Advances in Cancer Research, Volume 23

Citation preview

ADVANCES IN CANCER RESEARCH VOLUME 23

Contributors to This Volume

Joseph Alroy

Frederick 6. Merk

W. A. Andiman

G. Miller

A. Frank

Minako Nagao

Harold S. Ginsberg

Richmond T. Prehn

W. E. Heston

Takashi Sugimura

Miroslav Hill

Ronald S. Weinstein

lana Hillova

C. S. H. Young

ADVANCES IN CANCER RESEARCH Edited by

GEORGE KLEIN Department of Tumor Biology Karolinska lnrtitutet Stockholm, Sweden

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

Consulting Editor

ALEXANDER HADDOW Chester Beatty Research Institute Institute of Cancer Research Royal Cancer Hospital London, England

Volume

23 - 7976

ACADEMIC PRESS

New York

A Subsidiary of Harcourt Brace Jovanovich, Publishers

San Francisco

London

COPYRIGHT 1976, BY ACADEMIC PRESS,INC. ALL RIGHTS RESERVED. NO PART OF THIS PUBLICATION MAY BE REPRODUCED O R 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 by ACADEMIC PRESS, INC. (LONDON) LTD. 24/28 Oval Road. London NW1

LIBRARY OF CONGRESS CATALOG CARD NUMBER: 52-13360 ISBN 0-12-006623-8 PRINTED IN

THE UNITED STATES OF AMERICA

CONTENTS CONTRIBUTORS

TO

VOLUME 23

.

.

.

.

.

.

.

.

.

.

.

ix

. .

. .

. . . .

1 3

The Genetic Aspects of Human Cancer

.

W E. HESTQN

I. Introduction . . I1. Cancer of the Breast . I11. Leukemia . . . IV. Colorectal Cancer . V. Gastric Cancer . . VI. Retinoblastoma . . VII. Xeroderma Pigmentosum VIII . Albinism . . . IX . Lung Cancer . . References . . .

. . . . . . . . . .

. . . . . . . . . .

. .

. . . . . . . .

. .

. . . . . . . .

. .

. . . . . . . .

. .

. . . . . . . .

. .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

7 10 12 13

15 16 17 19

The Structure and Function of Intercellular Junctions in Cancer

RONALDS . WEINSTEIN.FREDERICK B. MERK.AND JOSEPHALROY

I. Introduction . . . . . . . . . . I1. Membrane Ultrastructure . . . . . . . . 111. Cell Junction Classification . . . . . . . . . . IV. Occurrence of Cell Junctions in Tumors . V. Cell-to-Cell Communication and Growth Control . . . VI . Cell Junctions in Embryonic Development . VII . Cell Junctions and the Biological Behavior of Tumor Cells VIII. Intercellular Adhesion in Tumors . . . . . . IX . Transepithelial Permeability and Malignant Transformation . . . . . . . . . . X. Summary . References . . . . . . . . . . .

. .

.

. . . . . . . . . . .

.

.

. . . . . . . . .

. . . . . . . . .

.

.

.

.

23 26 32 43 61 70 71 75 77 77 79

Genetics of Adenoviruses

HAROLD S. GINSBERGAND C. S. H . YOUNG I. Introduction . . . . I1. Isolation of Adenovirus Mutants HI. Genetic Characterization . .

.

.

.

.

.

.

.

. . . . . . . . . . . . . . . . . . V

91 101 105

vi

CONTENTS

IV . Phenotypes of Adenovinis Mutants . . . . V. Summing Up . . . . . . References .

.

.

. .

.

. .

. .

.

. .

.

. .

.

. .

.

. .

.

. .

116 123 126

Molecular Biology of the Carcinogen. 4-Nitroquinoline 1 -Oxide

MINAKONAG.40

AND

TAKASHI SUGIMURA

I. Introduction . . . . . . . . . . . . . I1. Mutagenic Activity of 4-Nitroquinoline 1-Oxide on Organisms . . . 111. Chromosome Aberrations . . . . . . . . . . . IV . Repair of 4-Nitroquinoline 1-Oxide-Damaged DNA . . . . . V . Interaction of 4-Nitroquinoline 1-Oxide and Its Derivatives with Nucleic . . . . . . . . . . . . . . Acids . 1’1. Interaction of 4-Nitroquinoline 1-Oxide and Its Derivatives with Protein VII . Recent Information on Carcinogenesis by 4-Nitroquinoline 1-Oxide . . VIII . 4-Nitroquinoline 1-Oxide and Microbial Screening Method for Carcinogen IX . Conclusions . . . . . . . . . . . . . References . . . . . . . . . . . . . .

132 133 140 143

151 156

157 163 163 164

Epstein-Barr Virus and Nonhuman Primates: Natural and Experimental Infection

A. FRASK.W . A . ASDIMAN.

AND

C . MILLER

I. Introduction . . . . . . . . . . I1 . EBV Reactive Antibodies in Nonhuman Primates . . 111. Lymphoblastoid Cell Lines ( L C L ) from Nonhuman Primates IV . Experimental Infection of Nonhuman Primates with EBV . V . Summary and Conclusions . . . . . . . . . . . . . . . . . References .

.

.

.

. .

. .

. .

. .

. .

177 186 . 197 . 1%

. .

.

.

171

. 172

Tumor Progression and Homeostasis

RICHMOXDT. PREHN I. Introdrrction . . . . . . . . I1. Initiation: The First Step in Tumor Progression . I11. The Subsequent Steps in Tumor Progression . IV . Immunity as a Homeostatic Mechanism . . V . Concluding Remarks . . . . . . References . . . . . . . . .

. .

. .

. .

. .

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

203 204 . 213 . 218 . 233 . 233

Genetic Transformation of Animal Cells with Viral DNA of RNA Tumor Viruses

Mmosmv HILL AKD I . Introduction I1. Endogeneous

.

.

Viruses

. .

. .

. .

JAXA

. .

. .

HILLOVA

. .

. .

. .

. .

. .

. .

237 239

vii

CONTENTS

I11. Virus-Specific DNA in Virus-Infected and Uninfected Cells IV. Infectivity of the Viral DNA . . . . . . . V . Sizing the RNA Genome in Virus Particles . . . . VI Search for Transforming Genetic Material . . . . . VII. Conclusions . . . . . . . . . . References . . . . . . . . . . .

.

SUBJECTINDEX . CONTENTS OF

.

.

.

. PREVIOUS VOLUMES.

. .

. . .

. .

. . . . . . . .

. . . . .

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

243 246 271 274 287 289 299 304

This Page Intentionally Left Blank

CONTRIBUTORS

TO VOLUME 23

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

ALROY, * Department of Pathology, Tufts University School of Medicine, Boston, Massachusetts (23)

JOSEPH

W. A. ANDIMAN,Departments of Pediatrics and Epidemiology and Public Health, Yale University School of Medicine, New Haven, Connecticut ( 171) A. FRANK,Departments of Pediatrics and Epidemiology and Public Health, Yale University School of Medicine, New Haven, Connecticut ( 171 )

HAROLD S . GINSBERG, Department of Microbiology, College of Physicians and Surgeons, Columbia University, New York, New York (91)

W . E. HESTON,Laboratory of Biology, National Cancer Institute, National lnstitutes of Health, Bethesda, Maryland ( 1 ) MIROSLAVHILL, Department of Cellular and Mokculur Biology and Equipe de Recherche No. 148 du C.N.R.S., lnstitute of Cancerology and Immunogenetics, Villejuif, France (237) JANA HILLOVA, Department of Cellular and Molecular Biology and Equipe de Recherche No. 148 du C.N.R.S., lnstitute of Cancerology and lmmunogenetics, Villejuif, France (237) FREDERICK B. MERK,Department of Pathology, Tufts University School of Medicine, Boston, Massachusetts ( 23) G. MILLER,Departments of Pediatrics and Epidemiology and Public Health, Yale University School of Medicine, New Haven, Connecticut ( 171 ) MINAKO NAGAO, National Cancer Center Research Institute, Chuo-ku, and lnstitute of Medical Science, University of Tokyo, Minato-ku, Tokyo, Japan (131) RICHMONDT. PREHN,The lnstitute for Cancer Research, The Fox Chase Cancer Center, Philadelphia, Pennsylvania ( 203 )

’Present address: Rush Medical College and Rush-Presbyterian-St. Luke’s Medical Center, Chicago, Illinois. ix

X

CONTRIBUTORS

TAKASHI SUGIMURA, Nationul Cancer Center Research Institute, Chuo-ku, and Institute of Medical Science, University of Tokyo, Minuto-ku, Tokyo, Japan ( 131)

RONALD S . WEINSTEIN,'Department of Pathology, Tufts University School of Medicine, Boston, Massachusetts ( 2 3 ) C . S . H. YOUNG, Department of Microbiology, College of Physicians and Surgeons, Columbia University, New Ymk, New York (91)

Present address: Rush Medical College and Rush-Presbyterian-St. Luke's Medical Center, Chicago, Illinois.

THE GENETIC ASPECTS OF HUMAN CANCER W. E. Heston Loboratory at Biology, National Cancer Institute. National Institutes of Health, Bethesda, Maryland

I. Introduction . . . . . . . 11. Cancer of the Breast . . . . . 111. Leukemia . . . . . . . . A. Studies on Inheritance . . . . B. Chromosomal Aberrations and Leukemia IV. Colorectal Cancer . . . . . . V. Gastric Cancer. . . . . . . VI. Retinoblastoma . . . . . . . VII. Xeroderma Pigmentosum . . . . V I E Albinism . . . . . . . . . . . . . . IX. Lung Cancer . References

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

. . .

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

.

.

. . .

.

.

. . .

.

.

.

.

. . . . . .

. . . . . .

.

.

.

.

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

1 3

7 7 9

10 12 13 15 16 17 19

1. Introduction

This chapter is concerned with the genetic transmission of factors influencing the occurrence of cancer in man. It is not a complete review of the subject. Such a review would require a large volume with half a volume of references. Cancers that have received greater emphasis in this area of investigation are discussed, and an assessment is made of the kinds of evidence in support of genetic influences. Selected references are given. An effort was made not to slight early investigations as is often done in reviews today. There was a peak of emphasis on the genetics of human cancer during the 1940s and early 1950s, stimulated in large part in the United States by a very informal Conference on Parental Influence in Relation to the Incidence of Human Cancer conceived and organized by Dr. C. C. Little and held as a part of the Fifteenth Anniversary Celebration of the Roscoe B. Jackson Memorial Laboratory at Bar Harbor, Maine, in 1944. It was concluded at this conference that evidence at that time on parental factors influencing occurrence of cancer in experimental animals, particularly mice, together with information already known regarding genetic influences in respect to cancer in man, made it imperative that geneticists extend their knowledge on the genetic influences in cancer in man, and place such knowledge in proper balance 1

2

W. E. HESTON

with the increase in information on the chemical and physical agents and other nongenetic factors. Several large studies were initiated soon thereafter. One on breast cancer was developed by Dr. Madge T. Macklin at Ohio State University. Another also on breast cancer was developed by Dr. C. P. Oliver at the Dight Institute of the University of Minnesota. Dr. Oliver later transferred to the University of Texas, where he developed a program in the same area; the Dight Institute program was carried on principally by Dr. Elving Anderson. At the same time a very significant program, principally on the genetics of gastrointestinal cancer, was being developed at the University of Utah by Drs. F. E. Stephens, Eldon J. Gardner, and Charles M. Woolf. This group had the advantages of access to the large polygamous families of the Mormons and the extensive family records in the Church's Genealogical Library in Salt Lake City. Another program deserving special mention was the one in Denmark involving the University Institute of Human Genetics under the direction of Dr. Tage Kemp, the Danish Cancer Registry under the direction of Dr. Johannes Clemrnenson, and the University Institute of Pathologic Anatomy directed by Dr. Julius Engelbreth-Holm. Extensive studies on genetics of leukemia and of breast cancer were carried out there. Productive programs were carried out elsewhere by other investigators, one of whom was Dr. R. P. Martynova in U.S.S.R. She deserves special mention not only because of the early work she did on the genetics of breast cancer in women, but also for her signal role in keeping the genetics of cancer alive in the U.S.S.R. during the Lysenko era. These programs, which utilized for the most part comparison of incidence of cancer in relatives of cancer probands with that in relatives of control probands, established that genetic factors were iiivolved in many kinds of cancer in man and could be demonstrated provided enough data were collected. The studies also defined some of the limitations of this approach. These studies are reviewed because of their own merit and so that today we may distinguish between new discoveries and extension or confirmation of observations made at that time. The concept of human cancer as a somatic mutation disease does not receive special emphasis here, but not because somatic mutation is not involved. It surely is, as some form of change in the genetic material of the cell, but the subject has recently been covered thoroughly bv Knudson (1975). Furtheimore, no attempt has been made to give a complete review of cytogenetics and cancer. This subject has been covered thoroughly in a recent volume edited by German (1974). Some attempt is made to forecast what future investigations might

GENETIC A S P E W OF HUMAN CANCER

3

reveal from what is being discovered in experimental animals. Of principal interest here is that many kinds of cancer in experimental animals are now known to be induced by viruses, and that many of these viruses are endogenous. These are of particular concern to the geneticist because of evidence that they are transmitted vertically as a part of the host genome, much as genes are transmitted. If viruses are involved in the induction of cancer in man, and they most surely will be shown to be involved, the geneticist must take a very active role in their discovery. It. Cancer of the Breast

Studies to date indicate an inherited influence in the etiology of breast cancer that is especially prominent in the case of premenopausal cancer. From a large proband study of cancer in Holland, Wassink (1935) observed that when the proband had cancer of the breast there was a significant increase of cancer among female relatives owing to an increase in the homologous form of cancer. This was followed by Martynova’s (1937) rather extensive study in the U.S.S.R. From her data on 201 breast cancer family histories, she concluded that hereditary factors play a definite role in predisposition to cancer of the breast in women. She also concluded that predisposition to cancer of the breast is in some way connected with predisposition to cancer in general. Jacobsen (1946), working at the University Institute of Human Genetics in Copenhagen and in collaboration with the Danish Cancer Registry, compared relatives of 200 breast cancer probands with like relatives of 200 controls and found what he termed an indubitable excess incidence of breast cancer among the female relatives of the breast cancer probands with exception of grandmothers. He interpreted these results as indicating that the hereditary predisposition was a major factor in the development of breast cancer. It was noted that the curve of the age distribution at the first manifestation of the disease in the 200 breast cancer probands had two peaks, at ages 45-49 and 60-64. It was further observed that probands with stronger evidence of hereditary iduence were in the early age group. This is comparable with the difference between pre- and postmenopausal breast cancer described recently by Anderson ( 1972). In 1955, Woolf reported on his study of breast cancer in the Utah population. He,selected 216 patients who had died from breast cancer as probands and collected cancer data on their relatives. The number of deaths from cancer in these relatives was compared with an expected number based on proportionate mortality rates from the general popula-

4

W. E. HESTON

tion. Also, by the sequential analysis method he compared the. frequency of cancer in the families of the breast cancer probands with the frequency in a control sample. In addition to data on mothers and sisters of the probands, he also collected data on fathers and 'brothers since he was interested in whether his data would confirm Martynova and others who considered that susceptibility to cancer of the breast was only one manifestation of susceptibility to cancer in general or confirm Penrose et al. (1948) and others who considered the predisposition to be organ specific. Female relatives of the breast cancer probands had a higher incidence of cancer of the breast than the female relatives of the control probands and higher than expected from the general population, confirming that there was an inherited predisposition. The fact, however, that the frequency of other types of cancer in all four groups of relatives g a s no greater than in the control sample indicated that the inherited predisposition was organ specific. Later studies of Penrose et al. (1948), Anderson et aZ. (1958), Oliver ( 1959), and Macklin ( 1959) have been published with similar conclusions. Since Macklin's study was probably the most comprehensive, it will be discussed here. She collected complete data on mothers, grandmothers, aunts, and sisters of three groups of probands: ( 1 ) women with diagnosed breast cancer, ( 2 ) women with some cancer other than of the breast, and (3) women who had had no known cancer. Breast cancer occurred 2 or 3 times as frequently among relatives of the breast cancer probands as would have been expected from mortality rates or proportionate death rates. Among relatives of the probands with some other cancer, breast cancer occurred even slightly less frequently than expected, especially on mortality rates, thus failing to support any possible genetic relationship between breast cancer and other types. There was no difference between the frequency of breast cancer among relatives of the probands with no cancer and what would have been expected on mortality rates or proportionate death rates. From these data, Macklin concluded that there was some factor or factors that caused the relatives of breast cancer patients to have significantly more breast cancer than would have been expected if they had experienced the same risk as the population with which they were compared. She pointed out that these factors might be in the environment or the genes or both. Macklin compared paternal aunts of the breast cancer probands with their maternal aunts and found no difference in frequency of breast cancer. She suggested that this would tend to rule out environmental factors and indicates the action of genetic factors, since it would be

GENETIC ASPECTS OF HUMAN CANCER

5

unusual to find the same environmental factors influencing the fathers’ mothers and sisters in the same way they influenced the mothers’ mothers and sisters. Comparison of paternal and maternal grandmothers of the breast cancer probands was of particular importance in view of the maternally transmitted mammary tumor virus (MTV) of the mouse. If there were a milk-transmitted breast cancer virus in women as in mice, one could expect a higher frequency of breast cancer in the maternal grandmothers than in the paternal grandmothers. From the fact that there was no difference between the two groups of grandmothers, she concluded that, if there is a milk-borne virus for breast cancer in women, it must be ubiquitous and some other agent is the deciding factor. These are very significant observations in relation to our present knowledge of transmission of mammary tumors in mice (for review, see Heston, 1973). While Macklin did not rule out the possibility of a breast cancer virus transmitted from parents to offspring, if such a virus exists in women the best experimentaI model is probably not the strain C3H model, where the virus (MTV) is primarily transmitted through the milk. The best model may be the strain C3HfB model where the virus, in this case nodule-inducing virus (NIV), is transmitted through the male as readily as through the female and is not transmitted through the milk; or the strain GR model, where the virus is transmitted through the male as readily as through the female but can also be transmitted through the milk. Present efforts are directed toward trying to find a virus in the milk of women, but this seems to be primarily because the milk is a convenient place to look for it, It must be pointed out that some suggestive, although far from conclusive, evidence for the presence of virus has been found. However, if any breast cancer virus in women is like NIV, it would not be expected to be in the milk, at least in detectable amounts. There is considerable evidence that the mammary cancer virus in mice is transmitted as a provirus, i.e., that the viral information is integrated in the genome of the mouse. It may prove even to be transmitted as a dominant gene ( Bentvelzen, 1972). Information thus far from studies of human breast cancer would suggest that if there is a breast cancer virus, it too is probably transmitted genetically and thus the virologists are going to need the assistance of geneticists in discovering it. Again one is reminded that geneticists made the original discovery in respect to the MTV in mice (Jackson Laboratory Staff, 1933; Korteweg, 1934). In relation to some of the earlier observations of Jacobsen (1946) referred to above, Anderson (1972) separated his breast cancer probands

6

W. E. HESTON

into premenopausal and postmenopausal cases. Breast cancer was increased about 3-fold in the relatives of the premenopausal group, but was not increased in the relatives of the postmenopausal group. He also noted a 5-fold increase in relatives of bilateral breast cancer patients; and in relatives of patients with both premenopausal and bilateral cancer the rate was increased 9-fold. He concluded that genetic factors must play a more important role in patients with early onset of multiple disease than in patients with late onset of a single tumor. The same situation could be expected if a vertically transmitted virus were a factor in inducing breast cancer. Anderson's observations are in line with those made in the mouse, where a strong genetic component or a strong viral factor, or both, results in early mammary tumors, the females often having multiple mammary tumors. It is thus in these patients with premenopausal and bilateral cancers that one would expect to have the greatest chance of demonstrating any breast cancer virus. The possibility of a genetic relationship between breast cancer and other forms of cancer suggested by the early works of Martynova (1937) and others is finding support in certain family studies reported recently. Li and Fraumeni (1969) described four families showing a concordance of soft tissue sarcomas, leukemia, breast cancer, and an apparent excess of multiple primary malignant neoplasms. Later Lynch et a2. (1973~) reported a study of 34 families in which two or more first- or seconddegree relatives had breast cancer. Of these 34 families, 11 had firstor second-degree relatives with associated soft tissue sarcomas, leukemia, or brain tumors, or combinations of these malignant neoplasms. In another study of 22 families, Lynch et al. (1973b) observed an association of gastrointestinal and breast cancer. Through three generations of two families reported by Lynch et al. (1974), there was an apparent association between breast -and ovarian cancer. One might expect the association between breast and ovarian cancer to be caused by hormonal factors. The administration of estrogen results in neoplasms in several organs of the endocrine and reproductive systems in mice. Woolley et al. (1952) induced adrenal, pituitary, and mammary gland neoplasms in certain hvbrid mice by the hormonal imbalance resulting from early castration. *In the absence of administered hormonal factors, one might expect such endocrine influences to be basically genetic. Such associations of different forms of cancer presumably also could result from viral or genetic factors. A vertically transmitted virus with the oncogenic capacity of the polyoma virus (Stewart et al., 1958) could result in such combinations. On the other hand, a single gene, like the Av or AZ'V genes of the mouse that increase the occurrence of hepa-

GENETIC ASPECTS OF HUMAN CANCER

7

tomas, mammary tumors, pulmonary tumors, and leukemias (Heston and Vlahakis, 1968), if present in human beings could also result in such associations. 111. Leukemia

A. STUDIESON INHERITANCE One of the most extensive proband studies on the genetics of leukemia was carried out in Denmark by Videbaek (1947). Data were collected on 209 leukemia probands and their 4041 relatives and on 200 sound control probands and their 3641 relatives with good agreement between the age distribution of the two groups of relatives. Videbaek reported among the relatives of the leukemia patients an excessive incidence of cancer, but this was due to high incidence of all forms of the disease. In the families of the 209 leukemia patient probands, however, 17 had at least one other verified case of leukemia whereas in the families of the 200 control probands there was only one case of leukemia. Thus, there was significantly more leukemia among the relatives of the leukemia probands than among the relatives of the controls. From analysis of these 17 families and others from the literature making a total of 39, it was concluded that genetic factors had a role in the occurrence of leukemia, but the mode of inheritance could not be determined. It appeared that genes were controlling a predisposition to the disease, leaving open the possibility of the additional influence of chemical or physical carcinogens or viruses. There was no evidence of sex limitation or sex linkage and no evidence of extrachromosomal or maternal inheritance, as had been shown by that time for mammary cancer in mice. Genetically, leukemia appeared as an entity with the various types occurring among the relatives of the probands. Chronic lymphogenous leukemia, and probably also acute leukemia and chronic myelogenous leukemia, tended to occur earlier in the familial cases than in the sporadic ones. These observations of Videbaek, with the exception of the increase in cancer in general in families of leukemia probands, might be expected from what we have observed of the occurrence of leukemia in certain inbred strains of mice and from the results of the classic studies of Cole and Furth (1941) and MacDowell and co-workers (1945). They demonstrated that genetic factors were involved in mouse leukemia, and these appeared to be multiple. Although Videbaek's observations were not confirmed by a more recent study of leukemia in man by Steinberg (1960), the excess of leukemia in sibships has been confirmed

8

W. E. HESTON

for childhood leukemia by Stewart (1961) and Miller (1963). The frequency is about four times normal expectation. A classical approach to the ident&cation of genetic factors is the comparison of concordance in identical or monozygous twins with that in fraternal or dizygous twins. Since nongenetic factors would be about as nearly alike in the dizygous twins as in the monozygous twins and since the latter would be identical genetically except for mutations occurring after the splitting of the zygote, a greater concordance would be expected between the monozygous pairs if genetic factors were involved. MacMahon and Levy (1964) have reported a concordance rate of about 25%for childhood leukemia among monozygous twins while in the literature only three concordant sets were described as dizygotic and none was well documented. If a twin child falls ill with leukemia the monozygous mate has one chance in four or five of also developing the disease and usually within weeks or a few months. Although the data from twins are strong evidence for genetic factors in childhood leukemia, Clarkson and Boyse (1971) have suggested as an alternative explanation the possibility that high concordance in the monozygous twins may be due to fusion of placentas with common circulation permitting the formation of hematopoietic chimerism. They are suggesting that the neoplastic change may occur before birth and that many cases of concordance may represent only one occurrence of leukemia, not two. They further point out that whether or not this is the case might be shown through cytogenetic studies. This evidence for inherited influences, especially in childhood leukemia, is in line with the high incidence and early appearance of leukemia in certain inbred strains of mice where the genetic influence is strong, but it is also generally true in mice where a potent leukemia virus is involved. The significant demonstration by Gross (1951) of a vertically transmitted leukemia virus in the mouse eventually led to the concept of genetically transmitted C-type leukemia viruses put forth by Huebner and his associates (Huebner and Todaro, 1969) as their oncogene theory. This postulates, like the provirus theory of the transmission of the B-type mammary tumor virus, that the C-type oncogenic RNA viral information is transmitted as DNA in the host genome. Actual identificaton and location of the locus or loci has come forth from works of others, particularly Rowe and associates (1972). From rapidly accumulating information on leukemia viruses in mice and other experimental animals, it appears likely that a leukemia virus will eventually be found in man. If so, it probably also will be genetically transmitted, and thus here geneticists working in collaboration with virologists will again be able

GENETIC ASPECTS OF HUMAN CANCER

9

to make a real contribution to out understanding of the transmission of the disease.

B. CHROMOSOMAL ABERRATIONS AND LEUKEMIA In the 1930s, Dr. Warren H. Lewis was observing that neoplastic cells had more chromosomal morphologic irregularities than did normal cells. He asked the basic question whether these changes were the cause of the neoplasia, the result of the neoplastic change, or a manifestation of a basic factor causing both the neoplasia and the chromosoma1 changes. We still do not have the final answer to his question although certain chromosomal traits are found to be of value in diagnosis and in determining cancer risks. Dr. Lewis’ basic observations stimulated a vast number of karyotypic studies of tumors of all kinds, which for the next two decades were relatively unproductive. The changes did not appear to have much uniformity in their patterns of manifestation or in their causation. The picture changed in 1960 when Nowell and Hungerford reported that a minute chromosome replaced one of the smallest autosomes in cells of seven patients with chronic granulocytic leukemia which they had studied. This minute chromosome was not seen in cells of four cases of acute granulocytic leukemia in adults or of six cases of acute leukemia in children. There were no other frequent or regular chromosomal changes in the cells of the chronic granulocytic leukemia patients, and all patients had many cells with a normal karyotype. Thus, a chromosomal marker for chronic granulocytic leukemia, that was to be confirmed many times, had been identified. This minute chromosome is now commonly referred to as the Philadelphia chromosome (Phl) (Sandberg et al., 1964). This observation of a definite chromosomal change associated with a particular kind of neoplasm has given a renewed stimulus to cytogenetic studies of neoplasia, particularly the leukemias and other reticulum cell neoplasms. The observation that Bloom’s syndrome is associated with increased susceptibility to acute leukemia is of particular interest to the geneticist because the syndrome itself is inherited through an autosomal recessive gene (Sawitsky et al., 1966). The syndrome is characterized by photosensitive telangiectasia of the face. Data indicate that one of eight persons with the syndrome will develop leukemia during the first 30 years of life. It is thought that the causation of leukemia is related to the observation of excessive chromosomal breakage and rearrangement in cultured cells from patients with the syndrome. Similarly, the recessively inherited Fanconi’s aplastic anemia ( Bloom

10

W. E. HESTON

et al., 1966) and ataxia-telangiectasia (Hecht et al., 1966) are also associated with an abnormal number of chromosomal aberrations, and both diseases also appear to be associated with an increased risk for leukemia and other neoplasms. Children with Down’s trisomy syndrome, mongolism, have an increased incidence of leukemia (Jackson et al., 1968). They have an extra chromosome number 21 of the G group of autosomes. Males are invariably infertile, but female mongols have been reported with offspring. The female mongol produces two kinds of gametes, one with 24 chromosomes which when fertilized by a normal sperm results in a mongol, and the other with 23 chromosomes which when fertilized produces a normal zygote (see Penrose and Smith, 1966). To this list could be added numerous less well defined chromosomal aberrations that appear to be associated with increased leukemia. It appears from these observations that there is an association between unbalanced karyotypes and neoplasia, particularly leukemia. It may be that such cells are more susceptible to oncogenic viruses. Todaro and co-workers (1966; Todaro and Martin, 1967) have shown that cultured fibroblasts from patients with Fanconi’s anemia or mongolism show increased malignant transformation when exposed to SV40. Hirschhorn and Block-Shtacher (1970) have evidence that these cells are also more susceptible to malignant transformation when treated with some of the chemical carcinogens. Hirschhorn ( 1970) concludes that these chromosomal aberrations lead to a selective advantage causing the increase in cell number called leukemia and he postulates that the advantage may be due in part to membrane changes leading to alterations in pseudoimmunological recognition mechanisms. How far have we come since the 1930s in answering Dr. Lewis’ questions on the relationship of these chromosomal aberrations to the neoplastic process? We have made progress and have opened up new areas for research, but we do not have the final answers. In the meantime we have discovered some markers to aid in prediction and diagnosis. IV. Colorectal Cancer

While there is abundant evidence that colorectal cancer may be familial, the primary evidence of genetic factors is found in the adenocarcinoma of the colon arising from inherited polyposis. It is well established that in individuals with this condition some of the polyps progress into carcinomas. Dukes (1952) gave an extensive review of the subject, and a recent review has been published by Burdette (1970). The classical condition is multiple polyposis of the colon first shown by Cockayne (1927) to be due to a single dominant gene. This has

GENETIC ASPECTS OF HUMAN CANCER

11

been confirmed by data from many more recent family pedigrees. Reed and Nee1 (1955) in a study of multiple polyposis of the colon estimated that in the state of Michigan, the minimum frequency at birth of individuals with the dominant gene for the trait is about 1 in 8300. Relative fitness of individuals with the gene derived from relative reproductive span was estimated to be 0.78. From these estimates of frequency and relative fitness and with consideration of known biases involved, the mutation rate of the multiple polyposis dominant gene was estimated to be within the range 1 to 3 x 105/geneper generation. A second polyposis important in the genesis of colorectal carcinoma is that described by Gardner (1955, 1962) and now usually referred to as the Gardner syndrome. Here the gastrointestinal polyps occur in association with osteomas, fibromas, and sebaceous cysts. Since these associated lesions occur earlier than the intestinal polyps or the resultant carcinomas, they are of diagnostic value. Analysis of the original kindred brought to the attention of Gardner and Stephens (1950) because it contained 9 cases of cancer of the digestive tract, 8 of which were colorectal, indicated that like multiple polyposis, the Gardner syndrome is inherited as a single dominant gene. Starting with probands obtained from death certificates of the state of Utah and family histories from the records of the Latter-Day Saints Genealogical Society in Salt Lake City, Woolf (1958) carried out a study to test the hypothesis that a genetic component, independent of the gene associated with multiple polyposis, existed for colorectal cancer. Individuals selected as probands were white persons born or raised in Utah with diagnosed carcinoma of the large intestine. Individuals with any indication that multiple polyposis had been present were excluded. In the study involving 242 families, he observed death certificates of 145 fathers, 142 mothers, 309 brothers, and 167 sisters of the probands or a total of 763. Of these, 26 had cancer of the large intestine as compared with 8 among the relatives of well-matched controls. Furthermore, there was a significant increase in all four classes of relatives over controls. There was no difference between relatives of the intestinal cancer probands and controls in respect to cancer at other sites. Woolf concluded that his results were compatible with the hypothesis of a genetic component, but he pointed out that his study, like so many comparable studies for other types of human cancer, did not distinguish between genetic and familial factors. Macklin (1960) studied families of 167 patients with gastric cancer and 145 patients with cancer of the large intestine where inherited multiple polyposis was lacking. Her aim was to see whether there was evidence for a genetic basis for these two forms of cancer, and if so was

12

W. E. HESTON

there evidence of any genetic factors in common with the two types. Gastric cancer was found more frequently among relatives of gastric cancer probands than in the general population, but cancer of the large intestine was not. Likewise, cancer of the large intestine was found more frequently among relatives of the probands with cancer of the large intestine than in the general population, but gastric cancer was not. This indicated that both types had familial factors favoring induction that did not influence the development of the other type. The fact that husbands and wives were not affected more often than expected on the basis of random distribution whereas parent and child and two siblings were, indicated that the familial factors probably were genetic. No simple genetic pattern for inheritance was revealed in the families with gastric cancer or those with cancer of the large intestine where there was no multiple polyposis. She concluded that the genetics of both was probably polygenic in contrast with the single dominant gene inheritance of multiple polyposis. From a genetic study of colon cancer based upon cancer in probands and first-degree relatives with matched controls, Lynch et al. (1973a) have reported site-specific colon cancer, breast cancer, and multiple primary malignant neoplasms occurring in significant excess among firstdegree relatives of the colon cancer probands compared with the controls. This suggests a common familial etiologic factor, which they suggest may be a gene or genes or a familial occurring oncogenic virus. While this is an interesting observation, it is not supported by the previous genetic studies of cancer of the colon by Woolf or those by Macklin. Furthermore, these observations do not find general support in the work on genetics of cancer or of viral oncogenesis in experimental cancer. The only virus definitely shown to induce mammary cancer is the B-type RNA mouse mammary tumor virus with its variations, and this induces only mammary cancer and only in mice. The other known oncogenic C-type RNA viruses induce only leukemias and sarcomas but have not been shown to induce carcinomas. The results observed by Lynch et al. (1973a) would require some unusual virus like the so-called polyoma virus that is a DNA virus ( Stewart et al., 1958). V. Gastric Cancer

A number of investigators have searched for evidence of a genetic factor in gastric cancer (Videbaek and Mosbeck, 1954; Hagy, 1954; Woolf, 1955, 1956; Macklin, 1960). The approach has been the usual one of comparison of gastric cancer in relatives of gastric cancer probands with that in comparable relatives of controls. In general there has been found to be more gastric cancer in the relatives of the probands

GENETIC ASPECTS OF HUMAN CANCER

13

than in those of the controls, although in some studies the difference has been barely significant. Woolf conducted an extensive study from the Utah families including 173 fathers, 168 mothers, 390 brothers, and 260 sisters of gastric cancer probands and a like number of controls. More gastric cancer occurred in each group of relatives of probands than in the comparable groups of control relatives, the total numbers being 66 compared with 32. That the tendency was site-specific was indicated by the fact that deaths due to other types of cancer in the families of the probands totaled 94 compared with 82 in the controls. The difference between these two incidences was not considered significant. Here again he indicated that it is difficult to tell whether the factor is genetic or familial. Interesting associations between gastric cancer and other nonneoplastic diseases have been revealed. Mosbeck (1953) compared the incidence of gastric cancer among relatives of 234 patients with pernicious anemia with that in relatives of 226 control patients and also with that in the general population of Denmark, Incidence of gastric cancer was significantly greater in the relatives of the pernicious anemia patients than in either of the other groups, but there was no difference in incidence of cancer at other sites. Achlorhydria has also been shown to be associated with gastric cancer (Comfort et al., 1947), and Videbaek and Mosbeck (1954) have suggested that the genetic factor is that causing achlorhydria, which in turn predisposes to both pernicious anemia and gastric cancer. The incidence of gastric cancer seems also to be higher in people of blood group A than in those of the other groups (Aird and Bentall, 1953). A number of genes that control other traits in the mouse also increase or decrease the occurrence of various kinds of cancer, but these genes usually have an effect on growth of the animal and those that increase growth increase occurrence of cancer while those that decrease growth decrease occurrence of cancer ( Heston et al., 1952, Heston and Vlahakis, 1961). These genes, however, have not as yet been associated with gastric cancer in the mouse. The histocompatibility H-2 locus of the mouse has been associated with mammary tumors (Miihlbock and Dux, 1974) and with leukemia of the mouse (Lilly and Pincus, 1973), but in each case the effect of H-2 is on the respective oncogenic virus. VI. Retinoblastoma

Retinoblastoma is a highly malignant tumor of the eye that usually occurs in children from a few months to 4 years of age. In the past the disease was almost invariably fatal, and at present it is considered to be responsible for about 1%of all deaths from cancer in early childhood

14

W. E. HESTON

and 5%of all blindness in children (Smith and Sorsby, 1958). In early publications a few cases were reported, however, of individuals who survived the disease and transmitted it to one or more of their offspring. With recent advances in therapy, the number of such cases is increasing. Through the years retinoblastoma frequently has been cited as an example of an inherited neoplasm in man, and it was generally considered to be due to a single dominant gene (see Neel and Falls, 1951). Neel and Falls made a complete survey of the state of Michigan and found 49 sibships with one or more children with retinoblastoma born to normal parents. On the assumption that these were due to a dominant mutation involving a single locus and with the total number of births in the state during that period, they calculated the mutation rate per gene to be 2.3 x This was considerably higher than the rate of 1.4 x which they cited as having been determined by Philip and Sorsby from material from London and which suggested to them that the rate might be different in different geographical areas. Recent work indicates that in considering the genetics of the disease, the bilateral cases must be separated from the unilateral cases. Sorsby (1972) has reported from his own records and from the literature 19 survivors of sporadic bilateral retinoblastoma with a total of 39 offspring of whom 17 were affected in both eyes and 3 in one eye. This confirms the single dominant gene inheritance of the bilateral disease. It has even been determined through cvtogenetic studies that this locus is on the long arm of a D chromosome (see Orye et al., 1971). The much more common unilateral cases apparently are not transmissible to the offspring escept for 5-10% that appear to be due to the same inherited dominant gene but are unilateral because of reduced expression. The major portion of the unilateral cases, if genetic, would be due to a somatic mutation presumably at the same locus. These are factors to be considered in genetic counseling which is becoming more urgent as chemotherapy and other fonns of treatment become more successful. For the bilateral cases one would expect 50% of offspring to have the trait. One would expect the majority of the unilateral cases not to be transmitted. Difficulty will arise in identifying the 5 1 0 %phenotypically unilateral cases that in fact represent incomplete expression of the germinal mutation for bilateral retinoblastoma. These individuals, if thev could be identified, could be advised that they also could expect to transmit the disease to 50% of their offspring. Sorsby (1972) suggests that there is a possibility that these few hereditary unilateral cases may be identified histologically. Cumings and Sorsby ( 1944) observed that unilateral neoplasms originated exclusively from the outer nuclear laver of the retina whereas the bilateral ones had

GENETIC ASPECTS OF HUMAN CANCER

15

a more diffuse origin. It remains to be seen whether the hereditary unilateral disease also has a diffuse origin, VII. Xeroderrna Pigmentosurn

Xeroderma pigmentosum is a hereditary disease found in one person in approximately 250,000 of the general population. Its occurrence is worldwide in all races. Most patients with the disease have an acute sensitivity to sunlight manifested early in life. Freckles develop on sunexposed areas where the skin becomes dry and scaly, hence the name “dry pigmented skin.” Later neoplasms of the skin appear. These are usually basal cell carcinomas and less frequently squamous cell carcinomas. In patients in which early neoplasms receive effective treatment, up to 50% may later develop malignant melanoma. Abnormalities of other organ systems, including the nervous system, are often associated with the cutaneous lesions. By analyzing family histories in which xeroderma pigmentosum occurred, Cockayne (1933) cdncluded that inheritance was by a single autosomal recessive gene, a conclusion supported by additional family histories published by Macklin ( 1936). Haldane’s analysis ( 1936) indicated to him that the true pattern of inheritance was incomplete sex-linkage, the gene being borne in that particular region of the sex chromosome common to X and Y. Later, in a critical review of autosomal and partial sex-linkage in man, Morton (1957) showed that the inheritance could not be partial sex-linkage and presented evidence. that the earlier tests had confounded sex-bias with partial sex-linkage. Since heterozygotes show the freckling and since homogygotes seldom reproduce, the trait sometimes has been referred to as semidominant sublethal. Recent work of El-Hefnawi et al. (1965) indicates that the locus for xeroderma pigmentosum is linked with the genes for the ABO blood types with a recombination frequency of about 18%.This would refute the claim that it was on the sex chromosome and reestablish it as autosomal. There was, however, an unusually strong nonrandom association between the disease and blood group 0. Eighty percent of the patients in this study were 0 whereas on a random basis one would expect only about 35%.This suggested to the authors some selective mechanism favoring survival of xeroderma pigmentosum homozygotes who are type 0 as compared with those who are A, B, or AB. A possible lead to the mechanism of gene action in xeroderma pigmentosum carcinogenesis appeared in a signal discovery of Cleaver (1968). His observation was that whereas normal skin fibroblasts in culture can repair ultraviolet radiation damage to D N A by inserting new bases into DNA, in fibroblasts with the xeroderma pigmentosum

16

W. E. HESTON

gene such repair replication of DNA is either absent or much reduced. He concluded that this faulty DNA repair must be related to carcinogenesis. Soon thereafter, Epstein and co-workers (1970) showed that this defect in DNA synthesis occurred in the epidermal cells in vivo of patients with the disease. They postulated that since defects in DNA repair result in increased mutation rate induced by ultraviolet light in bacteria, a similar increase in mutation rate in the skin of these patients after exposure to the sun might cause the subsequent neoplasms of the skin. This has been followed by extensive biochemical studies of the disorder not only in fibroblasts of xeroderma pigmentosum patients, but also in their lymphocytes, where ultraviolet-stimulated thymidine incorporation has been investigated (see Robbins et al., 1974). Cell fusion studies have been carried out that suggested heterogeneity of the genetic lesion. Fibroblasts from certain pairs of patients when fused show complementation overcoming the defect in DNA repair in each member of the pair. Yet all these studies have not actually linked the faulty DNA repair to carcinogenesis in these patients. In fact, studies by Robbins and Burk (1973) of one xeroderma pigmentosum patient whose cells failed to show faulty DNA repair have suggested that some mechanism other than enhancement of UV carcinogenesis by defective DNA repair may be responsible for the skin tumor carcinogenesis in this patient and possibly all patients with xeroderma pigmentosum. Despite the apparently normal DNA repair in the fibroblasts and lymphocytes, this patient developed numerous cutaneous neoplasms. The ,possibility that the xeroderma pigmentosum gene is increasing carcinogenesis by increasing oncogenic virus-induced transformation in cells of patients with the disease, has been suggested. Fibroblasts from patients with Down’s syndrome and Fanconi’s anemia, both of which are associatcd with a high incidence of neoplasia, show a marked increase in susceptibility of SV40-induced transformation as noted in the above section on leukemia. In a studv by Key and Todaro (1974) neither xeroderma pigmentosum fibroblasts, which show n o m d susceptibility to SV40 transformation, nor normal skin fibroblasts showed any increase in susceptibility to the virus-induced transformation following exposure to UV radiation. Their results, however, are contrary to some preliminary reports referred to by these authors. VIII. Albinism

In certain aspects albinism in man can be compared with xeroderma pigmentosum. Albinism has even been more clearly shown to be due

GENETIC ASPECJX OF HUMAN CANCER

17

to a simpIe MendeIian recessive gene than has xeroderma pigmentosum, and in this respect it is like albinism in other species of mammals. However, whereas albinism is one of a series of alleles in other species, such as the mouse, definite evidence for other alleles has not been found for man. Albinism occurs much more frequently than xeroderma pigmentosum: the incidence in Europeans is estimated as about one in 10,000, and in the Navajo Indians, it appears twice in 10,000. It has been studied extensively by Keeler (1963, 1964) among the Cuna Indians of the San Blas islands in the Caribbean. Here is found the highest incidence in the world, now being about 45 per 10,000. Albinos have snow-white hair and a great reduction of pigmentation in the eyes and skin. This causes them to be sensitive to the ultraviolet rays of the sun, but in the temperate zone they get along comparatively well by wearing dark glasses and protective clothing and they seldom develop skin cancer. In the tropics, however, where the Cuna Indians are, the intense sun produces a premalignant actinic keratosis in practically all albino children by two or three years of age. Frequently, the terminal episode is metastatic carcinoma. All skin cancer in the Cuna albinos is squamous cell carcinoma, in contrast with the xeroderma pigmentosum patients, in which the basal cell carcinoma is the usual cutaneous neoplasm. The abnormal DNA repair of UV damage of the cells of xeroderma pigmentosum cited in the previous section raises the question whether the cells of albinos are similarly affected. The carcinogenic action of the UV rays in the albino individuals, however, may not be unlike that in the usual Caucasians in whom excessive sunlight] such as is experienced in the Southwestern states, induces neoplasms of the skin. In the albinos the UV rays are probably more effective because of lack of pigmentation, just as Hereford cattle are more susceptible to sunlightinduced cancer eye than other breeds of cattle with pigmentation around the eyes (see Anderson, 1959). However, this still does not explain the mechanism by which UV rays induce neoplasms. Gene mutation is undoubtedly involved. IX. lung Cancer

One of the neoplasms in experimental animals for which genetic factors have been clearly shown is the lung tumor in the mouse. As early as 1926, Lynch published data showing an hereditary influence. Later, in a rather extensive study of both induced and spontaneous lung tumors, Heston (1942a,b) showed that the tumors resulted from multiple genetic and nongenetic factors with additive effects, the tumor

18

W. E. HESTON

appearing when the total effect passed a certain threshold. Thus, a carcinogen did not wipe out the genetic effect, but its influence was added to the genetic influence. For years geneticists avoided any genetic study of lung cancer in man, pointing out that its incidence was so low that a genetic study would be difficult and also that the disease in man was not comparable with the inherited disease in mice. While lung tumors in mice are alveolar in origin, most lung cancer in man is bronchogenic. With the great increase in lung cancer due to cigarette smoking and probably other factors, such as fumes from industry and automobiles, a genetic study became feasible. In 1963, Tokuhata and Lilienfeld reported on a comparison of relatives of 270 lung cancer probands with race-sex-age-residence matched controls. They observed a significant excess in lung cancer mortality among the proband relatives that could not be accoiinted for by age, sex, generation, and cigarette-smoking factors. The effect of familial factors among nonsmokers was similar in both men and women. The smoking factor appeared to be greater than the familial factor among men, but among women the familial factor appeared to be greater than the smoking factor. Their conclusion was that genetic factors may play a role in the etiology of lung cancer together with such environmental factors as cigarette smoking-a conclusion that might have been predicted from the experimental studies. Questions arose as to the possible relationship between the genetic susceptibility of the individual to lung cancer and his smoking habits. Several investigators in arguing against any causal relationship between cigarette smoking and lung cancer suggested that cigarette smoking and lung cancer might have a common genetic basis. Fisher (1958) concluded from his observations on smoking habits among twins that such was the case. To obtain further information on this question, Tokuhata (1963) obtained lifetime smoking experience data on his 270 lung cancer probands and their relatives and on the 270 controls and their relatives. He found evidence for a familial predisposition of the individual to the smoking habit, which he observed both in relation to and independentlv of the familial history of lung cancer. He found that lung cancer aggregates in families independentlv of smoking habit and that smokers aggregate in families independently of lung cancer. From this he concluded that the host susceptibilitv to cigarette smoking and the like susceptibility to cancer of the lung, both of which appeared to have a genetic basis, did not share a common genetic basis. Thus far, no virus has been shown to be causally related to lung tumors in mice, so one might not expect to find a virus in the etiology of cancer of this organ in man. If a lung cancer virus does occur in

GENETIC ASPECTS OF HUMAN CANCER

19

man, however, it may well be related to carcinogens that are involved. It may even be activated by them. It will certainly be controlled by the genetics of the individual. Therefore, the most likely individual in which to look for such a virus probably is the younger lung cancer patient with a strong familial factor and with a strong carcinogenic factor, such as cigarette smoking.

