3329. Gallimore, P. H., McDougall, J. K., and Chen, L. B. (1977). Cell 10, 669. Gallin, J . I., Sandler, J. A , , Clyman, R. I., and Manganiello, V. C., and Vaughan, M. (1978).J . fmmunol. 120, 492. Gazitt, Y., Ohad, I., and Loyter, A. (1976). Biochim. Biophys. Acta 436, I . Gery, I., Gershon, R. K., and Waksman, B. H. (1972). J . Exp. Med. 136, 128. Gillette, R. W., and Boone, C. W. (1975). Cancer Res. 35, 3774. Gillis, S., Ferm, M. M., Ou, W., and Smith, K. A . (1978). J. fmmunol. 120, 2027. Glaser, M., Kirchner, H., and Herberman, R. B. (1975). f n t . J . Cancer 16, 384. Glaser, R., Mumaw, V., Farrugia, R., and Munger, B. (1977). Cancer Res. 37,4420. Glucksmann, A. (1948). Br. J . Radio/. 21, 559.

186

PAUL H. BLACK

Goetzl, E. J. (1975). Immunology 29, 163. Goldberg, A. L., Howell, E. M., Li, J. B., Martell, S. B., and Prouty, W. F. (1974). Fed. Proc. 33, 1 1 12. Goldman, L. I., Flaxrnan, B. A., Wernick, G., and Zabriskie, J. B. (1974). Surgery 76, 50. Goldman, R. D. (1971). J . Cell B i d . 51, 752. Goldman, R. D., and Follett, E. A. C. (1969). Exp. Cell Res. 57, 263. Goldstein, I . M., Lind, S., Hoffstein, S., and Weissmann, G. (1977). J . Exp. Med. 146, 483. Golub, S. H. (1975). In “Cancer 4. Biology of Tumors: Surfaces, Immunology, and Comparative Pathology” (F. F. Becker, ed.), p. 259. Plenum, New York. Gordon, S. (1976). Ann. N . Y . Acad. Sci. 278, 176. Gordon, S . (1978). Fed. Proc. 37, 2754. Gordon, S., and Werb, Z. (1976). Proc. Natl. Acad. Sci. U.S.A. 73, 872. Gordon, S., Unkeless, J. C., and Cohn, Z. A. (1974). 1.Exp. Med. 140, 995. Gospodarowicz, D., and Moran, J. S . (1976). Annu. Rev. Biochem. 45, 531. Goth, A., Adams, H. R., and Knoohuizen, M. (1971). Science 173, 1034. Goto, M., Kataoka, Y., Kimura, T., Goto, K., and Sato, H . (1973). Exp. Cell Res. 82,367. Granelli-Piperno, A., Vassalli, J.-D., and Reich, E. (1977). J. Exp. Med. 146, 1693. Grant, C. W. M., and McConnell, H. M. (1974). Proc. Natl. Acad. Sci. U.S.A. 71,4653. Gratzl, M., and Dahl, G. (1976). FEES Lett. 62, 142. Gratzl, M., Dahl, G., Russell, J. T., and Thorn, N. A. (1977). Biochim. Biophys. Acta 470, 45. Green, D., Ryan, C., Malandruccolo, N., and Nadler, H. L. (1971). Blood 37, 47. Greenberg, H. B., Pollard, R. B., Lutwick, L. I., Gregory, P. B., Robinson, W. S. and Merigan, T. C. (1976). N . Engl. J . Med. 295, 517. Greene, M. I., Dorf, M. E., Pierres, M., and Benacerraf, B. (1977a). Proc. Natl. Acad. Sci. V . S . A . 74, 5118. Greene, M . I., Pierres, A., Dorf, M. E., and Benacerraf, B. (1977b). J. Exp. Med. 146, 293. Greengard, P. (1978). Science 199, 146. Grosser, N., and Thornson, D. M. P. (1975). Cancer 35, 2571. Habener, J. F. (1977). Biochemistry 16, 3910. Hagins, W. A., and Yoshikami, S. (1974). Exp. Eye Res. 18, 299. Hall, J. E., and Simon, S. A. (1976). Biochim. Biophys. Acta 436, 614. Halper, J., Fu, S. M., Wang, C. Y., Winchester, R., and Kunkel, H. G. (1978).J . Immunol. 120, 1480. Hamilton, J., Vassalli, J.-D., and Reich, E. (1976). J . Exp. Med. 144, 1689. Harker, L. A., and Slichter, S . J. (1972). N . Engl. J . Med. 287, 999. Harris, A. (1973). Exp. Cell Res. 77, 285. Harris, A. K., (1976). Nature (London) 263, 781. Harris, E. D., Jr. (1976). Arthritis Rheum. 19, 68. Harris, E. D., Jr., and Krane, S. M. (1971). Arthritis Rheum. 14, 669. Harris, E. D., Jr., DiBona, D. R., and Krane, S. M. (1%9). J . Clin. Invest. 48, 2104. Hart, C. A., Fisher, D., Hallinan, T., and Lucy, J. A. (1976). Biochem. J . 158, 141. Hashimoto, K., Yamanishi, Y., and Dabbous, M. K. (1972). Cancer Res. 32, 2561. Hashimoto, K., Yamanishi, Y., Maeyens, E., Dabbous, M. K., and Kanzaki, T. (1973). Cancer Res. 33, 2790. Haslam, R. J., and Lynham, J. A. (1977). Biochem. Biophys. Res. Commun. 77, 714. Hatcher, V. B., Oberman, M. S., Wertheim, M. S., Rhee, C. Y., Tsien, G., and Burk, P. G . (1977). Biochem. Biophys. Res. Commun. 76, 602.

CELL SURFACE SHEDDING

187

Hatcher, V. B., Wertheim, M. S., Rhee, C. Y., Tsien, G., and Burk, P. G. (1976). Biochim. Biophys. Acta 451, 499. Hauser, H., Phillips, M. C., and Barratt, M. D. (1975). Biochim. Biophys. Acta 413, 341. Hayman, M. (1978). Virology 85, 475. Hellstrom, I., and Hellstrom, K. E. (1969). Int. J . Cancer 4, 587. Hellstrom, K. E., and Hellstrom, I. (1970). Annu. Rev. Microbiol. 24, 373. Hellstrom, K. E., and Hellstrom, I. (1978). Proc. Natl. Acad. Sci. U . S . A . 75, 436. Hendil, K. B. (1977). J. Cell. Physiol. 92, 353. Herberman, R. B., Aoki, T., and Nunn, M. E. (1973). J . Natl. Cancer Inst. 50, 481. Herman, I. H., Musgrave, D. S., and Dennis, M. V. (1977). Arthritis Rheum. 20, 922. Hershko, A., Mamont, P., Shields, R., and Tomkins, G. M. (1971). Nature (London) New Biol. 232, 206. Herzog, V., and Farquhar, M. G. (1977). Proc. Natl. Acad. Sci. U . S . A . 74, 5073. Hilgard, P., and Thornes, R. D. (1976). Eur. J . Cancer 12, 755. Hirano, T., Kuritani, T., Kishimoto, T., and Yamamura, Y.(1977). J . Imrnunol. 119, 1235. Hirsch, M. S., and Black, P. H. (1974). Adv. Virus Res. 19, 265. Hitchcock, S. E. (1977). J . Cell Biol. 74, 1. Hoffmann, M. K., Koenig, S., Mittler, R. S., Oettgen, H. F., Ralph, P., Galanos, C., and Hammerling, U. (1979). J. Imrnunol. 122, 497. Hogg, N. M. (1974). Proc. Natl. Acad. Sci. U . S . A . 71,489. Hollyfield, J. G., Besharse, J. C., and Rayborn, M. E. (1977). J. Cell Biol. 75, 490. Horwitz, A. L., Kelman, J. A., and Crystal, R. G. (1976). Nature (London) 264, 772. Howie, S., and Feldmann, M. (1978). Nature (London) 273, 664. Huang, C.-C., Tsai, C.-M., and Canellakis, E. S. (1974). Biochim. Biophys. Acta 332, 59. Hubbard, A. L., and Cohn, Z. A. (1975a). J. Cell Biol. 64, 438. Hubbard, A. L., and Cohn, Z. A. (1975b). J . Cell Biol. 64, 461. Hughes, R. C., Sanford, B., and Jeanloz, R. W. (1972). Proc. Natl. Acad. Sci. U . S . A . 69, 942.

Hull, R. N., Cherry, W. R., Huseby, R. M., and Clavin, S. A. (1977). Thromb. Res. 10, 669.

Hunter, E., Hayman, M. J., Rongey, R. W., and Vogt, P. K. (1976). Virology 69, 35. Hynes, R. 0. (1973). Proc. Natl. Acad. Sci. U . S . A . 70, 3170. Hynes, R. 0. (1974). Cell 1, 147. Hynes. R. 0.. and Bye, J. M. (1974). Cell 3, 113. Hynes, R. O., and Destree, A. (1977). Proc. Natl. Acad. Sci. U . S . A . 74, 2855. Hynes, R. 0..and Pearlstein, E. S. (1976). J . Supramol. Struct. 4, I . Hynes, R. O., Wyke, J. A., Bye, J. M., Humphryes, K. C., and Pearlstein, E. S. (1975). I n “Proteases and Biological Control” (E. Reich, D. Rifkin and E. Shaw, eds.), p. 931. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York. Ikeman, R. L., Sullivan, A. K., Kositsky, R., Bartok, K., and Jerry, L. M. (1978). Nature (London) 272, 267. Ip, C., and Dao, T . (1978). Cancer Res. 38, 723. Jackson, R. C., and Blobel, G. (1977). Proc. Natl. Acad. Sci. U . S . A . 74, 5598. Jacobs, B. B. (1976). Transplantation 21, 433. Jacobs, B. B., and Uphoff, D. E. (1974). Science 185, 582. Jaken, S., and Black, P. H. (1979a). Proc. Natl. Acad. Sci. U . S . A . 76, 246. Jaken, S., and Black, P. H. (1979b). J . Cell. Biol. (in press). Jaken, S., and Mason, M. (1978). Proc. Natl. Acad. Sci. U . S . A . 75, 1750. Jamasbi, R. J., Nettesheim, P., and Kennel, S. J. (1978). I n t . J . Cancer 21, 387. Jamieson, J. D., and Palade, G. E. (1971). J . Cell Biol. 48, 503.

188

PAUL H . BLACK

Jamieson, J. D., and Palade, G. E. (1977). I n “International Cell Biology, 1976-1977” (B. R. Brinkley and K. R. Porter, eds.), p. 308. The Rockefeller Univ. Press, New York. Jockusch, B. M., Burger, M. M., DaPrada, M., Richard, J. G., Chaponnier, C., and Gabbiani, G. (1977). Nature (London) 270, 628. Johnson, H. M., Blalock, J. E., and Baron, S. (1977). Cell. lmmunol. 33, 170. Johnson, J. D., and Epel, D. (1975). Proc. Natl. Acad. Sci. U.S.A. 72, 4474. Jones, G. (1973). J. Immunol. 110, 1526. Jones, P. A., Laug, W. E., and Benedict, W. F. (1975). Cell 6, 245. Jones, P. A., Laug, W. E., Gardner, A., Nye, C. A., Fink, L . M., and Benedict, W. F. (1976). Cancer Res. 36, 2863. Judah, J. D., and Quinn, P. S. (1978). Nature (London) 271, 384. Kantor, F. S. (1975). N . Engl. J. Med. 292, 629. Kapeller, M., Gal-Oz, R., Grover, N. B., and Doljanski, F. (1973). Exp. Cell Res. 79, 152. Kaplan, A. P., Goetzl, E. J., and Austen, K. F. (1973). J. Clin.Invest. 52, 2591. Kaplan, J., and Moskowitz, N. (1975a). Biochim. Biophys. Acta 389, 290. Kaplan, J., and Moskowitz, N. (1975b). Biochirn. Biophys. Acta 389, 306. Karasaki, S., Simard, A., and delamirande, G. (1977). Cancer Res. 37, 3516. Karnovsky, M. L., and Lazdins, J. K. (1978). J. Immunol. 121, 809. Katz, F. N., Rothman, J. E., Knipe, D. M., and Lodish, H. F. (1977).J. Supramol. Struct. 7, 353. Keller, H. U., Wilkinson, P. C., Abercrombie, M., Beckers, E. L., Hirsch, J. G., Miller, M. E., Ramsey, W. S., and Zigmond, S. H. (1977). Clin. Exp. Immunol. 27, 377. Kellner, B., and Sugar, J. (1967). I n “Endogenous Factors Influencing Host-Tumor Balance” (R. W. Wissler, T. L. Doa and S. Wood, Jr., eds.), p. 239. Chicago Univ. Press, Chicago. Kennel, S . J. (1977). J. Virol. 22, 168. Kessel, D., Sykes, E., and Henderson, M. (1977). J. Natl. Cancer I n s t . 59, 29. Killion, J . J., Wallenbrock, M. A., Rogers, J. A., 111, Sansing, W. A., and Cantrell, J. L. (1976). Nature (London) 261, 54. Kim, U., Baumler, A., Carruthers, C., and Bielat, K. (1975). Proc. Natl. Acad. Sci. U . S . A . 72, 1012. Kimura, H., Tateishi, H., and Ziff, M. (1977). Arthritis Rheum. 20, 1085. Klagsburn, M. (1976). Biochim. Biophys. Acta 451, 170. Klein, G. (1975). Harvey Lect. 69, 71. Klenk, H.-D., Rott, R., and Orlich, M. (1977). J . Gen. Virol. 36, 151. Kletzien, R. F., Miller, M. R., and Pardee, A. B. (1977). Nature (London) 270, 57. Klimetzek, V., and Sorg, C. (1977). Eur. J. Immunol. 7, 185. Kloppel, T. M., Keenan, T. W., Freeman, M. J., and Morre, D. J. (1977). Proc. Natl. Acad. Sci. U.S.A. 74, 3011. Knecht, M. E., and Lipkin, G. (1977). Exp. Cell Res. 108, IS. Knipe, D. M., Baltimore, D., and Lodish, H. F. (1977a).J. Virol. 21, 1128. Knipe, D. M., Lodish, H . F., and Baltimore, D. (1977b). J . Virol. 21, 1121. Koch, G. L. E., and Smith, M. J. (1978). Nature (London) 273, 274. Koopman, W. J., Farrar, J. J., Oppenheim, J. J., Fuller-Bonar, J., and Dougherty, S. (1977). J . Immunol. 119, 55. Koopman, W. J., Farrar, J. J., and Fuller-Bonar, J. (1978). Cell. Immunol. 35, 92. Kornfeld, S., and Ginsburg, V. (1966). Exp. Cell Res. 41, 592. Kosower, N. S . , Kosower, E. M., and Wegman, P. (1975). Biochirn. Biophys. Acra 401, 530.

CELL SURFACE SHEDDING

189

Kosower, E. M., Kosower, N . S . , and Wegman, P. (1977). Biochim. Biophys. Acta 471, 311. Kottgen, E., Reutter, W., and Gerok, W. (1976). Biochem. Biophys. Res. Commun. 72,61. Koza, E. P., Wright, T. E., and Becker, E. L. (1975).Proc. Soc. Exp. Biol. Med. 149,476. Kraemer, P. M . ( 1967).J . Cell. Physiol. 69, 199. Kraemer, P. M., and Tobey, R. A. (1972). J. Cell Biol. 55, 713. Krane, S. M . (1974). Arthritis Rheum. 17, 306. Krishan, A . , and Frei, F. (1975). Cancer Res. 35, 497. Kuettner, K. E., Soble, L., Croxen, R. L., Marczynska, B., Hiti, J., and Harper, E. (1977). Science 1%, 653. Kuppers, R. C., and Henney, C. S. (1977). J. Immunol. 118, 71. Lacy, P. E., Klein, N. J., and Fink, C. J. (1973). Endocrinology 92, 1458. Laki, K., and Yancy, S . T. (1968). In “Fibrinogen” (K. Laki, ed.), p. 359. Marcel Dekker, New York. Lamont, J. T., Perrotto, J. L., Weiser, M. M., and Isselbacher, K. J. (1974). Proc. Natl. Acad. Sci. U . S . A . 71, 3726. Lamont, J. T., Gammon, M. T., and Isselbacher, K. J. (1977). Proc. Natl. Acad. Sci. U . S . A . 74, 1086. Laug, W. E., Jones, P. A., and Benedict, W. F. (1975). J . Nail. Cancer Inst. 54, 173. Lawson, D., Raff, M. C., Gomperts, B., Fewtrell, C., and Gilula, N. B. (1977). J. Cell Biol. 72, 242. Lazarides, E., and Weber, K. (1974). Proc. Natl. Acad. Sci. U . S . A . 71, 2268. Ledbetter, J . , and Nowinski, R. C. (1977). J. Virol. 23, 315. Lee, L.-S., and Weinstein, I. B. (1978). Nature (London) 274, 6%. Lin, H.-S., and Stewart, C. C. (1974). J . Cell. Physiol. 83, 369. Lindahl, P., Leary, P., and Gresser, 1. (1973). Proc. Nail. Acad. Sci. U . S . A . 70, 2785. Lindahl, P., Gresser, I., Leary, P., and Tovey, M. (1976). Proc. Natl. Acad. Sci. U . S . A . 73, 1284. Lindahl-Magnusson, P., Leary, P., and Gresser, I. (1971). Proc. Soc. Exp. Biol. Med. 138, 1044. Linder, E., Vaheri, A., Ruoslahti, E., and Wartiovaara, J. (1975). J . Exp. Med. 142, 41. Liotta, L. A . , Kleinerman, J., Catanzaro, P., and Rynbrandt, D. (1977). J . Natl. Cancer Inst. 58, 1427. Lipesky, P. E., and Rosenthal, A. S. (1975). J . Exp. Med. 141, 138. Little, S. P., and Huang, A. S. (1977). Virology 81, 37. Little, S. P., and Huang, A. S. (1978). J . Virol. 27, 330. Lloyd, C. W., and Cook, G. M. W. (1974). J . Cell Sci. 15, 575. Lolley, R. N., and Farber, D. B. (1976a). Ann. Opthalmol. 8, 469. Lolley, R. N., and Farber, D. B. (1976b). Exp. Eye Res. 22, 477. Lolley, R. N., Brown, B. M., and Farber, D. B. (1977). Biochem. Biophys. Res. Commun. 78, 572. Long, J. C., Hall, C. L., Brown, C. A , , Stamatos, C., Weitzman, S. A , , and Carey, K. (1977). N . EngI. J . Med. 297, 295. Loor, F., Forni, L., and Pernis, B. (1972). Eur. J. Immunol. 2, 203. Lopez, L. R., Vatter, A. E., and Talmage, D. W. (1977). J . Immunol. 119, 1668. Lopez, M. J . , and Thomson, D. M. P. (1977). In!. J . Cancer 20, 834. Loskutoff, D. J., and Edgington, T. S. (1977). Proc. Natl. Acad. Sci. U . S . A . 74, 3903. Loughbridge, L. W., and Lewis, M. G. (1971). Lancet 1, 256. Loyter, A., Ben-Zaquen, R., Marash, R., and Milner, Y. (1977). Biochemistry 16, 3903.

190

PAUL H. BLACK

Lucy, J. A. (1970). Nature (London) 227, 815. Luft, J . H . (1971). Anat. Rec. 171, 347. Luftig. R. B., and Yoshinaka, Y . (1978). J . Virol. 25, 416. Lundgren, E., and Roos, G. (1976). Cancer Res. 36, 4044. Lunney, J . , and Ashwell, G. (1976). Proc. Nail. Acad. Sci. U . S . A . 73, 341. Lutz, H. U., McMillan, P., and Wehrli, E. (1977). J. Cell Biol. 74, 389. Macarak, E. J., Kirby, E., Kirk, T., and Kefalides, N. A. (1978). Proc. Narl. Acud. Sci. U.S.A. 75, 2621. Mackanass, G. B. (1964). J. Exp. Med. 120, 105. McDougal, J. S., Cort, S. A., and Gordon, D. S. (1977). J. Immunol. 119, 1933. MacLennan, I . C. M. (1972). Clin. Exp. Immunol. 10, 275. McNutt, N . S . , Culp, L . A., and Black, P. H. (1971). J. Cell Biol. 50, 691. McNutt, N. S ., Culp, L. A., and Black, P. H. (1973). J. Cell B i d . 56, 412. Maeda, Y . , Kim. J., Mekada, E., Shiowkawa, Y., and Okada, Y. (1977). Exp. Cell Res. 108, 95. Majno, G., Shea, S. M., and Leventhal, M. (1969). J. Cell Biol. 42, 647. Malaise-Lagge, F., Greider, M. H., Malaisse, W. J., and Lacy, P. E. (1971). J . Cell Biol. 49, 530. Manery, J. F. (1966). Fed. Proc. 25, 1804. Mann, D. B., Abelson, L., Harris, S., and Amos, D. B. (1976). Nature (London) 259, 145. Manor, Y., Treves, A. J., Cohen, I. R., and Feldman, M. (1976). Transplantarion 22, 360. Maynard, J . R., Heckman, C. A., Pitlick, F. A., and Nemerson, Y. (1975). J. Clin. Invest. 55, 814. Maynard, J. R., Freyer, B. E., Stemerman, M. B., and Pitlick, F. A. (1977). Blood 50, 387. Meedel. T. H., and Levine, E. M. (1978). J . Cell. Physiol. 94, 229. Meister, A., and Tate, S. S. (1976). Annu. Rev. Biochem. 45, 559. Melcher, U., Eidels, L., and Uhr, J. W. (1975). Nature (London) 258, 434. Melchers, F., and Cone, R. E. (1975). Eur. J. Immunol. 5 , 234. Mikkelsen, R. B. (1978). Prog. Exp. Tumor Res. 22, 123. Mikulski, S. M., Billing, R., and Terasaki, P. I. (1977). J . Narl. Cancer Inst. 58, 1485. Miller, H . C., and Esseiman, W. J. (1975). J. Immunol. 115, 839. Miller, S. C., Hay, E. D., and Codington, J. F. (1977). J. Cell Biol. 72, 511. Molenaar, J. L., vanGalen, M., Hannema, A. J., Zeijelmaker, W., and Pondman, K. W. (1977). Eur. J . Immunol. 7, 230. Molnar, J., Teegarden, D. W., and Winzler, R. J. (1965). Cancer Res. 25, 1860. Moore, L., and Pastan, I. (1977). J. Cell. Physiol. 91, 289. Mooseker, M. S., and Tilney, L. G. (1975). J. Cell Biol. 67, 725. Moroney, J . , Smith, A., Tomei, L. D., and Wenner, C. E. (1978). J. CeN. Physiol. 95, 287. Morre, D. J. (1977). In “Cell Surface Reviews” ( G . Poste and G. L. Nicolson, eds.), Vol. 4, p. I . Elsevier, Amsterdam. Morrison, T. G., and Lodish, H. F. (1975). J . Biol. Chem. 250, 6955. Morrison, T. G . , and McQuain, C. 0. (1978). J . Virol. 26, 1 15. Mosesson, M. W., Colman, R. W., and Sherry, S. (1968). N . Engl. J . Med. 273, 815. Mosher, D. F., and Vaheri, A. (1978). Exp. Cell Res. 112, 323. Moskal, J . R., Gardner, D. A., and Basu, S. (1974). Biochem. Biophys. Res. Cornrnun. 61, 751.

Munro, A. J., and Taussig, M. J . (1975). Nature (London) 256, 103. Murray, E., Ruygrok, S., Milton, G. W., and Hersey, P. (1978). Inr. J . Cancer 21, 578. Murray, R. L., and Dubin, M. W. (1975). J. Cell Biol. 64, 705. Nagai, Y., and Klenk, H.-D. (1977). Virology 77, 125.

CELL SURFACE SHEDDING

191

Nagy, B., Ban, J., and Brdar, B. (1977). In!. J. Cancer 19, 614. Naso, R. B., and Brown, R. L. (1977). Virology 82, 247. Nelson, K., Pollack, S. V., and Hellstrom, K. E. (1975). I n t . J . Cancer 16, 932. Nemerson, Y. (1%6). Biochemisrry 5, 601. Nepom, J . T., Hellstrom, I., and Hellstrom, K. E . (1976). J. Immunol. 117, 1846. Nepom, J. T., Hellstrom, I . , and Hellstrom, K. E. (1977). Proc. Natl. Acad. Sci. U.S.A: 74,4605.

Neurath, H., and Walsh, K. A. (1976). Proc. Natl. Acad. Sci. U.S.A. 73, 3825. Nicolson, G. L. (1976a). Biochim. Biophys. Acta 457, 57. Nicolson, G. L. (1976b). Biochim. Biophys. Acta 458, 1. Nicolson, G. L., and Painter, R. G. (1973). J . Cell Biol. 59, 395. Nicolson, G. L., Poste, G., and Ji, T. H. (1977). I n “Dynamic Aspects of Cell Surface Organization” ( G . Poste and G. L. Nicolson, eds.), p. 1. Elsevier, Amsterdam. Niemetz, J. (1972). J. Clin. Invest. 51, 307. Nogueira, N., Gordon, S. , and Cohn, Z. (1977). J . Exp. Med. 146, 157. Noonan, F. P., Halliday, W. J., Wall, D. R., and Clunie, G. J. A. (1977). Cancer Res. 37, 2473. Nordquist, R. E., Anglin, J . H., and Lerner, M. P. (1977). Science 197, 366. North, R. J . (1978). J. Immunol. 121, 806. Nowotny, A., Grohsman, J., Abdelnoor, A., Rote, N., Yang, C., and Waltersdorff, R. (1974). Eur. J . Immunol. 4, 73. Nowotny, A,, Butler, R. C., Grohsman, J., and Keebler, C. (1976). Ann. N.Y. Acad. Sci. 276, 106. O’Brien, E. T., Thornes, R. D., O’Brien, D., and Hogan, B. (1968). Lancet 1, 173. O’Day, W. T., and Young, R. W. (1978).J. Cell B i d . 76, 593. Odili, J . L . , and Taylor, G. (1971). Br. Med. J . 4, 584. Okada, Y., Kim, J., Maeda, Y., and Koseki, I. (1974). Proc. Natl. Acad. Sci. U . S . A . 71, 2043. Old, L. J., Stockert, E., Boyse, E. A., and Kim, J. H. (1%8). J . Exp. Med. 127, 523. Olden, K., and Yamada, K. M. (1977). Cell 11, 957. Oldstone, M. B. A. (1975a). J. Natl. Cancer Inst. 54, 223. Oldstone, M. B. A. (1975b). Prog. Med. Virol. 19, 84. Oldstone, M. B. A., Theofilopoulos, A. N., Gunven, P., and Klein, G. (1974). Intervirology 4, 292. Olmsted, J . B., and Borisy, G. G. (1973). Annu. Rev. Biochem. 42, 507. O’Meara, R. A. (1958). Ir. J . Med. Sci. 394, 474. O’Meara, R. A., and Thornes, R. D. (1961).Ir. J. Med. Sci. 423, 106. Onodera, K., and Sheinin, R. (1970). J. Cell Sci. 7, 337. Orci, L., Gabbay, K. H., and Malaisse, W. J. (1972). Science 175, 1128. Oroszlan, S., Henderson, L. E., Stephenson, J. R., Copeland, T. D., Long, C. W., Ihle, J . N., and Gilden, R. V. (1978). Proc. Natl. Acad. Sci. U.S.A. 75, 1404. Orr, T. S. C., Hall, D. E., and Allison, A. C. (1972). N a m e (London) 236, 350. Ossowski, L., Quigley, J. P., Kellerman, G. M., and Reich, E. (1973). J . Exp. Med. 138, 1056. Ossowski, L., Quigley, J . P., and Reich, E. (1974). J . Biol. Chem. 249, 4312. Otsuka, H., and Moskowitz, M. (1978). Exp. Cell Res. 112, 127. Owen, F. L., and Fanger, M. W. (1975). J. Immunol. 115, 765. Palade, G. (1969). In ”Subcellular Particles’’ (T. Hayashi, ed.), p. 64. Ronald, New York. Palade, G. (1975). Science 189, 347. Panem, S., and Kirsten, W. H. (1973). J . Nut/. Cancer Inst. 50, 563.

192

PAUL H. BLACK

Panem, S., and Schauf, V. (1973). J . Virol. 13, 1169. Papahadjopoulos, D., Poste, G., Schaeffer, B. E., and Vail, W. J. (1974).Biochim. Biophys. Acta 352, 10. Papahadjopoulos, D., Vail, W. J., Newton, C., Nir, S., Jacobson, K., Poste, G., and Lazo, R. (1977). Biochim. Biophys. Acta 465, 579. Pardee, A. B., Dubrow, R., Hamlin, J. L., and Kletzien, R. F. (1978). Annu. Rev. Biochem. 47, 715. Parish, C. R., and McKenzie, I. F. C. (1977). Cell. Immunol. 33, 134. Parks, R. C. (1975). J. Natl. Cancer Inst. 54, 1473. Paronetto, F., and Vernace, S. (1975). Clin. Exp. Irnmunol. 19, 99. Paskind, W. P., Weinberg, R. A,, and Baltimore, D. (1975). Virology 67, 242. Paul, D. C., Babbitt, J . L., Hull, R. N., and Williams, D. C. (1978). J. Cell Biol. 79, 77a. Pearlstein, E. (1976). Nature (London) 262, 497. Pearlstein, E., and Waterfield, M. D. (1974). Biochim. Biophys. Acta 362, 1. Pearlstein, E., Hynes, R. O., Franks, L. M., and Hemmings, U. J. (1976). Cancer Res. 36, 1475. Perdue, J. F. (1973). J. Cell Biol. 58, 265. Perrin, L. H . , and Oldstone, M. B. A. (1977). J . Irnmunol. 118, 316. Peters, T. J., Seymous, C. A., Wells, G., Fakunle, F., and Neale, G. (1977). Br. Med. J . 2, 1576. Peterson, A. R., Mondal, S., Brankow, D. W., Thon, W., and Heidelberger, C. (1977). Cancer Res. 37,3223. Peterson, H.-I., Kjartansson, I., Korsan-Bengtsen, K., Rudenstam, C.-M., and Zettergren, L. (1973). Actu Chir. Scand. 139, 219. Petitou, M.,Tuy, F., Rosenfeld, C., Mishal, Z., Paintrand, M., Jasnin, C., Mathe, G., and Inbar, M. (1978). Proc. Natl. Acad. Sci. U . S . A . 75, 2306. Phillips, D. R., and Morrison, M. (1971). Biochemistry 10, 1766. Pierce, C. W., Kapp, J. A., and Benacerraf, B. (1976). J . Exp. Med. 144, 371. Piessens, W. F., Remold, H. G., and David, J. R. (1977). J . Immunol. 118, 2078. Pinder, J. C., Bray, D., and Gratzer, W. B. (1977). Nature (London) 270, 752. Plate, M. D. (1976). Nature (London) 260, 329. Pluznik, D. H., and Sachs, L. (1965). J. Cell. Comp. Physiol. 66, 319. Podolsky, D. K., and Weiser, M. M. (1975). Biochem. Biophys. Res. Commun. 65, 545. Podolsky, D. K., Weiser, M. M., Westwood, J. C., and Gammon, M. (1977). J. Biol. Chem. 252, 1807. Podolsky, D. K., Weiser, M. M., Isselbacher, K. J., and Cohen, A. M. (1978). N . Engl. J . Med. 299, 703. Pollack, A. (1971). Am. J. Clin. Pathol. 56, 155. Pollack, R., Risser, R., Conlon, S., and Rifkin, D. (1974). Proc. Natl. Acad. Sci. U.S.A. 71, 4792. Porter, C. W., and Bernacki, R. J. (1975). Nature (London) 256, 648. Porter, K., Prescott, D., and Trye, J. (1973a). J . Cell B i d . 57, 815. Porter, K. R., Todaro, G. J., and Fonte, V. (1973b). J. Cell Biol. 59, 633. Poste, G. (1972). Inr. Rev. Cytol. 33, 157. Poste, G. (1977). In “Cancer Invasion and Metastasis: Biologic Mechanisms and Therapy” (S. B. Day, ed.), p. 19. Raven, New York. Poste, G., and Allison, A. C. (1973). Biochim. Biophys. Acta 300, 421. Poste, G., Papahadjopoulos, D., Jacobson, K., and Vail, W. J. (1975). Nature (London) 253, 552. Poupon, M.-F., Kolb, J.-P., and Lespinats, G. (1976). J . Narl. Cancer Insr. 57, 1241.

CELL SURFACE SHEDDING

193

Prather, S. O., and Lausch, R. N. (1976). I n t . J . Cancer 17, 380. Price, M. R., and Baldwin, R. W. (1977). I n “Dynamic Aspects of Cell Surface Organization” (G. Poste and G. L. Nicolson, eds.), p. 423. Elsevier, Amsterdam. Proctor, J . W., Rudenstam, C. M., and Alexander, P. (1973). Nature (London) 242, 29. Pross, H. F., and Baines, M. G. (1976). I n t . J. Cancer 18, 593. Puck, T. T. (1977). Proc. Natl. Acad. Sci. U . S . A . 74, 4491. Purcell, R. H., London, W. T., McAuliff, V. J., Palmer, A. E., Kaplan, P. M., Gerin, J. L., Wagner, J. A . , Popper, H., Lvovsky, E., Wong, D. C., and Levy, H. B. (1976). Lancet 2 , 757. Quigley, J. P. (1976). J. Cell Biol. 71, 472. Quigley, J. P., and Goldfarb, R. H. (1978). J. Cell Biol. 79, 73a. Quigley, J. P., Ossowski, L., and Reich, E. (1974). J . Biol. Chem. 249, 4306. Racevskis, J., and Sarkar, N . H. (1978). J. Virol. 25, 374. Rahman, A. F. R., Liao, S. K., and Dent, P. B. (1977). In Vitro 13, 580. Rakoczi, I., Wiman, B., and Collen, D. (1978). Biochim. Biophys. Acta 540, 295. Ramseier, H. (1974). J. Exp. Med. 140, 603. Ramseier, H . (1975). Eur. J. Immunol. 5, 23. Ran, M., Eshel, I., Witz, I. P., and Klein, G. (1975). J . Nail. Cancer Inst. 55, 843. Rao, V. S., and Bonavida, B. (1976). Cancer Res. 36, 1384. Rao, V. S., and Bonavida, B. (1977). Cancer Res. 37, 3385. Rasmussen, H. (1977). I n “Cell and Tissue Interactions” (J. W. Lash and M. M. Burger, eds.), p. 243. Raven, New York. Rasmussen, H . , and Goodman, D. B. P. (1977). Physiol. Rev. 57,421. Raz, A., and Goldman, R. (1976). Biochim. Biophys. Acta 455, 226. Rees, D. A., Lloyd, C. W., and Thom, D. (1977). Nature (London) 267, 124. Reisfield, R. A., Allison, J. P., Ferrone, S., Pellegrino, M. A,, and Poulik, M. D. (1976). Transplant. Proc. 8, 173. Remold, H. C., and Rosenberg, R. D. (1975). J. Biol. Chem. 250, 6608. Revel, J. P., Hoch, P., and Ho, D. (1974). Exp. Cell Res. 84, 207. Reynolds, F. J., Jr., Hanson, C. A., Aaronson, S. A., and Stephenson, J. R. (1977). J. Virol. 23, 74. Richert, N. D., and Ryan, R. J. (1977). Proc. Natl. Acad. Sci. U . S . A . 74, 4857. Ridway, E. B., Gilkey, J. C., and Jaffe, L . F. (1977). Proc. Narl. Acad. Sci. U . S . A . 74, 623. Rieber, M., Bacalao, J., and Alonso, G. (1975). Cancer Res. 35, 2104. Rifkin, D. B., and Pollack, R. (1977). J . Cell Biol. 73, 47. Rittenhouse, H. G., Rittenhouse, J. W., and Takemoto, L. (1978). Biochemistry 17, 829. Ritzi, E., Martin, D. S., Stolfi, R. L., and Spiegelman, S. (1976). Proc. Narl. Accid. Sci. U . S . A . 73, 4190. Ritzi, E., Martin, D. S., Stolfi, R. L., and Spiegelman, S. (1977). J . Exp. Med. 145, 999. Robbins, P. W., Wickus, G. G., Branton, P. E., Gaffney, B. J., Hirschberg, C. B., Fuchs, P., and Blumberg, P. M. (1974). Cold Spring Harbor Symp. Quant. Biol. 39, 1173. Roberts, R. M., and Yuan, B. 0 . (1974). Biochemistry 13, 4846. Roberts, R. M., and Yuan, B. 0. (1975). Arch. Biochem. Biophys. 171, 234. Roblin, R. (1978). I n “Biological Markers of Neoplasia: Basic and Applied Aspects” (R. W. Ruddon, ed.), p. 421. Elsevier, Amsterdam. Roblin, R., Albert, S . 0.. Gelb, N. A., and Black, P. H. (1975a). Biochemistry 14, 347. Roblin, R. O., Chou, I. N . , and Black, P. H. (1975b). Adv. Cancer Res. 22, 203. Rohrlich, S. T., and Rifkin, D. B. (1977). J. CelI Biol. 75, 31. Rollins, B. J., and Culp, L. A. (1979). Biochemistry 18, 141.

194

PAUL H. BLACK

Ronquist, G., and Agren, G. K. (1975). Cancer Res. 35, 1402. Ronzio, R. A., and Mohrlok, S. H. (1977). Arch. Biochem. Biophys. 181, 128. Roos, E., and Dingeman, K. P. (1979). Biochim. Biophys. Acta 560, 135. Rose, G. G., and Robertson, P. B. (1977). J. Dent. Res. 56, 416. Rosenthal, A. S., Barcinski, M. A., and Blake, J. T. (1977). Nature (London) 267, 156. Rossen, R. D., Reisberg, M. A., Hersh, E. M., and Gutterman, J. U. (1977). J . Natl. Cancer Inst. 58, 1205. Rossignol, B., Herman, G., and Keryer, G. (1972). FEES Lett. 21, 189. Roth, S., and White, D. (1972). Proc. Natl. Acad. Sci. U . S . A . 69, 485. Rothman, J. E., and Lenard, J. (1977). Science 195, 743. Rothman, J. E., and Lodish, H. F. (1977). Nature (London) 269, 775. Rubin, R. P. (1974). I n “Calcium and the Secretory Process” (R. P. Rubin, ed.), Plenum, New York. Rudland, P. S . , Seeley, M., and Seifert, W. (1974). Nature (London) 251,417. Rudman, D., DelRio, A., Akgun, S. , and Frumin, E. (1%9). A m . J . Med. 46, 174. Ruoslahti, E., and Vaheri, A. (1975). J. Exp. Med. 141, 497. Rutenberg, W. D., Al-Khalidi, V., Rosen, F. S . , and Merler, E. (1979). J . Immunol. 122, 723.

Rutishauser, U., Thiery, J.-P., Brackenbury, R., Sela, B.-A., and Edelman, G. M. (1976). Proc. Natl. Acad. Sci. U . S . A . 73, 577. Ryan, J. L., Arbeit, R. D., Dickler, H. B., and Henkart, P. A. (1975). 1.Exp. Med. 142, 814.

Sabatini, D., and Blobel, G. (1970). J. Cell B i d . 45, 146. Sabatini, D. D., Tashiro, Y.. and Palade, G. E. (1966). J. Mol. B i d . 19, 503. Sack, G. H., Jr., Levin, J., and Bell, W. R. (1977). Medicine 56, I . Sakuragawa, N., Takahoshi, K., Hoshiyama, M., Jimbo, C., Ashizawa, K., Matsuoka, M., and Ohnishi, Y. (1977). Thrornb. Res. 10, 457. Salmon, S. E., and Hamburger, A. W. (1978). Lancet 1, 1289. Sarmay, G., Istvan, L., and Gergely, J. (1978). Immunology 34, 315. Sauerheber, R. D., and Lordon, L. M. (1975). Proc. Soc. Exp. Biol. Med. 150, 28. Schabel, F. M. (1975). Cancer 35, 15. Schachter, H., Jabbal, I., Hudgin, R. L., Pineric, L., McQuire, E. J., and Roseman, S. (1970). J . Biol. Chem. 245, 1090. Schauf, V., and Panem, S. (1976). Virology 71, 619. Schimke, R. T. (1975). I n “Methods in Membrane Biology” (E. D. Korn, ed.), Vol. 3, p. 201. Plenum, New York. Schimpl, A., and Wecker, E. (1972). Nature (London) New B i d . 237, 15. Schlessinger, J., Barak, L. S., Hammes, G. G., Yamada, K. M., Pastan, I., Webb, W. W. and Elson, E. L. (1977). Proc. Natl. Acad. Sci. U . S . A . 74, 2909. Schmidt-Ullrich, R., Wallach, D. F. H., and Ferber, E. (1974). Biochim. Biophys. Acta 356, 288.

Schnebli, H. P. (1972). Schweiz. Med. Wschr. 102, 1194. Schnebli, H. P., and Burger, M. M. (1972). Proc. Natl. Acad. Sci. U . S . A . 69,3825. Schnebli, H. P., and Haemmerli, G. (1974). Nature (London) 248, 150. Schneider, A. L., Herz, R., and Rosenheck, K. (1977). Proc. Narl. Acad. Sci. U.S.A. 74, 5036.

Schober, R., Nitsch, C., Rinne, U., and Morris, S. J. (1977). Science 195,495. Schochetman, G . , Fine, D. L., and Massey, R. J. (1978). Virology 88, 379. Schreiner, G. F., Fujiwara, K., Pollard, T. D., and Unanue, E. R. (1977). J. Exp. Med. 145, 1393.

CELL SURFACE SHEDDING

195

Schroder, E. W., Chou, I. H., Jaken, S. , and Black, P. H. (1978). Nature (London) 276, 828. Schroit, A. J., Geiger, B., and Gallily, R. (1973). Eur. J . Immunol. 3, 354. Schuel, H . , Wilson, W. L., Chen, K., and Lorand, L. (1973). Dev. Biol. 34, 175. Schuel, H.,Troll, W., and Lorand, L. (1976). Exp. Cell Res. 103, 442. Schulman, H.,and Greengard, P. (1978). Nature (London) 271, 478. Schwartz, R. H., David, C. S. , Sachs, D. H., and Paul, W. E. (1976). J. Immunol. 117, 531. Seeman, P., Chau-Wong, M., and Moyyen, S. (1973). Nature (London) New Biol. 241, 22. Sefton, B. M., and Rubin, H. (1970). Nature (London) 227, 843. Seifert, W. E., and Rudland, P. S . (1974a). Nature (London) 248, 138. Seifert, W. E., and Rudland, P. S. (1974b). Proc. Natl. Acad. Sci. U . S . A . 71, 4920. Shainoff, J. R., and Page, I. H.(1%2). J . Exp. Med. 116, 687. Shapiro, S. Z . , and August, T. (1976). Biochim. Biophys. Acta 458, 375. Sherline, P., Lee, Y.-C., and Jacobs, L. S. (1977). J. Cell Biol. 72, 380. Sherman, M. K., Strickland, S., and Reich, E. (1976). Cancer Res. 36,4208. Sherton, C. C., Evans, L. H.,Polonoff, E., and Kabat, D. (1976). J. Virol. 19, 118. Shied. A . , and Choppin, P. W. (1975). I n "Proteases and Biological Control" (E. Reich, D. B. Rifkin and E. Shaw, eds.), p. 645. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York. Shimamoto, T. (1974). Throm. Diath. Haemor. 60, 5 . Shoji, M., Morris, H. P., Davis, C. W., Brackett, N. L., and Kuo, J. F. (1977). Biochim. Biophys. Acta 500, 419. Shreffler, D. C., David, C. S., Cullen, S. E., Frelinger, J. A., and Niederhuber, J. E. (1976). Cold Spring Harbor Symp. Quant. Biol. 41,477. Siekevitz, P. (1972). J. Theor. Biol. 37, 321. Sindelar, W. F., Tralka, T. S., and Ketcham, A. S. (1975). J . Surg. Res. 18, 137. Singer, S. J. (1974a). Annu. Rev. Biochem. 43, 805. Singer, S. J. (1974b). Adv. Immunol. 19, I . Singer, S. J. (1977). J . Supramol. Struct. 6, 313. Singer, S. J., and Nicolson, G. L. (1972). Science 175, 720. Sinkovics, J. G., Cabiness, J. R., and Shullenberger, C. C. (1972). Cancer 313, 1428. Skinnider, L. F., and Ghadially, F. N. (1977). Br. J. Cancer 35, 657. Skipski, V . , Katopodis, N., Prendergast, J. S., and Stock, C. C. (1975). Biochem. Biophys. Res. Commun. 67, 1122. Smets, L. A., and DeLey, L. (1974). J. Cell. Physiol. 84, 343. Smith, J. F . , and Brown, D. T. (1977). J . Virol. 22, 662. Smith, R. J., and Ignarro, L. J. (1975). Proc. Natl. Acad. Sci. U . S . A . 72, 108. Solomon, J., Chou, I. N., and Black, P. H. (1979). (in preparation). Spataro, A. C., Morgan, H. F., and Bosmann, H.B. (1976). J. Cell Sci. 21, 407. Spudich, J. A , , and Lin, S. (1972). Proc. Natl. Acad. Sci. U . S . A . 69,442. Stackpole, C . W., Jacobson, J. B., and Lardis, M. P. (1974). J . Exp. Med. 140, 939. Stadler, J. K., and Adelberg, E. A. (1972). Proc. Natl. Acad. Sci. U . S . A . 69, 1929. Stanley, E. R., Cifone, M., Heard, P. M., and Defendi, V. (1976). J. Exp. Med. 143, 631. Stanley, E. R . , Chen, D.-M., and Lin, H . 3 . (1978). Nature (London) 274, 168. Stanley, E. R., and Heard, P. M., (1977).J. Biol. Chem. 252, 4305. Stathakis, N. E., Mosesson, M. W., Chen, A. B., and Galanakis, D. K. (1978). Blood 51, 1211. Steele, G., Jr., Sjogren, H. O., Rosengren, J. E., Lindstrom, C., Larsson, A., and Leandoer, L. (1975). J. Natl. Cancer Inst. 54, 959.

196

PAUL H . BLACK

Stefanovic, V., Ciesielski-Treska, J., Ebel, A., and Mandel, P. (1974). FEBS Lett. 49,43. Stein, W. D. (1972). Ann. N . Y . Acad. Sci. 195, 37. Steinhardt. R . , Zucker, R., and Schatten, G. (1977). Dev. Biol. 58, 185. Steinman, R. M., Silver, J. M., and Cohn, Z. A. (1974). J. Cell Biol. 63, 949. Stenman, S ., Wartiovaara, J., and Vaheri, A. (1977). J. Cell Biol. 74, 453. Stone, K. R., Smith, R. E., and Joklik, W. K. (1974). Virology 58, 86. Stradler, J., and Franke, W. W. (1974). 1.Cell Biol. 60,297. Strand, M., and August, J. T. (1976). Virology 75, 130. Strander, H., and Einhorn, S. (1977). Int. J . Cancer 19, 468. Straub, P. W., Riedler, G., and Frick, P. G. (1967). J. Clin. Pathol. 20, 152. Strauch, L. (1972). In "Tissue Interactions in Carcinogenesis" (D. Tarin, ed.), p. 399. Academic Press, New York. Strickland, S., and Beers, W. H. (1976). J . Biol. Chem. 251, 5694. Strickland, S., Reich, E., and Sherman, M. I. (1976). Cell 9, 231. Stuhimiller, G. M., and Seigler, H. F. (1977). J. Natl. Cancer Inst. 58, 215. Sugamura, K., and Smith, J. B. (1977). Cell. Immunol.30, 353. Sugarbaker, E. V., and Cohen, A. M. (1972). Surgery 72, 155. Sugarbaker, E. V., and Ketcham, A. S. (1977). Semin. Oncol. 4, 19. Sugiyama, K . (1977). Nature (London) 270, 614. Svanberg, L., and Astedt, B. (1975). Cancer 35, 1382. Synder, H. W., Jr., Stockert, E., and Fleissner, E. (1977). J. Virol. 23, 302. Tada, T., Taniguchi, M., and Takemori, T. (1975). Transplant. Rev. 26, 106. Tada, T., Takemori, T., Okumura, K.,Nonaka, M., and Tokuhisa, T. (1978). J. Exp. Med. 147,446. Tanaka, K., and Ichihara, A. (1977). J. Cell. Physiol. 93, 407. Taniguchi, M., Hayakawa, K.,and Tada, T. (1976). J . Immunol. 116, 542. Targowski, S. P., Abeyounis, G. J., and Milgrom, F. (1977). Proc. SOC.Exp. Biol. Med. 154, 365. Taussig, M. J. (1974). Nature (London) 248, 234. Taussig, M. J., and Finch, A. P. (1977). Nature (London) 270, 151. Temple, R . , Williams, J. A., Wilber, J. F.,and W O E , J. (1972). Biochem. Biophys. Res. Commun. 46, 1454. Teng, N. N. H., and Chen, L. B. (1975). Proc. Natl. Acad. Sci. U.S.A. 72, 413. Theofilopoulos, A. N., Andrews, B. S., Urist, M. M., Morton, D. L., and Dixon, R. J. (1977). J . Immunol. 119, 657. Thomas, D. W., Yamashita, U . , and Shevach, E. M. (1977). J. Immunol. 119, 223. Thompson, J. E., Chambers, J. A., and Semple, N. L. (1978). Exp. Cell Res. 113, 127. Thomson, D. M. P., Eccles, S., and Alexander, P. (1973a). Br. J . Cancer 28, 6. Thomson, D. M. P., Steele, K., and Alexander, P. (1973b). Br. J. Cancer 27, 27. Ting, C.-C., and Rogers, M. J. (1977). Nature (London) 266, 727. Todd, A. S. (1964). Br. Med. Bull. 20, 210. Toh, B. H., and Hard, G. C. (1977). Nature (London) 269, 695. Tokes, Z. A., and Sorgente, N. (1976). Biochem. Biophys. Res. Commun. 73, 965. Tokes, Z. A., Bruszewski, W. B., and O'Brien, R. L. (1978). Birth Defects Orig. Artic. Ser. XIV, 195. Toneguzzo, F., and Ghosh, H. P. (1978). Proc. Natl. Acad. Sci. U.S.A. 75,715. Tong, M. J . , Wallace, A. M., Peters, R. L., and Reynolds, T. B. (1975). N . Engl. J . Med. 293, 318. Tou, J.-S., and Stjernholm, R. L. (1975). Biochim. Biophys. Acra 392, I . Toyama, S., Toyama, S., and Uetake, H. (1978). Virology 86, 138.

CELL SURFACE SHEDDING

I97

Treves, A. J., Cohen, I. R., and Feldman, M. (1976). J. Natl. Cancer Inst. 57, 409. Truding, R., and Morell, P. (1977). J. Biol. Chem. 252, 4850. Tsung, P.-K., Kegeles, S. W., and Becker, E. L. (1977). Biochim. Biophys. Acta 499, 212. Tsung, P.-K., Kegeles, S. W., and Becker, E. L. (1978). Biochim. Biophys. Acta 541, 150. Tupper, J. T., and Zorgniotti, F. (1977). J . Cell B i d . 75, 12. Tupper, J. T.. DelRosso, M., Hazelston, B., and Zorgniotti, F. (1978).J. Cell. Physiol. 95, 71. Tweto, J., and Doyle, D. (1976). J . Biol. Chem. 251, 872. Tweto, J., and Doyle, D. (1977). In “Cell Surface Reviews” ( G . L . Nicolson, ed.), Vol. 4, p. 138. Elsevier, Amsterdam. Unanue, E. R., and Kiely, J.-M. (1977). J . Immunol. 119, 925. Unanue. E. R., and Schreiner, G. F. (1977). In “Dynamic Aspects of Cell Surface Organization” ( G . Poste and G. L . Nicolson, eds.), p. 619. Elsevier, Amsterdam. Unanue, E. R., Kiely, J.-M., and Calderon, J. (1976). J . Exp. Med. 144, 155. Unkeless, J. C., Tobia, A., Ossowski, L., Quigley, J. P., Rifkin, D. B., and Reich, E. (1973). J . Exp. Med. 137, 85. Unkeless, J., Dano, K., Kellerman, G. M., and Reich, E. (1974a).J. Biol. Chem. 249, 4295. 4295. Unkeless, J. C., Gordon, S., and Reich, E. (1974b). J. Exp. Med. 139,834. Urovitz, E. P., Czitrom, A. A., Langer, F., Gross, A. E., and Pritzker, K. P. H. J. Bone Joint Surg. SS-A, 308. Vaage, J. (1973). Cancer Res. 33, 493. Vaheri, A., and Mosher, D. F. (1978). Biochim. Biophys. Acfa 516, I . Vaheri, A., and Ruoslahti, E. (1974). I n / . J. Cancer 13, 579. Vaheri, A., and Ruoslahti, E. (l975a). J . Exp. Med. 142, 530. Vaheri, A,, and Ruoslahti, E. (1975b). I n “Proteases and Biological Control” (E. Reich, D. Rifkin, and E. Shaw, eds.), p. 967. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York. Vaheri, A., Ruoslahti, E., Hovi, T., and Nordling, S. (1973). J. Cell. Physiol. 81, 355. VanBlitterswijk, W. J., Emmelot, P., Hilkmann, H. A. M., Oomen-Meulemans, E. P. M., and Inbar, M. (1977). Biochim. Biophys. Acta 467, 309. VanObberghen, E., Somers, G., Devis, G., Vaughan, G. D., Malaise-Lagge, F., Orci, L., and Malaisse, W. J. (1973). J. Clin. Invest. 52, 1041. VanRood, J . J., vanleeuiven, A., Keuning, J. J., and Termijtelen, A. (1977). Scand. J. Immunol. 6, 373. Vassalli, J.-D., and Reich, E. (1977). J . Exp. Med. 145, 429. Vassalli, J.-D., Hamilton, J., and Reich, E. (1976). Cell 8, 271. Vassalli, J.-D., Hamilton, J., and Reich, E. (1977). Cell 11, 695. Verbert, A., Cacan, R., and Montreuil, J. (1976). Eur. J. Biochem. 70, 49. Vetterlein, D., Young, P. L., Bell, R. E., and Roblin, R. (1979). J . Biol. Chem. 254, 575. Viaje, A., Slaga, R. J., Wigler, M., and Weinstein, 1. B. (1977). Cancer Res. 37, 1530. Vial, J. D., and Garrido, J. (1976). Proc. Natl. Acad. Sci. U . S . A . 73, 4032. Vitetta, E. S . , and Uhr. J. W. (1972). J. Immunol. 108, 577. Vitetta, E. S., and Uhr. J. W. (1975). Biochim. Biophys. Acta 436, 253. Vlodovsky, I., Inbar, M., and Sachs, L. (1973). Proc. Natl. Acad. Sci. U . S . A . 70, 1780. Vogel, K. G., and Kelley, R. 0. (1977). J. Cell. Physiol. 92, 469. Volsky, D., and Loyter, A. (1977). Biochim. Biophys. Acta 471, 243. Vorbrodt, A,, and Koprowski, H. (1969). J. Natl. Cancer Inst. 43, 1241. Vos, J., Ahkong, Q. F., Botham, G. M., Quirk, S. J., and Lucy, J. A. (1976). Biochem. J. 158, 65 1.

198

PAUL H. BLACK

Voyles, B. A., and Moskowitz, M. (1976). Biochim. Biophys. Acta 441, 269. Wands, J . R., Perrotto, J. L., Alpert, E., and Isselbacher, K. J. (1975).J . Clin. Invest. 55, 921. Ward, P. A., and Becker, E. L. (1968). J . Exp. Med. 127, 693. Ward, P. A,, and Becker, E. L. (1970).J . Immunol. 105, 1057. Warren, B. A. (1973). J . Med. 4, 150. Warren, L., and Glick, M. C. (1968).J . Cell Biol. 37, 729. Wartiovaara, J . , Linder, E., Ruoslahti, E., and Vaheri, A. (1974).J. Exp. Med. 140, 1522. Watson, J . (1975). J . Exp. Med. 141, 97. Webb, G. C., and Roth, S . (1974). J . Cell Biol. 63, 796. Weiner, H. L., Scribner, D. J., and Moorhead, J. W. (1978). J. Immunol. 120, 1907. Weinstein, I . B., and Troll, W. (1977). Cancer Res. 37, 3461. Weiser, M. M. (1973). J . Biol. Chem. 248, 2542. Weiser, M. M., Podolsky, D. K., and Isselbacher, K. J. (1976). Proc. Narl. Acad. Sci. U.S.A. 73, 1319. Weiser, M. M., Neumeier, M. M., Quaroni, A., and Kirsch, K. (1978). J . Cell Biol. 77, 722. Weiss, G., and Beller, F. K . (1969). Am. J . Obst. Gynecol. 103, 1023. Weiss, L. (1977). Semin. Oncol. 4, 5 . Wells, V., and Mallucci, L. (1978). Exp. Cell Res. 116, 301. Werb, Z. (1978). J . Exp. Med. 147, 1695. Werb, Z., and Aggeler, J . (1978). Proc. Natl. Acad. Sci. U . S . A . 75, 1839. Werb, Z., Mainardi, C. L., Vater, C. A., and H a m s , E. D., Jr. (1977). N . Engl. J . Med. 296, 1017. Wheeler, G. L., and Bitensky, M. W. (1977). Proc. Natl. Acad. Sci. U . S . A . 74,4238. White, J . G . , and Estensen, R. D. (1974). Am. J. Pathol. 65, 45. Whitfield, J . F., MacManus, J. P., Boynton, A. L., Gillan, D. S., and Isaacs, R. J. (1974). J . Cell. Physiol. 84, 445. Whitfield, J. F., MacManus, J. P., Rixon, R. H., Boynton, A. L., Youdale, T., and Swierenga, S . H. H . (1976). I n Vitro 12, 1 . Whitney, R. B., and Sutherland, R. M. (1972a). J. Cell. Physiol. 80, 329. Whitney, R. B., and Sutherland, R. M. (1972b). Cell. Imrnunol. 5, 137. Whitney, R. B., and Sutherland, R. M. (1973). J . Cell. Physiol. 82, 9. Wickus, G. G., Branton, P. E., and Robbins, P. W. (1974). I n "Control of Proliferation in Animal Cells" (B. Clarkson and R. Baserga, eds.), p. 541. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York. Wigler, M., and Weinstein, I. B. (1976). Nature (London) 259, 232. Wigler, M. DeFeo, D., and Weinstein, I. B. (1978). Cancer Res. 38, 1434. Williams, D. A., Boxer, L. A., Oliver, J . M., and Baehner, R. L. (1977). Nature (London) 267, 255. Willingham, M. C., Yamada, K . M., Yamada, S. S., Pouyssegur, J., and Pastan, I. (1977). Cell 10, 375. Wilson, L., Bamburg, J . R., Mizel, S. B., Grimsham, L. M., and Creswell, K. M. (1974). Fed. Proc. 33, 158. Wiman, B., and Collen. D. (1978). Nature (London) 272, 549. Witte, 0. N., and Baltimore, D. (1978). J . Virol. 26, 750. Witte, 0. N . , Tsukamoto-Adey, A., and Weissman, I. L. (1977). Virology 76, 539. Wolf, B. A., and Goldberg, A. R. (1976). Proc. Natl. Acad. Sci. U . S . A . 73, 3613. Wolf, 8 . A., and Goldberg, A. R. (1978). Proc. Natl. Acad. Sci. U . S . A . 75, 4%7. Wolff, I . , Timpl, R., Pecker, I., and Steffen, C. (1967). Vox Sang. 12, 443.

CELL SURFACE SHEDDING

199

Wood, S. (1971). In “Pathobiology Annual” (H. L. Ioachin, ed.), p. 281. Appleton, New York. Woodin, A. M., and Wieneke, A. A. (1964). Biochem. J. 90, 498. Woody, J. N . (1977). Nature (London) 269, 61. Yamada, K . M., and Olden, K. (1978). Nature (London) 275, 179. Yamada, K . M., Olden, K., and Pastan, 1. (1978). Ann. N . Y . Acad. Sci. 312, 256. Yamada, K . , Yamada, S. S., and Pastan, I. (1977). J. Cell Biol. 74, 649. Yamanishi, Y., Dabbous, M. K., and Hashimoti, K. (1972). Cancer Res. 32, 2551. Yamasaki, H., Fibach, E., Nudel, U., Weinstein, I. B., Rifkind, R. A., and Marks, P. A. (1977). Proc. Natl. Acad. Sci. U . S . A . 74, 3451. Yefenof, E., Lundin, L., and Klein, G. (1978). Eur. J . Immunol. 8, 190. Yoshinaka, Y., and Luftig, R. B. (1977a). Cell 12, 709. Yoshinaka, Y., and Luftig, R. B. (1977b). Proc. Natl. Acad. Sci. U . S . A . 74, 3446. Youn, J. K . , LeFrancois, D., and Barski, B. (1973). J. Natl. Cancer I n s t . 50, 921. Young, R. W. (1971). J. Cell Biol. 49, 303. Young, R. W. (1973). Ann. Ophrhalmol. 5, 843. Zacharski, L. R., and McIntyre, 0. R. (1971). Nature (London) 232, 338. Zacharski, L. R., and McIntyre, 0. R. (1972). Proc. Soc. Exp. Biol. Med. 139, 713. Zacharski, L. R., and McIntyre, 0. R. (1973a). J . Med. 4, 118. Zacharski, L. R., and McIntyre, 0. R. (1973b). Blood 41, 679. Zacharski, L. R . , Bowie, E. J. W., Titus, J . L., and Owen, C. A . , Jr. (1969). Mayo Clin. Proc. 44, 784. Zakai, N., Kulka, R. G., and Loyter, A. (1976). Nature (London) 263, 696. Zakai, N.,Kulka, R. G., and Loyter, A. (1977). Proc. Natl. Acad. Sci. U . S . A . 74, 2417. Zetter, B. R., Chen, L. B., and Buchanan, J. M. (1976). Cell 7, 407. Zucker, R., and Nolte, J. (1978). Nature (London) 274, 78. Zurier, R. B., Hoffstein, S., and Weissmann, G. (1973). Proc. Narl. Acad. Sci. U . S . A . 70, 844. Zwaal, R. F . A. (1975). I n “Cell Surfaces and Malignancy” (P. T. Mora, ed.), p. 39. Fogarty International Conference Center Proceedings, #24. U. S. Government Printing O f i c e , Washington, D.C.

This Page Intentionally Left Blank

TUMOR ANTIGENS ON NEOPLASMS INDUCED BY CHEMICAL CARCINOGENS AND BY DNA- AND RNA-CONTAINING VIRUSES: PROPERTIES OF THE SOLUBILIZED ANTIGENS

Lloyd W. Law, Michael J.

Rogers,

and Ettore Appella

Laboratory of Cell Biology, National Cancer Institute. Bethesda. Maryland

I. Introduction ................................................ 11. Soluble Antigens from Chemically Induced Tumors ........................ A. p53, A Common Transformation-Related Antigen ...................... B. Relationship of Tumor Rejection Antigens (TATA) and H-2 Antigens . . . . C. Isolation and Fractionation of TATA from the C-l Sarcoma Bearing Unique

203 208 211 2 I4

111. TATA of Neoplasms

............................. IV. TATA of Neoplasms (Leukemias) Induced by RNA Tumor Viruses . . . . . . . . . V . Soluble Antigens and Immune Deviations of the Host ...................... VI. Concluding Remarks ............................. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

216 219 225 229 232

I. Introduction

Antigens detected on neoplastic cells are of diverse types and have been detected by various methods in vivo and in vitro. Many of the antigens that have been identified are not strictly tumor specific, and except for the tumor rejection antigens of the transplantation type (TATA), their precise biologic functions are not known. The best examples of tumor specific surface antigens are ( 1 ) the TATA of chemically induced sarcomas, (2) the TATA of DNA-containing virus-induced neoplasms, (3) the TL antigen of murine leukemias (MuL) (Old and Stockert, 1977), (4) the transformation-specific antigens of the avian and feline oncornavirus-induced neoplasms (Kurth and Bauer, 1972; Essex et al., 1971), and (5) the MuLV-induced TATA on Friend, Moloney, Rauscher virus (FMR-induced leukemias) (Ting et a l . , 1974). The broad array of other antigens found on neoplastic cells would include the differentiation alloantigens of the Ly series, Thy- I . MuLV-structural and MuLV-related antigens, derepression antigens, Gross cell surface antigen (GCSA), GIX, etc., species-specific antigens, receptor site antigens, tissue or organ type specific antigens, and fetal or embryonic antigens (see Hauschka, 1973; Law and Appella, 1975; Old and Stockert, 1977). Many of these antigens and the numerous ones described for human neoplastic tissues are de20 I ADVANCES IN CANCER RESEARCH, VOL. 32

Copyright @ 1980 by Academic Press, Inc. All rights of reproduction in any form reserved ISBN 0-12-006632-7

202

LLOYD w . LAW

et al.

tected by cell-mediated or humoral lysis, delayed hypersensitivity reactions, and immunofluorescence; they are not well characterized nor is their relevance to the antitumor immune response known. Studies of specific tumor-rejection antigens are carried out in a completely syngeneic system. In such a system, genetic differences between host and tumor are reduced to a minimum or are eliminated. This system also provides the most critical and convincing evidence for the presence and functioning of TATAs. The assay of tumor inhibition is cumbersome and time consuming. As a consequence, other in vivo and in vitro assays have been used in studies of membrane-expressed tumor antigens and the immune responses they induce. However, when tumor antigens are demonstrated by other in vivo reactions or by serological methods in vitro, it should be borne in mind that the observed immune reactions may be directed against membrane determinants other than the antigens uniquely expressed by the tumor. These other determinants may reside on fetal antigens, differentiation antigens, tissue-specific antigens, and thereby, may be totally irrelevant to the antigen under study; indeed, the immune response may be nonspecific. When relying strictly on an in vitro assay, it is therefore necessary to examine the specificity of each reaction very carefully and, more importantly, to obtain a good correspondence between the specificity of the rejection-inducing response and its in vitro counterpart. The value of many in vitro assays for assessing the rejection potential of a host challenged with a tumor is not clear; in fact, the recent literature is contradictory on this point. The findings from colony inhibition and microcytotoxicity assays and from observing the effects of “blocking” and “deblocking” factors are often puzzling; these methods do not have as yet reliable in vivo counterparts. Thus, studies restricted to in vitro assays must be interpreted with caution as to their relevance to a specific host-tumor immune response. It is not always wise to assume without question that one is using a completely syngeneic system in the study of TATAs. Yet, the present highly developed techniques and available specific reagents for the serologic identification of major histocompatibility (MHC) products (particularly H-2 and HLA) and the present knowledge of the biochemistry of H-2 and other MHC antigens permit detection of anomalous phenotypic changes in the tumor. Therefore, appropriate analyses can be carried out to resolve such a problem, at least with murine tumors. Antigens expressed at the cell surface with specificities identical to those of TATAs have also been demonstrated by in vitro reactions with sensitized lymphocytes or with serum antibody from tumor-immune donors. Although there is no proof that these surface antigens possess the

TUMOR ANTIGENS O N NEOPLASMS

203

same determinants as those functioning to produce tumor rejection, they represent specific tumor markers and will be discussed in the context of this article. A widely held concept is that tumor antigens are membrane proteins. This has led recently to a major emphasis on attempts to detect, isolate, and biochemically characterize membrane-associated proteins with antigenic specificity. A variety of solubilization techniques have been employed including use of sonic energy (Holmes et al., 1970), limited enzymatic digestion (Baldwin et d., 1973; Drapkin et af., 1974), chaotropic salts (Meltzer et al., 1971; Pellis et af., 1974), and detergents such as Nonidet P40 (NP40) and sodium deoxycholate (Prat et a/., 1975; Rogers et al., 1977b). Perhaps the most popular agents for membrane solubilization are chaotropic salts (e.g., 3 M KCl, which is used mainly for preparing crude extracts). However, none of these 3 M KCl extracts has been extensively purified and the quantities of antigen solubilized have not always been estimated. Similarly, the use of proteases has in a few cases resulted in biologically active components, but the yields obtained are quite low (Henriksen et af., 1977). These drawbacks are largely overcome by the use of detergents. Isolated plasma membranes are commonly prepared prior to protein solubilization. However, autoenzymatic degradation of solubilized TATAs is a constant problem, since isolated cell membranes exhibit a number of enzymatic activities. Serious efforts have to be made to inhibit autoenzymolysis in order to obtain optimum quantities of solubilized, biologically active membrane proteins. In this article, we will summarize the results of recent attempts to identify and isolate tumor antigens. We shall concentrate on TATAs, because these antigens are demonstrably important to the host antitumor response. Antigens detected by other than the tumor rejection response will be considered only peripherally. In those cases where actual data are presented, we have relied mainly on experiments from our own laboratory.

II. Soluble Antigens from Chemically Induced Tumors

The first studies on the appearance of neoantigens on cells transformed by chemical carinogens were recorded in early 1950. Polycyclic hydrocarbon-induced sarcomas and carcinomas in the mouse, rat, and guinea pig have been investigated most often. The most extensively studied are 3-methylcholanthrene (MC)-induced sarcomas and 4-dimethylaminoazobenzene (DMAB)-induced hepatomas. The immune response against

204

LLOYD w . LAW

et al.

tumor associated transplantation antigens (TATA), which are responsible for tumor rejection, are evoked by pretreatment of the syngeneic host with tumor cells that, for example, have been irradiated to prevent them from progressive growth. In general, these antigens are stable and appear to be individually distinct for each transplanted tumor. Even different primary tumors arising in a single host are antigenically dissimilar. The frequency of cross-reactivity, however, cannot be defined, since a test of a large number of tumors at any one time has not been carried out. There are reports of the existence of strong cross-reacting antigens (Leffell and Coggin, 1977; Economu et al., 1977) and of weaker cross-reacting TATAs (Burton and Warner, 1978; Hellstrom et al., 1978), in addition to the strong unique TATAs, but this has not been a general finding. These observations do not preclude the existence of other cross-reacting antigens; they show only that such antigens are rarely demonstrable in tumor rejection tests. The magnitude of tumor rejection responses elicited by individual tumors in a given strain or by tumors of different morphology in the same species is also quite variable. There is evidence that the immunogenicity of the various chemically induced sarcomas in the mouse depends on the dose of carcinogen: low doses produce tumors of low immunogenicity (Prehn, 1976). Further studies have revealed that with tumors induced in the rat by aminoazo dyes, aromatic amines, and alkylnitrosamines, the expression of tumor rejection antigen does not necessarily occur. With the 2-acetylaminofluorene-induced rat tumors, tumor transplantation resistance could only be demonstrated in a small proportion of tumors (Baldwin and Embleton, 1971). Spontaneously occurring tumors are also less frequently able to initiate tumor rejection reactions when compared to experimentally induced tumors. This has led to some criticism of all the emphasis given to chemically induced tumors; however, the nonimmunogenicity of some tumors might indicate that TATAs do not always elicit a cellular response or that the sensitivity of the immunological assay used in the analysis might not be adequate. Humoral antibodies to cell surface antigens are detectable in the sera of hosts that are immunized by inoculation of irradiated tumor cells or that have undergone tumor excision. These antibodies are demonstrable, for example, by immunofluorescence (Baldwin et al., 1971), colony inhibition, microcytotoxicity assays (Hellstrom et al., 1968; Baldwin and Embleton, 1974), and complement-dependent cytotoxicity (Takasugi and Klein, 1970). Although there is no indication that these results have any relevance to tumor growth or inhibition in vivo, or that the antigens detected by in vivo and in vitro procedures are the same, they have been

TUMOR ANTIGENS O N NEOPLASMS

205

valuable for the analysis of antigenic specificity among MC-induced sarcomas and DMAB-induced hepatomas. Attempts to isolate tumor antigens have been based on procedures developed for the isolation of histocompatibility antigens. These efforts have been characterized by low yield of antigenic material and poor biological activity. TATA of guinea pig hepatomas has been solubilized by extraction with 3 M KCl. Recovery of antigenic activity ranged from 1540% of that present on viable hepatoma cells; however, the activity in tumor rejection tests has been minimal (Meltzer ef a / . , 1971). In MCinduced sarcomas and aminoazo dye-induced hepatomas of the rat, tumor-specific antigens have been isolated in a soluble form by a variety of procedures. The yields were low, and the membrane or soluble preparations did not elicit transplantation immunity, although they evoked a humoral antibody response that was tumor specific. Biochemical procedures involving ion-exchange and gel filtration chromatography and immunoadsorption techniques have led to some degree of purification (Baldwin et al., 1978). Pellis and Kahan (1975) have showed transplantation activity of 3 M KCl extracts from several MC-induced sarcomas of C3H mice. However, the extracts were weakly immunogenic and effective only in a narrow dose range. Recent findings of Pasternak et a / . (1978) and Bubenik et a / . (1978) are similar. In our laboratory TATA from the membranes of an MC-induced sarcoma, Meth A, from BALB/c mice has been solubilized with the detergent NP40 and partially purified by conventional chromatographic procedures including gel filtration, lectin affinity chromatography, and column electrophoresis (Natori et a/., 1977a). Meth A, originally described by Old et a / . (1962) was maintained in ascitic form and was developed also into a tissue culture-passaged line. Most of the work involving solubilization of the antigen from membranes and tumor challenge was performed with the in vivo-passaged line. Meth A has a strong TATA. This was demonstrated by the fact that complete protection in syngeneic BALB/c mice against a challenge of 2 x lo4 cells (200 x TD5&is achieved through a single immunization with lo5 irradiated Meth A cells. Cross-reactivity in vivo has not been observed with any other neoplasms assayed. The TATA elicits specific cellular immune responses by T cells (observed in adoptive transfer studies). Another characteristic of Meth A that makes it particularly useful is its availability in ascitic form, which provides an excellent source of cell membranes and effective target cells for cytotoxic assays (DeLeo et a/., 1977). The ascites form also does not express murine leukemia virus (MuLV) antigens or alien H-2 antigens, which could otherwise be detected, either

206

LLOYD

w. LAW et al.

serologically or by tumor rejection studies using various F1 hybrid mice (Appella et al., 1978). MuLV and its antigens d o add serious complications to serologic studies, although these antigens do not appear to be functioning as TATA; otherwise, cross reactions among chemically induced neoplasms would be more common. A new preparation of Meth A TATA, released from the membranes of ascitic sarcoma cells by NP40, has been recently characterized (Law et al., 1978a). Fractionations on an Ultrogel AcA22 column, on a Lens culinaris lectin affinity column, on a WGA (wheat germ agglutinin) column, and finally on an Ultrogel AcA54 column were carried out. The striking immunity provided by the material that did not bind to the lectin affinity column is shown in Table I. Immunization in group 1 was accomplished with the freshly prepared lectin-unbound material (LcH-u), whereas groups 4 and 5 were immunized with the same preparation stored at -20°C for 4 months. There is some indication of loss of immunogenic potential of the stored material, although a complete dose-response study was not done. Greater than 95% tumor inhibition was achieved with as little as 1.0 pg (0.5 p g x 2) of the lectin-unbound fraction (group 2). The Sloan-Kettering subline of Meth A, found to be negative for MuLV and MuLV-related antigens (DeLeo et al., 1977), was also strikingly inhibited (group 3). LcH-u and the NP40-solubilized crude material (NP40-CS) showed specificity in the induction of immunity. LcH-u still contained

TABLE I IMMUNOGENIC POTENTIAL OF LcH-UNBOUND, SOLUBILIZED TATA FROM METHA CELLMEMBRANES Percentage of control tumor volume at varying total dosages (pg) Group" I 2 3b

4c

5

1.0

1.6

2.0

_

20

10

_ 8

_ _ - 21 - _

0.5

_ 12 1

6

50

1.1

0.6

4

Cld 2 -

CI -

-

-

100

CI I5

-

At least 8 BALB/c mice in each group: mice were immunized twice, 10 days apart, and challenged at 10 days after last immunization with 2 x lo4 Meth A ascites cells. Sloan-Kettering subline of Meth A. Negative for MuLV and MuLV-related antigens (DeLeo et al., 1978). Only one preparation was used throughout this study, but LcH-u material for groups 4 and 5 had been stored for 4 months at -20°C. CI, Complete inhibition of growth.

TUMOR ANTIGENS O N NEOPLASMS

207

5% H-2 activity and showed on SDS gels only a few bands of low and high molecular weight. Chromatography of the LcH-u material on a WGA column completely removed H-2 antigens (assayed by a sensitive radioimmunoassay employing broadly reacting sheep anti-H2 antiserum), but tumor rejection activity was found to be present in both the bound and unbound fractions. Serological analyses for the presence of viral structural proteins was negative. The detergent was removed from the unbound fractions of the WGA column and the material was then chromatographed on an Ultrogel AcA54 column. The material in the 60,000 MW range, representing the major protein peak, was assayed for tumor rejection activity at 0.5, 1.0, and 5.0 pg doses (Table 11). Two immunizations, 10 days apart, were followed by challenge with 2 X lo4 Meth A cells. Tumor rejection activity was retained in this major peak (Law el al., 1978a). It has been shown (Natori el al., 1977a) that the crude membrane (CM) and NP40-CS (solubilized) materials from Meth A induced immunity only against Meth A and not against other syngeneic neoplasms such as mKSA (an SV40 induced sarcoma) and LSTRA (an M-MuLV-induced lymphoma). Studies against two additional MC-induced sarcomas, CI-3 and CI-4 of BALB/c strain mice, which have their own strong TATA, also show specificity. In this study, also, immunizations were done with the material that was not bound by the lectin column. The unbound fraction (LcH-u) induced strong immunity but only against the Meth A sarcoma; thus, the LcH-u material retained the specificity of the intact Meth A cells and the crude membrane and crude soluble preparations. In the Meth A system, the results clearly show that the material solubilized is immunogenic in extremely small amounts and possesses the same specificity as the original tumor. It has not been possible, however,

TABLE I1 IMMUNOGENICITYOF MATERIALS FOLLOWING WGA Groupsa la Ib Ic 2 3a 3b 3c

Immunization material 0.5 pg x 2 WGA (unbound) 1.0 pg x 2 WGA (unbound) 5.0 pg x 2 WGA (unbound) 10.0 pg x 2 WGA (unbound) 0.5 pg x 2 AcA-54 (11) 1.0 pg X 2 AcA-54 (11) 5.0 pg x 2 AcA-54 (11)

AND

AcA54 CHROMATOGRAPHY

Percentage of control tumor volume 27 21 7 6

33 44 17

See Table I footnotes for details of immunization and challenge.

208

LLOYD

w.

LAW

et al.

to make a rigorous assessment of the recovery of antigen (as was possible with H-2) due to the lack of a specific antiserum. The in vivo assay does not lend itself to quantitation of TATA recovered in the various purification steps. Some attempt to estimate yield and specific activity of various purification steps was reported by Natori et al. (1977a). Antigen recognition assays, (e.g., delayed type hypersensitivity reactions and inhibition of macrophage migration), although showing tumor-specific responses to very small amounts of antigen, did not prove adequate for monitoring the fractionation of TATA (Dean et al., 1975, 1977; McCoy et al., 1977; Padarathsingh et al., 1978). The results reported by Baldwin et al. (1978) and Price and Baldwin, (1974a,b) (who have studied rat hepatomas and sarcomas induced by chemical carcinogens) are quite distinct from the results obtained in our laboratory concerning the biologic activities of isolated and solubilized tumor antigens from plasma membranes. Isolated plasma membranes of rat neoplasms, membranes solubilized by papain, ionic, and nonionic detergents, and fractions obtained by gel filtration retained a capacity to induce a specific humoral immunity detected by indirect membrane immunofluorescence, complement-dependent cytotoxicity, and radioisotopic antiglobulin assays. Tumor rejection immune responses were not usually induced, however, except with the intact or y-irradiated hepatoma or sarcoma cells. These differences in immune responses to soluble tumor antigens in the mouse and rat must reflect some basic difference in these antigens. A recent study was undertaken to resolve these differences. Treatment of irradiated Meth A cells of the mouse with glutaraldehyde and elevated temperatures [procedures known to abolish the in vivo protective capacity of D23 rat hepatoma cells in tumor challenge assays (Price et al., 1979a)l did not affect the tumor rejection capacity of Meth A sarcoma (Price et a/., 1979b). These same treatments in the rat system did not, however, impair antibody induction; the tumor-specific antigenicity as detected by serologic assays was maintained. These results highlight a difference between mouse and rat TATA in providing immunoprotection against tumor cell challenge. Extension of these studies may throw light on the question of equivalence of tumor,rejection determinants and serologically defined determinants and also on the roles of antigen presentation and processing for efficient responses to tumors (see Price et al., 1979b). A. ~ 5 3 A, COMMONTRANSFORMATION-RELATED ANTIGEN Although tumor rejection antigens (TATAs) that are distinct for each chemically-induced neoplasm have been defined for a large number of

TUMOR ANTIGENS ON NEOPLASMS

209

neoplasms, antibodies to these unique and distinct antigens have been extremely difficult to demonstrate. This situation stands in contrast to the detailed knowledge obtained of the surface antigens of neoplastic lymphoid cells of the mouse (Old and Boyse, 1973). Recently, however, DeLeo et al. ( 1978) have described syngeneic cytotoxic antisera that can detect antigens specific for the Meth A sarcoma or for the CMS4 sarcoma. These are new antigens not detected on any other carcinogen-induced neoplasm or normal tissue. In Table 111, a summary is presented of absorption studies performed to define the specificity of the syngeneic anti-Meth A antiserum. Since these serologic studies have defined a new antigen on Meth A sarcoma cells, we were interested in determining whether the serologically defined Meth A antigen was identical or related to the highly active and specific TATA of Meth A solubilized and partially purified in our laboratory. The molecules bearing determinants of Meth A and CMS4 were detected by immunoprecipitation of [35S]methionine-labeled extracts of cells (Fig. 1). From labeled extracts of Meth A sarcoma cells, a component of 53,000 daltons (p53) could be immunoprecipitated with anti-Meth A serum but not with normal mouse serum. A broad distribution of transformed and neoplastic cells of BALB/c origin were tested for the presence of p53. These immunoprecipitation assays have shown the presence of a cross-reacting antigen that has been further characterized as being transformed-related (DeLeo et a / . , 1979; Jay et al., 1980). The p53 antigen recognized by antisera prepared against Meth A or CMS4 was follnd in cells transformed by chemicals, radiation, viruses, or in spontaneously transformed cells. Of all normal tissue examined, it was found in low amounts only in the thymus (Jay et al., 1980). Although p53 is widely represented among mouse neoplasms, antibodies to p53 were found only in antisera prepared against Meth A and CMS4 and in sera obtained from either mice or hamster SV40-induced tumors. In studies to determine whether p53 of cells transformed by the DNA or R N A tumor viruses and by MC are identical, the p53 antigens immunoprecipitated from [35S]methionine-labeled extracts of different neoplastic cell lines have been compared using two-dimensional tryptic peptide analyses. Whereas the p53 fingerprints are similar for oncogenic virus-induced neoplasms (in transformed cells), the tryptic peptide analyses of four different MC-induced sarcomas demonstrate peptides that appear unique for each sarcoma (Jay et a / . , 1980). This latter finding resembles the findings of immunologically distinct TATA and individually distinct tumor surface antigens (TSSA), which are detected by specific cytotoxic antisera. It is not yet clear, however, if p53 does play a role in providing this kind of specificity. Experiments are in progress to study the interrelationships of p53, TATA, and TSSA

TABLE 111 ABSORPTION ANALYSIS OF (BALBIc x C57BLl6) F, ANTI-METHA SERUM" Positive absorption

Negative absorption Source of cells

Source of cells In vivo Meth A ascites Meth A solid

In vivo, normal Spleen

In vitro Meth A(a) Meth A(s)

In vivo, tumor Leukemias Chemically induced sarcomas

A,AKR,BALB/c, C57BU6,C3H/ He, DBA/Z,LP,MA/MY, NZB,RL,RF,SJL, SwissSWR ASLl ,RADAI, AKSL2,K36, RL 1, EL4,E GG2,ERLD, SL2,EARADl CMSl,3,4,5,7,14,21; B6MS1,2,4,5; BP8

In vitro, normal Adult lung fibroblasts

In vitro, tumor Chemically induced sarcomas

A,AKR,BALB/c, C3WH3,MNMy. PL

CMSl,2,3,4,5,7,8,9,10,11,13,14, 17,18,20,21; B6Ms2,5; Meth 4

Absorbed antisera assayed for residual cytotoxicity on Meth A(a) cultured cells. See DeLeo et al. (1977).

211

TUMOR ANTIGENS ON NEOPLASMS

FIG. 1. Autoradiogram of [35S]methionine-labeledproteins immunoprecipitated from extracts of Meth A sarcoma and adult fibroblasts by normal serum (NMS) or Meth A antiserum (aMeth A) and analyzed by SDS-polyacrylamide gel electrophoresis (See DeLeo et (11.. 1979).

in MC-induced neoplasms and to study the role of p53 in inducing and maintaining neoplastic transformation.

B. RELATIONSHIP OF TUMOR REJECTIONANTIGENS(TATA) ANTIGENS

AND

H-2

Some recent evidence suggests that genes of the H-2 complex of the mouse represent a source of genetic information for the tumor-specific

212

LLOYD

w.

LAW

et al.

rejection antigens (TATA) of chemically induced tumors. Derepressed H-2 components or H-2 components altered by mutation may function as TATA. Immunization with normal tissues from an incompatible strain has led to transplantation resistance against the syngeneic C- 1 fibrosarcoma, a methylcholanthrene-induced neoplasm of BALB/c mice (Parmiani and Invernizzi, 1975; Parmiani et d.,1978) and against several transplacentally produced lung tumors of C3Hf mice (Martin et al., 1976); in addition, transplantation immunity could not be induced against C- 1 in (BALB/c x C3H)FI mice but could be induced in other F, hybrids, suggesting that the H-2 antigens (or related antigens) of C3H mice, expressed on C-I, play a role in the host-antitumor immune response. We have investigated the properties of the antigens solubilized from purified cell membranes of the C-1 sarcoma of BALB/c mice in this laboratory. This tumor provides the most convincing evidence of the expression of alien H-2 antigens. Although originating in an H-2d strain mouse, this sarcoma apparently expresses, in addition to its normal complement of H-2 antigens, the antigenic specificities associated with the H-2Kk (H-2K. I , 5 , 11,23, and 25) and H-2Dk (H-2D.32) gene products (see Meschini et al., 1977; Rogers et al., 1979a, 1980). Appropriate absorption experiments have ruled out the possibility that the ability of these cells to absorb anti-H-2k antisera was due to viral antigens or to previously undetected public specificities. Furthermore, BALB/c mice immunized against C- 1 produce antisera having properties identical with anti-H-2K antisera produced in congenic mice. Membranes of C-I sarcoma were prepared by a modification of the method of Snary et al. (1974). These preparations were used to inhibit the precipitation of a purified H-2" molecule (Tanigaki et al., 1973). Complete inhibition of precipitation, especially by antisera against the private specificity, 23, provides rigorous proof that a structural antigen identical to H-2Kk is expressed on C-I, albeit at a somewhat lower density than on normal cells. Inhibition of the precipitation of labeled, purified H-2" in a radioimmunoassay (RIA) was used to monitor the fate of the alien antigen and of the normal H-2Dd antigen. As shown in Table IV, the specific activity and yield of the C-l alien molecule as assayed with specificity 25 (and also with H-2.23, I and 5) was strikingly lower than the normal H-2Dd antigen; decreasing the papain concentration did not improve the yield. Since papain digestion could not be used to solubilize the alien antigen from C-l sarcoma membranes, the combined methods of deoxycholate solubilization, gel filtration, and lectin affinity chromatography were used (Rogers ef ul., 1979a,b). These methods have been successfully used to purify H-2 and HLA antigens. At the same time, H-2k antigens were purified from a mixture of a BALB/c (H-2d)

213

TUMOR ANTIGENS ON NEOPLASMS

TABLE IV FATEOF H-2Dd A N D H-2.25 D U R I N G PAPAINDIGESTION OF MEMBRANES"

Protein Membranes

Fraction

(mg)

C-l Tumor

CMb Papain extract Papain pellet CM Papain extract CM Papain extract

11.6 2.0

C3H Liver mKSA Tumor

3.0 13 2.2 10 1.6

Specific activity

Yield

Units H-2 H-2Dd H-2.25

H-2Dd H-2.25

H-2Dd H-2.25

1392 758 348 -

1940 950

217
89 1664 732 -

120 378 116 -

194 593

18.7 <5

30 128 332

-

(%)

100

100

54 25

100

<5 41 100 44 -

49

-

-

-

See Rogers et ( I / . ( 1979a) for details. CM. crude membrane.

sarcoma ST2 and a C3H (H-2k) sarcoma C3H-7. The instability of the alien antigen dictated that purification be completed by immunoprecipitation. Although the molecular weight of the alien molecules was found to be identical to the normal H-2 molecules, alien and normal H-2 antigenic specificities did not coprecipitate, confirming the findings with papain digestion that the alien specificities are expressed on molecules distinct from the normal molecules. Except for the diminished expression of the C-1 molecule on the cell membrane and its sensitivity to papain, all other lines of evidence suggest that the alien H-2k molecules on the C-1 tumor are identical to their normal counterparts on the C3H-7 sarcoma. Most importantly, the alien molecule expresses the structural H2k antigens, which must indicate considerable sequence homology with a normal H-2 molecule. A summary of the biochemical and biological properties of the alien and normal H-2 molecules of C-l tumor cell membranes has been published (Rogers et l i l . , 1979a). The specific genetic mechanisms that could account for the appearance of alien phenotypes in allelic systems have been discussed by Bodmer (1973); the present evidence tends to favor a derepression mechanism. It should be stressed that the expression on tumor cells of alien antigens identified with the major histocompatibility locus of unrelated haplotypes (characteristic of the C-1 sarcoma) is probably a rare event. In this laboratory and in the laboratory of Old (DeLeo et al., 1978), alien specificities have not been detected by serologic or transplantation assays in a series of twelve methylcholanthrene-induced BALB/c sarcomas. In several instances, the apparent presence of inappropriate haplotypes on

214

LLOYD

w. LAW et al.

tumor cells, has been traced to artifacts in the assay systems, to the presence of C-type viral structural proteins (Klein, 1975), or to the existence on the target cells of previously undetected public specificities (see, for example, Prat et al., 1978; Flaherty and Rinchik, 1978).

OF TATA FROM C. ISOLATION A N D FRACTIONATION SARCOMA BEARING UNIQUE H-2 ANTIGENS

THE

C-I

There is considerable indirect biological evidence suggesting that the TATA of the C-1 tumor is not the alien H-2 antigen. Thus, in certain F, hybrid animals, an immune response was elicited against TATA in the absence of a humoral response against alien H-2 antigens (Parmiani et al., 1978). Also, alien H-2 antigens are lost during in vitro passage of C1, whereas TATA is retained (Carbone et al., 1978). Despite this evidence that alien H-2 molecules on the C-1 sarcoma are not functionally the tumor rejection antigen (TATA), more direct proof was necessary. We have, therefore, attempted to solubilize, separate, and enrich the membrane antigens from C-1 using the same techniques used successfully in separating TATA from H-2 antigens of leukemic cells (Rogers et al., 1978) and of the MC-induced sarcoma, Meth A (Natori et al., 1977a; Law et al., 1978a). Membranes prepared from solid tumor were solubilized with deoxycholate (DOC) and applied to an Ultrogel AcA34 gel filtration column. Both alien and normal H-2 antigens were assayed by inhibition radioimmunoassays using highly purified, radiolabeled H-2" antigen (H2Kk, H-2Dd) and appropriate antisera for detecting normal H-2'I and alien H-2k antigenic determinants (see Rogers et a l . , 1979a for details). Nine fractions from the Ultrogel AcA34 column were assayed for tumor rejction activity against C-1 challenge in BALB/c recipients. Inhibition of tumor growth, in the range of 75 to 90%, was obtained only with fractions that did not contain the bulk of the alien H-2 antigenic specificities. The lower molecular weight fractions also did not inhibit the growth of C-1. In a separate study, membranes of C-1 were subjected to NP40 solubilization and the NP40-soluble materials were fractionated on an Ultrogel AcA34 column. Fractions were checked for protein content and for both H-2" and alien H-2k activity. Pool I1 material that contained peak activities for H-2 alien (H-2k) and normal H-2 (H-2'9 antigens was subjected to Lens culinaris lectin chromatography. The material that did not bind and contained no detectable H-2 activity by RIA provided greater than 50% tumor inhibition against C-1 cell challenge. A fraction, Pool 111, which coeluted with '251-labeled bovine serum albumin and contained

215

TUMOR ANTIGENS ON NEOPLASMS

only very small amounts of normal and alien H-2 antigen activity, showed a striking tumor inhibition of C-1 (Table V). Neither the lectin-unbound material nor the NP40-soluble material induced tumor inhibition of Meth A, which has its own unique TATA, or against another MC-induced sarcoma of BALB/c mice, CI-4. Thus the specificity shown for immunization with intact or irradiated cells was also preserved in the NP40solubilized, chromatographed material. Thus, these studies (Law et ( I / . , 1980) show a clear separation of the normal and alien H-2 antigens from TATA activity. In this respect, the TATA from C-1 closely resembles the characteristics of the unique TATA isolated from MC-induced Meth A sarcoma, which is known not to contain alien antigens at least of the H-2k and H-2" haplotypes. The alloantigenic determinants H-2Dk and H-2Kk present on C-1 , therefore, do not appear to be suitable determinants for effective tumor immunity. Most recent studies by Parmiani ( 1979) have examined the concept that H-2-linked immune response genes at least partially control the ability of the host to respond to the TATA of C-1 sarcoma; the responses of different F, hybrids to C-1 immunization and the use of congeneic mice with recombinant H-2 haplotypes (including Ia regions) provide evidence of a genetic control by some Ia subregions of the in vivo responsiveness to TATA.

TABLE V TUMORINHIBITION ACTIVITY OF NP40 SOLUBILIZED A N D CHROMATOGRAPHED FRACTIONS OF C-1 SARCOMA

Group NP40 soluble Pool 111' Lectin (unbound) Controls

Immunization Ratio of number with dosage tumor to number challenged 2.5- 15b 2.5- 15 2.5- 10 None

11/32 16/32 13/24 16/16

Percentage of control tumor Immunogenic volume" indexd 18 29 45 100

5.6 3.5 2.3 -

Results at 20 days after challenge with 5 x 105 in vivo-passaged C-l cells. No striking differences in protection for 2.5, 5.0, 10.0 fig groups; results are therefore combined. Pool 111 material that coeluted on Ultrogel AcA34 column with 1Z51-labeledbovine serum albumin; M W ,60,000-70,000. Immunogenic index is defined as mean volume of control tumor per mean volume of experimental tumor

216

LLOYD

w.

LAW

et al.

Ill. TATA of Neoplasms Induced by the DNA Oncogenic Viruses, SV40 and Polyoma Virus

Most virally induced neoplasms express antigens directly coded for by the viral genome and, in addition, express other antigens associated with the transformed phenotype. The identification and isolation of antigens from these tumors has generally involved one of two approaches. In the first approach, a viral gene product has been identified as a component of the transformed cell, purified, and tested for antigenicity, using a variety of in vivo and in vitro assays of immunity. In the second approach, tumor antigenicity is demonstrated in a functional test such as tumor rejection, cell-mediated cytotoxicity or complement-dependent cytotoxicity. This functional test then becomes the assay used to monitor the isolation and purification of antigen. Both approaches have been used with varying degrees of success in a number of systems. In this discussion, we will concentrate on sarcomas induced by SV40 and polyoma virus, where the results have been most encouraging and best demonstrate the merits of each approach. The discussion will be largely confined to tumor-associated antigens of the transplantation type, TATA. The TATA of SV40 were originally demonstrated by immunization of hamsters with virus and protection against subsequent challenge with a transplantable tumor or by the prevention of the viral induction of tumors in hamsters by immunization with killed tumor cells (Giradi, 1965; Habel, 1962; Defendi, 1963). Cells transformed by SV40 and by polyoma virus also express a nuclear antigen, T antigen, identified by its reaction with sera from tumor-bearing hamsters. The failure to observe host cell-dependent changes in TATA or T antigen suggested that they were coded for the viral genome. SV40 TATA has been distinguished from T antigen by its different temperature sensitivity, its behavior in studies with temperature sensitive mutants (Black et al., 1963; Anderson et al., 1977a,b), and its separation from T antigen in the adenovirus-2-SV40 hybrid designated Ad2+-ND2 (Lewis and Rowe, 1973). Nevertheless, the observation that both T antigen and TATA were coded for by the early region of the SV40 genome (Lewis and Rowe, 1973; Girardi and Defendi, 1970) suggested that they might be related. Early attempts at purification of TATA from SV40-transformed cells began with cell fractionation and solubilization and used tumor rejection as an assay (Coggin et al., 1969; Tevethia and Rapp, 1966; Smith et al., 1970; Drapkin et al., 1974; Appella et al., 1976; Henriksen et al., 1977; Natori et al., 1977b). The latter two groups of investigators attempted to use quantitative assays for estimating recovery and specific activity. Crude membranes from the tissue culture-adapted mKSA sarcoma of BALB/c mice were solubilized with papain and the soluble material

TUMOR ANTIGENS ON NEOPLASMS

217

further fractionated through DEAE-cellulose, G- I50 Sephadex, and Con A-Sepharose columns. The most active fraction was that obtained from the Sephadex (3-150 column (Fraction I) with a molecular weight range of 40,000-50,000. Less than 2 pg of protein of this material provided greater than 70% tumor inhibition, and reasonable dose-response curves were obtained. H-2 activity was completely separated from tumor rejection activity by Con A-Sepharose fractionation. The data indicated the existence of fragments of different molecular weights after papain solubilization. There was also a lowered yield of activity in the solubilized material: 2-3% of that obtained from membranes compared to 34% for H-2 antigens. Nevertheless, the solubilized, chromatographed material had the immunogenic characteristics of the TATA of the intact cell in that a strong tumor rejection activity that was specific for the mKSA sarcoma was retained (Henriksen et al., 1977). The detergent Nonidet P40 was also used in our laboratory to solubilize the TATA from crude membranes of mKSA (Natori et al., 1977b). The solubilized material, applied to an Ultrogel AcA22 column, yielded a fraction that was strikingly immunogenic and had a higher specific activity than the crude solubilized material. This fraction, as well as the solubilized antigen, showed a clear dose-response relationship, and specific immunization was obtained only against mKSA. Thus, although the starting material (crude membranes) contained only a part of the TATA, the activity was stable and the specific activity increased during gel filtration. The TATA activity of this fraction was completely separated from H-2 activity following Lens culinaris lectin affinity chromatography. Recently, the molecular weight of immunoprecipitated T antigen was shown to be between 70,000 and 100,000 (Ahmad-Zadeh et al., 1976; Carroll and Smith, 1976; Tenen et a l . , 1977; Del Villano and Defendi, 1973). The same observation was also made for the T antigen of the closely related polyoma virus (Ito et al., 1977a). Since this molecular weight accounts for the entire coding capacity of the early region of the SV40, both T antigen and TATA must be structural antigens on the same polypeptide chain (Anderson et al., 1977a). This suggestion was supported by the following observations: ( I ) Nuclear preparations from SV40-transformed cells that contain all of the T antigen contained far more TATA than plasma membrane preparations (Anderson et al., 1977a; Rogers et al., 1977~).(2) A temperature sensitive SV40-transformed rat embryo cell lost T antigen and TATA activity in a parallel manner when shifted to the nonpermissive temperature. (3) The 28,000 MW fragment of T antigen expressed in the adenovirus-SV40 hybrid, Ad2+NDl, was shown to possess TATA activity (Jay et al., 1978). In retrospect then, it is clear that the early approaches to the purification of SV40 TATA were not ideal since they began with membrane

218

LLOYD

w . LAW et al.

fractions, and most of the TATA was probably in the nucleus. With the realization that TATA and T antigen were on the same molecule, the strategy changed. T antigen could be purified using complement fixation or immunoprecipitation as assays; these assays were certainly simple and more quantitative than the immunization and tumor rejection assay for TATA. T antigen-enriched fractions could then be assayed for TATA activity. This approach has in fact been taken (Chang et al., 1977; Anderson et al., 1977a; Mora et a l . , 1977) using the purified materials of Henderson and Livingston (1974). TATA and T antigen have been shown to copurify through several steps of purification, and the purest available preparations induce tumor immunity in mice at low concentrations (Chang et a / . , 1979). The genetic and biochemical evidence to date, therefore, support the conclusion that a single 100,000 MW protein coded for by the early region of the SV40 genome comprises both T antigen and TATA. The immunogenic properties of the soluble antigen obtained without resorting to solubilization procedures have also been thoroughly described (Rogers and Law, 1979). The association of TATA with a protein found in the nucleus of SV40 transformed cells raises certain conceptual difficulties, since the TATA is demonstrably immunogenic and immunosensitive, as shown by tumor rejection and tumor neutralization assays, and therefore must be present on the plasma membrane to interact with the immune system. Indeed, although as mentioned above, antiserum to T antigen does not react with the membranes of polyoma or SV40-transformed cells, antigens can be demonstrated on the surface of these cells by their sensitivity to cellmediated cytotoxicity (Trinchieri et a / . , 1976: Warnatz and Krapf, 1976; Gooding, 1977; Tevethia et ul., 1970; Walia and Lamon, 1977), to complement-mediated cytotoxicity (Luborsky et al., 1976), and to immunofluorescence (Meyer, 1971; Barra et al., 1977; Tevethia and Tevethia, 1975). To resolve this dilemma, it has been suggested that the T antigen is synthesized in the cytoplasm and then transported to the plasma membrane or the nucleus. It is postulated that when the protein is transported to the plasma membrane, an additional processing step that results in the loss of T antigen but retention of TATA occurs (Anderson et al., 1977a). Alternatively, it is possible that the unaltered molecule is present on the membrane at a concentration high enough for TATA to stimulate the immune system but too low to permit detection of T antigen by any of the assays available (Rogers et al., 1977~). Some support for the former hypothesis came from the work of Deppert and Walter (1976), who observed that peptides coded for by the early region of SV40 and having molecular weights of 46,000 and 52,000 could be found in the plasma membranes of HeLa cells infected with adenovirus-SV40 hybrids. Soule and Butel (1979), using careful cell fraction-

TUMOR ANTIGENS O N NEOPLASMS

219

ation procedures, have reported detection of low levels of a protein on the plasma membrane that was indistinguishable from nuclear T antigen. It has also been observed (Ito et a / . , 1977b; Silver et al., 1978) that cells infected or transformed by polyoma virus expressed in their plasma membranes a 50,000 MW peptide coded for by the early region of the viral genome. Interestingly, in both cases the peptides were precipitated by anti-T antigen sera from tumor-bearing animals, although the plasma membranes of SV40 or polyoma-transformed cells do not show immunofluorescence with the same antiserum in the classic assay for T antigen. It would be of great interest to determine if these plasma membrane peptides show TATA activity. In the case of polyoma virus, it has already been reported (Rogers et al., 1977d) that the plasma membranes of polyoma-transformed cells contain a greater proportion of the cellular TATA than the plasma membranes of SV40-transformed cells. This observation correlates with the fact that peptides coded for by the early region of the viral genome are present in the plasma membrane of cells transformed by wild type polyoma virus, but have not yet been detected in wild type SV40-transformed cells. The availability of deletion mutants of polyoma virus that do not express these membrane proteins (Ito et al., 1977b) should permit a more critical test of whether they are the TATA. It is clear from the above discussion that a central problem in the study of papova virus TATA is to identify the molecular configuration of the protein that is coded for by the early region of the viral genome and exists on the plasma membrane. However, there are other important questions that also remain to be answered. For example, is this virtually coded product the only molecule functioning as TATA, or are there other antigens as well? Also, it has recently been reported that T antigen and TATA of SV40-transformed cells contain species-specific determinants (Law et al., 1978b; Beth et al., 1977; Deichman et al., 1977). This observation raises the intriguing possibility that host cell-dependent posttranslational modification of the early region gene product may influence its antigenicity. Thus, although SV40-induced tumors stand alone as the only instance in which a single polypeptide has been unequivocally shown to function as TATA, the description of the TATA, even in this relatively simple system, is far from complete. IV. TATA of Neoplasms (Leukemias) Induced by RNA Tumor Viruses

Leukemic cells and tumor cells that have been induced by type C RNA viruses express viral structural antigens on their plasma membranes. These antigens have been demonstrated on cells transformed by murine

220

LLOYD

w. LAW et al.

and avian oncornaviruses (Ikeda et al., 1974; Bauer, 1975) and are expressed whether or not the cells are producing infectious virus (Bilello et al., 1974). Several investigators have demonstrated that these antigens are usually expressed on high-molecular-weight precursor polypeptides when they are found on the plasma membrane (Ledbetter et al., 1977; Lejneva et al., 1976; Tung et al., 1976). A great deal of work has focused on the role of these structural antigens in the antitumor immune response. Both cellular and humoral immunity against p30 and gp70 have been demonstrated in the mouse primarily by using in vitro assays that measure tumor-cell killing (Levy and Leclerc, 1977). Similar results have been obtained in the rat using model tumors induced by the Gross murine leukemia virus (G-MuLV) (Knight et al., 1975). Active and passive immunization against gp70 was found to protect mice against massive infection with Friend or Rauscher virus (Hunsmann et al., 1975). Passive immunization with antiserum to feline viral gp70 protected kittens from developing sarcomas induced by feline sarcoma virus (FeSV) (DeNoronha et al., 1977). These results suggest that viral structural antigens can function a s targets for the immune system. On the other hand, attempted immunizations of mice with p30 and gp70 against subsequent challenge with transplantable lymphocytic leukemias induced by Friend or Rauscher MuLVs have been unsuccessful (Rogers et al., 1977a). In addition, no correlation was found between the expression of viral structural antigens and an increase in immunosensitivity observed during passage of Rauscher murine virus-induced leukemia cells (RBL-5) in vitro (Liu et al., 1977). Repeated immunoselection of a Moloney murine leukemia virus (M-MuLV)-induced lymphoma led to the loss of a Moloney cell surface antigen (MCSA) (recognized by a syngeneic antiserum produced against irradiated cells), but retention of viral structural proteins, suggesting that the host also recognizes a virally induced antigen different from the known viral structural antigens (Fenyo and Klein, 1976; Fenyo, 1978). Membranes from these lymphoma cells were subsequently solubilized with detergents, and MCSA (in several molecular forms) was separated from the known viral structural antigens by standard biochemical techniques (Siegert et al., 1977). Distinctions between cell-surface antigens recognized by the host immune system and viral structural antigens have also been noted in the feline leukemia (Stephensod et al., 1977) and avian sarcoma (Kurth and Bauer, 1972) systems. Thus, although the viral structural antigens may be recognized by the host immune system, these results suggest that virally induced antigens distinct from the viral structural proteins are also recognized. We have published data to support this contention; we have demon-

TUMOR ANTIGENS O N NEOPLASMS

22 1

strated that the tumor-specific transplantation antigen (TATA) of leukemia RBL-5 can be solubilized and separated from the viral structural proteins p30, gp70, and p15E, as well as from the host histocompatibility antigens (H-2). RBL-5 is an R-MuLV-induced lymphocytic neoplasm (McCoy et al., 1967) that has strong cross-reacting antigens [including tumor rejection antigen (TATA)] with other Friend-Moloney-Rauscher (FMR) group-induced lymphomas. Plasma membranes were prepared from RBL-5 ascites cells using a Stansted cell disruptor and differential centrifugation as described by Snary et al. (1974). Solubilization was carried out immediately after preparation of membranes using 2% sodium dexoycholate, with more than 90% solubilization of the membrane protein. The solubilized membranes were applied to an Ultrogel AcA34 column. Column fractions were individually assayed for H-2, gp70, and p30 and then were arbitrarily pooled into nine fractions and diluted to appropriate protein concentrations for immunization of mice to study tumor rejection activity and in addition to study inhibition of the cytotoxicity induced by a syngeneic antiserum. TATA-containing fractions from the Ultrogel AcA34 column were pooled and applied to an affinity column prepared from Lens culinaris lectin coupled to Agarose, and proteins that bound to the lectin were collected. H-2 antigen, p30, and gp70 were determined by radioimmunoassay. TATA activity was assessed by subcutaneously immunizing groups of six to eight 2-month-old female BALB/c x C57BL/6 F, (CBF,) mice. Two immunizations were given at 10-day intervals. Ten days after the final immunization, the mice were challenged subcutaneously with 5 X 105 RBL-5 tissue culture cells. The TD,, of this in vim-adapted line was 1 x lo4 cells. These cells were found to have decreased oncogenicity compared to the ascites cells (Liu et al., 1977) and gave more reproducible results in the in vivo assay. The Ultrogel AcA34 fractions were also assayed for their ability to inhibit the cytotoxicity of our syngeneic anti-RBL-5 crude membrane antiserum. This antiserum was obtained by hyperimmunization of CBF, animals with crude membranes from RBL-5 ascites cells, and its reactivity has been described in detail elsewhere (Rogers et a / . , 1977b). Separation of TATA, H-2, p30, gp70 and the Target Antigen of Syngeneic Anti-RBL-5 Membrane Antiserum by Gel Filtration. The purity and yield of the plasma membranes were estimated by assaying for H-2 antigen at each stage of purification. The overall yield of H-2 was 2540% and the specific activity was increased 20- to 30-fold over cells.