REFERENCES Aird, I., and Bentall, H. D. (1953). Brit. Med. J . 1,799-803. Anderson, D. E. ( 1959). In “Genetics and Cancer” (Staff of M.D. Anderson Hospital and Tumor Inst., eds.), pp. 364-374. Univ. of Texas Press, Austin. Anderson, D. E. (1972). J. Nut. Cancer Inst. 48, 1029-1034. Anderson, V. E., Goodman, H. O., and Reed, S. C. (1958). “Variables Related to Human Breast Cancer,” Univ. of Minnesota Press, Minneapolis. Bentvelzen, P. ( 1972). In “RNA Viruses and Host Genome in Oncogenesis” (P. Emmelot and P. Bentvelzen, eds.), pp. 309-337. North-Holland Publ., Amsterdam. Bloom, G. E., Warner, S., Gerald, P. S., and Diamond, L. K. (1966). N . Engl. J , Med. 274, 8-14. Burdette, W. J. ( 1970). In “Carcinoma of the Colon and Antecedent Epithelium” ( W. Burdette, ed. ) pp. 78-97. Thomas, Springfield, Illinois. Clarkson, B. D., and Boyse, E. A. (1971). Lancet 1, 699-701. Cleaver, J. E. ( 1968). Nature (London) 218,652-656. Cockayne, E. A. (1927). Cancer Reo. 2,337447. Cockayne, E. A. (1933). “Inherited Abnormalities of the Skin and its Appendages.” Oxford Univ. Press, London and New York. Cole, R. K., and Furth, J. ( 1941). Cancer Res. 1, 957-965. Comfort, M. W., Kelsey, M. P., and Berkson, J. ( 1947). J . Nut. Cancer Inst. 7, 367-373. Cumings, J. N., and Sorsby, A. (1944). Brit. 1. Ophthalmol. 28, 533-537. Dukes, C. E. ( 1952). Ann. Eugen., London 17, 1-29. El-Hefnawi, H. S., Smith, S. M., and Penrose, L. S. (1965). Ann. Hum. Genet. 28, 273-290. Epstein, J. H., Fukuyama, K., Reed, W. B., and Epstein, W. L. (1970). Science 168, 1477-1478. Fisher, R. A. ( 1958). Nature (London) 182, 108. Gardner, E. J. ( 1955). In “Novant’anni della Leggi Mendelione” (L. Gedda, ed.), pp. 321329. Istituto Gregorio Mendel, Rome. Gardner, E. J. (1962). Amer. J . Hum. Genet. 14, 376-389. Gardner, E. J., and Stephens, F. E. (1950). Amer. J. Hum. Genet. 2, 4148. German, J. ( 1974). “Chromosomes and Cancer.” Wiley, New York. Gross, L. ( 1951). Proc. SOC. Exp. Bid. Med. 78,342348. Hagy, G. W. (1954). Amer. J. Hum. Genet. 6,434-447. Haldane, J. B. S. (1936). Ann. Eugen. 7 , 2 8 5 7 . Hecht, F., Koler, R. D., Rigas, D. A., Dahnke, G. S., Case, M. P., Tisdale, V., and Miller, R. W. (1966). Lancet 2, 1193. Heston, W. E. (1942a). 1. Nut. Cancer Inst. 3,69-78. Heston, W. E. (1942b). J . Nut. Cancer Imt. 3, 79-82.

20

W. E. HESTON

Heston, W. E. ( 1973). Methods Cancer Res. 7, 115-129. Heston, W. E., and Vlahakis, G. (1961). J. Not. Cancer Inst. 26, 969-983. Heston, W. E., and Vlahakis, G. (1968). J. Nat. Cancer Inst. 40, 1161-1166. Heston, W. E., Deringer, M. K., Hughes, I. R., and Cornfield, J. (1952). J. Nut. Cancer Inst. 12, 1141-1157. Hirschhorn, K. (1970). Proc. Not. Cancer Conf. 6th, 1968 pp. 107-112. Hirschhorn, K., and Bloch-Shtacher, N. (1970). Proc. 23rd Annu. Symp. Fundam. Cancer Res. I969 pp. 191-204. Huebner, R. J., and Todaro, G. (1969). Proc. Not. Acad. Sci. US.64, 1087-1094. Jackson, E. W., Turner, J . H., Klauber, M. R., and Norris, F. D. (1968). J. Chronic Dis. 21, 247-253. Jackson Laboratory Staff. ( 1933). Science 78, 465-466. Jacobsen, 0. (1946). “Heredity in Breast Cancer,” Vol. 11. Lewis, London. Keeler, C. E. (1963). Nut. Cancer Inst., Monogr. 10,349359. Keeler, C . E. ( 1964). J. Hered. 55, 115-120. Key, D. J., and Todaro, G. J. (1974). J . Znoest. Dennutol. 62,7-10. Knudson, A. G. (1975). In “Cancer: A Comprehensive Treatise 1,” (F. F. Bedker, ed.), pp. 59-74. Plenum, New York. Korteweg, R. ( 1934). Ned. Tiidschr. Geneesk. 78, 240-245. Li, R. P., and Fraummi, J. F. ( 1969). Ann. Intern. Med. 71,747-752. Lilly, F., and Pincus, T. (1973). Adoan. Cancer Res. 17,231-277. Lynch, C. J. ( 1926). J . Exp. Med. 43,339-355. Lynch, H. T., Guirgis, H., Swartz, M., Lynch, J,, Krush, A. J., and Kaplan, A. R. ( 1973a). Arch. Surg. (Chicago) 106, 66-75. Lynch, H. T., Krush, A. J., and Guirgis, H. (1973b). Amer. J. Gastroentefol. 59, 31-40. Lynch, H. T., Krush, A. J., Harlan, W. L., and Sharp, E. A. ( 1 9 7 3 ~ ) .Amer. Surg. 39, 199-206. Lynch, H. T., Guirgis, H. A., Albert, S., Brennan, M., Lynch, J., Kraft, C., Pocekay, D., Vaughns, C., and Kaplan, A. (1974). Surg., Gynecol. Obstet. 138,717-724. MacDowell, E. C., Potter, J. S., and Taylor, M. J. (1945). Cancer Res. 5, 65-83. Macklin, M. T. (1936). Arch. Dermatol. Syph. 34, 656-675. Macklin, M. T. (1959). I.Nut. Cancer Inst. 22,937-952. Macklin, M. T. (1960). J. Nut. Cancer Znst. 24,551571. MacMahon, B., and Levy, M. A. (1964). N . Engl. J . Med. 270, 1082-1085. Martynova, R. P. ( 1937). Amer. J . Cancer 29,530-540. Miller, R. W. (1963). N. EngZ. J . Med. 268,393-401. Morton, N. E. ( 1957). Amer. J . Hum. Genet.9,55-75. Mosbeck, J. (1953). O p . Dom. Biol. Hered. Hum. 34, 108. Muhlbock, O., and D u x , A. ( 1974). J . Nut. Cancer Inst. 53,993-996. Ned, J. V., and Falls, H. F. ( 1951). Science 114, 419-422. Nowell, P. C., and Hungerford, D. A. (1960). Science 132, 1497. Oliver, C. P. (1959). In “Genetics and Cancer” (Staff of M.D. Anderson Hospital and Tumor Inst., eds.), pp. 427-438. Univ. of Texas Press, Austin. Orye, E., Delbeke, M. J., and Vandenabeele, B. ( 1971). Lancet 2, 1376. Penrose, L. S., and Smith, G. F. (1966). “Down’s Anomaly.” Little, Brown, Boston, Massachusetts. Penrose, L. S., Mackenzie, H. S., and Kam, M. N. (1948). Ann. Eugen., London 14, 234-266. Reed, T. E., and Neel, J. V. ( 1955). Amer. J. Hum. Genet. 7, 236-263.

GENETIC ASPECIS OF HUMAN CANCER

21

Robbins, J. H., and Burk, P.G. ( 1973). Cancer Res. 33, 929-935. Robbins, J. H., Kraemer, K. H., Lukner, M. A., Festoff, B. W., and Coon, H. G . (1974). Ann. Intern. Med. 80,221-248. Rowe, W. P., Hartiey, J. W., and Bremmar, T. (1972). Science 178, 860-862. Sandberg, A. A., Ishihara, T., Kikuchi, Y.,and Crosswhite, L. H. (1964). Ann. N.Y. ACUU!. Sci. 113,663-716. Sawitsky, A., Bloom, D., and German, J. (1966). Ann. Intern. Med. 65, 487495. Smith, S. M., and Sorsby, A. (1958). Ann. Hum. Genet. 23,50. Sorsby, A. (1972). Brit. Med. J. 2,580583. Steinberg, A. G. (1960). Cancer 13, 985-989. Stewart, A. ( 1961). Brit. Med. J. 1,452-460. Stewart, S. E., Eddy, B. E., and Borgese, N. (1958). J . Nut. Cancer Inst. 20, 1223-1243. Todaro, G. J., and Martin, G. M. (1967). Proc. SOC.Ezp. Bid. Med. 124,1232-1236. Todaro, G. J., Green, H., and Swift, M. R. (1966). Science 153,1252-1254. Tokuhata, G. K. (1963). J. Nut. Cancer Inst. 31, 1153-1171. Tokuhata, G. K., and Lilienfeld, A. M. (1963). J. Nut. Cancer Inst. 30, 289412. Videbaek, A. ( 1947). “Heredity in Human Leukemia and its Relation to Cancer.” Lewis, London. Videbaek, A., and Mosbeck, J. (1954). Acta Med. Scand. 149,137-159. Wassink, W. F. ( 1935). Genetica 17, 103-144. Woolf, C. M. (1955). “Investigations on Genetic Aspects of Carcinoma of the Stomach and Breast,” Pub. Health, Vol. 11, pp. 265-350. Univ. of California Press, Berkeley. Woolf, C. M. (1956). Amer. J. Hum. Genet. 8, 102-109. Woolf, C. M. (1958). Amer. J . Hum. Genet. 1 0 , 4 2 4 7 . Woolley, G. W., Dickie, M. M., and Little, C. C. (1952). Cancer Res. 13, 231-245.

This Page Intentionally Left Blank

THE STRUCTURE AND FUNCTION OF INTERCELLULAR JUNCTIONS IN CANCER

.

Ronald S Weinstein. * Frederick

B . Merk.

and Joseph Alroy’

Departments of Pathology. Tufts University School of Medicine. Boston. Massachusetts and Rush Medical College. Chicago. Illinois

I. Introduction . . . . . . . . . . I1. Membrane Ultrastructure . . . . . . . A . Principal Electron Microscopy Techniques . . . B. General (Nonjunctional) Plasma Membranes . . . I11. Cell Junction Classification . . . . . . . A . Occludentes Junctions . . . . . . . B. Gap (“Nexus”) Junctions . . . . . . . C . Adherentes Junctions . . . . . . . . D. Junctional Complexes . . . . . . . . E. Miscellaneous Junctions . . . . . . . N . Occurrence of Cell Junctions in Tumors . . . . A. Solid Tumors (and Nonmalignant Growth Disorders) . B. Tissue Culture . . . . . . . . . V. Cell-to-Cell Communication and Growth Control . . . A. Introduction . . . . . . . . . . B. Intimate Communication at Gap Junctions . . . C . Control of Gap Junction Permeability . . . . D. Metabolic Coupling at Gap Junctions . . . . E. Coupling between Tumor Cells . . . . . . F. Genetic Correlations . . . . . . . . VI. Cell Junctions in Embryonic Development . . . . VII . Cell Junctions and the Biological Behavior of Tumor Cells . A Contact Inhibition of Movement (Locomotion) . . B. Postconfluence Inhibition of Growth (Cell Division) . C. Invasion and Metastases . . . . . . . VIII . Intercellular Adhesion in Tumors . . . . . . A Adhesion at Cell Junctions . . . . . . . B. Tumor Dissemination . . . . . . . . IX. Transepithelial Permeability and Malignant Transformation X . Summary . . . . . . . . . . . References . . . . . . . . . . .

.

.

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

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

. .

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

23 26 26 30 32 32 35 39 42 43 43 43 60 61 61 62 64 66 67 69 70 71 71 74 74 75 75 77 77 77 79

I. Introduction

One of the fundamental problems in cancer research is to determine the cell products coded for by the genes of neoplastic transformation * Present address: Rush Medical College and Rush.Presbyterian.St . Luke’s Medi-

cal Center. Chicago. Illinois.

23

24

RONALD S . WEINSTEIN, FREDERICK B. MERK, AND JOSEPH ALROY

and to identify the product(s) that accounts for the autonomous and destructive growth that is characteristic of malignant tumors. The cell surface membrane, which in large measure is a product of gene expression, probably plays a crucial role in the control of normal growth (Pardee, 1961; Pardee et al., 1974; Rapin and Burger, 1974). Malignant tumors are characterized by abnormalities in the regulation of their growth and by their invasiveness. Changes in the cell surface membrane may be related either indirectly or directly to these properties of tumors. A broad spectrum of structural, compositional, and functional membrane alterations appear in the course of malignant transformation, and it is becoming increasingly evident that individual instances of malignant transformation may involve a multiplicity of these membrane changes (Pollack and Hough, 1974; Nachbar et al., 1974). This multiplicity substantially complicates the task of correlating a single membrane change with any specific aspect of biological behavior. Cell junctions, a clearly recognizable set of membrane components, are frequently altered in the course of malignant transformation. These alterations may account for some significant properties of malignant cells. At present, research on the normal function and disposition of cell junctions and on the relationships of junctional changes to malignancy is still at a data-gathering stage. In our opinion, it is premature to attempt a theoretical consolidation of the tumor cell-junction field. Instead, our purpose in this review will be to summarize available data on junctions, and to outline and examine some of the hypotheses that have been used to interpret these data. Cell junctions are defined as structurally specialized domains that are formed at regions of contact between two cells and to which both cells contribute a part. A number of different types of cell junctions have been identified: adherens, occludens, “gap,” septate, etc. It is standard practice to consider these various cell junctions together since they are all components of the cell surface membrane, provide a structural means for cell-cell interactions, and enable cells to form compex multicellular structures. Cell junctions are customarily subclassified on the basis of their ultrastructure (Farquhar and Palade, 1963; Brightman and Palay, 1963; Weinstein and McNutt, 1972), as will be discussed in Section 111. Cell junctions perform the following functions: 1. They provide strong structural links between cells. The cells in turn form mechanically coherent tissues. 2. They serve as conduits through adjoining membranes of cell pairs. By this mechanism, junctions mediate direct communication between cells by allowing the transfer of ions and small molecules from cell to cell without leakage into the extracellular space.

INTERCELLULAR JUNCI’IONS IN CANCER

25

3. They seal cells together into a coherent tissue that can act as a highly selective barrier to diffusion. 4. They mediate the unidirectional propagation of electrochemical impulses from one cell to another. Some of these functions are the responsibility of a combination of specific types of cell junctions. There is a considerable body of evidence to support the notion that cell junctions are particularly important with respect to the cancer problem. First, many studies show that cell junctions contain structural components that are probably gene products, and that junction formation is genetically controlled (Campbell and Campbell, 1971; Azarnia et al., 1974). Second, all categories of cell junctions contribute to cell-to-cell adhesion ( McNutt and Weinstein, 1973), a property that is often altered in malignant tumors (Coman, 1944, 1961). Third, cell junctions may be involved in growth regulation (Loewenstein and Penn, 1967; Loewenstein, 1968a,b), and alterations in junctions could conceivably account for some of the growth abnormalities in tumors (Jamakosmanovic and Loewenstein, 1968; Loewenstein, 1974). Last, one type of junction, the occludens (i.e., “tight”) junction, modulates transepithelial permeability. Structural and functional changes in occludens junctions induced by exposing tissues to any one of a number of noxious stimuli could increase the flow of certain carcinogens, promoters, or cofactors to target cells within an epithelium. To date there is no irrefutable proof that cell junctions or, for that matter, any cell membrane components are causally related to malignant behavior. However, comparison of some of the properties of normal and malignant tissues inevitably focuses attention on cell junctions as possible candidates for an important role in malignancy since cell junctions provide a primary mechanism for various cell-cell interactions in tissues. In this review, we shall consider the ultrastructure, biochemistry and function of cell junctions in both normal and malignant tissues, emphasizing information that may be particularly pertinent to the cancer problem. Recent advances have been made in the development of methodology for isolating, purifying, and characterizing cell junctions. This work is of vital importance and is relevant to the cancer effort since it generates fundamental information about the molecular organization of cell junctions and also provides a framework for mechanistic studies on cell junction formation in tumors. Therefore, we will survey significant advances in cell junction physical-chemical characterization. We will also discuss the frequency of occurrence and potential significance of junctional abnormalities in tumors in relation to tumor biological behavior. At the outset we shall briefly outline data relating to the biochemical

26

ROSALD S. WEINSTEIN, FREDERICK B. MERK, AND JOSEPH ALROY

ultrastructure of cell junctions, since progress recently has been made in this important area and many nonmembranologists working in the cancer field may be unacquainted with the electron microscopy literature. The reader is referred to several comprehensive reviews for more detailed discussions of cell junctions, their ultrastructure and function ( McNutt and Weinstein, 1973; Satir and Gilula, 1973; DeHaan-Sachs, 1973; Staehelin, 1974; Overton, 1974 ).

It. Membrane Ultrastructure

A. PRINCIPAL ELECTRON MICROSCOPY TECHNIQUES The majority of reports dealing with cell junctions in tumors are electron microscopy studies. Three principal electron microscopy preparative procedures, thin sectioning, negative staining, and freeze fracturing, are commonly used to study the ultrastructure of the general plasma membrane and cell junctions. 1. Thin-Section Electron Microscopy For thin sectioning, tissue blocks are first stabilized by various chemical fixatives, dehydrated, and then embedded in plastic. A two-dimensional image representing a single plane within the specimen is obtained by examining stained sections of embedded material with a transmission electron microscope. Thin sections of most biological materials in an unstained state are virtually electron lucent. To enhance image contrast, specimens are exposed to heavy-metal ions (“electron dense” stains). The heavy-metal ions bind to cell structures that have an affinity for the stains. Most of the image contrast visible in the transmission electron microscope is due to electron scattering caused by the heavy-metal stains. Thin-section electron microscopy of the general plasma membrane of many cells reveals a characteristic triple-layered structure at the cell periphery. Membranes can be resolved into two 2.5-3.5 nm electronopaque lines separated by an electron-lucent space that varies in thickness from 3.5 to 7.5 nm, depending on the cell type, This trilaminar structure is generally referred to as the “unit” membrane, a term introduced by Robertson to emphasize that all three layers in the 7-15 nm structure are components of a single membrane, and also to suggest that membranes of diverse biological systems share this structure ( Robertson, 1959, 1964 ). Many investigators have attempted to deduce information about the molecular architecture of membranes from the unit-membrane image. However, it is now generally acknowledged that

INTERCELLULAR JUNCTIONS IN CANCER

27

the triple-layered image is essentially an artifact of electron microscopy preparative techniques and that it cannot serve as a basis for a precise interpretation of membrane structure at the molecular level (Korn, 1966; Stoeckenius and Engelman, 1969). The “unit” membrane image is a useful indicator of the location of the main permeability barrier at the cell surface, and it may give a rough approximation of the thickness of the diffusion barrier ( McNutt and Weinstein, 1973). In thin sections, the membrane components of cell junctions of all types appear as triple-layered unit membranes. Many recent studies employing the freeze-fracture technique show substantial heterogeneity in the central plane of junctional membranes. The discrepancy between these observations on freeze-fractured junctions and thin-section data may be attributed to section thickness, and to the disorganization and/or removal of membrane components during fixation, dehydration, and the embedding processes (Singer, 1962; Korn and Weisman, 1966; Lenard and Singer, 1968; Moretz et al., 1969).

2. Negative Staining The negative staining technique is useful for studying selected aspects of membrane fine structure. For negative staining, membranes are isolated and purified, washed, and then dried onto thin support films in the presence of a highly soluble heavy-metal salt, such as phosphotungstic acid. During the final phase of drying, the heavy-metal salt precipitates as extremely fine, clectron dense crystals. The appearance of negatively stained membrane structures is related to the distribution of the small stain crystals at the membrane. In general, membrane components which are impermeable to the salt appear electron lucent and stand out in negative contrast to adjacent permeable regions that contain stain. Negative staining is particularly useful for visualizing membrane components that protrude from the membrane surface but may reveal other components as well. The interpretation of negative stain data is frequently hampered by a lack of knowledge about the degree of penetration of the stain into the membrane. Since some globular structures are seen exclusively in an en face view in negative stain preparations the exact location of these components within the membrane can be difficult to resolve. Interpretation of negative stained preparations is further complicated by artifacts that may be introduced during isolation of cell membranes and the subsequent dehydration that occurs while labile membranes are exposed to increasing concentrations of the heavy metal salt during the staining procedure itself. The negative staining technique has demonstrated novel components in general plasma mem-

28

RONALD S.

WEINSTEIN, FREDERICK B.

MERK, AND JOSEPH ALROY

branes (Haggis, 1969), as well as a number of structural elements in cell junctions (Benedetti and Emmelot, 1968a,b; Goodenough and Revel, 1970).

3. Freeze-Fracturing The freeze-fracture technique circumvents some of the technical limitations of thin-sectioning and negative staining ( Bullivant, 1970). It is particularly useful as a tool for probing biological membrane ultrastructure since it provides a unique view of the internal organization of membranes. In freeze-fracturing, biological specimens are rapidly frozen and then cleaved at low temperatures (-loo0 to -196°C). Replicas of the fracture faces are prepared in oucuo by deposition of heavy metal (e.g., platinum) at an oblique angle onto the frozen and freshly cleaved surface of the specimen (Moor, 1966). The result is a coherent metal film that is reasonably representative of the ultrastructural topography at the fracture face. After casting of the replica the original specimen is first digested and then removed by washing. The replica is examined in a conventional transmission electron microscope ( Moor, 1966; Bullivant, 1973). An ancillary step variously referred to as “freeze-etching,” “heat etching,” or “deep etching” can be added to the freeze-fracture procedure. This step is equivalent to freeze-drying. It is used to sublime water-ice away from membranes and to demonstrate the true outer surfaces of membranes at the specimen-vacuum interface. In general, freeze-fracturing splits open membranes along planes of low mechanical resistance within their hydrophobic interior ( Branton, 1966; Deamer and Branton, 1967). The cleaving process generates novel fracture faces that represent aspects of the internal structure of the membrane ( Branton, 1969; Wehrli et ul., 1970; Weinstein et a?.,1970a; Chalcroft and Bullivant, 1970) at a resolution close to the molecular level. McNutt and Weinstein (1970) introduced several terms to describe membrane interior faces and the natural surfaces that are revealed by freeze-fracturing and freeze-etching. This nomenclature has emerged as the popular convention for referring to membrane fracture faces and heat-etched surfaces ( McNutt and Weinstein, 1973; Staehelin, 1974; Gilula, 1974; Bullivant, 1974), although at the time of this writing, a new nomenclature for freeze-fractured membranes is being devised ( D. Branton, personal communication ).I According to McNutt and Weinstein (1970), the fracturing process divides the cell membrane into two lamel‘ A new nomenclature for freeze-etching has now been published (Branton et al., 1975). According to this system: Face-A becomes Face-PF; Face-B becomes FaceEF; Face C becomes Surface-€%; Face-D becomes Surface-ES; Lamella-1 becomes the P-half membrane: and Lamella-2 becomes the E-half membrane.

29

INTERCELLULAR JUNCXIONS IN CANCER

lae: Lamella-1 ( LM-1 ), the juxtacytoplasmic membrane IamelIa that is in contact with the cell interior; and Lamella-2 (LM-2), which is in contact with the extracellular compartment ( Fig. 1).Two novel fracture faces ( A and B ) are generated by membrane fracturing, and the natural surfaces or faces ( C and D ) are revealed exclusively by freezeetching. Fracture face A is the fractured surface of the inner lamella LM-1 and fracture face B is the fractured surface of the outer lamella, LM-2. The natural surfaces of the membrane are: face-C, the inner (juxtacytoplasmic) surface, and face-D, the outer surface of the membrane which abuts the extracellular milieu. Fracture faces A and B of metabolically active nonjunctional plasma membranes bear a population of small particles called membrane-associated particles or intramembrane particles (MAP or IMP). As a rule,

--

Lm, Lmp Lmz Lm, -41

Face C -

-Face A -Face

B

-Face

c

Face D

f l t

PCP Intercellular PCP compartment

I

Intercellular compartment

FIG.1. These diagrams illustrate the relative positions of membrane lamellae and surfaces that are demonstrated at gap (nexus) cell junctions by the freeze-fracture technique. Left: An uncleaved junction in cross section. When each junctional membrane is fractured along the potential cleavage plane (PCP), which lies within the interior of the membrane, it is split into two lamellae, LMI and LM,. As seen on the right, fracture faces A and B are generated by the fracturing process. Reprinted from McNutt and Weinstein (1970), with permission of the Rockefeller University Press.

30

RONALD S. WEINSTEIN, FREDERICK B. MEW, AND JOSEPH ALROY

metabolically active enzyme-rich membranes contain more of these particles than relatively inert membranes, such as myelin (Branton and Deamer, 1972). It has been proposed on the basis of a considerable body of indirect evidence that smooth areas of fracture faces correspond to lipid bilayer regions of membranes and that the MAP correspond to proteins intercalated within the bilayer (Branton, 1969; Hong and Hubbell, 1972; Singer and Nicolson, 1972).

B. GENERAL ( NONJUNCXONAL) PLASMA MEMBRANE Cell junctions are structurally distinct domains embedded within the general plasma membrane. However, they share certain ultrastructural characteristics with the general plasma membrane ( McNutt and Weinstein, 1973). This is predictable, since junctions are formed by the stepwise modification of the general plasma membrane (Overton, 1962; Campbell and Campbell, 1971; Krawczyk and Wilgram, 1973; Johnson et al., 1974). Since cell junctions are specializations within the general plasma membrane and are structurally related to it, a brief description of general plasma membrane architecture will serve as a useful point of departure for our discussion of cell junctions. Many contemporary ideas about the general plasma membrane are summarized in the fluid-mosaic membrane model (Fig. 2) of Singer and Nicolson (1972). This model depicts the major membrane components, lipids, proteins, and oligosaccharides (not shown), in a lowest free-energy configuration. It incorporates one of the central themes of the earlier Danielli-Davson-Robertson membrane models by placing a lipid bilayer within the membrane. The hydrophobic groups of amphipathic lipids (mainly phospholipid) are sequestered within the interior of the bilayer while hydrophilic groups reside at exterior (natural) surfaces, where they can interact strongly with the aqueous phase. Singer (1971) introduced two terms to describe proteins associated with the membrane lipid bilayer : integral proteins and peripheral proteins. Integral proteins are amphipathic, having an ionic exterior segment in contact with water at the membrane surface and a hydrophobic segment intercalated within the membrane lipid matrix. Presumably, the extent to which an integral protein penetrates the membrane is determined by thc amino acid sequence and covalent structure of the protein molecule as well as by its interactions with the surrounding microenvironment. Some but probably not all integral proteins extend through the entire thickness of the lipid bilayer within the membrane (Weinstein and McNutt, 1970b; Marchesi et al., 1972; Steck, 1974). Peripheral proteins associate with the membrane surface by electrostatic interactions and other weak bonds. Removal of integral proteins from membranes by

INTERCELLULAR JUNCTIONS IN CANCER

31

FIG.2. Fluid-mosaic membrane model. The bulk of the phospholipids are organized in a discontinuous lipid bilayer (solid circles represent polar head groups; and wavy lines, their fatty acid chains). Integral proteins are embedded in the lipid bilayer but can protrude from it. Peripheral proteins may bind to phospholipid polar headgroups or to the membrane via protein-protein interactions. The arrow shows the position of a natural cleavage plane within the membrane. Adapted from Singer and Nicolson ( 1972 ) .

experimental manipulation requires drastic treatment with chemicals, such as detergents and organic solvents, whereas peripheral proteins can usually be dissociated from membranes by the addition of chelating agents or other mild treatments. On the basis of thermodynamic considerations, Singer and Nicolson ( 1972) suggested that, under physiological conditions, membrane lipids and some integral proteins may be in a fluid state although there probably are constraints on the lateral mobility of at least some integral proteins (Nicolson, 1973; Elgsaeter and Branton, 1974; Peters et al., 1974; Marikovsky et al., 1976). Examples of integral proteins (Bretscher, 1972; Steck et al., 1971; Marchesi et al., 1972) and peripheral proteins (Marchesi and Steers, 1968) have been purified from red cell ghosts. Membrane oligosaccharide residues are present at the exterior surface of the plasma membrane as constitutents of proteoglycans and glycolipids ( Winder, 1970). General plasma membranes, and presumably junctional membranes, are constantly in a state of flux owing to synthetic and degradative processes (Gurd and Evans, 1973). As recently reviewed by Rapin and Burger (1974), the composition and molecular architecture of the general plasma membrane varies with mitotic cycle, level of differentiation, cell microenvironment, metabolic state, and other factors. These variables

32

RONALD S. WEINSTEIN, FREDERICK B. MERK, AND JOSEPH ALROY

must be considered in efforts to interpret quantitative membrane changes in pathological states. 111. Cell Junction Classification

Cell junctions can be grouped into two major categories (Weinstein and McNutt, 1972). The first category includes junctions at which surface membranes of neighboring cells come into direct contact (e.g., occludentes junctions, “gap” or nexus junctions, and probably septate desmosomes). These junctions can be further subclassilied according to differences in the components within the interior of the junctional membranes. The second general category includes junctions where the surface membranes of adjacent cells are separated by a 15-35-nm interspace (e.g., adherentes junctions). This interspace typically contains electron-dense proteinaceous material. Junctions in this category can be subclassified on the basis of the morphology of cytoplasmic and extracellular matefial associated with the membrane at the junction. Cell junctions in both of these general categories are further subclassified on the basis of their overall shape and size. Latin terms, introduced into the cell junction literature by Farquhar and Palade, are used to describe the shape and total area of membrane-to-membrane contact at junctions. The term “zonula” (plural, zonulae) is used to describe junctions that extend as a belt around the entire cell. “Fascia” (plural, fasciae) is used to describe a junction that forms an extensive sheetlike area of attachment that does not completely encircle the cell. A junction that is a single spot or disk-shaped area of attachment is termed a “macula” ( Farquhar and Palade, 1963) . Cell junctions in normal tissues and in tumors are generally indistinguishable at the ultrastructural level although they may differ with respect to overall size, distribution at the cell surface, level of development, and numerical density ( McNutt and Weinstein, 1969; Martinez-Palomo, 1970a; Wiernik et al., 1973). Because of the apparent similarities of junctions in normal cells and cancer cells, their ultrastructure, biochemistry, and function will be discussed together. In view of available information on the many compositional and structural alterations in the general plasma membrane that occur during malignant transformation, it seems reasonable to anticipate that biochemical differences will be found for tumor cell junctions as well. However, the development of cell junction isolation methods is in its infancy (Benedetti and Emmelot, 1968a; Goodenough, 1974; Skerrow and Matolsty, 1974a), and, to date, available isolation methods have not been used to compare normal junctions with tumor cell junctions.

INTERCELLULAR JUNCITONS IN CANCER

33

A. OCCLUDENTES JUNCTIONS

Occludentes junctions occur in many epithelia and endothelia. 1. Ultrastructure

Unit membranes of adjacent cells come into immediate contact at zonulae, fasciae, and maculae occludentes junctions, and the outer leaflets of the membranes appear to “fuse,” thereby eliminating the extracellular space ( Figs. 3 and 7). Freeze-fracture replicas of occludentes junctions reveal fibrils 6-8 nm in diameter within the interiors of the junctional membranes (Weinstein et d.,1970b; Chalcroft and Bullivant, 1970) that correspond in their distribution to the lines of union of the membranes (Chalcroft and Bdlivant, 1970; Staehelin, 1973). An occludens junction is the only type of cell junction known to contain long intramembrane fibrils. Single intramembrane fibrils have been further resolved into two 3-4-nm filaments which lie side-by-side in the plane of the membrane ( McNutt and Weinstein, 1973). In many epithelia (including some endothelia), the occludens junction forms a belt that completely encircles the perimeter of the cell and is therefore called a zonula occludens (Farquhar and Palade, 1963). The occludentes junctions of some tissues are discontinuous and are thus designated maculae or fasciae occludentes. Comparative ultrastructural studies on unfixed and aldehyde-fixed tissues provide indirect evidence that the intramembrane fibrils at occludentes junctions may be integral membrane proteins. In unfixed tissue, membrane splitting by freeze-fracturing breaks the fibrils into a series of short segments that can be attached to either fracture face A or B. After prefixation with glutaraldehyde, a bifunctional reagent that cross-links membrane proteins at amino, imino, and guanidino groups, individual fibrils resist fragmentation during freeze-fracturing and selectively remain attached to fracture face A. This strengthening of their structure by glutaraldehyde fixation supports the notion that the fibrils contain proteins (Weinstein et al., 1970b; McNutt and Weinstein, 1973; Staehelin, 1973). 2. Isolation

Occludentes junctions have not been isolated. 3. Functions The primary function of the zonulae occludentes is to control transepithelial permeability. A secondary function is to contribute to cell-cell adhesion. Because the junction occludes the interspace between adjacent cells, it forms a seal between them preventing “bypass diffusion.” Zonulae

34

RONALD S. WEINSTEIN, FREDERICK B. B R K , AND JOSEPH ALROY

. A OCCLUDENS

A

ADHERENS

.A

ADHERENS

J l1 NCTI ON

FIG.3. Highly schematic representation of the intercellular junctions between two simple columnar epithelial cells. Surface “A” is a representation of freeze-fractured plasma membrane A-face. The general ( nonjunctional plasma membrane bears a sparse population of membrane-associated particles ( MAP) which may represent integral membrane proteins. The A-face fine structure of four types of intercellular junctions are sho\vn. (See text for details.) Plane X represents the thin section appearance of the same types of cell junctions. Adapted from Weinstein and McNutt (1972).

occludentes form a physical barrier between the lumen of cavitary organs and the adluminal compartments. In some tissues this barrier may be totally impermeable to molecules, small ions, and even water whereas in other tissues zonulae occludentes are relatively leaky. The extent of occludens junction formation (Farquhar and Palade, 1963; Friend and Gilula, 1972b) is tissue specific, and junction ultrastructure correlates well with transepithelial bypass diffusion in some tissues (Claude and

INTERCELLULAR JUNCTIONS IN CANCER

35

Goodenough, 1973) but not others ( Martincz-Palomo and Erlij, 1975). Zonulae occludentes junctions, together with mcrnbrane pumps, enable sheets of cells (e.g., intestinal epithcliurn) to create and maintain steep chemical and clcctrical gradients between compartments in various organs. Maculae and fasciae occludentes are usually found in endothelia. They morphologically resemble zonulae occludentes ( Simionescu et d.,1975; Staehelin, 1975), but form an incomplete network at the perimeter of cells (Karnovsky, 1967; Weinstein and McNutt, 1970a). The discontinuities in maculae or fasciae occludentes junctions provide shunt pathways for bypass diffusion of molecules and ions around endothelial cells (Karnovsky, 1967; Fromter and Diamond, 1972; Whitternburg and RawIins, 1971; Diamond, 1974; Simionescu et uZ., 1975). Experimentally induced changes (see Section I X ) , in the configuration and distribution of occludens junction structural elements (Wade and Kamovsky, 1974) can have a profound influence on the overall permeability across epithelium (Wade et al., 1973).

B. GAP (“NEXUS”)JUNCTIONS Gap junctions occur in many normal tissues and in tissue culture systems. They may join homotypic, heterotypic (Michalke and Loewenstein, 1971), or heterologous cells in coculture (Stoker, 1967). In these settings their junction ultrastructure is the same (Johnson et al., 1973). They also occur in benign and malignant tumors (see Section I V ) . 1. Uttrastructure A number of studies provide detailed information on gap junction fine structure (Robertson, 1963; Revel and Kamovsky, 1967; Benedetti and Emmelot, 1968a,b; McNutt and Weinstein, 1970; Goodenough and Revcl, 1970, 1971; Chalcroft and Bullivant, 1970; Goodenough and Stoeckenius, 1972; Merk et al., 1973; Peracchia, 1973a,b). Mch’utt and Weinstein ( 1970) have attempted to summarize the detailed ultrastructurar data obtained with several electron microscopy techniques in the form of a model. The data upon which the model was originally based have been reviewed in detail ( McNutt and Weinstein, 1973) and rccent studjes provide additional support for the model (Goodenough and Gilula, 1974). For our present purpose it shall suffice to briefly mention selccted aspects of gap junction ultrastructure and survey the essential features of the McNutt-Weinstein model. The reader is referred to recent reiriews for a complete discussion of the gap junction ultrastructure literature ( McNutt and Weinstein, 1973; Staehelin, 1974; Gilula, 1974). In conventional thin sections, the gap junction bears a superficial re-

36

RONALD S .

WEINSTEIN,

FREDERICK B. MEW, AND JOSEPH ALROY

semblance to the occludens junction since both appear as regions where the plasma membranes of neighboring cells come into intimate apposition. However, there are differences. The occludens junction has an overaIl thickness of less than twice that of the general plasma membrane whereas the gap junction has a thickness of more than twice that of the general plasma membrane ( Weinstein and McNutt, 1972). The increased thickness of the gap junction is due to the presence of a minute 2-nm “gap” between the membrane outler leaflets (Revel and Karnovsky, 1967). Special tracer studies show that the so-called “gap” is actually spanned by membrane components ( McNutt and Weinstein, 1973) which provide the structural basis of membrane-to-membrane contact. Freeze-fracturing demonstrates that the internal organization of gap junctional membranes is different from that of the general plasma membrane in that it contains ordered globular subunits (Chalcroft and Bullivant, 1970; Goodenough and Revel, 1970; McNutt and Weinstein, 1970). In the McNutt-Weinstein model, the gap junction is pictured as a bipartite interlocking array of subunits which receives equal structural contributions from each partner of the cell pair (Fig. 4 ) . As seen by freeze-fracturing, gap-junction membranes contain particles 4-5 nm in diameter on fracture face A. These particles often appear in a hexagonal array with a center-to-center spacing of approximately 9-10 nm. A 2.0-2.5 nm central depression is frequently observed at the apex of the particles. McNutt and Weinstein suggest that this central depression may represent a segment of a hydrophilic channel which is thought to extend through the gap junction ( McNutt and Weinstein, 1970; Johnson et al., 1974). Face B of gap junction membranes displays a polygonal arrangement of pits which house the face-A particles prior to membrane fracturing. In some tissues the particles are attached to the B face and the pits are located on the A face (see Staehelin, 1974, for a discussion of additional nomenclature). Face D of the outer lamella (LM-2) of the junctional membrane is organized as foot processes that bridge the extracellulax space and join foot processes projecting from the surface of the neighboring cell. A network of channels, which are an extension of the extracellular compartment, course around the foot processes ( Payton et al., 1969; McNutt and Weinstein, 19’70).When sectioned transversely (in the plane of the membrane) the foot processes are circular, approximately 7 nm in diameter, and have essentially the same center-to-center spacing as the intramembrane particles in freeze-fracture replicas. Morphological variants of the typical gap junction have been described (Staehelin, 1973; Peracchia, 1973a,b; Gilula, 1972; Satir and Gilula, 1973; Raviola and Gilula, 1973; Albertini and Anderson, 1974; Kogon and Pappas, 1975).

’

INTERCELLULAR JUNCTIONS IN CANCER

37

FIG.4. The McNutt-Weinstein Model: An interpretation of the dtrastructure of the Hap junction. (Left) The junctional membranes (Ma and M,) are shown pulled back to demonstrate the appearance of foot processes that normally project from each membrane and meet in the extracellular space at the midline. (Right) Aspects of the gap junction are shown as they appear in freeze-fracture replicas (see text for details). Intramembrane particles, which partially ( ? or completely) span the junctional membranes are present on the A face. The inset illustrates an interpretation of the position of hydrophilic channels within the gap junction. These channels may provide the structural basis for low-resistance electrotonic coupling of neighboring cells. The model is not drawn to exact scale. Reprinted from Mch’utt and Weinstein (1970), with permission of the Rockefeller University Press.

38

RONALD S .

WEINSTJ3N, FREDERICK

B. MERK, AND JOSEPH ALROY

2. Zsolation and Characterization Gap junctions are remarkably resistant to chemical and mechanical stress (Berry and Friend, 1969), a characteristic that is exploited in gap junction isolation methods ( Benedetti and Emmelot, 1968a,b; Goodenough, 1974). Benedetti and Emmelot (1968a) were the first to successfully concentrate gap junctions. They isolated liver plasma membranes according to the method of Neville (1!360), and then, by means of a brief treatment with Ahe detergent deoxycholate, preferentially solubilized the general plasma membrane while leaving the gap junctions intact. These isolated junctions were useful for ultrastructural studies, although they were contaminated with amorphous nonjunctional material which limited their usefulness for biochemical analysis. Other isolation procedures have been described by ,Evans and Gurd ( 1972), Zampighi and Robertson (1973), and Dunia et al. ( 1974). In 1972, Goodenough and Stoeckenius developed a method for preparing mouse hepatocyte gap junctions in a more highly purified form. They treated cell membranes with collagenase and then selectively solubilized the nonjunctional membrane with the detergent Sarkosyl NK-97 and ultrasonification. They separated the gap junctions from the amorphous nonjunctional debris by sucrose gradient ultracentrifugation. Goodenough and Stoeckenius produced only small amounts of purified gap junctions by this method, but their yield was adequate for a preliminary analysis of gap-junction biochemistry and structure. Thin-layer chromatography of the gap-junction fractions revealed one major phosopholipid and some neutral lipids. Low-angle X-ray diffraction of both wet and dry isolated gap junctions showed reflections which index on an 8.6 nm center-to-center hexagonal lattice, corresponding to the center-to-center spacing of intramembrane gap junction particles. Protein analyses with polyacrylamide-gel electrophoresis of sodium dodecyl sulfate ( SDS) solubilized gap junctions showed one major and two minor protein bands ( Goodenough and Stoeckenius, 1972). Recently, Goodenough ( 1974) modified the Goodenough-Stoeckenius method and achieved yields of up to milligram quantities. Goodenough's procedure employs continuous-flow centrifugation and a discontinuous sucrose gradient in a zonal rotor. The fraction of purified gap junctions is solubilized in SDS and then analyzed for protein content in electrophoretographs. Analysis of the gap-junction fraction shows a rather simple protein profile as compared to that of the hepatocyte nonjunctional membrane. There are two major protein peaks, one at 34,000 daltons and the second at 18,000 daltons plus two minor bands at 10,000 daltons. After exposure to disulfide reducing agents, most of the protein migrates as a doublet to the 10,000-dalton position. On the basis of these findings, Goodenough suggests that gap-junction proteins are composed of two

INTERCELLULAR JUSCTIONS IN CANCER

39

small polypeptides with molecular weights of about 10,000 daltons and that the polypcptides arc linked covalently either by interpeptide or intrapeptide disulfide bonds. The names “Connexin-A” and “-B” have been proposed for thc two 10,000-dalton components ( Goodenough, 1974). These peptides may conncct adjacent cells at gap junctions (Goodenough, 1974) although probably not by covalent bonds, since the two junctional membranes can be separated by immersion of tissue in a hypertonic disaccharide solution that “unzips” the junction in the central plane (Goodenough and Gilula, 1974). Dunia et d.( 1974) succecded in isolating lens-gap junctions and also found a 34,000-dalton major protein. It is noteworthy that the existence of the 18,000- and 10,000-dalton proteins in gap junctions in the native state is open to question because of the possibility that these low-molecular-weight polypeptides may represent proteolytic breakdown products of a large-molecular-weight subunit. Gilula ( 1974) examined additional properties of isolated gap junctions and found an absence of both negatively charged surface groups and concanavalin A binding sites. No endogenous enzyme activity has been consistently associated with gap junctions.

3. Functions Gap junctions mediate ionic metabolic coupling between cells by providing conduits across cell membranes. By definition, cell pairs are “ionically” ( electrotonically ) coupled when they freely exchange ions along concentration gradients without leakage into the extracellular compartment. They are “metabolically coupled” when nutrients or intermediate metabolites of low molecular weight are exchanged by diffusion. Gap junctions also contribute to strong intercellular adhesion ( Muir, 1967; Goodenough and Stoeckenius, 1972).

C. ADHERENTESJUNCTIONS In this section, we shall focus our attention on the macula adhercns, since this is the most frequently encountered type of adherens junction and it has received particular attention in cancer research. The ultrastructure of fasciae and zonulae adherentes is summarized elsewhere (McNutt and Weinstein, 1973). Maculae adherentes occur in nearly all epithelia. As in the case of gap junctions, maculae adherentes may join either homotypic or heterotypic cells (Hay, 1961; Overton, 1974, 1975). 1. Ultrastructure of the Macula Adherens ( D e m o s o m e ) In transverse sections, the membrancs of thc fully developed macula adhcrens junction are parallcl to one another and are separated by a

40

RONALD S. WEINSTEIN, FREDERICK B. M E R K , AND JOSEPH A L R O Y

2535-nm interspace (Fig. 5 ) . Within the extracellular zone of the junction there is a condensation of proteinaceous material that sometimes forms a central dense stratum in a plane equidistant from, but parallel to, the junctional membranes. A dense fibrillar plaque is located within the cytoplasm of each cell subjacent to the junctional membrane (Fawcett, 1961; Farquhar and Palade, 1963). These plaques are attachment sites for bundles of cytoplasmic tonofilaments. The tonofilaments do not terminate in the dense plaques, but instead loop through them and then back into the cytoplasmic matrix (Kelly, 1966). Freeze-fracturing reveals a nonpolygonal arrangement of closely packed granules and short

FIG.5. hpdculae adherentes junctions ( desniosomes ) in Fischer rat urinary bladder.

This micrograph shows an area of squamous cell differentiation in a transitional cell carcinoma induced with the carcinogen N-[4- ( S-nitro-2-furyl)-2-thiazolyl]formamide. At each inacula adherens, cytoplasmic tonofilaments converge on the dense plaques ( D P ) of the junction. A gap junction ( C J ) is illustrated. B. U. Pauli, R. S. Weinstein, and S. M. Cohen, unpublished micrograph. ~ 7 0 , 0 0 0 .

INTERCELLULAR JUNCI’IONS I N CANCER

41

filaments approximately 8-10 nm in rcplica diameter within the interior of the adherens junction membranes ( McNutt et al., 1971; Breathnach et al., 1972). The granules and filaments may represent segments of filaments arising out of the dense fibrillar plaques ( McNutt and Weinstein, 1973). The filaments may intercalate into the hydrophobic interior of the membranes. Junctions resembling one half of a macula adherens lie along the basal surface of many epithelia where they attach epithelial cells to the connective tissue substratum. These junctions are called “hemidesmosomes” ( Kelly, 1966) although there are structural differences between these junctions and intraepithelial desmosomes (Kelly, 1966; Hay and Revel, 1!369). Another type of adherens junction resembles the hemidesmosome in that it is an asymmetric adherens junction. Such junctions are intraepithelial, unlike true hemidesmosomes, and may represent imperfectly formed maculae adherentes junctions (Weinstein et d.,1974). We propose the name “ d a adherens imperfecta” (or “imperfect desmosome”) for this junction. The macula adherens imperfecta (see Fig. 7 in Weinstein et al., 1974) has a fully developed dense plaque subjacent to the junctional membrane of one cell but none in its neighbor, unlike asymmetric adherentes junctiofis in embryonic tissues, which have a well developed dense plaque in one cell and a poorly developed plaque in the neighboring cell (Hay, 1961) . Tonafilaments, when present, loop through the single dense plaque of a macula adherens imperfecta, and the extracellular condensation is moderately well developed. Macula adherens imperfecta could arise through one of several mechanisms: (1) in uiuo shearing of defective symmetrical adherentcs junctions; ( 2 ) incomplete formation of adhcrentes junctions due to the failure of one cell to assemble its contribution to the junction; or ( 3 ) as part of a cell-cell dissociationreassociation cycle (Wcinstein et al., 1974). Two other forms of the adhcrcns junction, namely zonula and fascia adherens, are usually included for convenience in the general adhercns category although their ultrastructure is significantly different from that of the macula adherens ( McNutt and Weinstein, 1973). These junctions have not been investigated with respect to cancer. Histochemical and cytochemical studies have dcmonstrated that the cytoplasmic dense plaques and the intercellular condensations of desmosomes are susceptible to proteolysis (Overton, 1968; Douglas et al., 1970; Borysenko and Revel, 1973). The junctional interspace is rich in carbohydrate residues (Kelly, 1966; Luft, 1971; Rambourg, 1969), and these may be components of membrane glycoprotein ( Staehelin, 1974). Selec-

42

RONALD

s. WEINSTEIN,

FREDERICK B. MERK,AND JOSEPH ALROY

tive removal of divalent cations with the chelator EDTA diminishes intercellular zone material and results in a widening of the intercellular space (Muir, 1967; Overton, 1968). The EDTA experiment may indicate that some maculae adherentes are adhesive structures which operate by calcium bridging (Sedar and Forte, 1964). Borysenko and Revel ( 1973) recently demonstrated that maculae adherentes from different normal tissues have varying sensitivities to proteolysis, chelators, and detergents, suggesting that adherentes junctions can have different chemical compositions. This heterogeneity suggests that intercellular adhesion at adherentes junctions in different tissues may be accomplished by different mechanisms. 2. Isolation and Chracterization Skerrow and Matoltsy ( 1974a,b) have isolated maculae adherentes junctions from cow snout epidermis at a level of purity suitable for biochemical analysis. Their purification method utilizes selective solubilization by citric acid-sodium-citrate buffer, pH 2.6, and discontinuous sucrose density gradient centrifugation. The isolated maculae adherentes contain protein, carbohydrate, and lipid at a ratio of 76:17:10. The protein is rich in nonpolar amino acid residues. Gel electrophoresis of junctional proteins demonstrates 24 bands with mobilities corresponding to a molecular weight range of 15,OOo-230,000. Two proteins, with molecular weights of 210,000 and 230,000, comprise 28% of the desmosome weight. Skerrow and Matoltsy suggest that the high-molecular-weight proteins may derive from the dense fibrillar plaque. Two other bands are periodic acid-Schiff positive and constitute 23%by weight of the macula adherens protein. These bands may derive from the carbohydrate-rich material at the junction interspace ( Skerrow and Matoltsy, 1974b). 3. Function

Adherentes junctions play an important role in cell-to-cell adhesion. They and their associated tonofilaments are elements in a cytoskeletal system that may mediate a complex form of cooperation between individual cells for the regulation of mechanical properties of the tissue as a whole.

D.

JUNCTIONAL

COMPLEXES

At regions of cell-to-cell apposition, several types of cell junctions can occur in groups called junctional complexes. A classic example is the “terminal bar” region of intestine columnar epithelium ( Farquhar and Palade, 1!363), in which a series of junctions whose order is invari-

LVTERCELLULAR JUNCTIONS IK CANCER

43

able appears at the apical margin of the lateral surfaces of columnar epithelial cells (Fig. 3 ) . A zonula occludens is the apical junction of this complex, and this is closely followed by a zonula adherens and then by a macula adherens. In other tissues, several types of junctions may occur in close juxtaposition but not arranged into an identifiable pattern ( McNutt and Weinstein, 1973). Unusual combinations of junctions are sometimes observed in tumors. For example, Ahoy and Weinstein (1976) studied the ultrastructure of an adenoacanthoma arising spontaneously in canine mammary gland. An adenoacanthoma is characterized by bidirectional differentiation; i.e., cells within the tumor have structural characteristics of both columnar and squamous epithelial cells. Carcinoma cell membranes in adenoacanthomas have prominent occludentes junctions that resemble those of nonneoplastic columnar cells. Interspersed among the intramembrane fibrils and grooves of the occludentes junctions are fully developed maculae adherentes (Fig. 7B), which resemble the adherentes junctions of normal squamous cells. Thus, the columnar and squamous cell phenotypes are both represented within a single region of the membrane.

E. MISCELLANEOUS JUNCTIONS Several additional types of Cell junctions have been characterized (Lasansky, 1969) including septate desmosomcs (Wood, 1959; Locke, 1965; Gilula et al., 1970), septatelike contacts (Friend and Gilula, 1972a), etc. These and other unusual junctions (Breton-Gorius et d., 1975) are either restricted in their occurrence to one or two organs or are rarely encountered in mammals. Since these junctions have not been related to the cancer problem, we will not consider them further. IV. Occurrence of Cell Junctions in Tumors

A. SOLIDTUMORS(AND NONMALIGSANT GROWTHDISORDERS) The descriptive literature regarding tumor ultrastructure is voluminous. Surgical pathologists recognize that cell junctions can be useful

for classifying certain tumors which present difficult diagnostic problems, and therefore cell junctions are mentioned in many pathology papers. Unfortunately, only a few of the many papers on tumor ultrastructure contain quantitative information on specific types of cell junctions. 1. Descriptions of Cell ]unctions in Tumors We have surveyed the ultrastructure literature and in Table I have listed representative reports that describe cell junctions in many types

I I)ESCHIPTIONS OF CF:LL JUNCTIONS I N TUMOHW-~ TABLE Cell junction

Organ Adrenal

Bone

Breast

Tumor (species)

Terminal barc

Occludens

“Gap”

Adherens

+

+

Kovaca el at. (1974) Sharmaand Hashimoto (1972) Brown el al. (1972) Steiner el al. (1973) Hou-Jensen el al. (1972) Steiner et al. (1972) Gonealez-Licea el al. (1907) Martinez-Palomo (1970b) Tobon and Price (1972)

+, HI)

Tobon and Price (1972)

Adenoma (human) Carcinoma (rat)

+

Pheochromocytoma (human) Chondrosarcoma (human) Ewing’s sarcoma (human) Giant cell tumor (human) Osteogenic sarcoma (human)

+ + + +

Adenocarcinoma (mice) Carcinoma, lobular, in s i t u , epithelium (human) Carcinoma, lobutar, in situ, myoepithelium (human) Carcinoma, mucinous (human) Carcinoma, scirrhous (human) Carcinoma, tubular (human) Cystosarcoma phylloides (human) Fibroadenosis (human) Mixed tumor myoepithelium (canine)

1

0

+

1

1

1

1

f

Reference

+ +

+ 1

Ahmed (1974) Ahmed (1974) Erlandson and Carstens (1 972) Toker (1968) Barton (1964) von Bombard and von Sandersleben (1973)

m

+, HI) +

Mixed tumor myoepithelium (canine) Sclerosing adenosis (human) Carotid body Craniopharyngioma Ear Epididymis Eye

Adenoma (human) Human Ceruminous gland adenocarcinoma (human) Adenomatoid tumor (human) Corneal “preinvasive” cancer (human) Retinoblastoma (human) Retinoblastoma-like tumor (rat)

Fallopian tube Hypopharynx Intestine

Kidney

Larynx Liver

Adenofibroma (human) Squamous cell carcinoma (human) Adcnocarcinoma (human) Papilloma (human) Transitional cloacogenic carcinoma (human) Villous adenoma (human) Congenital mesoblastic nephroma (human) Mesenchymal renal tumor (human) Renal cell carcinoma (clear cell) (human) Renal cell carcinoma (granular cell) (human) Wilms’s tumor (human) Papillomatosie (human) Adenoma, liver cell (human) Hepatoma (human)

+, I n )

+

+

+

+

+

1, HD

+

+

1

+ +

+ 1

+

+

+ + + + + + + 1

+ or 1 +, HD + +

Pulley (1973) Wellings and Roberts (1963)

Welsh el al. (1972) Ghatak et ol. (1971) Welti et al. (1972) Mackay et al. (1971) Tripathi and Garner ( 1972)

Ts’o el al. (1969) Kobayashi and Mukai (1974)

Kanbour el al. (1973) Rangan (1972) Imai and Stein (1963) Fisher and Sharkey (1962)

Fisher (1969) Ioachim el ol. (1974) Fu and K a y (1973) Favara et al. (1968) Tannenbaum (1971) Tannenbaum (1971) Tannenbaum (1971) Svoboda el al. (1963) K a y and Schatzki (1971)

Wills (1968) (Continued)

$

TABLE I (Continued) Cell junction

Organ

Tumor (species)

Terminal barc

m 0

z

Occludens

‘‘Gap” Adherens

Reference

E

r 8

Liver (Cont’d)

Hepatoma BH3 (mouse) Hepatoma BC3 (mouse) Hepatoma BRL (mouse) Hepatoma BNL (mouse) Hepatoma H-31780 (rat) Hepatoma, Morris 9121 (rat) Hepatoma, Novikoff (rat)

Lung

1

Hepatoma (human)

Hepatoma, Yoshida ascit,es, AH602 (rat,) Hepatoma, Yoshida ascites, AH7974 (rat) Hepatoma, Yoshida ascites, AH130 (rat,) Hepatoma, Yoshida ascites, AH13 (rat) Bile duct adenocarcinoma (rat) Acinic cell tumor (human) Adenomatosis (sheep) Carcinoid, bronchial (human) Carcinoma, anaplastic (hamster) Carcinoma, bronchoalveolar (human) Carcinoma, oat cell (human) Carcinoma, squamous cell (human)

+

+ 1 + 0

+

+ 1

1 1

+ 1 1

+ +

+ + 1

1

0

+ +

+

+ 0 + 0

n

+ + +

Ma and Blackburn (1973) Malick (1972) Malick (1972) Malick (1972) Malick (1972) Martlnez-Palomo (1970b) Urban el al. (1972) Babai and Tremblay (1972) Hoshino (1963) Hoshino (1963) Hoshino (1963) Hoshino (1963) Ma and Webber (1966) Fechner et al. (1972) Perk et at. (1971) Hage (1973) Harris et al. (1973) Kuhn (1972) Bensch et al. (1968) Lupulescu and Boyd (1972)

M

1

9 q

8

1 m

z

“E

1 8 !

w

4

E

2

Mouth

Nasopharynx Nervous System

Ovary

+

Adenomatoid tumor (human) Ameloblastoma (human) Calcifying epithelial odontogcnic tumor of Pindborg (human) Carcinoma, spindle-cell (human) Granular cell tumor (human) Carcinoma (human) Astrocytoma (human) Astrocytoma, glioblastoma multiformis (human) Choroid plexus papilloma (human) Ependymoma (human) Germinoma, in tracranial (seminoma) (human) Medulloblastoma (human) Medulloepithelioma (rat, mouse, hamster) Meningioma (human) Adenomatoid tumor (human) Brenner tumor (human) Cystadenocarcinoma (human) Cystadenofibroma (monkey) Cystadenoma (human) Clear cell tumor (human) Dysgerminoma (human) Gonadoblastoma (human) Granulosa cell tumor (human) Granulosa theca cell tumor (human) Sertoli-Leydig cell tumor (human) Surface papilloma (monkey)

+, HI)

+

+, HD 0

+ + or 1 + 1

+ +

+

1

+ + + + + +

+

+ + + +

+ +

+ + 1 + +

+

+ + 1 +

1

0

Taxy et al. (1974) -Minter and McGinnis (1972) Anderson el al. (1969) Leifer el al. (1974a) K a y el al. (1971) Lin et al. (1969) Rubinstein el al. (1974) Tani et al. (1973) Carter et al. (1972) WolR et al. (1972) Tani el al. (1974) Rubinstein el al. (1974) Mukai et al. (1974) Popoff el al. (1974) Ferencsy et al. (1972) Bransilver el al. (1974) Gondos (1971) Amin et al. (1974) Gondos (1971) Salazar et al. (197.4) Lynn el al. (1967) AMackayel al. (1974) Hamlett el al. (1971) Bransilver el al. (1974) Murad et a / . (1973) Amin el al. (1974)

13F:

5 z

2

$

2

v)

2

2

3z

.h

(Continued)

=j

b b 06

TABLE I (Continued) Cell junction

Y

Organ Pancreas

Parathyroid Pineal Prostate Salivary glands

Skin

Tumor (species) Acinar cell carcinoma (human) Alpha cell tumor (human) Beta cell tumor (human) Carcinoma, infantile type (human) Giant cell tumor (human) Adenoma (human) Carcinoma, acinar (human) Large cell tumor (human) Carcinoma (human) Acinar cell tumor (human) Mixed tumor, chondroid pattern (human) Mixed tumor, (epithelial, myoepithelial, and myxoid patterns) (human) Myoepithelioma (human) Sebaceous carcinoma (human) Basal cell carcinoma (human) Bowen disease (human) Keratoacanthoma (human) Melanoma (human) Pilomatrixoma (human) Squamous cell carcinoma (rat) Squamous cell carcinoma (human)

Terminal barc

Occludens

“Gap”

Adherens

Reference

+ + + +

Burns et al. (1974) Goldenbcrg et al. (1969) Bencosme et al. (1963) Frable el al. (1971) Rosai (1968) Marshall et al. (1967) Echevarria (1967) Ramscy (1965) Mao et al. (1966) Erlandson and Tandler (1972) Welsh and Mcyer (1968) Welsh and Meyer (1968)

I! YJ 5:

+

+

+

0

+ or 1 + 0

+

+ + 1

1, HL)

0

1

+, HD + + 1 1

Leifer et al. (1974b) Akhtar el al. (1973) Flaxman (197’2) Yeh (1973) Fisher et al. (1972) Mishima (1967) McGavron (1965) Martinez-Palomo (1970b) Fisher el al. (1972)

3

“g

i F! w

li

”

1

LI

8

u

3:

b

Soft tissues

+

Squamous cell carcinoma (human) Alveolar soft part sarcoma (human) Fibromyxosarcoma (human) Fibroxanthoma, malignant (human) Hemangioendothelial sarcoma (human) Hemangiopericytoma (human) Histiocytosis (human) Lciomyoblastoma (human) Leiomyosarcoma (human) Lymphoangiosarcoma (human) Mesothelioma (human) Mesothelioma (rat)

+ 1

+ 1 0

Ithabdomyoma (human) Sarcoma, clear cell (human) Sarcoma, epithelioid (human)

Stomach Testes

1

+ +

Adenocarcinoma (human) Embryonal carcinoma (human) Leydig cell tumor (rat) Leydig cell tumor (rat) Seminoma (human)

+

Seminoma, spermatocytic (human)

0

0 0

Sarcoma, Moloney virus (rat) Sarcoma, synovial (human) Sarcoma 180 (Crocker’s tumor) (mouse) Sclerosing hemangioma (human)

Scminoma (human)

+ + +

+

+

+,HD

+ + + +

+ or 1 +

T

+ +

+ + +

Hashirrioto el al. (1973) Welsh et a/. (1972) Leak et al. (1967) Merkow el al. (1971) Steiner and Dorfrnan (1972) Battifora (1973) Hou-Jensen el al. (1973) Cornog (1969) Wang et al. (1974) Merrick et al. (1971) Ferenczy et ul. (1972) Shin and Firminger (1973) Battifora et al. (1969) Kubo (1969) Fisher and Horvat (1972) Stanton el ul. (1970) Kubo (1974) Zuckerberg (1973) Hill and Eggleston (1972) Ming et ul. (1967) Pierce (1966) Roth el al. (1970) lteddy et al. (1973) Holstein and Korner (1974) J. Ahoy et al. (unpublished) Rosai el al. (1969)

(Continued)

TABLE I (Continued) Cell junction

Organ Testes (Cont’d) Thymus Thyroid

Ureter Urinary bladder

Uterus

Tumor (species) Sertoli cell adenoma (human) Teratocarcinoma (mouse) Yeminoma, thymic (human) Thymoma (human) Adenoma, follicular (human) Carcinoma, follicular (human) Carcinoma, papillary (human) Hurthle cell tumor (human) -Medullary carcinoma (human) AMedullarycarcinoma (human) Ultimobranchial adenoma (bovine) Transitional cell carcinoma (human) Transitional cell carcinoma, Grade I (human) Transitional cell carcinoma, Grade 11-111 (human) Squamous cell carcinoma (human) Carcinoma, adenosquamous (human) Clear cell adenocarcinoma (human) Papillary adenocarcinoma (human) Squamous cell carcinoma, cervix, in situ (human)

Terminal barc

Oceludens

‘‘Gap’’ Adherens

f

+

+ 1 +

+ + + +

+ + + 1 + +

1 +or1

1 or 0

+ +

+ t

1 T

+ +

f, H D

+

1 1

11

1

lleferen ce Able and Lee (1969) Pierce el al. (1967) Levine (1973) Leviqe (1973) Lupulescu and Boyd (1972) Tonietti el al. (1967) Tonietti el al. (1967) Feldman et al. (1972) Horvath el al. (1972) Bordi et al. (1972) Black el al. (1973) Flaks et al. (1970) R. S. Weinstein et al. (unpublished) It. S. Weinstein el al. (unpublished) R. S. Weinstein et al. (unpublished) Aikawa and Ng (1973) Rorat et al. (1974) Hameed and Morgan (1972) McNutt et al. (1971)

m

Squamous cell carcinoma, cervix, (human) Sauamous cell carcinoma, corpus (rat) Vagina

Clear cell carcinoma (human)

11 or 0

1

McNutt et al. (1971)

+, HI) +

Baba and von Haam (1971)

Silverberg and DeGiorgi (1972)

+

= junctions are present at a numerical frequency similar to t h a t in the tissue of Semiquantitation of intercellular junctions; tumor origin; 0 = junctions are absent; f = junctions are increased in number; 1 = junctions are moderately reduced in number; 11 = junctions are markedly reduced in number; HD = hemideamosomes are present. * Blank spaces do not indicate t h a t specific types of junctions are necessarily absent; rather they indicate t h a t the junction has not been described in relevant papers. c Terminal bar refers to the junctional complex at the luminal border in many cuboidal and columnar epithelia. Usually, the terminal bar region contains three types of junctions: zonuls occludens, zonula adherens, and macula adherens. In many papers, the terminal bar is mentioned, b u t individual component junctions are not identified.

1 C

$

0

Z

v)

2!

9

E

52

HONALD

s. WEINSTEIN,

FREDERICK B. MEXK, AND JOSEPH ALROY

of solid tumors. It is noteworthy that many of the cited reports omit from their descriptions some types of junctions. There are a number of reasons for this. Certain types of junctions, particularly gap junctions and tight junctions, are difficult to identify in surgical or autopsy material, even under the best preparatory conditions. Positive identification of gap junctions requires the use of special specimen preparation techniques that are generally not used in surgical pathology laboratories. Also, until recently, the classification of junctions was a matter of controversy. The earlier tumor ultrastructure literature contains many inaccurate descriptions of cell junctions. We can draw several conclusions about the occurrence of cell junctions in tumors from the data in Table I. These data show that the various types of cell junctions are present in many types of epithelial and mesenchymal tumors although the occurrence of these junctions often is less frequent than in normal tissue. The type of cell junction most frequently reported in solid tumors is the macula adherens (desmosome). This may be related more to the relative ease with which desmosomes are visualized in the electron microscope than to their relative abundance or importance. 2. Qtiuntitative Studies of Junctions in Solid Tumors

Systematic quantitative studies of cell junctions have been performed for two solid tumor systems, squamous cell carcinomas of human cervical epithelium ( McNutt and Weinstein, 1969; McNutt et al., 1971; Wiernik et al., 1973) and uroepithelial tumors (i.e., transitional cell carcinomas) arising in human urinary bladder ( Weinstein et al., 1974). a. Gay ]unctions. blcNutt et al. (1971) compared the population frequency of gap junctions in normal human cervical squamous epithelium, squamous metaplasia and dysplasia (two reversible nonmalignant states), carcinoma in situ, and invasive squamous cell carcinoma. Human uterine cervical epithelium is particularly well suited for quantitative studies of gap junction frequencies, since gap junctions are remarkably abundant in normal cervix ( McNutt and Weinstein, 1969). In addition, the neoplastic state carcinoma in situ, which arises spontaneously in cervical cpithclium, can serve as a positive control in studies correlating ultrastructure with biological behavior, since carcinoma in situ is a well recognized preinvasive malignant state which can persist in the cervix for years before the appearance of stromal invasion (Richart and Barron, 1969). Structural changes arising during the protracted in situ stage may be either early events in a series that leads to invasiveness or are unrelated to invasiveness. In either case, the in situ carcinoma is a control that is clearly noninvasive.

53

INTEHCELLULAR JUNCTIONS IN CANCER

The intermediate cells of normal cervical epithelium have up to 225

gnp junctions per cell (Table 11). In contrast, invasivc squamous cell carcinomas in uterinc cervix (Fig. 6 ) arc markedly deficient in gap junctions ( McNutt and Weinstein, 1969). In relatively well differentiated areas of human tumors, up to four gap junctions pcr cell arc present, but in poorly differentiated areas nonc are found. This variance in the distribution between well differcntiatcd and poorly differentiated areas may help to explain thc variations in electrotonic coupling that are found in different regions within solid tumors (Sheridan, 1970). Very fcw gap junctions arc prescnt in carcinoma in situ of the cervix (Table 11). This allows us to infer that there is a poor temporal correlation betwecn the development of severe gap-junction deficiencies and tumor invasion, although the possibility remains that the loss of gap junctions is one of several prerequisites required for stromal invasion. In reversible nonmalignant states, including squamous metaplasia and moderate dysplasia, there are statistically significant decreases in gapjunction frequency although junctions remain plentiful ( McNutt et a]., 1971). 11. Maculae Adherentes ( Desmosomes) . Numbers of maculae adherentes (desmosomes) have been estirnatcd for many types of tumors (Table I ) and have been precisely quantitatcd for two solid tumors, human cervical and urinary bladder carcinomas. McNutt and Weinstein

Basal zone

State Normal squsmous cpithelium Sorinal squamous cpithcliuni Squnmous mctaplasia l)ysplssia, mild to moderate Ilysplasia, severe carcinoma in situ Invssivc carcinoma

I

53 f 26.‘ 11 f 4 d 10 f 1‘ 11 5‘

*

Intermediate zone 187 f 40 -

+

60 8 72 f 31

2 f 1“ 0 . :P

6

a Adapted from h l c S u t t el nl. (1971). The ENF was estimated hy dividing the n u m h r of nexuses per section (n) by thc product of thc number of rell profiles pcr section (c) niultiplicd by the fraction of cell surface included per section (f),where:

j = section thickncss;/average cell diameter

* Value

and ESF = n / ( c f )

? one standard deviation. Basal zone, consisting of npproximalely thrce cell layem. d Bmal laycr, consisting of the single layer of cells adjacent to tho basement membfanc.

54

RONALD S. WEINSTEIN, FREDERICK B. MERK, AND JOSEPH ALROY

INTERCELLULAR JUNCXIONS IN CANCER

55

(19sS) were the first to observe that maculae adherentes are less frequent in invasive cervical squamous cell carcinomas than in normal cervix. In a more detailed study, Wiernik and his associates found a small yet statistically significant decrease in the total number of maculae adherentes per cell with malignant transformation. Their data also show that maculae adherentes in cervical invasive carcinomas are smaller than those in normal cervix. They found no significant difference in the number of maculae adherentes per unit length of membrane (Wiernik et al., 1973). Comparable studies of maculae adherentes have not yet been reported for preinvasive cervical carcinoma, although it is known that their numbers are reduced in Bowen’s disease, a form of preinvasive intraepidermal squamous cell carcinoma of the skin (Yeh, 1973; Yeh et al., 1974). Recently, studies of the occurrence of maculae adherentes in tumors have been extended to include transitional cell carcinomas arising in human bladder uroepithelium (Fulker et al., 1971; Weinstein et al., 1974; Ahoy et al., 1976). These studies provide some of the strongest circumstantial evidence for an association between abnormalities in maculae adherentes and malignant behavior. Relatively few maculae adherentes are present in normal human bladder uroepithelium (Richter and Moize, 1963; Battifora et aZ., 1965). In 1971, Fulker and FIG 6. Electron micrographs of cell junctions in normal and malignant epithelium of the human uterine cervix. ( A ) Normal cervical epithelium. Intermediate layer cells are connected to each other by gap junctions (nexuses ) and desmosomes. At gap junctions, the cell membranes are very closely apposed. At desmosomes, the neighboring cell membranes are attached by filamentous extracellular material, often showing a distinct central stratum (S).~37,420. ( B ) Normal cervical epithelium. Cap junction ultrastructure as demonstrated by lanthanum tracer impregnation. Four cell processes (PI-,) are attached at gap junctions. Electron-opaque lanthanum hydroxide fills thin channels of extracellular space around subunits (foot processes) which join the junctional membranes. Gap junctions, viewed en fuce, appear as a closely packed array of subunits outlined with lanthanum. They have a 9-10-nm center-to-center spacing. In cross section, this subunit pattern is obbcured and the so-called “gap” appears as a dark central 7-nm line ( S ). XSil20. Inset is a detailed en face view of a subunit array. Individual subunits (arrow) are 7 nm in diameter. x 120,960. (C)Squamous carcinoma of the cervix. Processes of adjacent cells are attached by desmosomes, but gap junctions are infrequent. Some tumor desmosomes lack a centml dense stratum. ~38,800. ( D ) Squamous carcinoma of the cervix. A rare gap junction connects two tumor cells. This junction has been infiltrated with lanthanum, revealing in cross section the characteristic five-layered appearance of a lanthanum-stained gap junction ( X ) and in en face sections a closely packed subunit array. ~62,640. A-D: From McNutt and Weinstein (1989);copyright 1969, American Association for the Advancement of Science.

56

RONALD

s. WEINSTEIN,

FREDERICK B. MERK,AND JOSEPH ALROY

his associates reported that the numbers of maculae adherentes often increase above normal in low-grade papillary transitional cell carcinomas and decrease below normal in moderate and high-grade tumors (Fulker et al., 1971; Cooper, 1972). These observations have since been confirmed ( Weinstein et al., 1974). Since low-grade papillary transitional cell carcinomas are infrequently invasive, unlike higher grade bladder tumors, an obvious question is whether loss of junctions facilitates invasion in higher-grade tumors. This question can be tested since low-grade bladder tumors occasionally invade stroma, and since a considerable percentage of higher-grade tumors are either noninvasive or pass through a preinvasive stage in the course of development. These variants have been used in a preliminary study to determine whether the number of maculae adherentes in bladder carcinomas show a stronger correlation with tumor grade per se or with invasive behavior. We recently quantitated maculae adherentes in low-grade human papillary transitional cell carcinomas which had invaded stroma and muscle, and in several higher grade tumors which were noninvasive at the time of surgical removal. Junctions were decreased in frequency in the low-grade invasive tumors and increased in frequency in the noninvasive higher-grade tumors. It would therefore seem that maculae adherentes frequency in bladder tumors is indeed more strongly correlated with invasive behavior than with tumor grade (Alroy et aE., 1976). If these observations are confirmed for a larger series of biopsies, they could have a considerable impact on diagnostic pathology since numbers of demosomes in transitional cell carcinomas may turn out to be a useful prognosticator of tumor behavior. Hruban et aZ. (1972) quantitated maculae adherentes in 35 different transpIantabIe Moms hepatomas which exhibited a broad spectrum of growth rates. Their data show no correlation between maculae adherentes frequency and growth rate. c. Occludentes Junctions. Quantitation of zonulae occludentes junctions requires more than simple counting of junctional sites. Since many normal epithelial cells typically have a solitary zonula occludens junction, quantitation is generally accomplished by counting the number of parallel intramembrane fibrils within the junctional zone (Claude and Goodenough, 1973). Although this measurement correlates well with junctional sealing capacity, simple counting of occludentes intramembrane fibrils in tumors can be meaningless since the numbers of fibrils are frequently extremely variable along the surfaces of individual tumors cells. Attenuation or loss of zonulae occludentes is a common occurrence in anaplastic carcinomas (see Table 11; also Martinez-Palomo, 1970b;

INTERCXLLULAR JUNCTIONS IN CANCER

57

Weinstein et al., 1974). These changes were systematically examined in spontaneous transitional-cell carcinomas arising in human urinary bladder. As is found in normal bladder uroepitheliums, in low-grade papillary tumors, zonulae occludentes are confined to the apical regions of the lateral surfaces of superficial tumor cells (Fig. 7A). In nonneoplastic states, the junctions contain four or five parallel intramembrane fibrils, but in low-grade tumors they are focally reduced to the width of a single intramembrane fibril. In invasive tumors, occludentes junctions are discontinuous (i.e., they become maculae occludentes ) and are markedly attenuated. They are found at all surfaces of individual superficial tumor cells as well as at the surfaces of cells deep within the tumors (R. S. Weinstein, F. B. Merk, and J. Alroy, unpublished observations). The presence of occludentes junctions on cells deep within tumors may be a manifestation of loss of cell polarity which is commonly observed in anaplastic tumors. The finding suggests that invasive tumor cells may be capable of retrograde migration. Loss of occludentes junctions also results in a decrease in intercellular adhesion, an increase in leakiness across the epithelia, and possibly enhances the loss of cellular polarity. The first two functional changes are discussed elsewhere (Sections VIII, A and IX). Differentiated epithelial cells are highly polarized with respect to their biochemical surface topography. For example, in renal proximal tubules, leucine aminopeptidase activity is located exclusively at the luminal border and Na-K-MgATPase is located exclusively at the basal border of the epithelial cells. In normal tissues and low-grade tumors, the intramembrane fibrils of zonula occludens may serve as a physical barrier to the lateral diffusion of individual membrane components in the plane of the membrane and, by this mechanism, keep membrane components segregated along specific membrane surfaces. Disruption zonulae occludentes may permit membrane components to migrate over the entire surface of the cell. This would result in a haphazard arrangement of mobiIe membrane components at the cell surface (Porter et al., 1974) and thus deprive the cell of its capacity to perform specialized functions. Other mechanisms are also contributors to the control of cell surface topography ( NicoIson, 1973; Elgsaeter and Branton, 1974). 3. Cell ]unctions between Blood Cells Normal erythrocytes, granulocytes, and small lymphocytes do not form morphologically detectable cell junctions. Small lymphocytes can be induced by immunologically specific as well as nonspecific stimulants to shift from a resting state to an activated state with high metabolic activity. In the activated state, lymphocytes apparently are able to form

58

RONALD S. WEINSTEIN, FREDERICK B. MERK, AND JOSEPH ALROY

INTERCELLULAR JUNCXIONS IN CANCER

59

low-resistance junctions. This was shown by Hulser and Peters (1972), who incubated bovine lymphocytes with the nonspecific stimulant phytohemagglutinin ( PHA ) and successfully demonstrated electrotonic coupling within 1 minute of initiation of stimulation. Also, Sellin et al. ( 1971, 1974) demonstrated fluorescein dye transfer from anti-immunoglobulin-stimulated lymphocytes to unstimulated lymphocytes and macrophages. R. S. Weinstein and M. Stadecker (unpublished observations) were unable to demonstrate any type of intercellular junctions between PHA-stimulated lymphocytes by either freeze-fracture or thinsection electron microscopy. Their finding is curious since electrotonic coupling and gap junctions coexist in all other cell systems that are capable of ionic communication, with the obvious exception of cell pairs joined by intercellular bridges (Woodruff and Telfer, 1973). Possible explanations for this discrepancy are that: the junctions in stimulated lymphocytes form transiently and may not be preserved by electron microscopy methods; or that gap junction subunits are widely dispersed over the lymphocyte cell surface and therefore are not identifiable as components of gap junctions since they fail to organize into the characteristic geometric array of the gap junction. It is also noteworthy that Cox et a2. (1974) were unable to demonstrate metabolic coupling between lymphocytes in continuous culture. There is evidence that unusual cell junctions may be present in some lymphomas (Behrens et al., 1974) and leukemias (Sane1 and Serpick, 1970) and in bone marrow of some patients with ineffective erythropoiesis ( Breton-Gorius et al., 1975).