222

LLOYD w . L A W

et al.

Since the plasma membrane comprises about 2% of the total cellular protein (Steck and Fox, 1972), about 50-fold purification is expected. The yields and purification of p30 and gp70 were lower because, unlike H-2, these antigens are not confined to the plasma membrane, but may also be present on cytoplasmic precursors (Barbacid et al., 1976; Arcement et al., 1977; Karshin et al., 1977) and on virus particles lost during isolation of the plasma membrane. Solubilization of these membranes with 2% DOC resulted in 95% solubilization of protein and 80% solubilization of H-2, p30, and gp70, as judged by radioimmunoassay . The TATA activity was also solubilized, but the recovery after solubilization could not be assessed precisely because of difficulty in quantitating the in vivo assay. However, in two independent experiments, the lowest immunizing dose (0.5 pg) of irradiated RBL-5 cells, plasma membranes, and solubilized plasma membranes provided comparable protection against tumor challenge, resulting in >90% reduction in mean tumor volume compared to controls. The yield after solubilization of the antigen which is the cytotoxic target of the syngeneic anti-RBL-5 membrane antiserum, was 30% as judged by inhibition of cytotoxicity. The behavior of all these antigens during gel filtration chromatography of the solubilized plasma membranes on a Ultrogel AcA34 has been described in detail elsewhere (Rogers et a / . , 1977b), and a summary is presented in Table VI. The TATA-containing fractions were clearly separated from fractions containing H-2, p30, gp70, and the cytotoxic target antigen of the syngeneic serum. Significantly, the cytotoxic activity of the syngeneic antiserum could be completely absorbed by R-MuLV, but not by purified gp70 (Rogers et al., 1977b). In addition, this antiserum was capable of lysing R-MuLV in the presence of complement, an activity previously restricted to antisera prepared against gp70 and p 15E (OroszIan and Gilden, 1977). On the basis of these data, it was tentatively concluded that the target of the syngeneic anti-RBL-5 membrane antiserum was p15E. Thus, in a single chromatographic step, the TATA of RBL-5 was separated from the bulk of H-2 antigen and from the three viral structural antigens most often implicated in antitumor immune responses: p30, gp70, and pl5E. The TATA obtained from the Ultrogel AcA34 column retained the characteristics of the TATA present on irradiated RBL-5 cells. It was specific for RBL-5 and other leukemias induced by viruses of the FriendMoloney-Rauscher group such as LSTRA and YAC and showed no cross-reactivity with other tumors of viral, chemical, or spontaneous origin such as Meth A, mKSA, EdG2 and P1798; the latter two are leukemias, but not of the FMR group.

TABLE VI ACTIVITIES OF DOC SOLUBILIZED RBL-5 MEMBRANES OBTAINED FROM ULTROGEL AcA34 COLUMNCHROMATOGRAPHY

AcA34 Pooled fraction

Average apparent molecular weight

I 2 3 4

>350,000 250,000 200,000 130,000 w000 80,000 70,000 30,000 12,000

Percentage of reduction of control mean tumor Percentage of total volumeo recovered protein ~

5 6 7 8 9

9.3 13.9 11.8

10 3 0 0 0

8.7 37.8 4.9 7.4 4.7 1.4

Percentage of total units recovered from AcA34 columnb H-2

p30

gp70

0.9 3.5 4.6 6.7

0.4 3.6 2.4

0 0 1.3

~~

Syngeneic serum target antigen ~~

7.6

<. I <.01 <.01

4.5

<. 1

6.8 2.4 0

15.8 0.3

2.2 0

In this experiment, 0.5 pg of each fraction was used to immunize. The control mean tumor volume was 320 mm3.

6 2.4

* For protein, H-2, p30, and gp70, the area under the elution profile was calculated for each fraction. For the inhibition of the syngeneic antiserum, a dose response was done for each fraction. The fractions enclosed in boxes comprise the peak(s) of activity of each of the antigens.

224

LLOYD

w. LAW et al.

Although the bulk of H-2, p30, and gp70 antigens were separated from TATA by gel filtration, a small percentage of each antigen was present in fractions containing TATA. Thus, it is possible that a small fraction of one of these antigens (perhaps somehow complexed to other proteins) could function as TATA. To investigate this possibility further and to continue the purification of the TATA, fractions 2, 3, and 4 from the AcA34 column were pooled and applied to a Lens culinaris lectin affinity column. Fraction 1 was omitted because the TATA in this fraction seemed to be aggregated (Rogers et al., 1977b). The results of this affinity column have been summarized (Rogers et al., 1978). Seventy-eight percent of the applied protein did not bind to the lectin, and the remainder was bound and eluted with 3% a-methylmannoside. Both the bound and unbound fractions contained TATA activity. Furthermore, radioimmunoassays demonstrated that more than 90% of the H-2 and gp70 present in the pooled material from Ultrogel AcA34 was retained by the lectin column and eluted with a-methylmannoside, while more than 90% of the p30 was not retained. Estimates of the molecular weight are complicated by the fact that TATA is either incompletely insolubilized or tends to aggregate, as judged by the fact that a relatively large fraction of the TATA was always found at the excluded volume after gel filtration, whereas the other antigens assayed were not. Attempts to eliminate this aggregation by mild reduction, that is, by treatment of the solubilized membrane with 2 m M dithiothreitol (DTT) for 30 minutes at room temperature, and chromatography in 2 m M DTT resulted in the loss of more than 90% of the TATA activity. These results suggest that the TATA may contain easily reducible sulfhydryls that are important for antigenic activity. Our findings to date may be summarized as follows. Previous work from this laboratory (Rogers et al., 1977b; Chang et al., 1975) strongly suggested that the TATA was distinct from the major viral structural antigens. TATA could be solubilized from crude membranes of the lymphocytic leukemias RBL-5 and FBL-3 using papain digestion. The trace amounts of gp70 and p30 present in the papain extract could not account for the TATA activity of this fraction. Indeed, in spite of the.fact that the extract retained considerable TATA activity, more than 90% of gp70 and p30 activity was destroyed and the rest reduced to fragments of MW less than 20,000. The data presented from our later studies confirm and extend these conclusions. TATA can be clearly separated from the bulk of H-2, p30, gp70, and probably p15E by gel filtration. More precisely, the TATA of RBL-5 is different from the 70,000 and 90,000 MW species of gp70, the 90,000, 70,000 and 30,000 MW species of p30 and a low-molecular-weight form of p15E. However, the possibility remains that the small amounts

TUMOR ANTIGENS O N NEOPLASMS

225

of viral antigen present in TATA-containing fractions are somehow complexed with other proteins to yield the TATA. The behavior of the TATA on the Lens cufinaris lectin affinity column argues against this possibility. Virtually all of the H-2 and gp70 are retained on the column and a substantial amount of TATA activity was not retained; gp70 and H-2, therefore, cannot be part of the unbound fraction of TATA. Similarly, since virtually all of the p30 is not retained on the column, that portion of the TATA retained by the column cannot contain p30. One could argue that there are two species of TATA: one a complex with gp70 and one a complex with p30. However, the most reasonable explanation of the data is that TATA contains no viral structural antigens and does not bind completely to the column for other reasons, e.g., incomplete glycosylation. A more recent study from our laboratory (Alaba et al., 1979) has addressed the question of the role of heterocomplexes involving gp70, p30, or H-2 antigens acting as the tumor rejection antigen. In the isolation of TATA, the DOC-solubilized material was sequentially chromatographed on anti-gp70, anti-p30, and anti-H-2b affinity columns. There was no observed loss of activity either in tumor rejection or in specific secondary sensitization assayed by CTL (cytotoxic T lymphocyte) activity. It should be stressed that the target of the syngeneic, cytotoxic antiserum produced against RBL-5 membranes is distinct from TATA. This is shown not only by the separate elution of this antigen and TATA from the gel filtration column, but also by the fact that the mild reducing conditions that drastically reduced TATA activity had little effect on the inhibitory activity of pool 8 material (Table VI). A possible indication of the identity of this antigen was provided by the virolysis experiments, which suggested that the target of the syngeneic antiserum was p15E (Rogers et al., 1978). The elution position of the fraction (pool 8) inhibiting the syngeneic serum from the Ultrogel AcA34 column is consistent with this possibility. It should be pointed out, however, that these results would also be obtained if the viral envelope contained tumor-specific cell components, distinct from viral proteins and TATA, but recognized by the anti RBL-5 membrane serum. In any case, the antigen recognized by this serum is clearly different from MCSA, the cell surface antigen recently partially purified by Siegert ef al. (1977). The molecular weight is lower (30,000 compared to 110,000), and the antiserum reacting with MCSA does not recognize p I5E. V. Soluble Antigens and Immune Deviations of the Host

Several investigators have reported immune deviations of the host following inoculations of solubilized antigens from membranes of tumor

226

LLOYD

w.

LAW

et al.

cells (see particularly Pellis and Kahan, 1975; Baldwin and Robins, 1977; Baldwin et al., 1973; Paranjpe and Boone, 1975; Rao and Bonavida, 1977; Bubenik et al., 1978). The following have been described: development of a "null" state following large doses of antigen; inhibition of growth attained only over a very restricted antigen dose range; a rapid decay of the immune state; facilitation (rather than inhibition) of tumor growth; and specific inhibition of cell-mediated immune responses. These immune "deviations" have been attributed to qualitative changes in the antigen resulting from processing or to inappropriate responses on the part of the host as a result of the differing form in which the antigen is presented; these deviations result in the induction of tumor-specific antibody rather than the appropriate cell-mediated response. Most of these soluble or solubilized preparations studied, however, did not produce tumor rejection immunity, were feebly immunogenic under restricted conditions, or were not assayed for in vivo immunogenic potential (see Price and Robins, 1978). It was of interest therefore to look for inappropriate immune responses in BALB/c mice with our solubilized antigens from the Meth A and SV40 sarcomas, since both were shown to be strikingly immunogenic in tumor challenge experiments. With Meth A soluble antigen, we did not observe "overloading" nor the development of a "null" state. Total doses of antigen ranging from 2.5 pg to 2000 pg protein per mouse all produced immunity to a challenge with 2 x lo4 Meth A cells (TDloo); nor was there an optimal or a restricted tumor dose at which the antigen was most, or only, effective. BALB/c mice immunized with 50 pg X 2 of NP40-solubilized Meth A material (CS) were shown to be solidly immune when challenged 60 days after the last immunization; this result indicates a stable immune state by soluble antigen. Spleen cells from mice, similarly immunized, were found to maintain the capacity to neutralize Meth A sarcoma cells (measured in the Winn assay) 150 days after immunization, without an intervening challenge of tumor cells (Law et al., 1978a). We did not observe any modification of the tumor rejection response through pretreatment of BALB/c mice with NP40-solubilized Meth A material prior to immunization; such a modification was reported by Baldwin et al. (1973), who employed soluble o r membrane preparations of a rat hepatoma. These preparations of Baldwin et af. did not, however, have tumor rejection activity. As mentioned previously, the protein of SV40-transformed cells containing both T antigen and TATA, is located primarily in the nucleus. Thus, vigorous disruption of dissociated cells will release the antigen into the high speed supernatant fraction. Solubilization with detergents or with proteolytic enzymes that might degrade or otherwise alter the prop-

227

TUMOR ANTIGENS O N NEOPLASMS

erties of TATA, was therefore not necessary in order to obtain soluble TATA for study. Table VII summarizes the fate and immunogenic activities during the disruption and differential centrifugation of mKSA sarcoma cells. Nearly complete inhibition was obtained with a single immunization of 1 pg protein of the high speed centrifugate; also, T antigen was found predominantly in this material and was enriched 4-fold over the activity of nondisrupted cells. This material, when subjected to gel filtration chromatography, yielded a fraction (MWin the range of 70,000lOO,OOO) that contained the bulk of T antigen activity and tumor rejection activity; 5 and 10 pg protein provided complete protection against tumor challenge and 1 pg provided greater than 95% protection. The host reactions studied here in response to the highly immunogenic and antigenically specific fraction of soluble antigen from mKSA do not give any evidence of unfavorable immune responses (Rogers and Law, 1979). Antigen “overloading” or the induction of a “null” state and a restricted range of activity of the antigen depending upon the burden of tumor challenge [specifically described by Pellis and Kahan (1975)] were not observed. Depression of anticellular immune responses have been reported by Paranjpe and Boone (1975), Baldwin and Price (1976), Embleton (1976), Rao and Bonavida (1977), and Bubenik et al. (1978) with solubilized materials in different assays, including the W r release cell-mediated assay, the microcytoxicity assay, the radioisotopic footpad assay (IFP),

FATEA N D ACTIVITY OF TATA

AND

TABLE VII T ANTIGENAFTER DISRUPTION OF MKSA CELLS

T Antigen activity Fraction Cells Low speed pelletb (nuclei) High speed pellet (membranes) High speed supernatant

Total units units/mg

TATA activity (%)” 1 pg

5 pg 10 p g

80,000 7.050

98 29

50

90


CI
S ,OOo

36

25

4


68,900

427

13


CI

Eight BALB/c in each group were immunized once with the indicated dose in PBS and challenged 10 days later with S x lo4 (SO0 X TD,,) mKSA cells. TATA activity is expressed as the percentage of mean tumor volume of controls. CI, complete inhibition. The nuclear and membrane pellet and high speed supernatant preparations used here had been stored for 6 months at -20°C. Similar results were obtained with three different preparations. (I

228

LLOYD

w. LAW

et

al.

or in studies of the interruption of tumor immunity in vivo. Inoculation of the extracted materials (in most instances, in large amounts) were stated to suppress in a specific fashion cellular immune responses against the tumor employed in preparing the extracts. In contrast to these findings, we were unable to detect any modification or impairment of cellmediated responses using the sensitive and specific IFP assay, nor did pretreatment in vivo with our highly immunogenic soluble fraction render mice incapable of producing a tumor rejection response when subsequently immunized. Activity of soluble antigen was virtually indistinguishable from the activity of intact cells when tested for immunogenicity in the tumor neutralization (Winn) assay. Specifically, the optimal time for sensitization was longer for soluble antigen than for cells, but multiple doses of soluble antigen shortened the time for sensitization rather than extending it further, or abrogating immunity altogether. There are several possible explanations for the contrasting results obtained with soluble antigen in our systems and in the other systems mentioned above. The trivial explanation is that the immunosuppressive effect of soluble antigen is not universal and that the mKSA and Meth A systems are unaffected. It should be noted however, that loss of specific tumor immunity (eclipse) is observed in animals bearing large mKSA tumors (Howell et al., 1975),and this is one observation that experiments with soluble antigen sought to explain. Another explanation is that the extraction procedures employed to generate soluble antigen in other systems somehow altered the antigen in a way that endowed it with immunosuppressive properties or with properties that stimulate unfavorable immune reactions. This seems unlikely, however, because in a tumor-bearing animal, the soluble antigen is ultimately generated by hostmediated processes, and these processes should be the same for all tumor antigens whether they are “shed” by the tumor o r administered in soluble form in an experimental situation. Moreover, if in vitro extraction does alter the antigen in a way that is not duplicated in the host, then the immunosuppressive activity resulting from such alterations would not be important in vivo. A third explanation is that the tumor antigen (or material) responsible for immunosuppression is distinct from the TATA. The soluble fraction utilized in the present studies might not contain such an antigen if it were found exclusively on the plasma membranes. This concept might also explain why many of the soluble antigens with immunosuppressive activity are not immunogenic. Rigorous proof of this concept requires the biochemical separation of the two antigens; this has not yet been achieved. Finally, it is possible that immunosuppressive effects generated against TATA are weak and would be observed only

TUMOR ANTIGENS ON NEOPLASMS

229

with tumors that are also weakly immunogenic. This notion is consistent with the weak or nonexistent immunogenicity of tumor antigens demonstrating immunosuppression, as well as with the potent immunogenicity of the TATA of the mKSA sarcoma or of Meth A, where no immunosuppression is observed.

VI. Concluding Remarks

The study of tumor-associated antigens presents two major problems for the tumor biologist. The first concerns their chemical structure, and the second concerns the basis for their unusual interaction with the immune system: for example, how does a tumor survive in a host that has developed a demonstrable antitumor immune response? As the foregoing review demonstrates, experimental approaches to these problems have often involved solubilization and partial purification of tumor antigens, in particular, TATA. Implicit in our purification protocols is the notion that a particular assay monitors one or a few major antigens. This is by no means proven, and it is certainly possible that a given assay detects a number of antigens. This is particularly true of virally-induced tumors, where the proteins coded for by the viral genome and the tumor specific proteins both may function as antigens. Heterogeneity of the antigen may explain why such great losses of activity are encountered during purification. For example, in our experienceawith a number of tumors, it is not unusual to find that 105 irradiated tumor cells will completely immunize against subsequent challenge. This represents about 10 wg of protein and about 300 ng of plasma membrane. Once the cell is broken, however, on the order of 10 p g of purified membranes are usually required to immunize the animal or about 3 x lo6 cell equivalents. As purification continues, the preparation frequently becomes even less active. Of course, there are other possible explanations for the loss of activity: for example, more rapid degradation of soluble protein in the host or loss of orientation effects of the membrane on intact cells. Regardless of the cause, this loss of activity coupled with the potential heterogeneity of tumor antigens poses a most serious probelm to their purification. Nevertheless, it is clear that tumor specific antigens, TATA or TSSA, from several types of neoplasms of the mouse (those induced by chemical carcinogens and those induced by the DNA-containing and RNA-containing oncogenic viruses) can be obtained in soluble form and maintain the specific immunogenic activity of the intact cell. Chromatographic fractionation of the solubilized material allows for separation of TATA

230

LLOYD

w.

LAW et

ul.

or TSSA from other antigens, for example, from H-2 and from the MuLVrelated antigens. Enrichment of TATA has been achieved: however, purification such as that obtained with the H-2" and H-2" antigens (Henriksen er ul., 1978; Rogers er a/., 1979) remains a formidable problem. The molecular nature of these tumor antigens is unknown at the moment, but the recent development of highly sophisticated techniques such as the production of monoclonal antibodies is likely to be useful in characterizing these antigens. The serological study of nonlymphoid tumors is still at a very early stage and undoubtedly additional classes of tumor antigens will be uncovered. A remarkable feature of TATAs of chemically induced tumors is their polymorphism. The genetic origin of these antigens has been amply discussed, and a commonly held view is that they represent products of derepressed or mutated genes. Since these antigens share many characteristics with normal histocompatability antigens, the question has been raised whether TATAs are entirely new cellular antigens or the outcome of spontaneous point mutations at the level of genes coding for histocompatibility (e.g., H-2) antigens. The alien H-2 antigens on the C-1 sarcoma have now been defined serologically and biochemically and the results lead to the conclusion that the alien H-2 antigens on the surface of C-1 cells do not function as TATA. Further, the three classes of TATA obtained, respectively, from cell membranes of MC-induced sarcomas Meth A and C-I; from an SV40-induced sarcoma mKSA; and from RBL5, an R-MuLV-induced lymphoblastic leukemia, are all separable from H-2 antigens. Although there are indications (using rather insensitive assays) that tumor-associated antigens on human neoplastic cells are associated with p2 microglobulin (p2m), recent investigations with the tumor rejection antigen of RBL-5 leukemia (Rogers, 1979) have shown no association of p2m with TATA either as an intrinsic subunit or as a part of complex between H-2 and TATA; 60% of the p2m associated with the plasma membrane is bound to H-2" and 40% exists as a monomer. The problem of the striking species differences in tumor rejection antigen activity, at least in chemically induced neoplasms of the rat and mouse, requires further study. Solubilized tumor antigens of rat neoplasms are not immunogenic with respect to causing tumor rejection; yet, serologic reactivity is retained for the membrane-associated tumor-specific antigens (TSSA) and for alloantigens. As pointed out by Price et al. ( 1979b), this qualitative difference may reflect a difference in presentation and processing of antigen (TATA) in the induction of specific immunity. The identification and characterization of a 53,000 MW antigen (designated p53) by a syngeneic antiserum made against Meth A sarcoma

TUMOR ANTIGENS ON NEOPLASMS

23 1

cells may be a very important finding. There is a correlation between unshared tryptic map peptides obtained for p53 of four methylcholanthrene sarcomas and the presence on each of these tumors of a specific TATA. If more definitive evidence could be obtained that this variable portion of the p53 antigen plays a role in providing specificity in transplantation rejections, then a reasonable approach to the definition of the molecular nature and of the immunological specificity of TATAs will be available. There is ample evidence that TATA of mKSA, an SV40 induced sarcoma, is coded by the viral gene product and that the two activities, T antigen and TATA, are related products of a single genetic locus. In order for T antigen to function in the capacity of a transplantation antigen, it must be expressed in the plasma membrane. However, hamster antisera with high titers to T antigens do not bind to the surface of the cell. The relationship of this surface antigen to T antigen and TATA remains unknown. The cell surface antigens (CSAs) induced by MuLVs on leukemia cells and identified by in vitro assays (for example, membrane immunofluorescence, cell-mediated cytotoxicity, or antibody-mediated cytotoxicity) require additional characterization, particularly with respect to their biologic functions in tumor rejection and in the host response during oncogenesis. GCSA (Gross cell surface antigen) and MCSA (Moloney cell surface antigen) have been characterized from leukemic cells of the mouse, and their activity was associated with several molecular species of MW 150,000- 193,000, 85,000-92,000, and 45,000-52,000. GCSA appears to be identical to precursor gag proteins, and MCSA, which is similar to the FOCMA antigen of cat leukemia, appears to be linked on the surface of the cell to virion structural proteins (Kurth et al., 1979). The antigen of RBL-5, however, which has been characterized by a specific cytotoxic antiserum, appears to be located on a molecular species of MW 30,000. This antigen does not have tumor rejection activity and is completely separated by gel filtration from the region containing the RBL-5 specific TATA. This activity (TATA) for RBL-5 is not related to virus proteins or its precursors. Finally, it is important to realize that only some of the many tumorassociated antigens can act as TATA. One must also realize, however, that antigens that do not act as TATA may be useful markers. Further analysis of the nature of the immune response to tumor antigens, with emphasis on their molecular structure, will undoubtedly provide new methods for an immunological approach to the understanding and control of human cancer.

232

LLOYD

w. LAW et al.

REFERENCES Ahmad-Zadeh, C., Allet, B., Greenblatt, J., and Weil, R. (1976). Proc. Natl. Acad. Sci. U . S . A . 73, 1097. Alaba, O . , Rogers, M. J., and Law, L. W. (1979). Inr. J. Cancer 24, 608. Anderson, J . L., Martin, R. G., Chang, C., Mora, P. T., and Livingston, D. M. (1977a). Virology 76, 420. Anderson, J. L., Martin, R. G., Chang, C., and Mora, P. T. (1977b). Virology 76, 254. Appella, E., Law, L. W., and Henriksen, 0. (1976). Cancer Res. 36, 3539. Appella, E., DuBois, G. C., Natori, T., Rogers, M. J., and Law, L. W. (1978). In ”Biological Markers of Neoplasia: Basic and Applied Aspects’’ (R. W. Ruddon, ed.), p. 213. Elsevier, Amsterdam. Arcement, L. J., Karshin, W. L., Naso, R. B., and Arlinghaus, R. B. (1977). Virology 81, 284. Baldwin, R. W., and Embleton, M. J. (1971). Isr. J . Med. Sci. 7, 144. Baldwin, R. W., and Embleton, M. J. (1974). Isr. J . Med.Sci. 13, 433. Baldwin, R. W., and Price, M. R. (1976). Ann. N . Y . Acad. Sci. 276, 3. Baldwin, R. W., and Robins, R. A. (1977). In “Contemporary Topics in Molecular Immunology ” (H. N. Eisen and R. A. Reisfeld, eds.), Vol. VI, p. 177. Plenum, New York. Baldwin, R. W., Embleton, M. J., and Moore, M. (1973a). Br. J. Cancer 28, 389. Baldwin, R. W., Harris, J. R., and Price, M. R. (1973b). Int. J . Cancer 11, 385. Baldwin, R. W., Price, M. R., and Moore, V. E. (1978). In “Biological Markers of Neoplasia: Basic and Applied Aspects” (R. W. Ruddon, ed.), p. I I . Elsevier, Amsterdam. Baldwin, R. W., Glaves, D., Harris, J . R., and Price, M. R. (1979). Transplant. Proc. 3, 1189. Barbacid, M., Stephenson, J . R., and Aaronson, S . A. (1976). Nature (London) 262, 554. Barra, Y., Berebbi, M., Reynier, M., and Meyer, G. (1977). Colloq. Inst. SanrP Recherche Med. 69, 337. Bauer, H. (1975). Adv. Cancer Res. 20, 275. Beth, E., Cikes, M., and Schloen, L. (1977). Int. J . Cancer 20, 551. Bilello, J . A., Strand, M., and August, J . T. (1974). Proc. Natl. Acad. Sci. U . S . A . 71, 3234. Black, P. H . , Rowe, W. P., Turner, H. C., and Huebner, R. J . (1963). Proc. Narl. Acad. Sci. U . S . A . 50, 1148. Bodmer, W. F. (1973). Transplant. Proc. 5, 1471. Bubenik, J . , Indrova, M., Nemeckova, S., Malkovsky, M., VonBroen, B., Palek, U., and Anderlikova, J . (1978). Int. J . Cancer 21, 348. Burton, R. C., and Warner, N. L. (1978). Br. 1. Cancer 37, 159. Carbone, G., Invernizzi, G., Meschini, A., and Parmiani, G. (1978). Int. J . Cancer 21, 85. Carroll, R. B., and Smith, A. E. (1976). Proc. Natl. Acad. Sci. U . S . A . 73, 2254. Chang, C., Luborsky, S. W., and Mora, P. T. (1977). Nature (London) 269, 438. Chang, C., Martin, R. G., Livingston, D. M., Luborsky, S. W., Hu, C.-P., and Mora, P. T. (1979). J. Virol. 29, 69. Chang. K. S. S., Law, L. W., and Appella, E. (1975). Int. J . Cancer 15, 483. Coggin, J. H., Elrod, L. H., and Ambrose, K. R. (1969). Proc. Soc. Exp. B i d . Med. 132, 328. Dean, J. H., McCoy, J . L., Lewis, D., Appella, E., and Law, L. W. (1975). Int. J. Cancer 16, 465. Dean, J. H., Lewis, D. D., Padarathsingh, M. L., McCoy, J. L., Northing, J. W., Natori, T., and Law, L. W. (1977). Int. J . Cancer 20, 951.

TUMOR ANTIGENS ON NEOPLASMS

233

Defendi, V. (1963). Proc. Soc. Exp. Biol. Med. 113, 12. Deichman, G. I . , Klochareva, T. E., and Kashkima, L. M. (1977). Int. J . Cancer 20, 616. DeLeo, A. B., Shiku, H., Takahashi, T., John, M., and Old, L. J. (1977). J . Exp. Med. 146, 720. DeLeo, A. B., Shiku, H., Takahashi, T., and Old, L. J . (1978). In "Biological Markers of Neoplasia: Basic and Applied Aspects" (R. W. Ruddon, ed.), p. 25. Elsevier, Amsterdam. DeLeo, A. B., Jay, G. Appella, E., DuBois, G. C., Law, L. W., and Old, L. J. (1979). Proc. Natl. Acad. Sci. U . S . A . 76, 2420. Del Villano, B., and Defendi, V. (1973). Virology 51, 34. DeNoronha, F., Baggs, R., Schafer, W., and Bolognesi, D. P. (1977). Nature (London) 267, 54. Deppert, W., and Walter, G. (1976). Proc. Natl. Acad. Sci. U . S . A . 73, 2505. Drapkin, M.S., Appella, E., and Law, L. W. (1974). J. Natl. Cancer Inst. 52, 254. Economou, G. C., Takeichi, N., and Boone, C. W. (1977). Cancer Res. 37, 37. Embleton, M. J . (1976). Inf. J . Cancer 18, 622. Essex, M., Klein, G., Snyder, S. P., and Harrod, J . B. (1971). Nature (London) 233, 195. Fenyo, E. M. (1978). In "Biological Markers of Neoplasia: Basic and Applied Aspects" (R. W. Ruddon, ed.), p. 47. Elsevier, New York. Fenyo, E. M.. and Klein, G. (1976). Nature (London) 260, 355. Flaherty, L., and Rinchik, E. (1978). Nature (London) 273, 52. Girardi, A. J. (1965). Proc. N a f l . Acad. Sci. U . S . A . 54, 445. Girardi, A . J . , and Defendi, V. (1970). Virology 42, 688. Gooding, L. R. ( 1977).J . Immunol. 118, 920. Habel, K. (1962).J. Exp. Med. 115, 181. Hauschka, T. S . (1973). In "Principles of Immunology" ( N . R. Rose, F. Milgrom and C. J . Von Oss, eds.), p. 417. Macmillan, New York. Hellstrom, I., Hellstrom, K. E., and Pierce, G. E. (1968). Int. J . Cancer 3, 467. Hellstrom, K. E., Hellstrom, I., and Brown, J . P. (1978). Int. J . Cancer 21, 317. Henderson, C., and Livingston, D. M. (1974). Cell 3, 65. Henriksen, 0.. Law, L. W., and Appella, E. (1977). J. Natl. Cancer Inst. 58, 1785. Henriksen, O., Robinson, E. A., and Appella, E. (1978). Proc. Narl. Acad. Sci. U . S . A . 75, 3322. Holmes, E. C., Kahan. B. D., and Morton, D. L. (1970). Cancer 25, 373. Howell, S . B., Dean, J. H.. and Law, L. W. (1975). I n t . J. Cancer 15, 152. Hunsman, G., Moennig, V., and Schafer, W. (1975). Virology 66, 327. Ikeda, H., Pincus, T., Yoshiki, T., Strand, M., August, J. T., Boyse, E. A,, and Mellors, R. C. (1974). J. Virol. 14, 1274. Ito, Y., Spurr, N., and Dulbecco, R. (1977a). Proc. Natl. Acad. Sci. U . S . A . 74, 1259. Ito, Y., Brocklehurst, J . R., and Dulbecco, R. (1977b). Proc. Natl. Acad. Sci. U . S . A . 74, 4666. Jay, G., Jay, F. T., Chang, C., Friedman, R. M., and Levine, A. S. (1978). Proc. Natl. Acad. Sci. U . S . A . 75, 3055. Jay, G., DeLeo, A. B., Appella, E., DuBois. G. C., Law, L. W., Khoury, G., and Old, L. J . (1980). In "Symposia on Quantitative Biology.'' Cold Spring Harbor Laboratory, Cold Spring Harbor, New York (in press). Karshin, W. L., Arcement, L. J., Naso, R. B., and Arlinghaus, R. B. (1977). J. Virol. 23, 787. Klein, P. A. (1975). J. Immunol. 115, 1254. Knight, R. A., Mitchison, N . A., and Shellam, G . R. (1975). I n t . J. Cancer 15, 417. Kurth. R., and Bauer, H . (1972). Virology 47. 426.

234

LLOYD

w.

LAW

et al.

Kurth, R., Fenyo, E. M., Klein, E., and Essex, M. (1979). Noture (London) 279, 197. Law, L. W.. and Appella, E. (1975).I n "Cancer-A Comprehensive Treatise" (F. F. Becker, ed.), Vol. 4, p. 135. Plenum, New York. Law, L. W. Appella, E., and DuBois, G. C. (1978a). In "Biological Markers of Neoplasia: Basic and Applied Aspects" (R. W. Ruddon, ed.), p. 35. Elsevier, Amsterdam. and Ting, R. C. (1978b). I n ! . Law, L. W., Takemoto, K. K., Rogers, M. J . , Henriksen, 0.. J . Cuncer 22, 3 15. Law, L. W.. DuBois, G. C., Rogers, M. J., Appella, E., Pierotti, M . A . , and Parmiani, G. ( 1980). 7'ranspltrnt. Proc. (in press). Ledbetter, J., Nowinski, R. C., and Emery, S. (1977). J. Virol. 22, 65. Leffell. M. S., and Coggin, J. H . (1977). Cancer Res. 37, 41 12. Lejneva. 0. M., Abeleo, G. I., Dorfman, N. A.. Strand, M., and August, J . T. (1976). Virology 75, 28 I . Levy, J . P., and Leclerc, J. L. (1977). Ad\*. Crincer Res. 24, 2. Lewis, A . M., and Rowe, W. P. (1973). J. Virol. 12, 836. Liu. W. T., Rogers, M . J . , Law, L. W., and Chang. K . S. S. (1977). J . Nut/. Cuncer I n s t . 58, 1661. Luborsky, S. W., Chang, C., Pancake, S. J., and Mora, P. T. (1976). Biochem. Biophys. Res. Commun. 71, 990. McCoy, J . L . , Fefer, A., and Glynn, I. P. (1967). Cuncer Res. 27, 1743. McCoy, J. L.. Padarathsingh, M. L., Dean, J . H., Henriksen, 0.. Natori, T . , and Law, L . W. (1977). J . Immunol. 119, 306. Martin, W. J., Gipson, T . G.. Martin, S. E., and Rice, J . M. (1976). Science 194, 532. Meltzer, M . S., Leonard, E. J., Rapp, H. J . , and Borsos, T. (1971). J . Nail. Cancer I n s t . 47, 703. Meschini, A., Invernizzi, G., and Parmiani, G. (1977). I n t . J . Cuncer 20, 271. Meyer, G. (1971). A h . Cuncer Res. 14, 71. Mora, P. T.. Chang, C.. Khoury, G., Kuster, J. M., Luborsky, S. W . , and McFarland, V. W. (1977). Colloq. Inst. Sunte Recherche Med. 69, 327. Natori, T., Law, L. W., and Appella, E. (1977a). Cancer Res. 37, 3406. Natori, T., Law, L. W., and Appella, E. (1977b). J . Nut/. Cancer Inst. 59, 1331. Old, L. J . , and Boyse, E. A. (1973). Harvey Lect. 67, 273. Old. L. J . , and Stockert, E. (1977). Ann. Rev. Genet. 17, 127. Old, L. J., Boyse, E. A., Clarke, D. A., and Carswell, E. (1962). Ann. N . Y . Acad. Sci. 101,ao. Oroszlan, S., and Gilden, R. V. (1977). In "Advances in Leukemia Research" (P. Bentvelzen. J . Hilgers, and D. s. Yohn, eds.), p. 90. Elsevier, Amsterdam. Oroszlan, S., Summer, M. R., Foreman, C., and Gilden, R. V. (1974).J. Virol. 14, 1559. Padarathsingh, M . C., Dean, J . H., Lewis, D. D., Northing, J. W., Natori, T., and Law, L. W. (1978). J . Imrnunol. 120, 1981. Paranjpe, M. S., and Boone, C. W. (1975). Cancer Res. 35, 1205. Parmiani, G. (1979). Imrnunogenetics 7, 271. Parmiani. G., and Invernizzi, G. (1975). Int. J . Cancer 16, 756. Parrniani, G., Meschini, A., Invernizzi, G., and Carbone, G. (1978). J . Narl. Cancer Inst. 61, 1229. Pasternak, L.. Pasternak, G., and Karsten, U . (1978). Cancer Imrnunol. Irnrnunother. 3, 273. Pellis, N . R., and Kahan, B. D. (1975).J. Imrnunol. 115, 1717. Pellis, N. R., Tom, B. H . , and Kahan, B. D. (1974). J . Imrnunol. 113, 708. PraL M.. Tarone, G.. and Comoglio, P. M. (1975). Itnmunochemis/ry 12, 9.

TUMOR ANTIGENS ON NEOPLASMS

235

Prat, M . . Rogers, M. J.. and Appella. E. (1978). J . N a i l . Cancer I n s t . 61, 527. Prehn. R. T. (1976). A d r . Ctrricer Res. 23, 203. Price, M. R.. and Baldwin, R. W. (1974a). Br. J. Cuncer 30, 394. Price. M . R., and Baldwin, R. W. (1974b). B r . J. Crrncer 30, 382. Price. M. R.. and Robins. R. A. (1978).I n "Immunologic Aspects of Cancer" (J. E. Castro, ed.), p. 155. Univ. Park Press, Baltimore, Maryland. Price, M . R., Dennick, R. G.. Robins, R. A., and Baldwin, R. W. (197%). Br. J . Crincer 39, 62 I . Price. M. R.. Dennick. R. G.. and Law, L. W. (1979b). Br. J . Cancer 40, 663. Rao, V. S.. and Bonavida, b. (1977). Ctrtrcer Res. 37, 3385. Rogers, M. J. (1979). 7rrinspl~iritrrrion(in press). Rogers, M. J . , and Law, L. W. (1979). I n t . J. Concer 23, 89. Rogers, M. J.. Law, L . W . . Appella. E.. Oroszlan. S., and Ting, C. C. (1977a). I n r . J. Crrricer 20, 303. Rogers, M. J.. Law. L. W.. Prat. M.. Oroszlan. S.. and Appella, E. (l977b). I n / . J . Cancer 21, 246. Rogers, M. J.. Law, L. W., and Appella, E. ( 1 9 7 7 ~ )J.. Nurl. Cancer Inst. 4, 1291. Rogers. M. J . , Law. L. W.. and Appella, E. (1977d). Colloq. I n s t . Sunte Recherche M e d . 69, 349. Rogers. M . J . , Law, L. W . . and Appella, E. (1978). I n "Biological Markers of Neoplasia: Basic and Applied Aspects" (R. W. Ruddon, ed.), p. 53. Elservier, Amsterdam. Rogers, M. J.. Appella, E.. Pierotti, M. A., Invernizzi, G.. and Parmiani, G. (1979a). Proc. N u t / . Acud. Sci. U . S . A . 76, 1415. Rogers, M. J., Robinson, E. A., and Appella. E. ( 1979b). J. B i d . Chem. 254, 1 1126. Rogers, M . J.. Pierotti. M . A.. Parmiani. G.. and Appella, E. (1980). Transplant. Proc. (in press). Siegert, W.. Fenyo, E. M.. and Klein, G. (1977). Inr. J. Cancer 20, 75. Silver, J . , Shaffhamen, B., and Benjamin, T. (1978). Cell 15, 485. Smith, R. W., Morganroth, J., and Mora, P. T. (1970). Nature (London) 227, 141. Snary. D., Goodfellow. B., Bodmer, W. F., Hayman, M. J . , and Crumpton, M. J . (1974). Nurare (London) 247, 457. Soule, H. R . , and Butel, J. S . (1979). J. Virol. 30, 523. Steck. T. L., and Fox, C. F. (1972). I n "Membrane Molecular Biology" (C. F. Fox and A. Keith, eds.), p. 27. Sinauer, Stamford, Connecticut. Stephenson. J. R., Essex. M., Hino, S.. Hardy, W. D., and Aaronson, S. A. (1977). Proc. N u l l . Acud. Sci. U . S . A . 74, 1219. Takasugi, M., and Klein, E. (1970). 7runsplunratiori 9, 219. Tanigaki, N . . Miyakawa, Y ., Yagi. Y ., Kruter, V. P., and Pressman, D. (1973).J. Imniuno/. M d h . 3, 109. Tenen. D. G., Garewal. H.. Haines, L. L., Hudson. J . , Woodward, V., Light, S., and Livingston, D. M. (1977). Proc. Nurl. Accid. Sci. U . S . A . 74, 3745. Tevethia. S . S., and Rapp, F. (1966). Proc. Sor. E x p . Biol. Med. 123, 612. Tevethia. S . S., and Tevethia, M. J. (1975). Irr "Cancer-A Comprehensive Treatise" (F. F. Beaker, ed.), Vol. I V , p. 185. Plenum, New York. Tevethia. S . S.. Grouch, N . A., Melnick. J . L., and Rapp, F. (1970). I n t . J. Corrcer 5 , 176. Ting. C. C.. Shiu, G . , Rodrigues, D., and Herberman. R. B. (1974). Cuncer Res. 34, 1684. Trinchieri. G., Aden, D. P.. and Knowles. B. B. (1976). Nrrrure (London) 261, 312. Tung, J . S . , Yoshiki, T., and Fleissner, E. (1976). Cell 9, 573. Walia. A. S., and Lamon, E. W. (1977). J. Ncirl. Cancer I n s t . 58, 1671. Warnatz, H., and Krapf, F. (1976). J. Imnruno/. 117, 981.

This Page Intentionally Left Blank

NUTRITION AND ITS RELATIONSHIP TO CANCER

Bandaru S. Reddy. Leonard A. Cohen, G. David McCoy, Peter Hill, John H. Weisburger, and Ernst L. Wynder Naylor Dana Institute for Disease Prevention, American Health Foundation, Valhalla, New York

I . Introduction . . . . . . . . . . . A. B.

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

Epidemiology . . . . . . . . . . Etiology .................................

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

..........

........................ D. Experimental Studies . . .

B.

Etiology.. . . . ...

E. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

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

E. Conclusions

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

VIII. Conclusions

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

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

238 24 1 24 1 245 249 251 210 27 1 27 1 211 211 279 282 282 282 283 287 290 29 1 291 293 293 294 294 295 295 302 3 12 3 18 323 3 24 3 24 326 3 26 329 329 3 29 332

237 ADVANCES IN CANCER RESEARCH, VOL. 32

Copyright @ 1980 by Academic Press, Inc. All rights of reproduction in any form reserved ISBN 0-12-006632-7

238

BANDARU

s. REDDY et (11.

I. Introduction

During the past two decades, epidemiologic studies have elaborated the influence of the environment on the development of certain forms of cancer and have led to the conclusion that cancer may not be an inevitable consequence of aging (Wynder and Gori, 1977). This is an encouraging observation that suggests that as we identify the ways in which environmental factors increase the risk of cancer we may be in a position to manipulate the environment and reduce to a minimum the risk of cancer in future generations. Doll (1967) and Higginson (1975) suggest that the lowest reported cancer rates for a given site should be considered the baseline or natural rate of occurrence. Increases in rates, therefore, can be ascribed to environmental factors, or factors that originate almost totally outside the host’s body. Based on comparisons of high and low rate areas, Higginson (1975) noted that approximately 90% of human cancers are associated with environmental factors. Epidemiologically, the effect of environmental factors on cancer incidence has been investigated by studies of incidence patterns between and among population groups, differences in the rates of the disease between the sexes, changes in disease rates over time, demographic and socioeconomic distributions of diseases, effects of migration, and the dietary habits of different populations. Studies of mortality rates also indicate a relationship between disease development and environmental factors. The proportion of deaths for both sexes that have been related to environmental factors are summarized in Tables I and 11. Although the accuracy of mortality and incidence data may be suspect in some areas of the world, reliable data obtained from select and scientifically rigorous cancer registries indicate large differences between high-risk and low-risk populations (Doll et al., 1970). The consistency of the findings suggest that environmental rather than racial or genetic factors play a predominant role in the etiology of cancer in man. Although the concept that diet and nutrition might influence cancer is not new (Tannenbaum, 1940), this relationship has received surprisingly little detailed attention. During the 1930s, a number of laboratories were interested in the possible influence exerted by nutritional factors on susceptibility to cancer, but the question soon lost the interest of both the scientific and lay community. Even though the pioneering studies indicated that dietary restriction reduced the incidence of mammary as well as lung tumors in mice (Tannenbaum, 1942) and that underfeeding of rats led to a lower spontaneous tumor incidence than ad libitum feeding

239

NUTRITION AND CANCER

TABLE I AGE-ADJUSTED MALEMORTALITY RATIOSFOR HIGH-A N D LOW-RISKPOPULATIONS AND FOR U.S. WHITESBY CANCER SITE" Mortality ratios

Cancer risk

Low-risk country

Trachea, bronchus lung S t omach Prostate

Portugal U.S. (for whites) Japan

Intestine except rectum Esophagus Larynx Rectum Buccal cavity, pharynx Pancreas

Japan Norway Sweden Chile Israel Italy

Bladder Skin Liver, biliary passage Thyroid Leukemia

Japan Japan Norway New Zealand Japan

High-risk country Scotland Japan U . S . (for nonwhites) Scotland France France Denmark France U.S. (for nonwhites) S. Africa Australia Japan Switzerland Denmark

High-risk U.S. and low-risk Whiteb country' 3.6 6.9

7. I 7.8 12.4

3.9 1.3 3.8 2.4 3.1 I .9

4.3 5.9 19.8 4.8 7.2 2.3

2.1 3.4 1.6 I .4 2.0

3.3 5.5 5.3

-

5.5 2. I

From Wynder and Gori (1977). mortality rate: low-risk country mortality rate. ' High-risk country mortality rate:low risk country mortality rate.

* U.S.

(Ross and Bras, 1965), it is apparent that additional study of the relationship between dietary factors and human carcinogenesis is in order. Following up on leads advanced by epidemiologists, experimentalists have found that nutrition, in general, is related to the development of cancer in three ways: (1) Food additives or contaminants may act as carcinogens, cocarcinogens, or both. (2) Nutrient deficiencies may lead to biochemical alterations that promote neoplastic processes. (3) Changes in the intake of selected macronutrients may produce metabolic and biochemical abnormalities, either directly or indirectly, which increase the risk for cancer. Although the relationship of nutrition and cancer is complex and sometimes perplexing to those who visualize carcinogenesis in terms of a specific carcinogen, it is important to understand that specific carcinogens play only a minor role as initiators in the relationship between nutrition and the development of cancer.

240

BANDARU S. REDDY

et a / .

TABLE I1 AGE-ADJUSTED FEMALE MORTALITY RATIOSFOR HIGH-A N D LOW-RISKPOPULATIONS A N D FOR U.S. WHITES BY CANCER SITE, 1966-1970" Mortality ratios

Cancer risk

Low-risk country

Stomach Breast Uterus (all parts)

U.S. (for whites) Japan Israel

Intestine (except rectum) Trachea, bronchus, lung Rectum Pancreas

Japan

Esophagus Bladder

Austria Japan

Skin Buccal cavity, pharynx

Japan Fed. Republ. of Ger. Norway New Zealand, Fed. Repub. of Germany Japan

Larynx Liver, biliary passage

Ovary, fallopian tube, broad ligament Thyroid Leukemia

Portugal Chile Italy

Australia Japan

High-risk country

High-risk U.S. and low-risk Whiteb countryC

Japan Netherlands U.S. (for nonwhites) Scotland

5.5 I .9

7.8 6.6 4.7

3.7

4.3

Scotland Denmark U.S. (for nonwhites) Chile U.S. (for nonwhites) Australia N. Ireland

2.4 I .6 1.9

4.2 3.3 2.4

1.0 1.5

6.9 2.4

2.5 2.4

4.2 3.8

Ireland Federal Republic of Germany Denmark

4.4 1.6

16.8 4.1

3.9

5.9

Austria Israel

1.1 1.6

4.0 I .9

From Wynder and Gori (1977).

* U.S. mortality rate:low-risk country

mortality rate. High-risk country mortality rate :low-risk country mortality rate.

Several examples suggest that diet, rather than industrialization and contaminants, is the significant factor in the etiology of certain types of cancer. For instance, colon cancer has shown only a slight upward trend in incidence in the United States in the last 40 years (Cutler and Young, 1975). Cancer of the breast and cancer of the prostate have exhibited similar slow increases in rate in the United States. The logical conclusion is that the pronounced alterations in our environment, such as industrial pollution, intentional and inadvertant food additives, and food contaminants, are not directly or indirectly associated with the development of

NUTRITION A N D CANCER

24 1

these three types of cancer in man. In Japan, a country with a high degree of industrialization, these three types of cancer have a low incidence. Recently, however, an increasing trend has appeared, associated with the progressive Westernization of the Japanese dietary intake since 1945. This also provides some evidence that the dietary pattern, rather than industrial activity is one of the important factors in relation to causative mechanisms for these types of human cancer. However, it must be recognized that the correlation between diet and certain forms of cancer does not prove causation. Many factors may be necessary for cancer causation, but the modification of only one of the contributing factors, such as diet, may be sufficient to retard the chain of causative events. This review covers six types of cancer. In four of these-breast, large bowel, stomach, and head and neck-the epidemiologic evidence is overwhelming that nutritional factors have a major etiological role. Indeed, the epidemiologic data on diet and nutrition in these four cancers provided the leads for metabolic and animal model studies that now fully support their epidemiologic inspiration. Dietary factors are also implicated in the etiologies of the two remaining types of cancer surveyed in this article-pancreas and prostate-but the epidemiologic evidence is presently not overwhelming. As a result, more attention is devoted in this article to the four types of cancer that we know most about with respect to variables of diet and nutrition. However, future research will undoubtedly produce abundant evidence for the etiological role of nutrition and diet in ull of the types of cancer included in this article. This article also presents an evaluation of the current status of the relationship between nutrition and cancer in man, the use of animal models to determine if the etiological factors established for man can be modified in an experimental setting, and we will question the inconsistencies and make recommendations for additional research and possible preventive measures. It. Dietary Factors and Cancer of the Large Bowel

A. EPIDEMIOLOGY Cancer of the large bowel has been the subject of several epidemiologic reviews (Wynder and Shigematsu, 1967; Wynder et al., 1969; Bjelke, 1974; Correa and Haenszel, 1978; Burkitt, 1971; Wynder, 1975b; Weisburger et ul., 1975). In addition, the major features of inter- and intra-

242

BANDARU

s. REDDY et al.

country distribution of cancer of the colon and rectum have been detailed in recent publications (Waterhouse et al., 1976; Fraumeni, 1975; Weisburger et a / . , 1975; Correa and Haenszel, 1978). The highest incidence rates are found in North America, New Zealand, and Western Europe, with the exception of Finland (Fig. 1). Intermediate rates are found in Eastern Europe and the Balkans, whereas the lowest incidences are found in Africa, Asia, and Latin America, with the exception of Uruguay and Argentina, where mortality rates are similar to those found in North America. The mortality data for most of the countries appear to be consistent with the incidence data. In general, the more economically developed a society and its agricultural industries, the greater the incidence of colon cancer, although not necessarily cancer of the rectum (Wynder, 1975b; Waterhouse et a/., 1976). In Israel, colorectal cancer incidence in Jews from Yemen and North Africa is half that of Jews from Europe or North America (Doll et al., 1966). The large difference in colon cancer incidence between Israelis

FEMALE RATE

n n i m . m pa

SCOIUND NEW ZUUM)

ENGLAND

L

WMES

u I , umwniif fRAW AUSIRlA SWIlZfRW

OllyAWI F R

mcinuumos w i n unlu swiDCm IlALV

WamAV )MlUWL W

L

FINLAND

cnni JUAN

h

FIG. 1 . Age-adjusted death rates for malignant neoplasms of intestine, except rectum, in different countries, 1966- 1967. (From Segi and Kurihara, 1972.)

NUTRITION A N D CANCER

243

born in Asia or North Africa and those born in Europe or North America probably reflects the influence of traditional dietary habits more than current lifestyles. Although some geographic and ethnic differences may be due to the well-known artifacts of inaccurate diagnosis and incomplete reporting, these artifacts can account for only a small proportion of the international variations. In Sweden, the methods of diagnosing colon cancer deaths are similar to those used in Denmark and Finland, and thus the differences in colon cancer mortality rates among these regions appear to be valid (Jensen ef al., 1974). In Sweden, there is a north-south gradient, with higher incidence in the south. Gastric cancer exhibits an opposite trend (Jensen et a l . , 1974). However, the difference in colon cancer mortality and incidence rates between the United States or Western Europe and Japan are also real, because the quality of Japanese medical facilities and vital statistics parallel those in the United States and Western Europe (Wynder et al., 1969). Urban populations generally have higher risks for colon cancer than rural populations (Levin et al., 1960; Teppo et al., 1975; Doll ef al., 1970; Staszewski, 1976; Clemmesen, 1974). Among American blacks, the rates are somewhat lower in certain rural parts of the South than the rates for blacks living in industrialized Northern cities (Haenszel and Dawson, 1965) who migrated from rural South to Northern cities. In recent years, the difference had tended to fade, corresponding to the equalization of standard of living, including diet. This is a strong argument for diet as a key element in etiology. 1. Colon-Rectum Differences

Generally, lower colon-rectum incidence ratios prevail in Asian and African populations, whereas the results for Japan, India, and Nigeria represent the prevailing pattern in that part of the world (Correa and Haenszel, 1978). The Eastern European registries report distinctly lower colon-rectum ratios, often below unity, and higher colon-rectum ratios are reported in Western Europe and North America than in low-risk countries. Rectal cancer is relatively more common than colon cancer in males (Wynder, 1975b). It is suggested that cancers of the rectum and colon may have different origins because rectal cancer does not vary as much in incidence in high- and low-risk populations, and it also exhibits a more pronounced sex-linked effect. The male/female ratio is about 1.4 for rectal cancer but is near unity for colon cancer. It can be postulated that the typical absence of fecal material in the

244

BANDARU S. REDDY

et a!.

lower rectum or the difference in histologic structure of mucosa between rectum and colon contributes to this variation in incidence between colon and rectum. 2 . Migrant Studies Further evidence for the importance of environmental factors in determining the geographic differences in large-bowel cancer incidence is provided by studies of migrants to the United States and Australia from equally industrialized Japan (Haenszel and Kurihara, 1968; Haenszel et al., 1973; Staszewski et al., 1971). Cancer incidence is higher in the first and second generation Japanese immigrants to Hawaii and California than in the native Japanese in Japan (Haenszel and Kurihara, 1968). A similar upward trend in colon cancer mortality has also been observed for Polish immigrants to Australia (Staszewski et al., 1971). Further support for environmental dietary influence and/or lifestyle in the development of large-bowel cancer is derived from a study of southern Asian migrants to Kenya (Chopra et al., 1975). Of equal interest is the fact that colon cancer seems to be increasing in Japan itself, a finding consistent with the increasing Westernization of the Japanese diet (Oiso, 1975; Hirayama, 1978). 3 . Socioeconomic Status

With the exception of Japan and Finland, large-bowel cancer is a disease of economically developed countries. Yet, no socioeconomic gradient in risk has been found by intrapopulation comparisons in highrisk countries such as North America and Western Europe. However, the situation in low-risk populations is quite different. The recent findings from Cali, Colombia, where the overall incidence of large-bowel cancer is one-fifth that reported by U.S. registries, indicate a 4-fold excess risk in the upper socioeconomic classes of both sexes in two different age groups (Haenszel et al., 1975). Collateral studies in Cali reveal that the rate of adenomatous polyps of the colon is minimal in the poorest socioeconomic class (Haenszel et al., 1975; Correa, 1975). Nutritional studies have shown a very large socioeconomic difference in meat consumption in Cali (Aragon, 1964). The absence of a socioeconomic gradient for the development of large-bowel cancer in the United States may be due to minimal differences in dietary intake of fat, protein, and carbohydrate. 4. Religious Differences

Comparative studies of religious groups have been motivated by the search for factors that link the lifestyle of individual groups within a small

NUTRITION A N D CANCER

245

geographical area to their site-specific cancer risks. Studies by Lemon and Walden (1966) indicated the total cancer mortality of California Seventh-Day Adventists to be about 60% of a comparable general population sample. The Seventh-Day Adventists who consume less meat and adhere to a lactoovo-vegetarian diet are reported to have 60-70% of the colon cancer death rate for total California (Lemon et al., 1964; Phillips, 1975). It can be postulated that a lactoovo-vegetarian diet protects against largebowel cancer (Phillips, 1975). The incidence of large-bowel cancer in Mormons (members of The Church of Jesus Christ of Latter-Day Saints) also is lower than other U.S. white populations (Enstrom, 1975b; Lyon et al., 1977). Traditionally, Mormons eat more whole grain breads, cereals, fruits, and vegetables. Jussawalla and Jain (1976) compared colon and breast cancer incidence rates in the Parsi community in Bombay, India, with that of the Hindu population of Bombay and other Indian communities and found that the Parsis exhibited a higher risk for colon and breast cancer, but not for rectal cancer, than the other religious groups of Bombay. Since the Parsis possess distinctive dietary habits, which lean more toward a Western diet, the difference in dietary habits between these two Indian populations could account for the differences in colon cancer rates.