4. Cell Junctions in Tumor Metastases The relationship between the occurrence of cell junctions and the ability of tumors to metastasize has not yet been systematically explored. FIG.7. Freeze-fractured occludentes junctions in tumors. ( A ) A zonula occludens (A face) is present at the boundary between the lateral and luminal surfaces of a cell in a Grade 1 noninvasive papillary transitional cell carcinoma in human urinary bladder. Microvilli (MV) are prominent at the luminal surface. x57,000. From Weinstein et al. (1974). ( B ) Unusual junctional complex, found in an adenoacanthoma which arose spontaneously in canine mammary gland. Intramembrane fibrils and grooves which are characteristics of occludentes junctions, partially encircle maculae adherentes ( MA). ~71,000.From Alroy and Weinstein ( 1976). ( C ) An occludens junction at fracture face B of an H-35hepatoma cell in culture. At this fracture face, the junction appears as an extensive network of grooves. Frequently, membrane-associated particles which presumably have been fractured away from the complimentary A-face membrane are present in these furrows. ~70,000. By courtesy of Mr. M. Porvamik.

60

RONALD

s. WEINSTEIN, FREDERICK B.

MEFX, AND JOSEPH ALROY

It is likely that studies on this topic will be complicated because of the presence of phenotypically different cells within tumors ( Bennington, 1969; Harris et al., 1970; Pierce and Wallace, 1971). An obvious approach is to determine whether the tumor cells with a capacity to metastasize have junctions by examining tumor cells at metastatic tumor sites. Data on this topic are fragmentary and inconclusive, although cell junctions are observed in metastatic tumors (Gondos, 1969; Tarin, 1970; Letourneau et al., 1975).

5 . Cell Junctions in Nonmalignant Growth Disorders The occurrence of cell junctions in nonneoplastic disease states is a topic of considerable interest, since many of the changes in tumor junctions may represent nonspecific reactions. Cell junctions are described in a few papers on tissue ultrastructure in nonmalignant growth states (Pitelka et al., 1973; Vogel and Narasimnan, 1974; Tice et al., 1975). For example, several studies have examined junctions in epithelial metaplasia. Metaplasia is a nonneoplastic, reversible form of abnomial tissue regeneration in which one adult cell type is replaced by another adult cell type. Harris et al. (1973) studied maculae adherentes in chemically induded squamous metaplasia of the tracheobronchial epithelium. They found that the maculae adherentes in metaplastic tissue are more fully developed than in normal respiratory epithelium. This finding is expected since maculae adherentes are more fully developed in normal squamous epithelium than in pseudostratified respiratory epithelium. Recently, Prutkin ( 1975) found that topical applications of vitamin A-acid to tumors, e.g., keratoacanthomas, arising in rabbit ear skin epithelium and to normal skin epithelium produced mucous metaplasia. Gap junctions form early in the process of mucous metaplasia both in tumors and in normal epithelium. Junctions have been examined in other growth states, such as hyperplasia, hypertrophy, and atrophy (Pitelka et al., 1973). R. TISSUE CULTURE Occludentcs junctions ( Xlartinez-Palonio, 19703; Lavin and Koss, 1971), gap junctions (Pinto da Silva and Gilula, 1972), and adherentes junctions (McNutt et al,, 1973) are present in some culture systems. Gap junctions occur in many fibroblastoid cell lines (Pinto da Silva and Gilula, 1972; Revel et al., 1971; OLague and Dalen, 1974), but, according to Hiisler, they are less frequently encountered in epithelioid lines of cultured cells (Hiilser and Demsey, 1973; Hulser and Webb, 1973) . I n general, positive correlation cannot be demonstrated between

INTERCELLULAR JUNCTIONS IN CANCER

61

tumorigenicity of cultured cells and the presence or the absence of gap junctions ( Hiilser and Webb, 1973). Cultured cells often manifest different junctional properties from the corresponding cells in uiuo; quite frequently, junctions are more conspicuous in culture systems. For example, malignant hepatocytes isolated from Novikoff hepatoma readily form low-resistance junctions ( Borek et al., 1969) although most of the hepatocytes in the tumor of origin are electrically uncoupled ( Loewenstein and Kanno, 1967). Cells in ethylnitrosourea-induced malignant neurinomas are uncoupled, but cell lines derived from the tumors are coupled. Occludentes junctions may be relatively inconspicuous in solid tumors but flourish in comparable culture systems (Lavin and KOSS,1971) (Fig. 7C). There are several possible explanations for the differences between the occurrence of junctions in vivo and in uitro. First, while several cell phenotypes may coexist in solid tumors, there may be preferential growth of one cell phenotype (i.e., junction formers) in culture ( Azarnia and Loewenstein, 1971). Another possibility is that factors in the cellular microenvironment may inhibit cell junction formation in uiuo. This mechanism could account for the finding that morphologically normal liver cells growing in close proximity to cancer cells display a marked reduction in low-resistance electrotonic coupling ( Loewenstein and Kanno, 1967). Alternatively, factors in the culture media (e.g., proteolytic enzymes activated by plasminogen activator) may stimulate junction proliferation in tissue cultures.

V. Cell-to-Cell Communication and Growth Control

A. INTRODUCTION The growth of constituent tissues is precisely balanced and coordinated in normal animal organs. In most adult tissues, the rate of new cell production offsets the rate of cell loss enabling the tissue mass to remain essentially constant (Castor, 1968; Vasiliev et al., 1969). Cell growth and differentiation in complex tissues has been extensively studied. While a number of theories have been advanced to explain the modus operundi of homeostatic mechanisms for the control of growth (Bullough, 1965; Loewenstein, 1968a; Roth, 1973), the identity and nature of regulatory systems that influence growth in embryonic and adult tissue remains a matter for speculation. Nonetheless, our knowledge of the general features of control systems lends support to the popular notion that cell junctions play a significant role in the regulation of growth. All

62

RONALD S. WEINSTEIN, FREDERICK B. MERK, AND JOSEPH ALROY

control systems consist ofdhree cardinal elements. These are: (1) the control signals, such as chemical mediators; ( 2 ) a conduit for the conveyance of signals; and (3) a signal receptor (and effector). Theoretically, defects in any of these elements could account for the loss of growth control that may be characteristic of malignancies. It is reasonable to expect that different etiologies of cancer might disrupt any one of these three elements and still produce an identical perturbation of growth control ( Loewenstein, 1974). Changes in cell junctions associated with neoplastic transformation may constitute a breakdown in one element of the growth control system, the conduit for the conveyance of signals. In mammals there are three levels of cell-to-cell communication by means of transfer of substances. The first two are long-range communication, which involves the exchange of diffusible regulatory substances, such as blood-borne tropic hormones, over long distances, and shortrange communication, which involves the transfer of signal molecules, such as embryonic inducer substances, between cells that are close to each other but not intimately joined. Both of these require that the signal molecules move through the intercellular space. The third, intimate communication, requires actual contact between cells for the passage of information. This is accomplished either by allosteric mechanisms involving interaction of membrane components (Roth, 1973) or by the direct exchange of ions and molecules either through permeable intercellular junctions (Loewenstein, 1968a,b, 1973) or, far less commonly, through true cytoplasmic bridges (Fawcett et al., 1959; Zamboni and Gondos, 1968; Woodruff and Telfer, 1973). Whereas the general plasma membrane may participate in long-range and short-range communication, intimate communication by the direct transfer of substances is mediated by cell junctions or bridges.

H.

INTIMATE COMAIUNICATION AT

GAP JUNCTIONS

Pioneer studies demonstrating the existence of intimate communication via the selective transfer of substances (Furshpan and Potter, 1959; Loewenstein and Kanno, 1964; Loewenstein, 1966) were conducted by inserting micropipettes into neighboring cells, passing a current through them, and then measuring the resting potential between the micropipettes. These studies showed low-resistance electrotonic coupling between manv types of cells (Furshpan and Potter, 1959; Loewenstein and Kanno, 1967) indicating that contact areas of cell membranes do not impede ionic flow from cell to cell. This type of ion transfer is related to the presence of gap junctions (Barr et aZ., 1965; Dreifus et al., 1966; Gilula et al., 1972) .

INTERCELLULAR JUNCTIONS IN CANCER

63

Low-resistance junctions are found in both excitable tissues ( Furshpan and Potter, 1959; Woodbury and Crill, 1961; Potter et al., 1966) and nonexcitable tissues ( Loewenstein et al., 1965; Loewenstein, 1966). Their occurrence in tissues that do not propagate action potentials raised the possibility that these junctions serve functions in addition to ionic coupling of cells. In addition to inorganic ions, small organic ions and molecules pass freely from cell to cell at gap junctions. This was first shown by Loewenstein and Kanno (1964), who followed the transfer of the organic dye fluorescein (MW 342) between cells. They injected fluorescein into Drosophila salivary gland cells with micropipettes and observed that the probe molecules diffuse throughout the cytoplasm of the injected cell and then pass with apparent ease into neighboring cells. This observation raised the possibility that biologically active molecules such as nucleoside triphosphates ( MW 482523 ) or cyclic 3’,5’-adenosine monophosphate ( CAMP; MW 349) may pass from cell to cell at gap junctions (Sheridan, 1971; Merk et al., 1972). Since small organic molecules pass freely through gap junctions, it is important to establish the maximum sizes of molecules that are transferred via this route. Gap junction channels have been probed with tracer molecules of graded size (Kanno and Loewenstein, 1966; Reese et al., 1971). In 1966, Kanno and Loewenstein reported the largest molecule capable of passing by diffusion from cell to cell in salivary gland epithelia is bovine serum albumin (BSA), which has a molecular weight of 69,000 and an equivalent hydrodynamic radius of 3.6 nm. With a micropipette they injected radioactively labeled BSA into cells and then examined neighboring cells for the presence of radioactivity. Although the radioactive label was present in neighboring cells, no effort was made to recover and analyze it, leaving the possibility that the label was attached to split products rather than to intact BSA. Loewenstein (1966) also claimed that BSA, conjugated with the fluorescent label fluorescein-isothiocianate, can diffuse through cell junctions, but here again injected BSA was not recovered. Reese et al. (1971) used peroxidase to probe gap junction channels in the lateral giant axon fiber of the crayfish. Peroxidases can be visualized by histochemical methods at the ultrastructural level. In order to monitor the spread of tracer molecules in the light microscope, they added fluorescein to the peroxidase solution prior to injection into the segmental axons. In the presence of electrical coupling and fluorescein dye transfer, horseradish peroxidase ( MW 40,000) failed to transfer to neighboring axons. A low-molecularweight microperoxidase ( MW 1800), however, succeeded in passing from cell to cell. Reese and his associates interpreted the results as

64

RONALD S .

WEINSTEIN, FREDERICK B. MERK, A N D

JOSEPH ALROY

showing that the hydrophilic channels traversing gap junctions can accommodate molecules with molecular weights of up to at least 1800. A criticism of their experimental design has been raised by McNutt and Weinstein (1973), who point out that microperoxidase may not be immobilized during fixation with glutaraldehyde and so may pass through low-resistance passageways after they are altered by fixation for electron microscopy. Thus, the upper size limit for molecules that can pass through junctional channels remains an open question, but it is probably around 1000 daltons ( Staehelin, 1974). Gap junctions in many different tissues are indistinguishable from one another in the electron microscope although there is compelling evidence that they may have different sieving properties. This evidence comes from studies on embryonic tissues. Certain embryonic cells that are electrotonically coupled, presumably at gap junctions, do not transfer fluorescein in detectable quantities (Slack and Palmer, 1969; Sheridan, 1971; Bennett et al., 1972). Further, cells in Xenopus laevis embryos are electrotonically coupled, but fluorescein dye does not pass between them during early developmental stages. However, at later stages of development both electrotonic coupling and dye transfer are observed. These data suggest that the hydrodynamic pore diameter of gap junction channels may not be fixed and may be modulated by an as yet unidentified control mechanism. Alternatively, gap junctions may represent a family of structures that differ with respect to fixed pore size (Bennett et al., 1972).

C . CONTROL OF GAP JUNCTION PERMEABILITY Intracellular calcium ( Ca?+), adenosine triphosphate ( ATP ), cyclic adenosine monophosphate (CAMP), and electrochemical potentials across the plasma membrane each seem to play a significant part in the regulation of gap junction permeability in normal intact tissues.

1. Intracellular Calcium and Adenosine Triphosphate Gap junctional permeability requires a cytoplasmic concentration of A4 (Loewenstein, 1967; Loewenstein et free Ca2+ below 10-5 to al., 1967; Payton et al., 1969; Rose and Loewenstein, 1971). Permeability falls markedly when free Ca?+ is injected into a cell through a micropipette, enters the cell from the extracellular compartment via a tear in the cell membrane, or diffuses into the cell sap from damaged mitochondria (Loewenstein, 1967; Politoff et al., 1969; Rose and Loewenstein, 1975a,b). When intracellular Ca2+levels equilibrate to levels of Ca2+typically present in extracellular fluids (i.e., lo-?to 1 0 - ~M ) junctional perme-

INTERCELLULAR JUNCTIONS IN CANCER

65

ability falls and becomes essentially the same as that of the general plasma membrane ( Oliveira-Castro and Loewenstein, 1971). Junctional permeability can be restored to normal by reestablishing low cytoplasmic free Ca2+levels (Rose and Loewenstein, 1975a). Maintenance of a low intracellular level of free Caz+ requires the adequate function of ATP-driven membrane pumps. Predictably, inhibition of ATP synthesis or excess ATP utilization raises intracellular Ca2+ levels and depresses junctional permeability (Politoff et al., 1969). The uncoupling of the gap junction by an elevation of intracellular Ca2+ may represent a significant early pathophysiologic response to cell injury since an influx of extracellular Caz+ into the cytoplasm of an injured cell will effectively insulate that cell from its noninjured neighbors ( Loewenstein and Penn, 1967; Loewenstein, 1972). Extracellular levels of Ca2+and Mg2+also influence junctional permeability by regulating the tightness of the junctional seals that insulate junctional conduits from the extracellular compartment ( Rose and Loewenstein, 1971). 2. 3s-Cyclic Adenosine Monophosphate (CAMP) cAMP is an intermediate or “second messenger” in the action of a number of hormones ( Robison et al., 1971). This nucleotide is of particular interest in light of the theme of this review because it probably is involved in the regulation of cell growth in normal cells and is capable of restoring contact inhibition of cell proliferation to transformed CUItured cells (Burk, 1968; Hsie and Puck, 1971; Otten et al., 1971; Sheppard, 1971; Burger et al., 1972; Seifert and Paul, 1972; Smets, 1972; Willingham et al., 1972; Tee1 and Hall, 1973). Whereas low intracellular levels of cAMP are present in actively growing cell cultures, high concentrations inhibit cell proliferation ( Sutherland, 1970; Frank, 1972; Froehlich and Rachmeler, 1972; Seifert and Paul, 1972). There is evidence that intimate communication, by the transfer of information carrying molecules at gap junctions, is a CAMP-mediated phenomenon (Hax et al., 1974a,b). Studies on salivary glands of the larvae of Drosophila hy&i in which intracellular cAMP levels are elevated by incubation of the gland in a medium containing either dibutyryl-CAMP, theophylline, or ecdysterone have shown that increases in intracellular cAMP are accompanied by increases in gap junctional permeability and decreases in the permeability of the general plasma membrane (Hax et al., 1974a,b).

3. Electrochemical Potential across the General Plasma Membrane Permeability of gap junctions is influenced by the potential across the general ( nonjunctional ) membrane. Evidence for a relationship be-

66

RONALD s. WEINSTEIN, FREDERICK B. MEW, AND JOSEPH ALROY

tlveen general plasma-membrane potential and gap-junction permeabdity comes from several observations. A depolarizing current uncouples lowresistance junctions in glandular epithelium (Socolar and Politoff, 1971), and junctional permeability can be restored by a repolarizing current under certain experimental conditions (Rose, 1971; Rose and Loewenstein, 1971) . Changes in junctional permeability that accompany changes i n membrane potential are probably due in part to intracellular Ca2+, since depolarization of the membrane produces an increase in Ca2+permeability and Ca2+enters the cell. 4. Suminary Based upon these observations and theoretical considerations, Hax and his associates have proposed that CAMPlevels may play a significant role in the molecular regulation of coupling phenomena (Hax et al., 1974a,b) . They suggest that hormonal activation of the adenylate-cyclase system, which results in an increase in intracellular cAMP and a concomitant decrease in ATP (Robison et al., 1971), may increase junctional permeability and thus enhance the flow of low-molecular-weight informational molecules from cell to cell. Conversely, decreased intracellular concentrations of cAMP would impede the transfer of substances at low-resistance junctions. According to their scheme, cAMP and intracellular Ca’+ may be components of a feedback system that regulates cell to cell communication. It is known that low intracellular concentrations of Ca?+ are responsible for a high level of adenylate-cyclase activity, which in turn results in an increased synthesis of CAMP (Bradham et al., 1970). Conversely, increased levels of intracellular Ca2+ inhibit adenylate-cyclase activation ( Bar and Hechter, 1969; Rasmussen, 1970; Bradham et al., 1970; Rubin et al., 1972). Therefore, cAMP and intracellular free Ca’+ appear to be mutually involved in a negative feedback control of their respective concentrations, although the precise mechanisms by which this is accomplished remain to be elucidated. Hax et al. (1974a) postulate that the same control system might regulate intracellular communication at gap junctions. They suggest that an elevated cAMP concentration may initiate the conversion of an inactive precursor of membrane-bound phospholipase A into its active form. A resulting increase of lysophospholipid of the junctional membrane may enhance intercellular communication ( Hax et al., 1974a).

D. METABOLICCOUPLING AT GAP JUNCTIONS Metabolic coupling (also called “metabolic cooperation”) of cells (see Section III,B ) has been demonstrated in several tissue-culture systems. Subak-Sharpe and his associates first demonstrated metabolic coupling

INTERCELLULAR JUNCTIONS I N CANCER

67

by showing that a defective phenotype of certain mutant cells in tissue culture can be corrected by intimate contact with normal cells (Burk et al., 1968; Subak-Sharpe et al., 1969). They found that mutant Chinese hamster fibroblasts, deficient in the enzyme inosinic guanylic pyrophosphorylase ( IPP- ) , are incapable of incorporating exogenous hypoxanthine into their nucleic acids when they grow alone, but that they do incorporate hypoxanthine when grown in contact with a wild-type cell (IPP'). The most probable explanation of these observations is that a low-molecular-weight molecule, such as a nucleotide or a nucleotide derivative, is transferred from the wild-type cell to the mutant cell, thus bypassing the enzyme block in the IPP- mutant (Subak-Sharpe et al., 1969; SubakSharpe, 1969). Another possible, but less likely, explanation is that the substance that passes to the mutant cell and endows it with the ability to incorporate [ 3H]hypoxanthine i s episomal DNA, informational RNA, or a regulatory substance that endows the mutant with the ability to synthesize or activate a functional inosinic pyrophosphorylase (Cox et al., 1970; Pitts, 1971). Gilula et al. (1972) examined the relationship between metabolic and ionic coupling and gap junctions. They studied several types of cocultured cells in various combinations in order to determine whether cell-tocell transfer of molecules could be correlated with the presence or the absence of gap junctions. Their test system included three cell lines: Don hamster fibroblasts, a normal cell type, which can incorporate exogenous [3H]hypoxanthine into nucleic acids; and two mutant lines, DA cells and A9 cells, which are both IPP- and therefore cannot incorporate [3H]hypoxanthine when cultured alone. The Don and DA cells are gapjunction formers, whereas the A9 cells are not. In these experiments Don:Don cell combinations formed gap junctions and were both ionically and metabolically coupled. A9:A9 combinations did not form gap junctions and were uncoupled. Don :DA cells formed gap junctions and were ionically and metabolically coupled. These observations elegantly demonstrate that ionic and metabolic coupling may occur if gap junctions are present but not in their absence (Gilula et al., 1972).

E. COUPLING BETWEEN TUMORCELLS Loewenstein and his colleagues provided the first evidence of a defect in low-resistance coupling in solid malignant tumors. They made electrical measurements on carcinomas in liver (Loewenstein and Kanno, 1967), thyroid (Jamakosmanovib and Loewenstein, 1968), and stomach (Kanno and Matsui, 1968) and found an absence of electrotonic coupling. Based on these observations and theoretical considerations, Loewenstein proposed that a genetically determined interruption of junc-

68

RONALD S. WEINSTEIN, FREDERICK B. MEXK, AND JOSEPH ALROY

INTERCELLULAR JUNCrIONS IN CANCER

69

tional communication may be one of many cazrses of cancerous growth (Loewenstein, 196813). Later, Sheridan ( 1970) succeeded in demonstrating electrotonic coupling between some tumor cells within Novikoff hepatomas, an observation consistent with Loewenstein’s prediction of coupling in some tumors of other etiologies (Loewenstein, 1974) as well as with other interpretations ( Sheridan, 1970; McNutt et al., 1971). Loewenstein and his co-workers also studied metabolic coupling in malignant cells and found a correlation between defective cell-to-cell transfer of [3H]hypoxanthine-derived material, and a lack of electrotonic coupling and the capability to transfer dye. They loaded three lines of cancer cells, two derived from the Morris 5123 hepatoma and one from an X-irradiated embryonic hamster cell, with [3H]hypoxanthine, and showed the absence of transfer of the labeled nucleotide to heterotypic IPP- mutant cells in coculture. These tumor lines also lack junctions that are permeable to either inorganic ions or fluorescein (Azarnia et al., 1972). F. GENETICCORRELATIONS In an elegant study, Azarnia and Loewenstein demonstrated a genetic correlation between the simultaneous occurrence of gap junctions, ionic coupling and contact ( density-dependent ) inhibition of growth ( Azarnia and Loewenstein, 1973; Azarnia et al., 1974). Further, they found evidence that the reinstatement of contact inhibition of growth by cell hybridization is accompanied by a parallel correction of gap junction effects. Cultured human Lesch-Nyhan fibroblast cells have gap junctions, are electrotonically coupled, and fluorescein freely diffuses from one cell to another (Fig. 8). They show contact inhibition of growth. CulFIG. 8. Demonstration of a correlation between the occurrence of gap junctions and cell-to-cell coupling in a human cell/mouse cell hybrid system. Left column: Figures a-d show the morphology of four cultured cell lines: ( a ) Lesch-Nyhan human parent cells; ( b ) mouse parent cells CI-1D derivative of an L-cell mouse line); ( c ) early hybrids between human cells and mouse cells; ( d ) revertant hybrids after loss of about 30 out of 46 human chromosomes. Right column: For a cell pair (cell I and 11) in each culture, current ( i = 1 x A) was injected into cell I and, after a 100-msec delay, another current was injected into cell 11. Resulting changes in membrane voltage ( V ) were measured simultaneously in the two cells with intracellular microelectrodes and displayed on an oscilloscope (inset). Simultaneously, fluorescein was injected into cell I, and the fluorescence was photographed in a dark field after a 5-minute interval. The relatively large V, in ( a ) and ( c ) show that there is cell-to-cell passage of small ions (the carriers of i ) . Fluorescein spread in ( a ) and ( c ) demonstrates cell-to-cell passage of small organic molecules. Azarnia et aZ. (1974), by permission.

70

RONALD s. WEINSTEIN, FREDERICK B. MERK, AND JOSEPH ALROY

tured mouse Cl-1D cells have a coupling frequency of zero, will not exchange fluorescein with neighboring C1-1D cells, lack gap junctions, and are not contact inhibited. Hybrids were prepared from Lesch-Nyhan and Cl-1D cells bv fusion with inactivated Sendai virus (Harris 1970). The resulting hybrid cells contain a nearly complete complement of chromosomes, are contact inhibited, have gap junctions, and are coupled with neighboring hybrid cells. This shows that the capacity to form cell associations necessary for coupling is expressed in hybrid membrane. Azarnia and Loewenstein (1973) have also shown that as hybrid cells lose human chromosomes with serial passage, clones appear among the segregants which have reverted to the noncoupling and junction-deficient trait of the mouse parent cells. Azarnia et al. (1974) postulate that the human cells may contribute a genetic factor to the hybrids that corrects the junctional deficiency of the mouse cells. The factor could be a junctional component or, alternatively, a component of the general plasma membrane which may be essential for cell-cell recognition or junction assembly. VI. Cell Junctions in Embryonic Developmeni

If gap junctions are essential for normal growth control processes, we would expect cell junctions to appear early in phylogeny and in embryogenesis of higher organisms (Trelstad et a?.1966, , 1967; Loewenstein, 1968a; Sheridan, 1970; Pappas et al., 1971; Bennett, 1973; Ducibella et al., 1975; Fisher and Linberg, 1975). Cell junctions have now been observed during cmbryogenesis in many systems. For example, blastomeric cells of Triturus pyrrhogaster (It0 and Loewenstein, 1969) and newt embryo cells at the morula stage of development are joined by gap junctions (It0 and Hori, 1966). Potter and his associates (1966) demonstrated junctional communication between dissimilar groups of cells during the later stages of squid embryo development, and, recently, the formation of gap junctions during amphibian neurulation has been described ( Decker and Friend, 1974). Adherens and occludens junctions also have been described in embryonic and fetal development (Hay and Revel, 1969; Revel et aL, 1973). The formation and proliferation of cell junctions in embryos cannot be considered synonymous with progressive differentiation since selective disappearance of junctions is often coincident with maturation. In the newt Xenopus 2aeui.s the eye becomes irreversibly polarized at developmental stage 32, after which it is insensitive to stimuli from tissue surrounding the eye (Hunt and Jacobson, 1972). Gap junctions are present throughout the embryonic neural retina and pigment epithelium up to

INTERCELLULAR JUNCXIONS I N CANCER

71

stage 31/32. At stage 32, gap junctions disappear from the central portion of the retina and from the pigment epithelium. This event appears to be correlated with the time of retinal specification (Dixon and CronlyDillon, 1973,1974).

VII. Cell Junctions and the Biological Behavior of Tumor Cells

A. CONTACT INHIBITION OF MOVEMENT( LDCOMOTION) Normal adult epithelial cells and some nonepithelial cells have a latent but restrained locomotor capacity ( Abercrombie and Heaysman, 1954; Oldfield, 1963). These locomotor restraints are diminished in the course of malignant transformation ( Abercrombie et al., 1957; Temin and Rubin, 1958; Abercrombie and Ambrose, 1962). In general, mobile nonneoplastic epithelial cells become quiescent late in embryonic development. However, they retain the ability to resume locomotion when tissues are damaged or when cells are excised and transferred to a tissue culture environment. Normal cells at saturation density in confluent culture exhibit locomotor restraints that may employ the same growth control mechanisms which are operative in intact tissues. Locomotor restraint manifests itself in culture systems by the distinct tendency of cells to uniformly occupy space on the culture substratum and to spread out as monolayers. This growth pattern was first examined in detail by Abercrombie and his associates. They observed that cell locomotion appears to be inhibited when cells collide with their neighbors (Abercrombie and Ambrose, 1958). The overall result is that cells spread out evenly and tend not to overlap ( Abercrombie and Heaysman, 1952, 1954; Abercrombie et al., 1957). The phenomenon of directional inhibition of locomotion in culture has been called “contact inhibition of movement” ( Abercrombie and Heaysman, 1954; Abercrombie, 1970). Many strains of malignant cells in culture are relatively insensitive to contact inhibition. However, this characteristic cannot be considered an invariant property of all malignant cells. It has been suggested that the loss of contact inhibition observed in vitro can be compared with loss of control over cellular movement, thought to exist in solid tumors ( Abercrombie and Ambrose, 1962). This line of reasoning das been extended to compare loss of contact inhibition with tumor invasiveness in vivo, a notion unsupported by current evidence. Several theories attempt to identify a single mechanism of contact inhibition of movement. However, there is a growing awareness that it is probably a multifactorial process (Martz and Steinberg, 1972; Stein-

72

RONALD

s. WEINSTEIN,

FREDERICKB. MF.RK, AND JOSEPH ALROY

berg, 1973). The three unitary theories that have received widest attention are: (1) the mutual adhesion theory; ( 2 ) the differential adhesion theory; and (3) the locomotory paralysis theory. Two of these theories, the mutual adhesion and the locomotory paralysis theory, could implicate cell junctions in the process of contact inhibition of movement.

1. Mutual Adhesion Theory The mutual adhesion theory suggests that contact inhibition results from the formation of strong and long-lasting adhesions between cell pairs. The observation that untransformed 3T3 fibroblasts form adhesions that last three times longer than adhesions between simian virus (SV40) transformed 3T3 cells has been interpreted as strong evidence for this theory (Gail and Boone, 1971). According to the mutual adhesion theory, the reduced susceptibility of transformed cells to contact inhibition is a reflection of the relatively short time that they are immobilized by their adhesions. Cell-to-cell adhesion in these culture systems is partially mediated by cell junctions. McNutt et al. (1973) found that untransformed 3T3 cells are joined together by many adherentes junctions, whereas junctions are less prominent in SV40-transformed 3T3 cells, an observation that may explain the differences in adhesion observed by others (Gail and Boone, 1971). The mutual adhesion theory fails to explain how cell-cell overlapping is prevented by contact inhibition. 2. Differential Adhesion Theary The differential adhesion theory proposes that contact-inhibited cells have a greater affinity for their glass or plastic supporting substratum than for one another. Contact-inhibited cultured cells, according to some experimental data, may adhere more strongly to the substratum than to each other, which would explain their tendency to form a monolayer (Abercrombie, 1961; Carter, 1967). Overlapping of cells would be minimized, since overlapping would require the loss of strong cell-substratum adhesions and the retention of weaker cell-to-cell adhesion. However, Harris (1973) has challenged the central assumption of the differential adhesion theory. He has made the observation that many contact-inhibited cells form coherent sheets and that considerable portions of these cell sheets often pull away spontaneously from the substratum. This would require a relatively greater strength of intercellular adhesiveness than would be present at the substratum front (Coman, 1961; Harris, 1973).

3. Locomotoy Paralysis Theoy According to the locomotory paralysis theory, cell locomotor mechanisms are paralyzed by a signal that emanates from the point of contact

INTERCELLULAR JUNCTIONS I N CANCER

73

with neighboring cells ( Abercrombie, 1961) . This theory attempts to account for the events that accompany contact inhibition of movement, but the theory cannot be evaluated in a detailed fashion at the molecular level because of broad gaps in our understanding of mechanisms by which cells propel themselves. Some proponents have speculated that contact inhibition by locomotory paralysis might be mediated by a chemical agent, perhaps in the form of a diffusible substance that is transferred from cell to cell at gap junctions. This idea was particularly attractive to cancer biologists, since many tumors are deficient in gap junctions (see Sections I V and V,E). Recent experimental evidence does not support the notion that molecular transfer from cell to cell at gap junctions can account for contact inhibition of movement. Ultrastructural and electrical studies reveal that gap junctions are absent at the initiation of contact inhibition in some culture systems (Flaman et al., 1969; Heaysman and Pegrum, 1973a,b; Goshima, 1969). On the other hand, some sarcoma cell lines, which are contact inhibited, are electrotonically coupled ( Furshpan and Potter, 1%8) and have gap junctions (Pinto da Silva and Gilula, 1972). Also, [3H]hypoxanthine-loaded Lesch-Nyhan skin fibroblasts can transfer nucleotides or their derivatives to either contacted-inhibited mouse 3T3 cells or contact-uninhibited transformed 3T3 cells with equal efficiency (Cox et al., 1974) showing that contact inhibition and metabolic coupling are unrelated phenomena. Thus, there is no correlation between the presence of functional gap junctions and contact inhibition of movement. Kolodny speculates that cell-to-cell transfer of macromolecules, such as RNA (Kolodny, 1971) is involved in contact inhibition of movement ( Kolodny, 1974). Exchange of large molecules would presumably occur at the general plasma membrane, since there is no evidence that these macromolecules are transferred through cell junctions.

4. Reevaluation of the Concept of Contact Inhibition Movement Steinberg and his associates have challenged the concept of contact inhibition of movement. They used a Nomarski optical system to prepare time-lapse films of 3T3 mouse cells in confluent cultures and found that so-called contact-inhibited 3T3 cells are capable of translocational mobility along the substratum and, in fact, move readily with respect to their neighbors, although they are restrained from moving over the surface of one another (Martz and Steinberg, 1972; Steinberg, 1973). Wiseman and Steinberg (1973) extended these observations to solid tissues. They monitored the movements of single cells within three-dimensional tissue masses under conditions where contact inhibition of movement would be expected to be present. They seeded radiolabeled embryonic cells onto the surface of embryonic heart and liver tissue

74

RONALD S .

WEINSTEIN, FREDERICK B. MERK, AND

JOSEPH ALROY

fragments (or aggregate$) and followed the positions of the seeded cells by autoradiography. They were able to demonstrate that single cells can penetrate solid tissues and can migrate for considerable distances within the tissue masses. These observations are inconsistent with the idea that cell-to-cell contact entirely curtails cell movement.

B. POSTCONFLUENCE INHIBITIONOF GROWTH(CELLDIVISION) In many nonmalignant culture systems, extensive cellular contact switches off net RNA and protein synthesis, and blocks the synthesis of new DNA (Stoker, 1967). This phenomenon has been called variously “contact inhibition of growth,” “density-dependent inhibition,” “contact regulation of cell division,” and “postconfluence inhibition of cell division‘‘ ( Martz and Steinberg, 1972). Contact inhibition of movement and postconfluence inhibition of cell division are operationally distinct and may be the manifestations of different control mechanisms (MacieiraCoelho, 1967; Stoker, 1967). However, postconfluence inhibition of cell division, like contact inhibition of movement, tends to bc markedly reduced in cultured tumor cells. Kinetic studies relating cell contact to inhibition of cell division suggest that the two processes are unrelated. Martz and Steinberg (1972) examined 3T3 cells at confluence and found that inhibition of cell division appears one cell generation after cells come into initial contact. From this they concluded that factors in addition to simple cell-to-cell “contact” must contribute to inhibition. They proposed several alternative mechanisms to explain the phenomenon. One of the mechanisms may implicate gap junctions. They suggested that areas of membrane contact may not become effective inhibitors of division until the contacts have “matured.” Stated another way, time may be required for fully functional cell junctions to form at the cell surface. Rates of formation for junctions are consistent with this idea since junctions take minutes or hours to become coupled (It0 et aZ., 1974a,b; Johnson et al., 1974). On the other hand, many cells that remain in contact probably never become uncoupled, even during division. Electrotonic coupling ( O’Lague et al., 1970) and gap junctions (Merk and McNutt, 1972) persist between mitotic and interphase cells, indicating that intercellular communication probably is not related to inhibition of division. C. INVASION AND METASTASES

Cell junctions are discussed in relation to tumor invasion in Sections 1\7,A,2,b; IV,A,4; and VIII,B. Several additional points about the invasion

INTERCXUULAR JUNCXIONS IN CANCER

75

phenomenon and its relation to growth control should be underscored for purposes of clarity. First, the control of growth (i.e., DNA synthesis) and the phenomenon of invasion are different processes and may well be unrelated to one another. Growth control is compromised in both benign and malignant tumors, but invasiveness is associated exclusively with cancer. Because of an abnormality of growth control, a benign tumor may grow expansively and may achieve enormous size, but the tumor will not invade the connective tissue stroma. On the other hand, in some malignan t tumors cell proliferation is reasonably well controlled, as evidenced by the tumors’ small size and low mitotic index, and yet such tumors may readily invade stroma and even metastasize widely throughout the body. Therefore, invasion and metastases, the aspects of tumor behavior which most frequently threaten life, are not the direct sequelae of loss growth control but result from other tumor cell properties and host factors. VIII. Intercellular Adhesion in Tumors

A. ADHESIONAT CELLJUNCTIONS A relatively low strength of adhesion between tumor cells may contribute to the biological behavior of malignancies (Coman, 1944, 1961) and may explain certain patterns of tumor growth (Steinberg, 1963). Intercellular adhesion in normal tissues as well as tumors is clearly a multifactorial process (Curtis, 1973). All classes of cell junctions, plus general plasma membrane components ( e.g., gIycoproteins, glycolipids, etc. ), microexudates, and divalent ions may all contribute to intercellular adhesion. Many methods have been developed to quantitate the strength of cell-to-cell adhesion in vitro (Berwick and Coman, 1962; Roth and Weston, 1967; Orr and Roseman, 1969; Curtis, 1970; Steinberg, 1970; Armstrong, 1971; Roth et al., 1971; Walther et al., 1973), but it is doubtful that each method measures the same molecular event (Walther et al., 1973). None of these quantitative methods have been used to estimate the relative contributions of each type of cell junction or other membrane components to adhesiveness. Since cell junctions form rapidly after cells come into contact (It0 and Loewenstein, 1969; Flaxman et al., 1969; Johnson et al., 1974; Ito et al., 1974a,b), it is likely that they are major contributors to cell adhesion as measured by these methods. Several electron microscopy studies provide qualitative information on the relative contributions of various plasma membrane components to adhesiveness under highly artificial conditions, including extracellular Caz+depletion (Sedar and Forte, 1964; Muir, 1967) or incubation of

76

RONALD S. WEINSTEIN, FREDERICK B. MERK, AND JOSEPH ALROY

tissue in a markedly hyperosomotic medium ( B a n et al., 1965, 1968; Goodenough and Gilula, 1974). In general, these studies suggest that cell-to-cell adhesion per unit area of surface membrane is considerably greater at cell junctional sites than in regions of nonjunctional close apposition of neighboring cells. Intercellular adhesion may be decreased in some solid tumors. This concept gained early support from the work of Coman and his associates, who attempted to compare the cell-to-cell adhesion of different tissues (Coman, 1944; McCutcheon et al., 1948; Coman and Anderson, 1955; Berwick and Coman, 1962). They devised a micromanipulation method to pull cells apart, using a calibrated microneedle. When pulling away an impaled cell, the degree of bend in the microneedle was used as an index of strength of adhesion (Coman, 1944, 1961). They found that carcinoma cells are less adherent than normal epithelial cells. This conclusion is consistent with the clinical observation that carcinoma cells exfoliate more readily than normal cells. However, many factors, including the absence of strongly interadherent superficial cells and increases in cell turnover rates, probably contribute to cell shedding from tumors. The reduction of intercellular adhesion described by Coman may arise at an early stage of malignant transformation (Coman, 1960). Analysis and comparison of intercellular adhesion within both normal stratified epithelium and tumors is a complex problem because of differences in intercellular adhesion at different levels in tissues. Heterogeneity of adhesiveness within epithelia is suggested by ultrastructural data. Basal cells in uterine cervical epithelium (McNutt et al., 1971) and in urinary bladder uroepithelium (B. U. Pauli and R. S. Weinstein, unpublished data) have fewer cell junctions than intermediate and superficial cells, suggesting that cell interadhesion may increase with maturation in these tissues. Further, a junctional complex (e.g., the terminal bar of columnar epithelia), consisting of several types of cell junctions in close juxtaposition along the membrane, forms near the apex of superficial cells in many normal epithelia and provides a very strong zone of intercellular adhesion. Junctional complexes are attenuated or absent in many malignancies ( Fulker et al., 1971; Weinstein et aE., 1974). These and other variations in ultrastructure should be considered when attempting to compare cell-to-cell adhesion in normal tissue with adhesion in tumors. Since typical tumor cells are usually incompletely differentiated, the only valid comparisons would be between strengths of adhesion between tumor cells and adhesion between cells from a level in normal epithelium at a comparable stage of differentiation. These comparisons would be exceedingly difficult to make in soiid tissues for technical reasons, including the position of comparable cell layers deep within normal epithelium.

INTERCELLULAR JUNCTIONS I N CANCER

77

B. TUMOR DISSEMINATION Since the turn of the century, it has been suspected that cancer cells may be disseminated into the general circulation by surgical trauma ( Tyzzer, 1913). Mechanical manipulation of malignant tumors during surgery presumably dislodges weakly adhering tumor cells. This idea provided the rationale for the development of the “no touch isolation” procedure for resecting colonic cancers, as described by Turnbull et al. (1967). They found that ligation of efferent tumor blood vessels prior to handling of the tumors improved survival rates. IX. Transepithelial Permeability and Malignant Transformation

Zonula occludens junctions (tight junctions) form a permeability seal between epithelial cells (see Section 111,A). Leakiness of zonula occludens junctions is increased by many factors, including electrical shock (Hirano et uZ., 1970), surgical trauma (Rhodes and Karnovsky, 1971), elevation of intravascular hydrostatic pressure (Pietra et d.,1969), immersion in hypertonic solutions (Erlij and Martinez-Palomo, 1972), inhalation of cigarette smoke (Simani et al., 1974), and exposure to chemical carcinogens, such as dibutylnitrosamine or methylnitrosourea ( Hicks et al., 1974). These and other noxous stimuli, which compromise the normal protective function of epithelia, may increase the exposure of basal cell layers to water-soluble carcinogens (Simani et al., 1974) and by this mechanism exert a cocarcinogenic effect. Attenuation of occludentes junctions ( Martinez-Palomo, 1970a; Fulker et al., 1971) (see Table I ) and a concomitant increase in transepithelial permeability (Hicks et aZ., 1974) is a common occurrence in carcinomas. The extent of junctional attenuation appears to correlate well with the level of tumor anaplasia (Weinstein et aZ., 1974), at least in some tumor systems. X. Summary

Intercellular junctions (cell junctions) are a set of structurally complex membrane components that are incorporated into the general plasma membrane at sites of close cell-to-cell apposition. The functions of some types of cell junctions are reasonably well understood. For instance, we now know that all types of junctions contribute to intercellular adhesion and that the zonula occludens endows epithelia with a sealing capacity against bypass diffusion. However, the primary function of some types of junctions, such as gap junctions, remains obscure; this is unfortunate since a considerable body of information on the occurrence, bio-

78

RONALD

s. WEINSTEIN, FREDERICK B. MERK, AND

JOSEPH ALROY

chemical ultrastructure, and physical properties of junctions suggests that they probably do play a central role in important biological phenomena. The pathology literature contains many references to cell junctions in benign and malignant solid tumors. Quantitative evaluations of the occurrence of junctions are subjective in most of these reports, although there are a few reports in which the data were obtained by quantitative electron microscopy techniques. Included in this review is a collation of much information on the occurrence of cell junctions in different types of tumors (Table I). These data show that junctional deficiencies are common in tumors, but they fail to demonstrate any consistent patterns of junctional deficiencies. This may be due to the imprecision of the data more than anything else. In spite of numerous reports to the contrary, there is neither concrete nor compelling circumstantial evidence which supports the popular notion that junctional defects contribute to those properties which are the hallmarks of malignant growth, namely, invasiveness and the ability to metastasize. In our introductory remarks, we noted that a fundamental problem in cancer research is to identify the cell products that are coded for by the genes of neoplastic transformation and to determine how these products can account for the biology of tumors. It is therefore germane to ask whether genes of neoplastic transformation may be related to qualitative and quantitative abnormalities in intercellular junctions, which are observed in some tumors. At the present time, we are unable to provide a definitive answer to this question. However, there are a number of different mechanisms by which gene products could influence intercellular junction formation and function. The most obvious mechanism involves the abnormal synthesis and/or assembly of junctional components. Evaluation of these possibilities must await the successful isolation of specific junctional components and comparative studies of these components in normal tissues and in tumors. Methodology for the isolation of junctional components is rapidly evolving. Another mechanism affectingjunctions probably involves alterations in the general ( nonjunctional) plasma membrane. As discussed in Section II,B, most cell junctions are formed by the structural modification of the general plasma membrane in a stepwise fashion. In tumors, abnormal gene products may modify the general plasma membrane and thus deprive the cell of normal junction assembly sites. This could account for quantitative changes in junctions as well as abnormalities in their biochemical ultrastructure and function. Such a mechanism might account for the linked changes in the occurrence of several types of junctions, as has been observed in several tumor systems (Weinstein et id.,1974). Alternatively, genetic modifications of cell surface components involved in cell-to-cell

INTERCELLULAR JUNCTIONS IN CANCER

79

recognition, either by modification of elements of the cellular machinery responsible for recognition or by masking of recognition sites, could interfere with the formation of junctions. Conversely, unmasking of sites may enhance junction formation. Thus, there are a number of ways in which genes of neoplastic transformation may influence intercellular junction structure and function. Additional research will be required to define the contribution of tumor gene products to the pathogenesis of junctional abnormalities in tumors and to elucidate the roles, if any, of these cell membrane defects in malignant growth.

ACKNOWLEDGMENTS We are grateful to Mrs. Marianne Alroy and Mr. Steven Halpern for editing the manuscript and Ms. Shirley Hunter and Ms. Renee Slack for typing the manuscript. This study was supported by National Cancer Institute Grants CA-14447 and CA16377 from the National Institutes of Health, United States Public Health Service.

REFERENCES Abercrombie, M. ( 1961). Exp. Cell Res., Suppl. 8, 188. Abercrombie, M. (1970). In Vitro 6, 128. Abercrombie, M., and Ambrose, E. J. (1958). Exp. Cell Res. 15, 332. Abercrombie, M., and Ambrose, E. J. (1962). Cancer Res. 22, 525. Abercrombie, M., and Heaysman, J. E. M. ( 1952). Exp. Cell Res. 5, 11. Abercrombie, M., and Heaysman, J. E. M. (1954). Exp. Cell Res. 6, 293. Abercrombie, M., Heaysman, J. E. M., and Karthauser, H. M. (1957). Exp. CeZl. Res. 13, 276. Able, M. E., and Lee, J. C. (1969). Cancer 23,481. Ahmed, A. (1974). J. Pathol. 112, 177. Aikawa, M., and Ng, A. B. P. (1973). Cancer 31,385. Akhtar, M., Gosalbez, T. G., and Brody, H. (1973). Arch Pathol. 96, 161. Albertini, D. F., and Anderson, E. (1974). J. Cell Biol. 63,234. Alroy, J., and Weinstein, R. S. ( 1976). J. Natl. Cancer Inst. In press. Alroy, J., Weinstein, R. S., and Pauli, B. U. (1976). Am. J. Path. In press. Amin, H. K., Ferenczy, A., and Richart, R. M. (1974). J. Comp. Pdhol. 84, 161. Anderson, H. C., Byunghoon, K., and Minkowitz, S. (1969). Cancer 24, 585. Armstrong, P. B. ( 1971). Wilhelm Roux' Arch. Entwicklungsmech. Organismen 168, 125. Azarnia, R., and Loewenstein, W. R. ( 1971). J. Membrane Biol. 6, 368. Azamia, R., and Loewenstein, W. R. (1973). Nature (London) 241, 455. Azamia, R., Michalke, W., and Loewenstein, W. R. (1972). J. Membrane Biol. 10, 247. Azarnia, R., Larsen, W. J., and Loewenstein, W. R. (1974). Proc. Nut. Acud. Sci. U.S. 71, 880. Baba, N., and von Haam, E. ( 1971). J. Nut. Cancer Inst. 47,675. Babai, F., and Tremblay, G. ( 1972). Cancer Res. 32,2765. Bar, H. P., and Hechter, 0. (1969). Biochem. Biophys. Res. Commun. 35, 681. Ban, L., Dewey, M. M., and Berger, W. (1965). 1. Gen. Physiol. 48, 797.

80

RONALD s. WEINSTEIN, FREDERICK B. MERK, AND JOSEPH ALROY

Barr, L., Berger, W., and Dewey, M. hl. (1968). J. Gen. Physiol. 51, 347. Barton, A. A. ( 1964). Brit. J. Cancer 18,682. Battifora. H. A. (1973). Cancer 31, 1418. Battifora, H. A., Eisenstein, R., Sky-Peck, H. H., and McDonald, J. H. (1965). J. Urol. 93, 217. Battifora, H. A., Eisenstein, R., and Schild, J. A. (1969). Cancer 23, 183. Behrens, U. J., Mashbum, L. T., Stevens, J., Hollander, V. P., and Lampen, N. ( 1974). Cancer Res. 34, 2926. Bencosnie, S. A., Allen, R. A., and Latta, H. (1963). Amer. J . Pathol. 42, 1. Benedetti, E. L., and Emmelot, P. (1968a). J. Cell Biol. 38, 15. Benedetti, E. L., and Emmelot, P. (1968b). In “The Membranes” (A. J. Dalton and F. Haguenau, eds.), pp. 51, 52, and 56. Academic Press, New York. Bennett, M. V. L. (1973). Fed. Proc., Fed. Amer. SOC. Exp. Bwl. 32, 65. Bennett, M. V. L., Spira, M. E., and Pappas, G. D. (1972). Deuelop. B i d . 29, 419. Bennington, J. L. (1969). Cancer Res. 29, 1082. Bensch, K. G., Conin, B., Pariente, R., and Spencer, H. (1968). Cancer 22, 1163. Berry, h4. N., and Friend, D. S. ( 1969). J. Cell Biol. 43, 506. Berwick, L., and Coman, D. R. ( 1962). Cancer Res. 22, 982. Black, H. E., Capen, C. C., and Young, D. M. ( 1973). Cancer 32, 865. Bordi, C., Anversa, P., and Vitali-Massa, L. ( 1972). Virchows Arch., A 357, 145. Borek, C., Higashino, S., and bewenstein, W. R. (1969). J. Membrane Bipl. 1, 274. Borysenko, J . Z., and Revel, J. P. (1973). Amer. J. Anat. 137, 403. Bradham, L. S., Holt, D. A,, and Sims, M. (1970). Biochim. Biophys. A d a 201, 250. Bransilver, B. R., Ferenczy, A., and Richart, R. M. (1974). Arch. Pathol. 98, 76. Branton, D. ( 1966). Proc. Nut. Acad. Sci. U.S. 55, 1048. Branton, D. (1969). Annu. Reu. Plorlt Physiol. 20,209. Branton, D., and Deamer, D. W. (1972). In “Membrane Structure,” p. 41. SpringerVerlag, Berlin and New York. Branton, D., Bullivant, S., Gilula, N. B., Karnovsky, M. J., Moor, H., Miihlethaler, K., Northcote, D. H., Packer, L., Satir, B., Satir, P., Speth, V., Staehelin, L. A., and Weinstein, R. S. (1975). Science 190, 54. Breathnach, A. S., Stolinski, G., and Gross, M. ( 1972). Micron 3,287. Breton-Gorius, J., Flandrin, G., Daniel, M. T., Chevalier, J., Lebeau, M., and Sanel, F. T. ( 1975). Virchows Arch B. Cell Path. 18, 165. Bretscher, M. S. ( 1972). Nature (London),New Biol. 236, 11. Brightman, M. W., and Palay, S. L. (1963). J. Cell Biol. 19, 415. Brown, W. J., Barajas, L., Waisman, J., and DeQuattro, V. (1972). Cancer 29, 744. Bullivant, S. ( 1970). In “Biological Techniques in Electron Microscopy” (D. F. Parsons, ed. ), p. 101. Academic Press, New York. Bullivant, S. ( 1973 ). In “Advanced Techniques in Biological Electron Microscopy” (J. K. Koehler, ed.), p. 67. Springer-Verlag, Berlin and New York. Bullivant, S. (1974). Phil. Trans. Roy. SOC.London, Set. B 268,5. Bullough, W. S. ( 1965). Cancer Res. 25,1683. Burger, M. M., Bombik, B. M., Breckenridge, B. McL., and Sheppard, J. R. (1972). Nature (London),New Biol. 239, 161. Biirk, R. R. (1968). Nature (London) 219, 1272.

INTERCELLULAR JUNCTIONS IN CANCER

81

Burk, R. R., Pitts, J. D., and Subak-Sharpe, J. H. (1968). Exp. Cell Res. 53,297. Burns, W. A., Matthews, M. J., Hamosh, M., van der Weide, G., Blum, R., and Johnson, F. B. (1974). Cancer 33,1002. Campbell, R. D., and Campbell, J. H. (1971). In “Origin and Continuity of Cell Organelles” (J. Reinert and H. Ursprung, eds.), p. 261. Springer-Verlag, Berlin and New York. Carter, L. P., Beggs, J., and Waggoner, J. D. ( 1972). Cancer 30, 1130. Carter, S. B. ( 1967). Nature (London) 213, 256. Castor, L. N. (1968). J. Cell. Physiol. 72, 161. Chalcroft, J. P., and Bullivant, S. (1970). J. Cell Biol. 47, 49. Claude, P., and Goodenough, D. A. (1973). J. Cell Biol. 58, 390. Coman, D. R. ( 1944). Cancer Res. 4, 625. Coman, D. R. ( 1960). Cancer Res. 20, 1202. Coman, D. R. ( 1961). Cancer Res. 21, 1436. Coman, D. R., and Anderson, T. F. (1955). Cancer Res. 15,541. Cooper, E. H. (1972). Ann. Roy. Coll. Surg. Engl. 51, 1. Cornog, J. L. (1969). Arch. Pathol. 87, 404. Cox, R. P., Krauss, M. R., Balis, M. E., and Dancis, J. (1970). Proc. Nut. Acad. Sci. US.67, 1573. Cox, R. P., Krauss, M. R., Balis, M. E., and Dancis, J. (1974). In “Cell Communication” (R. P. Cox, ed.), p. 67. Wiley, New York. Crick, F. ( 1970). Nature (London) 225,420. Curtis, A. S. G. (1970). J. Embryol. Exp. Morphol. 23,253. Curtis, A. S. G. (1973). Progr. Biophys. Mol. Biol. 27, 315. Deamer, D. W., and Branton, D. (1967). Science 158,655. Decker, R. S., and Friend, D. S. ( 1974). J. Cell Biol. 62, 32. DeHaan, R. L., and Sachs, H. G. (1973). Curr. Top. Deuelop. Biol. 8, 193. Diamond, J. M. (1974). Fed. Proc., Fed. Amer. SOC. Exp. Biol. 33, 2220. Dixon, J. S., and Cronly-Dillon, J. R. (1973). J. Embryol. Erp. Morphol. 28, 659. Dixon, J. S., and Crody-Dillon, J. R. (1974). Nature (London) 251, 505. Douglas, W. H. J., Ripley, R. C., and Ellis, R. A. ( 1970). J. Cell Biol. 44, 211. Dreifuss, J. J., Girardier, L., and Forssmann, W. G. (1966). PfEzregers Arch. Gesamte Physiol. Menschen Tiere 292, 13. Ducibella, T., Albertini, D. F., Anderson, E., and Biggers, J. D. (1975). Deuelop. Biol. 45, 231. Dunia, I., Sen Ghosh, C., Benedetti, E. L., Zweers, A., and Bolemendal, H. (1974). FEBS (Fed. Eur. Brochem. SOC.) Lett. 45,139. Echevarria, R. A. ( 1967). Cancer 20,563. Elgsaeter, A., and Branton, D. (1974). J . Cell Biol. 63, 1018. Erlandson, R. A., and Carstens, P. H. B. ( 1972). Cancer 29,987. Erlandson, R. A., and Tandler, B. (1972). Arch. Pathol. 93, 130. Erlij, D., and Martinez-Palomo, A. ( 1972). J. Membrane Biol. 9, 229. Evans, W. H., and Curd, J. H. (1972). Biochem. J. 138,1972. Farquhar, M. G., and Palade, G. E. (1963). 1. Cell Biol. 17, 375. Favara, B. E., Johnson, W., and Ito, J. ( 1968). Cancer 22, 845. Fawcett, D. W. (1961). Exp. Cell Res., Suppl. 8, 174. Fawcett, D. W., Slautterbach, D. L., and Ito, S. (1959). J. Biophys. Biochem. Cytol. 5, 453. Fechner, R. E., Bentinck, B. R., and Askew, B., Jr. (1972). Cancer 29,501. Feldman, P. S., Horvath, E., and Kovacs, K. (1972). Cancer 30, 1279.

82

RONALD S. WEINSTEIN, FREDERICK B. MERK, AND JOSEPH ALROY

Ferenczy, A., Fenoglio, J., and Richart, R. M. ( 1972). Cancer 30,244. Fisher, E. R. (1969). Cancer 24, 312. Fisher, E. R., and Horvat, B. (1972). Cancer 30, 1074. Fisher, E. R., and Sharkey, D. A. (1962). Cancer 15, 160. Fisher, E. R., McCoy, M. M., XI, and Wechsler, H. L. (1972). Cancer 29, 1387. Fisher, S. K., and Linberg, K. A. ( 1975). J. Ultmstruct. Res. 51,69. Flaks, B., Cooper, E. H., and Knowles, J. C. (1970). Eur. I. Cancer 6, 145. Flaxman, B. A. (1972). Cancer Res. 32,462. Flaxman, B. A., Revel, J, P., and Hay, E. D. ( 1969). Exp. Cell Res. 58,438. Frable, W. J., Still, W. J. S., and Kay, S. ( 1971). Cancer 27, 667. Frank, W. (1972). Exp. CeU Res. 71, 238. Friend, D. S., and Cilula, N. B. (1972a).J. Cell Biol. 53, 148. Friend, D. S., and Gilula, N. B. ( 1972b). J. Cell Biol. 53, 758. Froehlich, J. E., and Rachnieler, M. (1972). J. Cell Biol. 55, 19. Fromter, E., and Diamond, J. (1972). Nature (London), New Biol. 235, 9. Fu, Y. S., and Kay, S. ( 1973). Arch. Pathol. 96, 66. Fulker, M. J., Cooper, E. H., and Tanaka, T. ( 1971). Cancer 27, 71. Furshpan, E. J., and Potter, D. D. ( 1959). J. Physiol. (London) 145, 289. Furshpan, E. J.. and Potter, D. D. (1968). Curr. Top. Deoelop. Biol. 3, 95. Gail, M. H., and Boone, C. W. ( 1971). Exp. Cell Res. 65,221. Chatak, N. R., Hirano, A., and Zimmernian, H. M. (1971). Cancer 27, 1465. Gihla, N. B. (1972). J. Ultrastruct. Res. 38, 215. Gilula, N. B. (1974). In “Cell Communication” ( R . P. Cox, ed.), pp. 1-29. Wiley, New York. Gilula, N. B., Branton, D., and Satir, P. (1970). Proc. Not. Acad. Sci. 67, 213. Gilula, N. B., Reeves, 0. R., and Steinbach, A. ( 1972). Nature (London) 235,262. Goldenberg, V. E., Goldenberg, N. S., and Benditt, E. P. (1969). Cancer 24, 236. Condos, B. (1969). Cancer 24, 954. Condos, B. ( 1971).Cancer 27, 1455. Gonzales-Licea, A., Yardiey, J. H., and Hartmann, W. H. (1967). Cancer 20, 1234. Goodenough, D. A. ( 1974). J. Cell Biol. 61,557. Goodenough, D. A., and Cilula, N. B. ( 1974). J, Cell Biol. 61, 575. Goodenough, D. A., and Revel, J. P. (1970). 1. Cell Biol. 45,272. Goodenough, D. A., and Revel, J. P. (1971). 1.Cell Biol. 50,81. Goodenough, D. A., and Stoeckenius, W. ( 1972). J. Cell Biol. 54,646. Coshima, K. (1969).Exp. Cell Res. 58,420. Curd, J. W., and Evans, W. H. ( 1973). Eur. 3. Biochem. 36,273. Hage, E. (1973). Virchows Arch., A 361,121. Haggis, G . H. (1969). Biochim. Biophys. Acta 193, 237. Hameed, K., and Morgan, D. A. (1972). Cancer 29, 1326. Hamlett, J. D., Aparicio, S. R., and Lumsden, C. E. ( 1971). J. Pathol. 105, 111. Harris, A. ( 1973). Develop. Biol. 35, 97. Harris, C. C., Kaufman, D. G., Spron, M. B., and Staffiotti, U. (1973). Cancer Chemother. Rep., Part c 4,43. Harris, H. (1970). In “Cell Fusion,” p. 1. Harvard Univ. Press, Cambridge, Massachusetts. Harris, J. W., Meyskens, F., and Patt, H. M. (1970). Cancer Res. 30, 1937. Hashimoto, K., Yamanishi, Y., Maeyens, E., Dabbous, M. K., and Kanzaki, T. (1973). Cancer Res. 33, 2790.

INTERCELLULAR JUNCTIONS IN CANCER

83

Hax, W. M. A,, van Venrooij, G. E. P. M., and Vossenberg, J. B. J. (1974a). 1. Membrane Bid. 19, 253. Hax, W. M. A., Demel, R. A., Spies, F., Vossenberg, J. B. J., and Linnemans, W. A. M. (1974b). Exp. Cell Res. 89, 311. Hay, E. D. ( 1961). J . Biophys. Biochem. Cytol. 10,457. Hay, E. D., and Revel, J. P. (1969). Monogr. Deuelop. B i d . I, 1. Heaysman, J. E. M., and Pegrum, S. M. (1973a). E r p . Cell Res. 78, 71. Heaysman, J. E. M., and Pegrum, S. M. ( 1973b). Exp. Cell Res. 7,479. Hicks, R. M., Ketterer, B., and Warren, R. C. (1974). Phil. Tram. Roy. SOC. London, Ser B 268, 23. Hill, G. S., and Eggleston, J. C. (1972). Cancer 30, 1092. Hirano, A., Becker, N. H., and Zimmerman, H. M. (1970). J. Neurol. Sci. 10, 205. Holstein, A. F., and Korner, F. ( 1974). Virchows Arch., A 363,97. Hong, K., and Hubbell, W. L. (1972). Proc. Nut. Acad. Sci. U.S.A. 69, 2617. Horvath, E., Kavacs, K., and Ross, R. C. (1972). Virchows Arch., A 356, 281. Hoshino, M. ( 1963). Cancer Res. 23,209. Hou-Jensen, K., Priori, E., and Dmochowski, L. (1972). Cancer 29, 280. Hou-Jensen, K., Rawlinson, D. G., and Hendrickson, M. (1973). Cancer 32, 809. Hruban, Z., Mochizuki, Y., Slesers, A., and Morris, H. P. (1972). Cancer Res. 32, 853. Hsie, A. W., and Puck, T. T. (1971). Proc. Nut. Acad. Sci. U.S. 68, 358. Hiilser, D. F., and Demsey, A. ( 1973). Z . Natulforsch. C . 28, 603. Hulser, D. F., and Peters, J. H. ( 1972). E r p . Cell Res. 74, 319. Hulser, D. F., and Webb, D. J. (1973). E r p . Cell Res. 80,210. Hunt, R. K., and Jacobson, M. ( 1972). Proc. Nat. Acad. Sci. US.69,2860. Imai, H., and Stein, A. A. ( 1963). Gastroenterology 44, 410. Ioachim, N. J., Delaney, W. E., and Madrazo, A. (1974’). Cancer 34, 586. Ito, S., and Hori, N. (1966). 1. Gen. Physiol. 49, 1019. Ito, S., and Loewenstein, W. R. (1969). Develop. B i d . 19, 228. Ito, S., Sato, E., and Loewenstein, W. R. (1974a). J. Membrane B i d . 19,305. Ito, S . , Sato, E., and Loewenstein, W. R. (197413). J. Membrane Biol. 19, 339. Jamakosmanovih, A., and Loewenstein, W. R. ( 1968). J. Cell Biol. 38, 556. Johnson, R., Herman, W. S., and Preus, D. M. ( 1973). J. Ultrastruct. Res. 43,298. Johnson, R., Hammer, M., Sheridan, J., and Revel, J. P. (1974). Proc. N d . Acad. Sci. US. 71, 4536. Kanbour, A. I., Burgess, F., and Salazar, H. ( 1973). Cancer 31, 1433. Kanno, Y., and Loewenstein, W. R. (1966). Nature (London) 212, 629. Kanno, Y.,and Matsui, Y. (1968). Nature (London) 218, 755. Karnovsky, M. J. (1967). J. Cell Biol. 35, 213. Kay, S., and Schatzki, P. F. ( 1971). Cancer 28,755. Kay, S., Elzay, R. P., and Willson, M. A. ( 1971). Cancer 27,674. Kelly, D. E. ( 1966). J . Cell Biol. 28, 51. Kobayashi, S., and Mukai, N. (1974). Cancer Res. 34, 1646. Kogon, M., and Pappas, G. D. ( 1975). J . Cell Biol. 66, 671. Kolodny, G. M. ( 1971). E r p . Cell Res. 65,313. Kolodny, G. M. (1974). I n “Cell Communication” (R. P. Cox, ed.), p. 97. Wiley, New York. Korn, E. D. (1966). Biuchim. Biophys. Acta 116,317. Korn, E. D., and Weisman, R. A. (1966). Biochim. Biophys. Acta 116, 309.

84

RONALD S . WEINSTEIN, FREDERICK B. MERK, AND JOSEPH ALROY

Kovacs, K., Horvath, E., Delarue, N. C., and Laidlaw, J. C. (1974).Horn. Metab. Res. 5, 47. Krawnyk, W. S., and Wilgram, G. F. (1973). 1. Ultrustruct. Res. 45, 93. Kubo, T. ( 1969).Cancer 24,948. Kubo, T.(1974).Acta Pathol. lap. 24,163. Kuhn, C. (1972).Cancer 30,1107. Lasansky, A. ( 1969).I . Cell Biol. 40,577. Lavin, P., and Koss, L. G. ( 1971). I . Nut. Cancer Inst. 46, 597. Leak, L. V., Caulfield, J. B., Burke, J. F., and McKhann, C. F. (1967). Cancer Res. 27, 261. Leifer, C., Miller, A. S., Putong, P. B., and Min, B. H. ( 1974a).Cancer 34,597. Leifer, C., Miller, A. S., Putong, P. B., and Harwick, R. D. (1974b).Arch. Pathol. 98, 312. Lenard, J., and Singer, S. J. (1968).1. Cell Biol. 37, 117. Letoumeau, R. J., Li, J. J., Rosen, S., and Villee, C. A. (1975). Cancer Res. 35, 6. Levine, G . D. (1973).Cancer 31,729. Lin, H.S., Lin, C. S., Yeh, S., and Tu, S.-M. ( 1969).Cancer 23,390. Locke, M. (1965).J . Cell Biol. 25, 166. Loewenstein, W.R. (1966).Ann. N.Y. A d . Sci. 137,441. Loewenstein, W.R. ( 1967). J. Colloid. Interface Sci. 25, 34. Loewenstein, W. R. ( 1968a).Deuelop. B i d . , Suppl. 19,20. Loewenstein, W. R. (1968b).Perspect. B w l . Med. 11, 260. Loewenstein, W. R. (1972).Arch. Intern. Med. 129,299. Loewenstein, W. R. (1973). Fed. Proc., Fed. Amer. SOC. Exp. Biol. 32, 60. Loewenstein, W. R. ( 1974). In “Membrane Transformation in Neoplasia” (J. Schultz and R. E. Block, eds. ), p. 103.Academic Press, New York. Loewenstein, W. R., and Kanno, Y. ( 1964).1. CeU Biol. 22,565. Loewenstein, W.R., and Kanno, Y. ( 1967). 1. Cell B i d . 33, 225. Loewenstein, W.R.,and Penn, R. D. (1967).I . Cell Biol. 33,235. Loewenstein, W. R., Socolar, S. J., Higashino, S., Kanno, Y., and Davidson, N. ( 1965). Science 149, 295. Loewenstein, W. R., Nakas, M., and Socolar, S. J. (1967).1. Gen. Physiol. 50, 1865. Luft, J.H. (1971).Anat. Rec. 171,347. Lupulescu, A., and Boyd, C. B. ( 1972).Cancer 29, 1530. Lynn, J. A., Varon, H. H., Kingsley, W. B., and Martin, J. H. ( 1967). Amer. J. Pathol. 51, 639. Ma, M. H., and Blackburn, C. R. B. (1973).Cancer Res. 33, 1766. Ma, M. H., and Webber, A. J. (1966).Cancer Res. 26,935. McCutcheon, M., Coman, D. R., and Moore, F. B. (1948). Cancer 1, 460. McGavron, M.H. ( 1965).Cancer 18, 1445. Macieira-Coelho, A. (1967).Exp. Cell Res. 47, 193. Mackay, A. M., Pettigrew, N., Symington, T., and Neville, A. M. (1974). Cancer 34, 1108. Macka!., B., Bennington, J. L., and Skoglund, R. W. ( 1971).Cancer 27, 109. McNutt, N. S., and Weinstein, R. S. (1969).Science 165,597. McNutt, N. S., and Weinstein, R. S. (1970).1. Cell Biol. 47,666. McNutt, N. S., and Weinstein, R. S. (1973). Progr. Biophys. Mol. Biol. 26, 45. McNutt, N. S., Hershberg, R. A., and Weinstein, R. S. (1971).J . Cell Biol. 51, 805.

INTERCELLULAR JUNCTIONS IN CANCER

85

McNutt, N. S., Culp, L. A., and Black, P. H. (1973). 1. Cell Biol. 56,412. Malick, L. E. ( 1972). J. Nut. Cancer Inst. 49, 1039. Mao, P., Nakao, K., and Angrist, A. (1966). Cancer Res. 26,955. Marchesi, V. T., and Steers, E., Jr. (1968). Science 159, 203. Marchesi, V. T., Tillack, T. W., and Jackson, R. L. (1972). Proc. Nut. Acad. Sci. U S . 69, 1445. Marikovsky, Y., Brown, C. S., Weinstein, R. S., and Wortis, H. H. (1976). Exp. Cell Res. In press. Marshall, R. B., Roberts, D. K., and Turner, R. A. (1967). Cancer 20, 512. Martinez-Palomo, A. (1970a). In Vitro 6, 15. Martinez-Palomo, A. (1970b). Lab. Invest. 22,605. Martinez-Palomo, A. (1971). Pathobiol. Annu. I, 261. Martinez-Palomo, A., and Erlij, D. (1975). Proc. Nut. Acad. Sci., U S . 72, 4487. Martz, E., and Steinberg, M. S. (1972). J. Cell Physiol. 79, 180. Merk, F. B., and McNutt, N. S. ( 1972). J. Cell Biol. 55,511. Merk, F. B., Botticelli, C. R., and Albright, J. T. ( 1972). Endocrinology 90, 992. Merk, F. B., Albright, J. T., and Botticelli, C. R. (1973). Anut. Rec. 175, 107. Merkow, L. P., Frich, J. C., Slifkin, M., Kyreages, G., and Pardo, M. (1971). Cancer 28, 372. Merrick, T. A., Erlandson, R. A., and Hajdu, S. I. ( 1971). Arch. Pathol. 91,365. Michalke, W., and Loewenstein, W. R. (1971). Nature (London) 232, 121. Mincer, H. H., and McGinnis, J. P. (1972). Cancer 30, 1036. Ming, S. C., Coldman, H., and Freiman, D. G. (1967). Cancer 20, 1418. Mishima, Y. ( 1967). Cancer 20,632. Moor, H. (1966). Int. Reu. Exp. Pathol. 5, 179. Moretz, R. C., Akers, C. K., and Parsons, D. F. (1969). Biochim. Biophys. Actu 193, 1. Muir, A. R. (1967). J . Anat. 101,239. Mukai, N., Kobayashi, S., and Oguri, M. (1974). Acta Neuropathol. 28, 293. Murad, T. M., Mancini, R., and George, J. (1973). Cancer 31, 1440. Nachbar, M. S., Oppenheim, J. D., and A d , F. (1974). Amer. J. Med. Sci. 268, 122. Neville, D. M., Jr. (1960). J. Biophys. Biochem. Cytol. 8,413. Nicolson, G. L. (1973). J. Supramol. Struct. 1,410. OLague, P., and Dalen, H. ( 1974). Exp. Cell Res. 86,374. O’Lague, P., Dalen, H., Rubin, H., and Tobias, C. (1970). Science 170, 464. Oldfield, F. E. ( 1963). Exp. Cell Res. 30, 125. Oliveira-Castro, G. M., and Loewenstein, W. R. (1971). J . Membrane B i d . 5, 51. Orr, C. W., and Roseman, S. ( 1969). J . Membrane Biol. 1, 109. Otten, J., Johnson, G. S., and Pastan, I. (1971). Biochem. Biophys. Res. Commun. 44, 1192. Overton, J. (1962). Deuelop. Biol. 4,532. Overton, J. (1968). J. Exp. Zool. 168, 203. Overton, J. (1974). Progr. Surface Membrane Sci. 8, 161. Overton, J., and Kapmarski, R. (1975). J. Exp. 2001.192, 33. Pappas, G. D., Asada, Y.,and Bennett, M. V. L. ( 1971). J. Cell Biol. 49, 173. Pardee, A. B. (1964). Nut. Cancer Inst., Monogr. 14, 7. Pardee, A. B., Jimenez de Asua, L., and Rozengurt, E. (1974). In “Control of Proliferation in Animal Cells” (B. Clarkson and R. Baserga, eds. ), p. 547. Cold Spring Harbor Lab., Cold Spring Harbor, New York.

86

RONALD S . WEINSTEXN, FREDERICK B. MERK, AND JOSEPH ALROY

Payton, B. W., Bennett, M. V. L., and Pappas, G. D. (1969). Science 166,1641. Peracchia, C. (1973a). J. Cell Biol. 57, 54. Peracchia, C. ( 197313). J . Cell Biol. 57, 66. Perk, K., Hod, I., and Nobel, T. A. (1971). 1. Not. Cancer Inst. 46, 525. Peters, R., Peters, J., Tews, K. H., and Bahr, W. (1974). Biochim. Biophys. Acta 367, 282. Pierce, G. B., Jr. (1966). Cancer 19, 1963. Pierce, G. B., Jr., and Wallace, C. ( 1971). Cancer Res. 31, 127. Pierce, G. B., Jr., Stevens, L. C., and Nakane, P. K. (1967). J. Nut. Cancer Inst. 39, 755. Pietra, G. G., Szidon, J. P., Leventhal, M. M., and Fishman, A. P. (1969). Science 166, 1643. Pinto da Silva, P., and Gilula, N. B. (1972). Exp. CeU Res. 71,393. Pitelka, D. R., Hamamoto, S. T., Duafala, J. G., and Nemanic, M. K. (1973). J. Cell Biol. 56, 797. Pitts, J. D. ( 1971). Growth Contr. Cell Cult., Ciba Found. Symp., 1970 p. 89. Politoff, A. L., Socolar, S. J., and Loewenstein, W. R. (1969). J. Gen. Physiol. 53, 498. Pollack, R. E., and Hough, P. V. C. ( 1974). Annu. Reu. Med. 25, 431. Popoff, N. A., Malinin, T. I., and Rosomoff, H. L. (1974). Cancer 34, 1187. Porter, K. R., Fonte, V., and Weiss, G. ( 1974). Cancer Res. 34, 1385. Potter, D. D., Furshpan, E. J., and Lennox, E. S. (1966). Proc. Nut. Acad. Sci. U S . 55, 328. Prutkin, L. (1975). Cancer Res. 35, 364. Pulley, L. T. (1973).Amer. J. Vet.Res. 34, 1513. Rambourg, A. ( 1969). 1. Microsc. (Paris) 8,325. Ramsey, H. J. (1965). Cancer 18, 1014. Rangan, S. R. S. ( 1972). Cancer 29, 117. Rapin, A. M. C., and Burger, M. M. ( 1974). Aduan. Cancer Res. 20, 1. Rasmussen, H. (1970). Science 170,404. Raviola, E., and Gilula, N. B. (1973). Proc. Not. Acad. Sci. U.S. 70, 1677. Reddy, J., Svoboda, D., Azamoff, D., and Dawar, R. (1973). J. Nut. Cancer Inst. 51, 891. Reese, T. S., Bennett, M. V. L., and Feder, N. ( 1971). Anat. Rec. 169, 409. Revel, J. P., and Karnovsky, M. J. ( 1967). J. Cell Biol. 33, C7. Revel, J. P., Yee, A. G., and Hudspeht, A. J. (1971). Proc. Nut. Acad. Sci. US. 68, 2924. Revel, J. P., Yip, P., and Chang, L. L. ( 1973). Deuelop. Biol35,302. Rhodes, R. S., and Karnovsky, M. J. ( 1971). Lab. Inoest. 25, 220. Richart, R. M., and Barron, B. A. ( 1969), Amer. J. Obstet. Gynecol. 105,386. Richter, W. R., and Moize, S. M. (1963). J. Ultrastruct. Res. 9, 1. Robertson, J. D. (1959). Biochem. SOC. Symp. 16,3. Robertson, J. D. ( 1963). J . Cell Biol. 19,201. Robertson, J. D. ( 1961). In “Cellular Membranes in Development” ( M . Locke, ed.), p. 1. Academic Press, New York. Robison, G. A., Butcher, R. W., and Sutherland, E. W. (1971). “Cyclic AMP.” Academic Press, New York. Rorat, E., Ferenczy, A., and Richart, R. M. (1974). Cancer 33, 880. Rosai, J. (1968). Cancer 22, 333. Rosai, J., Khodadoust, K., and Saber, I. (1969). Cancer 24, 103.

INTERCELLULAR JUNCXIONS IN CANCER

87

Rose, B. (1970). Science 169,607. Rose, B. (1971). J. Membrane Biol. 5, 1. Rose, B., and Loewenstein, W. R. (1971). J. Membrane Biol. 5,20. Rose, B., and Loewenstein, W. R. (1975a). Nature 254,250. Rose, B., and Loewenstein, W. R. (1975b). Science 190, 1204. Roth, L. M., Spurlock, B. O., Sternberg, W. H., and Rice, B. F. (1970). Amer. 1. Pathol. 60, 137. Roth, S. (1973). Quart. Reo. B i d . 48, 541. Roth, S., and Weston, J. A. (1967). Proc. Nut. Acad. Sci. US. 58, 974. Roth, S., McGuire, E. J., and Roseman, S. ( 1971). J. Cell Biol. 51, 525. Rubin, R. P., Carchman, R. A., and Jaanus, S. D. (1972). Nature (London), New Biol. 240, 150. Rubinstein, L. J., Herman, M. M., and Hanbery, J. W. (1974). Cancer 33, 675. Salazar, H., Merkow, L. P., Walter, W. S., and Pardo, M. (1974). Obstet. Gynecol. 44, 551. Sanel, F. T., and Serpick, A. A. (1970). Science 168, 1458. Satir, P., and Gilula, N. B. ( 1973). Annu. Reu. Entoml. 18, 143. Sedar, A. W., and Forte, J. G. (1964). J. Cell Biol. 22, 173. Seifert, W., and Paul, D. (1972). Nature (London),New Biol. 240, 281. Sellin, D., Wallach, D. F. H., and Fischer, H. (1971). Eur. 1. Immunol. 1, 453. Sellin, D., Wallach, D. F. H., Weltzien, H. U., Resch, K., Sprenger, E., and Fischer, H. (1974). Eur. J. Immunol. 4, 189. Sharrna, R. K., and Hashimoto, K. ( 1972). Cancer Res. 32,666. Sheppard, J. R. ( 1971). Proc. Nut. Acad. Sci. U.S. 68, 1316. Sheridan, J. D. (1970). J. Cell Biol. 45,91. Sheridan, J. D. ( 1971). J. Cell Biol. 50, 795. Shin, M. L., and Firminger, H. I. ( 1973). Amer. J. Pathol. 70,291. Silverberg, S . G., and DeGiorgi, L. S. (1972). Cancer 29, 1680. Simani, A. S., Inoue, S., and Hogg, J. C. (1974). Lab. Invest. 31,75. Simionescu, M., Simionescu, N., and Palade, G. E. (1975). J . Cell Biol. 67, 863. Singer, S. J. (1962). Aduan. Protein Chem. 17, 1. Singer, S. J. (1971). In “Structure and Function of Biological Membranes” (L.I. RotMeld, ed.), p. 145. Academic Press, New York. Singer, S. J., and Nicolson, G. L. (1972). Science 175,720. Skerrow, C . J., and Matoltsy, A. G. ( 1974a).J. Cell Biol. 63,515. Skerrow, C. J., and Matoltsy, A. G. (1974b). J. Cell Biol. 63,524. Slack, C., and Palmer, J. F. ( 1969). Exp. Cell Res. 55,416. Smets, L. A. ( 1972). Nature (London), New Biol. 239, 123. Socolar, S. J., and Politoff, A. L. ( 1971).Science 172,492. Staehelin, L. A. (1973). J . Cell Sci. 13, 763. Staehelin, L. A. (1974). Int. Reu. Cytol. 39, 191. Staehelin, L. A. (1975). J. Cell S c i . 18,545. Stanton, M. F., Ting, R. C., and Miller, E. (1970). J. Nut. Cancer Inst. 45, 195. Steck, T. L. ( 1974). J. Cell Biol. 62, 1. Steck, T. L., Fairbanks, G., and Wallach, D. F. H. (1971). Biochemistry 10,2617. Steinberg, M. S. (1963). Science 137, 762. Steinberg;M. S. (1970). J. Ezp. ZOO^. 173, 395. Steinberg, M. S. (1973). Locomotion Tissue Cells, Ciba Found. Symp. p. 33. Steiner, G. C., and Dodman, H. D. (1972). Cancer 29, 122. Steiner, G. C., Ghosh, L., and Dorfman, H. D. (1972). Hum. Pathol. 3, 569.