B. ETIOLOGY 1. Correlation Analysis Since the risk of large-bowel cancer closely parallels the economic development of a given country, cross-national correlations between colon cancer frequency and diet have been used to select from possible hypotheses for testing in case-control and cohort studies. These studies have shown food preferences that appear to be associated with high- and low-risk populations. Such correlations may be spurious, but when a correlation such as that between fat intake and colon cancer mortality is supported by experimental evidence from animal models, and underlying mechanisms can be described, it seems worthy to consider the possible relevance (Reddy et al., 1975a). Wynder et al. (1969) and Wynder and Reddy (1973) proposed that colon cancer incidence is mainly associated with total dietary fat. They further suggested that dietary fat influences the metabolic activity of the fecal microflora and thus may be involved in the pathogenesis of cancer of the colon. A worldwide correlation between colon cancer incidence and total fat consumption has been established (Fig. 2).

246

BANDARU

a W

a n

... . . .. . . . . . . . . . . . . .. . 0..

W

**

l03 3

0 .

a

7w 3

s. REDDY et al.

0

5-

' . 2

.

0

.

0.

0

.

. r

0

1

1

FIG. 2. Correlation between age-adjusted mortality from colon cancer and per capita consumption of fat. (From Carroll and Khor, 1975.)

Gregor et al. (1969) analyzed cancer mortality and food consumption data in 28 countries and found a high correlation ( r = 0.81) between intestinal cancer and animal protein consumption. Dietary fat and fiber were not analyzed. They concluded that their data would be more compatible with a promoter rather than an initiator role for diet during development of the disease. Drasar and Irving (1973) analyzed United Nations Food and Agriculture Organization (FAO) diet data and colon cancer incidence data from 37 countries and showed that the incidence of colon cancer was highly correlated with dietary animal protein and bound fat. These two dietary items are themselves closely associated since much of the bound fat is of animal origin. Enig et al. (1978) examined the fat-cancer relationship in the United States and found an equally strong, significant positive correlation of colon cancer with total fat and vegetable fat; but no correlation was found between animal fat

NUTRITION A N D CANCER

247

and colon cancer mortality and incidence. These results support a role for total dietary fat in the incidence of colon cancer. Several investigators have systematically examined correlations between per capita consumption of food items based on F A 0 data and the incidence and/or mortality of colon cancer (Armstrong and Doll, 1975; Howell, 1975). Armstrong and Doll (1975) have shown that the dietary variables chiefly associated with large-bowel cancer rates were meat and other animal protein. Total fat, meat, and other animal protein are highly correlated. Howell (1975) pointed out that beef consumption was more closely related to colon cancer rates than was pork, poultry, or fish. Burkitt (1971, 1975) recognized the rarity of large-bowel cancer in most African populations and suggested that countries consuming a natural diet rich in fiber have a low incidence of large-bowel cancer, whereas those eating refined carbohydrates with little fiber have a higher incidence of the disease. It has been argued that large-bowel tumors are related to factors characteristic of modern Western society in which intestinal transit time is decreased, small firm stools produced, and the fecal bacteria flora altered. Slower transit was postulated to allow more time for gut bacteria to degrade intraluminal components and to produce carcinogens, allowing time for such carcinogens to act (Burkitt, 1971, 1975). There is, however, no support for the suggestion that longer transit time results in an increase in the degradation of substrates by gut bacteria (Walters et ul., 1975). A recent study comparing populations in Kuopio, Finland (low-risk) and in Copenhagen, Denmark (high-risk) (IARC Microecology Group, 1977), indicated that transit time and stool weight had few significant correlations with diet and defecation habits, but stool weights were higher in the population from Kuopio. Recent data also suggest that one of the factors contributing to the low-risk of large-bowel cancer in Kuopio appears to be that high dietary fiber intake leads to increased stool bulk, in effect diluting tumorigenic compounds in the colon (Reddy cr ul., 1978a). From these data, we can postulate a possible protective role of dietary fiber in the pathogenesis of large-bowel cancer in man. The question also arises of whether inhibitors present in the environment affect the response of humans to tumorigenic compounds. Several epidemiologic studies indicate an increased incidence of large-bowel cancer in man in geographic regions where selenium is deficient (Jansson er al., 1975; Shamberger and Willis, 1971). Shamberger and Willis (1971) also associated the amount of selenium in soil and forage crops and cancer mortality rates in U.S. and Canada and noted an inverse relationship between blood selenium levels and cancer mortality rates. Recently,

248

BANDARU

s. REDDY et al.

Schrauzer et al. (1977) found that selenium intake at naturally occurring levels varied inversely with breast and colon cancer incidence in many countries. From the apparent dietary selenium intake, estimates of mortality from breast and colon cancer in various countries were calculated. A close fit was observed between the calculated and observed values for these countries. 2 . Case-Control Studies Studies attempting to explain the frequency of large-bowel cancer have used both correlation and case-control studies. Wynder et a f . (1969) conducted a large-scale retrospective study on large-bowel cancer patients in Japan, which suggested a correlation between the Westernization of the Japanese diet and colon cancer. In an earlier large-scale retrospective study on large-bowel cancer patients in the United States, it was concluded that except for the established high risk for patients with ulcerative colitis and familial polyposis, no environmental dietary factors could be identified that differed significantly between the control and study populations (Wynder and Shigematsu, 1967). Based on these two studies, it is reasonable to assume that differences in dietary intake, especially those involving Westernization of the diet, could be determined by retrospective techniques such as a study in Japan (Wynder and Shigematsu, 1967). Such data gathering techniques do not, however, appear applicable to the United States where the total intake of various nutrients is now quite similar, except in special religious groups (Wynder, 1975b). Haenszel et al. (1973) demonstrated an association between colon cancer and dietary beef in 179 Hawaiian Japanese cases and 357 Hawaiian Japanese controls. Meat provided a striking example of a change in food practices between Japan and Hawaii; the rise in beef consumption paralleled the increase in colon cancer risk among Japanese migrants. However, Enstrom (1975a), in matching time trends and socioeconomic, urban-rural, and regional gradients in beef and fat consumption with the corresponding information on bowel cancer, argued that the data with respect to beef and colon cancer were incompatible. Enstrom (1975a,b) also pointed to a somewhat lower risk of colon cancer among Mormons who have no religious proscription against the use of beef. Some of these inconsistencies may be explained by the fact that (1) beef will differ in fat content depending on cattle age, on how long the animals were range-fed, and whether they were grain-fed in a feed lot, and (2) fiber intake may differ between population groups. In case-control studies in Israel, Modan et al. (1975) found that among a large variety of dietary constituents investigated, those that were lowest

NUTRITION AND CANCER

249

in the diets of patients with colon cancer as compared with controls were those containing fiber. Recently, Dales et al. (1979) conducted a casecontrol study of relationships of diet to colon cancer in American blacks and found that significantly more colon cancer patients than controls reported a high saturated fat-low fibrous foods eating pattern, as opposed to a low saturated fat-high fibrous foods diet. Bjelke (1974), who conducted diet interview study of hospitalized cases and controls in Minnesota and in Norway, found less frequent use of vegetables among colorectal cancer patients; and, in Minnesota, particularly less frequent use of cabbage. Another study on a series of cases and controls from Roswell Park Memorial Institute at Buffalo, New York, showed a lower risk of colon cancer for individuals ingesting vegetables such as cabbage, broccoli, and Brussels sprouts (Graham and Mettlin, 1979). The studies cited lead us to accept diet as a major etiologic factor in large bowel cancer. Diets high in total fat, low in fiber, and high in beef, are associated with an increased incidence of large-bowel cancer in man.

C. METABOLICEPIDEMIOLOGY Current evidence indicates that colon cancer may stem from the combined action of currently unidentified carcinogens, of cocarcinogens, and of promoting agents (Wynder and Reddy, 1973; Aries et al., 1969; Reddy et al., 1978b). To explain the relationship between dietary fat 'and colorectal cancer, it has been hypothesized that ( I ) the amount of dietary fat determines both the concentration of acid and neutral sterol substrates in the large bowel and also the composition of the microflora acting on such substrates; and (2) the gut microflora metabolize acid and neutral sterols to carcinogens active in the large bowel (Aries et al., 1969). Attention has been focused on the possible role of bacteria in altering the structure of colonic steroids. Investigators (Cook et al., 1940; Coombs et al., 1973; Haddow, 1970; Lacassagne et al., 1966; Hill, 1974) have examined the potential carcinogenic activity of certain bile acids because: (1) their overall structure is similar to carcinogenic polycyclic aromatic hydrocarbons (PAH); (2) they may be converted chemically to 3-methylcholanthrene; (3) full aromatization of the bile acid nucleus would yield a carcinogen metabolite based on cyclopentaphenathrene; (4) human gut flora have been shown to achieve partial aromatization of the sterol ring system; and ( 5 ) several bile acids induced sarcomas at the site of injection in experimental animals. It may be noted that such microflora-mediated reactions are unlikely to yield polycyclic aromatic hydrocarbons from

250

BANDARU

s. REDDY et al.

bile salts, but are much more likely to yield products that act as colon tumor-promoters or cocarcinogens rather than as complete carcinogens (Reddy et al., 1978b). Thus, a high-fat diet may not only change the composition of bile acids but also modify the activity of gut microflora, which may in turn produce tumor-promoting substances from bile acids in the lumen of the colon (Hill et al., 1971; Reddy and Wynder, 1973; Reddy et al., 1978b). High intake of dietary fiber of a certain type leads to an increase in stool bulk, thereby diluting carcinogens and promoters (Burkitt, 1975; Reddy et al., 1978a). Thus, we are concerned with two aspects: the search for carcinogens, and the search for modifying and, in particular, enhancing factors. Until very recently, the question of carcinogens affecting the large bowel could not be approached experimentally. Now, utilizing metabolic techniques and mutagenicity tests, it is experimentally feasible to test whether carcinogens affecting the colon can be isolated, identified, and quantitated. Investigations in man have also been carried out in several laboratories to determine (1) whether changes in the diet would alter the concentration of fecal bile acids and cholesterol metabolites, and the activity of fecal microflora, and (2) whether fecal constituents differ between high-risk and low-risk populations for colon cancer and between patients with colon cancer, familial polyposis, adenomatous polyps, and ulcerative colitis, and patients with no known large-bowel disorders. 1. Fecal Constituents of Populations with Diverse Dietary Habits

Aries et al. (1969) assayed fecal samples for anaerobic and aerobic microflora from British and Ugandan subjects, who constituted high- and low-risk groups for large-bowel cancer, respectively. They found that the British had a higher concentration of anaerobes and lower levels of aerobes than the Ugandans. Hill et al. (1971) confirmed these results and further observed a correlation between the death rate due to colon cancer and fecal anaerobes and fecal excretion of cholesterol and bile acid metabolites as well as their degradation by the gut flora. In addition, the feces from U.S. and English subjects contained higher levels of nuclear dehydrogenating clostridia (NDC) than subjects from African and Asian countries (Hill, 1974). The implication is that NDC is involved in the production of unsaturated steroids from the bile acid nucleus (Hill, 1974; Hill et al., 1975). Neither Moore and Holdeman (1974, 1975) nor Finegold et al. (1975) have seen significant differences i n the composition of fecal flora of high-risk Hawaiian Japanese from whom adenomatous polyps had been removed, high-risk healthy Hawaiian Japanese, high-risk North

NUTRITION A N D CANCER

25 1

Americans, and low-risk rural Japanese. Most of the species of bacteria encountered in North American polyp patients were also seen in the feces of Japanese and Africans, and there was little difference in the overall composite distribution of number and type of species in the two types of populations. Finegold et al. (1975) studied subjects with colonic polyps and Japanese-American groups with distinctive Japanese and Western diets, and found no significant differences in the distribution of fecal anaerobes. A recent comparison of vegetarian Seventh-Day Adventists with nonvegetarian Adventists and non-Adventists failed to show impressive differences among the populations, although there were 21 organisms or groups of organisms present with a significantly different frequency in the high-risk group (non-vegetarian) than in the low-risk group (Finegold and Sutter, 1978). Certainly these organisms should be studied for their capacity to produce carcinogens or cocarcinogens from appropriate substrates. [Some of the differences in fecal flora and discrepancies described by various investigators may be an artifact of procedures used to transport specimens from the several study areas to the laboratory for analysis (Moore and Holdeman, 1975).] Most diets in man exert little effect on the distribution of microflora until the host’s intestinal physiology changes in response to diet. The effect of dietary manipulations on fecal microflora profile has been studied by several investigators. Moore and Holdeman (1975) have shown that Americans consuming nothing but rice and tea or coffee for 3 days had relatively minor changes in the composition of fecal bacteria. Similar results were obtained when leafy vegetables replaced the rice, and, again, when only lean beef was consumed (Moore and Holdeman, 1975). In a study by Hentges et al. (1977) in which the meat content was altered without modifying the fat content, there was no great change in the fecal bacterial flora. Reddy et al. (1975b) compared the effect of a high-fat, mixed Western diet with a low-fat, nonmeat diet. Although differences in bacterial counts between the two diets were not striking, total, as well as some individual anaerobic counts were greater during the mixed Western diet than during the low-fat, nonmeat diet. The effect of dietary fiber supplementation on fecal flora was studied by Fuchs et al. (1976) who showed that total anaerobic counts were increased significantly while the subjects were on the high fiber diet. Draser et al. (1976) obtained no significant changes in fecal bacterial flora in subjects consuming a variety of fibers. It is evident that the identification and quantitation of bacterial species may be unproductive. We are more interested in metabolic and functional differences than in specific identification of fecal bacteria. Of primary

252

BANDARU

s. REDDY et al.

importance in investigating the etiology of large-bowel cancer is an understanding of the influence of dietary constituents on the enzymic (metabolic) activities of gut bacteria, irrespective of species. Various fecal bacterial enzymes, such as P-glucuronidase, 7a-dehydroxylase, cholesterol dehydrogenase, 7a-hydroxysteroid dehydrogenase, and nuclear dehydrogenase reflect not only the metabolic activity of the colonic bacteria but also the functional capabilities of colonic bacteria to produce putative carcinogens in the gut. Reddy and Wynder (1973) investigated fecal bile acids and neutral sterols, as well as the microbial P-glucuronidase activity, in the feces to assess the degree of microbial activity for the enzymic hydrolysis of various complete conjugates in the large-bowel of various population groups. (Americans on a high-fat mixed Western diet; Seventh-Day Adventists on a mixed Western diet without meat and less fat; Japanese; Chinese Americans consuming Chinese diet; strict vegetarians.) A significant increase in the excretion of total bile acid (deoxycholic acid, lithocholic acid, 12-ketolithocholic acid, 3& 12a-dihydroxy-5~-cholanic acid, 12a-hydroxy-3-keto-5~-cholanic acid, and 3-keto-SP-cholanic acid) and cholesterol metabolites (coprostanol and coprostanone) were found in Americans consuming a high-fat, mixed Western diet compared with other groups (Table 111). The fecal bacteria of groups consuming a mixed Western diet also had a higher P-glucuronidase activity. Since many exogenous and endogenous substances, including tumorigenic metabolites, are excreted via bile as glucuronide conjugates, the data might imply that colonic bacteria of high-risk groups are more active in hydrolyzing these conjugates. These differences are related to dietary composition, mainly a high content of total fat in the high-risk group. Macdonald et al. (1978) studied the fecal NAD- and NADP-dependent 7a-hydroxysteroid dehydrogenase (7a-HSDH), which converts hydroxybile salts to keto-bile salts, in vegetarian Seventh-Day Adventists and non-Seventh-Day Adventists consuming a mixed Western diet. The activity of fecal 7a-HSDH was lower in Seventh-Day Adventists compared to non-Adventists, suggesting that increased NAD- and NADP-dependent 7a-HSDH are associated with risk of large-bowel cancer. Thus, evidently a high-risk diet can alter the metabolic activity of gut microflora, and this effect may play an active role in the etiology of large-bowel cancer. Controlled studies comparing a high-meat, high-fat diet with a nonmeat, low-fat diet showed that the former resulted in elevated levels of fecal bile acid and cholesterol metabolites and increased bacterial P-glucuronidase activity (Reddy et al., 1974, 1975b). Our explanation for this difference is that the fecal bile acids and cholesterol metabolites and the activity of fecal microflora are obviously related to dietary composition,

TABLE I11 DAILY FECALEXCRETION OF NEUTRAL STEROLS A N D BILE ACIDSI N DIFFERENT POPULATIONS WITH VARIED RISK FOR COLON CANCER' North Americans on a mixed Western diet (40)' Neutral sterols Cholesterol Coprostanol Coprostanone Total neutral sterols Bile acids Cholic acid Chenodeoxycholic acid Deoxycholic acid Lithocholic acid Other bile acids Total bile acids

North American vegetarians (12)

North American Seventh-Day Adventists (25)

Japanese (25)

Chinese (25)

705t 104

67 2 17 231 ? 4 9 20 t 318 t 53'

60 f 20 201 f 26b 20 t 3b 281 ? 34b

90 ? 20 140 ? 2Sb 24 t 6b 254 t 35b

6 0 ? 18 129 t 2Sb 25 f 6b 214 f 36b

12 2 4 10 ? 3 1 1 5 2 18 9 0 f 10 48 f 9 275

7 2 6 6 5 2 32 f 6b 23 ? 5b 65 f 10 133 f 15O

8 5 1 62 I 30 f 5*

5 2 2 6 2 2 45 f 5 b 32 ? 3b IOf 2 98 f 6b

102 3 12 2 2 40 f 6b 38 ? 4b 2 2 1 102 2 lob

45 ? 10 520 ? 75 140 t 79

Averages 2 SEM. Daily fecal excretion expressed in mg/day. Significantly different from North Americans on a mixed Western diet (P< 0.5). Number of samples.

2 9 ? 3b 1723 90 f

254

BANDARU

s. REDDY et al.

mainly to a high dietary fat diet. Hentges et af. (1977) compared a diet series consisting of a control diet, a high-beef diet, and a meatless diet. There was no increase in the concentrations of bile acids and neutral sterols in the feces of subjects during the high-meat diet. Fat and fiber were essentially the same in these diets. These data indicate that high animal protein consumption produces no major effect on steroid composition of feces. Hill (1977), after review of several studies, also concluded that it is the fat, not the protein composition, in the meat that determines the effects on the fecal steroids. Recently, Reddy et al. (1978a) have undertaken a study of healthy controls in Kuopio (Finland), a low-risk population for the development of colon cancer. The dietary histories indicate that the total fat consumption is quite similar to the United States populations, but the major part of fat comes from milk and other dairy products, whereas in the U.S. the major source of fat is meat. The voluntary intake of meat in Kuopio is low, and complex cereal consumption is very high in comparison to the United States (Table IV). The daily output of feces is 3-fold higher in healthy controls from Kuopio than in the New York Metropolitan area. The concentration of fecal secondary bile acids, mainly deoxycholic acid and lithocholic acid, as well as p-glucuronidase activity is decreased in Kuopio, but the daily output remained the same in the two groups because of a 3-fold increase in the daily output of feces in Kuopio. The concentration of secondary bile acids was lower, indicating that fecal TABLE IV DIETARY INTAKE OF VARIOUS NUTRIENTS A N D FECALEXCRETION OF VARIOUS CONSTITUENTS I N MIDDLE-AGED MALEVOLUNTEERS FROM KUOPIO (FINLAND) AND NEWYORKMETROPOLITAN AREA^ Kuopio

(Inb

Dietary constituents Total protein Total fat Saturated fat Other fats Carbohydrates Total fiber Fecal constituents Fresh feces excreted Fiber Fecal dry matter Averages k SEM. Units: gm/day. Number of samples.

*4 *4

1152 3

3 3 4 3

66 rt 2 285 rt 4 14+ 2

93 110 59* 51 rt 320 5 32 2

211 26 61

New York (40)

5

20

5

8

*2

*2 49 * 2 89

7 6 2 12 9rt 1 22 rt 1

255

NUTRITION AND CANCER

TABLE V FECALBILEACIDSOF HEALTHY MALESUBJECTS FROM KUOPIO (FINLAND) A N D NEW YORKMETROPOLITAN AREA Bile acids :fecal material (mdgm) Bile acids Cholic acid Chenodeoxycholic acid Deoxycholic acid Lithocholic acid Ursodeoxycholic acid 3cr,p,12a-Trihydroxy-5pcholanic acid 12-Ketolithocholic acid Other bile acids Total bile acids

Daily excretion (mdday)

Kuopio (15)'

New York (20)

Kuopio (1.9

0.20 2 0.06" 0.13 f 0.03 1.72 2 0.16' 1.40 f 0.16b 0.08 f 0.02' 0.04 1 0.01'

0.24 2 0.04 0.23 f 0.03 3.74 2 0.26 3.27 f 0.15 0.13 f 0.01 9.12 f 0.01

12 2 2.9 8 f 1.3 104 f 12 84 f 5 5 f 1.1 2 f 0.8

77 f 4.5 3 2 0.3 3 2 0.3

0.06 f 0.02' 0.93 f 0.08b

0.13 f 0.01 3.8 2 0.26 11.7 f 0.54

4 f 1.0 56 f 5.p 277 f 22

3 f 0.2 89 f 6.0 275 f 14

4.59 2 0.42'

New York (20) 6

f

1.4

5 f 1.1 88 f 5.1

Averages f SEM. Significantly different from New York: P < 0.05, or better. Number of samples.

bulk diluted the fecal secondary bile acids (Table V) and fecal bacterial P-glucuronidase activity. This suggests that increased fecal bulk would dilute any suspected carcinogens or promoters in direct contact with the large bowel mucosa.

2 . Fecal Constituents of Patients with Colon Cancer and Adenomatous Polyps

In a case-control study, 80% of large-bowel cancer patients had fecal bile acid levels above an arbitrary cut-off level compared with only 17% of the controls; 80% of the cancer patients had fecal NDC compared with 43% of the comparable controls; whereas 70% of the cancer patients had a combination of high NDC and fecal bile acid levels compared with only 9% of the controls (Hill et u / . , 1975). These findings require confirmation in prospective studies, since the presence of cancer may have altered bacterial metabolism in the gut. We have carried out a study of fecal constituents of patients with colon cancer and nonhereditary adenomatous polyps (Reddy and Wynder, 1977). The fecal excretion of cholesterol metabolites and secondary bile acids (deoxycholic acid, lithocholic acid, and other microbially modified bile acids) was higher in patients with colon cancer or adenomatous polyps compared to healthy controls (Table VI). The total bile acid and

256

BANDARU

s. REDDY et al.

TABLE VI FECALNEUTRAL STEROLS A N D BILEACIDS EXCRETION I N PATIENTS A N D CONTROLS CONSUMING A MIXEDWESTERN DIET

Averages

f

HEALTHY

Patients with adenomatous POIYPS

Healthy controls

(1s)

(65)

6.4 f 0.8’ 19.6 f 3.2’ 4.0 f I . @ 30.0 f 3.1’

12.8 f 1.1 2.2 f 0.2 18.0 f 0.9

0.3 f 0.1 0.4 f 0.1 7.2 f 0.5’ 6.9 f 0.4’ 0.2 f 0.1

0.4 f 0.1 0.3 f 0.1 6.1 f 0.7’ 5.4 f 0.5’ 0.3 f 0.1

0.4 f 0.1 0.5 f 0.1 3.3 f 0.3 3.0 f 0.2 0.1 f 0.1

0.8 f 0.1 0.6 f 0.1

0.9 f 0.1 0.5 f 0.1 2.4 f 0.3 16.3 f 0.7b

0.2 f 0.1 0.5 f 0.1 2.3 f 0.2 10.3 f 0.8

Patients with colon cancer (50)‘ Neutral sterols Cholesterol Coprostanol Coprostanone Total Bile acids Cholic acid Chenodeoxycholic acid Deoxycholic acid Lithocholic acid 3a.P. 12a-Trihydroxy-5Pcholanic acid 12-Ketolithocholicacid Ursodeoxycholic acid Other bile acids Total bile acids

IN

10.4 f 1.4a3b 2.3’ 3.9 f 0.3b 37.1 f 2 . 9

22.8

2.2

*

f 0.2

18.6 t 0.9

3.0 f 0.2

SEM.

’ Significantly different from healthy controls: P < 0.05. Number of samples.

neutral sterol excretion was also higher in the above patients compared to controls. Our results also indicate that the activity of fecal bacterial 7a-dehydroxylase, which converts primary bile acids to secondary bile acids, was higher in patients with colon cancer or polyps than controls (Mastromarino et al., 1978). These data support the concept that patients with colon cancer or polyps are more able to convert primary bile acids into secondary bile acids in colonic contents than are the controls. The fecal bacteria of patients with colon cancer or polyps also contained higher cholesterol dehydrogenase activity, which converts cholesterol to coprostanol in colonic contents. Macdonald et al. (1978) have demonstrated that the fecal bacterial NAD- and NADP-dependent 7a-hydroxysteroid dehydrogenase activities are higher in large-bowel cancer patients as compared to controls. An association has been established linking colon cancer to higher dietary fat/meat and low dietary fiber, the metabolic activity of fecal flora, and fecal neutral sterol and bile acid metabolites. The extent to

NUTRITION A N D CANCER

257

which findings on bacterial enzymes may be used as indicators or diagnostic markers for high- and low-risk populations, as well as patients with colon cancer, requires further study-studies that should concentrate on identifying and validating such bacterial and/or biochemical indicators.

STUDIES D. EXPERIMENTAL Research on the mechanisms of cancer causation in the large bowel has been assisted by the discovery over the last 20 years of several animal models that mirror the type of lesions seen in man. These models are: (1) induction of large-bowel cancer in rats through chemicals such as 3-methyl-4-aminobipheny1,or 3-methyl-2-naphthylamine; (2) derivatives and analogs of cycasin and methylazoxymethanol (MAM) such as azoxymethane (AOM) and 1,2-dimethyIhydrazine (DMH), which work well in rats and mice of select strains; (3) intrarectal administration of direct-acting carcinogens of the type of alkynitrosoureas, such as methylnitrosourea (MNU) or N-methyl-N’-nitrosoguanidine(MNNG), which lead to cancer of the descending large bowel in every species tested so far; and (4)the oral administration of large doses of 3-methylcholanthrene, which leads to large-bowel cancer in select strains of hamsters (Bralow and Weisburger, 1976). 1. Effect of Dietary Far in Colon Carcinogenesis

Experimental studies have lent some additional support to the possible role of dietary fat in the induction of large-bowel cancer in man. Nigro er al. (1975) induced intestinal tumors in rats by AOM and compared animals fed a high beef fat diet and those fed a normal diet. Animals fed a high-fat diet developed more intestinal tumors and more metastasis than the rats fed a low-fat diet. Inasmuch as humans in various populations usually follow comparable dietary regimens over generations, Reddy et al., (1976a) designed experiments in which animals were exposed to a given regimen for two generations prior to treatment with a carcinogen. Virgin female rats fed diets containing 5% corn oil, 20% corn oil, 5% lard, or 20% lard were bred, and the litters were weaned t o the same diet consumed by the mothers. At 7 weeks of age, all second generation animals, except controls received 20 weekly subcutaneous (sc) doses of DMH (10 mg/kg body weight). Animals fed 20% lard or 20% corn oil were more susceptible to colon tumor induction by DMH than those in other groups (Table VII).

258

BANDARU

COLON TUMORINCIDENCE

Diet fat

Percentage in diet

Lard Lard Corn oil Corn oil Beef fat Beef fat Corn oil Corn oil Beef fat Beef fat Beef fat Beef fat Beef fat Beef fat

5 20 5 20 24 6 24 6 20 5 20 5 20 5

IN

s. REDDY et al.

TABLE VII RATS FEDDIETSHIGHIN FAT A N D TREATEDWITH CARCINOGENS

Protein Casein

Beef protein Beef protein Soybean protein Soybean protein Casein

Percentage in diet 25 25 25 25 40 IY

40 IY

22 22 22 22 22 22

Carcinogen DMH" DMH" DMH" DMH" DMH" DMH" DMH" DMH" DMHb DMHb MNU' MNU' MAM acetated MAM acetated

Percentage of rats with colon tumors 17 67 36 64 57 35 54 35 60 27 73 33 80 45

Female F344 rats, at 7 weeks of age, were given weekly sc DMH at a dose rate of 10 mgikg body weight for 20 weeks and autopsied 10 weeks later. Male F344 rats, at 7 weeks of age, were given a single dose of sc DMH, 150 mgikg body weight and autopsied 30 weeks later. Male F344 rats, at 7 weeks of age, were given ir MNU, 2.5 mgirat twice in one week and autopised 30 weeks later. Male F344 rats, at 7 weeks of age, were given ip MAM acetate, 35 mgikg body weight once and autopsied 30 weeks later.

The type of fat appears to be immaterial at the 20% level, although at the 5% fat level, there is a suggestion that unsaturated fat (corn oil) predisposes to more DMH-induced colon tumors than saturated fat (lard). Combinations of high beef protein (40%) and high beef fat (20%) or high soybean protein (40%) and high corn oil (20%) led to more DMH-induced colon tumors in F344 rats than control diets of beef protein (20%) and low beef fat (6%), or soybean protein (20%) and low corn oil content (6%) (Table VII; Reddy et al., 1976b). Further, F344 rats fed a diet containing 20% beef fat and treated intraperitoneally with MAM acetate, subcutaneously with DMH, or intrarectally with MNU, had a greater incidence of colon tumors than did rats fed a diet containing 5% beef fat and treated similarly (Reddy et al., 1977b) (Table VII). W/Fu rats fed a 30% lard diet had an increased number of DMH-induced large-bowel tumors compared to the animals fed the standard diet (Bansal et al., 1978). Broitman et al. (1977) showed

259

NUTRITION A N D CANCER

that rats fed a 20% safflower oil diet had more DMH-induced large-bowel tumors than those animals fed either the 5% or 20% coconut oil diets. However, these studies provide no evidence that dietary polyunsaturated fat per se is more effective than saturated fat in augmenting tumorigenesis by DMH. Rogers ez af. (1973) found that a diet, marginally deficient in lipotropes but high in fat, enhanced DMH-induced colon carcinogenesis in Sprague-Dawley rats. These results suggest that the total dietary fat, rather than the type or source of fat, may have a function in the pathogenesis of colon cancer. In order to understand the specifics of the mechanisms whereby dietary fat influences colon cancer, the effect of high dietary fat on biliary and fecal bile acid pattern was investigated. Biliary excretion of total bile acids, as well as cholic acid, P-muricholic acid, ursodeoxycholic acid, and deoxycholic acid, was higher in rats fed a diet containing 20% corn oil or 20% lard than in rats fed diets containing 5% corn oil or 5% lard. High fat (corn oil or lard at 20% level) intake was associated with an increased excretion of fecal neutral sterols and bile acids. The excretion of deoxycholic acid, lithocholic acid, and 12-ketolithocholic acid was increased in rats fed high-fat diets (Reddy et al., 1977a) (Table VIII). Recent studies indicate that the enhanced tumorigenesis in the animals

TABLE VIII EFFECTOF TYPEA N D AMOUNT OF DIETARY FATON FECAL BILEACIDSI N RATS"

5% Corn oil control (8) Cholic acid P-Muricholic acid 3a.P. 12a-Trihydroxy-SPcholanic acid Chenodeoxycholic acid Hyodeoxcholic acid Ursodeoxycholic acid Deoxycholic acid Lithocholic acid 12-Ketolithocholic acid 7-Ketodeoxycholic acid Other bile acids Total bile acids

20% Corn oil control (8)

0.68 ? 0.08**' 0.64 C 0;07' 0.82 ? 0.05' 0.98 t 0.08' 0.10 -t 0.01' 0.11 t 0.02'

*

0.12 0.01' 2.76 f 0.12' 0.10 f 0.2' 2.53 ? 0.18' 0.83 f 0. I I ' 0.44 f 0.03' 0.14 ? 0.02' 1.93 ? 0.10 10.45 ? 0.20'

0.15 ? 0.01' 2.73 t 0.16' 0.10 t 0.02' 4.80 t 0.23' 1.98 t 0.16' 0.77 t 0.07' 0.08 f 0.01' 2.52 t 0.25 14.86 -+ 0.41'

5% Lard control (8)

20% Lard control (8)

* 0.06' f 0.07' * 0.01'

0.86 t 0.10' 0.88 t 0.1 I ' 0.13 ? 0.01'

0.13 5 0.02' 3.14 2 0.18' 0.15 t 0.09' 2.61 ? 0.20' 1.00 -+ 0.10' 0.51 2 0.18' 0.16 f 0.02' 1.92 ? 0.16 11.24 t 0.49'

0.16 0.03' 2.73 0.17' 0.08 f O . O l l 4.54 t 0.30' 2.84 f 0.13' 0.77 f 0.02' 0.06 t 0.01' 2.51 ? 0.19 14.91 t 0.62'

0.74 0.80 0.10

*

*

Mean f SEM. Units: mgidayikg body weight. Mean with a common number superscript between groups in a horizontal row are not significant: P > 0.05.

260

BANDARU

s. REDDY et al.

fed the high-fat diet is due to promotional effects rather than alterations in carcinogen metabolism (Bull et al., 1979). While there is no human model for tumor promotion by high dietary fat, the above results in an animal model suggest that colon tumor promotion results through a mechanism involving increased colonic bile acid content. Since the intestinal bacteria contain many inducible enzymes, experiments were conducted to delineate the effects of various dietary factors on the metabolic activity of intestinal microflora in order to understand the relationship of colon cancer to diet-mediated changes in the intestinal bacteria. Goldin and Gorbach (1977) reported that rats fed a meat diet had higher levels of fecal bacterial p-glucuronidase, azoreductase, and nitroreductase activities than did grain-fed rats. After confirming these studies, Reddy et al. (1977b) extended the observation that not only a meat diet, but also a high-fat diet or high-protein high-fat diet, changes the bacterial p-glucuronidase activity in the large intestine. In an investigation of the effect of Lactobacillus acidophilus feeding on rat fecal enzymes, Goldin and Gorbach (1977) found that supplemental feeding of L. acidophilus significantly lowered the activity of fecal pglucuronidase, nitroreductase, and azoreductase in rats consuming a meat diet. The ability of L. acidophilus to reduce activities of these enzymes in rats fed a meat diet is of great interest, although its significance to carcinogenesis remains unestablished. Although it is premature to conclude from these studies that factors altering microflora enzymes have an effect on tumor formation in the large bowel, these changes in metabolic activity of microflora might alter the biological activity, toxicity, excretion, and reabsorption of many of the endogenous and exogenous compounds such as carcinogens and/or cocarcinogen metabolites. 2 . Effect of Dietaty Protein in Colon Carcinogenesis The possible role of dietary protein in DMH-induced colon carcinogenesis in Sprague-Dawley rats was studied by Visek et al. (1978). Protein source as a modifying factor in colon carcinogenesis was studied, using semi-purified diets containing 20% protein as freeze-dried raw beef, charcoal-broiled beef or soybean protein, and 20% beef fat. It was concluded that the source of protein was not a factor in the DMH-induced colon carcinogenesis in rats. In another study, rats fed 15% and 22.5% protein had a greater number of DMH-induced intestinal tumors than those fed 7.5% protein (Topping and Visek, 1976). However, the percentage of animals with colon tumors was not significantly different between various dietary groups. Whether the reduced number of tumors in rats fed 7.5% protein diet was due to suboptimal protein intake during the period of

NUTRITION A N D CANCER

26 1

rapid body growth could not be answered from this study, although the evidence seems compatible with the conclusion that the time of appearance of tumors and their size and number was influenced by protein intake (Topping and Visek, 1976). 3 . Effect of Fiber in Colon Carcinogenesis

It has been postulated that the protective effect of dietary fiber may be due to adsorption, dilution, and/or metabolism of cocarcinogens, promoters, and unidentified carcinogens by the components of the fiber. There is evidence that alfalfa, wheat straw, and some other fibers can bind considerable amount of bile acids (colon tumor promoters) in vitro (Kritchevsky and Story, 1974: Story and Kritchevsky, 1978; Kay, 1978). On the other hand, wheat bran, oat hulls, and synthetic fibers only bind negligible amounts of bile acids in tiitro. This indicates that the different types of nonnutritive fibers possess specific binding properties. Dietary fiber could also affect the enterohepatic circulation of bile salts (Kern et al., 1978). Fiber not only influences bile acid metabolism, thereby reducing the formation of potential tumor promoters in colon carcinogenesis, but also exerts a solvent-like effect in that it dilutes potential carcinogens and/or cocarcinogens by its bulking effect and ability to bind water, sterols, bile acids, and fat. Fiber may also influence the gut flora and cause decreased bacterial degradation of intraluminal carcinogens and/or carcinogens. Although the concept of fiber involvement in colon carcinogenesis is simple, attractive; and appears firmly based on logic, the data are often contradictory and confusing. The reasons for this discrepancy may be that the fiber terminology has been generally incorrectly used. Also, experimental design has failed to take into account the possible subtle effect of inhibitors, especially in relation to the promoting process. In evaluating the biological function of dietary fiber, information regarding the nature of the fiber often has been incomplete. Dietary fiber has been defined as that part of plant material taken in our diet that is resistent to the digestion by secretions of the gastrointestinal tract and that comprises a heterogeneous group of carbohydrates, including cellulose, hemicellulose, pectin, and a noncarbohydrate substance, lignin. Fibers can be generally classified into three groups according to Van Soest (1978): ( 1 ) vegetable fibers, which are highly fermentable with low indigestible residue: (2) brans, which are less fermentable; and (3) chemically purified fibers such as feed cellulose, which are relatively unfermentable. A class of soluble substances including pectins and gums may not be true fibers, but are considered part of the dietary fiber complex

262

BANDARU

s. REDDY et a / .

because of the similar effects they can elicit in the diet. Dietary fiber fractions of wheat bran consist principally of hemicellulose and smaller amounts of lignin and cellulose, whereas the dietary fiber fractions of vegetable and fruit fibers have a different percentage composition of hemicellulose lignin, and cellulose. The relationship between dietary fiber consumption and colon cancer has been studied in animal models. Sprague-Dawley rats fed a diet containing 20% corn oil or beef fat and 20% wheat bran had fewer DMHinduced colon tumors than those on a control diet containing 20% fat and no bran (Wilson et a / . , 1977). No differences in the incidence of colon cancer were found between rats fed corn oil and those fed beef fat. Recently, Freeman et u / . (1978) compared DMH-induced colon tumor incidence in Sprague-Dawley rats fed either a fiber-free diet or a 4.5% purified cellulose diet. Cellulose ingestion was associated with a reduced number of animals with colonic neoplasia, as well as a reduction in the total number of colon tumors. In addition, this protective effect appears to be time dependent and associated with a shift in tumor distribution from the proximal colon to a more distal site (Freeman et al., 1978). Although the precise mechanism for this apparent redistribution of the site of the tumors within the colon remains obscure, some change in the luminal physiochemical environment from proximal to distal colon or some inherent difference between the colonic mucosa itself from these two sites may be responsible for the observed differences. A recent study of Fleiszer et a / . (1978) indicates that the incidence of DMH-induced colon tumor in rats reduces as dietary fiber increases. In this study, some reduction in tumor incidence in the high-fiber group might be expected on the basis of reduced caloric intake. However, the data suggest that reduction of caloric intake alone cannot account for the significant protective effect of dietary bran. In another study, Cruse et al. (1978) reported that a diet containing 20% wheat bran had no effect on DMH-induced colon carcinogenesis in rats. However, the DMH dose levels in this experiment were so high that any protective effect of bran might have been unobservable. (One important concern in a study of the effect of diet on chemical carcinogenesis is to avoid too high a level of carcinogen for a prolonged period, as this may obscure more subtle changes induced by certain dietary modifications.) In fact, the data presented by Cruse er a / . (1978) suggest that a high-fiber diet reduces the number of DMH-induced early deaths in rats (Lowenfels, 1979). The effect of dietary alfalfa, pectin, and wheat bran at a 15% level on colon carcinogenesis by AOM or MNU was studied in F344 rats (Watanabe et a/., 1979). The animals fed the alfalfa diet and treated with MNU had a higher incidence of colon tumors than did those fed the control

263

NUTRITION A N D CANCER

diet containing only 5% alphacel or diets containing pectin or wheat bran (Table IX). There was no difference in MNU-induced colon carcinogenesis in rats fed diets containing pectin or wheat bran. The AOM-induced colon carcinogenesis in rats fed diets containing pectin or wheat bran was lower than that in rats fed the control diet or alfalfa diet (Table IX). The effect of alfalfa, wheat bran, and cellulose on AOM-induced intestinal tumor incidence was further studied in Sprague-Dawley rats fed diets containing 10% fiber and 35% beef fat or 20 or 30% fiber and about 6% beef fat (Nigro et af., 1979). The addition of 10% fiber to the high-fat diet did not reduce the intestinal tumor frequency. Apparently, the challenge of AOM with the high dietary fat was too great to be affected by the dietary fiber. The addition of 20% bran or cellulose or 30% of any fiber to the 6% fat diet significantly reduced the intestinal tumor frequency. In the proximal half of the large bowel, all dietary groups except the 20% alfalfa group showed a reduction in tumor frequency compared to the fiber-free groups. The concentration, but not the total daily excretion of fecal steroids, was significantly lowered in the groups with reduced tumor frequencies. Bauer et al. (1979) have indicated that the protective effect of dietary fiber in colon carcinogenesis is probably at the promotional stage rather than at the initiating period. Groups of rats were fed a fiber-free diet or diets containing 20% wheat bran, 20% carrot fiber, or 6.5% citrus pectin from 3 days prior to the first DMH administration until 14 days after the last injection. They were then transferred to a standard rat pellet diet for about 10- 12 weeks. There was no difference in the incidence of colorectal tumors among the groups fed a fiber-free diet and diets containing wheat TABLE IX COLONTUMORINCIDENCE I N FEMALE F344 RATS FED DIETSCONTAINING PECTIN, A L F A L F A , OR WHEAT BRANA N D TREATED W I TH AZOXYMETHANE OR METHYLNITROSOUREA Percentage of animals with colon tumors Diet Control Pectin Alfalfa Wheat bran

Azoxymethane treated

Methylnitrosourea treated

57 I 00 53 3 30

69

59 83b

60

Significantly different from the groups fed the control diet or alfalfa diet by

P < 0.05.

Significantly different from the other groups; P < 0.05.

x2 test;

264

BANDARU

s. REDDY et al.

bran or carrot fiber. However, it is possible in this study that the high tumor yield resulting from large doses of DMH failed to show any protective effect of dietary fibers. In addition, these results (Bauer et al., 1979) and those of others (Watanabe et al., 1979; Nigro et al., 1979; Wilson et al., 1977; Fleiszer e f al., 1978) suggested that continuous feeding of high-fiber diets had a protective effect on colon carcinogenesis, whereas transferring from high-fiber diet to low-fiber diet during postcarcinogen treatment resulted in no observable effect. These observations suggest that dietary fiber exerts a protective effect on tumorigenesis during the promotional phase. These results also indicate that the protective effect of dietary fiber in colon carcinogenesis depends on the source of fiber and type of carcinogen. The inhibition of tumor formation by dietary fiber may be due to the dilution of promoters in the lumen of the large intestine by the additional bulk. The protective effect of various fibers also depends on their capacity to bind bile acids in the intestinal tract, as well as their effect on colonic mucosa and indirect effects on the metabolism of the carcinogen. Although additional studies are warranted to elucidate the protective effect of various fibers in colon carcinogenesis, the human data and animal experiments suggest that increased intake of cereal fibers would, at least in part, reduce the risk for large-bowel cancer. 4. Bile Acids in Colon Carcinogenesis

The role of bile acids in colon carcinogenesis has received some support from studies in animal models. Nigro et al. (1973) observed that feeding of nonabsorbable resin, cholestyramine, which increases bile salt excretion, enhanced azoxymethane-induced colon tumors in rats. Although this effect was mediated through an increased fecal bile salt excretion, one could not exclude the possibility of some direct effect of cholestyramine itself in stimulating cell duplication. In another study, the carcinogenic effect of AOM in rats was increased by surgically diverting bile to the middle of the small intestine, which also raised the fecal excretion of bile salts (Chomchai et al., 1974). In an experiment pertinent to this area, B. S. Reddy (unpublished observations) has observed that cholestyramine-bound taurocholic acid or taurochenodeoxycholic acid, when incubated with mixed fecal cultures isolated from human stool, are deconjugated and further modified to free bile acids. The evidence of the importance of bile acids as colon tumor promoters came from studies by Narisawa et al. (1974) and Reddy et al. (1976c, 1977d; Reddy and Watanabe, 1979) (Table X). The development of adenomas significantly increased among those conventional rats initiated

265

NUTRITION AND CANCER

TABLE X COLONTUMORINCIDENCE I N GERMFREE A N D CONVENTIONAL RATSTREATED WITH INTRARECTAL MNNG AND/OR BILEACIDS Animals with tumors (percent) Germfree CA ( 10)"~b CDC (10) LC (10) MNNG (22) MNNG + CA (24) MNNG + CDC(24) MNNG + LC(24) Conventional CA (12) CDC (12) LC (12) MNNG (30) MNNG + CA (30) MNNG + CDA (30) MNNG + L C ( 2 4 )

0 0

0 27 50 54 71' 0 0 0

37 67' 70' 83'

Tumors per rat Total

Adenocarcinoma

Adenoma

0 0 0 0.27 0.63 1.08 1.04

0 0 0 0.14 0.29 0.29 0.33

0 0 0 0.13 0.34 0.79 0.71

0

0 0

0 0 0 0.32 0.63 0.96 1.50

0 0 0.55 0.87 1.23 1.83

0 0.23 0.24 0.27 0.33

CA, cholic acid: CDC, chenodeoxycholic acid; LC, lithocholic acid. Number of rats are shown in parentheses. CA, CDC, or LC group received intrarectally 20 mg of sodium salt of respective bile acid three times weekly for 48 weeks; MNNG group received intrarectally 2 mg of MNNG twice a week for two weeks followed by vehicle for 46 weeks; MNNG + CA, MNNG + LC, or MNNG + CDC group received intrarectally MNNG for two weeks and bile acid thereafter for 16 weeks. Significantly different from rats given MNNG alone by xz test, P < 0.05.

with limited amounts of intrarectal MNNG to give a definite low yield of colon cancer and administered with intrarectal lithocholic acid or taurodeoxycholic acid as promoters compared to the group that was given only the carcinogen. Deoxycholic acid applied topically to the colon increased MNNG-induced colon adenocarcinomas in germfree rats. The bile acids themselves did not produce any tumors. A recent study also indicates that the primary bile acids, cholic acid and chenodeoxycholic acid, also produced a MNNG-induced colon tumor promoting activity in rats (Reddy et af., 1977b). Cholic acid and chenodeoxycholic acid given intrarectally to conventional rats are subjected to bacterial 7a-dehydroxylation to deoxycholic acid and lithocholic acid, respectively. Cohen et af. (1979) reported that cholic acid in the diet increased MNU-induced colon carcinogenesis in rats. Total fecal bile

266

BANDARU

s. REDDY et al.

acids, particularly deoxycholic acid output was elevated in animals fed cholic acid compared to controls. This increase in fecal deoxycholic acid was due to bacterial 7a-dehydroxylation of cholic acid in the colonic contents. These studies demonstrate that these secondary bile acids have a promoting effect in colon carcinogenesis. The mechanism of action of bile acids in colon carcinogenesis has not been elucidated. Bile acids have been shown to affect cell kinetics in the intestinal epithelium, although the structural specificity of this effect has not been examined extensively (Bagheri et al., 1978; Ranken et al., 1971; Roy et al., 1975). In the intestine, the data do not permit a critical distinction between a direct effect of bile acids on cell division and an indirect or physiological stimulus secondary to increased cell loss from sloughing or damage (Bagheri et al., 1978). The cell renewal system is dynamic and may be influenced by changes in a number of factors including the composition of gut microflora (Matsuzawa and Wilson, 1965) and bile acids in the intestine (Meslin et al., 1974). It has been shown that, with proper modification of the microenvironment of the intestinal tract, it is possible to alter the cellular kinetics of the mucosa (Mastromarino and Wilson, 1976). Recently, Cohen et a f . (1979) reported an enhanced colonic cell proliferation in rats fed cholic acid, as well as in animals treated with intrarectal MNU. This increased cell population involved in DNA synthesis induced by cholic acid feeding would favor the expression of damage at a far higher level than with the carcinogen, MNU, alone, bringing about not only a greater overall incidence of MNUinduced colon tumors, but also an enhanced number of tumors in rats fed cholic acid. Lipkin (1975) demonstrated that, during neoplastic transformation of colonic cells, a similar sequence of changes leading to uncontrolled proliferative activity develops in colon cancer in humans and in rodents given a chemical carcinogen that induces colon cancer. Irrespective of the mechanism by which bile acids enhance cell proliferation and/or decrease the generation time of proliferating cells, the phenomenon may have important implications for colon carcinogenesis.

5 . Mutagens (or Presumptive Carcinogens) Until recently, the nature of the carcinogens responsible for colon cancer, as well as for other important types of cancer, was obscure. The hypothesis has been put forward that the dietary and body fat acts as a reservoir for environmental contaminants such as polycyclic aromatic hydrocarbons, PCB, PBB, and DDT. None of these contaminants have ever induced cancer in animal models in the colon, breast, or prostate. An important clue to the nature of carcinogens came from the studies

NUTRITION AND CANCER

267

of Nagao et al. (1977a,b) who demonstrated the presence of mutagens (presumptive carcinogens) in the charred surface of beef and fish. They speculated that this activity was the result of pyrolysis of the proteins and subsequently demonstrated mutagenicity of protein pyrolysates. Pyrolysis of individual amino acids was carried out, and the pyrolysate of tryptophan had the greatest mutagenic activity (Matsumoto et al., 1977). The active principle in this pyrolysate was a y-carboline derivative, a heterocyclic o-methylarylamine (Sugimura et a / ., 1977). Other o-methylarylamines (i.e., 3,2’-dimethyl-4-aminobiphenyl)have been shown to produce colon cancer in male and female, and breast cancer in female animals (Reddy and Watanabe, 1978). In view of the well established connection between mutagenic activity and carcinogenic activity, and the known properties of certain o-methylarylamines, especially in causing colon cancer in rats, it is possible that broiling and frying of meat may lead to carcinogens responsible for cancer of the colon (Weisburger and Spingarn, 1979). However, it is important to determine specifically whether such derivatives resulting from broiling and frying of meat can induce cancer in epithelial tissues, such as colon or breast. Recently, Commoner et a / . (1978) observed the formation of mutagenic activity after frying hamburgers. They also investigated the temperature dependence of this phenomenon and found a sharp rise in mutagen formation between 140 and 180°C (Dolara et af., 1979). Spingarn and Weisburger (1979), however, have shown that substantial mutagenic activity is formed whether the meat is fried, broiled, or boiled, despite the finding that surface temperatures of broiled patties do not exceed 130°C. The hypothesis that browning reactions, which are desirable for producing the flavor and aroma of cooked foods, are responsible for the mutagenic activity, has been investigated in a model system. The reactions between sugars and amines have long been used to investigate the reaction products in browning (Hodges, 1967), and these reactions have now been shown to produce mutagenic activity as well (Weisburger and Spingarn, 1979). The ubiquitous occurrence of browning reactions in cooking suggests that these mutagens are probably not restricted to cooked meats (Weisburger et a / . , 1980). Recently Bruce et al. (1977) found in the stools of some individuals mutagenic substances that they thought were N-nitroso compounds that might be responsible for colon cancer. It will be important to determine whether this mutagen stems from the metabolism of a mutagen from fried meats or whether it is derived from other precursors such as through a nitroso exchange reaction (Mandel et af., 1977) or from an as yet unknown pathway, including a nitrosation at the pH prevailing in the large intestine.

268

BANDARU

s. REDDY et al.

Bruce et al. (1977) also demonstrated that increased dietary fiber and decreased dietary fat and protein reduced fecal mutagen levels. A marked diminution of mutagen concentration could also be brought about by supplementing diets with either ascorbic acid or a-tocopherol. Autrup er al. (1978) demonstrated that human colonic mucosa can activate several types of carcinogens such as N-nitrosamines into forms that bind to DNA. Mandel et al. (1977) demonstrated that the stools of certain people contained a bacterial enzyme that can transfer a nitroso group from a nitrosamine to an amide with consequent production of what is presumably a direct-acting nitrosamide. More research in this area is necessary to determine the relevance of these findings to colon cancer in man.