88

ROSALD

s. WEINSTEIN,

FREDERICK B. MERK, AND JOSEPH ALROY

Steiner, G. C . , hfirra, J. M., and Bullough, P. G. (1973). Cancer 32, 926. Stoeckenius, W., and Engelman, D. M. (1969). J . Cell Bwl. 42,613. Stoker, hf. (1967). Curr. Top. Develop. BioZ. 2, 107. Subak-Sharpe, H. (1969). Homeostatic Regul., Ciba Found. S y m p . Subak-Sharpe H., Biirk, R. R., and Pitts J. D. (1969). J. Cell Sci. 4, 353. Sutherland, E. W. (1970). J. Amer. Med. Ass. 214, 1218. Svoboda, D. J., Kirchner, F. R., and Proud, G. 0. (1963). Cancer Res. 23, 1084. Tani, E., Nishiura, M., and Higashi, N. (1973). A d a Neuroputhol. 26, 127. Tani, E., Ikeda, K., Kudo, S., Yamagata, S., Nishiura, M., and Higashi, N. (1974). Acta Neuropathol. 27, 139. Tannenbaum, M. ( 1971). Pathobiol. Annu. p. 249. Tarin, D. (1970). J. Invest. Dermutol. 55, 26. Taxy, J. B., Battifora, H., and Oyasu. R. (1974). Cancer 34, 306. Tee], R. W., and Hall, R. G. ( 1973). Exp. Cell Res. 76,390. Temin, H. hl., and Rubin, H. (1958). Virology 6, 66a. Tice, L. W., Wollman, S. H., and Cartier, R. C. ( 1975). J . Cell Biol. 66, 657. Tobon, K , and Price, H. M. ( 1972). Cancer 30, 1082. Toker, C. (1968). Cancer 21,1171. Tonietti, G . , Baschieri, L., and Salabe, G. (1967). Arch. PuthoE. 84, 601. Trelstad, R. L., Revel, J. P., and Hay, E. D. (1966). J. Cell Biol. 31, C6. Trelstad, R. L., Hay, E. D., and Revel, J. P. (1967). Develop. Biol. 16, 78. Tripathi, R. C., and Gamer, A. ( 1972). Cancer Res. 32,90. Ts’o, M. 0. M., Fine, B. S., and Zimmerman, L. E. ( 1969). Arch. Pathol. 88, 664. Turnbull, R. B., Kyle, K., Watson, F. R., and Spratt, J. (1967). Ann. Surg. 166, 420. Tyzzer, E. E. (1913). J. Med. Res. 28, 1913. Urban, J., Kartenbeck, J., Zimber, P., Timko, J., Lesch, R., and Schreiber, G. (1972). Cancer Res. 32, 1971. Vasiliev, J. XI., Gelfand, I. M., Domnina, L. V., and Rappoport, R. I. (1969). Exp. Cell Res. 54, 83. Vogel, A., and Narasimnan, S. ( 1974). Dermutologica 148,201. von Bomhard, D., and Sandersleben, von J. (1973). Virchows Arch. A. 359, 87. Wade, 1. B., and Karnovsky, M. J. ( 1974). 3. Cell Biol.62, 344. Wade, J. B., Revel, J. P., and DiScala, V. A. ( 1973). Amer. J . Physiol. 224,407. Walther, B. T., Oehman, R., and Roseman, S. (1973). Proc. Nut. Acad. Sci. US. 70, 1569. Wang, X. S., Seeniayer, T. A., Ahmed, M. N., and Morin, J. (1974). Arch. Pathol. 98, 100. Wehrli, E., Muhlethaler, K., and Moor, H. (1970). Exp. Cell Res. 59, 336. Weinstein, R. S. ( 1974). In “The Red Blood Cell” ( D. M. Surgenor, ed.)., 2nd ed., Vol. 1, p. 213. Academic Press, New York. Weinstein, R. S., and McNutt, N. S. (197Oa). In “Microcirculation, Pedusion and Transplantation of Organs” (T. J. Malinin, B. S. Linn, A. B. Callahan, and W. D. Warren, eds.) p. 23. Academic Press, New York. Weinstein, R. S., and McNutt, N. S. (1970b). Semin. Hematol. 7,259. Weinstein, R. S., and McNutt, N. S. (1972). N . Engl. J. Med. 286, 521. Weinstein, R. S., Clowes, A. W., and McNutt, N. S. (1970a). P r w . SOC. E x p . Biol. Med. 134, 1195. Weinstein, R. S., McNutt, N. S., Nielsen, S. L., and Pinn, V. W. (1970b). Proc. 28th Meet. Electron Microsc. SOC. Amer. p. 108.

INTERCELLULAR JUNCI’IONS IN CANCER

89

Weinstein, R. S., Zel, G., and Merk, F. B. (1974). In “Membrane Transformation in Neoplasia” (J. Schultz and R. E. Block, eds.), p. 127. Academic Press, New York. Wellings, S. R., and Roberts, P. (1963). J. Nut. Cancer Inst. 30,269. Welsh, R. A., and Meyer, A. T. (1968). Arch. Pathol. 85, 433. Welsh, R. A., Bray, D. M., 111, Shipkey, F. H., and Meyer, A. T. (1972). Cancer 29, 191. Welti, C. V., Pardo, V., Millard, M., and Gerston, K. (1972). Cancer 29, 1169. Whittemburg, G., and Rawlins, F. A. ( 1971). Pjluegers Arch. Gesamte Physiol. Menschen Tiere 330, 302. Wiernik, G., Bradbury, S., Plaut, M., Cowdell, H., and Williams, E. A. (1973). Brit. J. Cancer 28, 488. Willingham, M. C., Johnson, G. S., and Pastan, I. (1972). Biochem. Biophys. Res. Commun. 48,743. Wills, E . J. (1968). Cancer 22, 1046. Winzler, R. J. (1970). Int. Reu. Cytol. 29, 77. Wiseman, L. L., and Steinberg, M. S. (1973). Erp. Cell Res. 79,468. Wolff, M., Santiago, H., and Duby, M. M. (1972). Cancer 30, 1046. Wood, R. L. (1959). J . Biophys. Biochem. Cytol. 67, 213. Woodbury, J. W., and Crill, W. E. (1961). In “Nervous Inhibition” ( E . Florey, ed.), p. 124. Pergamon, Oxford. Woodruff, R. I., and Telfer, W. H. (1973). J. Cell B i d . 58, 172. Yeh, S. ( 1973). Hum. Pathd. 4,469. Yeh, S., Chen, H. C., How, S. W., and Deng, C. S. (1974). J. Nat. Cancer Inst. 53, 31. Zamboni, L., and Gondos, B. (1968). J. Cell Biol. 36,276. Zampighi, G., and Robertson, J. D. (1973). J. Cell Biol. 56,92. Zuckerberg, C. ( 1973). Cancer Res. 33, 2278.

This Page Intentionally Left Blank

GENETICS OF ADENOVIRUSES

Harold S. Ginsberg and C. S.

H. Young

Department of Microbiology, Collegd of Physicians and Surgeons, Columbia University, New York, New York

I. Introduction . . . . . . . . A. The Virion . . . . . . . . B. Viral Replication . . . . . . . 11. Isolation of Adenovirus Mutants . . . . A. Types of Mutants . . . . . . B. Mutagenic Procedures . . . . . . C. Selection . . . . . . . . D. The Problem of Multiple Mutation . . . 111. Genetic Characterization . . . . . . A. Complementation . . . . . . . B. Recombination Tests . . . . . . C. Correlation between Genetic and Physical Maps D. Heterotypic Complementation . . . . IV. Phenotypes of Adenovirus Mutants A. Characterization . . . . . . . B. Functional Studies Using Adenovirus Mutants . . . . . . . . V. Summing Up References

.

.

.

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. .

. .

. .

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

. . . . . . . . . . .

. . .

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

. 9 1 92 95 .lo1 .lo1 .lo3 .lo4 .lo5 .lo5 .lo5 .lo9 . 113 . 115 . 116 .116 . 118 .123 .126

. .

1. Introduction

Adenoviruses infect many species of warm-blooded animals and effect varied clinical reactions including acute disease, latent infection, and the induction of malignancy. Similarly, in cell culture, the response to infection is varied, ranging from extensive cytopathic effects to cellular transformation. The establishment of a nuclear infection triggers this array of host responses, and offers an unusual potential to investigate the regulation of replication and transcription of DNA and the translation of mRNAs in eukaryotic cells. Studies with bacteriophage, however, vividly exposed the problem that biochemical techniques alone were inadequate to obtain detailed data on the controls governing many biosynthetic reactions. Selected conditionally lethal and deletion mutants were essential for the studies that yielded exquisite evidence on the mechanisms regulating bacteriophage replication. In recognition of the critical requirement of appropriate mutants, several laboratories ( see Table I ) have selected conditionally lethal temperature-sensitive mutants of adenoviruses, predominantly types 5 and 12, to study the genetic 91

92

HAROLD S. GINSBERG AND C. S. H. YOUNG

mechanisms of a DNA-containing virus and to investigate the reactions that control adenovirus replication. Since mammalian cells are so complex and contain of the order of 10‘ genes, a direct investigation of the molecular events regulating biosynthetic reactions and cell division is fundamentallv impossible at this time. Studies with a smaller genome, which contains a maximum of 50 genes and replicates in the nucleus of a eukaryotic cell, offers a simpler model that should yield evidence germane to the molecular biology of mammalian cells as well as to viral biosynthesis. The objectives of this article are to review the studies that described the mutagenesis and selection of adenovirus mutants, their use in formal genetic studies, their phenotypes, and their utility in experiments for understanding the regulation of adenovirus replication and cellular transformation. A number of reviews have described different aspects of adenovirus structure and replication ( Ginsberg, 1969; Schlesinger, 1969; Green, 1970; Philipson and Lindberg, 1974; Norrby, 1968; Philipson and Petterson, 1973), but to familiarize the nonspecialist with the data pertinent to this review, a brief summary of the essential evidence will be presented. A. THE VIRION The viral particle is an icosahedron with a mean diameter of 70-80 nm. Contrary to the initial prediction of Crick and Watson (1956), the isometric virus is not composed of identical repeating subunits, but rather the capsid is made of three unique major multimeric proteins and several minor proteins ( Ginsberg, 1969; Schlesinger, 1969; Green, 1970; Philipson and Lindberg, 1974; Norrby, 1968; Philipson and Pettersson, 1973; Maize1 et al., 1968a,b). It has been possible to describe the architecture of the virion and the topography of its components owing to the early demonstration that the virion can be dissociated into intact native subunits, that the capsid components are identical with the great excess ( approximately 10-fold) of unassembled “soluble antigens” present in infected cell extracts, and that the capsid proteins have the unusual feature of being soluble in aqueous media while remaining in their native functional forms (Wilcox, et d.,1963; Valentine and Pereira, 1965). The capsid contains 252 capsomers (Fig. 1 ) of which the faces and edges of the 20 equilateral triangles are comprised of 240 hexons, so termed because each has six neighbors (Ginsberg et d.,1966). At each of the 12 vertices of the icosahedron, the axis of 5-fold symmetry, is a penton (Valentine and Pereira, 1965), which consists of a base, the vertex capsomer, and a fiber. Three minor proteins of less than

93

GENETICS OF ADENOVIRUSES

MOLECULAR WEIGHT

HEXONASSOCI4TID PROTEIN

1 PENTON(

Fiber

H

PI

2S.SK

,\

FIG.1. Diagrammatic representation of a type 5 adenovirus particle partially disrupted to demonstrate the viral capsid and nucleoprotein core. A drawing of a representative pattern of the denatured virion proteins electrophoresed in a sodium dodecyl sulfate-polyacrylamide gel shows the relative sizes of the polypeptide chains. The nomenclature follows that of Maizel et al. (1968a) and Everitt et d. (1973). The molecular weights of the major type 5 virion polypeptide chains are presented, since most are significantly different from the values given for the comparable proteins of type 2 adenovirus (R. Kauffman, unpublished data).

25,000 daltons are reported to be associated with the hexons that make up the faces of the triangles, and one minor protein is said to be structurally related to the hexons surrounding the pentons, the so-called peripentonal hexons (Maizel et al., 1968a,b; Everitt et al., 1973). The minor proteins have been isolated in relatively constant amounts and may serve to assemble and stabilize the capsid, but until chemical evidence is presented showing that each is indeed a unique protein, the possibility must also be entertained that some are degradation products of one or more of the major virion proteins. Internally, there are two core proteins, closely associated with the viral DNA (Laver, 1970; Prage and Pettersson, 1971; Russell et aZ., 1971). The hexon, which induces the production of type-specific neutralizing antibodies ( Wilcox and Ginsberg, 1963a; Kjellkn and Pereira, 1968; Kasel et aZ., 1964) as well as broad cross-reacting antibodies which immunologically identify the adenoviruses as a group (Wilcox and Ginsberg, 1961; Pereira, 1960), is composed of three identical polypeptides of 100,000-120,000 daltons each, depending upon the type (Maizel et al., 1968a,b; Franklin et al., 1971; Cornick et al., 1973; Stinski and Ginsberg, 1975). The fiber varies morphologically in length in different types (Wil-

94

HAROLD S. GINSBERG AND C . S. H. YOUNG

cox et al., 1963; Valentine and Pereira, 1965; Norrby, 1968, 1969) and therefore probably in molecular weight. The types 2 and 5 fibers have molecular weights of 183,000 (Dorsett and Ginsberg, 1975) to 200,000 (Sundquist et al., 1973b) and consist of three polypeptide chains of 60,000 to 65,000 daltons each (Dorsett and Ginsberg, 1975; Sundquist d al., 1973b). The fiber, which is glycosylated (Ishibashi and Maizel, 1974b) and phosphorylated (Russell et al., 1972b), is either composed of two polypeptide chains which are chemically identical and one which is unique or three chemically different chains (Dorsett and Ginsberg, 1975). The penton base is relatively unstable and highly sensitive to proteolytic enzymes ( Pereira, 1958; Everett and Ginsberg, 1958), and it therefore has not yet been well characterized except to note that it consists of several polypeptide chains ( 4 or 5) of around 70,000-80,000 daltons (Maizel et al., 1968a,b). The fiber is associated with the base through noncovalent bonds, which are disrupted by 2.5 M guanidine HCl (Norrby and Skaaret, 1967), 8%pyridine (Pettersson and Hoglund, 1969), or 33%formamide ( Neurath et al., 1968). The viral genome is a linear, double-stranded DNA molecule of 20 to 25 X lo6 daltons (Green et al., 1967; van der Eb et al.,1969), depending upon the immunological type (there are 31 human, 23 simian, 8 avian, 6 bovine, 4 porcine, 2 canine, and 1 murine types reported). It is striking that the viral DNA is not terminally redundant nor circularly permuted, but, like the adeno-associated viral DNA (Koczot et al., 1973), each strand of the molecule contains an inverted terminal repetition (Garon et al., 1972; Wolfson and Dressler, 1972). Hence, when the viral DNA is denatured at a limited concentration, both strands can form single-stranded circles through hydrogen bonds between the complementary ends (Garon et al., 1972; Wolfson and Dressler, 1972). However, the function of this novel DNA structure is unclear. It has also been reported that a small protein molecule associated with the viral DNA maintains the genome in a circular form in the virion and when artifically released from it (Robinson et al., 1973). A similar circular genome has not been detected, however, after intracellular uncoating of the virion (Robinson et al., 1973) or during viral DNA replication (Sussenbach et al., 1972; Pettersson, 1973). Just as adenoviruses have been divided into subgroups according to biological properties, such as oncogenicity or hemagglutination ( Huebner et al., 1965; Rosen, 1960), the viral DNAs of different types may be similarly classified according to base composition (Piiia and Green, 1965) and DNA-DNA hybridization (Green, 1970; Garon et al., 1973). Thus adenoviruses have been arranged into three subgroups (A, B, and C ) according to their oncogenic potential in newborn hamsters (Huebner

95

GENETICS OF ADENOVIRUSES

et al., 1965). The viruses fall into similar subgroups when divided according to G C content and nucleotide sequence homology. The G C content of subgroup A, which consists of the most oncogenic adenoviruses (types 12, 18, and 31) is 48-49%; subgroup B, which contains weakly oncogenic viruses (e.g., types 3, 7, and 21) has a G C content of 49-52%;and subgroup C, which consists of viruses that are essentially nononcogenic (e.g., types 1, 2, and 5 ) has a G C content of'55-602 (Green, 1970; Piiia and Green, 1965). It is also striking that, between members of a subgroup, the nucleotide sequences are up to 95%homologous, whereas DNA-DNA hybridizations between selected members of different subgroups show only about 10%homology (Green, 1970; Garon et al., 1973 ) . Electron microscopic examination of heteroduplexes formed between DN As from different subgroup viruses reveal considerable heterology (Garon et al., 1973) while DNAs from viruses belonging to the same subgroup (types 1, 2, and 5 ) show almost complete homology (Garon et al., 1973; Bartok et al., 1974) except for two discrete regions of nonhomology representing a total of 16% of the genome (Bartok et al., 1974).

+

+

+

+

B. VIRALREPLICATION Production of infectious virions follows from an ordered series of reactions which is initiated by association of a viral particle with a susceptible cell and culminated by the assembly of virions from its subunits. Replication of types 2 and 5 adenoviruses has been most extensively studied because the yield is great (about lo4 PFU/cell) and the virion, as well as most of the components, can be easily purified and analyzed. Accordingly, the molecular biology of adenovirus infection is known in greatest detail with these types, and the description of adenovirus replication that follows will summarize the investigations with types 2 and 5 viruses. A single cycle of type 5 adenovirus multiplication with the accompanying biosynthetic events are diagrammatically summarized in Fig. 2. The temporal characteristics and viral yield vary, however, with different types of adenoviruses: for example, the eclipse period for type 12 adenovirus at 36OC is 16-18 hours (Gilead and Ginsberg, 1965), and biosynthesis of the viral DNA cannot be detected until 1 2 1 5 hours (Piiia and Green, 1969) after infection, whereas DNA replication of types 2 and 5 begins 6-7 hours after infection and their eclipse periods terminate 13-14 hours after infection (Pifia and Green, 1969; Ginsberg et al., 1967). The fiber is the virion's organ of attachment to receptor sites of susceptible cells (Levine and Ginsberg, 1967; Philipson et al., 1 W ) . The nature of the specific receptors, however, has not yet been identified. Attached

96

HAROLD S. GINSBERG AND C. S. H. YOUNG

lo’[ HOST PROTEINS

2 10 5

...-

- -- -

LATE PROTEINS L A T E / mRNAs V I R A L ~ D N A .__. .. .. -EARLY /PROTEINS.. . .-CLASS11 EARLY/m RNAs I EARLY/?RNAS CLASS .. -

___

J

,

___ I

virions enter susceptible cells by phagocytosis ( Dales, 1962; Chardonnet and Dales, 1970) or direct penetration of the plasma membrane (Morgan et cd., 1969). Whether either entry mechanism is preferred is still unclear, but even after phagocytosis the virions must directly penetrate the membrane of the phagocytic vacuole to gain access to the cytoplasm. Once within the cytoplasm, viral uncoating (i.e., viral eclipse) is rapidly initiated with disengagement of the pentons (Sussenbach, 1967), which is soon followed by dissociation of the capsid and release of the nucleoprotein core ( Sussenbach, 1967; Lawrence and Ginsberg, 1967; LonbergHolm and Philipson, 1969). Neither synthesis of proteins nor replication of nucleic acids is required for these initial uncoating reactions (Lawrence and Ginsberg, 1967; Lonberg-Holm and Philipson, 1969), which uncover the viral DNA so that it is susceptible to DNase although still associated with about 20% of the viral proteins, the core proteins (Lawrence and Ginsberg, 1967; Lonberg-Holm and Philipson, 1969). The partially uncoated virus appears to be transported in microtubules to nuclear pores (Chardonnet and Dales, 1970), where further dissociation of DNA from the internal proteins occurs at the nuclear membrane (Morgan et al., 1969) or in the nucleus (Lonberg-Holm and Philipson, 1969). But the extent and the mechanisms of the terminal uncoating reaction have not been determined. The description of viral penetration and uncoating presented carries with it a caveat. The ratio of total viral particles to infectious virions (i.e., plaque formers) varies from 10 to several hundred, according to the viral species (type), and all the experiments described employed

GENETICS OF ADENOVlRUSES

97

high multiplicities of infection. Consequently, the sequence of events that was observed, which was followed by biochemical analysis of isotope-labeled virus and electron microscopic observations, involved the large population of viral particles in each cell, but the virion or virions that established the infectious process may not necessarily have followed the same sequence.

1. Transcription Shortly after the viral DNA reaches the nucleus, 2-3 hours after infection, transcription begins (Thomas and Green, 1969; Lucas and Ginsberg, 1971; Parsons and Green, 1971). Since protein synthesis is not required (Parsons and Green, 1971) and the virion does not contain its own RNA polymerase, transcription apparently involves a host enzyme. This has been shown to be a-amanitin sensitive and therefore probably corresponds to the DNA-dependent RNA polymerase I1 (Ledinko, 1971; Price and Penman, 1972). Prior to viral DNA replication only a portion of the genome is transcribed, yielding early mRNAs (Thomas and Green, 1969; Parsons and Green, 1971; Lucas and Ginsberg, 1971; Tibbetts et al., 1974). However, the extent of the genome copied into early transcripts is still uncertain, with reports varying from 1&20% (Thomas and Green, 1969; Lucas and Ginsberg, 1971) to 4050%(Tibbetts et al., 1974). It is particularly important to note that the genes for early transcripts are not limited to a single region but are distributed throughout the genome (Craig et al., 1975). After the initiation of DNA replication, ‘‘late’’ transcripts appear (Bello and Ginsberg, 1969; Green et al., 1970; Lucas and Ginsberg, 1971 ) , and apparently the transcription of about 50% of the early mRNAs, the so-called Class I early mRNAs (Lucas and Ginsberg, 1971; Craig et al., 1975) ceases, although there is also not complete agreement on these findings (Thomas and Green, 1969; Tibbetts et al., 1974). The reactions that regulate the limited transcription of the parental viral genome, the onset of late transcription, and the switching of transcription from one strand to the other during late transcription (Tibbetts et al., 1974; Fujinaga and Green, 1970; Tibbetts and Pettersson, 1974) are still unknown. Indeed, there are several reactions by which the original transcripts are modified to produce functional mRNAs, and their mechanisms and control are also uncertain: ( 1 ) the primary transcripts are considerably larger than the polysomal mRNAs (McGuire et al., 1972; Wall et al., 1972), which are aIso heterogeneous in size, varying from 9 S to 27 S (Parsons and Green, 1971; Lindberg et al., 1972; Philipson et al., 1973; Craig et al., 1975); ( 2 ) only 70-804: of the RNA transcribed is processed and transported into the cytoplasm (McGuire et al., 1972; Lucas and Ginsberg, 1972); and ( 3 ) most if

98

HAROLD S. GINSBERG AND C. S. H. YOUNG

not all viral mRNAs are polyadenylated in the nucleus after transcription (Philipson et al., 1971). It is noteworthy that the heterogeneity and processing of adenovirus-specific RNAs resemble the characteristics of the uninfected host cell’s heterogeneous ( H n ) RNAs, and hence investigation of transcription of adenovirus mRNAs offers another convenient probe for study of the molecular biology of eukaryotic cells. 2. DNA Replication In productive infections, under conditions of multiple infection of each cell, replication of viral D N A begins about 6 hours after infection, reaching a maximum about 18 hours after viral attachment (Pifia and Green, 1969; Ginsberg et al., 1967). Because viral DNA has a higher G C content than that of the host, the two species can be effectively separated (Piiia and Green, 1969; Ginsberg et al., 1967). Accordingly, it is noted that synthesis of host DNA begins to decline about 6 hours after infection, when replication of viral DNA commences, and production of host DNA is effectively blocked by 12 hours (Ginsberg et al., 1967). But prior synthesis of either host or viral DNA is not essential for the inhibition of host DNA (Ginsberg et al., 1967). Like replication of most linear DNA molecules, little is known about the precise mechanism of duplicating adenovirus DNA. Although protein synthesis is required for the initiation of viral DNA replication (Wilcox and Ginsberg, 1963b), there is no evidence to indicate whether the polymerase is host or viral coded. Continued protein production, however, is not required after replication of viral DNA is established (Horwitz et al., 1973). Van der Vleit and Levine (1973) have identsed in type .j-infected cell extracts proteins of 72,000 and 48,000 daltons which bind preferentially to single-stranded DNA ( a so-called DNAbinding protein), similar to the T4 phage gene-32 protein ( Alberts and Frey, 1970). Peptide maps of the two suggest, however, that the 48,000-dalton polypeptide is a proteolytic breakdown product of the 72,000-dalton protein ( A. J. Levine, personal communication ) . Whether both proteins are functional is uncertain. Like T4 gene-32 protein, the DNA-binding protein in adenovirus-infected cells appears to be a viral gene product present in large numbers of copies per cells and essential for DNA replication (Alberts and Frey, 1970). Type 12 infected cells contain similar proteins of 60,000 and 48,000 daltons (Rosenwirth et QZ., 1975). Present in virus-infected cells are replicative forms of viral DNA, which sediment more rapidly and are denser than mature virion DNA (Belfett and Younghusband, 1972; van der Vleit and Sussenbach, 1972; van der

+

GENETICS OF ADENOVIRUSES

99

Eb, 1973; Pettersson, 1973). These replicative intermediates can also be isolated by the so-called “M-band” ( Sarkosyl) technique ( Pearson and Hanawalt, 1971; Shiroki et al., 1974; Yamashita and Green, 1974). In each instance, the replicative intermediate appears to consist of double-stranded DNA with single-stranded branches ( Bellett and Younghusband, 1972; van der Vleit and Sussenbach, 1972; van der Eb, 1973; Pettersson, 1973). Studying viral DNA replication in nuclei from infected cells, Sussenbach and colleagues have presented data which indicate that the DNA is replicated semiconservatively but asymmetrically, that the strand which has greater buoyant density in alkaline CsCl is initially replicated, displacing the complementary strand, and that the lighter strand is then copied in discontinuous segments that are subsequently joined (Sussenbach et al., 1972, 1973; Ellens et al., 1974). It should be stressed that although each viral DNA strand has the structure to generate a single-stranded circle (i.e., the strands have inverted terminal repetitious ends, as described above ) , such single-strand “circles” have not been observed in the replication complexes (Pettersson, 1973; Ellens et al., 1974).

3. Protein Synthesis Like host nuclear proteins, adenovirus proteins, whose transcripts are made in the nucleus, are produced in the cytoplasm and rapidly transported into the nucleus for assembly (Velicer and Ginsberg, 1968, 1970). Evidence cIearIy demonstrates that synthesis of early proteins precedes and is mandatory for replication of viral DNA (Wilcox and Ginsberg, 1963b; Horwitz et al., 1973). Thus far, only the T (Pope and Rowe, 1964; Gilead and Ginsberg, 1965) and P (Russell and Knight, 1967) antigens and the DNA-binding protein (van der Vleit and Levine, 1973; Rosenwirth et al., 1975) have been identified as probable early viral gene products. Translation of late proteins, which primarily consist of virion components, occurs on cytoplasmic polyribosomes of 180-200 S (Velicer and Ginsberg, 1968; Thomas and Green, 1966) and requires 1-2 minutes (Velicer and Ginsberg, 1970; White et al., 19sS). Kinetic data suggest that, after completion and release from polyribosomes, the nascent polypeptide chains are rapidly transported into the nucleus, where they are assembled into immunologically reactive multimeric capsid proteins ( Velicer and Ginsberg, 1970). However, cytoplasmic assembly of newly made protomers into capsid proteins can occur, as will be noted below with certain temperature-sensitive mutants (Kauffman and Ginsberg, 1975). Indeed, nascent polypeptides made in uitro can even self-assemble into capsid proteins in the cell-free reaction mixture ( Wilhelm and Ginsberg, 1972).

100

HAROLD S. GINSBERG AND C. S. H. YOUNG

Hexon protein, the major viral protein, is the first capsid protein to appear in the cell and is made in greatest amounts. All the capsid proteins, however, are produced in relative abundance since only about 10-20% of each of the proteins synthesized are assembled into virions ( Green, 1962; Wilcox and Ginsberg, 1 9 6 3 ~ ) .The excess viral proteins (often termed “soluble antigens”), along with similar amounts of excessive viral DNA, are the constituents of the characteristic intranuclear inclusion bodies seen in adenovirus-infected cells (Boyer et al., 1957, 1959; Morgan et al.,1960). 4. Assembly

It was generally accepted until several years ago that the capsid of isometric viruses self-assembled around previously formed viral nucleoprotein cores. Evidence has accumulated, however, showing that an RNA-deficient poliovirus procapsid may be initially assembled as the precursor of the infectious virion (Maizel et al., 1967; Phillips et aZ., 1968). With insertion of the viral RNA a procapsid protein is processed to form the stable capsid (Jacobson and Baltimore, 1968). Incomplete particles of SV40 have also been described as putative precursors of intact virions (Ozer, 1972; Ozer and Tegtmeyer, 1972). King and his collaborators studying the assembly of the SaZmonelZu typhimurium phage P22 reported convincing evidence that formation of a prohead is an initial phase of virion assembly. The prohead is a shell consisting of the capsid protein and a “scaffolding protein,” which exits from the precursor shell in concert with encapsulation of viral DNA (King and Casjens, 1974). The P22 proheads are similar to the T4 tau particles, which are precursors to T4 heads (Kellenberger et aZ., 1968; Simon, 1972). Data have been recently presented reporting that incomplete adenovirus particles (i.e., empty capsids) are also assembled as precursors into which viral DNA is inserted (Sundquist et al., 1973a; Ishibashi and Maizel, 1974a). With final assembly, two or three precursor viral proteins appear to be processed by cleavage and at least one disappears from the putative precursor particle (Sundquist et al., 1973a; Ishibashi and Maizel, 1974a; Anderson et al., 1973). Evidence obtained from the experiments summarized above, as well as from studies of abortive infections produced by elevated temperatures ( Okubo and Raskas, 1971), arginine deprivation ( Rouse and Schlesinger, 1972; Everitt et al., 1971), and temperature-sensitive mutants (Williams et al., 1974; Ensinger and Ginsberg, 1972), imply that assembly is controlled at several junctures during viral replication: (1) transport of nascent polypeptide chains or capsomers; ( 2 ) assembly of protomers

GENETICS OF ADENOVIRUSES

101

into capsomers; ( 3 ) formation of capsids; and (4)incorporation of viral nucleoprotein cores into virions.

II. Isolation of Adenovirus Mutants

A. TYPESOF MUTANTS In recent years, a number of investigators have applied to adenoviruses the techniques developed with bacteriophages in the hope of obtaining various kinds of mutants, with which to explore the interactions between virus and cell in both permissive and nonpermissive infections. The techniques and mutants obtained have also been employed to develop a genetic system. The next two sections will be concerned with the isolation and preliminary genetic characterization of adenovirus mutants. Theoretically, the types of mutants that may be sought include: 1. Plaque morphology mutants 2. Host range mutants-both those with an increased and those with a restricted host-cell range compared to that of the wild type 3. Conditional lethal mutants: At present this category is confined, in animal viruses, to temperature-restricted mutants, in particular, the so-called temperature-sensitive ( t s ) mutants that fail to replicate at high temperature while doing so at the lower. The other major class of conditional lethals, i.e., suppressor-sensitive mutants, which are suppressed by host cells containing UAG, UAA, or UGA suppressors, awaits the development of the appropriate cells and test systems 4. Drug-resistant mutants-those resistant to drugs that act specifically on viral replication 5. Virion structural mutants: These might include alterations to the stability or resistance of the virion to in uitro treatments, including heat or antiserum inactivation. Also, virions with altered buoyant density may be selected. Table I lists the publications dealing with the isolation of adenovirus mutants. It can be seen that all categories except ( 4 ) have been represented, but that ts mutants have been most actively searched for. The reasons for this are clear. Any gene that codes for a protein product is expected to be able to mutate to a t s form, if the temperature range over which the protein is active is mutable. Thus, t s mutants should represent lesions in a variety of functions essential to the replicative process. Furthermore, they are useful for several technical reasons: (1) the phenotype may be directly compared with wild type at both the restrictive and permissive temperatures and the functional lesion peculiar

T.4BLE I PIJHI.lChTION8 I)KSCRIBINO

Virus and mutant type H2 1s

H5 ts

H5

IS

Mutagenu

NN(; FIN02 NHzOH HNOz BrdUrd NHzOH HNO2 NNG NHzOH NNG BrdUrd NNG

H31 t s A1 Is H12 cyt Ad 2+ ND1 host range H 5 host range I15 heat stable

Killing or yield depression -10-4 -10-3 lo-' 10-5 10-2

gq 0.5 -

NHzOH Spontaneous

0.63 10-'-10- 3 10-*-10-3 10-2-10-3 10-~-10-3 5 x lo-' 10-1 10-1-10-4 8 x 10-4 -

TJV

10-'-10-~

uv

€I12 Is

ISOLATION O F

NH20H BrdUrd NNG HNO? NNG

uv

10-4

Spontaneous

-

ADKNOVIINJS SZUTANTS

Frequency of mutant (1s plaques/total) 7/172 5/170 14/146 8/95 2/355

References BBgin and Weber (1975) Williams el al. (1971) Ensinger and Ginsberg (1972)

0.01-0.10% 8/372 2/317 3/165 10/1440 45/700 2413370 2/260 171250

27/238d 5-&fold increase over spontaneous background 3/2-50 7/372 Only one mutant isolated

8

I"

Takahashi (1972) Rubinstein and Ginsberg (1974)* Lundholm and Doerfler (1971) Shiroki et al. (1972)

Ledinko (1974) Suzuki et al. (1972) Ishibashi. (1971) Takemori el al. (1968) Grodzicker el al. (1974s) Takahashi (1972) Young and Williams (197.5)

BrdUrd, bromodeoxyuridine ; NNG, N-methyl-N'-nitro-N-nitrosoguanidine;UV, ultraviolet irradiation. error, the wildtype virus used was type 5, rather than type 12 adenovirus as reported. I n this and the following table, the mutants isolated and characterized have been included as type 5 ts mutants. survivors. Range varied from 1/85 for -10-1 survivors to 7/144 for 2 X lo-' survivors to 3/94 for d ts mutants isolated from stocks which had been inactivated t.o give only from 0.5 to 3.0 % Is. a

* Owing to an unfortunate

E

8

E R

GENElTa OF ADENOVIRUSES

103

to the mutant may be identified; ( 2 ) the time and mode of action of the function may be identified by appropriate temperature-shift experiments; (3) they are eminently suitable for the performance of those genetic tests that depend on the ability to select a particular phenotype or phenotypic interaction.

PROCEDURES B. MUTAGENIC During serial passage of a virus, diverse spontaneous mutations will accumulate, giving rise to a stock of virus that is genetically heterogeneous, Thus most investigators have cloned their viral preparations by a number of plaque purifications before proceeding to isolate mutants from them, Since the starting material is thus relatively homogeneous and since there are frequently no selective methods to isolate particular classes of mutants (see below), a variety of mutagenic treatments have been used to enhance the frequency of mutants within the population (Table I). The mutagens have been of the kind that will predominantly induce “mis-sense” base-pair changes. This type of mutation is considered to be essential for the lesion that leads, for example, to a t s phenotype. The treatments fall into two experimental categories: ( 1 ) those used in uitro which are directed against the virion DNA (nitrous acid, hydroxylamine, UV irradiation) and ( 2 ) those used in the infected cell during viral replication ( nitrosoguanidine, bromodeoxyuridine) , I n uitm methods would be expected to induce lesions of the mismatched basepair type ( m / + ) , and it is perhaps surprising that many plaques isolated directly after such treatments have proved to be mutant (see, for example, Williams et al., 1971; Ledinko, 1974; BBgin and Weber, 1975) since it would be expected that an m/+ virus would segregate mlm and progeny during growth of the plaque. These mixed plaques might easily be scored as wild type under many of the screening procedures to be described below. The errors associated with treatment directed against virus replicating in uiuo arise from the fact that mutations occurring early will be replicated, and “separate” mutants isolated from the treated stock may prove to be siblings. This could lead to false conclusions concerning the mutation frequency of certain gene functions. The problem may be avoided by isolating single mutants from a number of separately treated stocks, and in general this has been the method employed. However, even where this precaution has not been taken, separate mutants isolated from one stock or mutagenized virus have often proved to be different. Thus Lundholm and Doerfler (1971) isolated mutants from a treated stock of type 12 adenovirus, several of which were different as judged by their temperature-sensitive phenotype.

+/+

104

HAROLD S . GINSBERG AND C . S . H. YOUNG

The frequencies of mutants induced by the various mutagens are listed in Table I. In no case, is there a clear indication of the optimal dose of mutagen or length of mutagenic treatment in induction of mutation; although Suzuki et aZ. (1972) have published a table showing the frequency of ts mutants after various doses of UV irradiation, the numbers of mutants obtained were limited by the difficulty of screening many hundreds of plaque isolates, and thus the dose versus mutant induction response was hard to assess accurately. C. SELECTION

All the phenotypic changes being looked for are the result of so-called “forward mutations, i.e., mutations from wild to mutant phenotype. In one case, the cytocidal mutants of adenovirus type 12 (Takemori et al., 1968), the search involved visual inspection of plaques; in another, the isolation of a heat-stable mutant of type 5 adenovirus followed three rounds of selection by heat inactivation (Young and Williams, 1975). More frequently, however, the change involves the loss of some function, e.g., the ability to grow on a certain host cell or to replicate at a particular temperature. As yet there have been no published reports of efforts to select adenovirus mutants on the basis of their inability to grow under restrictive conditions, while the wild type grows and incorporates some toxic substance such as a base analog or an amino acid analog. This procedure, which should lead to the selective destruction of wildtype virus in a mixed population, would be expected to enrich the nongrowing mutant phenotype. Thus, in the absence of selective methods, most mutant hunts have been random in the sense that plaques arising from viral stocks have been individually checked for the mutant characteristic. Naturally, this is a time- and material-consuming procedure. In the case of the ts mutants isolated by Williams et at. (1971), well separated plaques which had been grown at 31OC were picked and checked by plaquing at the restrictive and permissive temperatures. Plaques were tentatively classified as ts if they gave 50-fold less progeny plaques at the restrictive temperature compared with the permissive temperature. Essentially similar procedures were carried out by Ishibashi ( 1971) and Grodzicker et al. (1974a) in their searches for ts mutants of CELO adenovirus and of the nondefective SV40 adenovinis type 2 hvbrid AdPND1, respectively. Ensinger and Ginsberg ( 1972) employed a two-step, semiselective method to help speed the isolation procedure. Mutagenized virus was plated and incubated at the permissive temperature until small plaques appeared. Then, the plaques were marked and the plates were shifted

GENETICS OF ADENOVIRUSES

105

to the nonpermissive temperature. Plaques that failed to enlarge were picked and tested by infecting cells at a low multiplicity of infection to screen for inability to produce cytopathic effect (CPE) at the high temperature. Suzuki and Shimojo ( 1971), Shiroki et al. (1972), Takahashi ( 1972), and Ledinko (1974) screened plaque isolates directly by checking for CPE at the high temperature, while Lundholm and Doerfler (1971) picked plaques but transferred samples by sterile toothpick to plates already overlaid with agar, thus confining CPE to a small region and allowing more than one isolate to be screened per plate. BCgin and Weber (1975) have used the plaque enlargement and CPE methods and also have screened plaque isolates for the ability to make “inclusion bodies” at the high temperature. All these methods of selection tend to underestimate the frequency of ts mutants in the population either because ‘‘leaky’’ mutants will give CPE or because plaque enlargement may occur by cytocidal effects of defective virus or of viral proteins.

D. THE PROBLEM OF MULTIPLEMUTATION In all the cases where mutagenesis has been used to enhance the frequency of the mutant type being sought, those mutants which have been isolated will inevitably have more mutational lesions than the one that will give rise to the desired phenotype. No attempt has been made to determine whether or not all the phenotypic characteristics of a particular mutant are pleiotropic effects of one mutation or are caused by multiple mutations, by, for example, checking wild-type revertants or recombinants to see whether phenotypic characteristics are separable. Another potentially serious source of confusion, which was first pointed out for adenoviruses by Ishibashi (1971) and by Williams and Ustacelebi (1971b), lies in the isolation of mutants whose phenotype is caused by multiple mutations of the same class. For example, in a population where 10%of the survivors, following mutagenic treatment, are ts, 10% of the ts mutants are expected to arise from double mutations. This situation often can be revealed by complementation and recombination analysis which will be described below. 111. Genetic Characterization

A. COMPLEMENTATION The complementation test is designed to determine whether or not two mutants have mutations lying in the same or in different genes. Complementation tests of adenovirus ts mutants have been performed

106

HAROLD S . GINSBERG AND C . S . H. YOUNG

by infecting cells with a mixture of the pairs of mutants to be tested and with the mutants individually. The cells are then incubated at the restrictive temperature, and the yields from the mixed and single infections are compared by titration at the permissive temperature. To ensure that any increase is not caused by wild-type revertants or recombinants occurring in the mixed infections it is necessary to plaque the yields at the nonpermissive temperature as well, Various arbitrary criteria have been adopted for identifying positive complementation. In the case of the Ads t s mutants isolated by Williams et al. (1971), an increase of 10-fold in the yield of the mixed infection compared with either of the single infection controls was taken as being positive (Williams and Ustacelebi, 1971a), while in the tests of Ensinger and Ginsberg (1972) a continual increase with time in the ratio of mixed to single infection yields was evidence for complementation. In general, complementation has been found to be an efficient process both in terms of the increase in the yield in the mixed infection compared with singles and also in the absolute amount of virus produced, which may be as high as 5M (Ledinko, 1974) of the comparable yield of wild type grown at the restrictive temperature. It should be pointed out ( a ) that the efficiency of complementation is easy to determine if the t s mutants used are not leaky or do not have high levels of reversion and ( b ) that if complementation is low it is essential to determine whether or not the increase has been caused solely by wild-type virus arising in the mixed infection (see above). Table I1 summarizes the results published so far. It is important to point out the information that these results imply as well as the experimental and theoretical limitations that apply to them. 1. The results of Williams et al. (1974) can be interpreted to indicate that as many as 16 different genes have been identified, i.e., from onethird to one-half of the coding capacity of which a double-stranded DNA of 23 x lo6 MW is capable. Their results illustrate two points. ( a ) The accumulation of mutations in certain complementation groups implies that the mutagenic methods used are going to reveal further groups slowly. The increase from 22 mutants in 14 groups (Russell et al., 1972a), to 51 mutants in 16 groups (Williams et al., 1974) clearly indicates this point. Thus without rapid screening methods for complementation, as have been developed for SV40 (Chou and Martin, 1974), the task becomes progressively more laborious as each new ts mutant must be tested against all preexisting complementation groups. ( b ) The numbers of complementation groups may exceed the number of gene functions that have t s mutations located in them. Thus, in the complementation tests summarized by Williams et al. (1974), certain mutants

TABLE I1 COMPLEMENTATION GROUPSA N D THEIRPHENOTYPIC DEFECTS Complementation tests Virus

No. of No. of mutants groups

Mutants with single phenotypic defects” DNA Synthesis “Hexon”

“Hexon transport”

Fiber

15

13

0

0

0

Ad5

51

120

l(3)

I (24)C

l ( 7 ) ~ 3(3, 2, 1) 0

15

6

3

3

34

13

10 12

6 8 -

Ad12

Ad31 CELO

-

0

“Assembly”

3(1, 1, l ) b 8(6 X 1; 2 X 2)

Ad2

0

0

Penton base

0

4(3, 2, 1, 1)

1(1)

References BBgin and Weber (1975), Weber et al. (1975) Russell et al. (1972a, 74), Williams et al. (1974), Wilkie et al. (1973) Ensinger and Ginsberg (1972), Ginsberg et al. (1974a) Rubinstein and Ginsbergd (1974) Shiroki et al. (1972), Shiroki and Shimojo (1974) Ledinko (1974) Suauki et al. (1972) Ishibashi (1971)

a Some complementation groups have more than one phenotypic defect and are not included in this table (with the exception of DNA synthesis mutants, which fail to synthesize hexon, fiber, and penton base). Not all complementation groups have been examined phenotypically. b Numbers in parentheses refer to the numbers of mutants in each complementation group; e.g., 2(1, 1) indicates two complementation groups with one mutant each and (6 X l) indicates six groups with one mutant each. Not all mutants in each group have been examined phenotypically. c The publication by Williams et al. (1974) classified these mutants as hexon: 2(1, 1) and hexon transport: 4(20, 7, 1, 1). The reassignment is based on preliminary complementation data, as explained in the text. See also footnote in Section II1,A. d See footnote a in Table I. 8 Complementation was not performed with the CELO t s mutants.