6. Effect of Miscellaneous Dietary Factors in Colon Carcinogenesis Cruse et al. (1979) proposed that prolonged exposure to dietary cholesterol is cocarcinogenic for human colon cancer, since it facilitates the development, growth, and spread of the disease, and since dietary fats promote the action of several experimental carcinogens. Broitman et al. (1977) studied the effect of polyunsaturated fat and cholesterol on colon tumorigenesis and demonstrated that the interaction between dietary polyunsaturated fat and dietary cholesterol and/or tissue cholesterol may promote tumorigenesis compared with dietary saturated fat and cholesterol in the animal model. Our recent studies indicate that cholesterol does not act as a colon tumor promoter in the rat model (Reddy and Watanabe, 1979). There is considerable evidence to suggest that vitamin A, its synthetic and natural analogs such as retinol, and the esters and ethers of retinol, retinoic acid and synthetic analogs, can influence the development of some epithelial tumors (Sporn et al., 1976). Because vitamin A is necessary for the control of proliferation and the direction of differentiation of many epithelial tissues, its ability to act on colon carcinogenesis has received much attention. Rogers and Newberne (1973) found that hypervitaminosis A did not alter aflatoxin B,-induced colon tumors in rats. Chronic dietary deficiency of vitamin A slightly increased the incidence of colon tumors in rats receiving DMH by repeated intragastric administration (Rogers et al., 1973, 1974). They also found that high levels of vitamin A in the diet provided little or no protective effect on DMHinduced colon carcinogenesis. However, intrarectal injection of the direct-acting carcinogen, MNNG, gave a large-bowel tumor incidence in vitamin A-deficient animals that was one-half that of vitamin A-supplemented animals (Narisawa et al., 1976). These studies do not address the possibility that retinoic acid or its analogs might show some protective

NUTRITION .AND CANCER

269

effect. A recent study by Newberne and Suphakarn (1977) indicates that dietary supplementation with 13-cis-retinoic acid reduced the incidence of DMH-induced colon carcinogenesis in rats. Many synthetic antioxidants, such as butylated hydroxytoluene, butylated hydroxyanisole, and disulfiram, studied by Wattenberg (1978), Weisburger et (11. (1977), and Fiala (1977), provided one of the most consistent examples of inhibition of carcinogenesis. This subject has been reviewed recently by Wattenberg (1978). BHA is of interest because of its minimal toxicity and its extensive use as an additive in food for human consumption. Experimental studies on the inhibition of colon carcinogenesis by disulfiram, diethyldithiocarbamate, and bisethylxanthogen indicate that these compounds, when added to the diet, inhibit DMH-induced colon carcinogenesis (Wattenberg 1975, 1978). Disulfiram has also been found to inhibit AOM-induced colon carcinogenesis but to a considerably lesser extent than DMH (Wattenberg, 1978). Fiala (1977) demonstrated that carbon disulfide, a metabolite of disulfiram, inhibits the oxidation of DMH and AOM in vivo. The data suggest that carbon disulfide may be the chemical species responsible for the inhibitory action of disulfiram and related compounds. When incubated with microsomes, several thiono-sulfur-containing compounds, including disulfiram and diethyldithiocarbamate, produce a decrease in cytochrome P-450 (Hunter and Neal, 1975). This raises the possibility that thiono-sulfur-containing compounds as a group may have the capacity to modify cytochrome P-450so as to alter the microsomal metabolism of DMH, AOM, and other carcinogens in a manner that decreases their carcinogenicity (Wattenberg, 1978). There is some experimental evidence in animal models for the cancer inhibitory activity of selenite. Jacobs et a / . (1977a) reported that 4 ppm of selenium in drinking water inhibited DMH- or MAM acetate-induced colon tumors in rats. Selenium also reduced the mutagenic activity of 2AAF, N-OH-AAF, and N-OH-AF in the Ames test system (Jacobs et al., 1977b). The mechanisms by which these inhibitors operate is incompletely understood. In those instances in which information is available, the inhibitors may act by altering the metabolism of the carcinogen (Wattenberg, 1978). An evaluation of the current and potential role that these inhibitors may play is dependent on acquisition of further data on the range of compounds having the capacity to inhibit carcinogenesis and on mechanisms of inhibition. Despite the voluminous literature that exists on vitamin C, no adequate studies are available concerning the effect of ascorbic acid on the neoplastic process. Vitamin C intake may control the recurrence of largebowel neoplasia or adenomatous polyps in man after surgical interven-

270

BANDARU

s. REDDY et al.

tion. DeCosse et al. (1977) demonstrated that vitamin C given to patients with familial polyposis inhibited the formation of colorectal cancer by totally unknown mechanisms. Recently, Reddy et al. (1979) showed that the addition of 0.25 or 1.0% sodium ascorbate in the diet reduced the number of rats developing DMH-induced colon tumors from 26 to 0%. The above studies, in general, stress the importance of micronutrients such as selenium, vitamin A, and vitamin C as inhibitory factors in colon carcinogenesis. Except for the selenium, none of the other micronutrients have been studied in detail in colon carcinogenesis.

E. CONCLUSIONS In recent years, salient advances have taken place in our knowledge of factors in the etiology of large-bowel cancer in man. Through epidemiology and geographic pathology, we have learned that high fat and lack of fiber in the diet are involved in the genesis of this disease. Other factors indicated to be of importance by experimental observations include various micronutrients and antioxidants such as vitamins A and E, and selenium, although there is no significant epidemiologic basis for an influence of these compounds on colon tumor incidence. We have begun to appreciate the complexities through which diet translates its action in leading to colon cancer, involving mammalian and microbiologic enzymic, metabolic conversion steps of endogenous, and perhaps unidentified exogenous materials. Laboratory studies have shown the importance of the interaction of a high-fat diet and the production of bile acids as potentially relevant in the etiology of colon cancer. Other studies also indicate that a high intake of dietary fiber, in spite of high dietary fat, not only leads to an increase in stool bulk, thus diluting carcinogenic and/or promoters in the intraluminal contents, but also modifies the metabolism of these intraluminal compounds. These studies thus suggest that both high intake of fat and low intake of fiber may be necessary for the full expression of risk to colon cancer. Specific animals models have been developed that permit the detailed study of mechanisms. Animal model studies also indicate that amount of dietary fat, rather than type of fat, and certain types of dietary fibers play a role in colon carcinogenesis. More research in collateral areas, such as physiology and biology of colon and associated cell systems, has provided the information that the bile acids in the gut enhance cell proliferation and/or decrease the generating time of proliferating cells, the phenomenon of which may have important implications for colon carcinogenesis. Reports on the mechanisms of carcinogenesis have placed emphasis on carcinogens. Attempts have been made in various laboratories to isolate

NUTRITION A N D CANCER

27 1

and identify the carcinogens affecting the colon. Some irivestigators claim that broiling and/or frying of meat and fish, and the browning reactions between sugars and amines, yield mutagens (presumptive carcinogens) responsible for cancer of the colon. Others have reported the presence of mutagenic N-nitroso compounds in human feces that might be responsible for colon cancer. Although many substances are tumorigenic in experimental animals and a lesser number are carcinogenic in man, significant modifying factors enhance the effect of low-dose or low-potency carcinogens that by themselves would not suffice to induce cancers. In some instances, the modifying factor can be readily identified and removed, thereby eliminating or reducing the incidence of certain cancers. This may be difficult with initiating carcinogens. In the case of colon cancer, evidence has been presented that dietary fat exerts a promoting effect on tumorigenesis, whereas certain dietary fiber exerts a protective effect. Therefore, rather than concentrating on specific carcinogens, more attention must be given to modifying factors-cocarcinogens, promoters, and factors that influence the formation of endogenous tumorigenic compounds. With respect to further research, we need to continue and indeed to expand work in metabolic epidemiology and in laboratory studies to pinpoint the specific agents related to colon carcinogenesis. New approaches to prevention are urgently needed. However, while such studies are in progress, we think the time has come to make dietary recommendations to the public. Cardiovascular disease prevention has led to recommendations for a reduction in total fat as a prrident measure, to reduce the leading cause of death in the Western world. It would now appear that such a diet is also likely to lead to a reduction of colon cancer, particularly if this diet also includes a high intake of fibers. A reduction in the portion of total fat in average caloric intake, from 40% to about 25-30%, and a generous intake of dietary fiber in the form of whole-grain breads and cereals should be considered. Diet should be well balanced, including ample fresh fruits and vegetables to provide adequate vitamins and minerals. This was, in fact, recommended as one of the dietary goals for the United States by the Senate Select Committee on Nutrition and Human Needs.

Ill. Dietary Factors and Cancer of the Stomach

A. EPIDEMIOLOGY Gastric cancer shows widely varying cross-national incidence and mortality (Wynder et u / . , 1963; Hirayama, 1971, 1979; Haenszel and Correa,

272

BANDARU

s. REDDY et al.

1975; Bjelke, 1974; Modan et al., 1974). Areas with high incidence include Japan, Latin America west of the Andes, some parts of the Caribbean, Iceland, and Northern and Eastern Europe (Fig. 3). In contrast, Western Europe, the United States, Canada, Australia, New Zealand, and other Anglo-Saxon countries have low incidence. Age-adjusted death rates for stomach cancer in the United States were about 8/100,000 in white males and 17/100,000 in black males (Segi et al., 1969). However, this low incidence is only a recent development, for gastric cancer was the most common cancer in the United States 40 years ago. It seems that cancer of the glandular stomach with antecedent intestinalization has a different etiology and needs to be distinguished from diffuse stomach cancer. It may be that the diffuse kind is associated with blood group A, and possibly, pernicious anemia (Haenszel et al., 1976; Muiioz and Asvall, 1971; Muiioz and Connelly, 1971). In addition, recently, gastric stump cancer, noted some 15-30 years after partial gastrectomy for peptic ulcer, has been observed mainly after a Billroth FEMALE

FIG.3 . Age-adjusted death rates for malignant neoplasms of stomach in different countries, 1%6-1967. (From Segi and Kurihara, 1972.)

NUTRITION A N D CANCER

273

procedure. Domellof (1979) has reviewed this iatogenic disease entity and has noted that possible alkaline bile reflux may be an important promoting factor. The further discussions of gastric cancer refer mainly to the occurrence and mechanisms relevant to cancer in the glandular stomach and do not relate directly to diffuse stomach cancer. In addition to the marked international variation in risk, variation within countries is also observed, the general rule being that northern and/or colder regions have higher risks (Haenszel and Correa, 1975). One example is the higher stomach cancer rates in the mountainous region of Croatia, Yugoslavia, as contrasted with the lower risks in the Adriatic coastal zone. Other examples include the lower mortality rates in the Japanese prefectures of southern Kuyushu (Segi et al., 1965) and particularly high mortality in the northwestern part of Iceland (Sigurjonsson, 1966). Certain cities in the tropical zone of Latin America (Bogota, Cali, Guatemala City, and Lima) have elevated risks for stomach cancer; this indicates that populations in the mountainous central Andean region are at high risk, whereas residents of the tropical coastal zones in Latin America are at low risk (Haenszel and Correa, 1975). The incidence of stomach cancer is higher in males than in females within all populations, the overall female rate being roughly one-half to two-thirds of the corresponding male rate. The sex difference for stomach cancer has been investigated by Griffith (1968) and has been shown to be age-dependent. The male-to-female ratio is close to 1 at ages under 35, reaches a peak of about 2 at around 55, and thereafter declines to about 1.3 to 1.5 at the oldest ages. This pattern, which was peculiar to cancer of the stomach, was observed in mortality statistics from 24 countries, as well as in incidence data from a number of countries. The findings may be explained by histological differences in age- and sex-specific rates (Haenszel and Correa, 1975). The behavior of the sex ratios provides support for the thesis that gastric carcinoma is not a homogenous entity but is composed of at least two distinctive etiologic components (Haenszel and Correa, 1975). A marked inverse socioeconomic gradient in risk was a prominent characteristic of this disease, the risk for the lower class being roughly 2.5 times that for the higher socioeconomic class (Haenszel and Correa, 1975; Berndt et al., 1968; Graham et al., 1960; Hirayama, 1971). The major clues for the etiology of stomach cancer have come from utilization of the epidemiologic data of human migration (Haenszel, 1975). The United States foreign-born, migrating from countries with high risks for stomach cancer, continued to experience the risk characteristic of the population of origin. Risk characteristics of the host population did not appear until the succeeding generation was born in the United States

274

BANDARU

s. REDDY et al.

(Haenszel, 1961; Bjelke, 1974). Similar findings have been reported from Australia (Staszewski et al., 1971). A marked excess of stomach cancer cases in Cali, Colombia, has been observed among migrants born in the mountainous region bordering Ecuador (Correa et al., 1970). The case-control study of Hawaiian Japanese revealed that migrants from the Japanese prefectures with highest stomach cancer risks continued to experience an excess risk in Hawaii, but this effect did not persist among their Nisei (second generation) offspring (Haenszel et al., 1972). Lower risks were suggested for the Nisei, but not Issei (first generation) who adopted Western style diets (Stemmermann, 1977; Haenszel et al., 1972). These distinctions reinforced earlier inferences from migrant study data on the critical nature of exposures in early life (Haenszel, 1975). This decrease in stomach cancer mortality among second generation Hawaiian Japanese is probably not the result of improved diagnostic or treatment methods, but can be explained on the basis of a decreased consumption of traditional Japanese foods. Investigations of migrant populations benefited from close coordination in the collection of epidemiological and pathological observations (Haenszel and Correa, 1975). Mufioz et al. (1971) and Correa et al. (1970) classified stomach cancers in Latin American populations at high- and low-risk and reported that the intestinal type predominated among cases in high-risk areas. Typing of stomach cancers in Cali, Colombia, indicated that the intestinal type tumors accounted for most excess incidence in this subpopulation. The pathology studies in Miyagi prefecture and Hawaii also revealed that the incidence for diffuse carcinomas was substantially the same in both localities, but the incidence for intestinal, mixed, and other types was markedly lower in Hawaii (Correa et al., 1973). Several investigators have correlated national per capita average amounts of selected food items and nutrients available with mortality from cancer of the stomach, and the findings have been reviewed by Bjelke (1974), Haenszel and Correa (1979, Graham (1979, and Tulinius (1979). Internationally, negative correlations have been found between agestandardized mortality from stomach cancer and the consumption of fats and oils, animal protein, and sugar (Lea, 1967; Gregor et al., 1969) and a positive correlation with cereals (Hakama and Saxen, 1967). Recent data show a negative correlation between incidence and mortality rates of stomach cancer with total fat consumption and a positive association between incidence rates and fish consumption (Armstrong and Doll, 1975). However, the positive association for fish consumption is entirely dependent on the extreme values for Japan and Iceland. Fish consump-

NUTRITION A N D CANCER

275

tion is, therefore, unlikely to contribute significantly to international variation in gastric cancer, even though relevant to Iceland and Japan. In certain countries, high stomach cancer incidence in certain areas has been related to high intake of salted foods (Hirayama, 1971; Joossens, 1979) and certain smoked foods (Dungal and Sigurjonsson, 1967), and has been associated with widespread use of sodium nitrate as a fertilizer and high levels of nitrate in the drinking water (Hill et al., 1973; Zaldivar, 1978). In Japan, stomach cancer incidence is particularly low in some areas where sweet potatoes are a staple food (Segi et al., 1965). The relative abundance of sweet potatoes and other vegetables in the southernmost prefecture, Kagoshima, which has the lowest mortality from stomach cancer, has been pointed out (Segi et al., 1965). Dungal and Sigurjonsson (1967) observed higher intakes of vitamin C due to higher intakes of potatoes and rutabagas in the low mortality areas of Iceland. The low rates in Yugoslavian areas along the Adriatic Coast may also suggest a protective effect due to the availability of fruits and vegetables. A survey of dietary habits of the people of Chokai Village in rural Akita Prefecture, Japan (high death rates for stomach cancer) and people in Hawaii indicate that the specific food items that comprise the diet of the Hawaiian Japanese are very different from those that are characteristic of both the Hawaiian Caucasian and the indigenous Japanese (Stemmermann, 1977). The most conspicuous feature of the diet in Akita is a large intake of dried fish and miso soup, both of which have a high salt content. On the other hand, the Hawaiian Japanese have a higher intake of uncooked vegetables such as celery, lettuce, and tomato and fresh fruit juices (Stemmermann, 1977; Hirayama, 1979). Past case-control studies of diet and stomach cancer have been complicated by problems of informant recall and accuracy of histories and have not been rewarding, but diet inquiries have steadily improved. A case-control study of gastric cancer in Hawaiian Japanese by Haenszel er al. (1972) identified elevated gastric cancer risk with the consumption of pickled vegetables and dried salted fish and a rise in risk with increased frequency of these food items. Low risks were identified for several Western-type vegetables such as lettuce, celery, tomatoes, and corn, and the latter effects appeared to be independent of the associations with Japanese foods, suggesting possible protective effects. Although the companion case-control studies conducted in Hiroshima and Miyagi prefectures of Japan did not reproduce the associations with pickled vegetables and salted dried fish reported for Hawaiian Japanese, the greater use of lettuce and celery reported by controls in Japan rein-

276

BANDARU

s. REDDY et al.

forced the Hawaiian Japanese results on possible protective effects of Western-type vegetables (Haenszel and Correa, 1975). Case-control studies of stomach cancer were also conducted by Bjelke (1974) in Norway and in Minnesota, USA, for persons mostly of Scandinavian descent. In Norway, (1) recent use of cooked cereals and salted fish was higher among stomach cancer patients than among controls, and (2) more pronounced case-control differences were shown by a number of vegetables and fruits, which were used less frequently by the stomach cancer patients. The greatest deviations from controls were shown for the indices for total vegetables and vitamin C intakes, for which relative deviations were greatest among young patients and women. In Minnesota, (1) recent use of cooked cereals, smoked fish, and canned fruits was higher and intakes of lettuce and tomatoes were lower among stomach cancer patients than among controls, and (2) whereas total intakes of cereal products and fish were only slightly higher than among the controls, the index for total vegetables was considerably lower among the stomach cancer patients. In both Norway and Minnesota, the lower vegetable and vitamin C intakes among stomach cancer patients had persisted over a long period, were more pronounced among women, and, in both sexes were mainly a feature of the diffuse carcinomas (Bjelke, 1974).

Similarly, the diet of gastric cancer patients in the U.S. has been found to include fewer vegetables (Graham, 1975). Israeli case-control studies indicate a higher consumption of starches in gastric cancer patients (Modan et al., 1975). In Japan, which had the highest death rate due to gastric cancer, a correlation has been reported between the consumption of salted fish and the incidence of gastric cancer. Crude salt used for preserved fish may contain nitrate, which can be reduced to nitrite. Chile and some regions of Colombia and Costa Rica with high death rates due to gastric cancer possess large nitrate deposits and correspondingly elevated levels of nitrate in foods and drinking water (Cuello et al., 1978; Zaldivar and Robinson, 1973). A positive correlation between nitrate levels in drinking water and the incidence of gastric cancer has been established in Worksop, England, by Hill et al. (1973). In Worksop, the high nitrate content of the drinking water resulted in a weekly nitrate intake more than double that of people living in Paddington, England, a low-risk area for gastric cancer (Hill et al., 1973). The daily excretion of nitrate in the urine of Worksop people was three times higher than in Paddington. Similar results were obtained by Cuello et al. (1978) in Colombia. These factors alone, however, may not be sufficient to account for the complex etiology of gastric cancer.

NUTRITION A N D CANCER

277

In summary, the above studies indicate that a high-risk population consumes a diet high in carbohydrates and low in fat and limited protein, limited micronutrients on an annual basis, and of greater relevance, low levels of select micronutrients, especially vitamin C , on a seasonal basis. They also consume limited amounts of fresh fruits and vegetables. Other risk factors also include intake of pickled, highly salted, and smoked foods or foods grown in soils high in nitrate. Studies on migrants from high-risk to low-risk regions have recorded that people with lowered risk include lettuce and other fresh greens as part of their daily diet. B. ETIOLOGY The fact that gastric cancer has been shown in many studies to be strongly related to such social factors as ethnic background, occupation, and socioeconomic status, suggests that environmental factors play an overriding role in its etiology. The declining incidence and mortality rates observed worldwide in recent years is the most convincing evidence of this assertion. The weight of the evidence that environmental factors are of major importance came from migrant studies. The issue is to elaborate the relationship between these factors and nutrition and diet in the carcinogenic process that produces cancer of the stomach. The nitrosamines probably constitute the group of carcinogens most intensively studied (Drasar and Hill, 1974). The possible role of nitrosamines in human gastric cancer has excited much interest since nitrate is widely used by the food industry as a preservative, and is, in a n y case, naturally present in many foodstuffs and in sources of drinking water. Secondary amines are also present in many foodstuffs, although the documentation for this is somewhat scarce, but there is a body of circumstantial evidence compatable with the hypothesis that microflora can produce nitrosamines in v i m and that these are relevant to human gastric cancer. Nonetheless, as will be discussed, no nitrosamine is known to induce glandular stomach cancer in animal models, but nitrosamides do so.

C. METABOLICEPIDEMIOLOGY The question requiring definition is the identification of a carcinogen responsible for the induction of gastric cancer in man, which would also explain the difference in geographical distribution of this disease. Inasmuch as gastric cancer can be induced by alkylnitrosoureas in

278

BANDARU

s. REDDY et al.

animal models, the question could be asked whether gastric cancer in man might also be caused by exposure to this type of chemical (Druckrey, 1973, 1975; Sugimura and Kawachi, 1973, 1978; Mirvish, 1975; Raineri and Weisburger, 1975; Weisburger and Raineri, 1975). Sander et a / . (1973) made the important discovery that nitrosamines can be formed in vivo by nitrosation by nitrite of suitable substrate amines or amides in the stomachs of animals or man. Later, others established that cancer at various sites could be induced with identical results either by administration of a preformed nitrosamine or nitrosamide or by nitrite together with the corresponding amine or amide (Fan and Tannenbaum, 1973; Mirvish and Chu, 1973). Sander (1968) first demonstrated that certain strains of enterobacteria were able to N-nitrosate diphenylamine at neutral pH using nitrate as the source of the nitrous group. This work was confirmed and extended by Hawksworth and Hill (1971a,b) and by Hill and Hawksworth (1972). The reaction involving nitrate is only carried out using nitrate-reducing bacteria, but some strains of bacteria that do not produce nitrate are able to nitrosate secondary amines using nitrite as the nitrosating agent. The reaction could be brought about by an enzymic reaction, or it could be due to the bacteria producing the reactants and the conditions necessary for the reaction. Since bacteria are able to promote the catalysis of the N-nitrosation reaction at physiological pH values, the sites in which nitrosamines might be formed in vivo must be extended to include all those in which nitrate, secondary amine, and bacteria might coexist (Drasar and Hill, 1974). Since bacteria are able to reduce nitrate to nitrite, the important factor is no longer the amount of ingested nitrate (Drasar and Hill, 1974). Secondary amines are present in the stomach in small amounts dependent on the nature of the food consumed. The possible role of salivary nitrate or nitrite is being investigated by Tannenbaum et al. (1978a). Nitrite found in normal human saliva appears to be the product of microbial reduction of nitrate that is secreted by the salivary glands upon oral intake and absorption of nitrate. The data support the hypothesis that (1) nitrite formed in the oral cavity contributes to nitrosamine and nitrosamide formation in the normal stomach, (2) nitrosamides can be formed in the weakly acidic or even neutral stomachs, and (3) nitrosamines and nitrosamides are formed at other areas that normally contain bacteria or become infected (Tannenbaum et al., 1978b). Although these possibilities cannot be differentiated on the basis of their relevance to the etiology of gastric cancer in man, Tannenbaum et al. (1978a,b) suggest that both microorganisms and acidity are important for controlling nitrosation reactions in vivo. A situation in which the pH is not too low to prevent growth of bacteria and not too high to

NUTRITION A N D CANCER

279

prevent nitrosation formation would seem to represent the greatest hazard in the presence of catalysts (Tannenbaum er a/., 1979). In a recent study in Colombia, Tannenbaum et al. (1979) also reported that patients with diagnosed gastric pathology related to a precancerous state had high levels of nitrite in gastric contents with a pH above 5. These studies support, but do not prove, the hypothesis of bacterial reduction of nitrate in the stomach with concomitant formation of carcinogenic N-nitroso compounds. The question needs to be asked, however, whether the early events in gastric carcinogenesis would not occur in a high risk region during childhood or in young adults, perhaps before the onset, later in life, of hypochlorhydria. Migrant studies do suggest that the first 20 years of life may be the time of exposure or formation of gastric carcinogens.

D. EXPERIMENTAL STUDIES Because of the importance of gastric cancer in man in many parts of the world, serious efforts have been made by experimentalists to develop animal models (Bralow and Weisburger, 1976). Until recently, such efforts were relatively fruitless. While rodents, particularly mice, developed tumors of the forestomach upon treatment with certain of the classic carcinogens such as the polycyclic aromatic hydrocarbons or nitroaryl derivatives analogous to the carcinogenic aromatic amines, this lesion may not be a good model for the human disease, which generally originates in the glandular stomach. Stewart et af. (1961) discovered that 2,7bis-fluorenylacetamide could induce cancer of the glandular stomach in a relatively small, yet reliable, portion of rats at risk, but this procedure is not effective in all species under a variety of conditions. However, Sugimura made the important discovery that MNNG induced cancer of the glandular stomach in rats and other species in high yield (Sugimura and Kawachi, 1973, 1978). A number of investigations have now found this kind of agent to be the most reliable tool for studying model systems of gastric cancer. Druckrey er al. (1973) showed that N-methyl-”-acetylnitrosourea, nitrosobiuret, and similar substances also have this specific property. Thus, it seems that the molecular features of alkylnitrosamides can cause cancer of the glandular stomach. Salt intake in man (Joossens, 1979) and rats (Tatematsu er al., 1975) has a promoting effect. Inasmuch as the animal models of gastric cancer induced by alkylnitrosoureas have broad applicability across species lines, the question can be asked whether gastric cancer in man might also be caused by exposure to this type of chemical. Thus, it is important to establish whether such chemicals can enter man’s environment and, if so, under what conditions.

280

BANDARU

s. REDDY et al.

The well-established organ-specific carcinogenicity of nitrosamines and nitrosamides in animals (Magee and Barnes, 1956; Magee, 1971) is possibly of relevance to the development of human cancer, since the in vivo nitrosation of certain dietary amines and amides under the acidic conditions of the stomach could lead to the formation of these agents in human gastric juices, which have a pH range of 1-5 (Druckrey, 1975; Endo et al., 1974; Sander et al., 1973; Schoental and Bensted, 1969; Sugimura and Kawachi, 1978; Mirvish, 1975; Weisburger et al. (1980a,b). The formation of nitrite from nitrate, the conditions controlling the nitrosation of alkylamides, and the methods of inhibition of the nitrosation reaction have been studied by Weisburger and Raineri (1975). They have demonstrated that under realistic conditions, using foods eaten by populations typical of high-risk situations for gastric cancer, endogenous or added nitrate was converted in substantial amounts to nitrite when this food was stored at room temperature, but not in the refrigerator. Also, less nitrite was formed in the presence of high amounts of ascorbic acid, and in a model study, less carcinogenic methylnitrosourea was formed when nitrite and methylurea were combined in the presence of ascorbic acid. In relation to a working hypothesis that gastric cancer in man may result from the formation in the stomach of a locally active alkylnitrosamide, Weisburger and Raineri (1975) have established a possible source of nitrate (deliberately added or present as a result of geochemistry or agricultural practices) in food stored at room temperature. In the search for the etiologic factors responsible for gastric cancer, Marquardt e f al. (1977a,b) examined extracts of nitrite-treated food for mutagenic (presumptive carcinogenic) activity in the Ames Salmonella ryphimurium system. Extracts of Sanma fish, borscht, and beans, each of which is a dietary staple in Japan, Eastern Europe, or Latin America, areas with a high incidence of gastric cancer, showed mutagenic activity upon treatment with nitrite. In contrast, typical American foods such as hot dogs and beef failed to develop mutagenic activity with nitrite, perhaps because nitrite reacts preferentially with myoglobin. The formation of the mutagen was maximal at pH 3.0. In a dose-response study, incubation with 5000 ppm of sodium nitrite yielded the highest amount of mutagenic activity, but activity was also observed with 500 ppm. Moreover, and importantly, ascorbic acid prevented the formation of the mutagens in nitrite-treated foods. These data also suggest that the mutagen(s) may be of the alkylnitrosamide type. Pending the isolation and identification of the active mutagenic principle, Weisburger et al. (1980a) reported that the ether extract of the homogenate of Sanma fish treated at pH 3 with nitrite, when fed to Wistar rats, induced tumors mainly in the glandular stomach, and a few

28 1

NUTRITION A N D CANCER

in the pancreas and small intestine (Table XI). None of the animals fed the ether extract of Sanma fish alone showed any tumors. Thus, it would seem that the products obtained from pickling a specific kind of fish can induce glandular stomach cancer identical to that seen in man. Endo ef al. (1975) showed that nitrosation of methylguanidine under simulated gastric conditions produced a mutagenic principle identified as nitrosocyanamide. However, it has been found recently that fish and other foods probably do not contain significant amounts of methylguanidine (Fujinaka ef al., 1976). Also, although highly mutagenic, nitrosocyanamide induced mainly forestomach tumors in rats and thus failed to exhibit the specificity of inducing cancers of the glandular stomach, which are seen in man and which chemicals such as MNNG produce reliably. Thus, identification of mutagenic (presumptive carcinogenic) agents in nitrosated foods requires additional study. A thorough understanding of the conditions governing the formation of nitroso compounds, as well as an understanding of the natural occurrence of potential substrates, allows for a rational approach to the study of the etiology of gastric cancer in areas of high incidence.

TUMOR INCIDENCE

IN

TABLE XI RATS GIVENFISHEXTRACTWITH

WITRITE, OR

FISHALONE"

Number of animals ~~

Sites of tumors Effective number of rats Forestomach Papilloma Squamous cell carcinoma Glandular stomach Adenoma Adenocarcinoma Adenosquamouscarcinoma Pancreas Adenoma Adenocarcinoma Small intestine Adenocarcinoma From Weisburger er (I/. (1980).

Fish extract diet

Fish extract and NaNO, diet

8

10

0 0

0

0

2 2 I

0 0

2

0 0

2

0

1

1

282

BANDARU

s. REDDY et ul.

E. CONCLUSIONS Several investigators are now working on the experimental development of the hypothesis that gastric cancer in man may stem from in vivo nitrosation of an alkylamide type of substrate (Endo et al., 1974, 1975; Mirvish, 1975; Correa et al., 1975; Fan and Tannenbaum, 1973; Weisburger and Ranieri, 1975; Weisburger et al., 1980a,b). Current knowledge derived from epidemiology and animal experiments all support this hypothesis. If this concept could be borne out experimentally, it would appear that one relatively minor change in the human diet in certain high-risk population groups would prevent this important cancer. The required alteration would be to take at each meal fruits, vegetables, and salads as sources of vitamin C on a continuous rather than intermittent seasonal basis, as is presently the custom in many northern regions of several high-risk countries. It is necessary to emphasize the need for a continuous dietary intake of foods rich in vitamin C to prevent even an intermittent exposure to carcinogens, since in animal models, gastric cancer can be induced by relatively infrequent application of alkylnitrosamides. Also, epidemiologic data indicate that first generation migrants from high-risk countries like Japan, Poland, and Scandinavia maintain the risk for gastric cancer in their adopted country, suggesting that, once initiated, the reaction proceeds. Hence, there is a need to avoid formation of gastric carcinogens early in life and to continue this practice by minimizing the intake of actual or potential nitrite and optimizing the intake of foods containing ascorbate. Furthermore, in the light of the demonstrated enhancing effect of salt in experimental gastric cancer, and the known customary high salt intake in areas such as Japan, Iceland, and other high-risk regions, a reduction of dietary salt intake would also seem beneficial (Joossens, 1979).

IV. Dietary Factors and Cancer of the Upper Alimentary and Respiratory Tract

A . EPIDEMIOLOGY The epidemiology of upper alimentary tract cancer presents many unusual and interesting features. Striking variations in the incidence of esophageal cancer have been found within relatively small areas (Day, 1975). High rates of esophageal cancer have been reported in Central

NUTRITION A N D CANCER

283

Asia, namely, the Caspian littoral of Iran, Turkemenia, Kazakhstan, and Uzbekistan of the USSR, and Honan, Hopei, and Shansi Provinces of the Peoples Republic of China (Mahboubi ef al., 1973; Tuyns, 1970; Coordinating Group for Research on the Etiology of Esophageal Cancer of North China, 1974). The high-incidence region, called the esophageal cancer belt, begins in northern China, running from the Iranian Caspian littoral into Soviet Central Asia. High incidence of esophageal cancer is found in Singapore Chinese, in general, and particularly those speaking the dialects of Hokkien and Techew, who originate from Fukkien Province or the Swatow region in northern Kwantung Province (Shanmugaratnam and Wee, 1973). A case-control study of esophageal cancer among Chinese in Singapore indicated that those born in the Peoples Republic of China, irrespective of dialect group or sex, had a 3-fold increase in risk, as compared to those born in Singapore (deJong et al., 1970). In contrast, low incidences are found in West Africa. The incidence is low throughout Europe, with the exception of the northern regions of Brittany and Normandy in France, where esophageal cancer occurs predominantly among males. Both alcohol and tobacco have been shown to be associated factors. These data indicate that consumption of home-distilled apple brandy increases the risk for esophageal cancer. Although an association of alcohol consumption with esophageal cancer in France in the absence of tobacco consumption has been reported, the risk is dramatically increased when tobacco usage is concurrent with alcohol consumption (Tuyns and Masse, 1973; Tuyns et al., 1977). I n the United States, the incidence is higher in both male and female blacks compared to whites (Biometry Branch, National Cancer Institute, 1974). This increase appears most marked among the urban blacks in the north. Puerto Rico has a higher age-adjusted incidence rate of cancer of the mouth, pharynx, and esophagus than in the United States (Martinez et al., 1975) for every age group in both men and women. The predominance of malignant tumors of the upper alimentary tract in Puerto Rico reflects different nutritional and drinking habits from those in the United States survey areas (Martinez et al., 1975). In Utah, where the per capita consumption of alcohol and tobacco is well below the national average, the incidence of esophageal cancer was 55% lower than expected (Lyon et al., 1977).

B. ETIOLOGY Extensive epidemiological evidence has been available for some time now that indicates a strong association of both chronic alcohol and to-

284

BANDARU

s. REDDY et al.

bacco consumption with cancers of the head and neck area, specifically of the oral cavity (Wynder et ul., 1957), esophagus (Kamionkowski and Fleshler, 1965), and larynx (Wynder et al., 1956; Hinds et al., 1979) (Figs. 4 and 5). This association is supported further by the low incidences of these types of cancer in populations such as the Seventh-Day Adventists who are known to consume less alcohol than the general population (Phillips, 1975), as well as by the high incidences of these types of cancers in males. Because of the interaction of the smoking and drinking variables, it is difficult to assess the two independently since, most often, heavy drinkers are also heavy smokers. Even though some increased risk for cancer due solely to alcohol consumption is suggested by the data of Rothman and Keller (1972), a marked synergy is observed when tobacco usage is combined with chronic alcohol consumption. With the possible exception of esophageal cancer in France, the nature of the beverage consumed seems to be important only in terms of its ethanol content (Williams and Horn, 1977; Wynder and Stellman, 1977). The respective roles that alcohol and tobacco consumption play in the etiology of head and neck cancer have yet to be rigorously defined. However, it seems reasonable to assume that tobacco and/or tobacco

CASES= 239 CONTROLS' l r 7 2 5

21+

CIGARETTES PER DAY

I1

NONSMOKERS

0

1-6

7 +

02 OF ALCOHOL PER DAY

FIG.4. Relative risk of larynx cancer for daily consumption of alcohol and cigarettes for males. (From McCoy and Wynder, 1979.)

285

NUTRITION AND CANCER

CASES= 3EL CONTROLS= L 7 2 5

0

1-6

7-t

02 OF ALCOHOL PER DAY

FIG.5. Relative risk of oral cavity cancer for daily consumption of alcohol and cigarettes for males. (From McCoy and Wynder, 1979.)

smoke are the source of the initiating carcinogenic stimuli, and that ethanol facilitates the reactivity of some tobacco-associated initiator. As identified above, alcohol and tobacco are the two major risk factors for esophageal cancer in Western Europe and North America, but are of little importance in the esophageal belt of Central Asia. Epidemiologic studies in the Peoples Republic of China indicate a relation between esophageal cancer and wheat-eating, as opposed to rice-eating populations (The Coordinating Group for Research on Etiology of Esophageal Cancer in North China, 1975). A series of epidemiologic studies in the Caspian littoral of Iran indicate that the high risk for esophageal cancer is associated with the severely limited and probably irritant nature of the diet in conjunction with exposure to a carcinogenic agent derived either from opium tars or from wheat contaminants (Joint Iran IARC Study Group, 1977). Factors that have been associated with high risk include the consumption of wheat bread and a low intake of vitamin A, vitamin C, and riboflavin. Subsequent clinical studies have found signs, of only vitamin A and riboflavin deficiencies (McClaren and Siasi, 1978). A case-control study organized to follow up the results of the population investigation in Iran has shown no excess risk among patients for

286

BANDARU

s. REDDY et ul.

low intake of animal proteins, but has confirmed the risk attributed to a iow intake of fruit and raw vegetables and has shown that even among the generally poor rural communities of northern Iran, it is the lower social strata that are more severely affected by the disease (Cook-Mozaffari , 1979). A case-control study in Puerto Rico showed that the lower socioeconomic groups were most affected and that there was evidence that these lower socioeconomic groups had a diet deficient in good quality protein, total calories, vitamin A, and riboflavin (Martinez, 1969). Although the relationship between nutrient deficiency and esophageal cancer was not demonstrated on an individual basis, heavy consumption of alcohol and tobacco, together with hot beverages and spices were associated with the disease. High incidence of esophageal cancer found in Transkei, South Africa, has been linked to molybdenum deficiencies in the soil (Burrell et al., 1966), but no evidence has been produced that associates the possible deficiency with human disease (Rose, 1973; Warwick and Harrington, 1973). Recent studies by Fong et ul. (1977) in esophageal cancer patients in Hong Kong and in animal model studies show that zinc and copper are important in reducing the incidence of esophageal cancer. The possibility that nutrition might play a role in the etiology of esophageal cancer is also suggested by consideration of the association of Plummer-Vinson (Paterson- Kelly) syndrome and cancer of the upper alimentary tract. This disease, once prevalent among Swedish women, was shown to be associated with chronic iron and vitamin deficiencies. High rates of upper alimentary tract cancer were observed in the absence of exposure to tobacco or any other obvious source of carcinogen (Wynder and Fryer, 1958). Since the introduction of a national program of iron and B vitamins supplementation in Sweden in the early 1950s, a significant reduction in the number of cases of Plummer-Vinson syndrome with a subsequent reduction of upper alimentary tract cancer has occurred (Larsson et a l . , 1975). Thus, one way in which alcohol could increase the risk for cancer would be through the associated nutritional deficiencies commonly associated with alcoholism (Leevy et al., 1965). Since alcoholics often consume 900 or more calories a day from alcohol alone (deLint, 1975), it is not difficult to imagine that the rest of their dietary intake is insufficient for providing necessary nutrients. Cancers of the head and neck also seem to occur most commonly in those individuals who do not eat nutritionally balanced diets. (DHEW Publ. No. 74-124, 1974). Alcohol consumption can lead also to impaired absorption of nutrients and vitamins (Vitale and Coffey, 1971).

NUTRITION A N D CANCER

287

C. METABOLIC EPIDEMOLOCY A N D EXPERIMENTAL STUDIES Animal studies have provided evidence that suggests that nutrient deficiencies and dietary contamination may interact in esophageal carcinogenesis. Feeding N-nitrosodiethylamine to lipotrope-deficient rats increased esophageal carcinogenesis compared to those animals fed a control diet (Rogers et al., 1974). The tumors induced in these investigations were invasive squamous cell carcinomas morphologically similar to those in man. Feeding mice a riboflavin-deficient diet causes morphological alterations in skin and upper alimentary tract epithelium that are similar to those observed in patients suffering from Plummer-Vinson disease (Wynder and Klein, 1965). As the deficiency progresses, epithelial morphology progressively changes from atrophy to hyperkeratosis to, in several instances, hyperplasia. The experiments of Chan and Wynder (1970) have shown that, following initiation with benzo(a)pyrene and promotion with croton oil, riboflavin-deficient mice develop tumors more rapidly than control mice receiving a nutritionally adequate diet. In parallel studies, Chan et a / . (1972) have shown that basal levels of skin aryl carbon hydroxylase were slightly reduced in riboflavin-deficient mice. However, the skin activity of riboflavin-deficient animals was induced to a much greater extent following a single application of dimethylbenz(a)anthracene (DMBA). The work of Gerson and Meyer (1977) has shown that feeding rats diets deficient in zinc causes morphological changes of the buccal mucosa similar to the riboflavin-deficient condition. [Dietary zinc deficiency has been shown to increase the number of esophageal tumors and to decrease the latent period in rats exposed to methylbenzylnitrosamine (Fong et ( I / . , 1978).] Lower levels of zinc have been observed in hair and tissue samples from esophageal cancer patients (Lin er al., 1976). In animals made deficient for vitamin A, the tracheobronchial epithelium undergoes atrophic degenerative changes (Harris et ul., 1972; Salley and Bryson, 1957) that are quite similar to the changes observed in animals exposed to tobacco smoke (Dontenwill et al., 1973; Kobayashi el ul., 1974). Vitamin A-deficient animals have been shown to be more susceptible to polycyclic aromatic hydrocarbons or PAH carcinogenesis (Sporn et ( i / . , 1976). That ethanol is capable of decreasing Vitamin A levels is noteworthy in view of the participation of this vitamin in the regulation of epithelial cell differentiation (De Luca er al., 1969; Vaughn and Bernstein, 1976). Thus, currently available evidence suggests that nutritional deficiencies that arise from either undernutrition and/or as a direct consequence of

288

BANDARU

s. REDDY et af.

alcohol intake (impaired absorption or enhanced elimination) could play a role in the etiology of head and neck cancer. There are at least four possible models for correlating alcohol consumption and cancer: ( I ) alcohol as a solvent; (2) and (3) alcohol-induced increases or decreases in liver metabolism; or (4) alcohol-induced alterations in target tissue metabolism. 1. The simplest of the four possible models treats alcohol as a solvent. This model is based on the assumption that entry of tobacco-related carcinogens into target tissues is facilitated because of enhanced solubility and easier passage through cellular membranes of the carcinogen. Stenback (1969) demonstrated that administration of the carcinogen DMBA dissolved in ethanol resulted in a reduced latent period and increased skin tumor formation compared to mice treated with DMBA dissolved in acetone. Kuratsune et af. (1971) showed that although chronic treatment of mice or rats with various distilled beverages failed to cause tumor formation in either mice or rats, a tumor promoting activity for DMBA similar to that caused by croton oil was found in sake and its distillation residues. This model may apply to both oral and esophageal cancer, but certainly is not sufficient to explain the association of alcohol and cancer of the larynz, since this area does not come into direct contact with alcohol. 2. The extensive body of literature substantiating ethanol-induced alterations in liver metabolism requires that serious attention be given to the second model. Chronic ethanol consumption leads to, among other things, an enhancement in the liver microsomal drug metabolizing capabilities of both humans (Kater et al., 1969; Misra et al., 1971) and experimental animals (Misra et al., 1971; Rubin et al., 1968; Rubin and Lieber, 1968). In the absence of extensive destruction of liver tissue, increased production of a metabolite that is then delivered to the target tissues and further metabolized to its ultimate carcinogenic form could be envisioned. Radike et al. (1977) have shown that ethanol decreases the latent period for vinyl chloride carcinogenesis in rats, demonstrating a mechanism by which ethanol could increase the risk for the development of liver cancer. Recently, we have demonstrated that in vitro metabolism of the hepatocarcinogen N-nitrosopyrrolidine is increased in microsomal fractions isolated from ethanol-consuming animals and that postmitochondrial supernatants isolated from ethanol-consuming animals are capable of much greater conversion of N-nitrosopyrrolidine to a mutagen than control preparations (Fig. 6) (McCoy et al., 1979). Similar ethanol-associated increases in the in vitro metabolism of dimethylnitrosaminewere reported by Maling et af. (1975). This particular model must, of necessity, place

289

NUTRITION AND CANCER

mot

1

METHANOL MCONTROL

9 10

20

30

40

c

50

wMOLE /PLATE

bITROSOPYRROLIDINE] FIG.6. N'-Nitrosopyrrolidine mutagenicity of liver postmitochondrial supernatants from control and ethanol-consuming hamsters. Salmonella typhimurium TA 1535 was used as the indicator strain. Final protein concentration for both control and ethanol postmitochondrial supernatants varied from 1.3 to 1.5 mg/ml. Data expressed as mean It SE of six separate determinations. Data were corrected for spontaneous revertants, which ranged from 4 to 30 per plate. (From McCoy et a/., 1979.)

a restriction on the nature of the carcinogen in that the carcinogens and/ or their metabolites must have a high degree of site specificity, since all tissues in the body would be exposed to the results of changes in liver metabolism and yet only those in the head and neck area are at increased risk. The major objection to the involvement of enhanced liver metabolic activity in the causation of head and neck cancer is that this model excludes consideration of PAH participation. Though tobacco smoke contains nitrosamines that are site-specific for the upper alimentary tract (Boyland et al., 1964; Hoffmann et al., 1975) and upper respiratory tract (Hilfrich et ~ i l . ,1977; Lijinsky and Taylor, 1976) in animals, we feel that any model that fails to take into account the carcinogenic potential of PAH in tobacco smoke is a seriously compromised hypothesis. In addition, this model clearly cannot apply in the case of the severely damaged alcoholic liver. 3. An association between cirrhosis and the development of cancer of the liver in man has been observed (Lee, 1966; Leevy et af., 1964). In addition, Keller presented evidence that cancers of the head and neck

290

BANDARU

s. REDDY et al.

area are often found in cirrhotic patients (Keller, 1967, 1977). That this increased risk for liver cancer is due to components other than ethanol can be inferred from the work of Gibe1 et al. (1975). They demonstrated that several of the higher alcohols normally found in distilled beverages can be both hepatotoxic and carcinogenic. For example, they demonstrated that administration either orally or subcutaneously of isoamyl, isobutyl, or n-propyl alcohol to rats resulted in the appearance of both benign and malignant tumors of the liver, as well as other sites. Experimental evidence in favor of decreased liver metabolism resulting in increased risk for head and neck cancer was presented by Protzel et ul. (1964). In their studies, mice were treated with either ethanol or carbon tetrachloride and the buccal mucosa were swabbed with benzo(a)pyrene solutions; these mice developed more tumors with a shorter latency period than controls receiving only carcinogen treatment. Falk and Kotin (1963) demonstrated that the rate of bile clearance of benzo(a)pyrene was markedly decreased in rats whose livers were damaged by exposure to carbon tetrachloride. In patients suffering from cirrhotic liver disease, the rate of metabolic clearance of a number of drugs known to be metabolized by the microsomal drug metabolizing system was reduced (Chakraborty, 1978). The loss of metabolic capacity in cirrhotic liver would result in a decreased ability to detoxify tobacco-related carcinogens, thus causing a net increase in their systemic concentrations and resulting, in essence, in the exposure of target tissues to more carcinogen than would occur in individuals with uncompromised livers. 4. The fourth model is based on the hypothesis that chronic ethanol consumption alters intracellular metabolism of the epithelial cells at the target sites, resulting in enhanced metabolic activation of tobacco-associated carcinogens. Some of the more attractive features of this model are ( I ) that the cancer associated with the head and neck are epithelial in origin; (2) that these surface epithelial cells are mostly exposed to tobacco-associated carcinogens; and (3) that site specificity of the carcinogen need not be invoked since all sites, by virtue of their anatomical location, will be exposed to tobacco smoke. This third feature is an important point, because no class of carcinogen need be initially excluded from consideration since we cannot as yet say with any degree of certainty which of the many potential carcinogens in tobacco smoke are actually involved in the initiation of carcinogenesis.

D. CONCLUSIONS The epidemiologic data indicate a strong positive association of both chronic alcohol and tobacco consumption with cancers of the upper

NUTRITION A N D CANCER

29 1

respiratory and upper alimentary tract. One can hope that the increased national interest in both the prevention of alcoholism, as well as smoking cessation programs, will in and of themselves result in a decreased incidence of these cancers. Emphasis should be put on the education of the young as to the health consequences of chronic use of alcohol and tobacco. In view of the epidemiologic association of vitamin and mineral deficiencies with esophageal cancers, serious consideration must be given to the role that alcohol-related changes in nutritional status have in increasing the risk for cancer. It seems clear that much more experimental work is needed to translate nutrition as a risk factor into an understanding of the cause and effect relationship. Ethanol interacts with tissues in many ways (as a drug, energy source, or solvent) and can influence intracellular metabolism in many ways (enzyme induction, alteration in redox potential, as a metabolite). In addition, ethanol can cause marked changes in cellular and tissue metabolism because of alteration in hormonal status. It is hoped that evidence for the very strong association of nutritional deficiencies, as well as alcohol and tobacco usage, with cancers of the upper alimentary and upper respiratory tract will lead to more data and further insight into the role each of the known risk factors plays in the etiology of the disease.

V. Dietary Factors and Cancer of the Pancreas

A. EPIDEMIOLOGY Geographic differences in the incidence of pancreatic cancer are pronounced (Fig. 7; Segi et d . , 1972). The highest incidence of this cancer is among male Maoris of New Zealand and female natives of Hawaii. It is noteworthy that two groups of Polynesian descent, the New Zealand Maoris and native Hawaiians, are especially prone to this disease. According to Doll et a l. (1970), the lowest rates in both sexes are reported for Nigeria and for Bombay, India. Incidence data from the 1969 to 1971 Third National Cancer Survey (Biometry Branch, National Cancer Institute, 1974) reveal that pancreatic cancer predominates in blacks and in males. Factors underlying these racial differences are unclear, but an environmental influence is suggested by the much higher mortality in American Japanese than in native Japanese (Haenszel and Kurihara, 1968). Pancreatic cancer predominates in males by a ratio of 2 : 1 before age 50 and declines, although never reaching unity, after that age (Segi et al.,

292

BANDARU

s. REDDY et al. R4TL PER 100,000 POP.

R4TL PLR 100,000 POP.

D

-

FEMALE

MALE I

I0

D

-

U.S. NONWWTL

u s ,NoNwHIiC

SCOT L 4 N D

SWLDCN

SOUTM I f RlCA

DENM4RK

FlNL4ND

SCOTLIND

SWCDLN

FINLAND

C4N4DA

4USTRI4

U.S., WMlTC

U S ,WHITE

DENMIRY

lRLL4ND

IRLLAND

C4N4D4

r

-

1

-

IUSTRIA

NLTHCRL4NDS

CNGL4ND 6 W4LES

NCWZL4L4ND

NEW ZE4L4ND

SOUTNA$RlCA

-

1 - 1

4USTRALI4

ENGLANO~WUES

ISRIEL

ISR4CL

NETHCRLANDS

NORTH lRLL4ND

NORW4Y

SWITZLRL4ND

SWITZERLIND

AUSTRILIA

NORTH. IRELAND

OLRMAWV, $.I

GERMANY, F.R.

NORW4V

OLLGIUM

DELGIUY

FRANCE

FR4NCE

JAP4N

JAPAN

ITALY

I

a

&

lT4LY

FIG.7. Age-adjusted death rates for pancreatic cancer in different countries, 1966- 1967. (From Segi and Kurihara, 1972.)

1969; Wynder et al., 1973). The peak at ages 25-29 for U.S. whites was found to be related to an increased mortality for males. The data suggest that the different sex ratios for pancreatic cancer by country and age may be influenced by environmental factors. Studies on migrants have provided valuable information on the influence of environmental factors. A study by Smith (1956) on Japanese immigrants to the United States showed that the standardized mortality rates for pancreatic cancer was higher among Japanese Americans as compared with white Americans. Haenszel and Kurihara (1968) reported a higher mortality from pancreatic cancer for first generation Japanese migrants than for American-born Japanese. A similar effect of migration within the United States is suggested by a cancer mortality survey in Ohio; blacks born in the South have substantially higher rates than Ohioborn blacks (Mancuso, 1974). These findings support the hypothesis that environmental factors (most likely dietary in nature) increase the risk for pancreatic cancer.

NUTRITION A N D CANCER

293

Pancreatic cancer rates vary among religious groups. Newill (1961) and Seidman (1970) reported that pancreatic cancer was more frequent among Jews than among other religious groups. The rates among Seventh-Day Adventists are all in the vicinity of 50-75% of general rates (Phillips, 1975). In addition to abstinence from smoking and drinking, the most distinctive feature of the typical Adventist lifestyle is a unique diet: lactoovo-vegetarianism. Lyon et al. (1977) reported that in Utah the incidence and mortality for pancreatic cancer was 36 and 27% below expected, respectively. Mormon church doctrine prohibits the use of alcohol, tobacco, coffee, and tea.

B. ETIOLOGY Although the epidemiologic studies of pancreatic cancer point to an environmental influence, no environmental factor has been singled out with certainty (Fraumeni, 1975). Although this review is mainly concerned with nutrition and pancreatic cancer, a number of retrospective and prospective studies have shown an increased mortality rate from pancreatic cancer among cigarette smokers (Best, 1966; Weir and Dunn, 1970; Wynder et al., 1973). Cigarette smoking seems to account for the male predominance of pancreatic cancer, since no sex differential exists for this cancer among nonsmokers (Hammond, 1966). We can envision two ways in which tobacco components, or their metabolically activated forms, can reach the pancreas, namely, via reflux from the bile duct into the pancreatic duct or through the blood stream (Hansson, 1967; Wynder, 1975a). The migrant pattern of pancreatic cancer, the relatively high level of this disease in American Jews (although their cigarette consumption is relatively low as is well reflected in their lower rate of lung cancer), and the association of pancreatic cancer with fat consumption in various countries suggest that diet affects the development of pancreatic cancer (Wynder et al., 1973). Data on pancreatic cancer correlate with a general finding that overnutrition relates to a variety of cancers, including cancer of the colon, breast, and prcstate. In Japan, the climbing rates for pancreatic cancer have been linked to consumption of a Western diet (Wynder et d . , 1973).

C. METABOLIC EPIDEMIOLOGY As a working hypothesis on the etiology of pancreatic cancer, Wynder et a / . (1973) proposed that bile may contain carcinogens and/or cocarcin-

294

BANDARU

s. REDDY et al.

ogens and promoters (possibly originating in the diet, tobacco, and occupational environments) and that this bile, refluxed into the pancreatic duct, may cause pancreatic cancer. The effect of dietary fats is on the composition of the biliary bile acids, which have been shown to act as promoters (Reddy et al., 1978b). However, there are no metabolic, epidemiologic, and animal studies to test the relationship of nutrition to pancreatic cancer.

D. EXPERIMENTAL STUDIES Animal models have been developed to study pancreatic carcinogenesis. Pancreatic duct neoplasms, biologically and morphobiologically similar to those in man, were induced in Syrian golden hamsters after subcutaneous injection of N-nitrosobis(2-oxopropy1)amine (BOP) (Pour et al., 1976). Tumors also developed in the lungs, liver, gallbladder, and kidney. However, oral administration of BOP in drinking water resulted in a few pancreatic neoplasms and in a high incidence of intra- and extrahepatic bile duct neoplasms (Pour and Althoff, 1977). The significance of a bile reflux mechanism in pancreatic carcinogenesis was investigated by Pour and Donnelly ( 1978) using cholecystoduodenostomized and choledochostomized Syrian golden hamsters. The distribution and patterns of BOP-induced pancreatic neoplasms were not altered by bypassing the bile through the common duct. However, in this study, the possible regurgitation of bile-borne carcinogen from the duodenum into the pancreatic duct is most unlikely because in this species the pancreatic duct enters the common duct well before it opens into the duodenum.

E.

CONCLUSIONS

The comparison of the incidence of international high-risk groups, as well as intracountry variations and migrant studies, have yielded significant information on risk factors. Cancer of the pancreas, whose relation to dietary factors is less clear than that of cancers of the colon, breast, prostate, and stomach, is worth pursuing from the standpoint of metabolic epidemiology and animal studies. We want to be able to replace our tentative guesses about the risk factors with enough understanding to lower current levels of risk.

NUTRITION A N D CANCER

295

VI. Dietary Factors and Cancer of the Breast

A. EPIDEMIOLOGY Traditionally, epidemiologic studies have generated hypotheses for the etiology of breast cancer through international comparisons of incidence, case-control studies, migration studies, and correlation of selected variables with disease occurrence. Such studies have provided the basis for generalizations concerning the impact of environment and nutrition, in particular, on breast cancer incidence. 1, International Variation

Risk of breast cancer has been shown to vary according to geographical area; high incidence rates are found in North America and Western Europe, low rates in Asia, particularly Japan, and intermediate rates prevail in Finland, Southern Europe, and South America (Fig. 8). These differences suggest that environmental factors relate to this disease. It is noteworthy that in Western countries, breast cancer incidence continues to rise with increasing age after menopause, whereas it remains unchanged over a wide age span in postmenopausal Japanese. Recently Moolgavkar ct al. (1979) reported that incidence and mortality curves for postmenopausal Japanese and Western women increased with age, but the degree of increase was less in the former group than in the latter. These observations suggest that the action of the factors involved in breast cancer is more pronounced in the postmenopausal age group in high-risk Western populations. 2 . Migrant Studies

The strongest evidence for environmental factors in the etiology of breast cancer is found in the results of migrant studies. Within two to three generations, Japanese migrants to the U.S. experience an increase in cancer incidence rates from those common in Japan to those prevalent in the United States (Fig. 9). In fact, Buell (1973) reported that during the years 1969- 1971, the incidence of breast cancer in Japanese-American women had risen to one-half that of white American women and was five times that of age-matched native Japanese women. Staszewski and Haenszel (1965) showed that breast cancer incidence among Polish immigrants to the United States exceeded both urban and rural incidence rates in Poland. Of interest is the short time period in which these rates changed in comparison to those of Japanese immigrants. Possibly, Polish

296

BANDARU

s. REDDY et al.

34(

3a

m a‘

2 25( 0

NETHERLANDS

0

0-

0

a 2a w

n

w

ta

p””

I

t-

a

/

w

0

d

lo(

30. 35- 40- 45- 50. 34

39

44

49

54

5559

60. 65- 70- 15- 8064

69

14

79

04

A G E GROUPS

FIG.8. Female breast cancer death rates by age in four countries, 1966-1967. (From Segi and Kurihara, 1972.)

migrants assimilate more quickly and, therefore, adapt more rapidly than the Japanese. Although a number of reproductive variables have been associated with differences in rates of breast cancer (Hems, 1970; MacMahon et al., 1973; Hirayama, 1978), not one or any combination of variables can explain the changes in rates seen in migrant populations. Alterations in dietary practices, however, appear to be the environmental factor that best accounts for the increase in risk associated with migration from a low-risk to a high-risk country (Brit. Med. J . Editorial, 1974).

297

NUTRITION AND CANCER

3 . Correlariori Analyses Correlation studies have provided another source of evidence for nutrition's role in the etiology of breast cancer. A positive correlation between breast cancer mortality and daily per capita consumption of fat has been demonstrated by a number of researchers (Lea, 1967; Drasar and Irving, 1973; Carroll, 1975; Armstrong and Doll, 1975) (Fig. 10). Further, Carroll and Khor (1975) and a supporting study by Hems (1978) have shown that in various countries, a strong positive correlation exists between animal fat consumption and breast cancer incidence, but a similar association could not be demonstrated for vegetable fat. In an intracountry investigation, Hirayama ( 1978) correlated breast cancer incidence in 12 different districts of Japan with specific food consumption patterns. Of the food items studied, the highest positive correlation was found for pork, followed by total animal fat intake.

I

6t

70

6773

50

{JAPAN

HAWAII JAPANESE

,

,'60:64

k8-'72

,

,HAWAII, CAUCASIAN

FIG. 9. Age-adjusted breast cancer incidence rates in native Japanese (striped bars), Hawaiian Japanese (solid bars), and white Hawaiians (hatched bars). (From Wynder and Hirayama, 1978.)

298

BANDARU

s. REDDY et al. NETHERLANDS

kK' n

CANADA. SWITZERLAN~ IRELAND. BELGIUM.

OS. AFRICA

0

0

9 0 20-

ISRAEL.

0

-.

AUSTRla

5

ITALY

I

15-

HUNGARY. 'PORTUGAL

0 w 4

;10-

p : VENEZUELA PAN AM^

ln w

a

L

2 4

a .

HONG KONG

OFINLAND

eoLAND

&.ROMANIA GAR~A

~ A I N

*YUGOSLAVIA

gOLUMBIA

5-

PHlLlPPlNSS .MEXICO JAPAN.

n 0

.GERMANY

NORWAY .FRANCE

.CZECHOSLOVAKIA

0

a

E

"2

NEW Z E A L A N D

AUSTRALIAO.SWEDEN

w

$

DENMARK 0.

THAILANP I

'TAIWAN

CEYLON. E L SALVADOR I

1

I

I

1

1

FIG. 10. Correlation between age-adjusted death rates from female breast cancer and per capita consumption of fat. (From Carroll and Khor, 1975.)

4. 1Ynie- Trend A m lysis Analysis of U . S . food consumption patterns over time provide support for an association between fat disappearance and breast cancer. From 1909 to 1972, the estimated consumption of dietary fat increased from 125 to 156 grams per day per capita (Gortner, 1975). During the same period, breast cancer mortality rates among U.S. whites showed a small but gradual increase (Seidman et al., 1976). Over the past 40 years, American blacks have experienced a marked increase in breast cancer incidence, from 50/100,000 in 1935 to approximately 65/100,000 in 1970 (Cutler and Young, 1975; Seidman rt ul., 1976). During this time, large scale black migration from the rural South to the urban North occurred. Accompanying this migration were marked changes in socioeconomic status and lifestyle. Shifting patterns in breast cancer incidence over time in low-risk countries undergoing Westernization also support an association between high fat intake and increased breast cancer incidence. Hems (1978) found that the temporal changes in breast cancer rates for 20 countries were significantly correlated with total fat and animal protein intake and not obesity,

I

299

NUTRITION A N D CANCER

/

a

JAPAN

1935

1940

1945

1950 1955 1960 YEAR OF DIAGNOSIS

1965

1970

1975

FIG.1 1 . Changing breast cancer incidence rates in Connecticut (USA) and Japan, 19501975. (From Connecticut State Department of Health, 1977, and Hirayama, 1978.)

reinforcing the idea that dietary fat per se may serve as a risk factor in breast carci noge nesi s. During the past decade, breast cancer mortality and morbidity rates have sharply increased in Japan (Hirayama, 1978) (Fig. I I). This increase has been associated with a marked shift toward a more Western lifestyle, one in which intake of dietary fat has substantially increased (Fig. 12). For example, in the period 1957 to 1973, the estimated per capita consumption of fat increased from 23 to 52 grams per day. This increasing consumption of fat parallels the increase in the number of annual deaths from breast cancer (1572 in 1955 and 3262 in 1975). ( 1949.1 00 I

-

MILK MILK PRODUCTS MEAT EGG OIL FRUIT

10.00

m 0

e 0 I-

FISH VEGETABLES RICE

:1.00 W

-1 4

a W

POTATOES

W

X 4 t-

-5

0.10

SWEET POTATOES

0

4 t-

a

0.014

1950

:

:

1960 YEARS

:

! 1970

FIG.12. Change in intake of selected foods in Japan, 1949-1973. (From Hirayama, 1978.)

300

BANDARU

s. REDDY et al.

Although indicative of important trends, correlations involving worldwide fat consumption data must be viewed with caution. Fat consumption data, as compiled by the Food and Agricultural Organization of the United Nations (1971). are only rough approximations of the disappearance of dietary fat in a given population and not the actual per capita consumption. Moreover, they generally tend to correlate contemporary incidence or mortality data with recent food disappearance data, although migration studies suggest that an interval of 30+ years may be more appropriate. Despite these caveats, the correlation between breast cancer mortality and high fat intake is highly consistent both within and between country comparisons. 5 . Special Populations

Studies of populations that share a common gene pool with the general population but differ significantly in dietary practices, e.g., Seventh-Day Adventists, have indicated that the incidence of breast cancer in this group that eats little or no meat is lower than that found in the general population (Phillips, 1975). Although the difference was not statistically significant, the lower rates are real and this further implicates animal fat ingestion in the etiology of this disease. Moreover, Israeli Jews of Asian origin exhibit lower breast cancer incidence rates in comparison to the Israeli population of European origin (27/100,000 vs 124/100,000)(Gross et al., 1977).

6 . Case-Control Studies The value of case-control studies in the investigation of diet and disease is limited because of the difficulties inherent in obtaining accurate historical dietary information. Morgan et al. (1978) attempted to refine the methodology of such studies by utilizing three dietary assessment techniques, i.e., 24-hour recall, a four-day diet diary, and a detailed diet questionnaire. Major discrepancies were found among these methods, and it was suggested that the most reliable source of dietary information was obtained by the detailed dietary questionnaire. Recognizing such difficulties, Miller (1977) studied the nutritiodbreast cancer relationship in a carefully controlled retrospective study. He demonstrated that saturated fat intake was associated with breast cancer incidence in both preand postmenopausal women. In a recent large-scale prospective study of Hawaiian Japanese, dietary practices of the spouses of women with breast cancer were compared with those of the spouses of women who did not have breast cancer

NUTRITION A N D CANCER

301

(Nomura cf ul., 1975). Again, it was concluded that consumption of a high-fat Western diet contributed to increased breast cancer risk in these migrant Hawaiian Japanese women. The consequence of the findings from these two distinct methods of investigation lends credence to the relationships found. In a country such as the United States, however, dietary patterns are only partially dictated by custom, and thus accurate dietary histories are more difficult to collect. Some reported inconsistencies in results apparently relate to differences in the selection of the study population and/or the control group. Whole population comparisons, of course, are possible particularly when they reflect widespread differences among populations with regard to food traditions and food availability. We are in agreement with Miller (1977), Hems (1978), Gray ef ul. (1979), and Hirayama (1978) when we stress that variables reported to affect the risk of breast cancer (onset of menarche, age at menopause, age at first pregnancy, menstrual patterns, family history, fertility, height, or weight) do so at relatively low magnitudes, when compared to the major differences in risk between women living in high- and low-risk countries. We conclude that case-control studies, of the type conducted thus far are not likely to shed much additional light on the etiology of breast cancer. To summarize epidemiological findings: 1. Inter- and intranational studies, correlation studies, time-trend analyses, and analyses of special populations support the suggestion that an ubiquitous environmental factor may be significantly associated with breast cancer risk. Major dietary factors, particularly the intake of fat, affecting an entire population may represent that basis for the differences in breast cancer rates in various populations. 2. A strong argument for an environmental etiology of breast cancer comes from analysis of population groups migrating from countries of low incidence to countries of high incidence. Characteristically, these groups experience an increase in the incidence of breast cancer in successive generations, eventually approaching the incidence rates of the host population. 3. The concept that dietary fat intake, specifically, is an important determinant of breast cancer is supported by the following observations: (a) a significant positive correlation exists between estimates of dietary fat consumption (both saturated and unsaturated fats) and international mortality rates of breast cancer; (b) results of time-trend analysis in populations undergoing Westernization suggest a positive correlation between breast cancer incidence and consumption of dietary fat; and (c) analysis of special population groups suggest an association between

302

BANDARU

s. REDDY et al.

breast cancer incidence and dietary fat intake. The combined evidence from epidemiological studies, therefore, suggest that dietary fat is an important contributory factor in breast cancer. 4. Since fat and protein consumption parrallel one another in most human populations, on the basis of present information a role for dietary protein intake in breast carcinogenesis cannot be excluded.

B. EXPERIMENTAL STUDIES Through the efforts of Tannenbaum (1942), the field of nutritional carcinogenesis experienced a burst of creative activity in the early 1940s. In 1967, interest in this field was regenerated with the publications by Carroll and co-workers of nutritional studies using the 7, IZdimethylbenz( a)anthracene mammary tumor model (Carroll, 1975). Today there exists a great deal of knowledge upon which to develop and test mechanistic hypotheses of the relationship between nutrition and breast cancer. Tannenbaum’s early work (1957) brought to light four major facts concerning the relationship of dietary fat to breast cancer, facts which are of compelling interest today: (1) the quantity of dietary fat, holding carbohydrate and protein constant, influenced the subsequent development of mammary cancer; (2) the time of initiation of the diet influenced the course of tumor development; (3) the reproductive history of the animals was an important experimental variable; (4) caloric restriction has a profound retarding effect on the genesis of mammary tumors, as well as tumors at other organ sites. In 1947, he first demonstrated the initiating effect of dietary fat-spontaneous breast tumor incidence rates in female DBA mice were higher in those fed a high-fat diet than in those fed a low-fat diet. Further, tumor incidence was found to be greater in mice when the high-fat diet was initiated at 24 weeks of age than at 38 weeks. In the ensuing three decades, Tannenbaum’s findings have been confirmed by a variety of investigations in both rats and mice. Despite differences in the type or quantity of fat, the nature of the carcinogenic event (spontaneous or chemical), or the strain of animal used, one point stands out clearly: high intake of dietary fat increases the incidence of mammary cancer in rodents (Fig. 13). 1 . O h s i t y mnd Caloric Restriction

Experimentally, obese mice tend to have a higher tumor incidence than normal animals (Tannenbaum and Silverstone, 1957). However, mice

NUTRITION AND CANCER

303

DAYS AFTER MNU

FIG. 13. Effect of high- or low-fat diet on methylnitrosourea(MNU)-induced breast cancer incidence in female F344 rats. MNU was administered intravenously at 55 days of age and the rat was then fed high-fat (20% lard) or low-fat (5% lard) diet. The median latent period (time when 50% of tumor-bearing rats had developed tumors) was 83 days in the high-fat group and 103 days in the low-fat group. The tumor incidence differed significantly from the eightieth day on ( P< 0.01). (From Chan et a / . , 1977.)

kept on isocaloric high- and low-fat diets exhibit greater tumor incidence under high-fat conditions, suggesting that the fat effect is separate from the effect of high caloric intake (Tannenbaum, 1942). It has also been reported that rats fed a high-fat diet ad libitum have about the same rate of weight gain as rats fed a low-fat diet, but exhibit more tumors with a shorter latency period than animals maintained on a low-fat diet (Chan and Cohen, 1974). This experiment also indicates that the effect of highfat diets is independent of changes in body weight and caloric intake. As originally demonstrated by Tannenbaum, caloric restriction inhibits the development of breast cancer in rodents. Because the effects of changing the fat content of the diet were confined to skin and mammary tumors, whereas caloric restriction per se inhibited other tumors as well, Tannenbaum and Silverstone ( 1957) concluded that such effects were related to the specific action of fat rather than to the general influence of caloric restriction. 2 . Dktury Fat us u Tumor-Promoting Agent Since the spontaneous mouse mammary tumor model was used in Tannenbaum studies, it is impossible to discern whether the effect of

304

BANDARU

s. REDDY et a / .

nutrition was on the genesis of the tumor or on its subsequent growth and development: a key point from the standpoint of both etiological and preventive considerations. This question was partially answered by Carroll and Khor (1979, who showed that, in the DMBA model, the high fat effect could be demonstrated only when the diet was administered after initiation with DMBA. When a high fat diet was given before initiation and then replaced by a low fat diet, no enhancement was seen. Hence it appears that the dietary effect is exerted at the promotion, but not the initiation, phase of breast carcinogenesis. 3 . Effect of Carcinogen Dose Although this area has not been exhaustively studied, available evidence suggests that as the carcinogen dose increases, the promoting effect of dietary fat becomes less and less pronounced: a finding of considerable importance with regard to the human disease. The fact that ubiquitous environmental carcinogens are present at very low concentrations suggests that promoting factors may have a preponderant influence on the eventual outcome of the neoplastic process in humans. 4. Quantity und Quality of Dietary Fut

Using the DMBA model, Carroll and Khor (1971, 1975) also studied the effect of a variety of fats and oils on breast cancer development. When tumor incidence was the end point measured, high-fat diets promoted tumor development in most but not all cases. However, when total tumor yield was the end point measured, i.e., the total number of tumors per animal, consistent differences emerged. Animals fed fats that were deficient in, or contained low levels of, the essential fatty acid, linoleate, such as coconut oil and tallow, exhibited significantly lower total tumor yields than animals fed fats rich in linoleate, such as corn or sunflower seed oil. These results were confirmed under somewhat different experimental conditions by Nishizuka ( 1978) and Hopkins et al. (1976). The reason for the discrepancy between results obtained from tumor incidence measurements and tumor multiplicity measurements is unclear at present. Part of the problem may lie in the fact that the conditions under which these experiments were conducted differed particularly with regard to key intervening variables, such as carcinogen dose, animal strain, and diet composition. Another aspect of the high-fat effect on breast carcinogenesis is the finding in both the DMBA model and the transplantable mouse mammary tumor, that the tumorigenic response to fat intake is not linear in nature. When corn oil was the source of dietary fat, a threshold was reached

NUTRITION AND CANCER

305

beyond which increases in fat intake had no further influence on tumor incidence. When the corn oil content of the diet was increased from 5 to 20%, a significant difference in tumor incidence occurred. However, increasing dietary fat above 20% or decreasing it below 5% had no discernable effect on tumor incidence. Apparently, in the DMBA model, physiological responses are triggered only when fat comprises 8- 12% of the diet. The relationship between dietary linoleic acid intake and tumor yield is also not linear. Animals fed high fat diets containing lard or olive oil, containing 1 1 and 7% linoleate, respectively, develop almost as many tumors as those fed high-fat diets rich in linoleate, such as corn oil and sunflower seed oil (56 and 75% linoleate, respectively). Dayton et al. ( 1977) reported similar results. Using DMBA-treated animals fed high-fat diets of a wild-type and mutant safflower oil (70 and 12% linoleic acid, respectively), no differences in tumor incidence on total tumor yields were found. In a recent study by Hopkins and Carroll (1979), mixtures of linoleic acid-rich and linoleic acid-poor fats were added at 20% to the diet. It was found that DMBA-treated animals exhibited comparable tumor yields, provided greater than 2% linoleate was present in the diet. In addition, when carcinogen-treated animals were fed low-fat diets (3%) containing comparable amounts of linoleic acid as high-fat diets (20%), total tumor yields were significantly decreased in the low-fat group suggesting a certain amount of linoleic acid as well as a high’fat diet are required for optimal tumor growth enhancement. To summarize, five major themes emerge from this survey of experimental studies on dietary fat and breast carcinogenesis. 1 . Dietary fat exerts its effects on the promotional phase of breast carcinogenesis. 2. Promoting agents, such as excess dietary fat, play a more significant role in determining the eventual outcome of the neoplastic process when the initiating carcinogen dose is low, than when it is high. 3. The total quantity of dietary fat, irrespective of qualitative factors, is a central factor in the high-fat effect. 4. A certain threshold quantity of essential fatty acids is a necessary requirement for full expression of the high-fat effect. 5. The tumorigenic response to dietary fat consumption is saltatary, or discontinuous, rather than linear, or graduated, in nature.

5 . The Fat Eflect: Mechanism of Action Detailed reviews of the possible mechanism(s) by which dietary fat may exert its effect on breast cancer have been published by Hopkins

306

BANDARU

s. REDDY et al.

and West (1976), Hankin and Rawlings (1978), Alcantara and Speckmann (1976), and Miller (1977). A variety of mechanistic hypotheses have emerged from these reviews. They can be broken down into two broad categories: (a) direct effects at the level of the mammary gland, and (b) indirect effects mediated by host systems remote from the mammary gland. a . Direct Eflecfs. These models are based on the physical and chemical properties of fat, the formation of lipid peroxides, alterations in membrane structure and/or function, and enhanced prostaglandin synthesis. The role of adipose tissue in breast carcinogenesis (either acting as a reservoir or depot for carcinogens) has been a regularly reappearing theme over the past 30 years. Since adipose tissue rapidly takes up and slowly releases lipid-soluble agents, it has been proposed that the adipose tissue surrounding ductal cells may contribute to the susceptibility of tissue cells to cancer by prolonged exposure to lipid-soluble carcinogens of endogenous or exogenous origins (Beer and Billingham, 1978; Tannenbaum and Silverstone, 1957). Years ago, Dao et al. (1959) suggested that mammary adipose tissue functioned as a storage depot for hydrocarbons. However, this model appears unlikely since tumor incidence is not affected by a high-fat or low-fat diet given prior to treatment with DMBA and then placed on a high-fat diet (Carroll and Khor, 1971). Gamma1 ef al. (1968) demonstrated that uptake and clearance of radioactive DMBA in the mammary fat pad was the same, regardless of fat content in the diet. Since polyunsaturated fatty acids (PUFA) are converted by free radical reactions to lipid peroxides, a model involving breast cancer and lipid peroxidation has been advanced. Lipid peroxidation has been associated with a variety of pathological processes (Dormandy , 1978), including mutagenesis (Mukai and Goldstein, 1976) and carcinogenesis (Shamberger et al., 1974). It is possible that increased peroxidation of membrane lipids results in alterations in the function of transformed mammary cell membranes which, in turn, permit increased rates of growth (Hopkins and West, 1976). However, available evidence suggests that lipid peroxidation and free radical processes accompanying it are primarily' associated with the activation of procarcinogens (Bartsch ef al., 1976; Floyd et a / . , 1978) and not with the promotion of tumor development. Little evidence is available to suggest a role for peroxidation processes in the post transformation events, where dietary fat exerts its stimulatory effect. Rao and Abraham (1975) reported an unstable lipid composition in mouse mammary adenocarcinoma when the fat intake of the host was

NUTRITION A N D CANCER

307

altered. Within a relatively short period (8 weeks), changes in dietary fat were reflected in altered tumor lipid profiles. Similar findings were reported by Gamma1 er al. (1967) and Nishizuka er al. (1978). Tumor tissue tends to have greater proportions of PUFA than the normal gland (Rao and Abraham, 1975); this observation suggests that tumor growth is related to arachidonic availability through either prostaglandin (PG) synthesis or unknown membrane functions. Direct changes in the lipid composition of cell membranes induced by diet could have far-reaching structural and functional effects on membrane permeability and the activity of membrane-bound enzyme systems. The relevance of these membrane changes t o the stimulatory effect of a high-fat diet remains to be established however. Prostaglandins apparently act as intracellular regulatory molecules that, together with the cyclic adenosine monophosphate (CAMP) systems, govern the response of cells to hormonal stimulation. In human breast cancer, the capacity of tumor cells to produce prostaglandins increases from normal cells, to benign lesions, to primary tumors, to metastatic cancers. In addition, in animals, DMBA-initiated tumors contain more prostaglandins and a greater capacity to synthesize prostaglandins than normal rat mammary glands (Tan er al., 1974). However, daily injections of 1.5 mg/kg PGF% have reduced the growth of DMBA-induced mammary tumors over a 10-day period (Jacobson, 1974). h. Zndirecr Effects. Indirect mechanisms are those in which dietary fat secondarily stimulates mammary tumor growth by modifying the physiology of the host. Some evidence exists that dietary fat intake alters the function of at least three major systems that regulate the internal environment: ( 1) immune rejection responses; (2) mixed-function oxidases (MFO); and (3) endocrine control systems. These mechanisms are not necessarily mutually exclusive and may interact with each other. The experimental evidence is as follows: Mice fed a diet containing PUFA are more susceptible to the development of tumors following inoculation with cells from a transplantable tumor than mice fed a diet containing saturated fatty acids (Hopkins and West, 1976). PGE, and PGE2, which are synthesized from PUFA, are present, along with their precursor molecules, in greater amounts in the DMBA tumor than in normal mammary tissue (Tan er al., 1974); PGE, and PGE, (Plescia ef al., 1975) arachidonate and linoleate (Mertin er a/., 1973) inhibit cell-mediated immune responses in the lymphocyte test system. On the basis of this evidence, it has been proposed that increased amounts of dietary PUFA may suppress the cell-mediated immune rejection mechanism, either directly or after conversion to prostaglandins and, thereby, enhance the

308

BANDARU

s. REDDY et af.

capacity of transformed cells to proliferate (Hopkins and West, 1976). Quantitative analyses of the immunosuppressive influence of a high-fat diet are needed to confirm this hypothesis. Evidence suggests that nutritional factors can markedly influence the activity of microsomal MFO (Campbell, 1977). It is well established that mixed function oxidase systems are key in the biotransformation of chemical carcinogens (Weisburger and Williams, 1975) and steroids (Gustafsson et al., 1975). Steroids such as androgens and estrogens have been implicated in breast carcinogenesis, and alterations in steroid metabolizing enzyme systems such as MFO by diet could, in turn, influence breast cancer development. However, although the biotransformation of carcinogens to their active metabolites is a major role of MFO, it is unlikely that this aspect of MFO function is involved in the fat effect, since the influence of fat is exerted days or weeks after carcinogen administration. The work of Furth (1973), Pearson (1972), and Meites (1977) demonstrated clearly that although a variety of hormones are involved in mammary tumor growth, prolactin is the predominant factor in rodent breast cancer development. In addition, Smith et al. (1977) and Costlow and McGuire ( 1977) demonstrated unequivocally that the DMBA induced tumor and the R3230AC mammary tumor possess large quantities of specific prolactin binding receptors; and Kelly er a / .( 1978) showed that prolactin binding was highest in those tumors that showed the greatest growth response to prolactin injection. Meites (1977) showed in the DMBA model that experimental procedures that elevate serum prolactin (median eminence lesions, treatment with reserpine, or transplantation of secondary pituitaries under the renal capsule) increase tumor yield. It should be noted that tumor yield is enhanced only if the prolactin elevating procedure is introduced after, not before, DMBA administration-a response strikingly reminiscent of the fat effect on breast carcinogenesis. The discovery of hypothalamic neurotransmitters capable of regulating prolactin secretion by the anterior pituitary led to the concept that circulating prolactin concentrations may be regulated by environmental factors acting on the central nervous system (CNS). Meites (1977) found that administration of drugs that act by way of the CNS could elevate or depress serum prolactin levels and thereby promote or inhibit mammary tumor growth. In this regard, dietary fat may act via the CNS in a manner similar to that of reserpine, perphenazine, or thyrotropin releasing hormone, namely, by causing metabolic alterations (in the synthesis or action of biogenic amines) that result in the induction of prolactin secretion by the pituitary and ultimately in promotion of mammary tumor development. Early experimental evidence suggesting a CNS mechanism for the

NUTRITION AND CANCER

309

fat effect comes from the studies of Dunning et al. (1949). These investigators found that tumor incidence was elevated in rats fed a high-fat diet (40% Crisco) and that the pituitary glands were significantly heavier in high-fat than in low-fat animals. In addition, histological examination of the mammary glands of high-fat animals revealed signs of intense secretory activity. Similar results were reported by Benson et ul. (1956) and Cutler and Schneider (1974). Though unknown or ignored by most biomedical researchers, there is considerable literature from comparative physiology suggesting an important role for prolactin, and the regulation of fat metabolism has no doubt been obscured by the emphasis (exemplified by the very name of the hormone) placed on the relationship between prolactin and lactation. However, it is clear that prolactin and growth hormone share amino acid sequence homologies and probably evolved from a common ancestral protein (Niall et ul., 1971; Nicoll, 1975). Since growth hormone is intimately involved in the regulation of energy fuels such as fat (Fineberg et ul., 1972; Fraser and Blackard, 1977), it is not surprising that prolactin also plays a role in fat metabolism. Studies concerning the liporegulatory role of prolactin have been conducted principally by Meier (1977) in amphibians, birds, and lower mammals. It is clear from these works that mobilization and deposition of fat is regulated by diurnal (CNS-mediated) variations in serum prolactin concentrations and, furthermore, that these changes are temporarily associated with changes in serum corticosteroid concentrations. Insight into the mechanism by which prolactin regulates lipid mobilization and deposition is provided by the work of Scow and associates on the enzyme lipoprotein lipase (Nurririon Review editorial, 1975; Zinder et ul., 1974). Scow found that the activity of this enzyme in the rat was elevated during lactation specifically by prolactin. Since lipoprotein lipase activity represents the major route by which triglyceride fatty acids are cleared from the circulation, induction of this enzyme by prolactin substantiates a liporegulatory function for this hormone. Based on the foregoing considerations, Chan and Cohen (1974) designed a series of experiments to assess the relationship between high-fat intake, prolactin secretion, and mammary carcinogenesis. Drugs that antagonize estrogen action and block prolactin secretion retarded tumor development as expected. Only the antiprolactin drug abolished the differential in tumor incidence characteristic of animals fed high- and lowfat diets (Table XII). Though indirect, the results suggested that the fat effect was mediated by prolactin. It was also found that rats fed a high-fat diet exhibited significantly higher serum prolactin levels than rats fed a low-fat diet (Chan et a / . ,

3 10

BANDARU

s.

REDDY

et al.

TABLE XI1 EFFECTOF HORMONE ANIAGONISIS ON MAMMARY TUMORINCIDENCE I N DMBATREATEDRATSFED HIGH-FATA N D LOW-FATDIETS" Palpable tumorbearing rats per total number of rats

Control Control Antiestrogen' Antiestrogen Antiprolactid Antiprolactin (I

Total number of Adenocarcinomas

Diet*

Number

Percent

Palpable

HF LF HF LF HF LF

18/22 lOil9 7/18 1/13 9/21 8/22

82 52 39 8 42 36

39 15 7 I 15

12

Palpable Nonpalpable tumors per rat 12 7 0 2 9 7

I .77 0.78 0.38 0.08 0.70 0.54

From Chan and Cohen (1974).

* Hf diet, 20% corn oil; LF diet, 5% corn oil. ' Nafoxidine hydrochloride ( I mgikg body weight) administered sc three times. 2-Bromo-a-ergocryptine (3 mg/kg body weight) administered sc three times.

1975). Moreover, this increase was seen during the proestrus-estrus surge of the estrous cycle; no difference in prolactin levels occurred during the metestrus-diestrus stages, suggesting an influence of fat on the hypothalamic centers controlling circadian rhythms of prolactin secretion. It is also possible that dietary fat acts either at the ovary or at higher centers controlling ovarian steroidogenesis, thereby stimulating prolactin secretion and breast tumor development. This hypothesis is contradicted by evidence from experiments in which DMBA-treated rats were ovariectomized one month after initiation and then placed on high-fat or lowfat diets (Chan r t c d . , 1977). Ovariectomized animals fed a high-fat diet exhibited a marked increase in both tumor incidence and serum prolactin levels when compared to animals fed a low-fat diet (Fig. 14). On the other hand, these results do not indicate that a role for estrogens in the fat effect is completely eliminated. Estrogens synthesized by peripheral aromatization of androgens could influence prolactin secretion patterns and, thereby, indirectly influence mammary tumor development (Nufure (London) Editorial, 1975). Of importance from the standpoint of human breast cancer, this experiment may provide insight into dietary factors operating in postmenopausal women and deserves further detailed study. In a general sense, these results reinforce the idea that the enhancing effect of dietary fat on breast cancer development is not mediated by an

31 1

NUTRITION AND CANCER

ovarian mechanism either at the level of the ovary or at higher centers controlling the secretion of estrogens by the ovary. Direct proof that dietary fat alters circulating prolactin levels, but not estrogen levels, was obtained by simultaneous measurement of prolactin and estrogen levels in the serum of methylnitrosourea (MNU)-treated. animals 20 weeks after carcinogen treatment. The high-fat groups exhibited: (a) elevated prolactin levels; (b) unchanged total estrogens; (c) elevated prolactin to estrogen ratios at both proestrus-estrous, and metestrous-diestrous (though significance could be demonstrated only in the latter phases due to the small number of proestrus-estrous samples available) (Table XIII). These results are in keeping with a small but growing body of knowledge that suggests that the development of both normal rodent mammary gland and mammary tumors are regulated by the relative proportions of circulating prolactin and estrogen (Welsch and Nagasawa, 1977; Wuttke ef ul., 1976). Boyns ef nl. (1973) and Hawkins ef al. (1976) found that in three strains of rats with different genetically determined susceptibilities t o the carcinogenic action of DMBA, tumor yield was directly proportional to plasma

60

e

I6

24

32

40

WEEKS AFTER OVARIECTOMY

FIG.14. Effect of high-fat or low-fat diet on mammary tumor incidence in ovariectomized rats. Curves represent cumulative tumor incidence. DMBA ( 5 mg/kg) was administered on day 50 of age. Ovariectomy was performed on day 130 of age and the animals were then placed on either high-fat (20% lard) or low-fat (5% lard) diet. Experiment was terminated 9 months after ovariectomy. Numbers in parentheses represent serum prolactin concentrations (mean SEM). (From Chan et a / . , 1977.)

*

3 12

BANDARU

s. REDDY et al.

TABLE XI11 INFLUENCE OF DIETARY FATON SERUMHORMONAL PROFILES A N D MAMMARY TUMOR INCIDENCE I N METHYLNITROSOUREA-TREATED F344 RATS"

Diet

Tumor incidence (percent)

High fat (20% lard) Low fat (5% lard)

90 40

Estrous stage

Prolactin (ndml)

Total estrogenb (ng/100 ml)

P-E" M-D P- E M-D

237 ? 98* 100 29 140 f 79 38 3

23.1 22.2 23.2 18.9

* *

1.7

* 2.0 * 2.2 1.1

Prolactin and estrogen ratio 11.1 4.5 6.4 2.0

* 4.8

* 0.9 * 3.9 2

0.1

" From Chan el a / . (1977).

* Estradiol + estrone.

P-E, Proestrus-Estrus; M-D, metestrus-diestrus.

* Mean * SEM.

prolactin levels, whereas estrogen levels were unchanged-again suggesting that rat mammary tumor development is a function of the relative proportions of circulatory prolactin and estrogen. The evidence clearly shows that in the rat, dietary fat induced periodic increases in serum prolactin levels; although the mechanism by which this occurs is obscure. In summary, studies in model systems support the proposition based on epidemiological considerations, that increased dietary fat intake is positively associated with breast cancer risk. Although a variety of models, both direct and indirect, have been proposed to explain the fat effect, available evidence favors a model involving alteration of the neuroendocrine mechanism regulating the secretion of the pituitary hormone, prolactin. Prolactin is proposed to mediate the fat effect by virtue of its dual capacity as a liporegulatory hormone and as a promoter of mammary tumor development. At the level of the mammary tumor cell, the absolute levels of prolactin may be less important than the relative proportions of prolactin and estrogens.

C. METABOLIC EPIDEMIOLOGY As stated, leads established by epidemiological studies and tested in animal models must be confirmed in the human setting. Therefore, in the following section, the association in humans between hormones, particularly prolactin, and breast cancer risk will be discussed. Emphasis will be placed on nutritional factors as they relate to the endocrine system.

NUTRITION AND CANCER

3 13

Comprehensive reviews of this area have been published by MacMahon et ul. (1973), Zumoff r t ul. (1975), Lipsett (1975), and Hayward (1979). 1. Urinary Androgens and Breast Cancer Risk In the early 1960s, Atkins et ul. (1964) reported that the response of breast cancer patients to adrenalectomy was positively associated with the excretion of two urinary androgen metabolites, the 17-ketosteroids, etiocholanolone, and androsterone. In a large-scale prospective study involving women between the ages of 30-55 years, Bulbrook ct ul. (1962) found that women who eventually developed breast cancer exhibited significantly decreased excretion of urinary 17-ketosteroids when compared to women who did not develop the disease. On the basis of these early observations, hypothesis was advanced suggesting that breast cancer risk was increased in women who exhibited subnormal excretion of urinary androgen metabolites (Bulbrook et ul., 1962, 1971; Bulbrook and Hayward, 1967). However, not all case-control studies confirmed Bulbrook’s hypothesis. Urinary androgen metabolites in breast cancer patients were reported to be similar to (Wade et uf., 1977; Cameron et ul., 1979); and even higher than matched controls (Benard et a/., 1962). Case-control comparisons carried out in Japan and Poland also contradicted Bulbrook’s hypothesis. Tominaga e f ul. (1975) reported that Japanese breast cancer patients excreted significantly greater amounts of urinary dehydroepiandrosterone sulfate than healthy controls in a population of women (mean age 41). No differences were found in androsterone or etiocholanolone excretion. Sonka et a f . (1973) reported similar results in a population of postmenopausal Polish women (ages 56-63 years). When international comparisons were made, the relationship between subnormal androgen metabolite excretion and increased risk for breast cancer did not confirm the Bulbrook hypothesis. Healthy British women (high-risk) exhibited higher, rather than lower, urinary 17-ketosteroid excretion rates than healthy Japanese women (low-risk) (Bulbrook and Hayward, 1967; Hayward er a l . , 1978). Moreover, in a recent study comparing androgen metabolite excretion in four different ethnic groups with disparate risks for breast cancer, no differences in urinary 1 l-deoxy17-ketosteroids were found among the groups studied (Gross et ul., 1977). 2. Plusmu Androgens and Breust Cancer Risk Evidence has been presented suggesting that plasma levels of DHEA, a major adrenal secretion product, were decreased in postmenopausal breast cancer patients when compared to matched controls (Poortman et

3 14

BANDARU S . REDDY et

a/.

al., 1973). However, this relationship could only be demonstrated after mastectomy, and was not shown in an earlier study by Alqvist er al. (1968). More recently, Wang er a/. (1977) reported that no significant difference could be found between mean plasma levels of androstenedione (which is secreted in substantial amounts by the postmenopausal ovary as well as the adrenals) in normal Japanese women compared to breast cancer patients; and as Poortman er a/. (1973) previously found, subnormal values could be detected after mastectomy, but not before. If subnormal plasma androgens are associated with breast cancer, it is unusual that they cannot be demonstrated prior to surgery. In fact, the surgery itself may be responsible for the altered androgen concentrations noted. Hill et al. (1976b) compared plasma androgen profiles in healthy urban U.S. white (high-risk), Bantu, and Japanese women (low-risk). Among premenopausal women, no significant differences were found in plasma testosterone levels. Mean plasma levels of DHEA, the chief blood precursor of etiocholanolone and androsterone, were significantly higher in Bantu than in U.S. white and Japanese women; plasma androstenedione levels, however, were similar in the three populations. Among postmenopausal women, although mean plasma levels of DHEA were similar in the three populations, plasma androstenedione was markedly elevated in U.S. whites compared to both Japanese and Bantu women. The relevance of the latter finding to cancer risk is unclear at present. A recent study by Hayward et al. (1978) compared several steroid hormones in native Japanese, Hawaiian Japanese, and British women. Plasma androstenedione and urinary 1 1-deoxy ketosteroids and etiocholanolone levels in Hawaiian Japanese were similar to those in native Japanese women but significantly less than those in British women. Plasma DHEA sulfate levels in Hawaiian Japanese women resembled those of British women and were significantly higher than those in native Japanese women. In light of the conflicting results obtained from urinary and plasma androgen metabolite studies and the absence of information concerning the physiological role of androgens in breast carcinogenesis, the idea that risk of breast cancer is a direct result of adrenal dysfunction must be regarded with skepticism at present. Nonetheless, since changes in androgen metabolism are commonly found in high-risk groups, one cannot entirely exclude on the basis of present evidence a role for androgens in the etiology of breast cancer. 3. Urinary Estrogens and Breast Cuncer Risk The majority of studies in this area have centered around what is commonly known as the estriol hypothesis. Interest in the urinary me-

3 15

NUTRITION AND CANCER

tabolites of estrogen, estrone (El), estradiol (E2),and estriol (E3) centered initially on their use as prognostic aids for therapy. In 1966, Lemon et al. ( 1966) reported that premenopausal breast cancer patients exhibited reduced estriol excretion rates compared to matched controls. Later, Cole and MacMahon ( 1969) suggested that reduced estriol excretion early in reproductive life predisposed to future breast cancer. Stated briefly, the estriol hypothesis holds that the risk for breast cancer is inversely proportional to the ratio of urinary E3 to E, and E2 (Lemon, 1976). A number of studies, including case-control comparisons (Lemon et al., 1966), international comparisons (MacMahon et a/., 1971), migration studies (Dickinson et ul., 1975), within country comparisons (Gross et al., 1977), and animal experiments (Lemon, 1975), support the estriol hypothesis. However, the association between high risk and low urinary estriol ratios was not confirmed in several other case-control studies (Arguelles et ul., 1973; Hellman et a / . , 1971; Marmorston et a/., 1965a,b; Tominaga et a / ., 1975). The results of international comparisons have also been conflicting. Kumaoka et a f . (1973) reported, in contrast to MacMahon et a / . (1971), that differences in urinary estriol ratios in Japanese and British women aged 15 to 19 were small and that there was no difference in women aged 35 to 39. Dickinson et a / . (1975) reported that Hawaiian Japanese exhibited estriol ratios. Intermediate between native Japanese and white Hawaiians. A recent study by Hayward et ul. (1978) found no difference in estriol quotients in healthy Hawaiian Japanese compared to either British or native Japanese women. 4. Plasma Estrogens and Breast Cancer

Risk

The results of case-control studies are mixed. Plasma Ez levels have been reported to be unchanged (Jones et a / . , 1977; McFadyen et a/., 1976; Hill et a / . , 1976a), decreased (Hill and Wynder, 1976), and elevated (England et al., 1975; Malarkey et a / . , 1977a,b) in breast cancer cases when compared to matched controls. It is difficult to make any judgment concerning the relative validity of these studies since each was conducted under different conditions. Recently, Bulbrook et al. (1978) analyzed the plasma estradiol and progesterone levels in pre- and postmenopausal women at varying degrees of risk. The degree at risk was assessed on the basis of family history, age at menarche, and age at first birth. It was found that plasma estradiol values did not vary with risk. Subnormal plasma progesterone levels, however, were associated with increased risks. In an international comparison, Bulbrook et a / . (1976) reported that plasma EZ, El, and progesterone concentrations in healthy British and

3 16

BANDARU

s. REDDY et al.

Japanese women were similar at all ages, although Japanese adolescents exhibited higher E2levels during the luteal phase. These results confirm an earlier report by Kumaoka et a/. (1973) that plasma levels of E2 in British and Japanese women were indistinguishable. In contrast, Hill et al. (1976b) compared plasma E2concentrations in three populations with different risks for breast cancer. Premenopausal U.S. white women exhibited significantly lower plasma concentrations of E2 compared to matched Japanese and Bantu women; postmenopausally, no differences were observed. In a migrant study, Hayward et al. (1978) reported no differences in plasma estrogens in premenopausal native Japanese, Hawaiian Japanese, or British women. In postmenopausal women, however, plasma estradiol concentrations were decreased in Hawaiian Japanese women, compared to both native Japanese and British women. It is apparent that no consistent pattern can be discerned in reference to urinary or plasma estrogen metabolites and breast cancer risk. If anything, plasma studies suggest that low blood levels of E2,not high, are associated with increased risk. Aside from the inconclusive results, the validity of the estriol hypothesis has been questioned for a variety of other reasons. Lipsett (1975) pointed out the limited nature of information obtained from the study of urinary steroid metabolites. The three principal estrogens, for example, have completely different profiles in blood and urine; whereas E2is the dominant estrogen in blood, E3 is the dominant estrogen in urine. Hence, urinary estrogen profiles are not necessarily indicative of blood levels. Cole et a / . (1978) summed up the current status of the estriol hypothesis by stating that “useful data are sparse and inconsistent, and no conclusion can be drawn regarding the relationship between the estriol ratio and breast cancer risk.” The question of whether a high-fat diet increases the rate of conversion of androgens to estrogens is raised by Nimrod and Ryan (1975). To date, however, no experiments concerning the relationship between dietary fat intake and the rate of peripheral conversion in humans have been reported, nor are there any reported studies on changes in urinary or blood estrogen metabolites in animals fed high- or low-fat diets.

5 . Prolactin and Breast Cancer Risk Interest in the role of prolactin in human breast cancer is evidenced by the number of reviews that have appeared recently on the subject (Smithline e f al., 1975; Robyn, 1975; Van der Gugten et al., 1973; McGuire et al., 1978; Nagasawa, 1978).

NUTRITION A N D CANCER

317

Evidence for the role of prolactin in human breast cancer is largely indirect and no definitive proof of prolactin dependency in human breast cancer has yet been demonstrated. Nonetheless, indirect evidence from comparisons of Japanese and U.S. white women suggest that prolactin plays a role in breast cancer. No significant difference was seen in prolactin levels when single morning prolactin concentrations were averaged over the entire menstrual cycle. However, a significant difference was found when mean plasma prolactin levels taken over 5 consecutive days in healthy whites and healthy Bantu women were compared. In the migrant study by Hayward et al. (1978), no significant differences in mean basal plasma prolactin levels in British, Hawaiian Japanese, and native Japanese women, either premenopausally or postmenopausally, were found. The inconclusive nature of these studies does not necessarily mean that prolactin concentrations in high- and low-risk populations are indeed identical. As shown by experimental studies in rodent and in several human studies, analysis of nocturnal secretion profiles and/or the use of provocative tests of pituitary reserves may provide more meaningful measures of plasma prolactin in high- and low-risk populations. Since experimental studies in rodents suggest a link between dietary fat intake, prolactin hypersecretion, and breast cancer risk, Hill and Wynder ( 1976) conducted a series of dietary intervention experiments to determine whether a similar relationship exists in humans. In the first of three studies, volunteers were placed on a standard Western diet for 2 months and then switched to a strict vegetarian diet for 2 months. After each dietary rkgimen, overnight prolactin patterns were measured. The data indicate that there was a marked increase in plasma prolactin concentrations during periods of peak secretion (during deep sleep) in women on a high-fat standard Western diet compared to a low-fat vegetarian diet. These results show clearly that the amount and type of fat consumed in the diet can induce marked changes in the nocturnal secretion patterns of prolactin in humans (Fig. 15). In a second study involving healthy premenopausal women, morning plasma prolactin levels were measured every other day throughout the menstrual cycle after (1) 2 months on a Western diet and (2) 2 months on a vegetarian diet (Hill er ul., 1977). It was found that the mean prolactin levels averaged over the entire menstrual cycle was greater after 2 months on a Western diet than after 2 months on a vegetarian diet. To conclude, although evidence for the role of prolactin in human breast cancer is still largely indirect, there is a growing body of evidence to suggest that the prolactin secretion patterns are associated with increased risk for breast cancer. Moreover, dietary intervention studies

3 18

BANDARU

s. REDDY et al.

W

V

I V

V

‘t L

t

12 pm pm pm

10 II

2

am

4

am

5

am

9 am

TIME

FIG. 15. Twenty-four hour plasma prolactin profiles in four healthy premenopausal women taken at hourly intervals by indwelling catheter. Subjects were maintained on a Western diet (W) for 2 months and then transferred to a vegetarian diet for 2 months. After each 2-month period, nocturnal prolactin concentrations were assayed. Mean prolactin concentrations were significantly higher in the subjects at 2,4, 5 , and 9 AM ( P< 0.01) and at 10 PM (P < 0.05) when on a Western compared to a vegetarian diet. N o differences were seen at 12 PM. Bars represent difference between means. (From Hill and Wynder, 1976.)

conducted at our Institute indicate that the regulatory center controlling prolactin secretion is responsive to changes in dietary fat consumption in humans, as well as in rodent. The latter studies also suggest that international differences in breast cancer incidence and in dietary fat intake may be causally related to differences in prolactin secretion patterns.

D. ETIOLOGIC CONSIDERATIONS In a manner analogous to Berenblum’s two-stage mechanism for skin carcinogenesis (1959), breast carcinogenesis can be operationally divided

NUTRITION A N D CANCER

3 19

into two discontinuous stages: initiation and promotion. As will be seen in the following discussion, dietary fat may play a role either directly or indirectly in both aspects of this disease. Based primarily on the DMBA model, it appears that events that occur prior to or during puberty relate primarily to the initiation phase, while events that occur later in reproductive life increasingly relate t o the promotional phase of carcinogenesis. Viewed from this perspective, there are two points for preventive intervention in breast carcinogenesis: initiation, at or around puberty, and promotion, following puberty. Since initiation is considered to be irreversible, preventive action to block malignant transformation of ductal cells, will require identification of potential carcinogens. Promotion, on the other hand, is a reversible time-dependent process, and therefore, more accessible to preventive measures. Since preneoplastic lesions will not develop into clinically recognizable tumors in the absence of promoting factors, attentuating effects of promoting agents, whether of endogenous or exogenous origin can serve defucro to inhibit the development of breast cancer. As a general hypothesis, we propose that the promotion of human mammary tumor development is a direct function of the relative proportions of circulating prolactin and estrogen. Agents, whether exogenous or endogenous, that elevate the prolactitdestrogen (P/E) ratio serve to enhance breast tumor development; agents that lower the P/E ratio serve to suppress breast tumor development. A distinction is drawn between E l and E,. Because E, is a more potent estrogen, is present in plasma in higher concentrations, and drops more abruptly at menopause, the P/E ratio will be defined for purposes of this discussion as the P/E2 ratio. Moreover, since ductal fluid concentrates prolactin and probably E, as well, the P/E, ratio will apply for the present to both plasma and ductal fluid. Analysis of the histopathogenesis of breast cancer suggests that the promotion phase consists of a series of discontinuous steps rather than a smooth continuous transition from transformed cell to carcinoma (Haslam and Bern, 1977). Although it cannot be said with certainty whether mammary cells transform directly to neoplastic cells or first go through a preneoplastic stage, the weight of evidence suggests the latter. This is an important distinction, since a mechanism involving discrete steps implies that progress from one step to another can be blocked. The PIE2 ratio, for example, may act at one step in the sequence, and not at other steps. Welsch ( 1978), found that inhibition of spontaneous mammary tumorigenesis in mice by prolactin suppression was most effective in young nulliparous mice, less in older multiparous mice, and ineffective in old mice. These results imply that whereas early breast lesions in the

320

BANDARU

s. REDDY et al.

mouse are prolactin-dependent, at later stages of development, they lose this dependency. Since preneoplastic lesions in the mouse model do not parallel those in rat or human breast cancer in terms of their histopathogenesis, direct extrapolation of Welsch’s studies is not possible. Nonetheless, studies of the mouse model illustrate an important point, namely, that failure to demonstrate prolactin dependence in an advanced mammary carcinoma does not rule out a role for prolactin in the early developmental stages of the disease. Repetition of the study in the DMBA or MNU models would provide valuable insight into the hormone sensitivity of the early atypical lesions in human breast cancer. There is no direct evidence that intraductal hyperplasia in humans responds to either diet or drug-induced elevations in P/E2ratios. However, indirect evidence supporting such a proposition can be found in the study of Nomura et al. (1975) referred to previously. In this study, it was found that the percentage of proliferative-type, latent carcinoma in Hawaiian Japanese was twice that of native Japanese. Since breast cancer incidence rates and dietary fat intake have risen in parallel over time among Japanese migrants to Hawaii, these results suggest that environmental factors, and particularly dietary fat intake, may influence the development of latent carcinomas in humans. Measurement of nocturnal P/E2 ratios in Hawaiian Japanese and native Japanese will be necessary to determine whether or not this effect is mediated by changes in prolactin secretion patterns. In addition, histological analysis of terminal duct hyperplasia in DMBA-treated rats fed high- and low-fat diets, combined with hormone analysis, would provide experimental evidence that a fatprolactin mechanism is at work during the early phases of breast cancer developinent . The possibility that in situ lesions can be controlled by exogenous means is of considerable importance from the preventive standpoint. Should the initiation process prove too elusive or difficult to block, then intervention at a second, later, point in the progression from transformed cell to carcinoma may result in the same net effect: reduced tumor incidence. Further understanding of the steps involved in the formation of in situ lesions and the hormonal milieu that regulates their development is much needed at present. The key feature of the hypothesis under consideration is that it provides a plausible explanation for the previously unexplained variations in international incidence patterns noted by epidemiologists. In addition, it eliminates the necessity to postulate a dual etiology for breast cancer: one to account for premenopausal incidence patterns, the other to account for postmenopausal incidence patterns. A single mechanism, involving the relative concentrations of circulating (or ductal fluid) prolactin

NUTRITION A N D CANCER

32 1

and estradiol serves to explain the observed incidence patterns for all age groups. A P/Ez mechanism can also account for the presence of high-risk groups within a population, and the fact that survival rates in Japanese breast cancer patients are significantly higher than U.S. breast cancer patients, following surgery. International differences in breast cancer incidence rates are explained as follows: The difference in breast cancer incidence in the U.S. and Japan is 3: 1 premenopausally, and 9: 1 postmenopausally. In premenopausal women, the ovary actively synthesizes and secretes E2, throughout the menstrual cycle in a fairly regular cyclical pattern (Korenman et al.. 1978; Sherman et ul., 1976). While it has not been proven beyond doubt that E, stimulates prolactin secretion in humans as it does in rats, it has been demonstrated that Ez potentiates the effect of centrally acting prolactin-releasing drugs (Carlson et al., 1973; Buckman and Peake, 1973). Hence sleep-related increases in prolactin secretion are centrally or CNS-regulated events, whereas the Ez effects on prolactin secretion are peripherally regulated events. Within this context, it is proposed that in premenopausal Japanese women dietary fat exerts a small effect on the nocturnal prolactin peak. However, the presence of E, tends to potentiate any influence of dietary fat. Accordingly, the net difference in nocturnal P/E2 ratios in premenopausal Japanese and American women are not expected to be particularly marked. As a result, breast cancer incidence rates, though higher in the U.S., are not dramatically higher in the premenopausal age groups. At menopause, however, plasma E2 levels drop by approximately 80% of the premenopausal concentrations (Judd et al., 1974; Chakravarty et ul.,. 1976). Basal plasma prolactin levels, on the other hand, decrease by 40% according to some workers (Vekemann and Robyn, 1975), and according to Hill et al. (1977) and Kumaoka et al. (1976) may exhibit no decrease following menopause. Although the exact nature and extent of prolactin changes at menopause remain to be clarified, it is clear that sleep-related nocturnal peaks still occur (Rosencweig et ul., 1973; Rosen et ul., 1977; Malarkey et ul., 1977). Hence, as a natural consequence of the menopausal transition, the P/Ez ratio undergoes an abrupt increase in all postmenopausal women. It is after menopause, then, that central mechanisms (diethleep, drugs) controlling prolactin release predominate over peripheral mechanisms (Ez). Since U.S. white women consume three times the amount of fat consumed by Japanese women, the P/E2 ratio during the postmenopausal years is expected to be significantly higher in U.S. women compared to Japanese women. The subsequent increase in P/Ez concentrations in U.S. women may then stimulate the growth of in situ lesions. Stemmerman’s migrant data tends to support

322

BANDARU

s. REDDY et al.

such a prediction. Japanese migrants in the age group 50-59 showed a frequency of proliferative latent carcinoma similar to native Japanese. However, the frequency increased markedly in migrant women between the ages 60-69 and 70-79, compared to native Japanese women in the same age groups. The frequency of proliferative type, in situ lesions in premenopausal Japanese women in Hawaii and Japan has not been reported. On the basis of the above considerations, however, one could predict that the frequencies of proliferative type, in situ lesions in premenopausal Japanese migrants to Hawaii and native Japanese women would not be dramatically different. It can be seen from the foregoing that there is no need to postulate a dual etiology for breast cancer. A single mechanism based on the relative concentrations of plasma or ductal fluid prolactin and estradiol may be invoked to explain both pre- and postmenopausal disease patterns. The postmenopausal disease is not the result of a different mechanism of tumor promotion, but of an increase in its magnitude on women on a Western diet. Since most Western women, by definition, are at high risk, high-risk groups within a Western population can be viewed as extra-high risk individuals. Although most of the evidence is still tentative in nature, it appears that extra-high risk women exhibit disordered mechanisms governing the nycterohemeral secretion of prolactin. The net result of this effect is that women in these populations exhibit both temporal and quantitative changes in prolactin secretion. Analysis of nocturnal plasma and ductal fluid P/E2 ratios in women from high-risk groups, along with longitudinal studies, should shed further light on this important subject. Three independent studies have shown that the recurrence rate after breast surgery in postmenopausal breast cancer patients in Japan is markedly lower than in the U.S. Our experimental studies in ovariectomized animals (which partially mimic the postmenopausal condition) suggest a dietary-hormonal basis for these observations. It is proposed that postmenopausal Japanese patients, consuming a low-fat diet, have lower nocturnal ductal or plasma P/E ratios than their U.S. white counterparts. Assuming that some preneoplastic foci are present after surgery, the promoting influence of a high P/E ratio on this lesion in U.S. white women would be considerably greater than in Japanese women. As a result, the period of time elapsing between surgery and breast cancer recurrence would be shortened in U.S. white compared to Japanese women. Proof that such a diet-based mechanism is indeed occurring would require monitoring of nocturnal prolactin and estrogen levels in postmenopausal Japanese breast cancer patients following surgery. If the P/E2

NUTRITION A N D CANCER

323

hypothesis is corroborated, careful regulation of fat intake in postmenopausal U.S. patients may lower their recurrence rates to those exhibited by Japanese patients.