8

3

8 mG

2

z

m

v)

108

HAROLD S. GINSBERG AND C. S. H. YOUNG

defective in hexon antigen or in the transport of hexon from cytoplasm to nucleus gave low but positive values for complementation among themselves. It was suggested (Williams et al., 1974) that this might be accounted for by intracistronic complementation ( Fincham, 1966) between hexon polypeptides with defects located at different positions that might yield a partially active multimeric hexon capsomer. In this context it is important to distinguish between two types of hexon transport mutants-that which involves a separate gene function and that which involves the failure of the hexon polypeptide to assume the correct conformation for transport at the restrictive temperature. In complementation tests carried out by R. S. Kauffman (unpublished data) at a higher restrictive temperature (39.5”C compared with 38.S°C), several of the Williams ts mutants have been classified using Ad5 ts mutants whose phenotype has been pinpointed as being in either hexon polypeptide or in the separate function of hexon transport (Kauffman and Ginsberg, 1975). Preliminary data suggest that at least two of the Williams compiementation groups, one of which was previously classified as “hexon transport” (Russell et al., 1972a; Williams et al., 1974), fall into the hexon polypeptide class, not into the transport function class. One of the complementation groups previously described by these authors as “hexon transport” does fall into the transport function class ( R . S. Kauffman, unpublished data). Using these data and extrapolating that all the low complementating groups fall into the single hexon gene, an alternative classification of the Williams hexon and hexon transport mutants is included in Table 11.‘ 2. Some of the ts mutants that have been isolated fail to complement mutants from more than one group (e.g., in Suzuki et al., 1972). This strongly suggests that they are multiple mutants. If the mutant fails to adsorb, penetrate, or uncoat at the restrictive temperature, it should fail to complement with any other complementation group. In contrast to SV40, where such mutants have been found (Robb and Martin, 1972; Chou and Martin, 1974), there are as yet, no examples of such a class in adenoviruses. It can be stated, by contrast, that mutants that do complement all other groups are not doubles involving any other group 40 far isolated. 3. The availability of more than one mutant within a particular comRecent data describing the physical map and the transcriptional program of the genome suggest, however, that the assignment of mutations to proteins may be at variance with the interpretation based solely on phenotypic expression. Since the writing of this review, the interpretation based on the physical experiments places the hexon gene at the site previously referred (Fig. 3 ) to hexon transport and the mutants phenotypically defective in hexon antigen in the region of the lOOK protein.

GENETICS OF ADENOVIRUSES

109

plementation group has been useful in checking that the physiological defect found in one mutant of the group at the restrictive temperature is common to the group and is not peculiar to that mutant (Ensinger and Ginsberg, 1972; Russell et a,?., 1972a; Shiroki et al., 1972). This approach may be used instead of comparing the mutant with a ts+ revertant. In some cases, such revertants are difficult to obtain owing to the low reversion frequencies and to the cytotoxicity of the t s viral population on plaquing at high concentration during selection for ts' viruses. Despite the limitations to complementation analysis outlined above, the technique is useful in restricting the numbers of mutants that are to be examined physiologically, and it gives indications of the number of genes that may be easily identified by ts mutations. Furthermore, complementation tests have predictive value in the recombination and mapping analyses to be described below, in that mutants in the same complementation group are expected to map closely together. The use of heterotypic complementation tests will also be discussed.

B. RECOMBINATIONTESTS Recombination between adenovirus genomes has been observed using ts mutants (Williams and Ustacelebi, 1971a; Ensinger and Ginsberg,

1972; Ledinko, 1974; BBgin and Weber, 1975), cyt-kb mutants which may be considered a type of host-range mutant (Takemori, 1972), and a heat-stable mutant (Young and Williams, 1975). With the first two classes of mutants, the ability to select the recombinant wild type from a predominantly mutant population greatly facilitated the demonstration of genetic exchange. The phenomenon of recombination has been or may be exploited in a number of ways: (1) to arrange mutants in order, by constructing a genetic map based on the recombination frequencies (r.f.) that occur between them, the greater the r.f., the greater the genetic distance (Williams et al., 1974; M. J. Ensinger, R. S . Kauffman, and H. Ginsberg, unpublished); ( 2 ) to establish whether or not mutants in the same complementation group were identical, as for example the demonstration by Takemori (1972) that many cyt-kb mutants were separable by recombination; ( 3 ) to utilize recombination, as mentioned earlier, to investigate whether or not all phenotypic characteristics of a particular mutant are caused by a single base change, i.e., are pleiotropic expressions of a single mutational lesion; and ( 4 ) to construct new genotypes, as in the case of the transfer of a heat-stable marker from the t s mutant in which it was isolated to wild type and thence to other t s mutants (Young and Williams, 1975).

110

HAROLD S. CINSBERG AND C . S. H. YOUNG

Recombination tests using t s mutants are normally arranged as follows. Cells are infected with both parental strains of virus and parallel cultures of cells are infected with each virus alone. The infected cells are incubated for several days at the permissive temperature, and the yields are titrated at both the restrictive and the permissive temperatures, to select for ts+ rcombinants and to measure total viral yields, respectively. The recombination frequency, r.f., expressed as a percentage has commonly been calculated as: (titer at the restrictive temperature) / (titer at the permissive temperature) x 2 x 100. In general the frequency of revertants i n the single parent controls has been so low as to make a correction for reversion in thc mixed infection unnecessary. The factor of 2 in the expressip allows for the presence of the unmonitored double ts recombinant class which is assumed to arise at equal frequency to the ts' class. This has never been established with adenoviruses as it is a formidable technical task to screen t s offspring for double ts genotype by complementation analysis. Limited data from crosses using a heatsiable ts mutant, ts-hs, and wild type, ts+-hst, suggest that the two recombinant types, heat-stable wild type, ts+-hs, and the ts mutant ts-11s' occur at frequencies that are not grossly dissimilar (Young and Williams, 197.5). In the absence of evidence of the reciprocal class occurring in ts x ts crosses, some investigators have preferred to use r.f. as expressing only the frequency of ts+ (Ledinko, 1974). Some technical points should be noted. ( 1 ) The r.f. values for a particular cross vary over a considerable range (Williams and Ustacelebi, 1971a; Williams et al., 1974), and thus to obtain a reliable statistic, several separate tests should be run. ( 2 ) It is important to establish that the plaques appearing at the restrictive temperature are genuinely ts+ and are not caused by complementation of different t s mutants on the assay plate. This has been tested by picking plaques and checking that the progeny have a wild-type genotype, i.e., yield the same number of plaques at both restrictive and permissive temperatures (Williams and Ustacelebi, 1971a; Ensinger and Ginsberg, 1972; Ledinko, 1974; Begin and Weber, 1975). Although one study demonstrated that complementation on the plate may be a serious disturbance to r.f. values (Ensinger and Ginsberg, 1972), this has not been reported by others. ( 3 ) The mixed infections should be incubated sufficiently long to permit the exponential phase of viral replication to be completed, since the r.f. increases with viral increase (Williams et d., 1974). With adenoviruses, despite the detection and biochemical and biological characterization of defective viruses with deleted genomes ( Mak, 1971; Burlingham et al., 1974), no attempt has been made to clone such deletions and use them for genetic purposes. The ordering of genes

GENETIC3 O F ADENOVIRUSES

111

with deletion mutants would seem to be technically feasible. Similarly, although some mutants have been isolated that would be excellent third markers in ts x ts crosses, no advantage has been taken of them. In particular the cgt mutants of Ad12 (Takemori et uZ., 1968) and the host-range markers of Ad5 (Takahashi, 1972) would be eminently suitable since the yield from three-factor crosses could be tested directly by selection to screen both for ts+ and for the third marker. The only investigation to date which has involved the use of a third marker is that using a heat-stable mutant of Ad5. The data were limited by the since the necessity of picking ts+ plaques and checking for h and h+, yield showed phenlotypic mixing for the heat-stable phenotype (Young and Williams, 1975). In view of the lack of more precise mapping procedures, the data from two-factor ts X ts crosses have had to be used to generate a genetic map. To date the only published orders or maps for adenoviruses are those for human adenoviruses type 2 (Bkgin and Weber, 1975) and type 5 (Williams et al., 1974). The former is based upon data using twelve mutants from twelve complementation groups and yields a relatively unambiguous order. It should be pointed out that very high values for r.f. were obtained for some crosses by these investigators; e.g., values as high as 38%were obtained for the percentage of ts+ among the progeny. This is formally equivalent to the statement that the mutants are unlinked. The significance of the map positions of the twelve mutants in the crosses must await the discovery of their phenotypic characteristics, an investigation that is under way ( Weber et al., 1975). The type 5 adenovirus map of Williams et al. (1974) is reproduced in Fig. 3, top line. This map is a compilation of data obtained over a number of years and represents a “best fit” for the r.f. values. The additivity of r.f. is good in many instances, an observation that lends support to the view that most, if not all, of the ts mutants used were single mutations, but there are a number of anomalies, as the authors pointed out. Whether or not these anomalies are significant in terms of the molecular organization of the chromosome, or are related to specific marker effects, remains to be determined. For comparison, the map of Ad5, which has been constructed recently by M. J. Ensinger, R. S. Kauffman, and H. S. Ginsberg, (unpublished) using t s mutants which have been independently isolated and characterized, is shown in Fig. 3, line 2. Several features of the two maps are worth emphasizing. ( 1 ) Where direct cqmparisons can be made, the order of markers of known phenotypes and in corresponding complementation groups are identical. In many instances, the genetic distances (r.f.’s) are similar. ( 2 ) In both maps, it is apparent that there is no absolute segregation of early and

HAROLD S . GINSBERG AND C. S . H. YOUNG

112

I

420 19

---

--+ 49

125

ASSEMBLY

\/

? 1,8

36

!.

59

I

i

I I

I

24

56

ASSEMBLY

I

I I I

,-

I;

HEXON TRANsPoRT I I

!

j

i I

I

’,

ASSEMBLY

I

-- -

149

:

15

30

1

__

‘ 1

j ‘.,I

I

FlBEp 1.2,3 I

I

I I

I

I/

I I

125 122

147

I38

;

II

II

HFXON

I I

---

115

47

142 56

04_18

I

I I

I

?

- --

5

I

II

DNA SYNTHESIS

9

I

110 10 l cr u

I

I I

DNA SYNTHESIS II

65

I

I I

I

22

1:!!~3

w

~

FIG.3. The genetic map of type 5 adenovims based upon two-factor recombination frequencies between pairs of t s mutants. Adapted from Williams et al. (1974), upper line, and M. J. Ensinger, R. S. K a u h a n , and H. S. Ginsberg, (unpublished), lower line. Symbols:

49

=+a

I

122 115

1 . 5.6

I

FI+R I

-

mutant position on map; except for the “hexon” gene, each mutant represents one complementation group alleles of the “hexon” gene (see footnote in Section II1,A) recombination frequency between two mutants mutant phenotype; vertical dashed lines indicate that complementation tests have been performed between mutants isolated in the laboratories of J. F. Williams and H. S. Ginsberg, and the mutants have been assigned to particular complementation groups.

late functions. Both DNA synthesis genes, which are early acting, are bracketed by late functions. (3) The maps demonstrate that the two functions related to hexon phenotype are separated by as many as three other genes. In this context, it is worth pointing out that the five complementation groups that map closely together and are variously de-

GENETICS OF ADENOVIRUSES

113

scribed as hexon-antigen negative and hexon-transport defective (mutants 1, 3, 4, 17, and 20) probably consist of mutants that can undergo intracistronic complementation. ( 4 ) Although the data are not shown on the maps, alleles within the same complementation group map closely together and at approximately the same distance from alleles in other genes. This is to be expected if the r.f.'s reflect physical genome distances. C. CORRELATION BETWEEN GENETIC AND PHYSICAL MAPS

The value of genetic mapping rests on the assumption that the genetic order corresponds to the physical order of genes in the genome. The proof of this assumption in other organisms encourages one to believe that it will be true for adenoviruses, but it must be established, if genetic mapping is to be a useful predictive tool. The physical structures of the DNAs of many serotypes of adenovirus have been investigated using site-specific restriction endonucleases obtained from a variety of microorganisms. The characteristic fragments that are produced may be used to construct a physical map (Mulder et al., 1974) against which genetic and biochemical markers may be aligned. With the development of hybridization techniques for selecting messenger RNAs that bind to specific fragments, and in uitm translation systems for such exogenous messengers, it is possible to map early and late transcripts and to locate genes for specific viral polypeptides (Lewis et al., 1975). Thus genetic and physical maps may be compared to correlate the position of the fragment that codes for a particular viral polypeptide and the map position of the gene that appears to control this polypeptide. The pitfall in this approach is that t s mutations may not lie in the structural gene for the p~lypeptidewhose activity is altered. For example, the hexon may be altered in quantity or the processing of its mRNA or a protein precursor may be altered by a mutation lying outside its structural gene. To avoid this danger, methods must be devised that directly correlate the t s mutant with the physical genome. The use of ts mutants in different adenovirus serotypes, from which different restriction endonuclease fragments may be generated, allows the selection and physical description of intertypic ts+ recombinants. This approach has been developed by Grodzicker et d. (1974b) and Williams et al. (1975a); it will be described in some detail because it is a novel example of the power of a combined genetic and physical investigation and because it has led to the conclusion that the recombination map for type 5 adenovirus, despite its being based on somewhat ambiguous data, is a faithful representation of the physical structure underlying it. Ad2'NDl is a nondefective hybrid between adenovirus type 2 and

114

HAROLD S. CINSBERG AND C. S . H. YOUNG

simian virus 40 (SV40), which is capable of plaquing with equal efficiency on both human and monkey cells (Lewis et al., 1969) and contains 17% of the SV40 genome located 15%from the right-hand end of the adenovirus genome ( Kelly and Lewis, 1973). Temperature-sensitive and host-range mutants of this virus are now available (Grodzicker et d., 1974a,b). Types 2 and 5 adenovirus belong in the same subgroup of adenovirus serotypes, which are classified as nononcogenic and show considerable DNA base homology (Garon et al., 1973; Bartok et al., 1974). Probably because of this close relatedness, the ts mutants from Ad 5 and Ad2'NDl can complement each other and recombine with frequencies close to those obtained in homologous crosses with corresponding homotypic mutants. The ts' recombinants from heterologous crosses can be checked for their ability to plaque on monkey cells, the SV40 fragment acting as a host range third marker; more important, since the restriction fragment patterns of Ad5 and Ad2'NDl are distinct, the genetic contribution of both parents to the recombinant may be determined. Thus, when a crossover from one parent to another occurs, the event is often marked by an alteration in the fragment pattern. When several ts' recombinants from a particular cross are compared in this way, those crossovers that are seen to be present in all of them are taken to be necessary for production of the recombinant. The position of these invariant crossovers set limits on the positions of the ts markers that enter the cross. It is important to note that (1) some restriction endonucleases may not yield distinctive patterns of fragments from certain areas of the molecule, in which case alternative enzymes must be used, and ( 2 ) double crossovers within the limits of the heterologous endonucleolytic cleavages will not be observed, which will lead to ambiguity unless other recombinants from the cross are available for comparison. Using this technique, it has been possible to locate the positions of three mutants of Ad5 and one of Ad2'NDl in the right-hand end of the genetic map. The orders and distances of the physical and genetic maps corresponded remarkably well ( Grodzicker et al., 1974b; Williams et al., 1975a). Recent data suggest that this correspondence holds for several more markers examined, including the DNA synthesis negative mutants H5ts36 and H5ts125 (J. Sambrook, H. S. Ginsberg, and J. F. Williams, personal communication). Some further extensions of this heterotypic analysis should be mentioned. First, the recombinants may be tested with antisera directed against specific polypeptides from each of the two viruses. Thus, it has proved to be possible to locate the fiber and hexon using recombinants that are expectcd to have crossover points between them and thus to have the fiber of one serotype and the hexon of the other (V. Mautner

GENETICS O F ADENOVlRUSES

115

and J. F. Williams, personal communication). Second, polypeptides from recombinants may be compared with those of either parent on sodium dodecyl sulfate ( SDS ) -polyacrylamide geIs; where polypeptide differences between serotypes can be detected, the genes may be mapped. It should be noted that restriction endonuclease analysis of recombinants may be used also to examine the frequency and nature of genetic exchange in animal viruses. D. HETEROTYPIC COMPLEMENTATION As has been mentioned in Section I, adenoviruses fall into three groups, depending on their oncogenic potential. Heterotypic complementation analysis provides a method to study the functional differences between them to determine which replicative functions and which structures can be substituted by one serotype for another. In adenoviruses, it is apparent that complementation between a serotype from the weakly oncogenic group (e.g., Ad2 or Ad5), and from the highly oncogenic group (e.g., Ad12 or Ad31), would be worth examining since DNA hybridization (Green, 1970), heteroduplex studies (Garon et d.,1973) and partial denaturation studies (Doerfler and Kleinschmidt, 1970; Doerfler et aZ., 1972) have shown these viruses to have different DNA structures. Heterotypic complementation between Ad5 and Ad12 has been examined by Williams et a2. (1975b). The complementation tests involved the use of a strain of Ad12 (1131) which failed to produce plaques on HeLa cells, although capable of infecting and producing cell-associated virus. The complementation tests therefore were of the form Ad12 wild type x Ad5 ts, and the yields were plated on HeLa celIs at the permissive temperature. Eight complementation groups from Ad5 could be complemented but seven other groups could not. Phenotypic mixing, as measured by the ability of neutralizing antisera against Ad5 and Ad12 to inactivate the yield, was observed in some positive complementation tests but not in others (Williams et al., 197%). It should be emphasized that it has not been possible to demonstrate recombination between type 5 and type 12 adenoviruses (Williams et al., 1975b), in contrast with the ability of Ad5 and Ad2'NDl to recombine freely. In bacteriophage, a correlation has been drawn between the homology of the DNA between different members of a related series and their ability to recombine [see, for example, the T3, T7, $11 series studied by Hyman et aZ. (1973)]. Heterotypic complementation within oncogenic classes to classify mutants has been alluded to in Section III,C. Shiroki and Shimojo (1974) have also used this method to classify ts mutants of Ad12 with respect to a ts mutant of Ad31 known to be defective in viral DNA synthesis.

116

HAROLD S . CINSBERG AND C. S . H. YOUNG

IV.

Phenotypes of Adenovirus Mutants

A. CHARACTERIZATION 1. Temperature-Sensitive Mutants As many as 22 virus-induced, possibly virus-coded, proteins have been detected in extracts of adenovirus-infected cells (Maizel et al., 1968a,b; Anderson et al., 1973; Everitt et al., 1973) (Fig. 1).Several are precursors of virion proteins (Sundquist et al., 1973a; Anderson et al., 1973; Ishibashi and Maizel, 1974a) (e.g., pVI and pVII), and several minor components have not yet been shown to be unique polypeptides. Hence, 12-15 putative primary gene products have been identified (Maizel et al., 1968a,b; Everitt et al., 1973; Anderson et al., 1973) although potentially the adenovirus genome could code for as many as 50-60 proteins with the average size of 25,000 daltons. Of the known viral gene products only a few appear to be altered in the large number of adenovirus temperature-sensitive mutants that have now been characterized (Table 11). Perhaps this can be attributed to the fact that the precise function and immunological properties of the native proteins have been identified only for the major viral proteins, and these are the viral products that appear to be represented in most of the phenotypes thus far described. Thus, the best characterized protein phenotypes listed in Table I1 represent lesions in the most easily studied and assayed capsid proteins, the hexon and fiber proteins. In contrast, so-called “assembly mutants” constitute a large proportion of the ts mutants classified, but whose defects have not yet been identified. These mutants are termed “assembly” mutants because at the nonpermissive temperature all identifiable proteins are synthesized and all the proteins that can be assayed are immunologically functional (Ginsberg, 1969; Schlesinger, 1969; Green, 1970; Philipson and Lindberg, 1974; Norrby, 1968; Philipson and Pettersson, 1973). These mutants may represent defects in one of the minor virion components or in a nonvirion protein that is essential for modulating the assembly process. If the “assembly” mutants represent a melange of mutations in a number of genes coding for nonvirion proteins, such as the 100 K, 50 K, 27 K, 26 K, or virion proteins XIII-XI1 ( Fig. l ) , a comparable number of complementation groups should have been identified. Instead, a maximum of five “assembly” complementation groups for type 5 ts mutants have been described (until mutants isolated in different laboratories are compared, the precise number of complementation groups remains uncertain). Two putative assembly mutants belonging to different complementa-

GENETICS OF ADENOVIRUSES

117

tion groups, also fail to induce interferon in chick embryo fibroblasts ( CEF) at the nonpermissive temperature (Ustacelebi and Williams, 1972). Since interferon is not a virus-coded product, nor is its induction viral specific, one simple interpretation of this observation is that the uncoating of these two mutants does not proceed to completion in CEF at the nonpermissive temperature because structural components of the capsid are altered. The observation that, although the mutants complement in the infectious cycle in HeLa cells, they failed to complement in the induction of interferon on CEF (Ustacelebi and Williams, 1972) is consistent with this interpretation, as is the finding that the induction of interferon was thermosensitive only during the first 6 hours after infection but the ts lesion blocked the production of virus in HeLa cells at late times during infection (Ustacelebi, 1973). In addition, both mutants were found to be considerably more thermolabile than wild type in uitro (Ustacelebi, 1973; Young and Williams, 1975), which is not unexpected for ts mutations in capsid structural components (Fenner, 1969). Thermolability of some ts mutants has also been observed in type 12 adenovirus (Shiroki et al., 1972) and in type 2 (Weber et al., 1975). During viral replication, adenovirus DNA can be separated from host cell DNA by virtue of its higher G C content (Piiia and Green, 1969; Ginsberg et al., 1967). Thus it has been a relatively simple task to screen ts mutants for the ability to synthesize DNA at the restrictive temperature. So far, mutants with DNA synthesis defects have been isolated in type 5 (Ensinger and Ginsberg, 1972; Wilkie et al., 1973), type 12 (Ledinko, 1974; Rubenstein and Ginsberg, 1974; Shiroki aad Shimojo, 1974), type 31 (Suzuki et al., 1972), and CELO (Ishibashi, 1971). Several of these mutants have also been examined for alteration in the frequency of transformation of rat or hamster embryo cells; these investigations will be discussed later.

+

2. Plaque Morphology and Host-Range Mutants Takemori et al. (1968) isolated a large number of mutants of adenovirus type 12 which gave plaques that were larger and clearer than those given by wild-type virus on early-passage human embryo kidney (HEK) cells. These mutants were discovered to be of low tumorigenicity in newborn hamsters and failed to transform newborn hamster kidney cells in oitro. Some of the mutants, (“cyt-kb”),failed to propagate in one line of KB cells (KB-1) while others, (“cyt-kb”’),multiplied in them; both types propagated in KB-2 cells (Takemori et al., 1969). Although there is evidence that the product of the cyt gene is diffusible

118

HAROLD S. GINSBERG AND C. S. H. YOUNG

(see Section IV,B,4 on transformation and tumorigenesis), as yet it is not known what viral product is involved. Host-range mutants of human adenovirus type 5 have been isolated (Takahashi, 1972; T. J. Harrison and J. F. Williams, personal communication) that are capable of growing on human cells but unable to do so on hamster cells. As yet no specific lesion has been identified, but on SDS-polyacrylamide gels, it is clear that all virus-infected, cell-specific proteins are made, but in reduced quantity (T. J. Harrison and J. F. Williams, personal communication ) . As mentioned previously, the nondefective, adeno-SV40 hybrid AdB'NDl, is capable of growing on both human and monkey cell lines (Lewis et al., 1969). It was thought that host-range mutants that fail to grow on monkey cells would be mutated in the integrated SV40 fragment, which presumably promotes adenovirus lytic cycle functions in the normally semipermissive monkey cells. Accordingly, Grodzicker et al. (1974a) selected such mutants and examined the polypeptides made in both human ( HeLa) and monkey cells (CVI clone of AGMK) infected with Ad2"Dl wild type, a host-range mutant, and Ad2. In the HeLa cell line, a protein product of 30,000 daltons, which is characteristic of Ad2'NDl infection, was absent in both Ad2 and host-range mutant infected cells. In CV1 cells, the characteristic underproduction of many adenovirus late proteins was observed in both Ad2 and hostrange mutant infections. The authors also examined the production of the perinuclear SV40 U antigen and found that in permissive cells infected with the host-range mutant the appearance of maximal detectable antigen was delayed, whereas in CV1 cells, very few nuclei ever displayed antigen. Whether or not the host-range mutation lies in the structural gene for the 30,000 dalton protein or in a viral gene that induces a host-cell protein is not clear. It should be noted that host-range mutants can be absolutely defective if the function(s) that are mutable are required only for replication of the virus in the restrictive host. This is perhaps unlikely in the hamster/ human Ad5 mutants but is a distinct possibility in the Ad2'NDl hostrange mutants. Absolute defectives are of considerable interest, since, as a class, they include deletion mutants that may be used for mapping purposes and, where they delete part of a gene, may be used to identify specific gene products.

B. FUNCTIONAL STUDIESUSINGADENOVIRUS MUTANTS Since, during replication, conditionally lethal temperature-sensitive mutants can be shifted conveniently from conditions that permit full

GENETICS OF ADENOVIRUSES

119

expression of the viral genome to nonpermissive conditions that do not allow expression of a defective gene, the gene product affected can be identified and its functional role in viral synthesis can be explored. This approach to the study of viral replication and its regulation has had noteworthy success with bacterial viruses and offers similar promise for unraveling the intricacies of adenovirus synthesis. Studies taking advantage of adenovirus ts mutants are in effect just beginning, and significant progress has been made only investigating three central areas in viral replication: DNA synthesis (Suzuki and Shimojo, 1974; Shiroki and Shimojo, 1974; van der Vliet et al., 1975; Levine et al., 1974), transcription (Carter and Ginsberg, 1975), and transport of viral proteins (Kauffman and Ginsberg, 1975). In addition, both ts and absolute mutants are being employed to explore the mechanisms underlying the establishment and maintenance of adenovirus transformation ( Ledinko, 1974; Ginsberg et al., 1974a,b; Williams et al., 1974; Rubenstein and Ginsberg, 1974; Takemori et aZ., 1968; Takahashi et aZ., 1974). 1. Viral DNA Synthesis

Relatively few mutants that cannot replicate their DNA under nonpermissive conditions have been reported. But the mutants available thus far encompass two complementation groups for type 5 virus (Ginsberg et al., 1974a,b; Williams et aZ., 1974) and three unique complementation groups for type 12 virus (Shiroki and Shimojo, 1974). It is striking that all the mutants isolated appear to be defective in initiation (Shiroki and Shimojo, 1974; Ginsberg et al., 1974a,b). The methods employed, however, do not distinguish between a mutant that cannot initiate replication and one that can elongate but not engage in a final reaction that may be required to terminate the synthesis of a complementary strand in order to initiate a new round. Thus, at least three virus-coded gene products appear to be essential for adenovirus DNA replication, and each is probabIy required for a reaction in chain initiation. One virus-specific protein, which binds preferentially to single-stranded DNA (van der Vliet and Levine, 1973), in a similar manner to the T4 gene-32 protein ( Alberts and Frey, 1970), has been shown to be defective in H5ts125 (van der Vliet et al., 1975) and H12ts275 ( Rosenwirth et al., 1975). Moreover, the so-called DNA-binding protein made in H5tsl25-infected cells is degraded at 39.5"C and dissociates from single-stranded DNA at lower temperatures than the wild-type protein (van der Vliet et al., 1975). Although the function of the DNAbinding protein is uncertain, since the adenovirus replicating form contains extensive single-stranded regions ( Sussenbach et al., 1972, 1973; Ellens et al., 1974; Pettersson, 1973), the binding protein could serve

120

HAROLD S . GINSBERG AND C. S . H. YOUNG

to maintain the single strands and thus permit effective copying of the displaced strand (van der Vliet et al., 1975). 2. Transcription of the Viral Genome Evidence obtained using pyrimidine analogs to inhibit DNA replication showed that only early mRNAs could be transcribed, from the infecting parental genome, and that late transcripts could be made only after DNA synthesis was begun (Bello and Ginsberg, 1969; Lucas and Ginsberg, 1971). These conclusions were subject to criticism, however, since it was not possible to demonstrate unambiguously that the chemicals employed ( i.e., 5-fluoro-2-deoxyridine, and arabinosylcytosine), did not effect any other intracellular biosynthetic reactions. With the availability of appropriate mutants, it became possible to investigate more rigorously the relationship between viral DNA replication and viral transcription. Carter and Ginsberg (1975) used two DNA-minus mutants, H5ts125 and H5ts149, to infect KB cells at nonpermissive temperature (41°C) and employed hybridization techniques to measure DNA replication and RNA transcription. The data obtained confirmed the earlier studies that no late transcripts appeared if onset of DNA replication was not permitted, and that under the restrictive conditions all early mRNAs Classes I and I1 (Lucas and Ginsberg, 1971) were transcribed. It was further shown that both Class I and I1 mRNAs continued to be transcribed as long as 15 hours after infection at 41"C, indicating that the shutoff of Class I mRNA was the work of a late gene function. It is also striking to note that although the onset of DNA replication is essential for the switch to transcription of late messages, the continuous replication of viral DNA is not required for late transcripts to be made: i.e., when cells were infected with H5ts125 or H5ts149 at 32°C for 25 hours and then shifted to 41"C, the rate of DNA synthesis decreased rapidly for H5ts125 and slowly for H5ts149; but the rate of viral RNA synthesis after the shift up did not change for 3 to 4 hours for ts 125 and even continued to increase for ts 149; and within the limits of the hybridization-competition techniques employed, the data indicated that the mRNAs made consisted of all the late sequences and Class I1 early RNAs.

3. Transport of the Hexon Protein As noted earlier, adenovirus proteins are synthesized on cytoplasmic polyribosomes and rapidly transported into the nucleus (Velicer and Ginsberg, 1968, 1970), where about 10%are assembled into virions (Wilcox and Ginsberg, 1963~).The mechanism of protein transport has been difficult to investigate owing to the marked leakiness of the infected

GENETICS OF ADENOVIRUSES

121

nucleus for viral proteins (Velicer and Ginsberg, 1968), although it was considered likely that host-cell functions played a major role. The discovery of ts mutants in which one or more capsid proteins were made but accumulated in the cytoplasm rather than moving into the nucleus under nonpermissive conditions ( Ishibashi, 1970, 1971; Shiroki et al., 1972; Russell et al., 1972a; Ginsberg et al., 1974a,b), suggested that these may offer an opportunity to study the process of protein transport. The mutants of one complementation group, whose hexon proteins specifically cannot move into the nucleus (Russell et al., 1972a; Ginsberg et al., 1974a,b; Kauffman and Ginsberg, 1975), appear to be of particular value for this purpose. The hexon, like the other capsid proteins, is immunologically active and folds into its native, multimeric structure, but only the hexon is not transported (Kauffman and Ginsberg, 1975). These transport mutants are clearly distinct from the hexon antigen minus mutants, since genetic recombination analysis shows that at least one other gene separates their respective gene loci. Biochemical, immunological, and physical studies of H5ts147 indicate that unaffected hexons are assembled and accumulate in the cytoplasm at 39.5"C; and that upon shift-down to 32"C, the preformed, cytoplasmic hexons can be transported into the nuclei and assembled into virions if protein synthesis is permitted (Kauffman and Ginsberg, 1975). Preliminary data suggest that the precursor (pVI) to protein VI (Anderson et al., 1973), a putative hexon-associated protein (Everitt et al., 1973), is defective in H5ts147 replication at 39.5"C (Kauffman and Ginsberg, 1975). The precise function of the small (27,OOO daltons) pVI protein is still unknown. 4. Transformation Conditionally lethal mutants offer the hope that specific gene product( s ) that effect cellular transformation can be defined, provided that these products are essential not only for lytic growth, but also for transformation. This optimistic note appears to have some validity since specific temperature-sensitive mutants of Rous sarcoma virus ( Tooze, 1973) and SV40 (Martin and Chou, 1975; Tegtmeyer, 1975; Brugge and Butel, 1975; Osborn and Weber, 1975) either cannot transform or cannot maintain the transformed state at the nonpermissive condition. The results of transformation studies with adenovirus ts mutants do not clearly identify a specific gene function to be directly concerned with transformation. Data do suggest, however, that at least one viral protein serves in the regulation of transformation. Some, but not all, DNA minus mutants [H5ts125 (Ginsberg et al., 1974a,b), H12ts307 (Ginsberg et al., 1974a,b), and H12ts401 (Ledinko, 1974)l transform two to eight times more rat

122

HAROLD S. GINSBERG AND C. S . H. YOUNG

or hamster embryo cells than does wild-type 5 or 12 virus. In sharp contrast, H5ts149 (Ginsberg et al., 1974a) and H12ts406 (Ledinko, 1974), which are also DNA-minus mutants in different complementation groups from H5ts125 and H12ts401, respectively, transform at the same frequency as wild-type viruses. H5ts36 and H5ts37 (Wilkie et al., 1973) however, which belong to the same complementation group as H5ts149 (Williams et d.,1974; Ginsberg et al., 1974a,b) transform rat embryo cells at a 10- to 20-fold lower frequency than wild type at the nonpermissive temperature, although cells transformed by the mutants at 32.5"C do not lose their transformed characteristics at 38.5"C (Williams et d.,1974). (One clone of such cells, however, has the curious property of being unable to grow at 38.5"C although it still maintains its transformed morphology). The failure of the cells of all such clones to revert to a normal morphology on shift-up to the nonpermissive temperature, suggests that H5ts36 and 37 are defective in the initiation, rather than in the maintenance, of transformation (Williams et al., 1974). Temperature-shift experiments, during the establishment of transformation, suggest that the temperature-sensitive step occurs before 48 hours growth at 32.5"C (Williams et al., 1974). Type 5 adenovirus ts mutants have also been used to transform hamster embryo cells, which are normally permissive for Ad5 (Williams, 1973). By performing the transformation at 38.5"C, t s mutants were unable to enter the lytic cycle and transformed clones were obtained. The clones were found to be highly oncogenic in newborn hamsters (Williams, 1973), and, furthermore, sera taken from such animals were found by indirect immunofluorescence techniques to react with the transformed cell nuclei (Williams et al., 1974). Another approach to restricting the usual lytic cycle of type 5 adenovirus infection in hamster cells, has been taken by Takahashi et al. (1974) and T. J. Harrison and J. F. Williams ( personal communication ) . They have demonstrated that hostrange mutants that fail to replicate in hamster cells nevertheless can transform them. It is noteworthy that the portion of the genome that contains the H5ts125 gene (J. Sambrook and H. S. Ginsberg, unpublished), which codes for a DNA-binding protein (van der Vliet et al., 1975),and perhaps that which contains the H5ts36 gene (J. Sambrook and J. F. Williams, unpublished) are not present in that minimum segment of the viral genome that has been detected in adenovirus-transformed cells ( Gallimore et al., 1974). Thus, as little as 14%of the distal end of the Eco.R1 restriction endonuclease generated A fragment (Mulder et al., 1974) appears to be necessary to maintain the transformed state (Gallimore et al., 1974; Sambrook et al., 1974). Whether under conditions of viral infection,

GENETICS OF ADENOVIRUSES

123

all or only a small portion of the genome is initially integrated to induce cell transformation is unknown, but transformation can be effected by only a fragment of the viral DNA representing as little as 5%of the viral genome and consists of only a small piece, which is present in the large Eco.R, “A” fragment (Graham et al., 1974). The finding that H5ts125, as well as related type 12 DNA-minus ts mutants, transforms at an increased frequency suggests that the DNA-binding protein may normally play a role that modulates the viral genome-cell interaction to reduce the opportunity for transformation, perhaps for DNA integration. These high-efficiency transforming mutants do clearly demonstrate that transformation is not dependent on replication of the viral DNA. The use of nonconditionally lethal mutants of defined genotype to study transformation has so far been restricted to the cyt mutants of human adenovirus type 12 (Takemori et al., 1968).These mutants, which gave larger and clearer plaques on early-passage HEK cells than did the parental wild-type strain, were found to have lost the ability to transform hamster cells in z)itro and to cause tumors in newborn hamsters. The mutants cooperated with low tumorigenic cyt+ field strains of Ad12 and the weakly tumorigenic viruses Ad3 and Ad7, to produce high levels of tumorigenicity, which suggested that the gene product was diffusible (Takemori et al., 1968). The cyt mutants have also been examined for the percentage of defective particles which they and the parental strain generated, to determine whether there was a positive correlation between the frequency of transformation and the proportion of defectives (Ezoe and Mak, 1974). No such correlation could be found. This does not rule out the possibility, however, that a specific class of defectives is the agent of transformation. Such a demonstration may have to await the development of methods for cloning virus of known defective constitution in ways analogous to those devised for SV40, in which specific deletion mutants were complemented by specific ts mutants at the restrictive temperature (Brockman and Nathans, 1974). It should be pointed out that all mutants that have been examined for transforming abilities were selected, in the first instance, for alterations in the lytic cycle and only secondarily for changes in transformation. This method of screening precludes mutants that are deficient only in transformation, if indeed such exist. V. Summing Up

Beauty may be attributed to a perfect object, but observation of nature’s imperfections has often revealed the elegance of structure and function awarded to earthly creatures. Thus, the study of inheritable

I!%

HAROLD S . CINSBERG AND C.

S. H. YOUNG

defects noted in a vast variety of plants and animals has uncovered rules and mechanisms of genetic interactions as well as processes regulating differentiation, morphogenesis, and biosynthetic events. Viruses, bacterial and animal, have been similarly studied and shown to follow the same genetic game rules as plants and animals-perhaps the most telling point that viruses may be considered to be organisms in that perennial argument as to whether viruses are living. Adenovirus genetics has been reviewed in this paper; as predicted, these viruses participate in the same genetic interactions as other organisms, and their genetic imperfections ( i.e., mutations ) are elucidating viral structure and reactions regulating viral function. It is striking that although the adenovirus genome can potentially code for fifty or more proteins, unique mutations have been detected in only ten to twelve genes. And of the number of conditionally lethal, temperature-sensitive mutants isolated, there is considerable clustering of two phenotypes, hexon and “assembly” mutants. Hexon mutants segregate into a single complementation group in which there is some evidence for intracistronic complementation; “assembly” mutants probably comprise four or five nonoverlapping complementation groups involving several gene functions essential to the regulation of virion morphogenesis. It is not clear, however, why the number of unique mutants detected falls so far short of the genome’s coding potential. Nevertheless, the mutants isolated follow predictive intergenic reactions; and despite the relative imprecision of two factorial crosses, a reproducible linear map has been constructed (Williams et al., 1974; M. J. Ensinger, R. S. Kauffman, and H. S. Ginsberg, unpublished). Indeed, the recombination maps produced (Williams et al., 1974; M. J. Ensinger, R. S . Kauffman, and H. S. Ginsberg, unpublished) correspond remarkably well with two independent techniques of physical mapping ( Grodzicker et al., 1974b; Lewis et al., 1975). Viral genetics, in addition to furnishing means for investigating viral structure and function, has predictive values. For example, Hirst noted the relatively high frequency of recombination between influenza virus mutants and suggested from these findings that the viral genome may be fragmented (Hirst, 1962). No such dramatic prognostication can be made from evidence accumulated with adenovirus mutants. The intergenic reactions demonstrated, however, do permit certain hypotheses that explain the evolutionary development of numerous adenovirus types in many animal species, and the viruses’ potential for incorporating viral genetic materials into host genomes to produce cell transformation, and tumors, and possibly latent viral infections. As noted earlier ts mutants have proved to be of greatest value for

GENETIa OF ADENOVIRUSES

125

investigating the structure and function of adenovirus proteins. Thus, the grouping of hexon mutants in a single cistron confirms chemical and physical evidence that the hexon is formed from the assembly of three identical polypeptide chains (Franklin et al., 1971; Cornick et al., 1973; Stinski and Ginsberg, 1975). And the demonstration, using genetic techniques, that a mutant defective in the transport of hexon protein into the nucleus is distinct from the hexon antigen minus mutants, revealed evidence that synthesis, assembly, and transport of hexon capsomers require more than a single gene product (Kauffman and Ginsberg, 1975). Furthermore, the detection of three nonoverlapping complementation groups representing three genes essential for production of a functional fiber (Russell et al., 1972a) implies that the fiber consists of three rather than two unique species of polypeptides (Dorsett and Ginsberg, 1975). Studies using adenovirus ts mutants for discovering the viral proteins involved and their function( s ) in regulating viral biosynthesis are still in their infancy. Nevertheless, types 5 and 12 mutants have revealed a minimum of three viral proteins essential for initiation of DNA replication (Shiroki and Shimojo, 1974; Ginsberg et al., 1974a,b). But only one gene product has yet been identified, a DNA-binding protein which prefers single-stranded DNA (van der Vliet and Levine, 1973; van der Vliet et d.,1975; Rosenwirth et al., 1975). Two of these mutants, H5ts125 and H5ts149, have also been valuable for studying regulation of transcription (Carter and Ginsberg, 1975). Using these mutants, it has been possible to confirm the earlier finding (Bello and Ginsberg, 1969) that late transcripts can be made only after replication of viral DNA has begun (Carter and Ginsberg, 1975). The reduction in transcription of early class I mRNAs also requires viral DNA synthesis. Despite these demands, continued viral DNA replication is not essential for transcription of late genes, since change to the nonpermissive temperature and the consequent rapid cessation of viral DNA synthesis does not alter the rate or the quality of late transcription (Carter and Ginsberg, 1975). The ts mutants unable to replicate DNA at the restrictive temperature have also proved to be valuable for investigating cell transformation; it was found that H5ts125 ( Ginsberg et al., 1974a,b) H12ts307 (Rubenstein and Ginsberg, 1974), and H12ts401 ( Ledinko, 1974) transform significantly more cells than wild-type viruses or other ts mutants. H5ts125 has been most thoroughly characterized and, as noted above, is defective in a virus-specific, single-stranded DNA-binding protein ( van der Vliet et al., 1975). It is striking that the portion of the viral genome that codes for the ts 125 gene product (Fig. 3), which is approximately 0.65 physical map unit from the left-hand end of the genome (J. Sam-

126

HAROLD S. GINSBERG AND C. S . H. YOUNG

brook and H. S. Ginsberg, unpublished data), is rarely integrated in adenovirus-transformed cells ( Gallimore et al., 1974; Sambrook et al., 1974). The significance of these findings is emphasized by the fact that three independently isolated mutants present this same phenomenon of increased transforming capacity, whereas three other DNA-minus mutants, belonging to another complementation groups, H5ts36 and 37 (Williams et al., 1974) and H5ts149 (Ginsberg et al., 1974a), show either decreased or normal transformation frequency. It appears inescapable that the adenovirus DNA-binding protein, whatever its function is, plays a regulatory role in transformation as well as DNA replication. The beginnings of adenovirus genetics recounted in this review invite continued optimism, since the data thus far obtained, as predicted, not only reveal new mechanisms controlling viral replication, but also offer greater understanding of genetic interactions of DNA-containing animal viruses.

REFERENCES Alberts, B., and Frey, L. (1970). Nature (London) 227, 1313-1318. Anderson, C. W., Baum, P. R., and Cesteland, R. F. (1973). J. Virol. 12,241-252. Bartok, K., Garon, C. F., Berry, K. W., Fraser, M. J., and Rose, J. A. (1974). J. Mol. Biol. 87, 437-449. BCgin, M., and Weber, J. (1975). J. Virol. 15, 1-7. Bellett, A. J. D., and Younghusband, H. B. (1972). J. Mol. Bid. 72,691-709. Bello, L. J., and Ginsberg, H. S. ( 1969). J . Virol. 3, 106-113. Boyer, G. S., Leuchtenberger, C., and Ginsberg, H. S. (1957). J. Exp. Med. 105, 195-216. Boyer, G. S., Denny, F. W., Jr., and Ginsberg, H. S. (1959). J. Exp. Med. 110, 827-844. Brockman, W. W., and Nathans, D. (1974). Proc. Nut. Acad. Sci. U.S. 71, 942-946. Brugge, J. S., and Butel, J. S. (1975). J. Virol. 15, 619-635. Burlingham, B. T., Brown, D. T., and Doerller, W. (1974). Virology 60,419-430. Carter, T. H., and Ginsberg, H. S. ( 1975). J . Virol. (in press). Chardonnet, Y.,and Dales, S. ( 1970). Virology 40, 462-477. Chou, J. T., and Martin, R. G. (1974). J . Virol. 13,1101-1109. Cornick, G., Sigler, P. B., and Ginsberg, H. S. (1973). J. Mol. Biol. 73, 533-537. Craig, E. A., Zimmer, S., and Raskas, H. J. (1975). J . Virol. 15, 1202-1213. Crick, F. H. C., and Watson, J. D. (1956). Nature (London) 177, 473475. Dales, S. (1962). J. Cell Biol. 13, 303-322. Doerfler, W., and Kleinschmidt, A. D. (1970). J . Mol. B i d . 5 , 5 7 9 5 9 3 . Doerfler, W., Hellman, W., and Kleinschmidt, A. D. ( 1972). Virology 47, 507-512. Dorsett, P. H., and Ginsberg, H. S. ( 1975). J. Virol. 15,208-216. Ellens, D. J., Sussenbach, J. S., and Jansz, H. S. (1974). Virology 61, 427-442. Ensinger, M. J., and Ginsberg, H. S. ( 1972). J. Virol. 10,328-339. Everett, S. F., and Ginsberg, H. S. (1958). Virology 6,770-771. Everitt, E., Sundquist, B., and Philipson, L. (1971). J. Virol. 8, 742-753. Everitt, E., Sundquist, B., Pettersson, U., and Philipson, L. (1973). Virology 52, 130-147.

GENETIG OF ADENOVIRUSES

127

Ezoe, H., and Mak, S. (1974). J . Virol. 14, 733-739. Fenner, F. (1969). Curr. Top. Microbiol. Immunol. 48, 1-28. Fincham, J. R. S. ( 1966). “Genetic Complementation.” Benjamin, New York. Franklin, R. M., Pettersson, U., Akervall, K., Strandberg, B., and Philipson, L. ( 1971). J. Mol. Biol. 57, 383395. Fujinaga, K., and Green, M. (1970). Pruc. Nut. A c d . Sci. U.S.65,375482. Gallimore, P. H., Sharp, P. A., and Sambrook, J. (1974). J. Mol. Biol. 89, 49-72. Garon, C. F., Berry, K., and Rose, J. (1972). Proc. Nut. A c d . Sci. US. 69, 2391-2395. Garon, C. F., Berry, K. W., Hierholzer, J. C., and Rose, J. A. ( 1973). Virology 54, 414426. Gilead, Z., and Ginsberg, H. S. (1965). 1. Bucteriol. 90,120-125. Ginsberg, H. S. (1969). In “The Biochemistry of Viruses” (H. B. Levy, ed.), pp. 329-359. Dekker, New York. Ginsberg, H. S., Pereira, H. G., Valentine, R. C., and Wilcox, W. C. (1966). Virology 28, 782-783. Ginsberg, H. S., Bello, L. J., and Levine, A. J. (1967). In ‘The Molecular Biology of Viruses” (S. J. Colter and W. Paranchych, eds.), pp. 547-572. Academic Press, New York. Ginsberg, H. S., Ensinger, M. J., Kauffman, R. S., Mayer, A. J., and Lundholni, U. ( 1974a). C d d Spring Hurbor Symp. Quunt. Biol. 39,419-426. Ginsberg, H. S., Ensinger, M. J., Rubinstein, F. E., and Kauffman, R. S. (1974b). In “Viruses, Evolution and Cancer” (E. Kurstak and K. Maramorosch, eds. ), pp 167-181. Academic Press, New York. Graham, F. L., van der Eb, A. J., and Heijneker, H. L. ( 1974). Nature (London) 251, 687-691. Green, M. (1962). Cold Spring Harbor Symp. Quunt. Biol. 27, 219-235. Green, M. ( 1970). Annu. Rev. Biochem. 39,701-756. Green, M., Piiia, M., Kimes, R., Wensink, P. C., MacHottie, L. A., and Thomas, C. A., Jr. (1967). Proc. Nut. Acad. S c i . U.S. 57, 13021309, Green, M., Parsons, J. T., Piiia, M., Fujinaga, K., Caffier, H., and Landgraf-Leurs, I. (1970). Cold Spring Hurbor Symp. Quunt. Biol. 35,803-818. Grodzicker, T., Anderson, C., Sharp, P. A., and Sambrook, J. (1974a). J. Virol. 13, 1237-1244. Grodzicker, T., Williams, J., Sharp, P., and Sambrook, J. (1974b). Cold Spn’ng Hurbor Symp. Quunt. Biol. 39, 439-446. Hirst, G. K. (1962). Cold Spring Hurbor Symp. Qmnt. Biol. 27,303-309. Horwitz, M. S., Brayton, C., and Baum, S. G. (1973). J. Virol. 11, 544-551. Huebner, R. J., Casey, M. J., Chanock, R. M., and Schell, K. (1965). Proc. Nut. Acud. Sci. US.54, 381388. Hyman, R. W., Brunovskis, I., and Summers, W. C. (1973). J. Mol. Biol. 77, 189-196. Ishibashi, M. (1970). Proc. Nut. Acud. Sci. US.65, 304409. Ishibashi, M. ( 1971 ). Virology 45, 42-52. Ishibashi, M., and Maizel, J. V., Jr. ( 1974a). Virology 57,409-424. Ishibashi, M., and Maizel, J. V., Jr. (197413). Virology 58,345-361. Jacobson, M. F., and Baltimore, D. (1968). J. Mol. Biol. 33, 369-378. Kasel, J. A., Huber, M., Loda, F., Banks, P. A., and Knight, V. (1964). Prac. SOC. Erp. Biol. Med. 117, 186190. Kauffman, R. S., and Ginsberg, H. S. ( 1975). J . viTO1. (in press). Kellenberger, E., Eiserling, F., and Boy de la Tour, E. (1968). J. Ultrastruct. Res. 21, 335-360.

128

HAROLD S. CIh’SBERG AND C. S. H. YOUNG

Kelly, T. J., Jr., and Lewis, A. M., Jr. (1973). J. Virol. 12, 643-652. King, J., and Casjens, S. ( 1974). Nature (London) 251, 112-119. Kjellhn, L. and Pereira, H. C. ( 1968). J. Gen. Virol. 2, 177-185. Koczot, F. J,, Carter, B. J., Garon, C. F., and Rose, J. A. (1973). Proc. Nat. Acad. Sci. U.S. 70, 215-219. Laver, W. G. ( 1970). Virology 41, 488-500. Lawrence, W. C., and Ginsberg, H. S. ( 1967). J. Virol. 1,851-867. Ledinko, N. (1971). Nature (London), New Biol. 233, 247-248. Ledinko, N. (1974). J. Virol. 14, 457-468. Levine, A. J., and Ginsberg, H. S. ( 1967). I . ViroI. 1, 747-757. Levine, A. J., van der Vliet, P. C., Rosenwirth, B., Rabek, J., Frenkel, G., and Ensinger, hl. (1974). Cold Spring Harbor Symp. Quant. Biol. 39, 559-566. Lewis, A. M., Jr., Levin, M. J., Wiese, W. H., Crumpacker, C. S., and Henry, P. H. (1969). Proc. Nat. Acad. Sci. U.S.63, 1128-1135. Lewis, J. B., Atkins, J. F., Anderson, C. W., Baum, P. R., and Gesteland, R. F. ( 1975). Proc. Nat. Acad. Sd. US.72, 1344-1348. Lindberg, U., Persson, T.,and Philipson, L. (1972). I. Virol. 10, 909-919. Lonberg-Holm, K., and Philipson, L. ( 1969). J. Virol. 4,323338. Lucas, J. J., and Ginsberg, H. S. ( 1971). J. ViroZ. 8,203-213. Lucas, J. J., and Ginsberg, H. S. ( 1972). J . Virol. 10, 1109-1117. Lundholm, U., and Doerfler, W. (1971). Virology 45, 827-829. McGuire, P. M., Swart, C., and Hodge, L. D. (1972). Proc. Nut. Acad. Sci. U.S. 69, 1578-1582. hfaizei, J. V., Jr., Philiips, B. A., and Summers, D. F. ( 1967). Virology 32, 692-699. and Scharff, h4. D. ( 1968a). Virology 36, 115-125. Maizel, J, V., Jr., White, D. 0.. Maizel, J. V., Jr., White, D. O., and Scharff, hl. (1968b). Virology 36, 126-136. Mak, S. ( 1971). J. V i d 7, 426433. Martin, R. G., and Chou, J. Y. (1975). I . Virol. 15,599-612. Morgan, C., Godman, G. C., Breitenfeld, P. M., and Rose, H. M. (1960). J. E r p . Med. 112,373-382. Morgan, C., Rosenkranz, H. S., and Mednis, B. (1969). J. Virol. 4, 777-796. Mdder, C., Arrand, J. R., Delius, H., Keller, W., Pettersson, U., Roberts, R. J., and Sharp, P. A. ( 1974). Cold Spring Harbor Symp. Quant. Biol. 39, 397-400. Seurath, A. R., Rubin, B. A., and Stasny, J. T. (1968). 1. Virol. 2, 1086-1095. Norrby, E. (1968). Curr. T o p . Microbiol. lmmunol. 43, 1-43. Norrby, E. (1969). J. Gen. Virol. 5, 221-236. Norrby, E., and Skaaret, P. ( 1967). Virology 32, 489502. Okubo, C. K., and Raskas, H. J. ( 1971 ). Virology 46, 175-182. Osborn, hl., and Weber, K. (1975). 1. Virol. 15, 636-644. Ozer, H. L. (1972). 1. Virol. 9,41-51. Ozer, H.L., and Tegtmeyer, P. (1972). 1. Virol. 9 , 5 2 4 0 . Parsons, J. T., and Green, M. (1971). Virology 45, 154-162. Pearson, G . D., and Hanawalt, P. C. ( 1971). J. Mol. Biol. 62, 65-80. Pereira, H. G. (1958). Virology 6, 601-611. Pereira, H. C. ( 1960). Nature (London) 186, 571-572. Pettersson, U. (1973). 1. Mol. Biol. 81,521527. Pettersson, U.,and Hoglund, S. ( 1969). Virdogy 39,90-106. Philipson, L., and Lindberg, U. (1974). Compr. Virol. 3, 143-227. Philipson, L.,and Pettersson, U. (1973). Progr. Exp. Tumor Res. 18, 155. Philipson, L., Lonberg-Holm, K., and Pettersson, U. (1968). J. Virol. 2, 1064-1075.

GENETICS OF ADENOVIRUSES

129

Philipson, L., Wall, R.,Glickman, G., and Darnell, J. E. (1971). Proc. Nut. Acad. Sci. US.68, 2806-2809. Philipson, L., Lindberg, U., Person, T., and Vennstrom, B. (1973). Adoan. Biosci. 11, 167-183. Phillips, B. A., Summers, D. F., and Maizel, J. V., Jr. (1968). Virology 35, 216226. Piiia, M., and Green, M. (1965). Proc. Nut. Acad. Sci. US. 54, 547-551. Piiia, M., and Green, M. (1969). Virology 38, 573-586. Pope, J. H., and Rowe, W. P. (1964). J. Erp. Med. 120,577-588. Prage, L., and Pettersson, U. ( 1971) . Virology 45, 364-373. Price, R.,and Penman, S. (1972). J .ViroZ. 9, 621-626. Robb, J. A., and Martin, R. G. ( 1972). 1. Virol. 9,956-968. Robinson, A. J., Younghusband, H. B., and Bellett, A. J. D. (1973). Virology 56, 54-69. Rosen, L. ( 1960). Amer. J. H y g . 71, 120-128. Rosenwirth, B., Shiroki, K., Levine, A. J., and Shimojo, H. (1975). Submitted for publication. Rouse, H. C., and Schlesinger, R. W. ( 1972). Virology 48,463-471. Rubenstein, F . E., and Ginsberg, H. S. (1974). Interoirology 3, 170-174. Russell, W. C., and Knight, B. ( 1967). J. Gen. Virol. 1, 523-528. Russell, W. C., McIntosh, K., and Skehel, J. J. (1971). J. Gen. Virol. 11, 35-46. Russell, W. C., Newman, C., and Williams, J. F. (1972a). J. Gen. Virol. 17, 265-279. Russell, W. C., Skehel, J. J., Machado, R., and Pereira, H. G. (1972b). Virology 50, 931-934. Russell, W. C., Skehel, J. J., and Williams, J. F. (1974). J. Gen. Virol. 24, 247-259. Sambrook, J., Botchan, M., Gallimore, P., Ozanne, B., Pettersson, U., Williams, J., and Sharp, P. (1974). Cold Spring Harbor Symp. Quunt. Bwl. 39, 615-682. Schlesinger, R. W. (1969). Aduan. Virus Res. 14, 1-61. Shiroki, K., and Shimojo, H. (1974). Virology 61, 474-485. Shiroki, K., Irisawa, J., and Shimojo, H. (1972). Virology 49, 1-11. Shiroki, K., Shimojo, H., and Yamaguchi, K. (1974). Virology 60,192-199. Simon, L. (1972). Proc. Nut. Acad. Sci. US.69, 907-911. Stinski, M. F., and Ginsberg, H. S. (1975). J. Virol. 15, 898-90s. Sundquist, B., Everitt, E., Philipson, L., and Haglund, S. (1973a). J. Virol. 11, 449-459. Sundquist, B., Pettersson, U., Thelander, L., and Philipson, L. ( 1973b). Virology 51, 252-256. Sussenbach, J. S. (1967). Virology 33, 567-574. Sussenbach, J. S., van der Vliet, P. C., Ellens, D. J., and Jansz, H. S. (1972). Nature (London),New Biol. 239, 47-49. Sussenbach, J. S., Ellens, D. J., and Jansz, H. S. (1973). J. Virol. 12, 1131-1138. Suzuki, E., and Shimojo, H. (1971). Virology 43, 488-494. Suzuki, E., and Shimojo, H. (1974). J . Virol. 13, 538-540. Suzuki, E., Shimojo, H., and Moritsugu, Y. (1972). Virology 49, 426-438. Takahashi, M. ( 1972). Virology 49, 815-817. Takahashi, M., Minekawa, Y., and Yamanishi, K. (1974). Virology 57,300-303. Takemori, N . (1972). Virology 47, 157-167. Takemori, N., Riggs, J. L., and Aldrich, C. (1968). Virology 36, 575-586. Takemori, N., Riggs, J. L., and Aldrich, C. D. (1969).Virdogy 38, 8-15. Tegtmeyer, P. (1975). J. Virol. 15, 613-618. Thomas, D. C., and Green, M. ( 1966). Proc. Not. Acad. Sci. U.S. 56,243-246.

130

HAROLD S. CINSBERG AND C. S. H. YOUNG

Thomas, D. C., and Green, M. (1969). Virology 39, 205-210. Tibbetts, C., and Pettersson, U. (1974). J. Mof. Biol. 88, 767-784. Tibbetts, C., Pettersson, U., Johansson, K., and Philipson, L. (.1974). J. Virol. 13, 370-377. Tooze, J., ed. (1973). “The Molecular Biology of Tumour Viruses.” Cold Spring Harbor Lab., Cold Spring Harbor, New York. Ustacelebi, S. (1973). Ph.D. Thesis, University of Glasgow, Scotland. Ustacelebi, S., and Williams, J. F. (1972). Nature (London) 235, 5253. Valentine, R. C., and Pereira, H. G. (1965). J. Mol. Biol. 13, 13-20. van der Eb, A. J. (1973). Virology 51, 11-23. van der Eb, A. J., van Kesteren, L. W., and van Bruggen, E. F. J. (1969). Biochim. Biophys. Acta 182, 530-541. van der Vliet, P. C., and Levine, A. J. (1973). Nature (London), New Biol. 246, 170-174. van der Vliet, P. C., and Sussenbach, J. S. (1972). Etrr. J. Biochem. 30, 584-592. van der Vliet, P. C., Levine, A. J., Ensinger, M. J., and Ginsberg, H. S . (1975). 1. Virol. 15, 348-354. L’elicer, L., and Ginsberg, H. S. (1968). Proc. Nut. Acad. Sci. U S . 61, 1264-1271. Velicer, L., and Ginsberg, H. S. ( 1970). J. Virol. 5, 338-352. Wall, R., Philipson, L., and Darnell, J. E. ( 1972). Virology 50,2734. Weber, J., BCgin, M., and Khittoo, G. (1975).J. Virol. 15, 1049-1056. Scharff, M. D., and Maize], J. \’., Jr. ( 1969). Virology 38, 395-406. White, D. 0.. Wilcox, W. C., and Ginsberg, H. S. ( 1961). Proc. Nut. Acad. Sci. U.S. 47, 512-526. Wilcox, W. C., and Ginsberg, H. S. (1963a). Proc. SOC. Exp. Biol. Med. 114,3742. Wilcox, W. C., and Ginsberg, H. S. ( 1963b). Virology 20,269-280. Wilcox, W. C., and Ginsberg, H. S. ( 1 9 6 3 ~ ) J. . Erp. Med. 118, 295-306. Wilcox, W. C., Ginsberg, H. S., and Anderson, T. F. (1963). J. Exp. Med. 118, 307-3 14. Wilhelm, J. M., and Ginsberg, H. S. ( 1972). J. Virol. 9,973-980. Wilkie, N. M., Ustacelebi, S., and Williams, J. F. ( 1973). Virology 51, 499-503. Williams, J. F. ( 1973). Nature (London) 243, 162-163. Williams, J. F., and Ustacelebi, S. (1971a). J. Gen. Virol. 13, 345-348. Williams, J. F., and Ustacelebi, S. (1971b). Strategy Viral Genome, Ciba Found. Symp. pp. 275-290. Williams, J. F., Gharpure, M., Ustacelebi, S., and McDonald, S. (1971). J. Gen. Virol. 11, 95-101. Williams, J. F., Young, C. S. H., and Austin, P. E. (1974). Cold Spring Harbor Symp. Quant. Biol. 39, 427-437. Williams, J. F., Grodzicker, T., Sharp, P., and Sambrook, J. (1975a). Cell 4, 113-119. Williams, J. F., Young, C. S., and Austin, P. E. (1975b). J. Virol. 15, 675-678. Wolfson, J., and Dressler, D. (1972). Proc. Nat. Acad. Sci. U S . 69, 30543057. Yamashita, T., and Green, M. (1974). 1. Virol. 14, 412-420. Young, C. S. H., and Williams, J. F. (1975).J. Virol. 15, 1168-1175.

MOLECULAR BIOLOGY OF THE CARCINOGEN. 4-NITROQUINOLINE 1 -OXIDE Minako Nagao and Takashi Sugirnura

.

National Cancer Center Research Institute. Chu0.b. and Institute of Medical Science University of Tokyo. Minato.ku. Tokyo. Japan

I . Introduction . . . . . . . . . . . . . I1. Mutagenic Activity of 4-Nitroquinoline I-Oxide on Organisms . . . A . Historical Aspects of Mutations by 4-Nitroquinoline 1-Oxide . B. Base-Pair Change Mutations . . . . . . . . . . . . . . . . . . . C. Frameshift Mutations D . Deletion Mutations . . . . . . . . . . . E . Mitotic Gene Conversions . . . . . . . . . . F. Loss of the Cytoplasmic p Factor in Yeast . . . . . . G. Mutagenic Activities of 4-Nitroquinoline 1-Oxide and Related . . . . . . . . . . . . Compounds H. Phage Induction from Lysogenic Bacteria . . . . . . I11. Chromosome Aberrations . . . . . . . . . . A . Chromosome Aberrations in Cells with Normal Repair Function . . B. Chromosome Aberrations in a Repair-Deficient Strain of Human Cells . . . . . . . C. Endoreduplication of Chromosomes . IV. Repair of 4-Nitroquinoline 1-Oxide-Damaged DNA . . . . . A . Repair in Bacteria . . . . . . . . . . . B . Repair in Yeast . . . . . . . . . . . . C. Repair in Bacteriophages . . . . . . . . . . D. Repair in Plant Cells . . . . . . . . . . . E. Repair in Mammalian Cells . . . . . . . . . V. Interaction of 4-Nitroquinoline I-Oxide and Its Derivatives with Nucleic . . . . . . . . . . . . . . Acids . A . In Viuo Formation of Quinoline-Base Adducts . . . . . B . Enzymic Activation Steps Required for Modification of DNA by . . . . . . . . . 4-Nitroquinoline I-Oxide . C. Biological Activity of Modified DNA . . . . . . . D. Chemical Structure of Quinoline-Purine Adducts . . . . . E. Chemical Interaction of 4-Nitroquinoline 1-Oxide and Its Deriva. . . . . . . . . . . tives with DNA . VI . Interaction of 4-Nitroquinoline 1-Oxide and Its Derivatives with Protein VII. Recent Information on Carcinogenesis by 4-Nitroquinoline 1-Oxide . . A . I n Viuo Carcinogenesis . . . . . . . . . . B . Immunity . . . . . . . . . . . . . C. I n Vitro Carcinogenesis by, and Effect of, 4-Nitroquinoline 1-Oxide on in Viuo Viral Carcinogenesis . . . . . . . . D. Decarcinogenesis by 4-Nitroquinoline 1-Oxide . . . . . E. Other Biologic Effects of 4-Nitroquinoline I-Oxide . . . .

.

131

132 133 133 133 136 136 137 137

138 138 140 140 140 141 143 143 145 146 146 146

151 151 152 152

153 155 156 157 157 160 160 162 162

132

MINAKO NAGAO AND TAKASHl SUGIMUFlA

VIII. 4-Nitroquinoline l-Oxide and Microbial Screening Method for Carcinogen 163 IX. Conclusions . . . . . . . . . . . . . 163 References . . . . . . . . . . . . . 164

I. Introduction

Studies on chemical carcinogenesis have three main purposes. One is to elucidate carcinogenic mechanisms, another is to prevent possible carcinogenic hazard due to chemicals in our environment, and the third is to provide clinical model systems in experimental animals. The carcinogenicity of 4-nitroquinoline 1-oxide (4NQO) was first reported by Nakahara et al. (1957). However, the mutagenicity of 4NQ0 was reported even earlier, in 1955, by Okabayashi. Mutagenicity by a carcinogen was observed, among others, by Latarjet et al. (1949). Recent studies on chemical carcinogens and mutagens have shown that with a few exceptions carcinogens are mutagenic and mutagens are carcinogenic. 4NQ0 is an example of a compound that has long been known to be both carcinogenic and mutagenic. Many geneticists have used 4NQ0 in experiments on biological materials ranging from bacteriophages to mammals. In many studies, 4NQ0 has also been used as a positive control substance in mutagenic and carcinogenic experiments. 4NQ0 has been used in studies on cultured cells in vitro as well as on in vivo carcinogenesis. Very interesting results have been reported on its induction of certain types of gastric carcinomas, pancreatic tumors, and lung tumors that mimic cases of human tumors. The molecular mechanisms of the interaction of metabolites of 4NQ0 with nucleic acids have been fairly well clarified. Moreover genetic studies on the effect of 4NQ0 are more advanced than those on many other carcinogens. Owing to the relation between carcinogens and mutagens, it seems worthwhile to review recent advances in studies on 4NQO. A monographic review of studies on this compound entitled “Chemistry and Biological Actions of 4-Nitroquinoline l-Oxide” by Endo et al. (1971) has been published. Therefore, the present review principally aims to cover more recent reports on this compound. One advantage of using 4NQ0 as a carcinogen is that metabolic activation systems for it are commonly present in microbial and mammalian cells. Thus it differs from most typical carcinogens, like acetylaminofluorene, 3’-methyl-Ndimethyl-4-aminoazobenzene, and many aromatic hydrocarbons, which are metabolically activated only by mammalian microsomal enzymes, but not by any microbial enzyme. This has made possible many experiments using 4NQ0 on microbes. Only a few other compounds like 4NQ0

MOLECULAR BIOLOGY OF THE CARCINOGEN,

4NQO

133

are known. These include nitrofuran derivatives, which are now under intensive study. We hope that this review on 4NQ0 will be of general use for studies on chemical carcinogenesis. II. Mutagenic Activity of 4-Nitroquinoline 1 -Oxide on Organisms

A. HISTORICAL ASPEW

OF

MUTATIONSBY 4-NITROQUINOLINE 1-OXIDE

Okabayashi was the first to observe that 4NQ0 had biological activity, finding first antifungal activity ( Okabayashi, 1953) and subsequently mutagenic activities on fungi ( Okabayashi, 1955). 4NQ0 was found to cause mutation with regard to the pattern of colony growth (Okabayashi, 1955) and to induce biochemical mutants (Yamagata et al., 1956) of Aspergillus niger. It also caused reversion of amino acid-requiring auxotrophs of Streptomyces griseo@vus to their prototrophs ( Mashima and Ikeda, 1958). Okabayashi and Yoshimoto (1962) proposed that microorganisms, such as Escherichiu coli, Candidu utilis, Aspergillus niger, Brevibacterium liquefaciens, and Pseudomonas aeruginosa, reduced 4NQ0 metabolically to 4-hydroxyaminoquinoline l-oxide (4HAQO) and that the latter was further reduced to 4-aminoquinoline l-oxide (4AQO). Treatment of Aspergillus niger with the metabolic intermediate 4HAQO also induced morphological mutants and nutrient-requiring auxotrophic mutants ( Okabayashi et al., 1964). Subsequent analyses of the mutagenic actions of 4NQ0 on various kinds of microbes and their mutants have provided a clearer understanding of the mechanisms of the mutagenic actions of 4NQO.

B. BASE-PAIR CHANGE MUTATIONS Molecular analyses of mutagenesis induced by 4NQ0 were initiated

by Ishizawa and Endo (1970) using three types of rII mutants of coliphage T4. After treatment of Escherichiu coli infected with T4 phage mutants with 4NQ0, they demonstrated reversion patterns of rII mutants, as indicated in Table I. In vivo treatment of T4 wild type phage with 4NQ0 induced remarkable GC + AT transitions, but did not induce AT + GC transitions or cause reversion of frameshift mutants induced by proflavine. I n Vivo treatment of T4 phage with 4NQ0 preferentially yielded rII mutants having AT base pair at mutant sites, and none of them reverted on treatment with 4NQ0 or proflavine (Ishizawa and Endo, 1971). Ishizawa and Endo (1972) also reported the induction of amber suppressor mutants by 4NQ0 and 4HAQO in Escherichia coli B9601 (trpBa,, his,,). These

134

MINAKO NAGAO AND TAKASHI SUGIMURA

TABLE I REVERSION OF THREE TYPESOF rII MUTANTS BY 4NQ0° ~~

~

~~

Reversion index in units, 10-6

rIT mutant

Probable alteration at mutation site

Without 4NQOb (Io)

With 4NQ0 (30 pg/ml) (1)

GC GC GC GC AT

0.08 1.0 0.05 0.16 0.01 0.04 0.2 0.07 0.02 0.37

25.2 292 21.5 33.2 0.06 0.18 0.15 0.2 0.03 0.92

AT AT AT Frameshif t Frameshift a

b

Factor of increase (I/ZO) 316 292 430 207 6

4.5 0.8 2.9 1.5 2.5

Data from Ishizawa and Endo (1970). 4-NQO, 4-nitroquinoline I-oxide.

prototrophic reversions were mainly based on GC + TA transversion (SupD). In vitro treatment of T4 phage with 4NQ0 did not induce any mutation, and metabolic activation of 4 N Q 0 seemed to be essential for its mutagenic activity ( Ishizawa and Endo, 1967). Recently, extensive studies on the mutagenic specificity of 4NQ0 were carried out by Sherman and collaborators. They used various kinds of structural gene mutants of iso-l-cytochrome c of Saccharomyces cereuisiae (Prakash et al., 1974; Prakash and Sherman, 1974). The use of these mutants allowed differentiation of intragenic reversion from extragenic reversion, since intragenic revertants produced large colonies on lactate agar plates, whereas extragenic revertants ( suppressor revertants) produced small colonies. These results are summarized in Table 11. It can be seen from these results that there are three distinct responses to the mutagenic action of 4NQO: high-frequency reversion, low-frequency reversion, and no reversion. The cycl-131 strain, which is an initiator tester strain with valine codon GUG mutation from the initiation codon was reverted with high frequency by GC + AT transition. Three different types of mutants showed a low-frequency response. Type 1 included the initiator mutants, cycl-51, cycl-100, and cycl-181, in which the responses all involved the mutant leucine codon UUG or CUG at the initiation codon. Type 2 was also an initiator mutant cycl-133 in which the

TABLE I1 CODONA N D BASE-PAIRCHANGES ASSOCIATED WITH REVERSION OF cycl TESTER STRAINSA N D THEIRREVERSION FREQUENCIES’S* Reversion frequency mRNA change Initiation mutant

Base-pair change G+ -A C T

GUG-+ AUG

T

MU-

4NQ0

Re-

tant

(1 pg/ml)

sponse

811

High

~ycl-131

A

U

cycl-51

C

C

81 102 CYC~-181 55

Low

cycl-133

67

LOW

‘cycl-13 cycl-74 cycl-85 ,cycl-163

1 0

{cycl-~oo

or

UG -+ AUG

A

6’5; G+ -T -

AGG ---t AUG

C

A

T A

G C

-4-

or C - + -G G C or A+ -G T C

U AUC -+ AUG A

Ochre mutant

A

T

T+A

and A G

amino acid UAA +

T-+ c

codons

and A C

T’G

1 1 0 0

No

0

5 4

All the above, plus G T

Amber mutant

amino acid

E-+X

codons

and G + -C C G

UAG --+

Frameshift mutant

cycl-2 cycl-9 cycl-45 cycl-72 cycl-94 cycl-140 cycl-156

No

2 16

UA ?-+ deletion, etc. AAAA + - base, etc.

LA deletion, AT etc. AAAA + - base TTTT pair, etc. ~

AA TT

+

+

base pair, etc.

CYC~-179 92

cyc 1-31

0

CYC~-183 16 ~y~l-239

LOW

No

1

Data from Prakash et al. (1974) and Prakaah and Shermann (1974). Revertants per lo7 survivors after treatment with 4-nitroquinoline 1-oxide (4NQ0).

136

MINAKO NAGAO AND TAKASHI SUGIMURA

response involved arginine codon AGG mutated from the initiation codon. Type 3 was an amber mutant which had an UAG codon and might be reverted by any kind of base change. GC +JTA transversion was considered to be the molecular mechanism of these low-frequency responses. Mutants showing no reversion included initiator mutants, which contained the isoleucine codon, AUU, AUC, or AUA, replacing the initiator codon, and the ochre (UAA) mutant. Three types of frameshift mutants containing an AT base pair also showed no reversion. Thus a G - C base pair at the mutation site is a common character of all strains which showed efficient reversion on treatment with 4NQO. Studies using mutated codons with completely known nucleotide sequences clearly supported this conclusion. It could not be decided from these experiments whether 4 N Q 0 induced GC + CG transversion, but it certainly induced GC + TA transversion. Other mutagens besides 4NQ0 also preferentially induce GC + AT transition. These are ethylmethane sulfonate, diethyl sulfate, N-methyl-N’-nitro-N-nitrosoguanidine(MNNG), nitrosoimidazolidone, nitrous acid, and p-propiolactone ( Prakash and Sherman, 1973). However, these chemicals did not produce significant GC transversion. Thus 4 N Q 0 is a unique mutagen in this respect.

C. FRAMESHIFT MUTATIONS

4NQ0 did not induce a frameshift revertant of Saccharomyces cerevisiae, as previously mentioned (Table 11). However, Hartman et al.

(1971) and Ames et al. (1973) demonstrated that it induced frameshift mutations in Salmonella typhimurium. Three histidine requiring frameshift tester strains were examined: TA1536 with an unknown lesion, 1 base pair insertion and TA1538 with TA1537 with :ggEgr :$$?Egg:+ 2. Of these, only TA1538 was reverted by 4NQO. TA1538 also showed a high rate of reversion with 2-nitrosofluorene (Ames et al., 1973). The failure of 4NQ0 to induce reversed frameshift mutants of Saccharonyces cerevisiae could be due to the different base compositions of the frameshift mutants that were examined.

+

D. DELETION MUTATIONS

4NQ0 induces deletion mutations, as does UV-light. Yamamoto and Ishii (1974) examined deletions of chromosomes of Eschen’chia coli including the colB (sensitivity to colicin B ) , tonB (receptor for phage T1) and trp genes. Deletion from the coZB to tonB locus was “short” and deletion from the colB to trp locus was “long.” When E. coli B

MOLECULAR BIOLOGY OF THE CARCINOGEN,

4NQO

137

( H / r 30 argFam) and its repair-deficient derivatives, uurA- (with a defect in excision repair), poEA- (with a defect in DNA polymerase I ) , and recA- (with a defect in the recoinbinational mechanism) were treated with 4NQ0, the induced deletion mutants were “short” type, as on UV treatment. On the other hand, in spontaneous deletions the numbers of “short” and “long” deletions were similar. Deletion mutation of the uurA- strain was about 30 times higher than that of the parental strain. Like UV-light 4NQ0 did not induce deletion mutation of the recA- strain. Thus the modes of action of 4NQ0 and UV-light are quite similar.

E. Mrromc GENECONVEFSIONS In S . cerevisiae, mitotic gene conversions are induced by various agents including UV light (Roman and Jacob, 1957), alkylating agents (Zimmermann, 1971) , and acridine ( Fahring, 1970). Zimmermann and Schwaier (1967) reported that nitrite, MNNG, and other alkylating agents induced 103-foldmore mitotic gene conversions than reverse mutations. The D,, D4, and D, strains, which were established by Zimmermann, have frequently been used in gene conversion studies. Fahring (1973) reported that 4NQ0 induced gene conversion in high frequency in the D4 strain, which is diploid and heteroallelic at the a&, and frp5 loci. After treatment with 2 p M 4NQ0 for 4 hours, 45 convertants at each of these loci were found per 5 X los survivors. The D3 strain is heteroallelic at the a&, locus, and it also showed mitotic gene conversions on treatment with 4NQ0 (V. Simmon, Stanford Research Institute, personal communication). The genetic mechanism of mitotic gene conversion was relatively well elucidated by Mortimer and Manney (1971), but the molecular events involved are not completely understood. It is generally accepted that mitotic gene conversion is an indicator of repair of damaged DNA.

F. Loss

OF THE

CYTOPLASMIC p FACTOR IN YEAST

The formation of the respiratory system of yeast is controlled by both a nuclear gene and a cytoplasmic self-duplicating genetic factor p. A respiration-deficient ( RD ) mutant could be induced by various chemical and physical agents. Agents like acriflavine (Slonimski and Ephrussi, 1949), ethidium bromide (Slonimski et al., 1968), dimethyl sulfoxide (Yee et al., 1972), pinacyanol (Sugimura et al., 1969), and 5-fluorouraciI (Moustacchi and Marcovich, 1963) were found to induce the cytoplasmic

138

MIN'AKO NAGAO AND TAKASHI SUGIMURA

RD mutant with extremely high efficiency, especially from growing yeast. Nitrous acid induced exclusively the nuclear RD mutant (Schwaier et al., 1968) , and UV light, nitrosoimidazolidone, and nitrosomethylurethane produced both nuclear and cytoplasmic RD mutants. According to Mifuchi et aL ( 1963), RD mutants from a diploid strain of S. cereuisiae, Hansen 0209,induced by 4NQ0 consisted of cytoplasmic and/or nuclear RDs. These nuclear RD mutants showed the spectra of various types of cytochromes (Morita and Mifuchi, 1970). However, Nagai (1969) reported that 4NQ0 specifically induced cytoplasmic RD mutants with high efficiency. Epstein and St. Pierre (1969) also reported that 4NQ0 induced RD mutants of S. cerevisiae, but they did not identify these as cytoplasmic or nuclear. In synchronized yeast cultures, the frequency of induction of RD mutants by 4NQ0 was closely correlated with mitochondriogenesis (Morita and Mifuchi, 1974). A cytoplasmic RD, the N-1 mutant, which was derived from the Hansen 0209 strain by 4NQ0, showed deletion of mitochondria1 DNA (Morita and Mifuchi, 1974).

Epstein and St. Pierre (1969) investigated the induction of RD mutants of S. cerevisiae by compounds related to 4NQO and 4-nitropyridine l-oxide (4NPO). They found that the concentrations of chemicals that were required to induce RD mutants were low with carcinogenic compounds but high with noncarcinogenic compounds. Thus there seems to be a correlation between mutagenicity and carcinogenicity. Many substances showed simultaneous mutagenic, carcinogenic, and photodynamic activities as well as activities for free-radical production and induction of phages from lysogenic bacteria (Table 111).

H. PHAGEINDUCTIONFROM LYSOGENIC BACTERIA Phage induction by a carcinogen was first observed by Lwoff ( 1953). Endo et al. reported that 4NQ0 induced the lysogenic h phage of E. coEi ( 1963).Treatment of lysogenic Salmonella typhirnurium with 4NQ0 or 4HAQO resulted in the induction of phage P22 (N. Yamamoto et al., 1970). The induction of phage by these compounds was far greater from hcr- lysogen than from the wild-type lysogen, and no phage-induction from the rec- lysogen was observed. This phage-induction pattern of 4NQ0 was also very similar to that of UV light (Yamamoto, 1969).

TABLE I11 4-NITROQUINOLINE I-OXIDE (4NQO) DERIVATIVES: MUTAGENICITY O N Saccharomyces cerevisiae; CARCINOGENICITY, PHOTODYNAMIC ACTIVITY,FREE-RADICAL FORMATION, A N D PHAGE INDUCTION FROM LYSOGENIC BACTERIA"

Compound*

Conc. producing 50% growth inhibition (GO) (pglml)

Conc. causing double the spontaneous mutation rate (2SMR) Gg/ml)

Mutagenic potency (2SMR-1)

4NQ0 8-CHr4NQO 7-CH3-4N QO 2-CHs4NQ0 6-CHr4NQO 7-C1-4NQ0 4NQ 4HAQO.HCl 7-CI-4HAQO.HC1 4 HAPO.HC1 3-CH3-4NQ0 4NP0 3NQ0 5NQ0 6-NOr4HAQO.HCI

0.07 0.06 0.12 0.07 0.23 0.48 14 26 > 100 > 100 > 100 > 100 > 100 > 100 >200

0.06 0.10 0.13 0.27 0.31 1.5 9.0 26 > 100 > 100 > 100 > 100 >100 > 100 > 100

16.7 10.0 7.69 3.70 3.23 0.67 0.11 0.04 < 0.01 < 0.01 < 0.01 < 0.01 < 0.01 < 0.01 < 0.005

x Carcinogenicity

+ + ++ ++ ++ + + +d

-

+

Photodynamic activity 11.1 2.4 7.1 2.9 1.3 1.1 1.7 < 0.1

<