E. CONCLUSIONS Investigators concerned with the prevention of cardiovascular disease have, for some time, recommended a diet lower in calories, fats (in particular, saturated fats), and cholesterol than is typical of the average American diet. It now appears prudent to suggest a similar diet for the prevention of breast cancer, except that in this case data from animal experiments suggest a total fat reduction, including that of unsaturated fats. Most Western-conditioned individuals find it restricting to be placed on a low-calorie diet, a low-fat diet of 20% total calories, a low-cholesterol diet intake of 100 mg per day, as suggested by Connor and Connor (1972). Yet, it would appear that this is the type of diet that our sedentary bodies can adequately metabolize without creating an excess of lipids to overwhelm the metabolic capacities of our system. A prudent diet, even of 30% fat and 300 mg cholesterol per day, is difficult to establish for a sufficient number of women to test the nutritional hypothesis of breast cancer etiology in a prospective survey that would necessarily have to be carried out for at least one decade. While experimental studies need to continue, we propose several metabolic epidemiological studies to test whether prolactin output and particularly its concentration in the breast fluid can be modified by dietary alteration and whether this reduces the occurrence and the progression of breast cancer. We suggest, for these experiments, that the diet should not exceed 1800 calories, 20% total fat with an equal distribution of polyunsaturated and saturated fats, and cholesterol intake of 100 mg per day. At a time when so many types of clinical trials with a large variety of chemotherapeutic regimens are being carried out on cancer patients, it is time to consider both primary and secondary preventive nutritional trials as well. We hope that this review of existing evidence relative to the etiology of breast cancer will serve as a constructive stimulus to other engaged in etiological and preventive aspects of breast cancer. Epidemiologic evidence clearly suggests that breast cancer is not an inevitable consequence of aging. What has been done in this field so far, however, can be regarded as only a prologue. Those engaged in attempting to prevent breast cancer will not have succeeded until the incidence of breast cancer

324

BANDARU

s. REDDY et al.

is actually reduced. To realize this possibility requires greater attention and more interdisciplinary activity than we have given to the field of breast cancer etiology so far. VII. Dietary Factors and Cancer of the Prostate

A. EPIDEMIOLOGY Epidemiologic data on prostatic cancer have been reviewed by King et uf. (1963), Wynder et af. (1971), Hutchison (1976), and Blair and

Fraumeni (1978). Cancer of the prostate is common in western countries, including the United States, and uncommon in Japan and Africa (Doll et ul., 1970) (Fig. 16). Southern and Eastern European and Latin American countries have an intermediate incidence of prostatic cancer. The high rates seem to be associated with populations of Northern Europe or of Northern European origin. Among whites in the United States, mortality is elevated in areas with a high percentage of residents of Scandinavian

RATE PER 100,000 POP.

0

5

I0

15

20

25

I

US., NONWHITE SOUTH AFRICA SWEDEN NORWAY SWITZERLAND BELGIUM FRANCE NETHERLANDS AUSTRALIA NEWZEALAND AUSTRIA GERMANY, F.R. CANADA DENMARK US., WHITE SCOTLAND IRELAND ENGLAND a WALES FINLAND NORTH. IRELAND CHILE ITALY PORTUGAL ISRAEL JAPAN

FIG. 16. Age-adjusted death rates for prostate cancer in different countries, 1%6-1967. (From Segi and Kurihara, 1972.)

NUTRITION AND CANCER

325

descent (Blair and Fraumeni, 1978), whereas blacks have a higher incidence than whites in the United States. Compared with U.S. blacks, African blacks seem to have a low rate of cancer of the prostate. The differences suggest that environmental factors contribute to the high incidence of prostatic cancer among U.S. blacks. Although prostate mortality and morbidity vary by country, the rate appears to be affected by migration of population from low- to high-risk countries. Haenszel and Kurihara (1968) showed that the first generation immigrants in the U.S. from Japan have an increased mortality rate for prostatic cancer. An increased mortality rate was also found in Polish immigrants to the United States and Australia (Staszewski and Haenszel, 1965), and in foreign-born Caucasians compared to native-born (Newill, 1961). Wynder et al. (1971) found that an increase in mortality from prostatic cancer in migrants to the U.S. appeared only after 20 or more years residence in the U.S. Studies of migrants suggest that environmental factors, rather than genetic characteristics, account for a substantial part of the international variation in mortality rates for prostatic cancer. Since association between incidence of prostatic cancer and measures of socioeconomic status are not significant (Seidman, 1970; King et al., 1963), other environmental factors in the high-risk areas would appear to be involved. For example, mortality rates for U.S. blacks rose sharply with increasing population density, whereas the rates for blacks in sparsely populated countries did not (Blair and Fraumeni, 1978). This finding suggests that the special susceptibility of American blacks to prostatic cancer i's associated with urban residence (Blair and Fraumeni, 1978). From comparative studies of prostatic tissue in Japanese and Caucasian men, Akazaki (1973) postulated that environmental factors were responsible for the activation of latent lesions. These small latent lesions, which occur at a constant frequency in all areas, appear to progress to active lesions mainly in Western societies (Barnetson, 1954; Breslow er al., 1977). The next question is whether or not prostatic cancer has an epidemiology sufficiently similar to the epidemiologies of cancers of the breast in females and the large bowel that we should seek a common cause for them. Internationally, the incidence rates for the endocrine-dependent cancers (breast and prostate) follow fairly closely the rates of large-bowel cancer (Berg, 1975a). Although breast cancer incidence is linked particularly closely with large-bowel cancer, the association between cancer of the large bowel and of the prostate is statistically significant (Berg,

326

BANDARU

s. REDDY et al.

1975a). There is a positive correlation between prostatic cancer and female breast cancer in the United States, suggesting a common etiology for these two endocrine-related cancers (Wynder et al., 1967; Blair and Fraumeni, 1978).

B. ETIOLOGY International differences in prostatic cancer morbidity and mortality are substantial and suggest the influence of environmental factors (Wynder et al., 1967, 1971; Berg, 1975a). Interestingly, one striking difference between diets in high- and lowrisk areas is the fat intake (Fig. 17), which accounts for 40% of the daily calories in high-risk and 20% calories in low-risk areas. Blair and Fraumeni (1978) used national surveys to examine regional differences in the consumption of high-fat foods. The Central United States, including the high-rate North Central and Midwest regions, had a higher consumption of beef and milk products (5.9 and 14.4 Ib/week, respectively) per household than did the Northeast or South. For fats and oils, pork, and eggs, the South had a slightly higher consumption than did the Central United States, Northeast, or West. The total consumption of these high-fat foods was highest in the Central United States, intermediate in the Northeast and West, and lowest in the Southeast. This pattern parallels that of mortality for prostatic cancer among whites (Blair and Fraumeni, 1978). Assuming a positive correlation between dietary factors and incidence of prostatic cancer, what is the connecting link? Clinical studies have clearly shown that this disease is hormonally dependent (Fergusson, 1972) and that changes in hormonal metabolism occur in prostatic cancer patients (Marmorston er al., 1965a,b). Any factor that affects hormonal secretion, retention, and, in particular, the sensitivity of the target organ and/or cells influences the frequency of this cancer (Wynder ef af., 1971). Since diet may modify hormonal systems, it has the potential of inhibiting or enhancing tumorigenesis, i.e., diet or nutritional status may function as modifiers of prostatic tumorigenesis.

C. METABOLIC EPIDEMIOLOGY Concerning the relationship of lifestyle and diet to hormonal status, exercise alters androgen metabolism (Kuoppasalmi er al., 1976; Sutton et ul., 1973), whereas diet may directly modify hormonal activity (Edozien, 1960; Merimee and Fineberg, 1974) or may act indirectly through

2ot

:

CL

0

xQ 0

6.AFRICA .SWEDEN NORWAY. SWITZERLAND0 U S * AUSTRALIA. TT~RLANM AUSTRIA. F R A N ~Y N m A

\

a u

15HUNGARY@

I + a

.

w n 10w Icn 3

7

n

? w

COLWBIAO

5-

a Q

PH'LIPPtMS \

-

THAILAND

JAPPAN .TAWAN .CEYLON,

7

*IRELAND

.POLAND

.GREECE

.MEXICO E L SALVADOR

*

wLGARlqmYUGOSLAVIA

mU,K,

GERMANY.

*OR T UGAL .FINLAND PUERTO RICS ITALY QPA1N VENEZUELA. .CZECHOSLDVAKIA .ROMANIA ISRAEL. CHILE*AMA

W O N G KONG 1

I

I

I

NEW ZEALAND

3 28

BANDARU

s. REDDY et al.

neurotransmitters, biogenic amines, and alteration in sleep patterns (MacIndoe and Turkington, 1972; Phillips et al., 1975; Boyar ef al., 1972). In a comparative study of the hormonal levels in North American white and black men, high-risk groups, with South African black men, a lowrisk group, Hill et al. (1978) reported higher plasma estradiol levels in South African black men; testosterone levels were comparable in all three ethnic groups and human chorionic gonadotrophic produced a comparable release of testosterone in the three groups of men. However, in South African black men, the release of prolactin was significantly greater following TRH injection than in North American men. When North American black and South African black men were fed a vegetarian or a Western diet, respectively, urinary androgen and estrogen levels decreased in the South Africans (Hill er al., 1979) (Fig. 18). Hill and Wynder ( 1979) have also reported that a vegetarian diet decreases the overnight release of prolactin and the plasma level of testosterone. Evidence would

Ip Western

-5

I

t

-P

Customary

t

30

Vegetarian

0

c 0

c

C0

DIETS

\

-P

25 W

u

0

20 v)

W

* z a

a 15 4

= 2

10

+ 4

0

+

5

NAW

tB

SAB

LW

NAB

SAB

FIG. 18. Effect of vegetarian and Western diets on urinary estrogens and androgens in North American White and Black men and in South African Black men. On their customary diets, South African Black (SAB) men ( n = 21) had significantly lower tt(P < 0.01) urinary estrogen and androgen levels than North American Black (NAB) men ( n = 18). When fed a Western diet, urinary estrogen and androgen levels increased **(P< 0.01) in SAB men, while a vegetarian diet decreased **(P< 0.01) the urinary androgen and estrogen levels in NAB men. Levels of urinary androgen and estrogen levels for North American White (NAW) men ( n = 16) are also given. (From Hill et al., 1979.)

NUTRITION AND CANCER

329

therefore suggest that hormonal metabolism in men can be modified by dietary factors. Whether similar changes occur in men with prostatic cancer is currently being investigated by Hill and Wynder from our Institute. D. EXPERIMENTAL STUDIES Recently, Fingerhut and Veenema (1977) described an animal model to obtain prostatic animal carcinoma similar to that in man. However, there are as yet no studies to validate the model and to test the relationship of nutrition to prostatic cancer.

E. CONCLUSIONS Based mainly on epidemiologic studies, evidence suggests that nutrition plays an important role in the etiology of this disease. As with breast cancer, this disease is hormonally dependent and is, therefore, subject to manipulation by surgical ablation and by dietary modification. Added support for the importance of environmental factors in the etiology of this disease arises from the findings of Berg (1975a,b) who reported an association between large-bowel and prostatic cancer. Further study of the effect of specific dietary components on hormonal activity in men at different risk for this disease, complemented with in vitro studies, should elucidate the dietary role in the etiology of prostatic cancer. VI I I. Conclusions

At present, we have overwhelming evidence of remarkable variations in the overall cancer incidence and of the incidence of specific types between countries and within countries. None of the risk factors for cancer is probably more significant than diet and nutrition. The evidence from high- and low-incidence populations, worldwide, and the corresponding correlations with various dietary factors, the even more compelling picture that has emerged and continues to emerge from migrant studies and studies of singular populations (such as Seventh-Day Adventists in the United States), the analysis of parallel trends in incidence and dietary changes in certain well-characterized populations, and evidence arising from case-control and other studies on specific dietary

330

BANDARU

s. REDDY et af.

factors, all seem to provide highly defensible arguments for dietary implications in the causations of certain major human cancers (stomach, breast, and colon), and for corrective dietary recommendations toward prevention. The evidence is further reinforced by experimental animal studies that provided data strikingly supportive of the epidemiologic studies and that have provided leads to the probable mechanism of action. The various aspects of this article have provided the evidence that justifies the exclusion of environmental, occupational, or genetic factors as significant contributions to the etiology of these cancers. While the experimental scientist is rightly preoccupied with the need for a mechanistic understanding of the precise epidemiologic and experimental clues on hand, those with interest and responsibility in public health cannot fail to visualize the present opportunities for intervention, even before the detailed mechanistic picture is precisely and totally defined. With certain limits, dietary intervention seems to offer an exceptionally favorable ratio of risk and benefits, a situation where the population would have little to lose and probably much to gain. Thus, we have the research tools in hand to mount a concerted, effective effort to reduce the risk for the main premature killing diseases. It should be possible to design lifestyles and nutritional habits compatible with local customs and civilizations, not very different from those currently in effect in high-risk countries, for the types of neoplasms discussed in this article. Such modified lifestyles would ultimately serve to reduce the overall risk for these diseases, and hence provide a rational basis for the prevention of types of cancer now affecting literally millions of people around the world. For some time now, experts in the field of arteriosclerosis have recommended a modification of the typical western diet to reduce the prevalence of hyperlipidemia. To the extent that some aspects of this recommended prudent diet affect the major cause of death in most western countries, it may also relate to the reduction of certain types of cancers. When it was observed that certain types of cancer are prevalent in other western countries, a modification of the western diet along these lines may be necessary. A prudent diet should, therefore, lead to the prevention of nutrition-related cancers, namely, colon, breast, prostate, stomach, pancreas, upper alimentary tract, and the like. There are a number of studies that suggest that adequate amounts of vitamins A , C, and E and B vitamines, and certain trace minerals such as Se and Zn, in the diet might, in addition, prevent the development of certain cancers such as breast, esophagus, bladder, cervix, and lung.

NUTRITION A N D CANCER

33 I

It is beneficial for general health purposes to have an adequate amount of vitamins and trace minerals in our diet. The reduction of stomach cancer has been suggested to relate, at least in part, to an increased consumption of fresh fruits and vegetables. As we reduce our intake of fat and animal proteins, we need to replace these food categories with certain types of vegetables, including those containing vitamin C. The intake of salt and pickled food should be reduced, and of other foods with unnecessary substances, such as high levels of nitrites and nitrates. Additional recommendations would affect food preparation to avoid or reduce consumption of fried meats and proteins. Inasmuch as we and others have demonstrated that certain types of fibers act as a protective factor in colon carcinogenesis, it would seem that an increase in intake of foods containing fibers (mainly cereal fibers and also vegetable and fruit fibers) and a reduction in the fat calories from 40% to 20% are recommended. Since a high-fat diet has also been associated with cancers of the breast and prostate, the recommended diet may help prevent breast and prostate cancer. Since heavy alcohol consumption is associated with a high risk of cancer of the upper alimentary tract and larynx, reduction of alcohol intake to moderate levels is recommended. In conclusion, there would seem to be sufficient evidence to propose modifying the diet of Western countries to reduce total dietary fat, animal protein, and cholesterol, increase dietary fiber on the lines of above prudent diet, and increase certain food items, such as fresh vegetables and fruits. Diet should be well balanced to provide adequate amounts of vitamins and minerals. Although further research for more specific preventive measures is required, such measures are unlikely to be hazardous and can be advocated with a strong hope for benefits in the population. If these measures are taken and if, in addition, the readily preventable occupational and other environmental cancers are eliminated, we would enter an era where cancers of all types would no longer represent a major cause of death in man.

ACKNOWLEDGMENTS The authors thank Ms. Margaret Mushinski and Ms. Laurel Mathews for their expert editorial assistance and Ms. Arlene Banow for preparation of the manuscript. This work was supported in part by grants CA-16382 through the National Large Bowel Cancer Project, CA-12376, and CA-17613, and Contracts CP-85659, CP-95604, and CP-65818 from the National Cancer Institute.

332

BANDARU

s . REDDY et al.

REFERENCES Akazaki, K. (1973). I n “Host Environment Interactions in the Etiology of Cancer in Man” (R. Doll and I. Vodopija, eds.). I.A.R.C., Lyon, France. Alcantara, E. N., and Speckman, E. W. (1976). A m . J . Clin. Nutr. 29, 1035-1047. Alqvist, K. A., Jackson, A. W., and Stewart, J. G. (1968). Br. Med. J . 1, 217-221. American Cancer Society ( 1979). Cancer Statistics. American Cancer Society, New York. Aragon, L. A. (1964). “Estimacion del consumo de algunos alimentos basicos en la cuidad de Cali.” Tesis de Grado, Cali Universidad del Valle, Facultad de Ciencias Economicas, Cali, Colombia. Arguelles, A. E., Poggi, U. L., Saborida, C., Hoffman, C., Chekherdemian, M., and Blanchard, 0. (1973). Lancet 1, 165- 168. Aries, V., Crowther, J. S., Drasar, B. S.,Hill, M. J., and Williams, R. E. 0. (1969). Gut 10,334-335. Armstrong, B., and Doll, R. (1975). I n t . J. Cancer 15, 617-631. Atkins, H. J. B., Bulbrook, R. D.,Falconer, M. A., Hayward, J. L., MacLean, K. S., and Schunn, P. H. (1964). Lance? 2, 1133. Autrup, H., Harris, C. C., and Trump, B. F. (1978). Proc. Soc. Exp. B i d . Med. 159, 1 1 1115. Bagheri, S. A., Bolt, M. G., Boyer, J. L., and Palmer, R. H. (1978). Gastroenterology 74, 188- 192. Bansal, B. R., Rhoads, J. E., Jr., and Bansal, S. C. (1978). Cancer Res. 38, 3293-3303. Barnetson, J. (1954). Semain Hopital (Paris) 30, 129- 132. Bartsch, H., Traut, M., and Hecker, E. (1976). Biochim. Biophys. Acta 237, 556-566. Bauer, H. G., Asp, N. G., Oste, R.,Dahlquist, A., and Fredlund, P. (1979). Cancer Res. 39, 3752-3756. Beer, A. E., and Billingham, R . E. (1978). Lancet 2, 296. Benard, H., Bourdin, J. S., Saracino, R. T., and Seeman, A. (1962). Ann. Endocrinol. (Paris) 23, 525-532. Benson, T., Lev, M., and Grand, C. G. (1956). Cancer Res. 16, 135-137. Berenblum, I. (1959).Experientia 15, 285-289. Berg, J. W. (1975a). Cancer Res. 34, 3345-3350. Berg, J. W. (1975b). I n “Persons at High Risk of Cancer: An approach to Cancer Etiology and Control” (J. F. Fraumeni, Jr., ed.), pp. 385-398. Academic Press, New York. Berndt, H., Wildner, G. P., and Klein, K. (1968). Neoplasms 15, 501-515. Best, E. W. (1966). I n “A Canadian Study of Smoking and Health,” pp. 65-86. Department of National Health and Welfare, Ottawa, Canada. Biometry Branch, National Cancer Institute. (1974). Third national cancer survey (Advanced Three Year Report 1969-1971), DHEW Publ. No. (NIH) 74-637. National Cancer Institute, Bethesda, Md. Bjelke, E. (1974). Scand. J . Gastroenterol. 9, (Suppl. 31), 1-253. Blair, A., and Fraumeni, J. J. (1978). J . Natl. Cancer I n s t . 61, 1379-1384. Boyar, R., Finkelstein, J., Roffwarg, H., Kapen, S., Weitzman, E., and Hellman, L. (1972). N . Engl. J . Med. 287, 582. Boyland, E., Roe, F . J. C., Gorrod, J. W., and Mitchley, B. C. V. (1964). Br. J . Cancer 23, 265-270. Boyns, A. E., Buchan, R.,Cole, E. N . , Forrest, A. P. M., and Griffiths, K. (1973). Eur. J . Cancer 9, 169-171. Bralow, S. P., and Weisburger, J. H. (1976). Clin. Gastroenterol. 5, 527-542.

NUTRITION AND CANCER

333

Breslow, N., Chan, C. W., Dhom, G., Drury, R. A. B., Franks, L. M., Giellei, B., Lee, V. S., Lundberg, S., Sparke, B., Sternby, N. H., and Tulinius, H. (1977). Inr. J . Cuncer 20, 680-688. British Medical Journal, Editorial (1974). Br. M i d . J . 2, 134-135. Broitman, S. A., Vitale, J. J., Vavrousek-Jakuba, E., and Gottlieb, L. S. (1977). Cancer 40, 2455-2463. Bruce, W. R., Varghese, A. J., Furrer, R., and Land, P. C. (1977). In ”Origins of Human Cancer” (H. H. Hiatt, J. D. Watson and J. A. Winsten, eds.), pp. 1651-1656. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York. Buckman, M. T., and Peake, G . T. (1973).J . Clin. Endocrinol. 37, 977-980. Buell, P. (1973). J . Null. Cancer Inst. 51, 1479-1483. Bulbrook, R. D., and Hayward, J. L. (1967). Lancet 1, 519-522. Bulbrook, R . D., Hayward, J. C., Spicer, C. C., and Thomas, B. S. (1962). Lancet 2, 12351237. Bulbrook, R. D., Hayward, J. C., and Spicer, C. C. (1971). Lancer 1, 395-398. Bulbrook, R. D., Swain, M. C., Wang, D. Y., Hayward, J. C., Kumoaka, S., Takatani, O., Abe, O., and Utsumomiya, J. (1976). Eur. J. Cancer 12, 725-735. Bulbrook, R. D., Moore, J. W., Clark, G. M. G., Wang, D. Y., Tong, D., and Hayward, J. L. (1978). Eur. J . Cancer 14, 1369-1375. Bull, A. W., Soullier, B. K.,Wilson, P. S., Hayden, M. T., and Nigro, N. D. (1979). Cancer Res. 39, 4956-4959. Burkitt, D. P. (1971). Cuncrr 26, 3-13. Burkitt, D. P. (1975). J. Nutl. Cuncer Inst. 54, 3-6. Burrell, R. J. W., Roach, W. A., and Shadwell, A. (1966). J. Narl. Cancer I n s t . 36, 201214. Cameron, E. D. H., Griffiths, K., Gleave, E. N., Stewart, H., Forrest, A. P. M., and Campbell, H . (1979). Br. Med. J . 4, 768-779. Campbell, T. C. (1977). Clin.Pharmucol. Ther. 22, 699-706. Carlson, H. E . , Jacobs, L. S., and Daughaday, W. H. (1973). J. Clin.Endocrinol. 37,488490. Carroll, K. K. (1975). Cancer Res. 35, 3374-3383. Carroll, K. K . , and Khor, H. T. (1971). Lipid.\ 6, 415-420. Carroll, K. K., and Khor, H. T . (1975). Prog. Biochem. Pharmucol. 10, 308-353. Chakraborty, J. (1978). In “Clinics in Endocrinology and Metabolism” (V. Marks and J. Wright, eds.), pp. 273-2%. Saunders, London. Chakravarty, S., Collins, W. P., Forecast, J. D., Newton, J. R., Oram, D. W., and Studd, J. W. W. (1976). Br. Med. J . 2, 781-784. Chan, P. C., and Cohen, L. A. (1974). J. Nutl. Cuncer Inst. 52, 25-30. Chan, P. C., and Wynder, E. L. (1970). Cancer 16, 1221-1224. Chan, P. C., Okamoto, T., and Wynder, E. L. (1972). J . Nutl. Cancer Inst. 48, 1341-1345. Chan, P. C., Didato, F., and Cohen, L . A. (1975). Proc. Soc. Exp. Biol. Med. 149, 133135. Chan, P. C., Head, J . F., Cohen, L. A., and Wynder, E. L. (1977). J. Narl. Cuncer Inst. 59, 1279- 1283. Chomchai, C., Bhadrachari, N., and Nigro, N. D. (1974). Dis. Colon Rectum 17, 310-312. Chopra, S. A,, Linsell, C. A., Peers, F. G., and Chopra, F. S. (1975). I n t . J . Cancer 15, 684-693. Clemmesen, J. (1974). Acta. Pathol. Microhiol. Scand. (Suppl.) 247, IV. Clemmesen, J. (1977). Acta. Pathol. Microhiol. Scand. (Suppl.) 261, 65-75.

334

BANDARU

s. REDDY et al.

Cohen, B. I . , Raicht, R. F., Deschner, E. E., Takahashi, M., Sarwal, A. N., and Fazzini, E. (1980). J . N u t / . Crincer I n s t . (in press). Cole, P., and MacMahon, B. (1969). L(incc,! 1, 604-606. Cole, P., Cramer, D., Yen, S . , Paffenbarger, MacMahon, B., and Brown, J . (1978). Concer Res. 38, 745-748. Commoner, B., Vithayathil, A. J . , Dolara, P., Nair, S., Madyastha, P., and Cuca, G. C. (1978). Scirnc~e201, 913-916. Connecticut State Dept. of Health. (1977). “Cancer in Connecticut, Incidence and Rates 1945- 1965.” Connor, N. E., and Connor, S . L. (1972). Prev. Med. 1, 49-83. Cook, J . W., Kennaway, E. L., and Kennaway, N. M. (1940). Narure ( L < ~ n d o n145, ) 627. Cook-Mozaffari, P. (1979). Nutr. Concer 1, 51-60. Coombs, M. M., Bhatt, T. S. , and Croft, C. J . (1973). Conerr Res. 33, 832-837. Coordinating Group for Research on the Etiology of Esophageal Cancer of North China (1974). The Epidemiology of esophageal cancer in North China and preliminary results in the investigation of its etiological factors. Peking, Peoples Republic of China. Coordinating Group for Research on Etiology of Esophageal Cancer in North China (1975). Chinese. M r d . J . 1, 167-183. Correa, P. (1975). C(incer Res. 35, 3395-3397. Correa, P., and Haenszel, W. (1978). In “Advances in Cancer Research” (G. Klein and S. Weinhouse, eds.), pp. 1-141, Academic Press, New York. Correa, P., Cuello, C., and Duque, E. (1970). J . N d . Ctinr,er I n s t . 44, 297-306. Correa, P., Sasano, N., Stemmermann, G., and Haenszel, W. (1973).J . N u t / . Cuncrr Ins!. 51, 1449- 1459. Correa, P., Haenszel, W., Cuello, C., Tannenbaum, S. , and Archer, M. (1975). L ~ i n c r t2, 58-60. Costlow, M. E., and McGuire, W. L. (1977). J . Nor/. C m c w Ins!. 58, 1173-1175. Cruse, J . P., Lewin, M. R., and Clark, C. G. (1978). Loncet 2, 1278-1280. Cruse, J . P., Lewin, M. R., and Clark, C. G. (1979). Lrincet 1, 152-155. Cuello, C., Correa, P., Haenszel, W., Gordillo, G., Brown, C., Archer, M., and Tannenbaum, S. (1978). J . N u t / . Cnncrr I n s t . 57, 1015-1020. J / .451-459. Culter, M. G . , and Schneider, R. (1974). Food. Costnet. T O X ~ C Y12, Culter, S . J . , and Young, J . L. (1975). In “Persons at High Risk of Cancer” ( J . F. Fraumeni, ed.), pp. 307-342. Academic Press, Inc., New York. Dales, L. G., Friedman, G. D., Wry, H. K., Grossman, S . , and Williams, S. R. (1979). A m . J . Epidrmiol. 109, 132-144. Dao, T. L . , Bock, F. G., and Crouch, S. (1959). Pruc. Soc. ESP. B i d . M e d . 102, 635-638. Day, N. E. (1975). Concw Res. 35, 3304-3307. Dayton, N. E., Hashimoto, S. , and Wollman, J . (1977). J . N u t r . 107, 1353-1360. DeCosse, J . J . , Adams, M. B., Condon, R. E. (1977). Concer 39, 267-273. deJong, U . W., Breslow, N., Goh Ewe Hong, J . , Sridharan, M., and Shanmugaratnarn, K. (1974). In!. J . Conc~,r13, 291-303. deLint, J . (1974). P w v . M e d . 3, 24-35. DeLuca, L., Little, E. P., and Wolf, G. (1969). J . BkJ/. Chcun. 244, 701-708. DHEW Publ. (1974). No. (ADM) 74-124. Dickinson, L. E., MacMahon, B . , Cole, P., and Brown, J . B. (1975). N . En#/. J . Mrd. 291, 1211-1213. Dolard, P., Commoner, B . , Vithayathil, A., Cuca, G., Tuley, E., Madyastha, P., Nair, S., and Kriebel, D. (1979). M u t . Rrs. 60, 231-232.

NUTRITION AND CANCER

335

Doll, R. (1967). "Prevention of Cancer-Pointers from Epidemiology." Nuffield Hospital Trust, London. Doll, R., Payne, P., and Waterhouse, J. eds. (1966). "Cancer Incidence in Five Continents." Technical Report, International Union against Cancer, Berlin and New York. Doll, R., Muir, C., and Waterhouse, J., eds. (1970). "Cancer Incidence in Five Continents," Vol. 11. International Union against Cancer, Berlin and New York. Domellof, L. (1979). M e d . H y p o f h . 5(4), 463-476. Dontenwill, W., Chevalier, J. J . , Harke, H. P., LaFrenz, U., Reckzeh, G., and Schneider, B. (1973). J. Nirtl. Conccr Inst. 51, 1781- 1832. Dormandy, T. L. (1978). Lcrnc.rt 2, 647-650. Drasar, B. S., and Hill, M. J . , eds. (1974). "Human Intestinal Flora," pp. 201-212. Academic Press, New York. Drasar, B. S., and Irving, D. (1973). Br. J . Concer 27, 167- 172. Drasar, B. S . , Jenkins, D. J . A., and Cummings, J . H. (1976). J . M c d . Microhiol. 9, 423. Druckrey, H . (1973).I n "Topics in Chemical Carcinogenesis" (W. Nakahara, S. Takayama, T. Sugimura and S. Odashima, eds.), p. 73-103. Univ. Park Press, Baltimore, Maryland. Druckrey, H. (1975). Grrnn M o n o g r . 17, 107- 132. Dungal. N., and Sigurjonsson, J. (1967). B r . J . Concrr 21, 270-276. Dunning, W. F., Curtis, M . R., and Maun, M. E . (1949). Coricer R e s . 9, 354-361. Edozien, J . C. (1960). L ~ r i c r t1 , 258-259. Endo, H., Takahashi, K., Kinoshita, M., and Baba, T. (1974). Proc. Jn. Arcid. 50, 497502. Endo, H., Takahashi, K., Kinoshita, M., Utsunomiya, T . , and Baba, T. (1975). Gunn M o n o g r . 17, 17-30. England, P. C., Skinner, L . G., Cottrell. K. M., and Sellwood, R. A. (1975). Br. J. Surg. 63, 806-809. Enig, M. G., Mann, R. J . , and Keeney, M. (1978). Fed. Pro<.. 37, 2215-2220. Enstrom, J . E. (1975). Br. J. Concer 32, 432-439. Enstrom, J . E. (1978). Cancer 42, 1943-1951. Falk, H . L., and Kotin, P. (1963). Clin. Phormacol. Ther. 4, 88- 103. Fan. T. Y., and Tannenbaum, S. (1973). J . Agric. Fond Chern. 21, 237-240. Fergusson, J . D. (1972). I n "Endocrine Therapy in Malignant Disease" (B. A. Stoll, ed.), pp. 237-246. Saunders, London. Fiala, E. S. (1977). Cuncer 40, 2436-2445. Fineberg, S. E., Horland, A. A., and Merimee, T. J . (1972). Metabolism 21, 491-498. Finegold, S. M., and Sutter, V. L. (1978). A m . J. Clin. Nutr. 31, S116-S- 112. Finegold, S. M., Flora, D. J., Attebery, H. R., and Sutter, V. L. (1975). Cancer Res. 35, 3407-3417. Fineberg, S. E., Horland, A. A,, and Merimee, T. J . (1972). Metcibolism 21, 491-498. Fleiszer, D., MacFarlane, J . , Murray, D., and Brown, R. A. (1978). Lancet 2, 552-553. Floyd, R. A , , Soong, L. M., Stuart, M. A., and Reigh, D. L . (1978). Arch. Biochem. Biophys. 185, 450-457. Fong, L. Y. Y., Newberne, P. M., Lin, H., and Chan, W. (1977). Proc. Missouri Con$ Trtrce Substcrnc~rsHecrlfh.University of Missouri, Columbia, Missouri. Fong, L. Y. Y.,Sivak, A , , and Newberne, P. M. (1978).J. Natl. Cancer Inst. 61, 145- 150. Food Balance Sheets (1971). 1 9 6 4 1966-Average. Food and Agricultural Organization of the United Nations, Rome. Frdser, W. M., and Blackard, W. G . (1977). Horm. Mrrcrbol. Res. 9, 347-440.

336

BANDARU

s. REDDY et al.

Fraumeni, J. F., Jr. (1975). “Persons at High-Risk of Cancer. An Approach to Cancer Etiology and Control.” Academic Press, New York. Freeman, H. J., Spiller, G. A., and Kim, Y. S. (1978). Cancer Res. 38, 2912-2917. Fuchs, H. M., Dorfman, S., and Floch, M. H. (1976). A m . J. CIin. Nutr. 29, 1443-1447. Fujinaka, N., Masuda, Y., and Kuratsune, M. (1976). Cann 67, 679-683. Furth, J. (1973). I n “Human Prolactin” (J. L. Pasteels and C. Robyn, eds.), pp. 233-248. American Elsevier, New York. Gammal, E. B., Carroll, K. K., Plunkett, E. R. (1967). Cancer Res. 27, 1737-1742. Gammal, E. B., Carroll, K. K., and Plunkett, E. R. (1968). Cancer Res. 28, 384-385. Gerson, S. J . , and Meyer, J. (1977). J . Nutr. 107, 724-729. Gibel, W., Lohs, K. H., and Wildner, G. P. (1975). Arch. Geschwulstforsch. 45, 19-24. Goldin, B. R., and Gorbach, S. L. (1977). Cancer 40, 2421-2426. Gortner, W. A. (1975). Cancer Res. 35, 3246-3253. Graham, S. (1975). Cancer Res. 35, 3464-3468. Graham, S., and Mettlin, C. (1979). A m . J . Epidemiol. 109, 1-20. Graham, S., Levin, M., and Lilienfeld, A. M. (1960). Cancer 13, 180-191. Gray, G. E., Pike, M. C., and Henderson, B. E. (1979). Br. J . Cancer 39, 1 . Gregor, O., Toman, R., and Prusova, F. (1969). Gut 10, 1031-1034. Grifith, G. W. (1968). Cancer 30, 927-938. Gross, J . , Modan, B., Bertini, B., Spira, O., deWaard, F., Thijssen, J. H. H., and Veslergaard, P. (1977). J. Natl. Cancer Inst. 59, 7- 1 1 . Gustafsson, B. F., Einarsson, K., and Gustafson, J. A. (1975). J . Biol. Chem. 250, 84968502. Haddow, A. (1970). Ahbotempo 4, 8-11. Haenszel, W. (1961). J. Nail. Cancer Inst. 26, 37-132. Haenszel, W. (1975). I n “Persons at High Risk of Cancer. An Approach to Cancer Etiology and Control” (J. F. Fraumeni, Jr., ed.), pp. 361-371. Academic Press, New York. Haenszel, W., and Correa, P. (1975). Cancer Res. 35, 3452-3459. Haenszel, W., and Dawson, E. A. (1965). Cancer 28, 14-24. Haenszel, W., and Kurihara, M. (1968). J. Nail. Cancer Inst. 40, 43-68. Haenszel, W., Kurihara, M., Segi, M., and Lee, R. K. C. (1972). J. Natl. Cancer Inst.. 49, 969-988. Haenszel, W., Berg, J. W., Segi, M., Kurihara, M., and Locke, F. B. (1973). J. Nail. Cancer Iiisi. 51, 1765-1779. Haenszel, W., Correa, P., and Cuello, C. (1975). J . Natl. Cancerlnsi. 54, 1031-1035. Hakama, M., and Saxen, E. (1967). In!. J. Cancer 2, 265-268. Hammond, E. C. (1966). Nor. Cancer I n s r . Monogr. 19, 127-204. Hankin, J. H., and Rawlings, V. (1978). A m . J . Clin. Nutr. 31, 2005-2016. Hansson, K. (1967). Acta Chir. Scand. (Suppl.) 375, 1-120. Harris, C. C., Sporn, M. B., Kaufman, D. G., Smith, J. M., Jackson, F. E., and Saffioti, U. (1972). J . Natl. Cancer I n s t . 48, 743-761. Haslam, S. Z., and Bern, H. A. (1977). Proc. Nail. Acad. Sci. U . S . A . 74, 4920-4924. Hawkins, R. A,, Drewitt, D., Greedman, B., Killen, E., Jenner, D. A., and Cameron, E. H. D. (1976). Br. J . Cancer 34, 546-549. Hawkworth, G. M., and Hill, M. J. (1971a). Biochem. J . 122, 28. Hawkworth, G. M., and Hill, M. J. (1971b). Br. J . Cancer 25, 520. Hayward, J . L. (1979). Cancer Res. 24, 1-149. Hayward, J. L., Greenwood, F. C., Glover, G., Stemmermann, G. N., Bulbrook, R. D., Wang, D. Y., and Kumaoka, S. (1978). Eur. J . Cancer 14, 1221-1228.

NUTRITION AND CANCER

337

Hellman, L., Zumoff, B., Fishman, T., and Gallagher, T. F. (1971). J. Clin.Endocrinol. 33, 138-144. Hems, G. (1970). Br. J . Cunccr 24, 226-234.. Hems, G. (1978). Br. J . Cancer 37, 974-987. Hentges, D. J., Maier, B. R., Burton, G. C., Flynn, M. A., and Tsutakawa, R. K. (1977). Cuncer Res. 37, 568-571. Higginson, J. (1975). In “Persons at High Risk of Cancer. An Approach to Cancer Etiology and Control.” (J. F. Fraumeni, Jr., ed.), pp. 385-398. Academic Press, New York. Hilfrich, J., Hecht, S. S., and Hoffman, D. (1977). Cuncer Lett. 2, 169-176. Hill, M. J . (1974). A m . J . Clin. Nutr. 27, 1475-1480. Hill, M. J. (1977). I n “Origins of Human Cancer” (H. H. Hiatt, J. D. Watsow and J. A. Winsten, eds.), pp. 1627- 1640. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York. Hill, M. J., and Hawksworth, G. (1972). In ”N-Nitroso Compounds: Analysis and Formation” (C. P. Bogovski, R. Preussman and E. A. Walker, eds.), p. 116. International Agency for Research on Cancer, Lyon, France. Hill, M. J., Drasar, B. S., Aries, V. C., Crowther, J. S., Hawksworth, G. B., and Williams, R. E. 0. (1971). Lancet I, 95-100. Hill, M. J., Hawksworth, G., and Tattersall, G. (1973). Br. J. Cancer 28, 562-567. Hill, P., and Wynder, E. L . (1976). Luncet 2, 806-807. Hill, P., and Wynder, E. L. (1979). Cancer Lett. 7, 273-282. Hill, P., Wynder, E. L., Helman, P., Hickman, R., and Rona, G. (1976a). Cancer Res. 36, 1883- 1885.

Hill, P., Wynder, E. L., Kumar, H., Helman, P., Rona, G., and Kuno, K. (1976b). Cancer Res. 36, 4102-4106. Hill, P., Chan, P. C., Cohen, L. A., Wynder, E. L., and Kuno, K. (1977). Cancer 39, 1890-1896.

Hill, P., Wynder, E. L., Whitmore, J. F., Garnes, H., and Walker, A. R. P. (1978). Int. Cuncer C o n g r . , 12th. Bumos Aires Abstr. 1, Workshop 1’6. Hill, P., Wynder, E. L., Garbaczewski, L., Garnes, H., and Walker, A. R. P. (1979). Cuncer Res. 39, 5101-5105. Hinds, M. W., Thomas, D. B. and O’Reilly, H . P. (1979). Cuncer 44, 11 14- 1120. Hirayama, T. (1971). Gunn Monogr. 11, 3-19. Hirayama, T. (1975). Cancer ReS. 35, 3460-3463. Hirayama, T. (1978). Prev. Med. 7, 173-175. Hirayama, T. (1979). Nutr. Cancer 1(3), 67-81. Hodges, J. E. (1967). “Symposium on Foods: The Chemistry and Physiology of Flavors” (H. W. Schultz, E. A. Day, and L. M. Libbey, eds.), pp. 465-491. Avi, Westport, Connecticut. Hoffmann, D., Ranieri, R., Hecht, S. S., Maronpot, R., and Wynder, E. L . (1975). J. Natl. Cuncer Inst. 55, 977-981. Hopkins, G. J., and Carroll, K. K. (1979). J . Natl. C‘uncer I n s t . 62, 1009-1012. Hopkins, G. J., and West, C. E. (1976). Life Sci. 19, 1103- 11 16. Hopkins, G. J., West, C. E., and Hard, G. C. (1976). Lipids 2, 328-333. Howell, M. A. (1975). J. Chronic Dis. 28, 67-80. Hunter, A. L., and Neal, R. A. (1975). Biochi.rn. Pharrnacol. 24, 2199-2205. Hutchison, G. (1976). Semin. Oncol. 3, 151-159. IARC Intestinal Microecology Group (1977). Lancet 2, 207-21 1. Jacobs, M. M., Jansson, B., and Griffin, A. C. (1977a). Cancer Lett. 2, 133-138.

338

BANDARU

s. REDDY et al.

Jacobs, M. M., Jansson, B., and Griffin, A. C. (1977b). Cuncer Lett. 2, 319. Jacobson, H. I . (1974). Cuncer Chemothtv. Rcp. 58, 503-51 1. Jansson, B., Seibert, G. B., and Speer, J . F. (1975). CunccJr 36, 2373-2384. Jensen, 0. M., Mosbech, J., Salaspuro, M., and Jhamaki, T. (1974). I n t . J . Epidemiol. 3, 183-186. Joint Iran International Agency for Research on Cancer Study Group (1977). J . Nutl. Cancer Inst. 59, 1127- 1138. Jones, M. K., Ramsat, I. D., Collins, W. P., and Dyer, G. I. (1977). Eur. J . Cmcc,r 13, 1109- 1 1 12. Joossens, J. V., (1979). B r . Med. J . 1, 1145. Judd, H. L., Judd, G. E., Lucas, W. E., and Yen, S. C. C. (1974).J . Clin. Endocrinol. 39, 1020. Jussawalla, D. J . , and Jain, D. K. (1976). “Cancer Incidence in Greater Bombay, 19701972.” Indian Cancer Society-Bombay, Cancer Registry, Bombay. Kamionkowski, M. D., and Fleshler, B. (1965). A m . J. Med. Sci. 249, 696-700. Kater, R. M. H., Roggin, G., Tobin, F., Zieve, P., and Iber, F. L. (1969). A m . J . M e d . Sci. 258, 35-39. Kay, R. M. (1978). A m . J . Clin. Nutr. 31, 562-563. Keller, A. Z. (1967). Cancer 20, 1015- 1022. Keller, A. 2. (1977). A m . J . Epidemiol. 106, 194-202. Kelly, P. A., Ferland, L., and Labrie, F. (1978). I n ”Developments in Endocrinology” (C. Robyn and M. Harter, eds.), Vol. 2, pp. 59-68. Elsevier, Amsterdam. Kern, F., Jr., Birkner, H. J., and Ostrower, V. S. (1978). A m . J . Clin. Nutr. 31, ,9175S179. King, H., Diamond, E., and Lilienfeld, A. M., (1963). J . Chron. Dis. 16, 117- 153. Kmet, J. (1966). Cancer 19, 163-171. Kobayashi, N., Hoffmann, D., and Wynder, E. L. (1974). J . N u t / . C‘uncer Inst. 53, 10851093. Korenman, S. G., Sherman, B. M., and Korenman, J. C. (1978). I n “Clinics in Endocrinology” (G. T. Ross and M. Lipsett, eds.), pp. 625-643. Saunders, London. Kritchevsky, D., and Story, J. A. (1974). J . Nutr. 104, 458-462. Kumaoka, S., Abe, 0.. Utsunomiya, J., Bulbrook, R. D., Hayward, J . L., and Swain, M. D. (1973). I A R C 7, 131-135. Kuoppasalmi, J., Naveri, H., Rehunen, S . , Harkonen, M., and Aldercreutz, H. (1976). J . Stmlid Biochem. 7, 823-829. Kuratsune, M., Kohchi, S . , Horie, A., and Nishizumi, M. (1971). Cunn 62, 395-405. Lacassagne, A., Buu-Hoi, N. P., and Zajdela, F. (1966). Nature (London) 209, 1026-1027. Larsson, L. G., Sandstrom, A., and Westling, P. (1975). Cuncer Res. 35, 3308-3316. Lea, A. J . (1967). Ann. R . Coll. Surg. EngI. 41, 432-438. Lee, F. I . (1966). Cur 7, 77-85. Leevy, C. M . , Gellene, R., and Ning, M. (1964). Ann. N.Y. Acnd. Sci. 114, 1026-1040. Leevy, C. M . , Baker, H., tenHove, W., Frank, O., and Cherrick, G. R . (1965). A m . J . Clin. Nutr. 16, 339-346. Lemon, F. R., and Walden, R. T. (1966). J . A m . Mod. Assoc. 198, 117-126. Lemon, F. R . , Walden, R. T., and Woods, R. W. (1964). Cancer 17, 486-497. Lemon, H. M. (1975). C a n w r R c i . 35, 1341-1353. Lemon, H. M. (1976). C u n c w D C I P L Prevent. I. 1, 263-281. Lemon, H. M., Wotiz, H. H., Parsons, L., and Mozden, P. J. (1966). J . A m . Med. Assoc. 196, 1128- 1136.

NUTRITION AND CANCER

339

Levin, M. L., Haenszel, W., Carroll, B. J., Gerhardt, P. R., Handy, V. H., and Ingraham, S. C. (1960). J. Nutl. C a n c w Inst. 24, 1243-1257. Lijinsky, W., and Taylor, H. W. (1976). Cancer Ras. 36, 1988-1990 Lin, H. J . , Chan, W. C., Fong, L. Y. Y., and Newbern, P. M. (1977). Nutr. R e p . In!. 15, 635-643. Lipkin, M. (1975). Cancer 36, 2319-2344. Lipsett, M. B. (1975). Cancer Res. 35, 3359-3361. Lowenfels, A. B. (1979). Lancet 1, 108. Lyon, J. L., Gardner, J. W., Klauber, M. R., and Smart, C. R. (1977). Cancer 39, 26082618. McClaren and Siasi (1978). as quoted by Cook-Mozaffari, P., Food and Cancer Supplement 16, Naringsforskning, pp. 40-52. McCoy, G. D., and Wynder, E. L. (1979). Cnncer Res. 39, 2844-2850. McCoy, G. D., Chen, C. B., Hecht, S . S . , and McCoy, E. C. (1979). Cancer Res. 39,7937%. Macdonald, I. A., Webb, G. R., and Mahoney, D. C. (1978). A m . J. Clin. Nuir. 31, S233S238. McFadyen, I . J . , Prescott, R. J . , Groom, G. V., Forrest, A. P. M., Colder, M. P., and Fahmy, D. R. (1976). Lance1 1, 1100-1102. McGuire, W. L., Chamness, G. C., Costlow, M. E., and Shepherd, R. E. (1974). Metabolism 23, 75- 100. McGuire, W. L., Hormitz, K. B., Sava, D. T., Garola, R. E., and Chamness, G. C. (1978). Metabolism 27, 487-501. MacIndoe, J . H., and Turkington, R. W. (1972). J. Clin. Invesi. 52, 1973-1978. MacMahon, B., Cole, P., Aoki, K., Lin, T . M., Morgan, R. W., and Wood, N. C. (1971). Luncet 2, 900-902. MacMahon, B., Cole, P., and Brown, J . (1973). J. Nail. Cancer Insi. 50, 21-42. Magee, P. N. (1971). Food Cosrnrt. T o x i c d . 9, 207-218. Magee, P. N., and Barnes, J. M. (1956). Br. J. Cancer 10, 114-122. Mahboubi, E., Kmet, J., Cook, P. J., Day, N. E., Ghadirian, P., and Salmasizadeh, S. (1973). Br. J . Cancer 28, 197-208. Malarkey, W. B., Schroeder, L. I., Stevens, V. C., James, A. G., and Lanese, R. R. (1977a). Cancer Res. 37, 4650-4654. Malarkey, W. B., Schroeder, L. I., Stevens, V. C., James, A. G., and Lanese, R. R. (1977b). Cancer R r s . 37, 4655-5648. Maling, H. M., Stripp, B., Sipes, I. G., Highman, B., Sawl, W., and Williams, M. A. (1975). Toxicol. Appl. Phurmacol. 33, 291-308. Mancuso, T. R. (1974). J. Chronic Dis. 27, 459-474. Mandel, M., Ichinotsubo, D., and Mower, H. (1977). Nature (London) 267, 248-249. Marmorston, J., Crowley, L. G., Myers, S . M., Stern, E., and Hopkins, C. E. (1965a). A m . J . Obstet. Gvnrcol. 92, 460-467. Marmorston, J . , Lombardo, L. J., Myers, S. M., Gierson, H., Stern, E., and Hopkins, C. E. (1965b). J . Urol. 93, 287-295. Marquardt, H., Rufino, F., and Weisburger, J. H. (1977a). Science 196, 1000-1001. Marquardt, H., Rufino, F., and Weisburger, J . H . (1977b). Food Cosmet. Toxicol 15, 97100. Martinez, I.(1969).J. Nafl. Cancer Inst. 42, 1069-1094. Martinez, I . , Torres, R., and Friaz, Z. (1975). Cancer Res. 35, 3265-3271. Mastromarino, A,, Reddy, B. S . , and Wynder, E. L. (1978). Cancer Res. 38,4458-4462.

340

BANDARU

s. REDDY et al.

Mastromarino, W., and Wilson, R. (1976). Radiation Res. 66, 393-400. Matsumoto, T., Yoshida, D., Mizursaki, S., and Okamoto, H. (1977). Murat. Res. 48, 279. Matsuzawa, T., and Wilson, R. (1965). Radiation Res. 25, 15-24. Meier, A. H. (1977). In “Comparative Endocrinology of Prolactin” (H. D. Dellman, T. A. Johnson and D. M. Klachko, eds.), pp. 153-191. Plenum, New York. Meier, A. H., and Burns, J. T. (1976). A m . Zoo/. 16, 649-659. Meites, J. (1977). I n “Comparative Endocrinology of Prolactin” (H. D. Dellman, J. A. Johnson and D. M. Klachko, eds.), pp. 135-142. Plenum, New York. Merimee, T. J., and Fineberg, S. R. (1974). J . Clin. Endocrinol. 39, 385-386. Mertin, J., Shenton, B. C., and Fields, E. J. (1973). Br. Med. J . 2, 777-778. M e s h , J. C., Sacquet, E., and Raiband, P. (1974). Ann. B i d . Anim. Biochem. Biophys. 14, 709-720. Miller, A. B. (1977). Cancer 39, 2704-2708. Miller, A. B. (1978). Cancer Res. 38, 3985-3990. Mirvish, S. S. (1975). Toxicol. Appl. Pharmacol. 31, 325-351. Mirvish, S. S., and Chu, C. (1973). J . Natl. Cancer Inst. 50, 745-750. Misra, P. S., Lefevre, A., Ishii, H., Rubin, E., and Lieber, C. S. (1971). A m . J . Med. 51, 346-351. Modan, B., Lubin, F., Barell, V., Greenberg, R. A., Modan, M., and Graham, S. (1974). Cancer 34, 2087-2092. Modan, B., Barell, V., Lubin, F., and Modan, M. (1975). Cancer Res. 35, 3503-3506. Moolgavkar, S. H . , Stevens, R. G., and Lee, J. A. H. (1979). J. Nail. Cancer Inst. 62, 493-501. Moore, W. E. C., and Holdeman, L. V. (1974). Appl. Microbiol. 27, 961-979. Moore, W. E. C., and Holdeman, L. V. (1975). Cancer Res. 35, 3418-3420. Morgan, R. W., Jain, M., Miller, A. B., Choi, N. W., Matthews, V., Munan, L., Burch, J. D., Feather, J., Howe, G. R., and Kelly, A. (1978). A m . J. Epidemiol. 107, 488498. Mukai, F. M., and Goldstein, B. D. (1976). Science 191, 868-869. Muiioz, N., and Asvall, J. (197411nt. J. Cancer 8, 144-157. Muiioz, N., and Connelly, R. (1971). Inr. J. Cancer 8, 158-164. Muiioz, N., Correa, P., Cuello, C., and Duque, E. (1971). Int. J . Cancer 8, 158-164. Nagao, M . , Honda, M., Seino, Y., Yahagi, T., and Sugimura, T. (1977a). Cancer Lett. 2, 221. Nagao, M., Honda, M., Seino, Y., Yahagi, T., Kawachi, T., and Sugimura, T. (1977b). Cancer Lett. 2, 334. Nagasawa, H. (1978). Eur. J . Cancer 15, 267-279. Narisawa, T., Magadia, N. E., Weisburger, J. H., and Wynder, E. L. (1974). J . Nail. Cancer Inst. 55, 1093- 1097. Narisawa, T., Reddy, B. S., Wong, C. Q . , and Weisburger, J. H. (1976). Cancer Res. 36, 1379-1383. Nature (London) Editorial (1975). Androgens converted to estrogens. Nature (London) 256, 260. Newberne, P. M., and Suphakarn, V. (1977). Cancer 40, 2553-2556. Newill, V . A. (1961). J. Narl. Cancer Inst. 26, 405-417. Niall, H. D., Hogan, M. L., Sauer, R., Rosenblum, I. Y., and Greenwood, F. C. (1971). Proc. Narl. Acad. Sci. U . S . A . 68, 866-869. Nicoll, C. S. (1975). A m . Z o o / . 15, 881-903. , Nigro, N. D., Bhadrachari, N., and Chomchai, C. (1973). Dis. Colon Rectum 16, 438-443.

NUTRITION A N D CANCER

34 1

Nigro, N. D., Singh, D. V., Campbell, R. L., and Pak, M. S. (1975). J. Nail. Cancer I n s t . 54, 429-442. Nigro, N. D., Bull, A. W., Klopfer, B. A., Pak, M. S. and Campbell, R. L . (1979).J. Nut/. Cancer Insr. 62, 1097-1102. Nimrod, A., and Ryan, K. J . (1975). J. Clin. Endocrinol. Mercih. 40, 367-372. Nishizuka, Y. (1978). Prev. Med. 7, 218-224. Nomura, A., Stemmermann, G. N . , Rhoads, G. G., and Glober, G. A. (1975). Hawaii M d . J . 34, 309-316. Nutrition Review Editorial (1975). Nutr. Rcv. 33, 341-343. Oiso, T. (1975). Cancer Res. 35, 3254-3258. Pearson, 0 . H.(1972). I n "Current Research in Oncology" (C. B. Anfinson, M. Potter and A. N. Schechter, eds.), pp. 125- 144. Academic Press, New York. Phillips, F., Crisp, A. H., McGuiness, B., Kalucy, E. C., Chen, C. N., Koval, J., Kalucy, J., and Lacey, J . H. (1975). Lancet 2, 723-725. Phillips, R. L. (1975). Cancer Rcs. 35, 3513-3522. Plescia, 0.J . , Smith, A. H., and Grinwich, K. (1975). Proc. Nut/. Acad. Sci. U . S . A . 72, 1848- 185 1. Poortman, J., Thijssen, 3. H. H., and Schwarz, P. (1973). J. Clin. Endocrind. 37, 101-109. Pour, P., and Althoff, J. (1977). Cancer Lett. 2, 323-326. Pour, P., and Donnelly, T. (1978). Cancer Res. 38, 2048-2051. Pour, P., Althoff, J., Kruger, F. W., Schmahl, D., and Mohr, W. (1976). Canccu Lett. 1, 3-6. Protzel, M., Giardina, A. C., and Albano, E. H. (1964). Oral Surg. 18, 622-634. Radike, M. J . , Stemmer, K. L., Brown, P. B., Larson, E., and Bingham, E. (1977). Environ. Health Persptct. 21, 153-166. Raineri, R., and Weisburger, J. H. (1975). Ann. N . Y . Acad. Sci. 258, 181-189. Ranken, R., Wilson, R., and Bealmear, M. (1971). Proc. Soc. Exp. B i d . Med. 138, 270272. Rao, G. A , , and Abraham, S. (1976). J. Natl. Cuncer I n s t . 56, 431-432. Reddy, B. S . , and Hirota, N . (1979). Fed. Proc. 38, 714. Reddy, B. S ., and W a a n a b e , K. (1978). J. Natl. Cancer I n s f . 61, 1269- 1271. Reddy, B. S . , and Watanabe, K. (1979). Cancer Res. 39, 1521-1524. Reddy, B. S . , and Wynder, E. L. (1973). J. Nurl. Canccr Inst. 50, 1437-1442. Reddy, B. S . , and Wynder, E. L. (1977). Cancer 39, 2533-2539. Reddy, B. S . , Weisburger, J. H., and Wynder, E. L. (1974). Science 183, 416-417. Reddy, B. S . , Narisawa, T., Maronpot, R., Weisburger, J. H., and Wynder, E. L. (1975a). Cuncer Res. 35, 3421-3426. Reddy, B. S., Weisburger, J. H., and Wynder, E. L. (1975b). J . Nutr. 105, 878-884. Reddy, B. S., Narisawa, T., Vukusich, D., Weisburger, J . H., and Wynder, E. L. (1976a). Proc. Soc. Exp. B i d . Med. 151, 237-239. Reddy, B. S . , Narisawa, T., and Weisburger, J. H. (1976b). J . Nut/. Cancer I n s t . 57, 567569. Reddy, B. S . , Narisawa, T., Weisburger, J. H., and Wynder, E. L. ( 1 9 7 6 ~ )J.. N u t / . Cancer I n s t . 56, 441-442. Reddy, B. S . , Mangat, S., Sheinfil, A., Weisburger, J . H., and Wynder, E. L . (1977a). Cancer Res. 37, 2132-2137. Reddy, B. S . , Mangat, S., Weisburger, J. H., and Wynder, E. L. (1977b). Cuncer Res. 37, 3533-3536. Reddy, B. S., Watanabe, K., and Weisburger, J. H. (1977~).Cancer Res. 37, 4156-4159.

342

BANDARU

s.

REDDY et

al.

Reddy, B. S., Watanabe, K., Weisburger, J. H., and Wynder, E. L. (1977d). Ccrncw R e s . 37, 3238-3242. Reddy, B. S., Hedges, A. R., Laakso, K., and Wynder, E. L. (1978a). C a n w r 42, 28322838. Reddy, B. S., Weisburger, J . H., and Wynder, E. L. (1978b). I n “Carcinogenesis” (T. J. Slaga, A. Sivak and R. K. Boutwell, eds.), Vol. 2, pp. 453-464. Raven, New York. Robyn, C. (1975). Piithol. Biol. 23, 783-792. Rogers, A. E., and Newberne, P. M. (1973). Nrrturc, (London) 246, 491-492. Rogers, A. E., Herndon, B. J., and Newberne, P. M. (1973). Cancer Rrs. 33, 1003-1009. Rogers, A. E., Sanchez, O., Feinsod, F. M., and Newberne, P. M. (1974). Cancer R e s . 34, 96-99. Rose, E . R. (1973). J. Ncrtl. Cuncer Inst. 51, 7-16. Rosen, P. O., Ashikari, R., Thaler, H., Ishikawa, S., Hirota, T., Ake, O., Yarnarnoto, H., Beattie, E. J., Urban, J. A., and Mike, V. (1977). Cancer 39, 429-434. Rosencweig, H., Heuson, J. C., Bila, S., Sermite, M., and Robyn, C. (1973). Eur. J. Cancer 9, 657-664. Ross, M. H., and Bras, G. (1965). J. Nutr. 87, 245-260. Rothrnan, K., and Keller, A. Z . (1972). J. Chronic. Dis.25, 711-716. Roy, C. C., Laurendeau, G., Doyon, G., Chartrand, L., and Rivest, M. R. (1975). Proc. Soc. E x p . B i d . M e d . 149, 1000-1004. Rubin, E., and Lieber, C. S. (1968). Scic,nc,i, 162, 690-691. Rubin, E., Hutterer, F., and Lieber, C. S. (1968). Science 159, 1469-1470. Salley, J., and Bryson, W. (1957). J. Dent. Rps. 36, 935-944. Sander, J. (1968). Hoppc, S i y l m Z. Physiol. Chem. 349, 429. Sander, J., Burkle, G., and Schweinsberg, F. (1973). In ”Topics in Chemical Carcinogenesis” (W. Nakahara, S. Takayama, T. Sugirnura, and Odashima, S., eds.), pp. 292312. Univ. Park Press, Baltimore, Maryland. Sander, J., Schweinsberg, F., LaBar, J., Burkle, G., and Schweinsberg, E. (1975). Gann Monogr. 17, 145- 160. Schoental, R., and Bensted, J. P. M. (1969). Br. J. Cuncer 23, 757-764. Schrauzer, G. N., White, D. A., and Schneider, C. J. (1977). Bioinorg. Chem. 7, 23. Segi, M., Kurihara, M., Matsuyama, T., Takano, A., Ito, M., Nagano, Y., Arakura, S., Hatakeyarna, E., and Noda, K. (1965). “Cancer Mortality in Japan (1899- 1962).” Dept. of Public Health, Tobuku Univ. School of Medicine, Sendai, Japan. Segi, M., Kurihara, M., and Matsuyarna, T. (1969). “Cancer Mortality for Selected Sites in 24 Countries, No. 5 (1964-1965).” Dept. of Public Health, Tobuku Univ. School of Medicine, Sendai, Japan. Segi, M., and Kurihara, M. (1972). “Cancer Mortality for Selected Sites in 24 Countries, No. 6 (1966-1967):’ Japan Cancer Society, Tokyo. Seidrnan, H. (1970). Envirun. Rcs. 3, 234-250. Seidman, H., Silverberg, E., and Holleb, A. 1. (1976). Cn-C Ciincer J . Clin. 26, 2. Sharnberger, R. J., and Willis, C. E. (1971). Clin. L d . Sci. 2, 211-221. Shamberger, R. J., Andreone, T. L., and Willis, C. E. (1974). J. Nail. Cnncer Inst. 53, 1771-1773. Shanrnugaratnam, K., and Wee, A. (1973). In “Host Environment Interactions in the Aetiology of Cancer in Man” (R. Doll and I. Vodopiji, eds.), Publ. No. 7. I.A.R.C., Lyon, France. Sherman, B. M., West, J. H., and Korenrnan, S. G . (1976). J . Clin. Endocrinol. Metcihol. 42, 629-636. Sigurjonsson, J. (1966). J . Niitl. Crincer. Inst. 36, 899-907.

NUTRITION AND CANCER

343

Smith, R. D, Hilf, R., and Senior, A. E. (1977). Cancer Res. 37, 595-598. Smith, R. L. (1956). J. Nut/. Cuncer I n s t . 17,459-473. Smithline, F.,Sherman, L., and Kodony, H. D. (1975). N . Engl. J . M e d . 292, 784-792. Sonka, J . , Vitkova, M., Gregorova, I., Tomsova, Z., Hilgertova, J., and Stas, J. (1973). Endokrinologie (Lripzig) 62, 61-68. Spingarn, N . E., and Weisburger, J. H . (1979). Cancer Lett. 7, 259-264. Sporn, M. B., Dunlop, N . M., Newton, D. L., and Mith, J . M. (1976). Fed. Proc. 35, 13321338. Staszewski, J . (1976). “Epidemiology of Cancer of Selected Sites in Poland and Polish Migrants. ” Ballinger, Cambridge, Massachusetts. Staszewski, J., and Haenszel, W. (1965). J. Natl. Cancer Inst. 35, 291-297. Staszewski, J., McCall, M. G., and Stenhouse, N . S. (1971). Br. J . Cancer 25, 599-610. Stemmermann, G. N. (1977). Gann 68, 525-535. Stenback, F. (1969). Actci. Puthol. Microhiol. Scand. 77, 325-326. Stewart, H . L., Snell, K. C., Morris, H. P., Wagner, B. P., and Ray, F . (1961). Nail. Cuncer Inst. Monogr. 5, 105-139. Story, J. A., and Kritchevsky, D. (1978). A m . J. Clin. Nutr. 31, S199-S202. Sugimura, T., and Kawachi, T. (1973). In “Methods in Cancer Research” (H. Busch, ed.), pp. 245-308. Academic Press, New York. Sugimura, T., and Kawachi, T. (1978). In “Gastrointestinal Tract Cancer” (M. Lipkin and R. A. Good, eds.), pp. 327-342. Plenum, New York. Sugimura, T., Kawachi, T., Minako, N., Takie, Y., Seino, Y., Okamoto, T., Shudo, K., Kosuge, T., Tsuji, K. Wakabayashi, K., Iitaka, Y., and Itai, A . (1977). Proc. Jn. Acad. S c . 53, 58. Sutton, J . R., Coleman, M. J., Casey, J., and Lazarus, S. (1973). Br. Med. J . 1, 520-522. Tan, W. C., Privett, 0 . S., and Goldyne, M. E. (1974). Cuncer Res. 34, 3229-3331. Tannenbaum, A. (1940). Arch. Pathol. 30, 509-517. Tannenbaum, A. (1942). Cancer Res. 2, 468-475. Tannenbaum, A,, and Silverstone, H. (1957). I n “Cancer” (R. W. Raven, ed.), Vol. 1, pp. 306-334. Butterworth, London. Tannenbaum, S. R., Archer, M., Wishnok, J . S., and Bishop, W. W. (1978a). J. Nut/. Cuncer Ins/. 60, 251-253. Tannenbaum, S. R., Felt, D., Young, V. R., Land, P. C., and Bruce, W. R. (1978b). Science 200, 1487- 1489. Tannenbaum, S. R., Moran, D., Rand, W., Cuello, C., and Cornea, P. (1979). J. Natl. Cuncer I n s t . 62, 9- 12. Tatematsu, M., Takahasi, M., Fukushima, S., Hananouchi, M., and Shirai, T. (1975). J. Nrirl. Cuncer Inst. 55, 101-104. Teppo, L., Hakama, M., Hakulinen, T., Lehtonen, M., and Saxen, E. (1975). “Cancer in Finland 1953- 1970: Incidence, Mortality, Prevalence.” Munksgaard, Copenhagen. Tominaga, T., Tei, N., Kitamura, M.,Taguelie, T., Kato, Y., and Sato, S. (1975). Gann 66, 305-310. Topping, D. C., and Visek, W. J . (1976). J. Nutr. 106, 1583-1590. Tulinius, H. (1979). Nutr. Cuncer 1, 61-69. Tuyns, A. J. (1970).In “Cancer Morbidity and Mortality Data in the USSR” (A. F . Serenko and A. A. Romenski, eds.). IARC Internal Technical Report 70/003. Lyon, France. Tuyns, A. J . , and Masse, L. M. F. (1973). In!. J . Epidemiol. 2, 242-245. Tuyns, A. J . , Pequinot, G., and Jensen, 0 . M. (1977). Bull. Cancer 65, 59-64. Van der Gugten, A. A,, and Verstraeten, A. A. (1973). I n “Methods in Cancer Research” (H. Busch, ed.), Vol. 10, pp. 161-200. Academic Press, New York.

344

BANDARU

s. REDDY et al.

Van Soest, P. J. (1978). A m . J . Clin. Nurr. 31, 512s-520s. Vaughn, F. L., and Bernstein, I. A. (1976). Mol. Cell. Biochem. 12, 171-179. Vekemann, M., and Robyn, C. (1975). Er. Med. J . 4,738-739. Visek, W. J., Clinton, S. K., and Truex, C. R. (1978). Cornell Vet. 68, 3-39. Vitale, J . J., and Coffey (1971). I n “The Biology of Alcoholism” (B. Kissin and H. Begleiter, eds.), Vol. 1 , pp. 327-352. Plenum, New York. Wade, A. P., Tweedie, A. C. R., Davis, J. C., Clarke, C. A., and Haggart, B. (1977). Lancer 1, 853-857. Walters, R., Baird, I. M., and Davis, P. S. (1975). Br. Med. J . 2, 536-538. Wang, D. Y., Bulbrook, R. D., and Hayward, J. L. (1977). Eur. J . Cancer 13, 187-192. Warwick, G. P., and Harrington, J. S. (1973). A d v . Cancer Res. 17, 81-229. Watanabe, K., Reddy, B. S., Weisburger, J. H., and Kritchevsky, D. (1979). J . Natl. Cancer Insr. 63, 141- 145. Waterhouse, J., Muir, C., Correa, P., and Powell, J., eds. (1976). “Cancer Incidence in Five Continents,” Vol. 111. International Union against Cancer, Berlin and New York. Wattenberg, L. W. (1975). J . Nail. Cancer Inst. 54, 1005-1006. Wattenberg, L. W. (1978). In “Advances in Cancer Research” (G. Klein and S. Weinhouse, eds.), Vol. 26, pp. 197-226. Academic Press, New York. Weir, J . M., and Dunn,J . E., Jr. (1970). Cancer 25, 105- 112. Weisburger, E. K., Evarts, R. P., and We&, M. L. (1977). Food Cosmet. Toxicol. 15, 139. Weisburger, J . H. (1973). Seventh National Cancer Conference Proceedings, pp. 456-473. Lippincott, Philadelphia, Pennsylvania. Weisburger, J. H., and Raineri, R. (1975). Toxicol. Appl. Pharmacol. 31, 369-374. Weisburger, J. H., and Spingarn, N. E. (1979). In “Naturally Occurring CarcinogensMutagens and Modulators of Carcinogenesis” (J. A. Miller, E. C. Miller, T. Sugimura, T. Takayama and I. Hirono, eds.) in press. Univ. Park Press, Baltimore, Maryland. Weisburger, J. H., and Williams, G. (1975). In “Cancer: A Comprehensive Treatise” (F. D. Becker ed.), Vol. I. pp. 185-234. Plenum, New York. Weisburger, J. H., Reddy, B. S., and Joftes, D. L., eds. (1975). “Colorectal Cancer.” International Union against Cancer, Geneva. Weisburger, J. H., Reddy, B. S., Spingarn, N. E., and Wynder, E. L. (1980). In “Progress in Cancer Research” (S. J. Winawer, P. Sherlock and D. Schottenfeld, eds.) in press. Raven, New York. Weisburger, J. H., Marquardt, H., Hiroto, N., Mori, H., and Williams, G. M. (1980a). J . Natl. Cancer I n s r . 64, 163-167. Weisburger, J. H., Marquardt, H., Mower, F., Hirota, N., Mori, H., and Williams, G. (1980b). Prev. Med. (in press). Welsch, C. W. (1978). Cancer Res. 38, 4054-4058. Welsch, C. W., and Nagasawa, H. (1977). Cancer Res. 37, 951-963. Williams, R. R., and Horn, J. W. (1977). J . Narl. Cancer Inst. 58, 525-541. Wilson, R. B., Hutcheson, D. P., and Wideman, L. (1977). A m . J . Clin. Nurr. 30. 176-181. Wuttke, W., Dohler, K. D., and Gelato, M. (1976). J . Endocrinol. 68, 391-396. Wynder, E. L. (1975a). Cancer Res. 35, 2220-2225. Wynder, E. L. (1975b). Cancer Res. 35, 3388-3394. Wynder, E. L., and Fryer, J. H. (1958). Ann. I n i . Med. 49, 1106- 1128. Wynder, E. L., and Gori, G. B. (1977). J. Natl. Cancer Inst. 58, 825-832. Wynder, E. L., and Hirayama, T. (1977). Prev. Med. 6, 567-594. Wynder, E. L., and Klein, V. E. (1965). Cancer 18, 167-180. Wynder, E. L., and Reddy, B. S. (1973). J . Narl. Cancer Insr. 50, 1099- 1106. Wynder, E. L., and Shigematsu, T. (1967). Cancer 20. 1520-1561.

NUTRITION A N D CANCER

345

Wynder, E. L., and Stellman, S. D. (1977). Cancrr Rrs. 37, 4608-4622. Wynder, E. L . , Bross, I. J . , and Day, E. (1956). J. A m . M i d . Assoc. 160, 1384-1391. Wynder, E. L . , Bross, I. J., and Feldman, R. M. (1957). Cunccr 10, 1300-1323. Wynder, E. L., Kmet, J . , Dungal, N., and Segi, M. (1963). Cuncer 16, 470-487. Wynder, E. L., Hyams, L., and Shigematsu, T. (1967). Cunccr 20, 113- 126. Wynder,.E. L., Kajitani, T., Ishakawa, S., Dodo, H . , and Takano, A. (1969). Cancer 23, 1219-1220. Wynder, E. L., Mabuchi, K., and Whitmore, W. F., Jr. (1971). Cuncer 28, 344-360. Wynder, E. L., Mabuchi, K., Maruchi, N., and Fortner, J . G. (1973). J . N u / / . CancerIns/. 50, 645-667. Zaldivar, R. (1970) Z. Kri~hsforsch.75, 1- 13, Zaldivar, R . (1978). Z. Krehsforsch. 92, 215-216. Zaldivar, R., and Robinson, H. Z. (1973). Z. Krehsforsch. 80, 289-295. Zinder, 0.. Hamosh, M., Fleck, T. R. C., and Scow, R. 0. (1974). A m . J . Physiol. 226, 744-748. Zumoff, B., Fishman, J., Bradlow, H. L., and Hellman, L. (1975). C'uncc.r Res. 35, 33653373.

This Page Intentionally Left Blank

SUBJECT INDEX A

Adenomatous polyps, fecal constituents and, 255-257 Alcohol, cancer epidemiology and, 283, 286, 291 Alimentary tract cancer, dietary factors in, 282-291 Androgens, breast cancer risk and, 313 Anti-gelatin factor (AGF), CIG compared to, 151

C

Calcium cells stimulated by, shedding from, 119 mechanism, 147 cGMP, growth and, 123- 125 metabolism, role in cell membrane structure, 84-85 Cancer coagulation in, 153- 155 fibrinolysis in, 155 nutrition and, ,237-345 sites of, mortality ratios for, 239, 240 B Cancer cells B lymphocytes, shedding from, 130- 131 shedding from, 75- 199 BHA, role in colonic cancer inhibition, 269 cell phenotype and, 113- I14 Bile acids consequences, 148- 172 in feces, dietary fat, and, 259 invasion and metastases, 156- 161 in carcinogenesis, 264-266 malignancy, 149- 150 as tumor promoters, 16, 19 Carcinogenesis Bladder, two-stage carcinogenesis in, 31 promotion of, 1-74 Blebs, role in shedding, 147- 148 two-stage, 4-6 Blocking factors cell cultures, 33-35 immune complexes as, 164- 166 models, 16-31 tumor antigen and, 163- 164 viral, tumor promotion and, 60-62 Blood elements, tumor promoter effects Carcinogens, role in bowel cancer, 249on, 51 257 Bowel cancer, see Colon cancer Cell(s) Breast cancer activation of, 104- 105 androgen levels in risk for, 3 13-3 14 biochemical events of, 121- 125 case control studies of, 300-302 cancer, see Cancer cells correlation analyses of, 297 differentiation, altered, in tumor dietary factors in, 295-324 promotion, 45-51, 59-60 fat, 303-312 morphology of, phorbol diester effects epidemiology of, 295-302 on, 35-38 etiology of, 318-323 Cell cultures, tumor-promoter studies on, experimental studies on, 302-3 12 32-54 metabolic epidemiology of, 312-318 Cell membrane(s) migrant studies on, 295-2% calcium metabolism in, 84-85 obesity and, 302-303 interaction of, in tumor induction, 56 prolactin levels in risk for, 316-318 phorbol diester effects on, 26-27 in special populations, 300 protein factors maintaining structure of, time-trend analyses of, 298-300 79-85 Butylated hydroxytoluene (BHT) as tumor protein release from, 85-99 promoter, 16, 19 energy requirements, 93-95 347

348

SUBJECT INDEX

exocytosis, 88-92 gl ycoproteins, %- 97 secretion, 85-88 stimulus secretion coupling, 92 protein turnover in, 101- 104 shedding from, 75-199 structure of, 78-85 maintenance, 80-84 tumor promoter effects on, 38-40 Cell transformation, definition of, 77 cGMP, calcium, growth, and, 123- 125 Chalones, phorbol diester effects on, 24 Chemotaxis, surface proteases and, 174177 Cholesterol metabolites, bowel-cancer incidence and, 255-257 Chromosomes, exchanges among, tumor promoter effects on, 52 Cigarette smoke condensates, as tumor promoters, I5 Coagulation, in cancer, 153- 155 Cold insoluble globulin (CIG), as circulating form of fibronectin, 150151 Collagenase, role in tumor invasion, 157158 Colon cancer colon-rectal differences in, 243-244 dietary factors in, 241-271 epidemiology of, 241-245 etiology of, 245-248 metabolic etiology of, 249-257 migrant studies on, 244 miscellaneous dietary factors in, 268-271 mutagen role in, 266-268 religious differences and, 244-245 socioeconomic status and, 244 Colony stimulating factor (CSF), shedding of, 139-140 Corticosteroids, in prevention of shedding, 178 Croton oil, tumor promotors from, 6-7, 16 Cryofibrinogenemia, cell-shedding role in, 151- I53 Cyclamate, as tumor promoter, 16, 18 Cyclic nucleotides, phorbol diester effects on, 24-25

D

Daphnane esters as tumor promoters, 13 biological effects, 53 DDT, as tumor promoter, 16, 19 Deoxycholic acid, as tumor promoter, 16 Desquamation, see Shedding Diet, cancer and, 237-345 Disulfiram, in inhibition of colon carcinogenesis, 269 Diterpene esters, as tumor promoters, 13, 54 DNA repair, tumor promoter effects on repair of, 44-45 replicative, tumor promoter effects on synthesis of, 42-44 DNA oncogenic viruses, TATA from neoplasms induced by, 216-219 E

Egg fertilization, cell shedding in, 126- 127 Embryogenesis, cell shedding in, 127 Endoskeleton, of cell, structural features, 80-83 Energy requirements, of cell-surface secretion, 93-95 Enzymes, shed from tumors, 168- 171 Epidermal growth factor (EGF), as mitogen, cell shedding from, 119 Estrogens, levels of, breast cancer risk and, 314-316 Euphorbia esters, as tumor promoters, 13 Exocytosis, membrane fusion and, 88-92 Exfoliation, see Shedding Exoskeleton, of cell, structural features, 83-84 F

Fat (dietary) breast cancer and, 303-312 large-bowel cancer and, 245-247 experimental studies, 257-260 Fatty acid(s) oxygenated as tumor promoters, 16, 18 unsaturated as tumor promoters, 14- 15

349

SUBJECT I N D E X Feces, composition of, bowel cancer and, 250-257 Fertilization, of eggs, cell shedding in, 126- 127 Fiber, dietary, in colon carcinogenesis, 261-264,271 Fibrinolysis, in cancer, 155 Fibronectin current model of, 111 lack of, in cancer cells, 110-112 role in cancer-cell shedding, 150-151 tumor promoter effects on, 41, 54 Fish, browning of, mutagens from, 267 Flame retardants, as tumor promoters, I5 I-Fluoro-2,4-dinitrobenzene, as tumor promoter, 15 Fucosyl transferase, shed from tumors, 170

G G protein, of viral membranes, synthesis of, 97-99 Galactosyl transferase, shed from tumors, 170- 171 Genes, altered expression of, in tumor promotion, 56-59 Glycoproteins of cell membrane synthesis of, %-97 turnover, 101 in membranes tumor promoter effects on, 42 shedding of, mechanism, 144- 147 Glycosyl transferases, in normal and cancer cells, 168-172 Gnidia esters, as tumor promoters, 13 Growth, calcium, cGMP, and, 123- 125 Growth factors cells stimulated by, shedding from, 119 shedding of, 139- 140 H

H-2 antigens, TATA antigens relationship to, 211-214 Hamburgers, frying of, mutagens from, 267 Helper factors, in immune system, shedding of, 132- 134

Hepatitis B virus, shedding by, 172, 173 Herpesvirus, induction of synthesis of, tumor promoter effects on, 52-53 Histidase, phorbol diester effects on, 24 Hydrocarbons, nonaromatic, as tumor promoters, 15 Hyperplasia, role in tumor induction, 55 I

l a antigens, shedding from, 131-132, 134-

135 Immune system, cells of, shedding from, 130- 140 Inflammation, role in tumor induction, 55 Ingenane esters, as tumor promoters, 13 Interferon, in prevention of shedding, 178 Invasion, of cancer cells, shedding and, 156- 157 Iodoacetic acid, as tumor promoter, I5 J

Jews, pancreatic cancer in, 293 L

LETS protein, see Fibronectin Leukemia( s) compounds promoting, 16 TATA from, 219-225 Liver, two-stage carcinogenesis in, 29-31 Lymphocytic choriomeningitis virus, shedding in, 172, 173 Lymphoma cells, shedding by, 135 Lytic viruses, cells infected by shedding in, 143 M M protein, of viral membranes, synthesis of, 97-99 Macromolecular synthesis, phorbol diester effects on, 22-23 Macrophage growth factor (MGF), shedding of, 139 Macrophages shedding from, 135- 136 activation and, 136

350

SUBJECT INDEX

Malignancy, of cancer, shedding role in, 149- 150 Meat, browning of, mutagens from, 267 Membranes, structures of, role in shedding, 147- 148 Metastases, of cancer, shedding role in, 156- 157 Meth A, TATA from, 205-208 Mezerein as tumor promoter, 13 biological effects, 53 Microvilli, role in shedding, 147- 148 Mitogens, cells stimulated by, shedding from, 116-121 Mitotic cycle proteases and, 109 shedding and, 112- 113 Mormons, large-bowel cancer in, 245 Mouse, skin tumor promoters for, 14, 16-21 modifiers, 27-28 Mutagens, role in colon carcinogenesis, 266-268 Mutation, tumor promoter effects on, 52 N

Nitrates and nitrites, stomach cancer and, 275-282 Nucleic acid synthesis, phorbol diester effects on, 22 Nutrition cancer and, 237-345 of alimentary tract, 282-291 of breast cancer, 295-324 of large bowel, 241-271 of pancreas, 291-294 of prostate, 324-329 of respiratory tract, 282-291 of stomach, 271-282

0 Obesity, breast cancer and, 302-303 Oncoviruses cells infected by, shedding from, 142143 shedding from, 172 structure and replication of, 140- 141

Ornithine decarboxylase, phorbol diester effects on, 23-24, 40-41 Ovulation, cell shedding in, 125-126 P

P53 antigen, as transformation-related antigen, 208-21 1 Pancreatic cancer dietary factors in, 291-294 epidemiology of, 291-293 etiology of, 293 experimental studies on, 294 metabolic epidemiology of, 293-294 Parsis, large-bowel cancer in, 245 Phenobarbital, as tumor promoter, 16, 18 Phenolic compounds, as tumor promoters, 15 Phorbol and phorbol diesters biochemical effects of, 21-27, 38-45 biological effects of, 17-21, 45-53 cellular interaction and metabolism of, 12- 13 effects on cell morphology, 35-38 as tumor promoters, 7- 12, 16, 18 Phosphorylation, of membrane proteins, 90 Photoreceptor cells, shedding from, 127130 Pimelea esters, as tumor promoters, 13 Plasminogen activator (PA) activation and release of, 137- 139 from cancer cells, 106- 109 fibrinolysis and, 155 tumor promoter effects on, 41-42 Polychlorinated biphenyls, as tumor promoters, 16, 19 Polyoma virus, TATA from neoplasms induced by, 216-219 Prolactin, levels of, breast cancer risk and, 316-318 Prostaglandins, phorbol diester effects on, 25- 26 Prostate cancer dietary factors in, 324-329 epidemiology of, 324-326 etiology of, 326 experimental studies on, 329 metabolic epidemiology of, 326-329 Proteases in cancer cells, nature of, 109- 110

35 1

SUBJECT INDEX

cell shedding of, 106- 109 mitotic cycle and, 109 phorbol diester effects on, 24 Proteins dietary, in colon carcinogenesis, 260-261 shedding of, mechanism, 144- 147 Protein synthesis, phorbol diester effects on. 22-23 R

Resinifernol esters, as tumor promoters, 13 Respiratory tract cancer, 282-291 dietary factors in, 282-291 epidemiology of, 282-283 etiology of, 283-286 metabolic epidemiology of, 287-290 Rheumatoid arthritis, chemotaxis, shedding, and 174- 177 RNA tumor viruses, TATA from neoplasms induced by, 219-225 Rods (photoreceptor), shedding from, 127130

S Saccharin, as tumor promoter, 16, 18, 54 Salty foods, stomach cancer and, 275, 279 Second messenger, interaction of, in shedding process, 122 Secretion, of cell-membrane proteins, 8588 Selenium, role in large-bowel cancer, 247248, 269 Seventh-Day Adventists, large-bowel cancer in, 245 Shedding, from normal and cancer cells, 75- 199 activated specific cells and, 125-144 cancer cell phenotype and, 113- 114 cell activation and, 104- 105 characteristics of, 99- 100 coagulation, fibrinolysis, and cancer, 153- 155 in cryofibrinogenemia, 151- 153 of fibronectin, 110-112, 150-151 glycosyl transferase role in, 168- 172 by immune-system cells, 130- 140 interactions in, 136- 137 nature of material released by, 134 malignancy and, 149- 151

mechanism of, 144- 148 membrane structures in, 147- 148 in metastases, and invasion, 157- 159 mitotic cycle and, 112-113 from mitogen-stimulated cells, 116- 121 in ovulation, 125- 126 prevention of, 178- 179 of proteases, 106- 109 chemotaxis and, 174- 177 secretion with, 121 structures shed from, 114- 116 synthesis coupling and, 120- 121 in tumor immunity, 161- 168 in viral disease, 172- 173 from virus-infected cells, 140- 144 Sialyl transferase, role in tumor shedding, 168- 170 Skin tumor promoters for, 14- 15 two-stage carcinogenesis in, 29 Smoking, pancreatic cancer and, 293 Stimulus-secretion coupling, exocytosis in, 93 Stomach cancer dietary factors in, 271-282 epidemiology of, 271-277 experimental studies on, 279-281 metabolic epidemiology of, 277-279 Suppressor activity, of tumor antigen, 166168 Suppressor factors, in immune system, shedding of, 132- 134 Surfactants, as tumor promoters, 15 SV40 virus, TATA from neoplasms induced by, 216-219 Sweetening agents, as tumor promoters, 16, 18, 54 T

T cells, shedding from, 132-134 TATA antigen, 201 H-2 antigens relationship to, 21 1-214 isolation of, 214-215, 223-225 from leukemias induced by RNA tumor viruses, 219-225 from virus-induced neoplasms, 216-219 Thrombin, cells stimulated by, shedding from, 119- 120 T L antigen, 201

352

SUBJECT INDEX

Tobacco, cancer epidemiology and, 283285, 291 TPA, as tumor promoter, 16 DL-Tryptophan, as tumor promoter, 19 Tumor( s) antigens of, role in shedding, 163- 164 escape mechanisms for, 162- 163 progression of, 5 promotion of, see Tumor promoters Tumor antigens, 201-235 isolation of, 205, 221-225 soluble, 203-215 immune deviations and, 225-229 types of, 201 Tumor immunity, role in shedding of, 161168 Tumor promoters, 1-74 biochemical mechanisms of, 54cells stimulated by, shedding from, 117119 chemistry of, 6- 16 from croton oil, 6-7 experimental models of, 16-54 mouse kin, 16-28 phorbol diesters as, 7- 12

V

Vesicles, role in shedding, 147- 148 Vesicular stomatitis virus cells infected by, shedding in, 143 as model for membrane biosynthesis, 97- 99 Viruses(a1) carcinogenesis, tumor promotion and, 60-62 cells infected by, shedding from, 140144 cell transformations induced by, tumor promoter effects on, 53 disease, chronic, shedding in, 172- 178 Visual cell, diagram of, 128 Vitamin A, role in colonic cancer inhibit ion, 268- 269 Vitamin C in inhibition of colon and gastric carcinogenesis, 269-270, 282

2

Zeiosis, definition of, 114

CONTENTS OF PREVIOUS VOLUMES Volume 1 Electronic Configuration and Carcinogenesis C. A . Coulson Epidermal Carcinogenesis E . V . Cowdry The Milk Agent in the Origin of Mammary Tumors in Mice L . Dmochowski Hormonal Aspects of Experimental Tumorigenesis T. U.Gardner Properties of the Agent of Rous No. 1 Sarcoma R . J . C . Harris Applications of Radioisotopes to Studies of Carcinogenesis and Tumor Metabolism Charles Heidelberger The Carcinogenic Aminoazo Dyes James A . Miller and Elizabeth C. Miller The Chemistry of Cytotoxic Alkylating Agents M . C . J . Ross Nutrition in Relation to Cancer Albert Tannenbaum and Herbert Silverstone

Plasma Proteins in Cancer Richard J . Winder AUTHOR INDEX-SUBJECT INDEX

Volume 2 The Reactions of Carcinogens with Macromolecules Peter Alexander Chemical Constitution and Carcinogenic Activity G. M. Badger Carcinogenesis and Tumor Pathogenesis I. Berenblum

Ionizing Radiations and Cancer Austin M . Brues Survival and Preservation of Tumors in the Frozen State James Craigie Energy and Nitrogen Metabolism in Cancer Leonard D.Fenninger and G. Burroughs Mider Some Aspects of the Clinical Use of Nitrogen Mustards Calvin T. KIopp and Jeanne C. Bateman Genetic Studies in Experimental Cancer L . W . Law The Role of Viruses in the Production of Cancer C . Oberling and M. Guerin Experimental Cancer Chemotherapy C . Chester Stock AUTHOR INDEX-SUBJECT INDEX

Volume 3 Etiology of Lung Cancer Richard Doll The Experimental Development and Metabolism of Thyroid Gland Tumors Harold P. Morris Electronic Structure and Carcinogenic Activity and Aromatic Molecules: New Developments A . Pullman and B. Pulltnan Some Aspects of Carcinogenesis P. Rondoni Pulmonary Tumors in Experimental Animals Michael B. Shimkin Oxidative Metabolism of Neoplastic Tissues Sidney Weinhouse AUTHOR INDEX-SUBJECT INDEX

353

354

CONTENTS OF PREVIOUS VOLUMES

Volume 4 Advances in Chemotherapy of Cancer in Man Sidney Farber, Rudolf Toch, Edward Manning Sears, and Donald Pinkel The Use of Myleran and Similar Agents in Chronic Leukemias D. A. G . Galton The Employment of Methods of Inhibition Analysis in the Normal and TumorBearing Mammalian Organism Abraham Goldin Some Recent Work on Tumor Immunity P. A . Gorer Inductive Tissue Interaction in Development Clifford Grobstein Lipids in Cancer Frances L. Haven and W . R . Bloor The Relation between Carcinogenic Activity and the Physical and Chemical Properties of Angular Benzacridines A . Lacassagne, N . P . BuuHol, R . Daudel, and F. Zajdela The Hormonal Genesis of Mammary Cancer 0 . Muhlbock AUTHOR INDEX-SUBJECT INDEX

Volume 5 Tumor-Host Relations R. W . Begg Primary Carcinoma of the Liver Charles Berman Protein Synthesis with Special Reference to Growth Processes both Normal and Abnormal P . N. Campbell The Newer Concept of Cancer Toxin Waro Nakuhara and Fumiko Fukuoka Chemically induced Tumors of Fowls P. R. Peacock Anemia in Cancer Vincent E. Price and Robert E. Greenfield Specific Tumor Antigens L. A . Zilber

Chemistry, Carcinogenicity, and Metabolism of 2-Fluorenamine and Related Compounds Elizabeth K . Weisburger and John H . Weisburger AUTHOR INDEX-SUBJECT INDEX

Volume 6 Blood Enzymes in Cancer and Other Diseases Oscar Bodansky The Plant Tumor Problem Armin C. Braun and Henry N . Wood Cancer Chemotherapy by Perfusion Oscar Creech, Jr. and Edward T. Krementz Viral Etiology of Mouse Leukemia Ludwick Gross Radiation Chimeras P . C. Koller, A. J . S. Davies, and Sheila M . A . Doak Etiology and Pathogenesis of Mouse Leukemia J . F. A . P . Miller Antagonists of Purine and Pyrimidine Metabolites and of Folic Acid G . M . Timmis Behavior of Liver Enzymes in Hepatocarcinogenesis George Weber AUTHOR INDEX-SUBJECT INDEX

Volume 7 Avian Virus Growths and Their Etiologic Agents J. W . Beard Mechanisms of Resistance to Anticancer Agents R . W . Brockman Cross Resistance and Collateral Sensitivity Studies in Cancer Chemotherapy Dorris J . Hutchison Cytogenic Studies in Chronic Myeloid Leukemia W . M. Court Brown and lshbel M. Tough

CONTENTS OF PREVIOUS VOLUMES

Ethionine Carcinogenesis Emmanuel Farber Atmospheric Factors in Pathogenesis of Lung Cancer Paul Kotin and Hans L. Falk Progress with Some Tumor Viruses of Chickens and Mammals: The Problem of Passenger Viruses G. Negroni AUTHOR INDEX-SUBJECT INDEX

Volume 8 The Structure of Tumor Viruses and Its Bearing on Their Relation to Viruses in General A . F. Howatson Nuclear Proteins of Neoplastic Cells Harris Busch and William J . Steele Nucleolar Chromosomes: Structures, Interactions, and Perspectives M . J . KoDac and Gladys M . Mateyko Carcinogenesis Related to Foods Contaminated by Processing and Fungal Metabolites H . F. Kraybill and M . B. Shimkin Experimental Tobacco Carcinogenesis Ernst L . Wynder and Dietrich Hoffman AUTHOR INDEX-SUBJECT INDEX

Volume 9 Urinary Enzymes and Their Diagnostic Value in Human Cancer Richard Stambaugh and Sidney Weinhouse The Relation o f the Immune Reaction to Cancer Louis V . Caso Amino Acid Transport in Tumor Cells R . M . Johnstone and P. G. Scholefeld Studies on the Development, Biochemistry, and Biology of Experimental Hepatomas Harold P. Morris Biochemistry of Normal and Leukemic

355

Leucocytes, Thrombocytes, and Bone Marrow Cells I . F. Seitz AUTHOR INDEX-SUBJECT INDEX

Volume 10 Carcinogens, Enzyme Induction, and Gene Action H . V . Gelboin I n Vitro Studies on Protein Synthesis by Malignant Cells A . Clark GrifJin The Enzymatic Pattern of Neoplastic Tissue w, Eugene Knox Carcinogenic Nitroso Compounds P. N . Magee and J . M . Barnes The Sulfhydryl Group and Carcinogenesis J . S . Harrington The Treatment o f Plasma Cell Myeloma Daniel E. Bergsagel, K. M . Griffith, A . Haut, and W . J . Stuckley, Jr. AUTHOR INDEX-SUBJECT INDEX

Volume 11 The Carcinogenic Action and Metabolism of Urethan and N-Hydroxyurethan Sidney S . Mirvish Runting Syndromes, Autoimmunity, and Neoplasia D. Keast Viral-Induced Enzymes and the Problem of Viral Oncogenesis Saul Kit The Growth-Regulating Activity of Polyanions: A Theoretical Discussion of Their Place in the Intercellular Environment and Their Role in Cell Physiology William Regelson Molecular Geometry and Carcinogenic Activity of Aromatic Compounds. New Perspectives Joseph C . Arcos and Mary. F. Argus AUTHOR INDEX-SUBJECT INDEX CUMULATIVE INDEX

356

CONTENTS OF PREVIOUS VOLUMES

Volume 12 Antigens Induced by the Mouse Leukemia Viruses G. Pasternak Immunological Aspects of Carcinogenesis by Deoxyribonucleic Acid Tumor Viruses G. I . Deichman Replication of Oncogenic Viruses in VirusInduced Tumor Cells-Their Persistence and Interaction with Other Viruses H. Hanafusa Cellular Immunity against Tumor Antigens Karl Erik HellstrQm and Ingegerd Hellstrorn Perspectives in the Epidemiology of Leukemia Irving L. Kessler and Abraham M. Lilienfeld AUTHOR INDEX-SUBJECT INDEX

Volume 13 The Role of Immunoblasts in Host Resistance and Immunotherapy of Primary Sarcomata P. Alexander and J. G . Hall Evidence for the Viral Etiology of Leukemia in the Domestic Mammals Oswald Jarrett The Function of the Delayed Sensitivity Reaction as Revealed in the Graft Reaction Culture Haim Ginsburg Epigenetic Processes and Their Relevance to the Study of Neoplasia Gajanan V . Sherbet The Characteristics of Animal Cells Transformed in Vitro Ian Macpherson Role of Cell Association in Virus Infection and Virus Rescue J . Svoboda and I. Hloianek Cancer of the Urinary Tract D. B. Clayson and E. H . Cooper

Aspects of the EB Virus M. A. Epstein AUTHOR INDEX-SUBJECT INDEX

Volume 14 Active Immunotherapy Georges Mathe The Investigation of Oncogenic Viral Genomes in Transformed Cells by Nucleic Acid Hybridization Ernest Winocour Viral Genome and Oncogenic Transformation: Nuclear and Plasma Membrane Events George Meyer Passive Immunotherapy of Leukemia and Other Cancer Roland Motta Humoral Regulators in the Development and Progression of Leukemia Donald Metcalf Complement and Tumor Immunology Kusuya Nishioka Alpha-Fetoprotein in Ontogenesis and Its Association with Malignant Tumors G. I . Abelev Low Dose Radiation Cancers in Man Alice Stewart AUTHOR INDEX-SUBJECT INDEX

Volume 15 Oncogenicity and Cell Transformation by Papovavirus SV40: The Role of the Viral Genome J . S. Butel. S . S . Tevethia, and J . L. Melnick Nasopharyngeal Carcinoma (NPC) J . H. C. Ho Transcriptional Regulation in Eukaryotic Cells A . J. MacGillivray, J . Paul, and G. Threlfall Atypical Transfer RNA’s and Their Origin in Neoplastic Cells Ernest Borek and Sylvia 1. Kerr

CONTENTS OF PREVIOUS VOLUMES Use of Genetic Markers to Study Cellular Origin and Development of Tumors in Human Females Philip J . Fialkow Electron Spin Resonance Studies of Carcinogenesis Harold M . Swartz Some Biochemical Aspects of the Relationship between the Tumor and the Host V . S . Shapot Nuclear Proteins and the Cell Cycle Gary Stein and Renato Baserga AUTHOR INDEX-SUBJECT INDEX

357

Particular Emphasis on the Transkei, South Africa Gerald P. Warwick and John S. Harington Genetic Control of Murine Viral Leukemogenesis Frank L i l y and Theodore Pincus Marek's Disease: A Neoplastic Disease of Chickens Caused by a Herpesvirus K . Nazerian Mutation and Human Cancer Alfred G. Knudson, Jr. Mammary Neoplasia in Mice S . Nandi and Charles M . McGrath AUTHOR INDEX-SUBJECT INDEX

Volume 16 Polysaccharides in Cancer Vijai N . Nigam and Antonio Cantero Antitumor Effects of Interferon Ion Gresser Transformation by Polyoma Virus and Simian Virus 40 Joe Sambrook Molecular Repair, Wound Healing, and Carcinogenesis: Tumor Production a Possible Overhealing? Sir Alexander Haddow The Expression of Normal Histocompatibility Antigens in Tumor Cells Alena Lengerova 1,3-Bis(2-Chloroethyl)-I -Nitrosourea (BCNU) and Other Nitrosoureas in Cancer Treatment: A Review Stephen K . Carter, Frank M . Schabel, Jr., Lawrence E. Broder, and Thomas P . Johnston AUTHOR INDEX-SUBJECT INDEX

Volume 18 Immunological Aspects of Chemical Carcinogenesis R . W . Baldwin Isozymes and Cancer Fanny Schapira Physiological and Biochemical Reviews of Sex Differences and Carcinogenesis with Particular Reference to the Liver Yee Chu Toh Immunodeficiency and Cancer John H . Kersey, Beatrice D. Spector, and Robert A . Good Recent Observations Related to the Chemotherapy and Immunology of Gestational Choriocarcinoma K . D. Bagshave Glycolipids of Tumor Cell Membrane Sen-itiroh Hakomori Chemical Oncogenesis in Culture Charles Heidelberger AUTHOR INDEX-SUBJECT INDEX

Volume 17 Volume 19 Polysaccharides in Cancer: Glycoproteins and Glycolipids Vijai N . Nigam and Antonio Cantero Some Aspects of the Epidemiology and Etiology of Esophageal Cancer with

Comparative Aspects of Mammary Tumors J . M . Hamilton The Cellular and Molecular Biology of RNA Tumor Viruses, Especially Avian Leu-

358

CONTENTS OF PREVIOUS VOLUMES

kosis-Sarcoma Viruses and Their Relatives Howard M . Temin Cancer, Differentiation, and Embryonic Antigens: Some Central Problems J . H . Coggin, Jr. and N . G. Anderson Simian Herpesviruses and Neoplasia Fredrich W . Deinhardt, Lawrence A. Falk, and Lauren G. Wove Cell-Mediated Immunity to Tumor Cells Ronald B . Herberrnun Herpesviruses and Cancer Fred Rapp Cyclic AMP and the Transformation of Fibroblasts Ira Pastan and George S . Johnson Tumor Angiogenesis Judah Folkman AUTHOR INDEX-SUBJECT INDEX

Cell Death in Normal and Malignant Tissues E . H . Cooper, A . J . Bedford, and T. E. Kenny The Histocompatibility-Linked Immune Response Genes 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 Keen A . Rafferty, Jr. 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 20 Tumor Cell Surfaces: General Alterations Detected by Agglutinins Annette M . C . Rapin and Max M. Burger Principles of Immunological Tolerance and Immunocyte Receptor Blockade G. J . V . Nossal The Role of Macrophages in Defense against Neoplastic Disease Michael H . Levy und 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 SUBJECT INDEX

Volume 21 Lung Tumors in Mice: Application to Carcinogenesis Bioassay Michael 8. Shimkin and Gary D. Stoner

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 Proteolytic Enzymes, Cell Surface Changes, and Viral Transformation Richard Roblin, lih-Nan Chou, and Paul H . Black Immunodepression and Malignancy Osias Stutman SUBJECT INDEX

Volume 23 The Genetic Aspects of Human Cancer W . E. Heston The Structure and Function of Intercellular Junctions in Cancer Ronald S. Weinstein, Frederick B. Merk, and Joseph A h o y

CONTENTS QF PREVIOUS VOLUMES

Genetics of Adenoviruses Harold S. Ginsberg and C . S. H . Young Molecular Biology of the Carcinogen, 4-Nitroquinoline I-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 and J . C . Leclerc Organization of the Genomes of Polyoma Virus and SV40 Mike Fried and Beverly E. GrifJin p,-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 Jr., Michael J . Mas(rangelo, Ann M . Ainsworrh, David Berd, Robert E. Bellet, and Evelina A . Bernardino SUBJECT INDEX

Volume 25 Biological Activity of Tumor Virus DNA F. L. Graham

359

Malignancy and Transformation: Expression in Somatic Cell Hybrids and Variants Harvey L . Ozer and Krishna K . Jha Tumor-Bound Immunoglobulins: In Siru Expressions of Humoral Immunity Isaac 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 Role of Suppressive Factors in Cancer Isao Kamo and Herman Friedman Passive Immunotherapy of Cancer in Animals and Man Steven A . Rosenberg and William D. Terry SUBJECT INDEX

Volume 26 The Epidemiology of Large-Bowel Cancer Pelayo Correa and William Haenszel 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 Corynebacrerium parvum Luka Milas and Martin T. Scott SUBJECT INDEX

Volume 27 Translational Products of Type-C RNA Tumor Viruses John R . Stephenson, Sushilkumar G . Devare, and Fred H . Reynolds, Jr.

360

CONTENTS OF PREVIOUS VOLUMES

Quantitative Theories of Oncogenesis Alice S. Whittemore Gestational Trophoblastic Disease: Origin of Choriocarcinoma, Invasive Mole and Choriocarcinoma Associated with Hydatidiform Mole, and Some Immunologic Aspects J . I . 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 . Herberman and Howard T. Holden SL'BJECT INDEX

Volume 28 Cancer: Somatic-Genetic Considerations F. M. Burnet Tumors Arising in Organ Transplant Recipients Israel 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 lsaiah J . Fidler. Douglas M . Gersten, and Ian R . Hart Bovine Leukemia Virus Involvement in Enzootic Bovine Leukosis

A , Burny, F. Bex, H. Chanrrenne, Y . Cleuter, D. Dekegel, J . Ghysdael, R. Ketimann, M. Leclercq, J . Leunen, M. Mammerickx, and D. Portetelle Molecular Mechanisms of Steroid Hormone Action Stephen J . Higgins and Ulrich 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 Jose W e 1 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 Seppala Mammary Tumor Viruses Dan H . Moore, Carole A . Long, Akhil B. Vaidya, Joel B. Sheffield, Arnold S. Dion, and Etienne Y. Lasfargues Role of Selenium in the Chemoprevention of Cancer A . Clark Griffin SUBJECT INDEX

Volume 30 Acute Phase Reactant Proteins in Cancer E. H . Cooper and Joan Stone Induction of Leukemia in Mice by Irradiation and Radiation Leukemia Virus Variants Nechama Haran-Ghera and Alpha Peled On the Multiform Relationships between the Tumor and the Host V . S . Shapot

CONTENTS OF PREVIOUS VOLUMES

361

Volume 31

Role of Hydrazine in Carcinogenesis Joseph Bald Experimental Intestinal Cancer Research with Special Reference to Human Pathology Kazymir M. Pozharisski, Alexei J . Likhachev. Valeri F . Klimashevski. and Jacob D . Shaposhnikov The Molecular Biology of Lymphotropic Herpesviruses Bill Sugden, Christopher R. Kintner, and Willie Mark Viral Xenogenization of Intact Tumor Cells Hiroshi Kobayashi Virus Augmentation of the Antigenicity of Tumor Cell Extracts Faye C . Austin and Charles W . Boone

The Epidemiology of Leukemia Michael Alderson The Role of the Major Histocompatibility Gene Complex in Murine Cytotoxic T Cell Responses Hermann Wagner, Klaus Pfitenmaier, and Martin Rollinghoff The Sequential Analysis of Cancer Development Emmanuel Farber and Ross Camerson Genetic Control of Natural Cytotoxicity and Hybrid Resistance Edward A . Clark and Richard C . Harmon Development of Human Breast Cancer Sefton R. Wellings

SUBJECT INDEX

SUBJECT INDEX

This Page Intentionally Left Blank